THE MILLENNIUM DEVELOPMENT GOALS FOR HEALTH RISING TO THE CHALLENGES The Millennium Development Goals for Health RISING TO THE CHALLENGES The Millennium Development Goals for Health RISING TO THE CHALLENGES by Adam Wagstaff and Mariam Claeson THE WORLD BANK © 2004The International Bank for Reconstruction and Development /TheWorld Bank 1818 H Street, NW Washington, DC 20433 Telephone 202-473-1000 Internet www.worldbank.org E-mail feedback@worldbank.org All rights reserved. 1 2 3 4 07 06 05 04 The findings, interpretations, and conclusions expressed herein are those of the author(s) and do not necessarily reflect the views of the Board of Executive Directors of theWorld Bank or the governments they represent. TheWorld Bank does not guarantee the accuracy of the data included in this work.The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of theWorld Bank concerning the legal status of any territory or the endorse- ment or acceptance of such boundaries. Rights and Permissions The material in this work is copyrighted. Copying and/or transmitting por- tions or all of this work without permission may be a violation of applicable law.TheWorld Bank encourages dissemination of its work and will normally grant permission promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher,World Bank, 1818 H Street NW,Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. ISBN 0-8213-5767-0 Library of Congress Cataloging-in-Publication Data The millenium development goals for health : rising to the challenges /TheWorld Bank. p. cm. Includes bibliographical references and index. ISBN 0-8213-5767-0 1. Health promotion. 2. Health promotion--Cross-cultural studies. 3. Medical policy. 4. Medical policy--Cross-cultural studies. 5. Health planning. 6. Health Planning--Cross-cultural studies. I. World Bank. RA427.8.M556 2004 362.1--dc22 2004047781 Cover photo from theWorld Bank Photo Library Cover design by Chris Lester of Rock Creek Creative CONTENTS ix Foreword CHAPTER 5 x Acknowledgments 69 Households--Key but Underrated Actors in the xi Key Messages and Policy Highlights Health Sector xiii Statistical Highlights CHAPTER 6 xv Abbreviations 93 Improving Service Delivery 1 Overview: Rising to the Challenges CHAPTER 7 111 Tackling Human Resource and Pharmaceutical PART I Constraints The Millennium Development Goal Challenges CHAPTER 8 131 Strengthening Core Public Health Functions CHAPTER 1 CHAPTER 9 25 Backdrop to the Millennium Development Goals 145 Financing Additional Spending for the CHAPTER 2 Millennium Development Goals--In a 31 The Millennium Development Goals for Health: SustainableWay Progress and Prospects CHAPTER 10 CHAPTER 3 155 Applying the Lessons of Development Assistance 47 Effective Interventions Exist--They Need to for Health Reach More People Appendixes PART II APPENDIX A Rising to the Challenges 169 Data and Methods CHAPTER 4 APPENDIX B 55 Extra Government Health Spending Is 175 WhyTracking Progress toward the Health Necessary but Not Sufficient--Health Sector Goals Isn't Easy Strengthening Is Also Required, and Spending Needs to Be BetterTargeted 179 Index v Boxes 118 Box 7.7 The problem of counterfeit drugs 2 Box 1 The health-related Millennium 119 Box 7.8 Do we know how affordable drugs Development Goals really are? 27 Box 1.1 Worlds apart:The poor die earlier 119 Box 7.9 High drug costs inVietnam deter use 29 Box 1.2 The promising (and challenging) array and cause impoverishment of health-related partnerships 121 Box 7.10 Perverse provider incentives and 43 Box 2.1 Why the decline in maternal mortality tuberculosis drugs in Georgia is slowing 123 Box 7.11 Drug subsidies and drug insurance: 53 Box 3.1 Low coverage of HIV/AIDS Different countries, different policies interventions 124 Box 7.12 The attractions of pooled 62 Box 4.1 Coupling targeting with institutional procurement innovation through social investment funds 125 Box 7.13 The Accelerating Access Initiative for 63 Box 4.2 Marginal budgeting for bottlenecks antiretroviral drugs 71 Box 5.1 Key family practices for the production 126 Box 7.14 Global public-private partnerships to of child health and nutrition accelerate the introduction of new vaccines 72 Box 5.2 Why people can go hungry when food 134 Box 8.1 Increasing the supply of and demand is plentiful for insecticide-treated bednets 72 Box 5.3 Social solidarity in Côte d'Ivoire 135 Box 8.2 Reducing communicable diseases through disease surveillance in Brazil 74 Box 5.4 Blanket subsidies often benefit the better off most 135 Box 8.3 Monitoring progress toward the Millennium Development Goals in the 75 Box 5.5 Vouchers for sex workers in Nicaragua Dominican Republic 76 Box 5.6 Increasing coverage of key 136 Box 8.4 Core public health functions and the interventions through demand-side incentives case for public management 77 Box 5.7 Helping poor women protect 138 Box 8.5 Strengthening core public health themselves--India's SEWA functions to combat HIV/AIDS 79 Box 5.8 Working with the private sector to 148 Box 9.1 Raising tax levels in the developing improve hygiene behaviors world--hard but doable 80 Box 5.9 Radio dramas to promote 157 Box 10.1 The importance of country contraception commitment: How Bangladesh andThailand 81 Box 5.10 Argentina's water privatization have fared in reducing malnutrition program saved young lives 158 Box 10.2 Helping eradicate polio through IDA 94 Box 6.1 The changing mix of cure and care: credit buy-downs Who treats what--and where? 162 Box 10.3 Improving performance of the health 95 Box 6.2 Who delivers which care for the sector in Guinea Millennium Development Goals for health? 162 Box 10.4 Increasing health sector spending and 97 Box 6.3 Different management styles, different immunization coverage by reforming the budget countries--but just miles apart process in Mauritania 99 Box 6.4 Management in India'sTamil Nadu 164 Box 10.5 Harmonizing, monitoring, and Integrated Nutrition Program evaluating HIV/AIDS programs 100 Box 6.5 Innovative and effective management 165 Box 10.6 Key actions to accelerate progress on of public health workers in Ceara, Brazil the health, nutrition, and population 104 Box 6.6 Contracting health services in Millennium Goals:The Ottawa consensus Cambodia, Guatemala, and Pakistan 166 Box 10.7 Summary of recommendations for 113 Box 7.1 Ghana's loss of health sector workers action from the High-Level Forum on the Millennium Development Goals for Health held 114 Box 7.2 The devastating impact of HIV/AIDS January 8­9, 2004 on the health workforce 115 Box 7.3 Competing with the private sector in Bolivia Figures 116 Box 7.4 What do health workers in India want 2 Figure 1 The poorest countries suffer the most? highest burdens of premature mortality and 117 Box 7.5 Lack of drugs threatens the Millennium malnutrition Development Goals 5 Figure 2 Faster economic growth and other 118 Box 7.6 Inappropriate or "irrational" drug use changes outside the health sector will help move vi| The Millennium Development Goals for Health regions toward the targets, but in most cases it 48 Figure 3.1 Why children die will not get them there 50 Figure 3.2 The arsenal of effective interventions 6 Figure 3 Full use of existing interventions would against childhood killers reduce maternal deaths dramatically 51 Figure 3.3 Interventions for reducing maternal 8 Figure 4 Paths to better health, nutrition, and mortality population outcomes 52 Figure 3.4 Low-income countries lag behind on key preventive interventions for maternal and 11 Figure 5 What makes service providers child health accountable 52 Figure 3.5 Use of key preventive interventions 14 Figure 6 What health workers in Andhra for maternal and child health is lower in some Pradesh want from their jobs--and whether regions than others they get it 52 Figure 3.6 Some children with acute respiratory 35 Figure 2.1 Progress on malnutrition, under-five infections receive treatment--others don't mortality, and maternal mortality, by region and 53 Figure 3.7 Proportion of children age 12­23 mos income who received full basic immunization coverage-- 35 Figure 2.2 People on track to hit malnutrition, poorest 20 percent vs.population as a whole under-five mortality, and maternal mortality 54 Figure 3.8 Use trends for household-delivered and targets professionally delivered interventions 36 Figure 2.3 Countries on track to hit 54 Figure 3.9 Full use of existing interventions malnutrition, under-five mortality, maternal would dramatically cut child deaths mortality targets 54 Figure 3.10 Full use of existing interventions 36 Figure 2.4 Some countries are reducing would dramatically cut maternal deaths malnutrition quickly. In others malnutrition has 57 Figure 4.1 CPIA scores across Bank regions fallen less slowly, and in some it has increased. 58 Figure 4.2 The contributions of faster growth of 37 Figure 2.5 How the poor have fared on government health spending in countries with malnutrition reductions--absolutely and good policies and institutions compared with the less poor 65 Figure 4.3 Paths to better health, nutrition, and 37 Figure 2.6 How the poor have fared on under- population outcomes five mortality reductions--absolutely and 74 Figure 5.1 Who gets subsidies? compared with the less poor 74 Figure 5.2 Under Bolivia's National Maternal 38 Figure 2.7 Child mortality--the pace of decline and Child Insurance Program key maternal is too slow, and in the developing world it's health interventions rose fastest among the poor getting even slower 77 Figure 5.3 Poor women have less of a say in spending their own money 38 Figure 2.8 Child mortality--percentage of countries with a faster rate of decline in the 78 Figure 5.4 Fewer poor women have completed 1990s than in the 1980s the fifth grade 112 Figure 7.1 Doctors across the world in the 39 Figure 2.9 Trends in HIV prevalence among 1990s--how many and how much change pregnant women in Uganda,1990­2001 112 Figure 7.2 Doctors across Sub-Saharan Africa in 39 Figure 2.10 Tuberculosis DOTS detection rates, the 1990s--how many and how much change selected countries in Sub-Saharan Africa, 114 Figure 7.3 Stocks and flows in human resources 1995­2001 146 Figure 9.1 Government health spending is 39 Figure 2.11 Tuberculosis DOTS cure rates, higher in richer countries, but private health selected countries in Sub-Saharan Africa, spending is higher in low-income countries 1994­2000 147 Figure 9.2 Some countries spend considerably 39 Figure 2.12 Past performance is not necessarily a less than expected on government health good predictor of future performance: under- programs--and some spend more five mortality 149 Figure 9.3 Health absorbs a higher share of 40 Figure 2.13 Growth rates of per capita income: government spending,and general revenues actual 1990­2000 and forecasts for 2001­15 absorb a higher share of GDP in richer countries 41 Figure 2.14 Girls catching up with boys at 150 Figure 9.4 Small health shares and low secondary school level--growth needed in the government revenues cause some low-income share of female population over age 15 who countries to spend less than they can afford to on have completed secondary education to health eliminate the secondary education gender gap 151 Figure 9.5 General revenues as a share of GDP 41 Figure 2.15 In all but two regions access to rose significantly during the 1990s in some drinking water needs to grow faster to achieve countries and fell in others the water target Rising to the Challenges |vii 157 Figure 9.6 The share of GDP going to 59 Table 4.2 How far would the share of GDP government health spending during the 1990s devoted to government health spending need to rose significantly in some countries and fell in rise in well-governed countries to meet the others Millennium Development Goals? 152 Figure 9.7 Importance of external financing in 64 Table 4.3 Marginal budgeting for bottlenecks-- total health expenditures in selected countries in how targeting can raise government health Sub-Saharan Africa spending elasticities 156 Figure 10.1 Official development assistance and 86 Table 5.1 Reducing barriers facing households other flows for health are rising in the use of effective child health interventions 89 Table 5.2 Reducing barriers to the use of effective maternal health interventions by Tables households 3 Table 1 Progress toward selected health 91 Table 5.3 Reducing barriers to the use of Millennium Development Goals (percent) effective interventions for malaria, HIV/AIDS, 33 Table 2.1 Goals, targets, and indicators--many and tuberculosis by households of them for health 133 Table 8.1 Public health responsibilities and 42 Table 2.2 Economic growth and attainment of functions the gender and water targets will push countries 137 Table 8.2 Potential for intersectoral synergies to toward the health targets but leave many a long achieve the Millennium Development Goals for way from them health and nutrition 48 Table 3.1 Effective interventions for reducing 140 Table 8.3 Examples of public health functions illness, deaths, and malnutrition and infrastructure requirements for preventing 57 Table 4.1 Elasticities of Millennium and controlling communicable diseases Development Goal outcomes to government 152 Table 9.1 Some uncomfortable arithmetic-- health spending, as a percent of GDP, by quality how long to raise domestic resources for health? of policies and institutions viii| The Millennium Development Goals for Health Foreword The extent of premature death and ill health in the devel- of core public health functions:monitoring,evaluation,sur- oping world is staggering.In 2000 almost 11 million chil- veillance,regulation,social mobilization,and concerted dren died before their fifth birthday,an estimated 140 action beyond the health sector. million children under five are underweight,3 million died The report identifies what developing country from HIV/AIDS,tuberculosis claimed another 2 million governments can do to accelerate the pace of progress lives,and 515,000 women died during pregnancy or child while ensuring that benefits accrue to the poorest and birth in 1995,almost all of them in the developing world. most disadvantaged households. It also pulls together the Death and ill health on such a scale are matters of con- lessons of development assistance and country initiatives cern in their own right.They are also a brake on economic and innovations to improve the effectiveness of aid, based development.It was these twin concerns that led the inter- on a number of country case studies. It highlights some of national community to put health firmly at the center of the principles of effective development assistance: country the Millennium Development Goals when adopting them driven coordination; strategic coherence expressed in at the Millennium Summit in September 2000. comprehensive poverty reduction strategies, which fully This report focuses on the health and nutrition address the issues of health, nutrition, and population; Millennium Development Goals agreed to by over 180 financial coherence embodied in medium term expendi- governments. It assesses progress to date and prospects of ture framework; pooling of donor funds; and a common achieving the goals.The report argues that where progress framework for reporting and assessing progress. has been slow, the reason is not lack of technical solutions A process has begun that we hope will lead to more but that effective interventions are not used.The report coordinated donor actions to accelerate country level reviews the determinants to low access and use of these progress.This report aims to inform this process at the interventions at household and health systems levels, and global level, resulting in actions to which the international shows the role that policies and institutions play. community is firmly committed. But more importantly, The report argues that low coverage of MDG-related we hope that this report will stimulate and inform policy interventions reflects in part the low level of government dialogue and actions at national and sub-national levels, health spending but more fundamentally the numerous involving the many stakeholders--households foremost policy and institutional weaknesses across the entire health among them--that are critical for achieving the health system.It shows that additional spending on health would and population MDGs. Ultimately, this report will have accelerate progress toward the MDGs in countries with served its purpose if it generates action and change, and if good policies and institutions,while progress in countries it contributes to accelerate the pace toward achieving the where they are weak,requires a more targeted approach.It MDGs. Many lives are at stake.We know what needs to be requires first and foremost a focus on households--not only done. It is time to act. as users of services but as producers of health--and on JEAN-LOUIS SARBIB JACQUES BAUDOUY providers and their accountability. Accelerating progress SeniorVice President Director toward the MDGs in those contexts requires strengthening Human Development Network Health Nutrition and Population of input markets,such as people and pharmaceuticals,and Human Development Network ix Acknowledgments This report was prepared by a team led by AdamWagstaff coordinated by the Production Services Unit of theWorld and Mariam Claeson. Jumana Qamruddin and Henrik Bank's Office of the Publisher, under the supervision of Bjorn Axelsson served as research analysts.The team from Susan Graham and Monika Lynde. within and outside theWorld Bank--AlexandreV. Several consultations with partners provided helpful Abrantes, Anita Alban, Henrik Bjorn Axelsson, Florence guidance for the report:an external and internal technical Baingana, Ousmane Bangoura, Amie Batson, Eduard Bos, consultation with bilateral partners,agencies,and Logan Brenzel, Flavia Bustreo, Jillian Cohen, Isabella nongovernmental organizations to review the messages of an Anna Danel, Monica Das Gupta, François Decaillet, Pablo early draft,held inWashington in January 2003;a high-level Gottret, Davidson Gwatkin, April Harding, Robert policy consultation on the framework for accelerated Hecht, Eva Jarawan,Timothy A. Johnston, Rachel progress toward the health-related Millennium Kaufmann, Peyvand Khaleghian, Rudolph Knippenberg, Development Goals,cohosted by the Canadian International Christoph Kurowski, Rama Lakshminarayanan, Benjamin DevelopmentAgency,the U.K.Department for Loevinsohn, Elizabeth Lule, Joan MacNeil,Tonia Marek, International Development,and theWorld Bank,held in Milla McLachlan, Julie Mclaughlin, Saul Morris, Joseph Ottawa in May 2003;and the High-Level Forum on the Naimoli,Tawhid Nawaz, Kjeld Pedersen, Alexander Health MDGs,cohosted by theWorld Health Organization Preker, Jumana Qamruddin, Sangeeta Raja, G.N.V. and theWorld Bank,held in Geneva in January 2004. Ramana, Pablo Ribera, Juan Rovira, Nicole Schwab, Extensive comments were provided by partners,including Meera Shekar, Agnes Soucat,Susan Stout,Eldaw Suliman, the Canadian International DevelopmentAgency,the U.K. Emi Suzuki,Juan Pablo Uribe,DianaWeil,Harvey Department for International Development,the Netherlands Whiteford,and AlanWright--contributed in various ways Ministry of ForeignAffairs,the Swedish International to the background analysis,country case studies,and review Development CooperationAgency,the United Nations of operational experiences.They also played a part in draft- Children's Fund,and theWorld Health Organization.The ing and reviewing the manuscript.The work was carried analytical work was also informed by the team leaders' out under the general direction of Jacques Baudouy. participation in the Bellagio study group,which reviewed The peer reviewers of the report were Shantayanan the evidence for child mortality and published the results in Devarajan, Deon Filmer, Jonathan Halpern, Daniel Kress, the Lancet series on child survival in July 2003.A Samuel Lieberman, George Schieber, and Eric Swanson. consultation on the costing and financing of the health- An external panel provided advice and technical inputs in related Millennium Development Goals was cohosted by the the early drafting of the report. Bruce Ross-Larson and World Bank and the Disease Control Priorities project,in Meta de Coquereaumont were the principal editors, with Washington in November 2003.Some team members also the assistance of Elizabeth McCrocklin,Thomas Roncoli, participated in the Millennium Project taskforces,and the ChristopherTrott, and ElaineWilson. Book and cover analysis has benefited from interactions with them. design, copyediting, production, and printing were The work was supported in part by DutchTrust Funds. x KEY MESSAGES AND POLICY HIGHLIGHTS At the United Nations Millennium Summit in September most optimistic calculations, the combined effects of 2001,147 heads of state endorsed the Millennium these stimuli will be insufficient for most countries to Development Goals,half of which concern different make the difference between hitting and missing the aspects of health--directly or indirectly.This report assesses Millennium DevelopmentTargets. progress to date toward these goals and analyzes prospects · Developments within the health sector will tend to for the future.The report argues that faster progress can be slow down progress.Many countries cannot sustain the made with existing health interventions.It argues that rate of increase of coverage of attended deliveries they extra government health spending is not enough,and that saw during the 1990s--the inevitable slowdown will policies and institutions within and beyond the health sec- lead to a slower rate of decline for maternal mortality. tor need to be strengthened if faster progress toward the health millennium development goals is to be made.On Accelerating progress toward the Millennium these and other issues the report reaches clear-cut conclu- Development Goals for health is possible in all sions.These are summarized below. regions and countries. Progress toward the Millennium Development · Effective interventions exist for malnutrition,child Goals for health has been mixed. mortality,maternal mortality,and communicable diseases.But they are being used too little by the people · Some good news:nearly 80 percent of the world's peo- who can benefit from them--especially poor people.If ple live in a country that is on track to hit the malnutri- universal coverage rates were achieved for a handful of tion target.In the 1990s,38 percent of countries key child health interventions,the number of under- accelerated the rate of decline for under-five mortality five deaths worldwide would fall by nearly two-thirds-- despite HIV/AIDS.And in two regions--East Asia and the Millennium DevelopmentTarget.A three-quarters Pacific and the Middle East and North Africa--mater- reduction in the rate of maternal mortality--the nal mortality appears to have fallen swiftly in the 1990s. Millennium DevelopmentTarget--could be achieved · But plenty of bad news, too. On under-five mortality, by scaling up coverage rates of a handful of key progress has been much too slow--and has been maternal mortality interventions,including improved getting even slower. Sub-Saharan Africa is lagging access to comprehensive essential obstetric care.The behind badly across all the Millennium Development technology is available--it just needs to be used by all. Goals for health but especially on under-five mortality, for which no Sub-Saharan country is on track to hit Additional resources are required but will not be the target.All regions face challenges on at least some sufficient to reach the Millennium Development Millennium Development Goals, including regions Goals. with many middle-income countries. Overall, however, the poorest countries are progressing the · Increased government spending on health is a part of slowest. And at least for under-five mortality, poor the answer to getting effective interventions used more communities within countries are progressing the widely. But it is not the whole story. In countries with slowest. very poor policies and institutions, across-the-board increases in government spending will have little if any Progress in the second half of the 1990­2015 impact on Millennium Development Indicators. In window will not necessarily be swifter. these countries, improved policies and institutions-- within and beyond the health sector--are crucial if · There will be some stimuli for faster progress outside progress toward the Millennium Development Goals is the health sector. Economic growth is set to increase in to be accelerated. Even in countries with relatively all regions except East Asia and Pacific.The good policies and institutions, additional government Millennium Development Goal agenda may well speed health spending needs to be targeted. Preliminary esti- up progress toward universal primary enrollment, the mates suggest the returns to targeted spending--in elimination of gender gaps in secondary education, and reduced mortality--are much higher than the returns increased access to safe drinking water. But even on the to across-the-board increases in government health spending. xi Strengthening policies and institutions in the and incorporating the lessons into policymaking.And health sector requires working across several all too often intersectoral issues are poorly handled.Yet, interrelated areas. as is becoming increasingly clear, investments outside the health sector (in water and sanitation, and roads and · Stronger policies and better institutions for health transport) do not achieve their full impact on health require lowering the financial and nonfinancial barriers unless they are accompanied by well designed programs that households face in their dual roles as producers of that aim to change people's behavior (for example, health and users of health services. Price is key, hand-washing) or accessing services. especially for the poor, but knowledge and geographic access are also important. Some countries appear to spend less than they · Better policies and institutions within the health sector can afford. also entail improving the performance of health providers--on their quality,their responsiveness,and their · In these countries, governments should consider raising efficiency--through greater accountability.Within the share of government spending in GDP, or increas- provider organizations,stronger management involves ing the health share of total government spending. increasing the accountability of frontline providers to the Such reforms take time and may require technical and organization.But their performance can also be improved financial assistance from donors. Governments need by making them more accountable to the public,whether also to play a stronger role in encouraging or arranging along the direct route (for example,enabling community the pooling of out-of-pocket expenditures through organizations to exercise oversight of providers) or the insurance and prepayment mechanisms. indirect route (making providers more accountable to pol- icymakers through,for example,contracts or agreements, Donors can do better than they have done and making policymakers in turn more accountable to the in the past. public through greater democracy and openness). Reaching the poor requires new thinking about the most · Development assistance to health is too unpredictable, effective service delivery for the different kinds of basic and the transaction costs are too high. Development services:outreach,fixed facilities,and referral. assistance to health--like government spending itself-- · Human resource challenges face most developing is more productive in countries with sound policies countries, and policymakers need to work along several and institutions. Aid can help foster good policies and dimensions.Wages and monetary benefits are facilitate transition to them, but donors cannot force important elements of recruitment and retention policies on countries against their will.Aid is fungible, efforts. But there are other avenues that can usefully be at least in part: this points toward enhanced explored, such as increasing training opportunities, coordination, aid pooling, and putting countries in the enhancing promotion prospects, and so on. driver seat. Global partnerships in health add value, but · For medicines and essential supplies,governments have to contain risks. ensure that medicines reach and are affordable for people · At the High-Level Forum on the Millennium who need them most.Again,working along a variety of Development Goals for health in Geneva in January dimensions is required,including improved logistics,but 2004, there was broad agreement among a wide variety also better incentives,more strategic procurement of different actors--donors, international technical arrangements,and a more responsible role for industrial agencies, philanthropists, and developing countries-- governments in making existing drugs available at afford- on a number of points. It was agreed that the able prices and providing the appropriate incentive envi- Millennium Development Goals for health pose formi- ronment for research and development into diseases that dable challenges and that it is vital for all to rise to disproportionately affect poor countries. them.All agreed that a greater sense of urgency is required if 2015 is not to pass by with a large number Many governments lack credible and resource- of countries having missed the targets.The forum par- backed strategies for the prevention, treatment, and ticipants agreed on actions in the mobilization of control of communicable diseases. resources for health, aid effectiveness and harmonization, human resources, and in the monitor- · Too little effort is devoted to surveillance, monitoring ing of performance. and evaluation, learning the lessons of such exercises, xii| The Millennium Development Goals for Health STATISTICAL HIGHLIGHTS Why the Millennium Development Goals matter-- ied across regions--slowest in Sub-Saharan Africa,swiftest it's the world's poor who die earlier. in the Middle East and North Africa. · 60 percent of the people in the Middle East and North Rates of mortality and malnutrition tend to be much higher Africa are in countries on track to reach the goal for among the world's poor and,with the exception of deaths under-five mortality, 39 percent in Latin America and from HIV/AIDS,lower among better-off sections of the the Caribbean, 28 percent in Europe and Central Asia, world's population. 17 percent in East Asia and Pacific, 10 percent in South · 41 percent of children in the poorest quarter of the Asia, and 0 percent in Sub-Saharan Africa. world's population are underweight, 3 percent in the · 84 percent of the people in the Middle East and North richest quarter. Africa are in countries on track to reach the goal for · 114 deaths among children under five per 1,000 live births for the poorest quarter, 13 per 1,000 for the Figure 2 Progress toward the Millennium richest quarter. Development Goals has varied · 63 maternal deaths per 10,000 live births for the poor- est quarter, 4 per 10,000 for the richest quarter. 100 in 90 80 70 people Figure 1 It's the poor who die earlier 60 of country 50 200 40 180 on-track 30 160 20 Percentage 140 10 120 0 100 East Asia Europe Latin Middle South Sub- 80 and and America East Asia Saharan Rate/ratio 60 Pacific Central and the and North Africa Asia Caribbean Africa 40 20 Underweight 0 children Underweight Under-five Maternal AIDS deaths Tuberculosis Under-five children deaths per deaths per per 100,000 deaths per mortality rate (percent) 1,000 live 10,000 live people 100,000 Maternal births births people mortality ratio Poorest 25% Second poorest 25% Second richest 25% maternal mortality, 69 percent in East Asia and Pacific, Richest 25% 19 percent in Europe and Central Asia, 3 percent in Sub-Saharan Africa, 2 percent in Latin America and the Caribbean, and 0 percent in Sub-Saharan Africa. · 164 AIDS deaths per 100,000 people for the poorest quarter, 31 for the richest quarter. Even within a region the progress toward the Millennium Development Targets can vary Progress toward the Millennium Development Goal tremendously. targets has varied--across indicators and across regions. Many countries in East Asia and Pacific reached the required annual rate of reduction to achieve the target of Progress has been fastest for malnutrition,reflecting in part halving of malnutrition by 2015 in the 1990s, and the the lower target for this indicator (a halving of the rate region as a whole is on track to hit the target. But several between 1990 and 2015,compared with a two-thirds countries have not achieved the required rate of reduction in the case of maternal mortality).Progress has reduction. been slowest for under-five mortality.Progress has also var- xiii Figure 3 Varying progress in East Asia and Pacific Figure 4 The poorest countries are progressing toward the malnutrition goal slowest toward the Millennium Development Goals Under-five Maternal Rep. mortality mortality PDR Malnutrition rate ratio of China Korea, Dem. Malaysia Indonesia Vietnam Thailand Myanmar Cambodia Lao Philippines Mongolia 0 2 rate ­1 change 0 ­2 Target ­2 average (percent) ­3 percentage ­4 ­4 Target 1990s ­6 in ­5 Target average ­8 ­6 change ­10 ­7 Annual Population-weighted ­8 Low-income countries Lower middle-income countries Upper middle-income · China has been reducing malnutrition by 8.2 percent a countries year, Malaysia by 5 percent,Thailand by 2.7 percent (just enough to reach the goal). · Myanmar and Cambodia have been reducing malnutri- tion by only 1.1 percent a year, Lao PDR by 0.9 Most of the world lives in a country where under- percent, and the Philippines by 0.6 percent. five mortality fell slower in the 1990s than it did in · In Mongolia malnutrition has been increasing by 1.1 the 1980s. percent a year. Under-five mortality typically fell slower during the The poorest countries are progressing slowest 1990s than it did in the 1980s. toward the Millennium Development Goals. · In South Asia only 11 percent of the people lived in a country where the rate of decline quickened in the For malnutrition, under-five mortality, and maternal mor- 1990s,in East Asia and Pacific 20 percent,in Middle East tality the story is the same--the low-income countries and North Africa 34 percent,in Europe and Central Asia reduced their rates least quickly in the 1990s. 36 percent,in Latin America and the Caribbean 38 per- · Low-income countries have been reducing cent,and in Sub-Saharan Africa 51 percent. malnutrition by only 2.6 percent a year,lower middle- income countries by 7.6 percent,and upper middle- income countries by 4.6 percent. Figure 5 The pace of decline of under-five · Low-income countries have been reducing under-five mortality slowed during the 1990s mortality by 2.3 percent a year,lower middle-income countries by 2.5 percent,and upper middle-income 1990s 60 countries by 4.9 percent. in 50 · Low-income countries have been reducing maternal country in 40 mortality by 2.4 percent a year,lower middle-income increased countries by 4.9 percent,and upper middle-income 30 people countries 2.5 percent. of 20 decline of 10 rate 0 Percentage East Asia Europe Latin Middle South Sub- and and America East Asia Saharan where Pacific Central and the and North Africa These figures are culled from the text. Please see the text Asia Caribbean Africa for source material. xiv | The Millennium Development Goals for Health ABBREVIATIONS BASICS Basic Support for Institutionalizing Child Survival BCG bacillus Calmette-Guerin (vaccine for tuberculosis) DOTS directly observed treatment, short-course (treatment regimen for tuberculosis) DPT diphtheria, pertussis, and tetanus GAMET Global HIV/AIDS Monitoring and Evaluation SupportTeam GAVI Global Alliance forVaccines and Immunization GDP gross domestic product IDA International Development Association IMF International Monetary Fund NGO nongovernmental organization OECD Organisation for Economic Co-operation and Development PAHO Pan American Health Organization SARS severe acute respiratory syndrome SEWA Self-EmployedWomen's Association TRIPS Trade-Related Aspects of Intellectual Property Rights UN United Nations UNAIDS Joint United Nations Programme on HIV/AIDS UNDP United Nations Development Programme UNICEF United Nations International Children's Emergency Fund USAID United States Agency for International Development WHO World Health Organization xv OVERVIEW Rising to the Challenges The scale of death and ill health in the world is staggering. effects of illness on income may take time to appear and In 2000 more than 11 million children died before their be long lasting. If children are malnourished, they are less fifth birthday.1 Among the world's 613 million children likely to be in school, and they learn less when they are in under five, 140 million are underweight.1 In 1998, 843 school.As a result, they are less productive later in life. million people were classified as undernourished on the The devastating economic consequences of illness and basis of their food intake. In 2001, 3 million people died death are evident at the macroeconomic level, too. from HIV/AIDS.2, Tuberculosis claimed 2 million lives, 3 Demographic variables account for an estimated half of the and in 1995 half a million women died during pregnancy difference in growth rates between Africa and the rest of or childbirth. the world over 1965­90.6 Children require teachers to This heavy burden of death and suffering is heavily con- learn, and in some parts of Africa deaths from HIV/AIDS centrated in the world's poorest countries (figure 1). Just 1 are making a major dent in the stock of teachers--Zambia percent of the world's 11 million under-five deaths now loses half as many teachers as it trains to HIV/AIDS.7 occurred in high-income countries,with 42 percent occur- The AIDS epidemic has been estimated to knock 0.3­1.5 ring in Sub-Saharan Africa alone. Almost half the world's percentage points off rates of economic growth.8 underweight children--65 million children--live in South During the 1990s the international community became Asia. And 98 percent of the world's half million maternal more alarmed about the scale of ill health and death in the deaths took place in the developing world--252,000 in developing world and at the growth of HIV/AIDS and Sub-Saharan Africa alone.1 Within countries, too, it is the tuberculosis.The decade also saw major new global health poor who shoulder the lion's share of the disease burden. initiatives and partnerships, including UNAIDS (the Joint Death and disease matter in their own right, but they United Nations Programme on HIV/AIDS); the Global also act as a brake on poverty reduction.As Nobel laureate Alliance for Vaccines and Immunization; the Stop TB Amartya Sen recently put it, "health is among the most Partnership; the Global Fund to Fight AIDS, TB and important conditions of human life and a critically signifi- Malaria; and many others. With the 1990s drawing to a cant constituent of human capabilities which we have rea- close the international community decided that even son to value."4 But health also matters because it affects more needed to be done.At the UN Millennium Summit living standards--of households and countries. Health in September 2001, 147 heads of states endorsed the expenses can easily become burdensome for households: Millennium Development Goals (MDGs), nearly half of inVietnam alone they are estimated to have pushed some which concern different aspects of health--directly or 3 million people into poverty in 1993.5 indirectly (box 1).* Beyond its direct impact on a household's living stan- dards through out-of-pocket expenditures,ill health has an *Throughout the report,"health" is interpreted broadly. For example, nutri- indirect effect on labor income, through productivity and tion is both a key aspect of health in its own right and a key influence on the number of hours people can spend working. The other aspects of health. 1 Figure 1 The poorest countries suffer the highest Box 1 The health-related Millennium burdens of premature mortality and malnutrition Development Goals 180 160 · Goal 1: Eradicate extreme poverty and hunger.The target 140 is to cut in half the proportion of people who suffer from 120 hunger between 1990 and 2015, with progress measured in terms of the prevalence of underweight children under 100 five.The target implies an average annual rate of reduction 80 Rate/ratio of 2.7 percent. 60 · Goal 4:Reduce child mortality.The target is to reduce the 40 under-five mortality rate by two-thirds between 1990 and 20 2015, equivalent to an annual rate of reduction of 4.3 0 percent. Child Under-five Maternal AIDS Tuberculosis malnutrition mortality mortality mortality mortality per 100 per 1,000 per 10,000 per 100,000 per 100,000 · Goal 5: Improve maternal health.The target is to reduce population live births live births population population the maternal mortality ratio by three-quarters between Poorest 25% of world's population 1990 and 2015, equivalent to an annual rate of reduction 2nd poorest 25% of world's population of 5.4 percent. 2nd richest 25% of world's population Richest 25% of world's population · Goal 6: Combat HIV/AIDS, malaria, and other diseases. The target is to have halted and begun to reverse the Source: References 130, 131, and 123. spread of these diseases by 2015. · Goal 7: Ensure environmental sustainability.The target is to cut in half the proportion of people without sustainable access to safe drinking water by 2015. The health Millennium Development · Goal 8: Develop a global partnership for development. Goals: Progress and prospects The target is to provide access to affordable essential drugs Tracking progress toward the health goals is not straight- in developing countries. forward. Good-quality trend data are available for most Source: Reference 9. countries only for under-five mortality. Poorer-quality trend data are available for a limited set of countries for child malnutrition. Limited trend data are available for maternal mortality, so this study used a model to forecast rates in the 1990s and beyond.Trend data are also limited for communicable diseases, and no modeling was under- 1990s,well short of the target of 4.2 percent.For mater- taken for the report. nal mortality the average population-weighted decline was just 3.2 percent,far from the 5.4 percent target. A mixed score at half-time · Health targets are relevant for middle-income countries as Of the Millennium Development Goals for which trend well as low-income countries. Two of the most affluent data are available or estimated, the fastest progress has been regions--Europe and Central Asia and Latin America on malnutrition (table 1).The following factors are impor- and the Caribbean--have the smallest shares of coun- tant in interpreting the mixed score at half time: tries on track for the child mortality target (22 percent · The number of people living in on-track countries, not just and 10 percent). Only 4 percent of countries in Latin the number of countries, matters.Worldwide 77 percent of America and the Caribbean are on track to meet the the people living in the developing world live in maternal mortality target. countries that are on track to meet the malnutrition · Evidence on how the poor are faring within countries is mixed. target. In Sub-Saharan Africa only 15 percent of the For malnutrition, within countries the poorest 20 per- people live in an on-track country. cent of the population appears,on average,to have been · Different indicators show different levels of improvement. For experiencing broadly similar rates of reduction as the under-five mortality the developing world managed population as a whole. But for under-five mortality the only a 2.5 percent average annual reduction in the rate has been falling more slowly among the poor. 2 | The Millennium Development Goals for Health of in e track n.a. 8.56 8.81 1.6 4.48 0.0 2.7 1.3 7.06 2.1 1.3 0.0 2.03 0.0 8.32 Shar population living on- country ratio for of mortality e n.a. 4.2 7.3 5.6 0.0 4.5 track 14.3 25.0 64.3 12.5 10.3 32.6 10.0 17.4 15.6 nal Shar countries on Millennium Development rgetaT Mater centage ­5.4 ­4.5 ­4.3 ­1.5 ­6.9 ­2.9 ­1.6 ­2.4 ­4.9 ­2.5 ­1.9 ­2.5 ­3.2 ­2.5 ­3.1 Population- weighted average, 1990s (yearly per change) in of in e faster of n.a. 20.2 35.8 37.7 33.5 11.4 50.6 31.3 19.8 28.9 37.0 31.5 25.8 31.7 26.7 Shar population living country with rate decline 1990s of e of 1990s n.a. 56.5 33.3 33.3 26.7 37.5 41.7 39.7 40.7 35.3 21.9 54.2 39.1 35.7 38.2 Shar countries with faster rate decline in rate n.a. centage ­2.7 ­2.8 ­4.4 ­5.6 ­3.2 ­1.1 ­2.5 ­3.3 ­4.4 ­5.2 ­4.7 ­3.0 ­4.7 ­3.2 Population- weighted average, 1980s (yearly per change) of mortality in e n.a. 9.7 0.0 17.4 27.8 38.7 60.0 14.4 20.9 34.2 58.6 41.3 18.7 41.9 22.3 -five Shar people living on-track country cent) Under on (per of e for n.a. 0.0 6.3 26.1 22.2 10.0 46.7 25.0 22.2 23.5 34.4 62.5 15.9 46.4 24.2 Goals Shar countries track Millennium Development rgetaT centage ­4.3 ­2.7 ­2.5 ­3.7 ­3.6 ­2.6 ­0.3 ­2.3 ­2.5 ­4.4 ­5.7 ­4.1 ­2.5 ­4.1 ­2.8 Population- weighted average, 1990s (yearly per change) Development of in e n.a. n.a. n.a. n.a. n.a. 88.7 88.2 57.6 81.5 86.4 15.2 69.1 87.3 65.0 77.0 Shar population living on-track country Millennium evalence for 2. health pr of e track n.a. n.a. n.a. n.a. n.a. Shar countries on Millennium Development rgetaT 45.5 37.5 50.0 62.5 33.3 27.6 30.2 53.8 45.5 40.0 chapter see selected d Underweight towar centage n.a. n.a. n.a. n.a. ­2.7 ­6.7 ­9.6 ­4.1 ­6.3 ­3.5 ­0.2 ­2.6 ­7.6 ­4.6 ­5.0 calculations;f Population- weighted average, 1990s (yearly per change) staf ess ogr Bank Pr and orld and Asia 1 East middle middle applicable. W Asia America Asia Pacific the North income income income oup ope Not ableT Region/ income gr Millennium Development rgetaT orld East the Eur Central Latin and Caribbean Middle and Africa South Sub-Saharan Africa Low Lower income Upper income High (non­OECD) High (OECD) Developing countries Industrial countries W n.a. Source: The Challenges | 3 Will the second half go better? tries and regions firmly toward the targets, they will for the most part not get them there (figure 2). The picture is As a comparison of the child mortality experiences in the bleakest for under-five mortality--and for Sub-Saharan 1980s and 1990s demonstrates, past performance is not Africa. Even with Sub-Saharan Africa's extra percentage necessarily a good predictor of future performance. The point in annual reductions in under-five mortality coming fact that a country is on track on the basis of its perfor- from economic growth and achievement of the gender and mance in the 1990s does not guarantee that it will main- water goals, its projected rate of reduction of under-five tain the required annual rate of reduction of malnutrition mortality for 2000­15 is still only 1.6 percent a year. or mortality during 2000­15. It is also possible that coun- tries currently off track may get on track in the second half WHY THE PICTURE MAY BE BLEAKER The assump- of the Millennium Development Goal "window." tions underlying these calculations are probably overopti- mistic.The gender and drinking water targets may well be STIMULI FROM OUTSIDE THE HEALTH SECTOR CAN BE EXPECTED The World Bank estimates that missed. And contrary to what has been assumed, it is economic growth will fall somewhat in East Asia and the unlikely in the absence of these stimuli that the pace of Pacific in 2000­15, turn from negative to positive in decline of the Millennium Development Indicators Europe and Central Asia and Sub-Saharan Africa, and achieved in the 1990s will continue. The three variables increase somewhat in Latin America and the Caribbean, used to forecast maternal mortality in the model are likely the Middle East and North Africa, and South Asia.10 to be less conducive to reductions in maternal mortality in Primary education completion rates will also probably the new millennium than in the 1990s. grow faster in the new millennium as a result of the "Education For All" and "FastTrack" initiatives. But higher The goals matter for all countries rates of educational attainment among women of child- Given the likelihood that many countries will miss several bearing age will not show up until 2005 or so, and even of the goals,why should they be taken seriously?The goals then the first full round of effects on under-five mortality matter for several reasons. will not be felt until 2010. More relevant is the fact that gender gaps in secondary education may well narrow more · Faster progress is important even if targets are missed. A key quickly in the new millennium than in the 1990s as a result message of this report is that progress can be acceler- of the gender Millennium Development Goal.To achieve ated through a judicious mix of spending and policy parity with boys by 2015 in the proportion of the popula- and institutional reform. tion 15 and over who have completed secondary educa- · The goals facilitate benchmarking. By focusing on a limited tion, girls' completion rates will have to grow faster in the set of outcomes, the Millennium Development Goals new millennium than in the 1990s in most regions, espe- show what is achievable and where faster progress can cially in South Asia and East Asia and the Pacific.And if the be made. water Millennium Development Goal is to be reached, access rates will need to grow much faster in 2000­15, · The poor risk being left behind. One weakness of the especially in Sub-Saharan Africa. Millennium Development Targets is that they are national averages and so do not remind us automati- EVEN WITH ECONOMIC GROWTH AND FASTER cally that progress needs to be for everyone,not just the PROGRESS ON THE NONHEALTH GOALS, MANY better off. Progress has been uneven, with the poorer REGIONS WILL STILL MISS MANY OF THE countries lagging behind the rest. For under-five mor- HEALTH TARGETS The combined contributions to the tality, the poor within countries are lagging behind the decline in malnutrition and mortality of faster economic rest of the population. Progress needs to be monitored growth and achieving the gender and water goals could be and analyzed by income--and efforts directed toward significant. In Europe and Central Asia progress on these population groups that are being left behind. fronts might add as much as 1.4 percentage points to the rate of decline of under-five mortality, 1.1 percentage points to the rate of decline of maternal mortality, and just Effective interventions exist--they under 1 percentage point to the annual rate of reduction in need to reach more people underweight children. In South Asia they might add as much as 2.6 percentage points to the annual rate of reduc- Lack of interventions is not the obstacle to faster progress tion in maternal mortality, taking it from 2.9 percent a year toward the goals. It is the low levels of use--especially to 5.5 percent.While these contributions will push coun- among the poor--of interventions that work. 4 | The Millennium Development Goals for Health Figure 2 Faster economic growth and other changes outside the health sector will help move regions toward the targets, but in most cases it will not get them there Underweight 4% 2% rate 0% ­2% growth ­4% percentage ­6% Maternal mortality 0% ­8% Annual ­10% ­2% ­12% rate East Asia Europe Latin Middle East South Sub-Saharan ­4% and Pacific and America and North Asia Africa Central and the Africa growth Asia Caribbean ­6% Extra government health spending Extrasectoral contributions Current percentage ­8% Required 2000­15 Target Annual ­10% Under-five mortality 0% ­12% East Asia Europe Latin Middle East South Sub- ­1% and Pacific and America and North Asia Saharan Central and the Africa Africa Asia Caribbean rate ­2% Extra government health spending Extrasectoral contributions ­3% Current growth Required 2000­15 ­4% Target percentage ­5% ­6% Annual ­7% ­8% East Asia Europe Latin Middle East South Sub-Saharan and Pacific and America and North Asia Africa Central and the Africa Asia Caribbean Extra government health spending Extrasectoral contributions Current Required 2000­15 Target Source: World Bank staff. The Challenges |5 The array of existing effective interventions is impressive If use of all the proven effective childhood preventive and treatment interventions were to rise from current levels to The available interventions constitute a powerful arsenal 99 percent (95 percent for breastfeeding), the number of for preventing and treating the main causes of malnutrition under-five deaths worldwide could fall by as much as 63 and death (see chapter 3, table 3.1). Diarrhea, pneumonia, percent.11 Deaths from malaria and measles could be all but and malaria account for 52 percent of deaths among chil- eliminated. And deaths from diarrhea, pneumonia, and dren worldwide. For each of these causes, there is at least HIV/AIDS could be dramatically reduced. If coverage rates one proven effective preventive intervention and at least of the key maternal mortality interventions were increased one proven effective treatment intervention, each capable from current levels to 99 percent, 391,000 maternal deaths of being delivered in a low-income setting. In most cases, worldwide--74 percent of current maternal deaths--might several proven effective interventions exist. For diarrhea-- be averted. One intervention stands out as especially the second-leading cause of child deaths--there are no important: access to essential obstetric care, which accounts fewer than five proven preventive interventions and three for more than half the deaths averted (figure 3). proven treatment interventions. Effective interventions are underused, What countries need to do especially by the poor to rise to the challenges Why, then, are the high rates of malnutrition and death in If the lack of interventions is not keeping countries from the developing world so high? For one reason or another, achieving the goals,what is?What do countries need to do people do not receive the effective interventions that could to accelerate progress toward them? save their lives or make them well nourished.In upper mid- dle-income and high-income countries, 90 percent of chil- dren receive DPT3 vaccinations, more than 90 percent of babies are delivered by a medically trained person,and more than 90 percent of pregnant women make at least one ante- Figure 3 Full use of existing interventions would natal visit. In South Asia less than 50 percent of pregnant reduce maternal deaths dramatically women have antenatal checkups, and only 20 percent of babies are delivered by a medically trained person.The story Improved access to safe abortion services is similar for other childhood interventions--and for inter- ventions for other goals. Of the estimated 6 million people Improved access to comprehensive in low-income and middle-income countries currently essential obstetric care needing antiretroviral therapy, only 300,000 receive it. In Tetanus toxoid immunization Asia, where more than 7 million people are living with HIV/AIDS, no country has exceeded 5 percent antiretrovi- Treatment for iron deficiency ral therapy coverage. Just as shortfalls in coverage vary across countries, they Drugs for preventing malaria vary within countries, with the poor and other deprived groups invariably lagging behind.These groups are less likely Active management in third stage of labor to receive full basic immunization coverage,have their deliv- Hemorrhage eries attended by a medically trained person,or make at least Magnesium sulphate for pre-eclampsia Puerperal infection one antenatal care visit to a medically trained person.On the Eclampsia positive side, the poor for the most part are making faster Calcium supplements Obstructed labor during pregnancy Abortion complications progress in coverage, reflecting that the better off already Malaria have high coverage rates for many interventions.Progress has Antibiotics for treating Anemia bacterial vaginosis been more propoor in professionally delivered interventions. Tetanus Antibiotics for preterm Underuse of effective interventions costs lives rupture of membranes 0 0.1 0.2 0.3 0.4 The low use of effective interventions--in the developing Percentage of maternal deaths averted world in general and among the poor in particular--trans- lates into rates of mortality, morbidity, and malnutrition Source: World Bank staff estimates. that are far higher than they need be. 6 | The Millennium Development Goals for Health Extra government health spending alongside child mortality and uses the Bank's Country is needed, but it is not enough Policy and Institutional Assessment index to measure the quality of policies and institutions. Some argue that the cause of slow progress is the lack of government health spending.12 The evidence presented in How much would reaching the goals cost support of this view is not altogether compelling.Arbitrary in well-governed countries? assumptions are made about the links between scaling up Well-governed countries with good policies and institu- expenditures and intervention coverage rates, and no tions could, in principle, achieve the goals simply by scaling explicit assumptions are made about the links between up their expenditures on existing programs in proportion to coverage levels and health outcomes.There is no assurance current allocations. In practice, however, the amount of that incurring the extra government health spending extra spending required would likely be prohibitive. Take claimed to be necessary would reduce mortality at all--let East Asia and the Pacific. If economic growth proceeds as alone by the proportions required to reach the goals. expected and the other relevant Millennium Development In fact, some writers have argued that government health Targets are hit, the region would achieve the required rates spending has little impact at the margin on health outcomes, of reduction of underweight and maternal mortality even once the effects of other determinants have been accounted without additional government health spending. But it for.The reason? The many weak links in the chain running would miss the under-five mortality target.To hit this tar- from government spending to health outcomes.13­15 Their get, a minimum of five percentage points would need to be evidence has limitations. It refers simply to child mortality, added to the rate of growth of the government health share not to health outcomes in general.It indicates what happens of GDP.That would take the projected share of GDP spent to child mortality among the whole population, in an aver- on government health programs to 3.7 percent in 2015-- age country, and therefore hides significant spending effects more than twice what it would be if the 1990s pattern of among specific subpopulations and in well-governed coun- growth continues. tries.16­19 And it indicates what would happen to child mor- In Sub-Saharan Africa the conclusions are even starker. tality if additional government spending were to take the Even if faster economic growth materializes and the other form of a proportional scaling up of all government health targets are hit, the share of government health spending in programs, not what would happen if extra spending were GDP would need to grow by an additional 12.3 percentage focused on specific subpopulations or specific programs. points a year, taking the share to 12.2 percent in 2015. Policies and institutions mediate the impact Compare that with a 2000 figure of 1.8 percent and a 2015 of government health spending forecast of 2.2 percent based on the 1990s annual growth of just 1.2 percent in the government health share of GDP. In countries with good governance, additional government Poorly governed countries cannot expect to make health spending does reduce child mortality.19 This result is much progress toward the Millennium Development consistent with recent studies that find that the elasticity of Goals simply by scaling up their expenditures on existing infant mortality to development assistance depends on the programs in proportion to current allocations. And while quality of a country's policies and institutions.Development well-governed countries could in principle simply scale assistance has a stronger effect in countries with strong poli- up existing spending to reach the targets, this is unlikely to cies and institutions than in countries with only average be affordable--for them or for their donors. quality policies and institutions--and an insignificant effect What are the implications for health spending?The first in countries where policies and institutions are weak.20,21 is that targeting additional government spending is impor- The assertion is also consistent with the findings of a study tant for both sets of countries.The second is that building undertaken for this report, which includes other outcomes good policies and institutions is important for all coun- tries: it increases the productivity not just of additional spending but also of existing spending commitments. But The Country Policy and Institutional Assessment is an annual assessment what do better policies and institutions entail in the health byWorld Bank staff of the quality of International Bank for Reconstruction sector? Health systems are very broad--far broader than and Development and International Development Association borrowers' many people think.Weak policies and institutions can arise policy and institutional performance in areas relevant to economic growth at several points along the pathway from government and poverty reduction. It consists of 20 equally weighted criteria represent- health spending to health outcomes (figure 4).The report ing policy dimensions of an effective poverty reduction and growth strategy, uses this framework to tackle the difficult question of how such as economic management, structural policies, policies for social inclu- sion and equity, and public sector management. to build stronger policies and institutions. The Challenges |7 Figure 4 Paths to better health, nutrition, and population outcomes Millennium Health Sector and Government Development Households/Communities Related Sectors Policies and Actions Goals Health providers · Private: for-profit, Health Household nonprofit actions and risk · Public: hospitals, · Underweight factors Household assets primary care, · Under-five informal mortality rate · Health sector policies · Human · Maternal · Use of health at macro, health system mortality ratio services · Physical and micro levels · Communicable · Dietary, sanitary, · Financial · Government spending diseases and sexual Finance and inputs practices · Lifestyle · Public and private insurance; financing and coverage · Drugs Other MDG outcomes · Poverty Other policies Community Suppliers in factors related sectors · Infrastructure · Cultural norms · Food · Transport · Community · Energy · Energy institutions · Roads · Agriculture · Social capital · Water and · Water and · Environment sanitation sanitation Source: Reference 132. Improving expenditure allocations and tertiary infrastructure and personnel--despite low bed and targeting occupancy rates. Some governments have tried to scale back the share of hospital spending.Tanzania,for example,reduced Geographic targeting the share from 60 percent in 2000 to 43 percent in 2002. Simply reallocating the budget toward primary care In most countries, governments spend most of their need not result in higher payoffs to government health money in cities, and spending disproportionately benefits the better off.22,23 Resource allocation formulas can be spending in lower malnutrition and child and maternal used to reduce government spending gaps across regions.24 mortality, however.14,15 In many instances service providers have failed to deliver good-quality care and use resources In Bolivia, which has used such formulas since 1994 as efficiently.The trick is to couple expenditure reallocations part of its decentralization efforts, some fairly large--and with measures to improve the performance of primary care propoor--improvements in maternal and child health facilities and district hospitals--and measures to ensure that indicators have occurred. Targeting resources to poor households actually demand relevant interventions. regions and provinces may benefit from nontraditional mechanisms for priority-setting and implementation, such Targeting specific programs as social investment funds. A recent impact evaluation in Bolivia concluded that such funds were responsible for a Programs such as the directly observed treatment, short- decline in under-five mortality from 88.5 per 1,000 live course for tuberculosis (DOTS) and Integrated Manage- births to 65.6.25 ment of Childhood Illness for child health are good examples of programs that may yield high returns to gov- Changing the allocation of spending across levels of care ernment spending at the margin. Both are the subject of Developing country spending on health is characterized by a ongoing evaluation, but early results are encouraging.26,27 surprisingly high concentration of spending on secondary A recent World Bank study28 in India provides further 8 | The Millennium Development Goals for Health support for the idea that the way government spending is providers (if patients don't demand care, providers cannot allocated across programs makes a difference to its impact deliver it) and as producers of health through the delivery of on the Millennium Development Indicators. home-based interventions and in their everyday health behaviors (this is especially important for child health). In Targeting specific population groups both roles they face barriers. Policymakers need to be aware Many countries subsidize all government health services of those barriers and to formulate appropriate policies. for everyone.These blanket subsidy schemes fail to target Lowering financial barriers interventions that give rise to externalities, and they fail to disproportionately benefit the poor--this, despite the Low income is a barrier to the use of most health interven- stronger equity case for subsidizing their care and the fact tions.Economic growth is therefore an important weapon in that they tend to bear a disproportionate burden of mal- the war against malnutrition and mortality.31,32 But social nutrition and child and maternal mortality. protection programs are also important. South Africa's old- age pension scheme increased the height of under-five black Targeting spending to remove bottlenecks children by eight centimeters, half a year's growth.33,34 Also Another approach is to assess the health sector impediments important are informal community solidarity schemes,which to faster progress in a country,identify ways to remove them, often substitute for formal social protection programs.35 and estimate both the costs of removing them and the likely The other part of the affordability equation is price. impacts of their removal on Millennium Development Goal Higher money prices tend to reduce demand--especially outcomes.29 Work along these lines--sometimes referred to among the poor--unless accompanied by improvements as marginal budgeting for bottlenecks--has begun in several in service quality.36 In many cases out-of-pocket payments African countries and India.30 In Mali key bottlenecks were are informal rather than formal.37 And it is not just the identified for supporting home-based practices and deliver- payments to health providers that matter--users of health ing periodic and continual professional care.These included services incur other money costs in using health facilities, low access to affordable commodities and the need for com- including transportation costs. munity-based support for home-based care; low geographi- User charges for Millennium Development Goal inter- cal access to preventive professional care (immunization, ventions are to be discouraged.Why? Many of these inter- vitamin A supplementation, and antenatal care); shortages of ventions involve benefits that spill over to people who do qualified nurses and midwives; and an absence of effective not receive the intervention (immunization is a classic third-party payment mechanisms for the poor for profes- example). But an equity case can also be made for reduc- sional continuous care. ing prices facing the poor and near-poor, even where Combining estimates of the costs of measures to there are no spillovers. Subsidies should thus be targeted to remove these bottlenecks with estimates of the mortality services with spillovers and to the poor. In practice, they reduction on their removal gives an elasticity of mortality are often badly targeted in at least one respect if not both. to government health spending that compares very favor- There are exceptions, however. In Ifakara, Tanzania, a ably with the elasticities for untargeted government health voucher program for mosquito nets was launched success- spending--even those for well-governed countries with fully for pregnant women and children under five.38­40 good policies and institutions. Only time--and careful And in Indonesia a health card introduced during the eco- monitoring and evaluation--will tell whether these esti- nomic crisis increased use among the poor.41 mates turn out to be accurate. Some recent programs,especially in Latin America,have not simply made health care affordable for the poor--they Better policies toward households--as have made it profitable. Rather than simply reducing the producers and demanders of care cost of using specific interventions these programs provide cash payments to users,linked to specific interventions and Improving people's health is the overarching aim of the restricted to certain groups--often poor mothers and their health sector.But households are not just the endpoint of the children.The experience with these programs, in targeting sector's activity. They are major actors. Indeed, they play a and impact, is encouraging.42­47 dual role: as users of health services delivered by professional There is another reason for limiting user charges. Risk aversion coupled with the unpredictability of illness pro- vides a motivation for pooling risks through an insurance The term "household" is used here to refer to whatever grouping of peo- scheme. Lack of insurance and the consequent exposure to ple share responsibility for health. It is not limited to parents (caregivers of children might be grandparents, aunts, stepparents) and can encompass the the risk of medical expenses cause households to hold more broad array of kinship and household patterns around the world. wealth (and more of their wealth in liquid form) than they The Challenges | 9 otherwise would in the hope that they can smooth their and skills is through information dissemination and coun- consumption when health shocks occur.48 Evidence from seling in the health sector. rural China suggests that households fail in these efforts and One venue for delivering health messages is the public that the poor have the least success in self-insuring against facility. A lactation clinic at the Children's Hospital in income shocks.49 Insurance in the developing world is very Islamabad, Pakistan, promoted exclusive breastfeeding by limited, and those who are least able to smooth consump- altering women's perceptions of its importance and by tion without insurance are the least likely to have insurance counseling them on techniques.59 But a focus on public coverage.50 Governments have a role to play here. In Egypt health facilities is far from ideal, since many people, a school health insurance program for all children attending including many poor people, do not use them when they school resulted in larger increases in coverage among the fall ill--and may not even seek care at all. Programs in the poor than among other groups and achieved considerable community--where the conveyers of knowledge seek out impact on use and out-of-pocket expenditures.51 target groups--seem likely to have a better chance of reaching a broader group and reaching the poor. Empowering women There are several success stories here. In Brazil health Women exercise little control over household resources in knowledge among mothers and feeding practices improved many countries.All else constant, such women are less likely after health workers trained by the Integration of Childhood to receive antenatal care, to have antenatal visits, and to have Management Illness provided information and counseling at visits in the first trimester of pregnancy.52 Microcredit pro- health facilities and in the community.30 After only 18 grams aimed at poor women are thought to be one way of months the nutritional status of children in the area increasing women's financial autonomy. Whether they improved as well. Social marketing and media campaigns increase the use of maternal and health services is less clear.53 have also proved effective in some circumstances.38,39,60 Also important is a woman's ability to make decisions more Reducing time costs generally. In India, although contraceptives are readily avail- able in retail shops, community pressure or disapproval by Transportation systems, road infrastructure, and geography husbands often prevent women from using them. influence the demand for care delivered by formal providers through their impact on time costs, which can Providing information--enhancing knowledge be substantial.61­65 In rural communities, where roads are Lack of knowledge is a major factor behind poor health. It poor and transport unreliable, the time spent waiting for results in people not seeking care when they need it, transport is also a major cost.Time costs tend to be a major despite the absence of price barriers.In Bolivia a large frac- issue for maternal mortality. Health centers are unable to tion of poor babies are not delivered by a trained attendant provide essential obstetric care for a complicated delivery, even though the mothers are eligible for free care under and hospitals, which could provide the needed care, are the Maternal and Child Health Insurance program.54 Lack hard to reach. Road rehabilitation and other transport of knowledge also results in people, especially poor people, projects are important here.66 But so are subsidies linked seeking and receiving inappropriate care--and paying for to the use of health services. Malaysia and Sri Lanka pro- it.55 Ignorance may also result in people not getting the vide free or subsidized transportation to hospitals in emer- maximum health gain out of inputs they have available to gencies.67 Other options to tackle inaccessibility include them for use. Many people do not know that piped water engaging in outreach, building new public facilities in in many countries requires further purification or that underserved areas, and establishing partnerships between hand-washing confers much of the health benefit of piped government and nongovernmental organizations (NGOs), water. Not surprisingly, piped water has a much greater private providers, or community organizations. impact on the prevalence of diarrhea among the children Providing access to water and sanitation of the better off and better educated.56 Better-educated women--especially those with a sec- The availability of plenty of water and improved sanitation ondary education--achieve better health outcomes for are associated with better maternal and child health out- themselves and their children.57 They do this not by using comes, at least among the better educated, even after con- health-specific knowledge that they acquire at school but trolling for other influences.61,63,68­74 This isn't altogether by using general numeracy and literacy skills learned at surprising. Hand-washing is easier if the household has school to acquire health-specific knowledge later in life.58 piped water that provides readily available quantities of So,while better-educated girls will mean healthier women safe water. And the safe disposal of feces is easier if the and healthier children in years to come, a shorter and household has an improved form of sanitation.The devel- more direct route to increasing health-specific knowledge oping world lags well behind the industrial world in 10 | The Millennium Development Goals for Health both--and poor people fare especially badly.They are less leading from the citizen to the policymaker and thence to likely to be connected to a network, and the sources they the provider. Still others work along both routes simulta- rely on tend to be more costly per liter than the net- neously. worked services used by the better off.75 The two challenges from a health perspective are to Improving management--increasing accountability increase access to water and sanitation infrastructure and within provider organizations to ensure that people know how to get the maximum Much of the world's public sector is managed through health benefits from such investments. The first was what is, in effect, a command-and-control structure. addressed extensively in the World Bank's 2004 World Initiative and decisionmaking are exercised only at the Development Report: Making Services Work for Poor People.75 highest level. Problems at lower levels are passed up to The second is addressed in chapter 8 of this volume. higher levels for decisions.There are few managers in the true sense of the word. Instead, administrators execute Improving health service delivery decisions according to previously agreed protocols and rules, with little or no scope for autonomous decision- Health providers--in both the public and private sectors, making at facilities. and in both the formal and informal sectors--deliver inter- Management styles have recently begun to change, ventions of relevance to the Millennium Development though the impact is not altogether clear. Responsibility Goals. Many are efficient, deliver high-quality care, and are for tasks and decisionmaking is delegated to specific parts responsive to their patients. But many are not. As a result, of the organization and to specific individuals. Individual resources--public and private--are wasted and facilities sit accountability is emphasized, and there is a focus on per- underused. Patients often receive care--and pay for it, out formance--not inputs or processes but outputs and out- of very limited means--that is inappropriate to their needs. comes. Good performance is rewarded, financially or in They may also receive downright dangerous care. some other way.There is also a focus on clients and a belief Two things can make a difference. One is the quality of that an organization is ultimately accountable to its clients. management. Better management means a clearer delin- A client-oriented strategy emphasizes customer choice eation of responsibilities and accountabilities inside orga- and satisfaction. Business techniques enhance performance nizations, a clearer link between performance and reward, and are a standard part of strategic planning. and so on. Management means getting accountabilities The new approach is evident in several countries, right within an organization. The other thing that can including Malaysia. Elements of the approach are also evi- make a difference is getting accountabilities right between the organization and the public.75 Some strategies work dent in successful nutrition and child health programs. In the Tamil Nadu Integrated Nutrition Program, in India, along a "short route" leading directly from the patient to community nutrition workers were given clearly defined the provider (figure 5). Others work along a "long route" duties. Information on outputs enabled the community to keep them accountable, but it also enabled the nutrition workers to see how their program was working. In the Figure 5 What makes service providers Programa de Agentes de Saude, in Céara, Brazil, health accountable agents and nurse-supervisors were assigned clear tasks and The state given clear responsibilities.The program has been credited Politicians Policymakers with substantially reducing child mortality.76The intended outcomes of the program were emphasized to health workers and members of the public. Good team perfor- Voice Long route of accountability Compact mance was rewarded with a prize.And health agents were held accountable through community-based monitoring. Citizens and clients Providers Shortroute Coalitions and inclusion Increasing the accountability of Clientpower Management provider organizations to the public Nonpoor and poor Frontline Organizations GOVERNANCE PARTICIPATION Having community representatives participate in the governance and oversight Services of providers can improve the productivity and quality of Source: Based on reference 75. public sector providers. Relatively little is known about the impact, and several governments have had difficulty The Challenges |11 establishing meaningful participation. But where it has of performance expectations are easily measurable (as in been established, the changes have been for the better. In primary health care), and how far capacity-strengthening Burkino Faso participation of community representatives of the payer or funder is addressed as a central part of the in public primary healthcare clinics increased immuniza- initiative. tion coverage, the availability of essential drugs, and the Contracting with nonprofits is most common in low- percentage of women with two or more antenatal visits.77 income countries. Most cases have had positive impacts on In Peru comparisons of primary healthcare clinics with target outcome or output variables. In Bangladesh con- and without governance participation indicated that gov- tracts with nonprofits for the planning and implementation ernance was associated with decreases in staff absenteeism of an expanded program on immunization was credited and waiting times and increases in perceived quality by with a dramatic increase in immunization. In Haiti con- patients.22,78 The approach probably works best for pri- tracting for a primary healthcare package also significantly mary care and when strong technical and advisory support increased immunization coverage.84 In Bangladesh, Mada- is provided to community representatives. gascar, and Senegal significant reductions in nutrition rates were attributed to contracting initiatives.85 Only a few CONTRACTING Contracting can take the form of cases assess efficiency. Contracting with nonprofits works internal contracts within the public sector or external best when the contractors have well-functioning account- contracts between the public sector and the private sector, ability arrangements and strong intrinsic motivation. The whether nonprofit or for-profit. Contracts are between government needs to be capable of assessing, selecting, and the policymaker and the provider,specifying remuneration managing the ongoing relationship with contractors. It for the delivery of certain types of service,with or without must be able to fulfill its side of the deal (contractual agree- a quality threshold. Payment is tied to some measurable ments on funding are genuine) and not interfere with the aspect of performance. Some contracting arrangements running of the services. also strengthen the short route of accountability--by Results on contracting with for-profits are mixed. encouraging patients to choose between providers and Efficiency gains were achieved in contracting for high- having the payment follow the patient, for example. tech diagnostic services in Thailand.86 But experience Contracts with NGOs commonly include outreach in from the hospital sector warns that weak government urban slums for health promotion or education, services contracting capacity often allows the provider to capture for stigmatized or hard-to-reach groups, and social mar- efficiency gains or expand volume to generate more keting of priority health goods or services, such as con- income. In Zimbabwe the cost per service decreased, but doms and oral rehydration salts. In the for-profit private the lack of volume control led to an increase in total sector, contracting often focuses on primary care services cost.87 Other adverse outcomes are also possible. In Brazil and services for which the public sector lacks capacity, contracting with for-profit hospitals led to increases in such as diagnostic and high-tech services. access but also to false billing and cream-skimming to Evidence on the impact of contracting within the public avoid costly patients.88 These problems seem less pro- sector is mixed, but such contracting seems to work best in nounced in primary health care. In Peru and El Salvador middle-income countries. In several countries in Europe contracting with private primary healthcare providers and Central Asia, there is evidence of positive impact from performance-based payment at the primary care level.79­82 increased access, choice, and consumer satisfaction.89 Contracting with for-profit providers seems to work best In Argentina and Nicaragua social security institutes have when the government invests in the development of increased productivity by establishing capitation-based pay- capacity to manage the contracting process,90 when qual- ments for an integrated package of inpatient and ambula- tory services.83 In several countries in Europe and Central ity is at least as high in the private sector as in the public sector (ability to monitor quality is usually low), and Asia the introduction of capped case-based payments for when it involves primary care or other relatively observ- inpatient services was followed by an increase in services and a reduction in average length of stay.79 Quality has not able services (diagnostic services). deteriorated because of skimping on costly unobservable DECENTRALIZATION Decentralization can increase aspects of quality. In the former Soviet republics and in patient leverage because local governments are more easily Latin America, however, the results are more ambiguous. pressured than central ones. It strengthens policymaker- Key influences on the success of contracts within the pub- provider accountability because the government supervisor lic sector include whether the provider has the ability to gets closer to the provider.But the impact of decentralization respond, whether service commitments are congruent on the health sector has been mixed in low-income coun- with funding levels, whether output and key components tries. In Tanzania it improved the efficiency of and access to 12 | The Millennium Development Goals for Health primary healthcare services, but in Ghana efficiency sion if they graduate. They also encourage people in the declined.91 There is less evidence on the impact in middle- profession to think about leaving--to exit from the labor income countries,though in Colombia it has improved both force, join another profession, or leave for another coun- responsiveness and equity.92 One important factor influenc- try.102 And they encourage absenteeism, which makes for ing the success of decentralization is whether the other arm higher workloads for those left behind, further reducing of the long route of accountability (the voter-policymaker motivation and prompting further absenteeism and exits. link) is functioning well. Also important are building local Low rates of compensation in rural settings help explain government capacity--planning, supervision, budgeting, rural-urban imbalances. And compensation differentials expenditure and financial management--and setting up between the public and private sectors influence transitions mechanisms to ensure continuing capacity to cover core from the public sector to the private--and occasionally public functions. back again. Narrowing compensation differentials is clearly an Ensuring adequate human important potential policy tool. Thailand has attracted resources for health back medical professionals through a reverse brain-drain program offering generous research funding and monetary A common lament in international health is that faster progress incentives.103 Zambia more than doubled nursing salaries toward the health-related Millennium Development Goals is with the support of 16 development partners. Several being impeded by a variety of human resources problems. countries have experimented with bonus schemes for health workers working in rural areas, some successfully. The issues There are, however, limits to what can be achieved Human resource stocks in health are often low, and in through changes in compensation alone. Fortunately, some countries they are falling. In Europe and Central research suggests that nonpecuniary aspects of jobs also Asia there are on average 3.1 physicians per 1,000 popula- matter to people. A recent study from India shows what tion, in Sub-Saharan Africa just 0.1.Tanzania is projected public and private health workers hope to get from their to see its health workforce fall from 49,000 in 1994 to job and the extent to which they get it (figure 6).104 It 36,000 in 2015.93 Many developing countries appear to shows that in the public sector, health workers want--and face the double burden of low personnel inflows and high have--good working relationships with colleagues, free- personnel outflows.94­98 dom from political interference, and absence of bribes. Skills are often woefully inadequate, with misdiagnosis What they lack are the tools and materials they need to and mistreatment commonplace.99 Even if the correct apply their skills, training opportunities and opportunities treatment is administered, there is no assurance that it will to advance, employment benefits, good physical working be administered successfully. In the public sector in India conditions, income, and time for personal and family life. in the early 1990s less than 45 percent of patients diag- A recent study in Uganda found that medical staff work- nosed with tuberculosis were successfully treated.100 ing in religious nonprofit institutions work for a wage It is not just the level of skills that matter--it is also the below the market rate but that these institutions provide skill mix. A recent study in Tanzania found an excess of more propoor services than other providers and more ser- unskilled labor of 5,000 full-time equivalents and a short- vices with a public good element--dimensions of the job age of skilled labor of 8,000.93 that the underpaid staff presumably derive some satisfac- Low application to the job is another concern. Recent tion from.105 There would seem to be much scope for random surveys of primary health facilities in six developing using the results of such studies to build feasible and better countries found absenteeism rates of between 19 percent human resource policies. (Papua New Guinea) and 43 percent (India).75 In Tanzania Refining recruitment and training time-and-motion studies showed overall staff productivity in public facilities as low as 57 percent,with only 37 percent People raised in rural areas are more likely to practice in of staff time spent on patient care and as much as 10 percent rural locations and to choose family medicine.106 In spent on irregular breaks and social contacts.101 Thailand rural recruitment and training yielded some suc- cess. Evidence suggests that rural service of graduates Narrowing compensation differentials lengthened, with two-thirds of the graduates continuing Low wages in the medical professions relative to wages in their rural placement after their compulsory years.103 other professions discourage people from entering training Training opportunities are clearly valued by health work- institutions,completing their studies,and joining the profes- ers.But training makes health workers more marketable and The Challenges | 13 Figure 6 What health workers in Andhra Pradesh want from their jobs--and whether they get it Andhra Pradesh: Public health workers Good physical working conditions Good income 100 Knowing what is expected and being able to achieve it Good employment benefits 80 Freedom from political interference 60 Time for personal or family life Not needing to pay bribes 40 20 Desirable location Security of job Important to health workers? 0 Present in workplace? No interference by superiors Training opportunities Respected and trusted by clients Challenging work Superior recognizes good work Tools and materials to use skills fully on the job Good working relationship Good opportunities with colleagues to advance Source: Reference 104. more likely to leave the public sector for the private sector Getting drugs to the frontline or foreign countries. One way around this is to focus gov- Some countries have succeeded in improving the supply ernment spending on the development of specific skills and chain through better logistics. In Ghana, as part of the to leave health workers themselves to pay--in the form of "Strategies for Enhancing Access to Medicines" initiative lower wages--for any general training.This seems sensible funded by the Bill and Melinda Gates Foundation, part of in the light of the emerging evidence on the poaching and the drug distribution system has undergone innovative international migration of health workers. The countries reforms to improve mission hospitals'use of medicines,with that have emulated the training standards of industrial coun- good results. Better management is important. But without tries (such as Ghana) are those that have been most vulnera- ble to poaching by them.75 Training in areas of special institutional arrangements that give the right incentives to the different actors involved in drug distribution and pre- relevance to the goals--such as Integrated Management of Childhood Illness107,108--provides a good example of spe- scription,drug availability will improve only marginally. Where drugs are not in facilities, part of the reason-- cific training unlikely to be especially valued in industrial sometimes a major reason--is that providers have little countries. incentive to have them there. Surveys in Uganda suggest Realigning the skills mix represents another option. that on average about 70 percent of medical supplies and Many interventions for the Millennium Development drugs in public facilities were appropriated by staff for use Goals can be delivered by relatively low-skilled providers, in their private work.109­111 Where providers have an such as community health workers. In addition to being incentive to have the drugs available, they often are. The cheaper, these providers are more likely to be willing to concern then is that they will prescribe inappropriate or work in rural areas and less likely to lured away by the pri- poor-quality drugs. Limiting this through regulation and vate sector, urban provider organizations, or foreign health behavior change is not impossible, but it is not easy either. sectors. There may be scope in changing the incentive arrange- Ensuring appropriate and affordable ments in the distribution of drugs--by contracting out the medicines and other health supplies distribution of medicines to the private sector,for example. Information asymmetries and drug regulation Medicines and other commodities are key components of the arsenal of effective interventions against child and Health education in various forms (providing information maternal mortality, and against communicable disease on dosages and how to administer to patients) and building mortality. Here, too, are major concerns. trust between patients and prescribers have helped increase 14 | The Millennium Development Goals for Health patient compliance and reduce high levels of self-medica- the capacity to manufacture medicines could still use the tion.112 A mix of interactive group discussions with mothers, Doha compulsory licensing opportunities by contracting training seminars for various providers at the community with foreign firms. So far, the provision has been cited to level, and districtwide monitoring has reduced the "irra- increase the supply only of antiretroviral therapy.An addi- tional"use of injections for children.113 But behavior change tional route to making antiretrovirals affordable is the programs directed at providers and retailers will have only Accelerating Access Initiative, launched by five UN agen- limited effects if they have strong economic incentives not to cies and five pharmaceutical companies to reduce the high be responsible in their prescribing behavior. cost of antiretroviral drugs and increase access to All countries require some form of drug regulatory HIV/AIDS care and treatment. One way out of the framework. It is unreasonable and largely inappropriate to impasse on neglected diseases is to separate research and expect developing countries to set up a drug regulatory development from drug manufacturing and sales.114 authority with the resources and capacity of, say, the U.S. Industrial countries, donors, or foundations would commit Food and Drug Administration or the European Medicines to purchasing--for a sizable fee--the patent resulting from Agency. Policymakers can, however, usefully explore, iden- the development of a major new vaccine or drug and then tify, and set up a drug regulatory authority that implements make the patent available freely to drug manufacturers. the most essential functions in an efficient, transparent, and Another innovative idea is the global public-private part- affordable way. nership known as the International AIDSVaccine Initiative, a public-private partnership supporting and speeding vac- Policies toward drug prices and drug spending cine development in order to expand the number and Drug costs are often not covered by insurance at all, or quality of new vaccines. they are covered only partially.This reflects both the lack of insurance in the developing world and the tendency to Strengthening core public health functions exclude at least some (usually outpatient) drug costs from schemes that do exist. The exclusion of outpatient drug Vulnerable populations need to be informed, educated, costs makes little sense, since it deters people from both and protected from risks and damages. Public health regu- taking preventive care and seeking care from a low-level lations need to be established and enforced. Infrastructure provider as soon as they fall ill. They may well get even needs to be in place to reduce the impact of emergencies sicker later and end up in the hospital, where they incur and disasters on health.All this needs to be done through a large inpatient bills against which they are fully covered. public health system that is transparent and accountable. Governments have ways of influencing drug prices-- Developing country governments generally recognize both the prices they (and ultimately taxpayers) pay when that these public health functions are important, but they they subsidize the cost and the prices consumers pay out of often lack the capacity and financial resources to imple- pocket.They can exert a direct influence over retail prices ment them. Indeed, few low-income countries invest in by regulating them (through fixed prices, risk-sharing these public health functions. agreements, and reference-based pricing schemes). They National strategies for disease prevention, can also influence retail prices indirectly through their treatment, and control policies toward the domestic pharmaceutical industry.And they can make large bulk purchases from manufacturers By employing skilled public health professionals, the gov- and then sell drugs at wholesale prices to the private non- ernment can develop and enforce standards, monitor the profit sector.Whether governments sell the drugs they pur- health of communities and populations, and emphasize chase or keep them for use or sale in their own facilities, health education, public information, health promotion, they can exert a major influence over drug prices by and disease prevention. Public action can help improve engaging in strategic purchasing--purchasing from abroad, consumer knowledge and change attitudes so that private pooling procurement efforts with other groups (including markets can operate effectively to meet the needs of the international agencies),and focusing on generics. poor--through, for example, the social marketing of insecticide-treated bednets to reduce transmission of Research and development and intellectual property malaria or condoms to reduce transmission of HIV. The lengthening of patents under the Trade-Related Government-led monitoring and evaluation Aspects of International Property Rights agreement led to concerns that new drugs would become even less afford- Integrated disease surveillance, program assessment, and able to developing countries. In August 2003 it was agreed collection and analysis of demographic and vital registra- that countries facing public health emergencies without tion data are essential if governments and donors are to The Challenges | 15 ascertain whether policies and programs are having an Indoor air pollution is caused by the use of low-cost, impact on the health goals. Chapter 2 presents a list of traditional energy sources, such as coal and biomass (wood intermediate indicators and"proxies"for the goals that can and cow dung) for cooking and heating, the main source help monitor progress, test the impact of policies, and of energy for about 3.5 billion people.Indoor air pollution adjust programs going forward.There is a need for much is a major risk factor for pneumonia and associated deaths greater investments in systems to monitor these interme- in children and for lung cancer in women who risk expo- diate indicators. Disease surveillance then helps determine sure during cooking. Projects in China, Guatemala, and whether health outcomes are improving. India are under way to improve access to efficient and Some good practices in surveillance are being devel- affordable energy sources through local design, manufac- oped, in Brazil and elsewhere. But not all developing turing, and dissemination of low-cost technologies, mod- countries can afford to invest in the infrastructure required ern fuel alternatives, and renewable energy solutions.118 for strong surveillance systems. Most rely on alternative The community-based project in China was initiated by short- to medium-term solutions for data gathering, such the health sector,troubled by the leveling off of child mor- as intermittent household surveys, health facility surveys, tality reductions among the rural poor. and simplified facility-based routine reporting. A few Agricultural policies and practices influence food countries have made special efforts to improve the surveil- prices, farm incomes, diet diversity and quality, and house- lance of a specific intervention, such as tuberculosis treat- hold food security. Policies that focus on women's access ment or immunization, while others attempt to monitor to land, training, and agricultural inputs; on their roles in progress toward a specific Millennium Development Goal. production; and on their income from agriculture are Some governments are explicitly developing or modifying more likely to have a positive impact on nutrition than their monitoring and evaluation framework to focus on policies that do not specifically focus on women, particu- the goals. larly if combined with other strategies, such as women's education and behavior change.119,120 Intersectoral actions--going beyond the ministry of health Financing additional spending for the Significant potential exists for intersectoral synergies in goals--in a sustainable way meeting the Millennium Development Goals. Roads and transport are vital for health services, especially for reduc- Additional health spending will be required in many ing maternal mortality. But it is not just physical infra- countries to accelerate progress toward the health goals. structure that matters.Also important are the availability of But how should this extra spending be financed? transportation and the affordability of its use.115 Transport Encouraging risk-pooling and private spending and roads are complementary to health services.A 10-year study in Rajasthan, India, found that better roads and Health spending can be broken down into private (out- transport helped women reach referral facilities, but many of-pocket expenditures and private insurance), public women died anyway because there were no corresponding (expenditures financed out of general revenues and social improvements at household and facility levels.116 insurance contributions), and external sources (develop- Improved hygiene (hand-washing) and sanitation (using ment assistance). latrines, safely disposing of children's stools) are at least as Private spending absorbs a larger share of income in important as drinking water quality in shaping health out- poorer countries. In low-income countries it absorbs a comes, specifically on reducing diarrhea and associated larger share of GDP, on average, than domestically child mortality.117 Constructing water supply and sanita- financed public spending. And in low-income and lower tion facilities is not enough to improve health outcomes-- middle-income countries, private spending invariably sustained human behavior change must accompany the means out-of-pocket expenditures, not private insur- infrastructure investment. In collaboration with other sec- ance.50 This leaves many near-poor households heavily tors, the health sector can develop public health promo- exposed to the risk of impoverishing health expenses.The tion and education strategies and implement them in risk is clearly greater the poorer the country, since poorer collaboration with agencies that plan, develop, and manage countries tend to have larger shares of poor people.121 A water resources.The health sector can also work with the country's private share of health spending in GDP may private sector to manufacture, distribute, and promote not in practice be related to its per capita income. But on affordable in-home water purification solutions and safe poverty-reduction grounds, there are good reasons to wish storage vessels--and advocate for water, sanitation, and that it were.Governments thus have a major role to play in hygiene interventions in poverty-reduction strategies. helping shape effective risk-pooling mechanisms. 16 | The Millennium Development Goals for Health Getting governments to spend what they can afford countries. Development assistance for health is especially important in Sub-Saharan Africa: in all countries in the Government spending as a share of GDP is higher in region, external funding exceeded 35 percent of total richer countries. But at any given per capita income, there health expenditures in 2000.124 is a surprising amount of variation across countries in the Development assistance is not, however, without its share of GDP allocated to government health programs. drawbacks.There is no assurance that development assis- Countries that appear able to spend similar shares of GDP tance to the health sector will continue to grow. Many on government health programs end up spending quite donors require that assistance be kept in parallel budgets different amounts. outside the ministries of finance, which eliminates the How can extra domestic resources be mobilized where possibility of appropriate planning and targeting of countries are spending less than they can afford to? expenditures. Such off-budget expenditures make it Domestically financed government health spending comes impossible to properly target resources to particular inter- from general revenues, social insurance contributions, or ventions, geographic locations, or population groups.Yet both. The amount of general revenues flowing into the such targeting may be essential to improving the impact health sector is the product of the amount of general (tax of expenditures on outcomes and the probability of and nontax**) revenues collected by the government (the reaching the health goals. Last and most important, com- general revenue share) and the share of general revenues mitments of expenditures in health must be permanent, allocated to the health sector (the health share of govern- implying that any external financing must at some point ment spending).122 Low government health spending be substituted by additional domestic revenues or expen- could be due to either or both being low. In poorer coun- diture reallocations. tries both shares are typically lower than they are in richer countries. But there are differences across countries that cannot be explained by per capita income alone. What the development community Countries need to ascertain whether their low spending needs to do to rise to the challenges is due to unduly low general revenues or to unduly low allocations to health and explore ways of making appro- With the advent of new funding sources for health in priate adjustments. Bolivia managed to raise its general 2000­02--including the Gates Foundation; the Global revenue share in the 1990s, as the result of a sustained Fund to Fight AIDS,Tuberculosis and Malaria; the special reform process begun in 1983.The health sector there has U.S. financing for HIV/AIDS; and the use of World Bank been one of the beneficiaries of this growth of tax rev- International Development Association (IDA) grants-- enues: government health spending as a share of GDP development assistance to health from all external sources grew at an annual rate of nearly 10 percent in the 1990s.123 rose from an average of $6.4 billion in 1997­99 to about Raising domestic resources takes time. This does not $8.1 billion in 2002. What has this money achieved in mean, though, that countries that can apparently afford to recent years? Can it be spent in a way that will achieve a spend more out of their own resources should not be greater impact on the Millennium Development Goals? encouraged to start the process. Development agencies What are some of the lessons that can be carried forward have a role to play here--by providing technical support as the external financing envelope expands? of tax reform, helping develop government commitment to health in public expenditure allocations, and providing financial assistance, both to ease the adjustment costs and Learning the lessons of development to provide support while the gap between current and assistance for health affordable spending is being closed. Development assistance to health works-- Recognizing the limits of development assistance in a good policy environment Official development assistance tends to account for a Recent research suggests that development assistance to larger share of government health spending in poorer health does lead to better health outcomes.20,21 But it does not do so in countries where the policy environment is **Some countries with large public sector companies, especially in the poor.21 The productivity of aid is not a black and white Middle East and North Africa, South Asia, and Sub-Saharan Africa regions, issue--there are gradations of good policy, and as policy have important nontax revenues, which can represent as much as 9 percent gets better, the productivity of aid increases.This finding, of GDP. Increasing revenues from this source would require consideration of corroborated by the World Bank's experience with pro- the competitive environment in which the companies operate and the need for reinvestment in the companies. jects, is influencing IDA allocations. The Challenges |17 Aid can help improve health policies-- recipient countries can be huge and that individual project but only under certain conditions management units have not made sustainable contribu- tions. Donor-funded units have sometimes run parallel to Tying aid to the adoption of policy changes through con- local structures, fostered a sense that the project staff were ditions is widely used, but recent studies cast doubt on its accountable to the financier rather than the government, wisdom.125 If governments are committed to reform, con- and redirected the most qualified human resources away ditions can help by enabling governments to publicly from government employment toward employment in commit to certain reforms and thereby persuade private development assistance agencies. investors of their seriousness. But if governments are not committed to reform, conditions will not make them There is need to enhance coordination, reform, not least because disbursements often continue explicitly pool aid, and put countries in the driver seat even when the conditions are not met. Donors cannot There is a growing view that if aid is indeed fungible and force policies on governments, but they can help in policy earmarking imposes transaction costs on recipient coun- design. tries, donors should dispense with the fiction that they can Donors have recently begun to use innovative financing identify what their money buys.This view has encouraged a mechanisms to improve performance by linking disburse- search for broader development assistance mechanisms that ments to specific performance measures, including better recognize the importance of the entire expenditure pro- policies.This is the tack taken by the Global Fund to Fight gram. These mechanisms range from the Multi-Country AIDS,Tuberculosis and Malaria and by the Global Alliance AIDS Program in Africa to sectorwide approaches in health for Vaccines and Immunization, which disburses funds to and Poverty Reduction Support Credits that back a broad countries using a per capita payment for each additional public spending agenda. child fully vaccinated against a target schedule. Recent Several key principles are emerging: programmatic social adjustment loans to Brazil and Peru by the World Bank link the disbursement of large single · Countries,not donors,need to drive the coordination. tranches of funding to changes in key policies in the health · Poverty Reduction Strategy Papers and the health sec- sector to improve the targeting of public spending toward tor analysis that feeds into them can help achieve the poor. Performance-based lending is also at the heart of strategic coherence. the new scheme for IDA credit buy-downs for polio. · Medium-term expenditure frameworks and agree- There is substantial fungibility in ments that all donor funding will respect the overall development assistance for health spending plans and limits of the government can help Recent research suggests that aid is fungible--across sec- achieve financial coherence. tors and within sectors.125 This implies that when aid is · Donor funds are best pooled into a single account, and earmarked for primary health services and excludes ter- aid is best untied from procurement only from the tiary care, governments simply focus their resources on donor country. health services for the population served by public hospi- tals--a wealthier, urban population in many poor coun- · Programs have greater impact when the number of tries.This suggests that donors should not try to channel country coordination bodies is limited and when a their external funding to specific programs without common reporting and progress assessment frame- engaging in a dialogue with the government on basic work is used, with a strong focus on countries doing changes in the overall patterns of public spending for the monitoring and evaluation, learning from it, and health.If these changes occur,donors may be able to trans- using the information to make their programs more fer their financial assistance to the health sector as a whole, effective. knowing that they are likely to have a positive impact on Millennium Development Goal outcomes. Global partnerships can add value, but they involve risks The many global initiatives and partnerships in the health The transactions costs of aid are still too high sector speak different technical languages, have vastly dif- More than 20 donors involved in health--including bilat- ferent resource bases, and target different risk groups. But erals, multilaterals, global programs, foundations, and large they all seek to add value through a common array of NGOs--can operate in a single low-income country. functions, including national and global coordination, Donors are starting to acknowledge that the demands on strategy development and evaluation, global financing and 18 | The Millennium Development Goals for Health delivery mechanisms and new tools, and resource mobi- Applying the lessons lization, social mobilization, and advocacy. to the World Bank's work Recent external evaluations of some of the major health partnerships, such as Stop TB and the Global A review of recent trends and changes in the Bank's prin- Alliance forVaccines and Immunization, suggest that these cipal instruments suggests some encouraging develop- collaborations are adding significant value in these areas. ments in the way the institution is responding to the But several challenges call for caution in embarking on Millennium Development Goal challenge. While much new global partnerships for health instead of concentrat- more needs to be done in the coming years, some positive ing on improving the effectiveness of some of the existing signs are already visible. partnerships. Some partnerships lack strategic focus and Analyzing Millennium Development Goal trends, try to do too much. Others fail to engage adequately with prospects, and challenges country processes. Partnerships can easily ignore or even exacerbate the problems afflicting the entire health system, A study of the causes of the large interstate disparities in such as the lack of human resources for service delivery Millennium Development Goal outcomes in India con- and weaknesses in integrated monitoring and evaluation. cludes that the impact of additional public spending for improved child health and nutrition would be greatest in Getting funds to the frontline the poorer Indian states.28 It recommends steps to improve Central government funds can easily leak as they move the efficiency of public expenditures, by targeting immu- through the system to the periphery of the country. And nization and community-based nutrition activities and vil- in the absence of local initiative and the right incentives, lages and districts in the poorer states, where health and service provision can fail to reflect the views of local peo- nutrition indicators are the worst. In Egypt a recent ple. Effective development assistance for health needs to Millennium Development Goal analysis also found that channel technologies, ideas, finance, and technical assis- targeting publicly financed services and involving civil tance closer to households, health providers, and supervi- society in implementing programs and monitoring sory officials in ways that are consistent with national progress for the health goals are critical for success.126 policies and amenable to monitoring and reporting. The World Bank has collaborated with local analysts and Assistance is likely to be more successful if the following policymakers in Africa over the past two years to conduct are in place: studies on the causes of health trends, especially among the poorest households. Country status reports have been com- · decentralized systems of fiduciary and technical pleted for nine countries--Burkina Faso,Chad,the Gambia, management in the public sector Guinea, Malawi, Mauritania, Mozambique, Niger, and · financially sound NGOs and private providers Tanzania--and are under way in several others. · a government body equipped and charged with Helping incorporate the Millennium Development regulating the quality of public and private providers Goals in government policies and budgets · a balanced approach to community-driven develop- The findings of the country status reports have fed into ment in health, to ensure that social fund­type financ- country-led processes--for the Poverty Reduction Strategy ing for community health initiatives is sustainable Papers and the medium-term expenditure frameworks-- changing key policies in the health sector and reallocating Development assistance for health remains unpredictable public spending toward health-related services likely to have Development assistance for health depends on donor bud- an impact on the health goals. In Mauritania the results of gets, which are subject to the usual business and political the country status report and the linked marginal budgeting cycles.Assistance may go up or down yearly as a result of for bottlenecks exercise shaped the Poverty Reduction decisions by legislative bodies during budgetary processes. Strategy Paper and resulted in important policy shifts. Further work is needed in designing mechanisms that pro- Success in promoting policy change and increasing spend- vide greater assurance of sustained long-term financial ing for the health goals requires awareness and buy-in from support. The challenges are to overcome the factors that the ministries of health and finance, a strong donor coali- result in interruptions in long-term assistance, including tion, involvement of the International Monetary Fund in those stemming from changes in political leadership and budget discussions, high-quality technical assistance, and aid agency management that can lead to reneging on ear- information campaigns and the mobilization of local gov- lier agreements. ernment and civil society. The Challenges |19 Using the goals to assess World Bank An important feature in health lending is that much of it is country assistance being incorporated as health components in other sectors, such as transport,social protection,and water supply and san- To better align the Bank's country assistance strategies with itation.Of the $1.7 billion committed in fiscal 2003,some 44 the goals, informal assessments have been carried out in percent was in projects and programs outside the health sec- several regions.A portfolio ranking in Benin found that the tor.This pattern is consistent with the lessons of development Bank's activities were weakly related to the goals. It con- assistance, but it also raises new issues.The Bank's organiza- cluded that the Bank needed to do more--through tional structure and incentives make it difficult to bring Poverty Reduction Support Credits and other projects--to health specialists into project teams led by staff from other have a substantial impact on the health-related Millennium fields, and a different inhouse skills mix is required to work Development Goals.127 along these lines. Monitoring and evaluation also need to be Bank staff in the European and Central Asia region improved in order to assess such cross-sectoral projects. drafted a Millennium Development Goal "business plan," including a country-by-country analysis showing that offi- Using the Millennium Development Goals to build cial data on child mortality and malnutrition are question- monitoring and evaluation capacity able and that many countries in the region are unlikely to Effective national programs to pursue the health goals meet the goals for 2015.The business plan called for special depend on intermediate indicators to track progress--and efforts to strengthen capacity for monitoring and evalua- on national monitoring and evaluation systems. The Bank tion, align Bank projects to support achievement of the convened a meeting of technical experts in November 2001 goals,and expand multisectoral linkages. to review and agree on a framework of intermediate deter- Integrating the goals in sectorwide minants. Those determinants were published in a booklet and programmatic instruments now being used in identifying indicators for Millennium Development Targets in Poverty Reduction Strategy Papers The goals are increasingly providing the strategic underpin- and other national strategies and in developing monitoring ning of Bank assistance to countries in programs in health and evaluation systems.128 and multisectoral budget support.A recent health project in Strengthening the capacity for monitoring and evaluation the Dominican Republic used Millennium Development is taking different forms. In Albania a monitoring and evalu- Goal intermediate and outcome indicators to monitor ation template has been prepared to help four ministries, progress. The health information system was designed to including health, develop their own systems, and a special- capture data on these indicators, augmented by periodic ized advisory body on monitoring has been established. In household surveys.In Bolivia performance indicators for the Mali a health card has been developed to give government Bank-financed health project include coverage and quality officials and the public access to a snapshot of policy actions, of key child and maternal health and disease control services, health service indicators, and health outcomes for the coun- as well as changes in mortality and disease incidence. try and for income groups. One of the most important Monitoring will be at the municipal and national levels,with efforts to improve monitoring and evaluation for the the involvement of civil society, to increase local participa- Millennium Development Goals is in HIV/AIDS, and the tion and make local politicians and health care providers World Bank has been asked to take the lead role in coordi- more accountable for results.Some sectorwide health opera- nating the support to country monitoring. tions inAfrica embody a strong focus on the goals. Reorienting and increasing Bank loans and grants to achieve Millennium Development Goal outcomes Coordinating donor actions to accelerate progress toward the goals To underpin national efforts to improve health out- comes--through projects focused on specific diseases and The Framework for Action to Accelerate Progress on the population groups or through broader health sectorwide health, nutrition, and population Millennium Development approaches and other multisectoral operations--the Bank Goals, endorsed at the high-level policy meeting in May has expanded its financial commitments in health signifi- 2003 in Ottawa, Canada, holds promise for all stakeholders cantly over the past four years, consistent with the stated in coordinating their efforts.129 An important part of this goal of systematically increasing health lending from $1 framework is building stronger national health systems as a billion in fiscal 2001 to $2.2 billion in fiscal 2005. New platform for delivering essential services to the poor in lending commitments grew from $0.95 billion in fiscal pursuit of the goals.The framework lays out common prin- 2000 to $1.7 billion in fiscal 2003. ciples and describes a process for countries and donors to 20 | The Millennium Development Goals for Health work together in expanding and improving the effectiveness · Monitoring performance.Participants agreed on the need for of their investments in health systems. The framework a common set of intermediate indicators and measures of requires country actions, such as incorporating analysis of policy and institutional performance to gauge short- and Millennium Development Goal challenges and policy and medium-term progress toward the MDGs. funding gaps in Poverty Reduction Strategy Papers and simplifying donor coordination arrangements. It also calls for stronger efforts at the global level to invest in key public References goods with multicountry benefits, such as supporting 1. UNICEF. 2001. Progress Since theWorld Summit for Children:A research and development on new drugs and vaccines for Statistical Review. NewYork: UNICEF. AIDS, tuberculosis, and malaria and documenting and shar- 2. Food and Agricultural Organization of the United Nations. ing successful national efforts to achieve the goals for mal- 2000. The State of Food Security in theWorld. Rome: FAO. nutrition and maternal and child health. 3. UNAIDS. 2002. Report on the Global HIV/AIDS Epidemic. The top issues identified in Ottawa included increasing Geneva: UNAIDS. human resource and related system capacity; improving 4. Sen,A. 2002."Why Health Equity?" Health Economics 11 (8): 659­666. management and strategic planning;creating efficient health 5. Wagstaff,A., and E.Van Doorslaer. 2003."Catastrophe and information systems; providing safe, affordable, and pre- Impoverishment in Paying for Health Care:With Applications dictable supplies of drugs; understanding the social and cul- toVietnam 1993­1998." Health Economics 12 (11): 921­34. tural determinants of health, illness, and care-seeking 6. Bloom, D., and J. Sachs. 1998."Geography, Demography and behaviors and the related policy responses; and moving Economic Growth in Africa." Brookings Papers on Economic beyond the traditional public sector service delivery mode Activity 2: 207­295. to work with NGOs and the private sector in extending the 7. Grassly, N.C., K. Desai, E. Pegurri,A. Sikazwe, I. Malambo, C. Siamatowe, and D. Bundy. 2003."The Economic Impact of coverage of the interventions for core health, nutrition, and HIV/AIDS on the Education Sector in Zambia." AIDS 17 (7): population. It was decided to establish a High-Level Forum 1039­1044. on health MDGs to review progress on health goals, to 8. Bell, C., S. Devarajan, and H. Gersbach. 2003."The Long-Run monitor changes in donor commitments and behaviors in Economic Costs of AIDS:Theory and an Application to South Africa."World Bank,Washington, DC. moving toward better harmonization, and to act on issues 9. United Nations. 2004. Millennium Development Goals. holding back progress. www.un.org/millenniumgoals. The first High-Level Forum for health MDGs, held in 10. World Bank. 2003. Global Economic Prospects and the Developing January 2004 in Geneva, addressed several of these issues. Countries. Washington, DC:World Bank. Heads of development agencies, bilateral agencies, global 11. Jones, G., R.W. Steketee, R.E. Black, Z.A. Bhutta, and S.S. health initiatives, and ministers of finance and health met Morris. 2003."How Many Child Deaths CanWe PreventThis Year?" Lancet 362 (9377): 65­71. to informally discuss concrete actions to accelerate and 12. World Health Organization. 2001. Macroeconomics and Health: monitor progress toward the MDGs for health and nutri- Investing in Health for Economic Development. Report of the tion.They agreed on actions in four areas: Commission on Macroeconomics and Health. Geneva:WHO. 13. Filmer, D., and L. Pritchett. 1999."The Impact of Public · Resources for health and poverty reduction strategy papers. Spending on Health: Does Money Matter?" Social Science and Countries should have a single process leading to Medicine 49 (10): 1309­1323. one "MDG-responsive" Poverty Reduction Strategy 14. Filmer, D., J. Hammer, and L. Pritchett. 2000."Weak Links in Paper. Participants agreed on the importance of the Chain:A Diagnosis of Health Policy in Poor Countries." World Bank Research Observer 15 (2): 199­224. incorporating explicit reference to progress toward 15. Filmer, D., J. Hammer, and L. Pritchett. 2002."Weak Links in the MDGs in joint assessments of Poverty Reduction the Chain II:A Prescription for Health Policy in Poor Strategies. Countries." World Bank Research Observer 17 (1): 47­66. 16. Bidani, B., and M. Ravallion. 1997."Decomposing Social · Aid effectiveness and harmonization. A working group Indicators Using Distributional Data." Journal of Econometrics would follow up on country pilots in budget support 77 (1): 125­139. for health and link with other broader development 17. Gupta, S., M.Verhoeven, and E.R.Tiongson. 2003."Public efforts on harmonization,drawing out lessons on com- Spending on Health Care and the Poor." Health Economics 12 (8): 685­696. mon methods and instruments for health. 18. Wagstaff,A. 2003."Child Health on a Dollar a Day: Some · Human resources. Participants agreed to assess expen- Tentative Cross-Country Comparisons." Social Science and Medicine 57 (9): 1529­1538. ditures on human resources and conduct in-depth 19. Rajkumar,A., andV. Swaroop. 2002."Public Spending and human resources, and learning by doing studies in Outcomes: Does Governance Matter?" Policy Research selected countries. Working Paper 2840,World Bank,Washington, DC. The Challenges |21 20. Feyzioglu,T.N.,V. Swaroop, and M. Zhu. 1996."Foreign Aid's 38. Schellenberg, J.R., S.Abdulla, R. Nathan, O. Mukasa,T.J. Impact on Public Spending." Policy ResearchWorking Paper Marchant, N. Kikumbih,A.K. Mushi, H. Mponda, H. Minja, H. 1610,World Bank,Washington, DC. Mshinda, M.Tanner, and C. Lengeler. 2001."Effect of Large- 21. Burnside C., and D. Dollar. 2000."Aid, Growth, the Incentive Scale Social Marketing of Insecticide-Treated Nets on Child Regime and Poverty Reduction." In TheWorld Bank: Structure Survival in RuralTanzania." Lancet 357 (9264): 1241­1247. and Policies, eds. C.L. Gilbert and D.Vines, 210­227. 39. Abdulla,S.,J.A.Schellenberg,R.Nathan,O.Mukasa,T.Marchant, Cambridge: Cambridge University Press. T.Smith,M.Tanner,and C.Lengeler.2001."Impact on Malaria 22. World Bank. 1999. Peru: Improving Health Care for the Poor. Morbidity of a Programme Supplying Insecticide-Treated Nets in Washington, DC:World Bank. ChildrenAged under 2Years inTanzania:Community Cross- Sectional Study."British Medical Journal 322 (7281):270­273. 23. Ensor,T.,A. Hossain, Q.Ali, S. Begum, and A. Moral. 2001. 40. Armstrong Schellenberg, J.,A. Mushi, H. Mponda, and "Geographic Resource Allocation in Bangladesh." Research C. Lengeler. 2002."DiscountVouchers forTreated Nets in Paper 21, Ministry of Health and FamilyWelfare, Health Tanzania:Targeted Subsidy for Malaria Control." Presentation Economics Unit, Dhaka. at meeting on malaria and equity organized by theWorld Bank 24. Diderichsen, F., E.Varde, and M.Whitehead. 1997."Resource and the London School of Hygiene andTropical Medicine, Allocation to Health Authorities:The Quest for an Equitable London School of Hygiene andTropical Medicine, London. Formula in Britain and Sweden." British Medical Journal 315 41. Saadah, F., M. Pradhan, and R. Sparrow. 2001."The (7112): 875­878. Effectiveness of the Health Card as an Instrument to Ensure 25. Newman, J.L., M. Pradhan, L. Rawlings, G. Ridder, R. Coa, and Access to Medical Care for the Poor During the Crisis."World J.L. Evia. 2002."An Impact Evaluation of Education, Health, Bank,Washington, DC. andWater Supply Investments by the Bolivian Social 42. Mesoamerica Nutrition ProgramTargeting Study Group. 2002. Investment Fund." World Bank Economic Review 16 (2): 241­274. "Targeting Performance of Three Large-Scale, Nutrition- 26. Lambrechts,T., J. Bryce, andV. Orinda. 1999."Integrated Oriented Social Programs in Central America and Mexico." Management of Childhood Illness:A Summary of First Food Nutrition Bulletin 23 (2): 162­174. Experiences." Bulletin of theWorld Health Organization 77 (7): 43. International Food Policy Research Institute. 2002. Final 582­594. Report: Nicaragua Social Protection Network, Pilot Evaluation 27. Santos, I., C.G.Victora, J. Martines, H. Goncalves, D.P. Gigante, System, Impact Evaluation. Washington, DC: IFPRI. N.J.Valle, and G. Pelto. 2001."Nutrition Counseling Increases 44. International Food Policy Research Institute. 2003. Sexto Weight Gain among Brazilian Children." Journal of Nutrition Informe. Proyecto PRAF/BID Fase II: Impacto Intermedio. 131 (11): 2866­2873. Washington, DC: IFPRI. 28. World Bank. 2003."Attaining the Millennium Development 45. International Food Policy Research Institute. Estudo De Goals in India: How Likely andWhatWill ItTake?" Avaliação De Impacto Para O Programa Bolsa Alimentação. Relatório Washington, DC. 3: Ánalise De Impacto Final.Washington, DC: IFPRI. 29. Soucat,A.,W.Van Lerberghe, F. Diop, S. Nguyen, and R. 46. Gertler, P., and S. Boyce. 2001."An Experiment in Incentive- Knippenberg. 2002."Marginal Budgeting for Bottlenecks:A BasedWelfare:The Impact of PROGRESA on Health in New Costing and Resource-Allocation Practice to Buy Health Mexico." University of California, Berkeley CA. Results." World Bank,Washington, DC. 47. Morris,S.,R.Flores,P.Olinto,and J.Medina.2003."A Randomized 30. UNICEF andWorld Bank. 2003. Marginal Budgeting for Trial of Conditional CashTransfers to Households and Peripheral Bottlenecks: How to Reach the Impact Frontier of Health and Health Centers:Impact on Child Health and Demand for Health Nutrition Services and Accelerate Progress towards the MDGs:A Services."Paper presented at the Fourth International Health Budgeting Model and Application to Low Income Countries. New EconomicsAssociationWorld Congress,San Francisco. York: UNICEF;Washington, DC:World Bank. 48. Jalan, J., and M. Ravallion. 2001."Behavioral Responses to Risk 31. Pritchett, L., and L.H. Summers. 1996."Wealthier Is Healthier." in Rural China." Journal of Development Economics 66 (1): 23­49. Journal of Human Resources 31 (4): 841­868. 49. Jalan, J., and M. Ravallion. 1999."Are the Poor LessWell 32. Haddad, L., H.Alderman, S.Appleton, L. Song, andY.Yohannes. Insured? Evidence onVulnerability to Income Risk in Rural 2003."Reducing Child Malnutrition: How Far Does Income China." Journal of Development Economics 58 (1): 61­81. GrowthTake Us?" World Bank Economic Review 17 (1): 107­131. 50. Musgrove, P., R. Zeramdini, and G. Carrin. 2002."Basic 33. Case,A. 2001."Does Money Protect Health Status? Evidence Patterns in National Health Expenditure." Bulletin of theWorld from South African Pensions." Center for Health andWellbeing Health Organization 80 (2): 134­142. Working Paper,WoodrowWilson School, Princeton University, 51. Yip,W., and P. Berman. 2001."Targeted Health Insurance in a Princeton, NJ. Low Income Country and Its Impact on Access and Equity in 34. Case,A. 2001."Health, Income, and Economic Development." Access: Egypt's School Health Insurance." Health Economics 10 Paper presented at the AnnualWorld Bank Conference on (3): 207­220. Development Economics,Washington, DC. 52. Beegle, K., E. Frankenberg, and D.Thomas."Bargaining Power 35. Aye, M., F. Champagne, and A.P. Contandriopoulos. 2002. within Couples and Use of Prenatal and Delivery Care in "Economic Role of Solidarity and Social Capital in Accessing Indonesia." Studies in Family Planning 32 (2): 130­146. Modern Health Care Services in the Ivory Coast." Social Science 53. Pitt,M.M.,S.R Khandker,S.M.Mckernan,and M.Abdul Latif. and Medicine 55 (11): 1929­1946. 1999."Credit Programs for the Poor and Reproductive Behavior 36. Alderman, H., and V. Lavy. 1996."Household Responses to in Low-Income Countries:Are the Reported Causal Relation- Public Health Services: Cost and QualityTradeoffs." World Bank ships the Result of Heterogeneity Bias?"Demography 36 (1):1­21. Research Observer 11 (1): 3­22. 54. Koblinsky, M.A., and O. Campbell. 2003."Factors Affecting the 37. Lewis, M. 2000."Who Is Paying for Health Care in Europe and Reduction of Maternal Mortality." In Reducing Maternal Central Asia?" Report 20940,World Bank,Washington, DC. Mortality: Learning from Bolivia, China, Egypt, Honduras, Indonesia, 22 | The Millennium Development Goals for Health Jamaica, and Zimbabwe, ed. M.A. Koblinsky, Xiv, 132. 74. Behrman, J., and B.Wolfe. 1987."How Does Mother's Washington, DC:World Bank. Schooling Affect Family Health, Nutrition, Medical Care 55. Das,J.,and C.Sánchez­Párano.2003."Short But Not Sweet : Usage and Household Sanitation?" Journal of Econometrics 36 (1): New Evidence on Short Duration Morbidities from India."Policy 185­204. ResearchWorking Paper 2971,World Bank,Washington,DC. 75. World Bank. 2003.World Development Report 2004: Making 56. Jalan, J., and M. Ravallion. 2003."Does PipedWater Reduce ServicesWork for Poor People. Washington, DC:World Bank. Diarrhea for Children in Rural India?" Journal of Econometrics 76. Victora, C., F. Barros, J.Vaughan,A. Silva, and E.Tomasi. 2000. 112 (1): 153­173. "ExplainingTrends in Inequities: Evidence from Brazilian 57. Desai, S., and S.Alva. 1998."Maternal Education and Child Child Health Studies." Lancet 356 (9235): 1093­1098. Health: IsThere a Strong Causal Relationship?" Demography 35 77. Eichler, R. 2001."Improving Immunization Coverage in an (1): 71­81. Innovative Primary Health Care Delivery Model: Lessons from 58. Glewwe, P. 1999."Why Does Mother's Schooling Raise Child Burkina Faso's Bottom Up Planning, Oversight, and Resource Health in Developing Countries?" Journal of Human Resources Control Approach that Holds Providers Accountable for 34 (1): 124­59. Results."World Bank,Washington, DC. 59. Abbas, K.A. 1995."Counselling in a Hospital Setting. 78. Ewig, C. 2003."The Contributions of Community-Based Breastfeeding." Dialogue Diarrhoea (59): 3. Decentralization to Democracy: Peru's Local Health Administration Committees." Paper presented at the annual 60. Heerey, M., and A. Kols. 2003. Improving the Quality of Care: meeting of the American Political Science Association. Quality Improvement Projects from the Johns Hopkins University Bloomberg School of Public Health Center for Communication 79. Langenbrunner, J. 2003."Resource Allocation and Purchasing Programs. Center Publication No. 101, Johns Hopkins in the ECA Region:A Review."World Bank,Washington, DC. University, Bloomberg School of Public, Baltimore, MD. 80. Chawla, M., P. Berman,A.Windak, and M. Kulis. 1999. 61. Benefo, K., andT. Schultz. 1996."Fertility and Child Mortality "Provision of Ambulatory Health Services in Poland:A Case in Côte d'Ivoire and Ghana." World Bank Economic Review 10 Study from Krakow."World Bank,Washington DC. (1): 123­158. 81. Cotlear, D. 1999."Peru: Improving Health Care for the Poor." 62. Mwabu, G., M.Ainsworth, and A. Nyamete. 1993."Quality of LCSHD Paper Series 57,World Bank, Human Development Medical Care and Choice of MedicalTreatment in Kenya:An Department,Washington, DC. Empirical Analysis." Journal of Human Resources 28 (4): 838­862. 82. Vladescu, C., and S. Radulescu. 2001."Primary Health Services: 63. Lavy,V., J. Strauss, D.Thomas, and P. DeVreyer. 1996."Quality of Output-Based Contracting to Lift Performance in Romania." Care, Survival and Health Outcomes in Ghana." Journal of Public Policy for the Private Sector 239. http://rru.worldbank.org/ Health Economics 15 (3): 333­357. viewpoint/HTMLNotes/239/239Vlade-831.pdf. 64. Thomas, D.,V. Lavy, and D. Strauss. 1996."Public Policy and 83. Bitran, R. 2001."Paying Health Providers through Capitation Anthropometric Outcomes in the Côte d'Ivoire." Journal of in Argentina, Nicaragua andThailand: Output, Spending, Public Economics 61 (2): 155­192. Organizational Impact and Market Structure." USAID PHR Project,Washington, DC. 65. Wong, E., B. Popkin, D. Guilkey, and J.Akin. 1987. "Accessibility, Quality of Care and Prenatal Care Use in the 84. Eichler, R., P.Auxilia, and J. Pollock. 2001."Output-Based Philippines." Social Science and Medicine 24 (11): 927­944. Health Care: Paying for Performance in Haiti."World Bank, 66. Van DeWalle, D., and D. Cratty. 2002."Impact Evaluation of a Washington, DC. Rural Road Rehabilitation Project."World Bank,Washington, 85. Marek,T., I. Diallo, B. Ndiaye, and J. Rakotosalama. 1994. DC. "Successful Contracting of Prevention Services: Fighting Mal- 67. Pathmanathan, I., J. Liljestrand, J.M. Martins, L.C. Rajapaksa, nutrition in Senegal and Madagascar." Health Policy Planning 14 C. Lissner,A. De Silva, S. Selvaraju, and P.J. Singh. 2003. (4): 382­389. "Investing in Maternal Health: Learning from Malaysia and Sri 86. Tangcharoensathien,V., and N. Khongsawatt Nas. 1997. Lanka."World Bank, Health, Nutrition, and Population "Private-Sector Involvement in Public Hospitals: Case-Studies Department,Washington, DC. in Bangkok." In Private Health Providers in Developing Countries: 68. Merrick,T.W."The Effect of PipedWater on Early Childhood Serving the Public Interest? ed.A. Mills. London: Zed Books. Mortality in Urban Brazil,1970 to 1976."Demography 22 (1):1­24. 87. Mcpake,B.,and C.Hongoro.1995."Contracting Out of Clinical 69. Ridder, G., and I.Tunali. 1999."Stratified Partial Likelihood Services in Zimbabwe."Social Science and Medicine 41 (1):13­24. Estimation." Journal of Econometrics 92 (2): 193­232. 88. Slack, K., andW. Savedoff. 2001."Public Purchaser-Private 70. Lee, L­F., M. Rosenzweig, and M. Pitt. 1997."The Effects of Provider Contracting for Health Services: Examples from Latin Improved Nutrition, Sanitation, andWater Quality on Child America and the Caribbean." Sustainable Development Health in High-Mortality Populations." Journal of Econometrics DepartmentTechnical Paper Series 111, Inter-American 77 (1): 209­235. Development Bank,Washington, DC. 71. Esrey, S.A., and J.P. Habicht. 1988."Maternal Literacy Modifies 89. Fiedler, J.L. 1996."The Privatization of Health Care inThree the Effect ofToilets and PipedWater on Infant Survival in Latin American Social Security Systems." Health Policy Planning Malaysia." American Journal of Epidemiology 127 (5): 1079­1087. 11 (4): 406­417. 72. Jalan, J., and M. Ravallion. 2001."Does PipedWater Reduce 90. Mills,A., S. Bennett, and S. Russell. 2001. The Challenge of Diarrhea for Children in Rural India?" Policy Research Health Sector Reform:What Must Governments Do? NewYork: Working Paper 2664,World Bank,Washington, DC. St Martin's Press. 73. Wolfe, B., and J. Behrman. 1982."Determinants of Child 91. Gilson, L., and A. Mills. 1995."Health Sector Reforms in Sub- Mortality, Health and Nutrition in a Developing Country." Saharan Africa: Lessons of the Last 10Years." Health Policy 32 Journal of Development Economics 11 (1): 163­193. (1­3): 215­243. The Challenges | 23 92. Londoño, B, I. Jaramillo, and J.P. Uribe. 1999. 112. Cruz, O., K. Hanson, and A. Mills. 2001."Approaches to "Decentralization and Reforms in Health Services:The Overcoming Health Systems Constraints at the Peripheral Colombian Case."World Bank,Washington, DC, Level:A Review of the Evidence."WG5 15, CMHWorking 93. Oliveira-Cruz,V., C. Kurowski, and A. Mills. 2003."Delivery Paper Series.World Health Organization, Geneva. of Priority Health Services: Searching for Synergies within 113. Santoso, B., S. Suryawati, and J.E. Prawaitasari. 1996."Small theVerticalVersus Horizontal Debate." Journal of International Group Intervention vs. Formal Seminar for Improving Development 15 (1): 67­86. Appropriate Drug Use." Social Science and Medicine 42 (8): 94. Raufu,A. 2002."Nigeria Concerned over Exodus of Doctors 1163­1168. and Nurses." British Medical Journal 325: 65. 114. Weisbrod, B. 2003."Solving the Drug Dilemma." Washington Post, August 22. 95. Federation forAmerican Immigration Reform.2002."Brain Drain."http://www.Fairus.Org/Immigrationissuecenters/ 115. Eissen, E., D. Efenne, and K. Sabitu. 1997."Community Loan Immigrationissuecenters.Cfm?ID=1242&C=17. Funds andTransport Services for Obstetric Emergencies in Northern Nigeria." International Journal of Gynecology and 96. Marcelo, R. 2003."Hospital `Angels' Look for Heaven Obstetrics 59 (Suppl. 2): 48. Elsewhere." Indian FinancialTimes, October 13. 116. Pendse,V. 1999."Maternal Deaths in an Indian Hospital:A 97. Padarath,A., C. Chamberlain, D. Mccoy,A. Ntuli, M. Rowson, Decade of (No) Change?" Reproductive Health Matters. and R. Loewenson. 2003."Health Personnel in Southern Africa: Confronting Maldistribution and Brain Drain." 117. Esry, S., J. Potash, and L. Roberts, and C Schiff. 1991."Effects EQUINET Discussion Paper 3. of ImprovedWater Supply and Sanitation on Ascariasis, Diarrhea, Dracunculiasis, Hookworm Infection, 98. Upadhyay,A. 2003."Nursing ExodusWeakens Developing Schistosomiasis, andTrachoma." Bulletin of theWorld Health World." Inter Press Service News Agency. http://www. Organization 69 (5): 609­621. Ipsnews.Net/Migration/Stories/Exodus.Html. 118. World Bank. 2003."Public Health at a Glance Fact Sheet: 99. World Health Organization. 1998. CHD 1996­97 Report. Indoor Air Pollution." Health, Nutrition, and Population Geneva:WHO. Department,Washington, DC. 100. Ministry of Health and SocialWelfare. 2002. RNTCP Perfor- 119. Quisumbing,A.R. 1995."Gender Differences in Agricultural mance Report: India. 3rd Quarter. CentralTB Division, Delhi. Productivity:A Survey of Empirical Evidence." FCND 101. Kurowski,C.,S.Abdulla,andA.Mills.2003."Human Resources Discussion Paper 5: 1­71. International Food Policy Research for Health:Requirements andAvailability in the Context of Institute,Washington, DC. Scaling-Up Priority Interventions.A Case Study fromTanzania." 120. Johnson-Welch, C. 1999."Focusing onWomenWorks: London School of Hygiene &Tropical Medicine,London. Research on Improving Micronutrient Status through Food- 102. Holmas,T.H. 2002."Keeping Nurses atWork:A Duration Based Interventions." International Center for Research on Analysis." Health Economics 11 (6): 493­503. Women/Opportunities for Micronutrient Interventions 103. Wilbulpoprasert. 2002."Integrated Strategies toTackle Project,Washington, DC. Inequitable Distribution of Doctors inThailand: Four 121. World Bank. 2000. World Development Report 2000/2001: Decades of Experience." Attacking Poverty. Washington, DC:World Bank. 104. Peters,D.H.,A.S.Yazbeck,R.R.Sharma,G.N.V.Ramana,L.H. 122. Hay, R. 2003."The `Fiscal Space' for Publicly Financed Health Pritchett,and A.Wagstaff.2002.Better Health Systems for India's Care." Oxford Policy Institute Policy Brief. Poor:Findings,Analysis,and Options. Washington:World Bank. 123. World Bank. 2003. World Development Indicators 2003. 105. Reinikka, R., and J. Svensson. 2003."Working for God? Washington, DC:World Bank. Evaluating Service Delivery of Religious Not-for-Profit 124. World Health Organization. 2000. TheWorld Health Report Health Care Providers in Uganda." Policy ResearchWorking 2000. Health Systems: Improving Performance. Geneva:WHO. Paper 3058,World Bank,Washington, DC. 125. World Bank. 1998. Assessing Aid:WhatWorks,What Doesn't, and 106. Rabinowitz, H. 1999."A Program to Increase the Number of Why. Oxford: Oxford University Press. Family Physicians in Rural and Underserved Areas: Impact 126. El-Saharty, S., E. Richardson, and S. Chase. 2003. Egypt and the after 22Years." Journal of the American Medical Association 281 Millennium Development Goals.Washington, DC:TheWorld (3): 255­260. Bank. 107. Tulloch, J. 1999."Integrated Approach to Child Health in 127. Abrantes,A. 2003. Personal communication.World Bank, Developing Countries." Lancet 354 (Suppl. 2): SII16­ SII20. Washington, DC. 108. Gove, S. 1997."Integrated Management of Childhood Illness 128. World Bank. 2002. Annual Review of Development Effectiveness: by Outpatient HealthWorkers:Technical Basis and Overview." Achieving Development Outcomes:The Millennium Challenge. WHOWorking Group on Guidelines for Integrated Washington, DC:World Bank. Management of the Sick Child. Bulletin of theWorld Health 129. Claeson, M. 2003."A Framework for Action:Accelerating Organization 75 (Suppl. 1): 7­24. Progress to Meet the HNP MDGs."World Bank, Health, 109. Reinikka, R. 1999."Using Surveys for Public Sector Nutrition, and Population Department,Washington, DC. Reform." PREM Notes.World Bank,Washington, DC. 130. World Health Organization.2003.Global Database on Child 110. Mcpake,B.,D.Asiimwe,F.Mwesigye,M.Ofumbi,L.Orthenblad, Growth and Malnutrition.www.Who.Int/Nutgrowthdb/. P.Streefland,andA.Turinde.1999."Informal EconomicActivities 131. UNICEF. 2003. Statistics: Child Mortality. of Public HealthWorkers in Uganda:Implications for Quality and www.Childinfo.Org/Cmr/Revis/Db2.Htm. Accessibility of Care."Social Science and Medicine 49 (7):849­865. 132. Claeson, M., C.G. Griffin,T.A. Johnston, and others. 2001. 111. Ablo, E., and R. Reinikka. 1998."Do Budgets Really Matter? "Health, Nutrition and Population." In Poverty Reduction Evidence from Public Spending on Education and Health in Strategy Paper Sourcebook, 2nd ed., ed.World_Bank. Uganda." Policy ResearchWorking Paper 1926,World Bank, Washington, DC:World Bank. Washington, DC. 24| The Millennium Development Goals for Health CHAPTER 1 Backdrop to the Millennium Development Goals "It started in 1999 when my brother, Geogys, went away--I that they will go away, too.And now my youngest boy, Isaac, is mean, he died at that time. He was ill so many times with also starting. He is a year old now, but it's very easy with young fever. He was 28 when he went away. After two months, his children to see how they feel with their malaria.Their temperature first-born went away too, and three months later our two sis- rises very quickly and they start to sweat.For myself,it comes and ters, Margaret and Lucy, followed them. Malaria killed all of goes. Last year, I was affected almost all year, so I could not work them. And it keeps on going. Two weeks ago our aunt died well.The headaches were very bad.But this year,it is better." from malaria. She was ill onTuesday, complaining about a bad --GEORGE OSIGA, COBBLER, KENYA (adapted from a supplement in the Guardian, headache and fever, and onWednesday she was dead. It is not "Earth:Health check for a planet and its people under pressure," unusual like that.At least with AIDS it can take you a long August 22,2002) time to die, but malaria is quick. In the past three years, 25 of my family and friends have died in this way. The extent of premature death and ill health in the My village is in the highlands. Everyone is getting malaria developing world is staggering. In 2000 almost 11 mil- nowadays. Recently, we have buried 10 people from this dis- lion children died before their fifth birthday, 99 percent ease, including my aunt, because it is so bad. Everyone knows of them in the developing world, 4.5 million in Sub- mosquitoes are the danger, but the problem is that people are Saharan Africa alone.An estimated 140 million children poor. Even if you tell them about bednets, they are thinking under five are underweight, almost half of them in about how they will eat. Nobody uses nets. Instead we build South Asia. In 2001, 3 million people died from fires in our places to try to keep the mosquitoes out with HIV/AIDS, 99 percent of them in the developing smoke. But we still get infected.When the fever comes, it's bet- world.Tuberculosis claimed another 2 million lives.And ter to find the doctor quickly and think about how to pay later. 515,000 women died during pregnancy or childbirth in It can force you to sell a cow sometimes, or maybe the doctor 1995, 98 percent of them in the developing world. will take some hens or a sheep. But sometimes the medicine Death and poor health on such a scale are matters of does not work well, so maybe it is better just to pray. concern in their own right. But they also act as a brake Me, I am worried about my wife and my second-born, on economic development. It was these twin concerns Marcy.They are always complaining of fever.I have taken them to that led the international community to put health hospital, and they have taken different medicines, but still they firmly at the center of the Millennium Development have the headaches and the fever.What should I do? I'm worried Goals when adopting them in 2001. 25 points to one of the reasons--the close connection between Health, interventions, poverty, and people health and poverty, a third theme of this report. Poverty The Kenyan cobbler in the opening story records the makes people especially vulnerable to sickness. But it also deaths of no fewer than 25 family members and friends means they have few resources to spend on preventive over three years--deaths that are far from atypical. measures and on curative care. So they are more likely to be malnourished and to die early (box 1.1). · Almost 11 million children died before their fifth As George's story illustrates, people's ability to work birthday in 2000.1 Less than 1 percent of these deaths depends on their health. As long as they remain ill, their (79,000) occurred in high-income countries, com- earning capacity is reduced, and the income flow into the pared with 42 percent in Sub-Saharan Africa, 35 per- household remains limited. George and many like him in cent in South Asia, and 13 percent in East Asia. the developing world appear to be caught in a vicious cir- · Of the estimated 140 million children under the age cle: their poverty keeps them from taking appropriate pre- of five who are underweight, almost half (65 million) ventive measures and seeking health care when sick.This are in South Asia. keeps them sick--which keeps them in poverty. If the poor spend their limited resources on preventing disease · In 1998 an estimated 843 million people were consid- and restoring themselves to good health when they fall ered to be undernourished on the basis of their food sick,they risk leaving themselves and their loved ones with intake.2 too little for food and other necessities. · Of the 3 million people who died from HIV/AIDS in Illness and death in a household reduce income and 2001,3 almost all (99 percent) were in the developing lead to potentially impoverishing expenses on medical world--73 percent in Sub-Saharan Africa alone. care. In Vietnam health expenses are estimated to have pushed about 3 million people into poverty in 1993 and · Tuberculosis claimed 2 million lives, with the epi- 2.7 million in 1998.7 Voices of the Poor records the case of a demic worsening during economic and social crises. 26-year-old man in Lao Cai, in the north ofVietnam.As a result of the high health care costs of his daughter's severe · In 1995, 515,000 women died during pregnancy or illness, he moved from being the richest man in his com- childbirth: 1,000 in the industrial world, 252,000 in Sub-Saharan Africa.1 munity to one of the poorest.8 The vicious circle of ill health and poverty is evident at Death, disease, and malnutrition on this scale are matters for con- the country level also. With a daily per capita income of cern in their own right--the first theme of this report. As $1.17 (purchasing power parity of $5.60) in 2000, it is Nobel laureate Amartya Sen put it,"Health is among the hardly surprising that coverage levels of key preventive and most important conditions of human life and a critically treatment interventions are so low in low-income coun- significant constituent of human capabilities which we tries.9 Barely 40 percent of the births in low-income coun- have reason to value."4 Put another way, being poor is not tries take place with the assistance of a medically trained simply about having too little money. To be poor, as the person.Barely 60 percent of children receive DPT3 immu- 2001 World Development Report: Attacking Poverty notes, is nization.And only a third of infectious patients with tuber- "to be hungry, to lack shelter, to be sick and not cared for, culosis are using effective treatment. These low coverage to be illiterate and not schooled."5 rates translate into high rates of child and maternal mortal- George Osiga's story illustrates a second theme of the ity and low levels of nutrition.That, in turn, acts as a brake report. For the most part, it is not the lack of effective on economic growth and contributes to income poverty. interventions that is the principal cause of the large num- Children are less likely to be in school if ill and malnour- ber of deaths in the developing world. It is that the coverage ished, and they learn less when they are in school. Their of preventive and treatment interventions is too low. For exam- productivity in later life is reduced as a result. A recent ple, insecticide-treated nets are effective in preventing study estimates that $1 invested in an early child nutrition malaria, and effective antimalarials exist for treating it.The program in the Philippines would yield at least a $3 return scope for reducing mortality by scaling up the coverage in higher earnings thanks to better educational outcomes.10 rates of effective interventions is considerable: if the utiliza- But children require teachers to learn, and in some parts of tion rates of childhood interventions were to rise from cur- Africa deaths from HIV/AIDS are making a major dent in rent levels to 99 percent, the number of under-five deaths the stock of teachers--Zambia now loses half as many worldwide would fall by an estimated two-thirds.6 teachers as it trains to HIV/AIDS.11 This begs the question: why are effective interventions Illness and malnutrition reduce the productivity of so little used in the developing world? The opening story current workers, too--as well as the hours they can 26 | The Millennium Development Goals for Health Box 1.1 Worlds apart: The poor die earlier Between the world's poorest and richest people are large differ- scales of the denominators for the rates and ratios were chosen to ences in the risk of premature death and malnutrition. Box figure ensure that each gradient is visible.) 1 lines countries up by their per capita income and divides the The under-five mortality rate among the poorest quarter of world's population into four equal groups based on the per capita the world's population is 10 times that of the richest quarter (box income of the country they live in--the world's poorest quarter, figure 1). The figure for tuberculosis mortality is only slightly the world's second poorest quarter,and so on.The figure does not lower.The figure for maternal mortality is 20 times that of the capture the fact that people within countries vary in terms of per richest quarter. Within countries too, gaps exist--and persist-- capita income and mortality risks. But box figure 2 does. (The between the rich and the poor, as box figure 2 shows.14,15 The distribution of malnutrition and mortality Poor children die earlier than other children between poor and rich countries is unequal 160 180 births 140 160 live 140 120 120 1,000 100 100 per 80 rate 80 Rate/ratio 60 60 40 mortality 20 40 0 Child Under-five Maternal AIDS Tuberculosis 20 malnutrition mortality per mortality per mortality mortality per Under-five per 100 1,000 live 10,000 live per 100,000 100,000 0 population births births population population Bolivia Egypt Indonesia Poorest 25% of world's population Poorest 20% of children 2nd poorest 25% of world's population 2nd poorest 20% of children 2nd richest 25% of world's population Middle 20% of children Richest 25% of world's population 2nd richest 20% of children Richest 20% of children Source: References 21­23. Source: Reference 14. spend working.A recent study from Indonesia finds that decide whether to use mosquito nets or build fires, whether the loss of the ability to undertake a typical activity of to seek treatment when fever comes or simply to pray. If daily living leads to a loss of earnings equivalent to 10 treatment is sought, it is a health provider who makes the percent of average earnings, while a loss of ability to diagnosis and delivers the care.Households and providers are undertake any activity of daily living wipes out earnings fallible, and their knowledge and skills are often lacking. altogether.12 The effects of poor health and mortality Patients often request inappropriate drugs and tests in the show up in macroeconomic aggregates: health and mistaken belief that they are appropriate.And providers often demographic variables have been estimated to account knowingly administer or prescribe them, since they have a for as much as half the difference in growth rates financial incentive to do so.This contributes to drug resis- between Africa and the rest of the world between 1965 tance,to poverty,and in the extreme to death. and 1990.13 Low rates of economic growth and low ini- Providers can also be ill informed, lacking even rudi- tial per capita incomes translate into low coverage of key mentary diagnostic skills. In a survey in Burundi in 1992 interventions and high rates of mortality and malnutri- only 2 percent of children with diarrhea taken to a health tion.And so the cycle goes on. facility were correctly diagnosed, and of these only 13 George Osiga's story draws attention to a fourth theme of percent were correctly rehydrated.16 Households need the report: interventions are not dropped like manna from heaven; not always be ignorant of health matters, and providers they are demanded by people and delivered by people. Households can acquire the appropriate skills. InVietnam in 1997, 78 Backdrop to the Millennium Development Goals | 27 percent of children seeking care for diarrhea were cor- Global Alliance for Vaccines and Immunization, the Stop rectly assessed, and of those 67 percent were correctly TB Partnership, Roll Back Malaria, the Global Fund to rehydrated.16Vietnam is also one of the few high-burden Fight AIDS, Tuberculosis and Malaria, and most recently countries that have already surpassed global targets for the Global Alliance for Improved Nutrition (box 1.2).The 2005 for tuberculosis treatment and cure. scale of philanthropic involvement in international health This report argues for better understanding of what also increased, with the launch of the Bill and Melinda prompts households to engage in preventive activities and Gates Foundation and the Packard Foundation and the to seek care when sick--and what makes some providers greater attention to global health issues by such established but not others deliver good-quality care. Poor households foundations as the Rockefeller Foundation. These initia- and poor countries can break out of the vicious circle tives brought new funds to the fight against disease, death, linking poor health and poverty. Improving knowledge is and malnutrition in the developing world--as well as new one avenue. Lowering prices and other barriers to using ideas, new energy, and new coordination challenges and effective interventions is another. Sri Lanka started to mechanisms. achieve dramatic reductions in its maternal mortality ratio With the 1990s drawing to a close, the international as far back as the 1930s, when it was extremely poor.17 community decided that even more needed to be done to The exercise of understanding the factors behind house- address the array of development challenges facing the hold and provider behavior has to be done in the context poor. In September 2001 at the Millennium Summit, 147 of the health system,but it also has to go beyond the health heads of states endorsed the Millennium Development system. It has to be done in the context of government Goals. Later, at the UN General Assembly, the heads of policies for the health sector, but also for other sectors.And state of all 189 UN member states adopted the goals. it has to be done in the context of development assistance and international initiatives and partnerships. The rest of the report This report focuses on the Millennium Development Health gets on the international Goals,which all governments have agreed on.Other health community's radar screen--and the outcomes are equally or even more important in many Millennium Development Goals are born countries, especially in countries going through the demo- In the 1990s the international community recognized the graphic, epidemiological, and nutritional transitions and importance of health in development. In a period when those in which noncommunicable disease and injury con- official development assistance (ODA) was declining, tribute significantly to the burden of disease.The focus in development assistance to health rose--in real terms.18 It this report is on outcomes,on households,on health system also doubled as a share of International Development constraints and public health functions, and on investments Association (IDA) disbursements,which in turn accounted in improving policies and institutions.The agenda is rele- for a larger share of ODA (27 percent in 1996­98,up from vant for all countries, whether their targets are the Millen- 25 percent in 1990­92).* nium Development Targets, other national and subnational The 1990s saw a growing clamor in the industrial world targets, or some combination of the two.All countries have over debt in the developing world, fueled in no small mea- an interest in monitoring progress using national and sure by a perception that interest payments were constrain- context-specific intermediate indicators, in buying better ing government health expenditures in developing results with their limited resources, and in accelerating countries.19 The enhanced Highly Indebted Poor Country progress toward their goals. Initiative was explicitly geared to channeling freed resources into the health and other social sectors.20 The · Chapter 2 sets out the Millennium Development Goals for health and assesses progress to date and Poverty Reduction Strategy Paper was introduced for prospects of achieving the goals. developing country governments seeking debt relief or concessional IDA loans to set out their plans for fighting · Chapter 3 argues that where progress has been slow, poverty on all fronts,including health. the reason is not a lack of effective interventions but The 1990s also saw the development of major new global inadequate use of existing interventions. health initiatives and partnerships, including UNAIDS, the · Chapter 4 introduces the report's second part (chapters 5­10). It argues that low coverage rates partly reflect the low levels of government health spending but more *The International Development Association, part of theWorld Bank Group, provides loans to the world's poorest countries at zero interest with a fundamentally the numerous policy and institutional grace period of 10 years and maturities of 35­40 years. weaknesses across the entire health system. 28 | The Millennium Development Goals for Health Box 1.2 The promising (and challenging) array of health-related partnerships This list is not comprehensive, but it gives a sense of the large International Public/Private Partnership for Health number of partnerships, most of them formed in the past seven Global AIDS Monitoring and EvaluationTeam to eight years. Many involve formal governance structures with InternationalTreatment Acceleration Coalition international institutions, donors, governments, industry, and civil society. Region-specific partnerships Communicable diseases African Program for Onchocerciasis Control International Partnership against AIDS in Africa Global Fund to Fight AIDS,Tuberculosis and Malaria UNDP/World Bank/WHO Special Program for Research & Maternal and child health and nutrition Training inTropical Diseases Joint United Nations Program on HIV/AIDS (UNAIDS) Safe Motherhood Initiative Global Alliance forVaccines and Immunization Research in Human Reproduction Program International AIDSVaccine Initiative Global Alliance for Improved Nutrition Partnership for Child Development StopTB Partnership Child Health and Nutrition Research Initiative Global Drug Facility Global Alliance forTB Drug Development (TB Alliance) Health systems strengthening and financing Fund for Investment in New Diagnostics Roll Back Malaria Alliance for Health Policy and Systems Research Multilateral Initiative on Malaria European Observatory on Health Care Systems Medicines for MalariaVenture INDEPTH Health Surveillance and Experimentation Network InternationalTrachoma Initiative · Chapter 5 shows that additional government health References spending would accelerate progress toward the health Millennium Development Goals only in countries 1. UNICEF. 2001. Progress Since theWorld Summit for Children: A Statistical Review. NewYork: UNICEF. with relatively good policies and institutions. 2. Food and Agricultural Organization of the United Nations. · Chapter 6 argues that poor progress toward the goals 2000. The State of Food Security in theWorld. Rome: FAO. reflects weak household demand for the services of 3. UNAIDS. 2002. Report on the Global HIV/AIDS Epidemic. Geneva: UNAIDS. formal providers and weak household capacity to deliver key home-based interventions. 4. Sen,A. 2002."Why Health Equity?" Health Economics 11 (8): 659­666. · Chapter 7 argues that slow progress also reflects the 5. World Bank. 2001. World Development Report 2000/2001: inefficiency and unresponsiveness of provider organi- Attacking Poverty.Washington, DC:World Bank zations, which reflects poor management and weak 6. Jones, G., R.W. Steketee, R.E. Black, Z.A. Bhutta, and S.S. Morris. 2003."How Many Child Deaths CanWe PreventThis accountability to the public. Year?" Lancet 362 (9377): 65­71. · Chapter 8 discusses the role of input markets in holding 7. Wagstaff,A., and E. van Doorslaer. 2003."Catastrophe and Impoverishment in Paying for Health Care:With Applications back progress toward the Millennium Development toVietnam 1993­98." Health Economics 12 (11): 921­933. Goals,examining human resource shortages,inadequate 8. Narayan, D., R. Patel, K. Schafft,A. Rademacher, and S. Koch- skills,and unaffordable and often inappropriate drugs. Schulte. 2000. Voices of the Poor: Can Anyone Hear Us? New York: Oxford University Press. · Chapter 9 argues that many countries have a weak 9. World Bank. 2004. World Development Report 2004: Making public health infrastructure that leaves core public ServicesWork for Poor People.Washington, DC:World Bank. health functions uncovered. 10. Glewwe, P., H.G. Jacoby, and E.M. King. 2001."Early Childhood Nutrition and Academic Achievement:A · Chapter 10 discusses the role of the World Bank and Longitudinal Analysis." Journal of Public Economics 81 issues of interagency coordination and partnerships. (3):345­368. Backdrop to the Millennium Development Goals |29 11. Grassly,N.C.,K.Desai,E.Pegurri,A.Sikazwe,I.Malambo, 17. Pathmanathan, I., J. Liljestrand, J.M. Martins, L.C. Rajapaksa, C.Siamatowe,and D.Bundy.2003."The Economic Impact of C. Lissner,A. de Silva, S. Selvaraju, and P.J. Singh. 2003. Investing HIV/AIDS on the Education Sector in Zambia."AIDS 17 (7): in Maternal Health: Learning from Malaysia and Sri Lanka. 1039­1044. Washington, DC:World Bank. 12. Gertler, P., and J. Gruber. 2002."Insuring Consumption against 18. OECD Development Assistance Committee. 2000. RecentTrends Illness." American Economic Review 92 (1): 51­76. in Official Development Assistance to Health. Paris: OECD. 13. Bloom, D., and J. Sachs. 1998."Geography, Demography and 19. Kirby,A. 1999."UK Archbishop Heads Debt Chain." BBC Economic Growth in Africa." Brookings Papers on Economic Online, June 13. Activity 2: 207­295. 20. Gupta, S., B. Clements, M.T. Guin-Siu, and L. Leruth. 2002. 14. Gwatkin,D.,S.Rutstein,K.Johnson,R.Pande,and A.Wagstaff. "Debt Relief and Public Health Spending in Heavily Indebted 2000."Socioeconomic Differences in Health,Nutrition and Poor Countries." Bulletin of theWorld Health Organization 80 (2): Population."Health,Nutrition and Population Discussion Paper, 151­157. World Bank,Washington,DC.http://www.worldbank.org/ 21. World Health Organization.2003.Global Database on Child poverty/health/data/index.htm. Growth and Malnutrition.http://www.who.int/ 15. Wagstaff,A. 2000."Socioeconomic Inequalities in Child nutgrowthdb/. Mortality: Comparisons across Nine Developing Countries." 22. UNICEF. 2003. Statistics: Child Mortality. http://www. Bulletin of theWorld Health Organization 78 (1): 19­29. childinfo.org/cmr/revis/db2.htm. 16. World Health Organization. 1998. CHD 1996­97 Report. 23. World Bank. 2003. World Development Indicators 2003. Geneva:WHO. Washington, DC:World Bank. 30 | The Millennium Development Goals for Health CHAPTER 2 The Millennium Development Goals for Health: Progress and Prospects Nearly half the Millennium Development Targets concern situation at "half time" in the 1990­2015 period for the health--directly or indirectly.They vary in their ambitiousness, goals. with the maternal mortality goal of a three-quarters reduction The"second half"may go better.Per capita incomes are fore- between 1990 and 2015 being the most ambitious. Progress cast to grow faster in all but one region (East Asia and the has varied across the goals, in part reflecting the differences in Pacific). And reaching the targets for gender and environment ambitiousness. All World Bank regions except Sub-Saharan would imply faster progress in girls' education and access to safe Africa will hit the malnutrition targets, but many countries-- drinking water than in the 1990s for most regions.This impetus even in regions where malnutrition has fallen rapidly--will may push three regions to (or close to) the maternal mortality tar- not. Progress on child survival has been insufficient to leave get.But it will still leave all regions well short of the child mortal- any Bank region or any Sub-Saharan country on track to hit ity target and Sub-Saharan Africa well short of the malnutrition the target, but 40 percent of Sub-Saharan countries accelerated target. This chapter argues that the Millennium Development their pace of reducing child mortality in the 1990s.All but one Goals are useful because they focus on outcomes and remind us of Bank region will miss the maternal mortality target,albeit by a what is achievable. What is important is to accelerate progress smaller margin than for the child mortality target.That's the toward the targets,even if the targets will not be hit. 31 Health is prominent in the Millennium What's the score at half time? Development Goals Tracking progress toward the health goals is not straightfor- It is a reflection of the international community's concern ward because of differences in how the indicators are for health in the developing world that nearly half the goals defined,which measurement instruments are used,how fre- and targets concern--directly or indirectly--different quently information is collected, and how much effort has aspects of health.* Table 2.1 sets out the goals, targets, and been invested in developing measurement systems (see indicators, along with key additional monitoring indicators appendix B). And as with progress on anything, much recommended by a technical consultation in 2001.1 depends on how it is measured.Are we talking about coun- The first goal explicitly aims to improve the living stan- tries being on track to hit the Millennium Development dards of poor people.The other goals do not.It is statistically Targets or about how close countries come to hitting the possible for there to be no reduction in child malnutrition targets? Are we talking about the number of countries that among, say, the poorest 20 percent of a population but for are on track or the number of people living in on-track the population as a whole to achieve the required 50 per- countries? Are we talking about progress at the global level, cent reduction.Achieving such a result would require a 62.5 at the level of the World Bank region, or at the level of dif- percent reduction among the top 80 percent of the popula- ferent income groupings of countries (low,middle,or high)? tion, but it is a possibility with which policymakers need to Are we talking about progress for populations as a whole or be concerned. One bilateral aid agency recently expressed for specific sections of the population, such as the poor? concern over this, urging that the pursuit of the goals be Countries may, after all, hit the Millennium Development done through a propoor approach to ensure that the poor Target for the whole population and yet leave some groups see gains across all goals.2 This report presents new evidence (the poor,women,ethnic minorities) or regions behind. on whether progress toward the health-related Millennium Prospects for meeting the malnutrition Development Goals has left the poor behind. target are reasonably good There are,of course,links and interdependencies among the various goals: Goal 1 calls for halving the proportion of people suffering from hunger between 1990 and 2015, an average annual · Nutrition has an important influence on maternal and reduction of 2.7 percent.One of the two indicators tracking child health, so progress toward meeting goals 1, 4, and progress toward this goal is the prevalence of underweight 5 is likely to be related. children among children under five. · Nutrition also influences the severity of communicable Five of the six Bank regions achieved population- diseases, including HIV/AIDS and tuberculosis, reduc- weighted average annual reductions in underweight chil- ing people's ability to fight off such diseases. Similarly, dren of 2.7 percent or more in the 1990s (figure 2.1). diarrhea, measles, and malaria are major influences on Upper middle-income countries and particularly lower children's nutritional status (such as anemia). Success middle-income countries exceeded the target (see fig- on goals 1 and 6 is thus also likely to be linked. ure 2.1). Not surprisingly, then, the developing world as a whole is on track to reach the malnutrition target, having · Progress in securing access to safe water and satisfac- achieved a population-weighted average annual decline of tory sanitation (goal 7) and on securing access to 5 percent in the 1990s.More than 50 percent of the devel- affordable medicines (goal 8) is likely to influence oping world's population lives in an on-track country for progress toward goals 1 and 2. the malnutrition goal (figure 2.2). In East Asia and South · Progress on the health Millennium Development Asia, the figure is more than 85 percent, and in low- Goals is also likely to reflect progress on the nonhealth income countries, the figure is 64 percent. goals. Early child nutrition and health, for example, Against this good news is some inevitable bad news. affect educational outcomes for school-age children. Sub-Saharan Africa saw barely any change in child malnu- trition in the 1990s. Only 17 percent of countries there · Rapid progress on the income poverty goal would are on track to meet the malnutrition goal (figure 2.3),and speed progress on the health goals, as would rapid only 12 percent of people there live in an on-track coun- progress on the education and gender equality goals. try. Malnutrition has been falling more slowly in low- income countries than in middle-income countries.In the *Throughout the report we interpret "health" broadly. For example, nutri- tion is both a key aspect of health in its own right and a key influence on other aspects of health. See appendix A for a full explanation of data sources and methods used. 32 | The Millennium Development Goals for Health Table 2.1 Goals, targets, and indicators--many of them for health Millennium Development Goal Targets Indicators Additional monitoring indicators Goal 1: Eradicate Halve, between 1990 and extreme poverty 2015, the proportion of and hunger people whose income is less than $1 a day. Halve, between 1990 and Prevalence of underweight Core intermediate indicators 2015, the proportion of children under five years of age Percentage of children 6­59 months who people who suffer from Proportion of population below received one dose of vitamin A in the past six hunger. minimum level of dietary energy months; proportion of infants under six consumption months who are exclusively breastfed Optional indicators Low birth weight incidence rate; proportion of mothers receiving vitamin A supplementation by eight weeks postpartum Indicators requiring further research Timely complementary feeding; recuperative feeding Goal 4: Reduce Reduce by two-thirds, between Under-five mortality rate Core intermediate indicators child mortality 1990 and 2015, the under-five Infant mortality rate Proportion of infants under six months who are mortality rate. exclusively breastfed; proportion of surviving Measles immunization among infants who have received a dose of measles children under one vaccine by their first birthday; proportion of children with fast or difficult breathing in the past two weeks who received an appropriate antibiotic; proportion of children with diarrhea in the past two weeks who received oral rehydra- tion therapy; proportion of children under five who slept under an insecticide-treated net the previous night (in malarious areas); proportion of children with fever in the past two weeks who received an appropriate antimalarial (in malari- ous areas) Optional indicators Vitamin A supplementation; proportion of infants six- to nine-months-old receiving breast milk and complementary food; piped water and sanitation; female education; effects of income on mortality (such as access to water and sanita- tion, access to and use of cars); birth spacing Indicators requiring further research Perinatal and neonatal indicators (for example, deliveries attended by skilled personnel); indica- tors to capture determinants of death in the first days of life (such as identification and treatment of asphyxia, prevention and treatment of sepsis at birth) Goal 5: Improve Reduce by three-quarters, Maternal mortality ratio Core intermediate indicators maternal health between 1990 and 2015, Proportion of births attended by Contraceptive prevalence rate; percentage the maternal mortality ratio. skilled health personnel of women receiving any antenatal care; provision of emergency obstetric care; prevalence of syphilis in pregnant women and proportion properly treated; percentage of women receiving antenatal care who receive at least two intermittent preventive malaria treat- ments during pregnancy (in malarious areas) Optional indicators Prevention of mother-to-child transmission of HIV/AIDS; perinatal mortality rate; total fertility rate; female genital mutilation Indicators requiring further research Adolescent reproductive health; caesarean section rate The Millennium Development Goals for Health: Progress and Prospects | 33 Table 2.1 Goals, targets, and indicators--many of them for health (continued) Millennium Development Goal Targets Indicators Additional monitoring indicators Goal 6: Combat Halt by 2015 and begin to HIV prevalence among 15- to Core intermediate indicators HIV/AIDS, malaria, reverse the spread of 24-year-old pregnant women Percentage of people using a condom during and other diseases HIV/AIDS. Condom use rate of the most recent higher-risk sexual encounter; contraceptive prevalence rate percentage of sexually transmitted infection clients who are appropriately diagnosed and Number of children orphaned treated according to guidelines; percentage of by HIV/AIDS HIV-positive women receiving antiretroviral treatment during pregnancy to prevent mother- to-child transmission of HIV Halt by 2015 and begin to Prevalence and death rate Core intermediate indicators reverse the incidence of associated with malaria Percentage of patients with uncomplicated malaria and other major Proportion of population in malaria who received treatment within 24 diseases. malaria risk areas using effective hours of onset of symptoms; percentage of malaria prevention and children under age five sleeping under insecti- treatment measures cide-treated nets; percentage of pregnant women sleeping under insecticide-treated nets; percentage of pregnant women who have taken chemoprophylaxis or drug treatment for malaria Prevalence and death rates Core intermediate indicators associated with tuberculosis Percentage of estimated new smear-positive Proportion of tuberculosis cases tuberculosis cases registered under the DOTS detected and cured under DOTS approach Goal 7: Ensure Integrate the principles of environmental sustainable development sustainability into country policies and reverse loss of environmental resources. Halve by 2015 the proportion Proportion of population with of people without sustainable sustainable access to an improved access to safe drinking water. water source, urban and rural By 2020 achieve a significant Proportion of urban population improvement in the lives of with access to improved sanitation at least 100 million slum Proportion of households with dwellers. access to secure tenure Goal 8: Develop Develop further an open, a global rule-based, predictable, partnership for nondiscriminatory trading development and financial system. Address the special needs of the least developed, land- locked, and small island developing countries. Deal comprehensively with the debt problems of developing countries. Develop and implement strategies for decent and productive work for youth. In cooperation with pharma- Proportion of population with ceutical companies, provide access to affordable essential access to affordable essential drugs on a sustainable basis drugs in developing countries. In cooperation with the private sector, make available the benefits of new tech- nologies, especially informa- tion and communications. Source: Reference 18. 34 | The Millennium Development Goals for Health Figure 2.1 Progress on malnutrition, under-five Figure 2.2 People on track to hit malnutrition, mortality, and maternal mortality, by region and under-five mortality, and maternal mortality income targets Average annual percentage change Percentage of people on track 0 20 40 60 80 100 ­10 ­8 ­6 ­4 ­2 0 East Asia and East Asia and Pacific Pacific Europe and Europe and Central Asia Central Asia Latin America and Latin America and the Caribbean the Caribbean Middle East and Middle East and North Africa North Africa South Asia South Asia Under-five Under-fiveMaternal Sub-Saharan Sub-Saharan Maternal mortality mortality mortality Africa Africa mortality ratio rate rate Malnutrition ratio Malnutrition Developing Developing world world Low income Low income Lower middle Lower middle income income Upper middle Upper middle income income High income: High income: non-OECD non-OECD High income: High income: OECD OECD Source: References 12, 13, and 14. Source: References 12, 13, and 14. low-income countries, malnutrition has declined only 2.6 tries--Colombia and Tanzania--progress was substantially percent a year--just short of the target rate. Even in the faster among the very poor than among the whole popu- regions experiencing high average annual declines in mal- lation.But in several countries--including Bangladesh and nutrition, several countries have fallen short of the target Bolivia, often held up as examples of countries that have rate of 2.7 percent (figure 2.4). The East Asia and the implemented propoor policies--child malnutrition has Pacific region is well on target,for example,but nearly half been falling more slowly among the very poor. the countries there are currently off target. But things Child mortality: Poor progress could have been worse. That 17 percent of Sub-Saharan countries and 20 percent of low-income countries are on The targets call for a two-thirds reduction in infant and track provides some hope for other Sub-Saharan and low- under-five mortality rates between 1990 and 2015, an income countries. average reduction of 4.3 percent a year. But no Bank How have the poor fared in reducing malnutrition? For region achieved a population-weighted annual average the 17 countries with disaggregated trend data,the average reduction in excess of 4.3 percent in the 1990s (see figure rate of reduction of malnutrition in the 1990s was broadly 2.1). The only countries on target are the high-income similar for the poorest 20 percent of the population and countries--indicating that, contrary to what is often for the population as a whole (figure 2.5). In two coun- claimed, large percentage reductions in child mortality are The Millennium Development Goals for Health: Progress and Prospects |35 Figure 2.3 Countries on track to hit malnutrition, Figure 2.4 Some countries are reducing under-five mortality, maternal mortality targets malnutrition quickly. In others malnutrition has fallen less slowly, and in some it has increased. Percentage of countries on track 0 20 40 60 80 Average annual percentage change ­30 ­20 ­10 0 10 20 30 40 50 East Asia and Pacific Mexico Kazakhstan Europe and C. African Rep. Central Asia Tunisia Sudan Russian Federation Latin America and Iran, Islamic Rep. the Caribbean Chad Maternal Under-five Gambia, The Middle East and mortality mortality Guyana North Africa ratio rate China Dominican Rep. Malnutrition Chile South Asia Botswana Jamaica Egypt, Arab Rep. Sub-Saharan Bolivia Africa Indonesia India Developing world Vietnam Venezuela, RB Mozambique Ghana Low income Pakistan Philippines Lower middle Armenia income Lesotho Mongolia Nicaragua Upper middle El Salvador income Panama Maldives High income: Comoros non-OECD Eritrea Nigeria High income: Syrian Arab Rep. OECD Costa Rica Cameroon Yemen, Rep. Honduras Source: References 12, 13, and 14. Uruguay Azerbaijan Togo Albania Guinea Argentina possible even at low levels of mortality.The goals, which are defined in terms of percentage reductions rather than Source: Reference 13. absolute levels, are as relevant to middle-income and high- income countries as they are to low-income countries.As a group, the middle-income countries experienced slower tries are on track for the under-five mortality target, and annual reductions in child mortality than the high-income only 22 percent of the developing world's population lives countries,and they are off track for the child mortality tar- in an on-track country (see figures 2.2 and 2.3). Not a sin- get.The upper middle-income countries have fared better gle Sub-Saharan country is on track for the child mortality and are just on track, while the lower middle-income target.And in all regions--especially Latin America and the countries are badly off track. Caribbean, the Middle East and North Africa, and Sub- The widening gap between poor and better-off coun- Saharan Africa--the rate of decline of under-five mortality tries is mirrored at the country level.And gaps in child mor- was greater in the 1980s than in the 1990s (figure 2.7). tality between the poor and the better off are widening The news on child mortality is not all bad, however. within countries, too. On average, the child mortality rate Although mortality rates fell more rapidly, on average, in the among the poorest 20 percent of the population fell just half 1980s than in the 1990s, 38 percent of countries saw a faster as fast as for the whole population (figure 2.6 and appendix decline in the 1990s than in the 1980s (figure 2.8). For Sub- B).And there is more bad news. Only 16 percent of coun- SaharanAfrica the figure was 41 percent,and half the region's 36 | The Millennium Development Goals for Health Figure 2.5 How the poor have fared on Figure 2.6 How the poor have fared on malnutrition reductions--absolutely and under-five mortality reductions--absolutely compared with the less poor and compared with the less poor Average annual percentage change Average annual percentage change ­15 ­10 ­5 0 5 10 ­10 0 10 20 Egypt, Arab Rep. Cameroon 1991­1998 1995­2000 Nepal 1996­2001 Turkey 1993­1998 Zambia 1996­2001 Colombia 1995­2000 India 1992/93­1998/99 Guatemala 1995­1998 Ghana 1993­1998 Bolivia 1994­1998 Malawi 1992­2000 Peru 1996­2000 Zimbabwe 1994­1999 Nepal 1996­2001 Guatemala 1995­1998 Mali 1995/96­2001 Peru 1996­2000 Ghana 1993­1998 Bangladesh Zambia 1996­2001 1996/97­2000 Uganda 1995­2000/01 Nicaragua 1997­2001 Mali 1995/96­2001 India 1992/93­1998/99 Nicaragua 1997­2001 Malawi 1992­2000 Bangladesh Turkey 1993­1998 Poorest 20% 1996/97­2000 Tanzania 1996­1999 Population Cameroon 1991­1998 Population Haiti 1994/95­2000 Uganda 1995­2000/01 Poorest 20% Colombia 1995­2000 Haiti 1994/95­2000 Bolivia 1994­1998 Tanzania 1996­1999 Egypt, Arab Rep. Zimbabwe 1994­1996 1995­2000 Kazakhstan 1995­1999 Kazakhstan 1995­1999 Source: World Bank staff calculations and reference 15. Source: World Bank staff calculations and reference 15. people live in a country that accelerated its rate of decline of estimated annual reduction of 4.5 percent. Across the child mortality in the 1990s.This partly reflects the fact that developing world, 17 percent of countries are on track to Ethiopia and Uganda improved their performance in the achieve the maternal mortality target, but there are large 1990s.This is also a cause for hope:despite HIV/AIDS,many variations across regions, ranging from 4 percent in Latin Sub-Saharan countries have shown themselves able to accel- America and the Caribbean to 64 percent in the Middle erate the rate of decline of under-five mortality. East and North Africa. For the developing world as a whole, 32 percent of the population lives in an on-track Maternal mortality: Insufficient progress country. But the 12 percent of countries on track in South The Millennium Development Goals call for a three- Asia have a negligible share of that region's population. quarters reduction in the maternal mortality ratio between Communicable diseases:Too little progress 1990 and 2015,an average annual reduction of 5.4 percent. The developing world as a whole is off target, having The goals call for a reversal in the spread of HIV/AIDS and registered an average population-weighted decline of just other communicable diseases and the beginning of a 3.2 percent in the 1990s (see figure 2.1). Indeed,only one decline in incidence. According to UNAIDS, the AIDS region--the Middle East and North Africa--is on target, epidemic led to the deaths of an estimated 3.1 million peo- though East Asia and the Pacific comes fairly close, with an ple in 2002, and an estimated 5 million people were living with HIV infection in 2002. Some countries have been More accurately, a country's estimated rate of reduction is weighted by the devastated by HIV/AIDS. In Botswana the rate of increase number of births. in the number of new cases appears to have peaked in the The Millennium Development Goals for Health: Progress and Prospects | 37 Figure 2.7 Child mortality--the pace of decline is Figure 2.8 Child mortality--percentage of too slow, and in the developing world it's getting countries with a faster rate of decline in even slower the 1990s than in the 1980s Average annual percentage change Percentage of countries ­6 ­5 ­4 ­3 ­2 ­1 0 0 20 40 60 80 100 East Asia and Pacific East Asia and Pacific Europe and 1980s 1990s Europe and Central Asia Central Asia Latin America Latin America and and the Caribbean the Caribbean Middle East and Middle East and North Africa North Africa South Asia South Asia Sub-Saharan Africa Sub-Saharan Africa Developing world Developing world Low income Low income Lower middle Lower middle income income Upper middle Upper middle income income High income: High income: non-OECD non-OECD High income: OECD High income: OECD Source: Reference 16. Source: Reference 16. late 1990s--but not until the prevalence of HIV reached a Figures 2.10 and 2.11 show DOTS detection and cure staggering 44 percent in urban areas.Yet there are some rates for selected countries in Sub-Saharan Africa. hopeful signs. In Uganda HIV prevalence in Kampala has been declining steadily--from a high of 30 percent in 1990 Will the second half go better? to 11 percent in 2000--and similar rates of decline have been observed in other areas of the country (figure 2.9). In As the well-known stock market dictum warns--and as Côte d'Ivoire HIV prevalence in major urban areas has figure 2.12 shows for under-five mortality--past perfor- declined for the last three years for which data are available. mance is not necessarily a good predictor of future perfor- The number of countries implementing directly mance.The fact that a country is on track on the basis of observed treatment, short-course (DOTS) regimens for its performance in the 1990s does not guarantee that it tuberculosis has increased to 155 (of 210). More than 10 will maintain the required annual rate of reduction of mal- million patients have been diagnosed and treated in DOTS nutrition or mortality in 2000­15. It is also possible, of programs since 1995. While some countries have made course, that countries that are currently off track may get rapid progress in DOTS detection rates, those with high on track during the second half.What are the likely future tuberculosis burdens are not increasing case detection rates trends in the various country-level determinants of mal- toward the 70 percent target in DOTS areas. Treatment nutrition and mortality?Will countries grow more rapidly success under DOTS for 2000 was 82 percent on average in the new millennium than in the 1990s? If so,how much but substantially below average in Africa (72 percent). extra mortality reduction will this faster growth buy? 38 | The Millennium Development Goals for Health Figure 2.9 Trends in HIV prevalence among Figure 2.11 Tuberculosis DOTS cure rates, selected pregnant women in Uganda, 1990­2001 countries in Sub-Saharan Africa, 1994­2000 35 100 90 positive 30 HIV 80 directly are 25 70 who under 20 short-course 60 cured 50 women 15 cases treatment, 40 of pregnant 10 30 of observed 20 5 Percentage 10 Percentage 0 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 1993 1994 1995 1996 1997 1998 1999 2000 2001 Major urban area Outside urban area Côte d'Ivoire Ethiopia Ghana Nigeria Source: Reference 12. Source: Reference 17. Figure 2.10 Tuberculosis DOTS detection rates, selected countries in Sub-Saharan Africa, 1995­2001 the Middle East and North Africa, and South America 100 (figure 2.13). Given mortality's and malnutrition's close links with per capita income,4,5 the increase in growth will 90 help push countries toward the goals, especially in Europe directly 80 and Central Asia, where the change in growth is apprecia- ble. But even with faster growth, per capita incomes will under 70 rise at more than 2 percent a year in only two regions-- short-course 60 East Asia and the Pacific and South Asia. detected 50 Primary education completion will probably cases increase more rapidly in the new millennium treatment, 40 new of 30 The Millennium Development Goals call for 100 percent completion rates at primary level, and it is likely that observed 20 Education for All and the Fast Track Initiative will accelerate 10 Percentage progress toward universal primary completion.The health 0 payoffs to this in the second half are unclear.An increase in 1994 1995 1996 1997 1998 1999 2000 2001 2002 the primary completion rate in 2000 will not filter Côte d'Ivoire Ethiopia Ghana Nigeria through into higher levels of educational attainment among women of childbearing age until 2005 or so. And Source: Reference 17. the first full round of effects on under-five mortality will not be felt until 2010. Since only a small percentage of births are to 15-year-olds, the bulk of any child survival Economic growth will increase in the new payoffs associated with an increased primary completion millennium, except in East Asia rate in 2000 will come after 2010--mostly well after 2015. TheWorld Bank3 estimates that growth will fall somewhat Even beyond 2015 it is unlikely that increases in pri- in East Asia, turn from negative to positive growth in mary school completion rates will yield appreciable payoffs Europe and Central Asia and Sub-Saharan Africa, and in better health outcomes. A recent multicountry study6 increase somewhat in Latin America and the Caribbean, using Demographic and Health Surveys data found that The Millennium Development Goals for Health: Progress and Prospects |39 Figure 2.12 Past performance is not necessarily a Figure 2.13 Growth rates of per capita income: good predictor of future performance: under-five actual 1990­2000 and forecasts for 2001­15 mortality 7% 1990­2000 Average annual percent change 6% 2001­15 ­13 ­10 ­8 ­5 ­3 0 5% Egypt, Arab Rep. Malaysia 4% Czech Republic Oman Libya 3% Poland 1980s Indonesia 1990s 2% Tunisia Morocco 1% Ecuador Hungary 0% Bangladesh Turkey Slovak Rep. ­1% Chile Bhutan ­2% Philippines Iran, Islamic Rep. ­3% Samoa East Asia Europe Latin Middle East South Sub-Saharan West Bank & Gaza and and America and Asia Africa Vanuatu Pacific Central and the North Brazil Asia Caribbean Africa Macedonia, FYR Lao PDR Source: Reference 3. Mexico Bolivia Honduras Syrian Arab Rep. Saudi Arabia of quality are probably too low for girls to acquire the lit- Peru El Salvador eracy and numeracy skills needed to develop enough Yugoslavia, Fed. Rep. health-specific knowledge later in life. Cape Verde Nicaragua Comoros Gender gaps in secondary education may Romania well narrow in the new millennium Albania Nepal Croatia One of the Millennium Development Goals is to promote Cuba St. Kitts & Nevis gender equality and empower women. If the elimination Solomon Islands of gender disparities at all levels of education entails faster Maldives Dominica growth in secondary school completion rates among girls, Grenada it could have significant consequences for maternal and Source: Reference 16. child health outcomes. To achieve parity with boys by 2015 in the proportion of the population 15 and older that completes secondary education, girls' completion while secondary education can exert a significant effect on rates will have to grow faster in the new millennium than infant mortality, primary education does so in only a few in the 1990s in most regions, especially in East and South countries.§This is consistent with the finding7 that better- Asia (figure 2.14).** educated women achieve better health outcomes not by Attaining the drinking water and sanitation goals would using health-specific knowledge acquired at school but by entail faster growth in access in the new millennium using the general numeracy and literacy skills they are taught at school to acquire health-specific knowledge later The Millennium Development Goals (and the Johannesburg on in life. The implication is not that primary schooling goals) call for halving the proportion of the world's popula- necessarily has limited and possibly negligible payoffs for child and maternal health. It is that current average levels ** In this simulation it is assumed that the proportion of males in each World Bank region who complete secondary education continues to grow §This is consistent with cross-country econometric work undertaken for in 2001-15 at the same rate it grew during the second half of the 1990s.The this report that finds that while female secondary education significantly figure for females is then brought up to the level that males would reach in reduces maternal mortality, female primary education has no significant effect 2015, on the assumption that the figures for both males and females grow on any Millennium Development Goal outcome (see appendix A). linearly. 40 | The Millennium Development Goals for Health Figure 2.14 Girls catching up with boys at Figure 2.15 In all but two regions access to secondary school level--growth needed in the drinking water needs to grow faster to achieve share of female population over age 15 who have the water target completed secondary education to eliminate the 3.0 secondary education gender gap Extra growth required Current growth rate 6 2.5 Required growth rate Extra growth required 1995­2000 2.0 5 (percent) (percent) 1.5 4 rate rate 1.0 growth 3 growth 0.5 Annual annual 2 0.0 Average1 ­0.5 East Europe Latin Middle South Sub- Asia and America East Asia Saharan 0 and Central and and North Africa East Europe Latin Middle South Sub-Saharan Pacific Asia the Africa Asia and America East Asia Africa Caribbean and Central and and Pacific Asia the North Source: World Bank tabulations and reference 14. Caribbean Africa Source: World Bank staff calculations and reference 13. points to the annual rate of reduction of maternal mortality, taking the annual rate of decline from 2.9 percent to 5.5 tion without access to safe drinking water and sanitation percent. These contributions will push countries and between 1990 and 2015. If drinking water access rates grow regions firmly toward the Millennium Development linearly at the trend rates of the 1990s, only Latin America Targets, but for the most part they will not get them there. and the Caribbean and the Middle East and North Africa (An exception is maternal mortality in EastAsia.) In Europe will achieve the Millennium DevelopmentTarget.To achieve and Central Asia and South Asia the impetus is also large-- the water target, access rates would need to grow much but for the relatively slow progress in the 1990s, the antici- more rapidly, especially in Sub-Saharan Africa, where access pated faster economic growth and faster progress on girls' is currently growing by 0.43 percent a year but would need education might have allowed both regions to hit the to grow by 2.53 percent a year to meet the goal (figure maternal mortality targets.The picture is bleakest for under- 2.15). If access rates were to grow rapidly enough to achieve five mortality and for Sub-Saharan Africa. Even with the the water target, the under-five and maternal mortality rates one extra percentage point annual reduction in under-five could fall appreciably. mortality driven by economic growth and achievement of the gender and water targets, the projected reduction of Even with economic growth and faster progress under-five mortality for 2000­15 is still only 1.6 percent a on the nonhealth goals, many regions year in Sub-Saharan Africa. will still miss many of the health targets Why the picture may be bleaker Faster economic growth and achievement of the gender and water goals will lead to appreciable declines in mortality The assumptions about future trends in the determinants and malnutrition (see appendix A for methods). In Europe of the Millennium Development Goals used for the simu- and Central Asia progress on these three fronts might add as lations conducted here are probably overly optimistic. It is much as 1.4 percentage points to the rate of decline of assumed that progress on eliminating gender gaps in sec- under-five mortality, 1.1 percentage points to the rate of ondary education and expanding access to drinking water decline of maternal mortality, and just under 1 percentage is fast enough to achieve the relevant target. It is also point to the annual rate of reduction of underweight (table assumed that in the absence of faster economic growth 2.2). In South Asia it might add as much as 2.6 percentage and accelerated progress toward the gender and water The Millennium Development Goals for Health: Progress and Prospects |41 Table 2.2 Economic growth and attainment of the gender and water targets will push countries toward the health targets but leave many a long way from them Target Current Contributions from faster Millennium growth rate growth rate economic growth and Revised Development 1990­2015 1990­2000 progress toward gender growth Goal Region (percent) (percent) and water targets (percent) rate (percent) Under-five East Asia and the Pacific ­4.3 ­2.7 ­0.2 ­2.9 mortality rate Europe and Central Asia ­4.3 ­2.5 ­1.4 ­3.9 Latin America and ­4.3 ­3.7 0.0 ­3.7 the Caribbean Middle East and North Africa ­4.3 ­3.6 ­0.1 ­3.7 South Asia ­4.3 ­2.6 ­0.6 ­3.2 Sub-Saharan Africa ­4.3 ­0.3 ­1.0 ­1.3 Maternal East Asia and the Pacific ­5.4 ­4.5 ­1.7 ­6.3 mortality ratio Europe and Central Asia ­5.4 ­4.3 ­1.0 ­5.3 Latin America and the Caribbean ­5.4 ­1.5 0.0 ­1.5 Middle East and North Africa ­5.4 ­6.9 ­0.8 ­7.8 South Asia ­5.4 ­2.9 ­2.6 ­5.5 Sub-Saharan Africa ­5.4 ­1.6 ­2.4 ­4.1 Underweight East Asia and the Pacific ­2.7 ­6.7 0.2 ­6.6 Europe and Central Asia ­2.7 ­9.6 ­0.9 ­10.5 Latin America and the Caribbean ­2.7 ­4.1 0.0 ­4.1 Middle East and North Africa ­2.7 ­6.3 ­0.1 ­6.4 South Asia ­2.7 ­3.5 ­0.1 ­3.6 Sub-Saharan Africa ­2.7 ­0.2 ­0.3 ­0.5 Tuberculosis East Asia and the Pacific n.a. n.a. 0.2 0.2 mortality Europe and Central Asia n.a. n.a. ­1.1 ­1.1 Latin America and the Caribbean n.a. n.a. 0.0 0.0 Middle East and North Africa n.a. n.a. ­0.1 ­0.1 South Asia n.a. n.a. ­0.1 ­0.1 Sub-Saharan Africa n.a. n.a. ­0.3 ­0.3 n.a. Not applicable (no targets for tuberculosis mortality). Source: World Bank staff calculations. goals, the pace of decline would continue at the rate of the the gender and water goals is substantially accelerated, the 1990s. Both assumptions are probably overly optimistic. developing world is likely to wake up on the morning of Changes in three variables found to be good predictors of January 1, 2016, some way from the health targets; in Sub- maternal mortality are likely to be less conducive to Saharan Africa, it may be a long way. So why should the reductions in maternal mortality in the new millennium Millennium Development Goals be taken seriously? than they were in the 1990s (box 2.1). In the absence of Progress can be quickened, irrespective of the targets faster economic growth and accelerated progress on the nonhealth goals, maternal mortality will probably decline Accelerating progress on the various health goals is critical-- less quickly in the new millennium than in the 1990s. even if the targets will not be reached. Reversing negative trends in Sub-Saharan Africa and accelerating slow progress in other regions are vital.A key message of this report is that So why do the Millennium progress can be accelerated through a judicious mix of Development Goals matter? spending and reform of both policies and institutions. The upshot of all this:even if economic growth accelerates as Countries can set--and are setting--their own development Global Economic Prospects predicts,and even if progress toward targets, based on the Millennium Development Goals, but 42 | The Millennium Development Goals for Health Box 2.1 Why the decline in maternal mortality is slowing The Millennium Development Goal for maternal health is to Asia, or Africa; whether death registration is complete; and the reduce the maternal mortality ratio by three-quarters between prevalence of HIV among adults. 1990 and 2015. In addition to the challenge of accomplishing Skilled attendance at delivery has a negative correlation with such a large reduction, it will be difficult in many countries to maternal mortality (and is an indicator used to measure the measure whether the goal has been achieved or whether coun- Millennium Development Goals for maternal health by itself). tries have recently been making enough progress to be on track HIV prevalence and the resulting mortality from AIDS reduces to reach the goal.Trends in maternal mortality ratios are difficult the exposure to death from other causes, including maternal to establish for several reasons: mortality. Higher fertility rates tend to lead to higher maternal mortality ratios, because more high-risk pregnancies occur. · Measurement of maternal mortality is complex, partly Data for these variables for 1990, 1995, and 2001 are estimated because of the concept itself (a cause-specific mortality rate) fromWorld Bank and UN databases.Projections for 2015 are based and because of the rarity of maternal mortality. Estimates on extrapolated trends in the proportion of skilled attendants in the based on hospital studies suffer from a lack of representative- 1990s and on the assumption of gradually increasing or decreasing ness.Vital registration systems exist in many countries, but HIV prevalence,depending on trends in the 1990s.The general fer- the incompleteness of the systems is a major obstacle to their tility rate is estimated from projections of the number of births and use in epidemiology. Household surveys using respondents' the number of females between 15 and 49 (see appendix B). reports on the mortality of sisters have provided most of the The estimated proportion of deaths of women of reproduc- measured estimates of maternal mortality in developing tive age is applied to the overall envelope of deaths to women countries. But estimates based on this approach usually lack 15­49 (estimated and projected) to obtain the number of deaths precision and are typically for several years in the past. due to maternal causes for all countries.The maternal mortality · Maternal deaths are frequently misclassified. Even in coun- ratio is then calculated by dividing this number by the estimated tries with complete coverage of deaths in vital registration, or projected number of births. maternal deaths may be misclassified in 50 percent of cases.8 · Baseline estimates have not been collected in many coun- Results: 1990­2001 tries, and too few additional estimates are likely to be forth- Skilled attendance at birth rose, fertility declined, and HIV/AIDS coming. This is the most serious obstacle to monitoring increased (see table) in all Bank regions in the 1990s. In each case trends in maternal mortality. the effect was to reduce the proportion of deaths of women of To overcome the paucity of the data, UNICEF and the World reproductive age (PMDF).In East Asia and the Pacific,Europe and Health Organization (WHO) have developed models to predict CentralAsia,and the Middle East and NorthAfrica,the number of the maternal mortality ratio from a set of variables that are female deaths also fell, reinforcing the effect of the declining known to be related to maternal mortality. Such model-derived PMDF. However, the number of births fell in these three regions, estimates have been issued for 1990 and 1995, and a set of esti- putting upward pressure on the maternal mortality ratio. The mates for 2000 was released as this report was being completed. maternal mortality ratio is estimated to have fallen in all three Because the models use different variables,the results of the three regions, though by a smaller percentage than the PMDF in East exercises are not comparable,and the data cannot be used to ana- Asia and the Pacific and in Europe and CentralAsia.In the Middle lyze trends over time. East and North Africa, the decline in the number of births was small, and the maternal mortality ratio fell by a larger percentage New analysis than the PMDF. In Latin America and the Caribbean, Southeast Asia, and Sub-Saharan Africa, the number of female deaths rose Analysis for this report uses the model developed by WHO- during the 1990s, partly offsetting the fall in the PMDF. In Latin UNICEF for countries without vital registration data or reliable America and the Caribbean, Southeast Asia, and Sub-Saharan survey estimates to generate the 1995 estimates and extends the Africa, the maternal mortality ratio is estimated to have fallen less period back to 1990 and forward to 2015.8 Estimates of progress than in the other three regions, in part because of relatively small from 1990 to 2001 are made by calculating the rate of change in percentage reductions in the PMDF,in part because of increases in the maternal mortality ratio for this period, which is then com- female deaths. pared with the rate of change needed to achieve the Millennium Development Goal by 2015. Projections for 2001­15 The model estimates the proportion of deaths of women of reproductive age due to maternal causes from the following vari- The extrapolated trend in skilled attendance at birth will further ables: the proportion of births with a skilled health care worker; reduce the PMDF. In some countries, however, skilled atten- the general fertility rate; whether a country is in formerly social- ist Europe; whether a country is in Latin America, South orWest (continued) The Millennium Development Goals for Health: Progress and Prospects | 43 Box 2.1 Why the decline in maternal mortality is slowing (continued) dance has already reached high levels and has limited room for The result of these projected changes in the independent vari- future increases; in only two regions (Latin America and the ables is that the PMDF will keep falling, but at a slower rate than Caribbean and Southeast Asia) are attended deliveries expected in the 1990s, except in Southeast Asia. Beginning in 2001, female to increase at a faster rate than in the 1990s.The general fertility deaths are expected to rise in all regions except the Middle East rate is projected to continue to decline but at a slower rate than and North Africa, partly offsetting the fall in the PMDF.The net in the 1990s, except in Sub-Saharan Africa, where it is projected effect is likely to be a reduction in the number of maternal deaths, to decline more rapidly. HIV/AIDS prevalence will continue to but--partly because of changing numbers of births--the mater- increase or decline, depending on trends in the 1990s, but the nal mortality ratio is projected to decline more slowly than in the rate of increase or decline will slow. 1990s,except in Latin America and the Caribbean. Maternal mortality ratio trends and forecasts Latin America Middle East Sub- East Asia and Europe and and the and Southeast Saharan the Pacific Central Asia Caribbean North Africa Asia Africa 1990­2000 growth rates Attended deliveries 1.9% 1.1% 0.2% 3.7% 2.1% 1.8% Gross fertility rate ­0.6% ­0.9% ­0.5% ­0.7% ­0.4% ­0.2% HIV/AIDS 72.5% 74.7% 80.5% 61.6% 78.5% 101.4% Proportion of deaths of women ­4.6% ­5.6% ­3.4% ­5.4% ­2.7% ­2.1% of reproductive age Number of female ­1.1% ­0.1% 2.1% ­0.7% 0.8% 3.2% deaths Number of births ­1.6% ­2.2% ­0.2% ­0.3% ­0.1% 1.9% Maternal mortality ratio ­4.5% ­4.3% ­1.5% ­6.9% ­2.9% ­1.6% 2001­2015 growth rates Attended 1.6% 0.2% 0.5% 1.3% 2.8% 1.5% deliveries Gross fertility rate ­0.2% ­0.2% ­0.4% ­0.3% ­0.4% ­0.3% HIV/AIDS 1.4% 4.9% 1.2% 4.7% ­0.2% 0.8% Proportion of deaths of women ­2.9% ­2.6% ­3.0% ­2.8% ­3.2% ­2.1% of reproductive age Number of female 2.1% 1.3% 0.3% ­1.0% 2.1% 3.0% deaths Number of births ­0.3% ­0.6% ­0.6% 0.9% ­0.2% 1.1% Maternal ­1.5% ­1.1% ­1.9% ­4.8% ­2.5% ­1.3% mortality ratio they are tailoring the targets to local circumstances. In some five mortality--less so for child malnutrition--the poor are countries the targets may be more ambitious than the progressing less quickly than the better off.There is an urgent Millennium Development Targets. In others realism will need to ensure that countries do not progress toward the require the setting of less ambitious targets.The important Millennium Development Goals by focusing on the low- thing is that the goals be pursued and that realistic but chal- hanging fruit, leaving the poor behind in the process. lenging targets be aimed at. Countries and regions that are on track cannot be compla- cent. For example, the Middle East and North Africa region Progress needs to be for everyone, not just the better off has better prospects than most for reaching the targets, but it Progress toward the Millennium Development Targets is has some of the largest inequalities across wealth groups in the uneven within countries (see figures 2.5 and 2.6).For under- Millennium Development Indicators.The challenge there is 44 | The Millennium Development Goals for Health to ensure that progress for the whole population is accompa- True, the targets are ambitious. But if they had been set so nied by narrowing the gaps between the poor and the better low that developing countries could have achieved them off. In Egypt in the late 1990s under-five mortality fell more by making progress at current rates, they would have been rapidly among the poor than among the whole population criticized for being unambitious. In the 1990s, 20 percent (see figure 2.6).That should inspire other countries to reach of developing countries achieved the target 4.3 percent out to the poor in their efforts to hit the targets. annual reduction in under-five mortality. And several countries accelerated their annual reductions--Egypt Goals help focus on outcomes from 5 percent to 8.5 percent, Indonesia from about 3 A major attraction of the Millennium Development Goals percent to 6.25 percent,and Bangladesh from about 3 per- is that they provide an opportunity to refocus attention in cent in the 1980s to 5.5 percent in the 1990s. the health sector on outcomes. Focusing on a specific set What this suggests is that the world should not abandon of outcomes requires choosing a limited set of corre- hope for the countries off track on reaching the goals. sponding interventions and devising strategies to ensure Annual rates of under-five mortality reduction in excess of that everyone who could benefit from these interven- 4.3 percent are possible. Some developing countries with tions--whether affluent or poor--actually gets them.This impressive rates of decline in the 1980s accelerated those approach forces us to think about why coverage levels are rates in the 1990s.A final reason for hope: many countries often so low and about the role of household factors,com- not on track for the under-five mortality target achieved munity factors, and health system factors. (or were close to achieving) the required annual percent- The appropriate strategy will usually entail working at age reduction in the 1980s (see figure 2.12). All this sug- all levels.The role of households in initiating and deliver- gests that, while the targets are ambitious, they are not ing care cannot be ignored. The role of communities in impossible for the majority of developing countries. shaping demand and delivering interventions is often cru- Countries off track should be encouraged to ask how they cial, especially in poorer communities. Also crucial is the can get on track--or back on track. health system--providers and their links to one another, There is reason to hope that developing countries can the human and material inputs they have available to make faster progress toward the other Millennium Develop- them, the incentives they face, the organizational and reg- ment Goals as well.9 Several countries have achieved signifi- ulatory structure they work in. cant reductions in maternal mortality rates.Sri Lanka,a poor These factors have long been considered important in country,reduced the maternal mortality ratio from 555 to 24 the health sector.What is different about the Millennium per 100,000 live births between 1950 and 1995.10 Malaysia Development Goals is that they force us to take stock of reduced its maternal mortality ratio from 500 to 21 over the whether the general debates surrounding the health sector same period. And China, Egypt, and Honduras greatly have done enough to generate strategies for raising the improved maternal health outcomes in the 1990s.11 coverage levels of the very specific set of interventions of There are encouraging examples for communicable dis- relevance to the goals. ease, too. Tuberculosis case detection under the recom- Public policy debates in the health field too often focus on mended DOTS approach increased steadily through the the delivery of services in health facilities.Indeed,analysis and late 1990s--from 7.5 percent of tuberculosis cases in 1995 debate often center on the delivery of care in public facilities. to 33 percent in 2001, although the scaling-up is still too The Millennium Development Goals, with their focus on slow. Even so, large countries such as China, India, outcomes,provide a powerful reminder that such perspectives Indonesia, Pakistan, the Philippines, and the Russian are unduly narrow. The preventive care and treatment that Federation are all moving to scale up more quickly, with households themselves deliver are important to health out- some successes so far.The epidemiological impact of ser- comes. The care delivered outside the public sector is also vice interventions is measurable. In China the areas of 16 important. The Millennium Development Goals provide a provinces where DOTS had been applied, with World chance to refocus health policy on health outcomes and the Bank support, saw a 36 percent drop in tuberculosis preva- major determinants, core interventions, and delivery strate- lence between 1990 and 2000--a decline that was more gies--and in the process to forge closer links between health than 10 times greater than in nonintervention areas. In ministries and other organizations that can help achieve these Peru reaching cure and case detection targets was associ- outcomes,including other government ministries. ated with a rapid decline in incidence. Sustained reductions in HIV infection have been The goals remind us of what is achievable achieved in Thailand through government efforts to pro- The value of the Millennium Development Goals goes mote safer sex.The HIV prevalence rate has been reduced beyond getting policymakers to focus more on outcomes. among 13­19 year-olds in Masaka, Uganda.The spread of The Millennium Development Goals for Health: Progress and Prospects |45 HIV has been contained in Senegal through a broad-based 3. World Bank. 2003. Global Economic Prospects and the Developing response to keep HIV prevalence low among pregnant Countries.Washington, DC. women and steps to maintain a high level of condom use 4. Pritchett, L., and L.H. Summers. 1996."Wealthier Is Healthier." Journal of Human Resources 31 (4): 841­868. to prevent sexually transmitted infections among sex 5. Haddad, L. 2003."What Can the Analysis of Food Security and workers. In the effort to roll back malaria, free treatment Nutrition Add to Poverty Diagnosis?" World Bank, and insecticide-treated nets have reduced malaria deaths in Washington, DC. Vietnam and coastal Kenya, while insecticide spraying and 6. Desai, S., and S.Alva. 1998."Maternal Education and Child effective case management have reduced malaria in Health: IsThere a Strong Causal Relationship?" Demography 35 (1): 71­81. Azerbaijan. 7. Glewwe, P. 1993."Why Does Mother's Schooling Raise Child Health in Developing Countries?" Journal of Human Resources The rest of the report 34 (1):124­159. 8. Hill, K., C.Abou Zhar, andT.Wardlaw. 2001."Estimates of That some countries--and indeed some regions--have Maternal Mortality for 1995." Bulletin of theWorld Health achieved the target rates of improvement in the Organization 79 (3):182­193. Millennium Development Indicators points to the need to 9. World Health Organization. 2002. Health:A Key to Prosperity: understand what has been responsible for the successes Success Stories in Developing Countries. Geneva. and the failures. Chapter 3 reviews the interventions avail- 10. Pathmanathan, I., J. Liljestrand, J.M. Martins, L.C. Rajapaksa, able for reducing child and maternal mortality, malnutri- C. Lissner,A. de Silva, S. Selvaraju, and P.J. Singh.2003. Investing in Maternal Health: Learning from Malaysia and Sri Lanka. tion, and communicable disease mortality. It argues that, Washington, DC:World Bank. together, they constitute a powerful arsenal of weapons to 11. Koblinsky, M.A., ed. 2003. Reducing Maternal Mortality: Learning take on the Millennium Development Goals and that gaps from Bolivia, China, Egypt, Honduras, Indonesia, Jamaica, and in the arsenal have not, for the most part, held countries Zimbabwe.Washington, DC:World Bank. back. The report's second part argues that the causes of 12. UNAIDS. 2002. Report on the Global HIV/AIDS Epidemic. Geneva. slow progress lie elsewhere--in low levels of government 13. Barro, R.J., and J.-W. Lee. 2000."International Data on health spending and, even more, in poor policies and insti- Educational Attainment Updates and Implications." NBER tutions across the entire health sector. Working Paper 7911, National Bureau of Economic Research, NewYork. 14. UNICEF. 2003. Statistics:Water and sanitation. References http://www.childinfo.org/eddb/water.htm. 1. Claeson, M., and E. Bos. 2002. Health, Nutrition and 15. ORC Macro. 2003. Demographic and Health Surveys. Population Development Goals: Measuring Progress Using the http://www.measuredhs.com/. Poverty Reduction Strategy Framework. Report of aWorld Bank 16. UNICEF. 2003. Statistics: Child mortality. http://www. Consultation, November 28­29.World Bank, Health, Nutrition childinfo.org/cmr/revis/db2.htm. and Population Department,Washington, DC. 17. World Health Organization. 2003. GlobalTuberculosis Control. 2. Short, C. 1999."Better Health for the Poor of theWorld." Geneva. Speech to the King's Fund. London, Department for 18. United Nations. 2004."Millennium Development Goals." New International Development. York. 46 | The Millennium Development Goals for Health CHAPTER 3 Effective Interventions Exist--They Need to Reach More People Suneeta, a Bengali woman living in a rural village, falls ill with functioning information system may be in place to register that chronic cough, fever, weight loss, and other problems. Others a case was detected and treated. No supervisors visit the facility around her may have previously been ill with similar symptoms. to provide training or feedback on performance. She may delay seeking care due to other duties or a lack of funds It is not the lack of an effective intervention for tuberculosis to pay for travel or consults. She may purchase medicines from a that explains this sequence of events. It is the delays in seeking drug seller or visit a traditional healer or private physician. She care, the insufficiently skilled or inadequately paid staff, the lack may take drugs sporadically until her money runs out,her symp- of patient and community health education, the poor supply toms remain, and she becomes further incapacitated. Her family chains, the failure to report the tuberculosis case, the lack of fol- or neighbors may then bring her to a health center or hospital, low-up with the patient or her close contacts,and the poor super- where she sees a nurse auxiliary who may or may not refer her to vision in the use of the medicines that explain the risk of failed a physician.The physician may charge her for an X-ray or treat treatment, chronic disease, or death for the patient.This situation her with general antibiotics and send her home or refer her for a can also contribute to the spread of infection in the community free sputum smear exam if available. and to the emergence of drug-resistant organisms, which pass Suneeta is found to have active infectious tuberculosis.The from community to community and across borders. physician or nurse may tell her about tuberculosis--or due to a lack of interest or time,tell her only that she needs to take more It is not a lack of interventions that is the main obstacle medicines.They may give her a month's worth to take home or to faster progress toward the Millennium Development tell her to attend the health clinic every day, even though it is Goals--it is the low levels of use, especially among the two hours to get there from her home.Worse yet, they may tell poor, of existing effective interventions. Globally, the her to buy the medicines elsewhere,if no drugs are in stock.She use of effective interventions for child health is typically might get only two of the four medicines she needs, and family below 50 percent, and in many poor countries the fig- members who may also have become infected or ill may not be ure is much lower. Using all known interventions seen by the health provider.The health provider may not be appropriately--achieving 99 percent coverage rates-- able to follow up to see whether Suneeta tolerated the medi- could avert 63 percent of child deaths and 74 percent of cines, took them regularly, was cured, remained ill, or died. No maternal deaths. 47 Effective interventions exist-- Figure 3.1 Why children die for all the health targets 40 Proportion of deaths due Interventions are being developed and tested for each of the to being underweight (%) 35 Millennium Development Targets, with the potential to 30 substantially reduce mortality and ill health.That is for the deaths 25 future--though in some cases the not-too-distant future. child 20 What matters for the present is the impressive array of of 15 interventions now available to fight child malnutrition, 10 child mortality,maternal mortality,and communicable dis- * ease mortality (table 3.1). Not all the evidence is equally 5 Proportion strong, and not all interventions are equally effective. But 0 overall the interventions listed in table 3.1 constitute a Malaria Other AIDS powerful arsenal for preventing and treating the main Neonatal Diarrhea Measles disorders Pneumonia Unknown causes of death and malnutrition. Cause of death Take child mortality. Three causes of death--diarrhea, Note: Bars indicate uncertainty bounds. pneumonia, and malaria--account for 52 percent of *Work in progress to establish cause-specific contribution of being under- deaths worldwide (figure 3.1). Neonatal causes account weight to neonatal deaths. for another 33 percent, measles for 3 percent, and AIDS Source: Reference 2. for 1 percent. Malnutrition is an underlying cause in Reprinted with permission from Elsevier (The Lancet, 2003, vol. 361, page 6). Table 3.1 Effective interventions for reducing illness, deaths, and malnutrition Goal Preventive interventions Treatment interventions Reduce child Breastfeeding; hand-washing; safe disposal of stool; use of latrine; Case management with oral rehydration therapy mortality safe preparation of weaning foods; use of insecticide-treated nets; for diarrhea; antibiotics for dysentery, complementary feeding; immunization; micronutrient pneumonia, and sepsis; antimalarials for malaria; supplementation (zinc and vitamin A); antenatal care, including newborn resuscitation; breastfeeding; steroids and tetanus toxoid; antimalarial intermittent preventive complementary feeding during illness; treatment in pregnancy; newborn temperature management; micronutrient supplementation (zinc and nevirapine and replacement feeding; antibiotics for premature vitamin A) rupture of membranes; clean delivery Reduce Family planning, intermittent malaria prophylaxis, use of Antibiotics for preterm rupture of membranes, maternal insecticide-treated bednets, micronutrient supplementation (iron, skilled attendance (especially active management mortality folic acid, calcium for those who are deficient) of third stage of labor), basic and emergency obstetric care Improve Exclusive breastfeeding for six months, appropriate complementary Appropriate feeding of sick child and oral nutrition child feeding for next 6­24 months, iron and folic acid supplemen- rehydration therapy, control and timely tation for children, improved hygiene and sanitation, dietary treatment of infectious and parasitic diseases, intake of pregnant and lactating women, micronutrient supplemen- treatment and monitoring of severely malnour- tation for prevention of anemia and vitamin A deficiency for ished children, high dose treatment of clinical mothers and children, anthelminthic treatment in school-age children signs of vitamin A deficiency Prevent and Safe sex, including condom use; use of unused needles by drug Treatment of opportunistic infections, combat users; treatment of sexually transmitted infections; safe, screened cotrimoxazole prophylaxis, highly active HIV/AIDS blood supplies; use of antiretrovirals in pregnancy to prevent antiretroviral therapy, palliative care maternal to child transmission and after occupational exposure Prevent and Directly observed treatment of infectious cases to prevent Directly observed treatment to cure symptomatic combat transmission and emergence of drug resistant strains and treatment cases, including early cases of tuberculosis tuberculosis of contacts, BCG immunization. Prevent and Use of insecticide-treated nets, indoor residual spraying (in Rapid detection and early treatment of combat epidemic-prone areas), intermittent presumptive treatment of uncomplicated cases, treatment of complicated malaria pregnant women cases (such as cerebral malaria and severe anemia) Note: "Intervention" refers to the direct action that leads to prevention or cure. The act of immunization, not the vaccine itself, is the intervention. Case manage- ment, not medicines or vitamins, is the intervention. Counseling for safer sex is not an intervention in this sense--safer sex is. 48 | The Millennium Development Goals for Health nearly half of under-five deaths.For each of these causes of Goals (box 3.1). In short, there is substantial underuse of death, there is at least one proven effective preventive effective interventions in the developing world. intervention and at least one proven effective treatment Just as shortfalls in coverage vary across countries, they intervention (figure 3.2). And each can be delivered in a vary within countries.The poor and other deprived groups low-income setting. In most cases, there are several proven invariably lag behind the population in many preventive and effective interventions. For example, for diarrhea--the curative health interventions.They are less likely to receive second-leading cause of child deaths--there are no fewer full basic immunization coverage (bacillus of Calmette and than five proven preventive interventions (breastfeeding, Guerin [BCG],measles,and DPT) (figure 3.7);to have their complementary feeding, hygiene and use of safe water, deliveries attended by a medically trained person; or to have zinc, and vitamin A) and no fewer than three proven treat- at least one antenatal care visit to a medically trained person. ment interventions (oral rehydration therapy, antibiotics, The trends in use by wealth group are more encouraging. and zinc). Many of these interventions can be--and are-- For the most part, the poor are making more rapid progress delivered by households, but many are not. than the better off, who already have high coverage rates for Similar arguments can be made for the other goals. many interventions (see tables in appendix D). Progress has Some interventions for maternal mortality are delivered been more propoor in professionally delivered interventions by the household (nutrition, family planning), but most (skilled attendance, treatment of common childhood ill- are not (figure 3.3).And for those delivered by households, nesses) than in home-delivered interventions (breastfeeding, there is typically little--if any--scientific evidence avail- timely complementary feeding,and vitamin A supplementa- able on impact. tion) (figure 3.8). Because of the multiple and interrelated causes of child- hood and maternal deaths, integrated strategies, such as The underuse of effective Integrated Management of Childhood Illness (IMCI) and interventions costs lives Integrated Management of Pregnancy and Childbirth (IMPAC) have been introduced in many countries. The low use of effective interventions--in the developing Delivery of these bundled interventions has the potential world and among the poor--translates into rates of mor- to yield greater efficiency, synergy, and impact--if imple- tality, morbidity, and malnutrition that are far higher than mented on scale and not underused. they need be. Take child mortality.If the use of all the proven effective childhood preventive and treatment interventions were to Effective interventions are rise from its current level to 99 percent (95 percent for underused--especially by the poor breastfeeding), the number of under-five deaths world- The proximate cause of the high rates of death and malnu- wide would fall dramatically (figure 3.9). Deaths from trition in the developing world is that, for one reason or malaria and measles could be all but eliminated,and deaths another, people do not receive the interventions that could from diarrhea, pneumonia, and HIV/AIDS could be dra- save their life or make them well nourished. In upper mid- matically reduced. Overall, a reduction in childhood dle-income and high-income countries,90 percent of chil- deaths of 63 percent could be achieved by increasing cov- dren receive diphtheria, pertussis, and tetanus (DPT3) erage rates to 99 percent. vaccinations; more than 90 percent of babies are delivered Of these 6 million averted deaths worldwide,perhaps half by a medically trained person; and more than 90 percent of would result from increased coverage of home-delivered pregnant women have at least one antenatal visit (fig- interventions.*This fraction will vary according to the rela- ure 3.4). In low-income and lower middle-income coun- tive importance of different causes of death. In Brazil and tries, the figures are much lower--often dramatically so. In China, just under half of child deaths are neonatal,2 for South Asia less than 50 percent of pregnant women receive which the relevant effective interventions are professionally antenatal checkups,and only 20 percent of babies are deliv- delivered. In much of Sub-Saharan Africa, by contrast, ered by a medically trained person (figure 3.5). In some countries less than 20 percent of children with acute respi- *This is a preliminary estimate, based on the figures in table 2 of Jones and ratory infections are seen by a medically trained person others.1 It classifies breastfeeding, insecticide-treated materials, complemen- (figure 3.6)--despite the fact that pneumonia, the most tary feeding, hygiene and use of clean water, vitamin A supplementation, oral rehydration therapy, and antimalarials as home-delivered interventions. severe respiratory infection, can be treated with an antibi- The figure is a tentative estimate because unlike table 3 in Jones and others, otic.The story is similar for other childhood interventions no account is taken in table 2 of competing risks (in principle, a child could and for interventions for other Millennium Development be saved several times over in this calculation). Effective Interventions Exist--They Need to Reach More People | 49 1 2 1 1 sepsis Neonatal 1 1 Neonatal tetanus 1 1 2 2 1 Preterm delivery asphyxia 2 deaths Birth 1 under-five HIV/AIDS of killers Causes 1 1 2 1 Malaria evidence) childhood 1 2 1 1 Measles evidence) indirect evidence or no against or (sufficient (limited 1 1 1 12 1 1 2 impact Pneumonia Level Level No 1 1 1 1 1 1 1 1 1 2 interventions Diarrhea fective ef pregnancy of feeding in rapture 1. management ence interventions materials arsenal hygiene treatment interventions feeding therapy replacement eferr premature pneumonia sepsis dysentery The prev. steroids and for for for for omfr temperature resuscitation vaccine Preventive sanitation, A toxoid A delivery Treatment 3.2 vaccine rehydration e Adapted Breastfeeding Insecticide-treated Complementary Water, Hib Zinc Vitamin Antenatal Newborn Tetanus Nevirapine Antibiotics Clean Measles Antimalarial Oral Antibiotics Antimalarials Antibiotics Newborn Antibiotics Zinc Vitamin Figur Source: 50|The Millennium Development Goals for Health 2 2 Other Tetanus 2 2 2 2 1 Anemia 2 1 Malaria death 1 maternal Abortion of complications Causes 1 2 1 labour evidence) Obstructed mortality evidence) indirect 2 2 1 2 2 2 1 or maternal Eclampsia (sufficient (limited 1 2 2 1 1 1 2 impact reducing Puerperal infection Level Level No for 2 1 22 1 1 2 11 1 2 mortality Hemorrhage nal Interventions for of in and in the mater related to beyond) in hemorrhage hemorrhage pregnancy and pregnancy prevention weeks) or mild postnatal) pregnancy for pregnancy vaginosis 37 for pregnancy anticonvulsants supplements in during (skilled pregnancy weeks during weeks educingr nets other pregnancy in disorders during (before (36 bacterial during management postpartum postpartum 12 in energy therapy administration prophylaxis services and for deficiency and and treated drug primary pre-eclampsia iron caregivers preterm prelabor treating malaria delivery expectant labor of secondary hours for hypertensive sulphate with protein supplementation of for membranes for membranes for hypertension of Misoprostol of 24 Interventions supplementation A supplementation insecticide of of planning versus supplementation of birth stage rectal women Iron Folate Balanced pregnancy Treatment Calcium preventing problems Vitamin Use malaria Intermittent Family Continuity child Antibiotics rupture Antibiotics rupture Antibiotics pregnancy Antihypertensive moderate Professional attendant) Active third Management with Treatment (between Magnesium for Interventions 3. ence 3.3 care care clinical e care essential Refer Preventive essential Figur Primary Basic obstetric Comprehensive obstetric Source: Effective Interventions Exist--They Need to Reach More People | 51 Figure 3.4 Low income countries lag behind on Figure 3.6 Some children with acute respiratory key preventive interventions for maternal and infections receive treatment--others don't child health Level (%) 100 0 20 40 60 80 100 90 Tanzania 1999 Jordan 1997 80 South Africa 1998 70 Zambia 2001 Vietnam 1997 60 Indonesia 1997 Namibia 1992 50 India 1999 Uganda 2000 40 Pakistan 1990 Egypt, Arab Rep. 2000 30 Peru 2000 20 Philippines 1998 Nicaragua 2001 10 Kenya 1998 Comoros 1996 0 Zimbabwe 1999 Exclusive DPT3 Safe Antenatal care Gabon 2000 breastfeeding immunization deliveries Brazil 1996 to four Bolivia 1998 months C. African Rep. 1994­95 Mauritania 2000 Proportion of Low income Upper middle income Guinea 1999 children with acute respiratory Lower middle income High income Côte d`Ivoire 1994 Mozambique 1997 infections seen Haiti 2000 medically Source: Reference 4. Guatemala 1998 Nigeria 1990 Eritrea 1995 Madagascar 1997 Cambodia 2000 Mali 2001 Figure 3.5 Use of key preventive interventions Cameroon 2000 for maternal and child health is lower in some Yemen, Rep. 1997 Bangladesh 2000 regions than others Malawi 2000 Ghana 1998 100 Togo 1998 Nepal 2001 90 Niger 1998 Chad 1996­97 80 Morocco 1992 Ethiopia 2000 70 Rwanda 2000 60 Source: World Bank staff calculations and reference 5. 50 40 30 Spectacular reductions in mortality could also be 20 achieved if coverage rates of the key maternal mortality 10 interventions in figure 3.3 were increased from current 0 levels to 99 percent. It is estimated that perhaps as many as Exclusive DPT3 Safe Antenatal breastfeeding immunization deliveries care 1391,000 deaths worldwide--74 percent of current mater- to four months nal deaths--could be averted, all from increased coverage East Asia and Pacific Middle East and North Africa of professionally delivered interventions. Four stand out as Europe and Central Asia South Asia Latin America and the Caribbean Sub-Saharan Africa especially important (figure 3.10):access to essential obstet- ric care (52 percent of deaths averted), access to safe abor- Source: Reference 4. tion services (16 percent), active rather than expectant management in the third stage of labor (10 percent), and the use of magnesium sulphate and other anticonvulsants neonatal deaths account for about a quarter of child deaths, for women with pre-eclampsia (8 percent). and malaria and diarrhea--two causes of death for which household-delivered interventions could be important-- These are preliminary estimates.The details will be spelled out in a separate together account for more than half of child deaths.2 technical note. 52 | The Millennium Development Goals for Health Box 3.1 Low coverage of HIV/AIDS Figure 3.7 Proportion of children age 12­23 mos interventions who received full basic immunization coverage-- poorest 20 percent vs. population as a whole Most countries have developed strategic frameworks for HIV Utilization level (%) prevention, but only a fraction of people at risk have meaning- 0 20 40 60 80 100 ful access to basic prevention and treatment. Egypt, Arab Rep. 2000 Uzbekistan 1996 Risk reduction behavior among young people Rwanda 2000 Kyrgyz Rep. 1997 Survey results indicate that condom use with nonregular part- Kazakhstan 1999 Guatemala 1998 ners is higher in urban than in rural areas and higher among Armenia 2000 young men than among young women. Data also suggest that Malawi 2000 Zambia 2001 condom use varies considerably across countries, ranging from Zimbabwe 1999 2 percent to 88 percent in Sub-Saharan Africa. In this region, Nicaragua 2001 Philippines 1998 15­20 percent of young people report having had sexual inter- Peru 2000 course before the age of 15, with young women reporting an Brazil 1996 Nepal 2001 earlier median age of first sex than males. Namibia 1992 Morocco 1992 Tanzania 1999 Management of sexually transmitted infections South Africa 1998 Bangladesh 2000 Because untreated sexually transmitted infections increase the Colombia 2000 Ghana 1998 risk of HIV transmission by several orders of magnitude, con- Kenya 1998 trol of sexually transmitted infection is a fundamental element Indonesia 1997 Vietnam 1997 of effective HIV prevention. Yet from limited information Comoros 1996 Population received, only one in four countries in Sub-Saharan Africa Bolivia 1998 Cambodia 2000 Poorest 20% reports that at least 50 percent of sexually transmitted infection Turkey 1998 patients are appropriately diagnosed, counseled, and treated. Uganda 2000 Haiti 2000 Eritrea 1995 Prevention of mother-to-child transmission Cameroon 1998 Pakistan 1990 Four years after research indicated that a relatively inexpensive Togo 1998 Madagascar 1997 single dose of nevirapine to mother and newborn significantly Jordan 1997 reduced the odds of HIV transmission to the infant, prevention India 1999 Paraguay 1990 of mother-to-child transmission (PMTCT) coverage remains Mozambique 1997 virtually nonexistent in many heavily affected countries.Apart Mali 2001 C. African Rep. 1994­95 from Botswana, where coverage reached 34 percent by the end Guinea 1999 of 2002, PMTCT remains extremely low in countries hardest Côte d`Ivoire 1994 Mauritania 2000 hit by HIV/AIDS. Nigeria 1990 Yemen, Rep. 1997 Ethiopia 2000 Antiretroviral therapy Gabon 2000 Niger 1998 While an estimated 6 million people in low- and middle- Chad 1996­97 income countries currently need antiretroviral therapy, only Source: World Bank staff calculations and reference 5. 300,000 were obtaining such therapy by the end of 2002. Although coverage remains low in Sub-Saharan Africa, Botswana, Cameroon, Nigeria, and Uganda have made serious efforts to increase antiretroviral therapy coverage through the The challenge public and private sectors.Caribbean countries report coverage of less than 1 percent. In Asia, where more than 7 million peo- If the lack of interventions is not holding countries back ple are living with HIV/AIDS,no country has exceeded 5 per- from achieving the goals, what is? Part II of the report cent coverage. investigates this question. Chapter 4 argues that a lack of Source: World Bank. government health spending is only part of the story of low coverage rates. It makes the case that governments need to improve their targeting of health spending, but even more they need to improve their policies and the institutions of the health sector--very broadly defined. The rest of part II shows how. Effective Interventions Exist--They Need to Reach More People |53 Figure 3.8 Use trends for household-delivered Figure 3.9 Full use of existing interventions and professionally delivered interventions would dramatically cut child deaths Average annual percentage change 12 Unpreventable with existing 0 10 20 30 40 interventions Preventable with existing 10 Exclusive interventions breastfeeding 8 Oral rehydration (millions) therapy Household- delivered 6 deaths Complementary of feeding 4 Skilled Number attendance 2 Treatment of fever 0 Full immunization Other Total Diarrhea Malaria HIV/AIDSMeaslesNeonatal Professionally Pneumonia disorders delivered Antenatal care Poorest 20% Source: Reference 1. Treatment of Population acute respiratory infections Figure 3.10 Full use of existing interventions Treatment of would dramatically cut maternal deaths diarrhea Improved access to safe abortion services Tetanus toxoid Improved access to comprehensive essential obstetric care Tetanus toxoid immunization Source: World Bank staff calculations and reference 5. Treatment for iron deficiency Drugs for preventing malaria Active management in third stage of labor References Magnesium sulphate for pre-eclampsia 1. Jones, G., R.W. Steketee, R.E. Black, Z.A. Bhutta, and S.S. Calcium supplements during Morris. 2003."How Many Child Deaths CanWe PreventThis pregnancy Year?" Lancet 362 (9377): 65­71. Antibiotics for treating bacterial vaginosis 2. Black, R.E., S.S. Morris, and J. Bryce. 2003."Where andWhy Antibiotics for preterm rupture of Are 10 Million Children Dying EveryYear?" Lancet 361 (9376): membranes 2226­2234. 0 10 20 30 40 3. Ramana, G. 2003. Personal communication. Senior Public Percentage of maternal Health Specialist, Health, Nutrition and Population deaths averted Department,World Bank,Washington, DC. Hemorrhage Abortion complications 4. World Bank. 2003. World Development Indicators 2003. Puerperal infection Malaria Washington, DC:World Bank. Eclampsia Anemia 5. ORC Macro. 2003. Demographic and Health Surveys. Obstructed labor Tetanus http://www.measuredhs.com/. Source: World Bank staff estimates. 54| The Millennium Development Goals for Health CHAPTER 4 Extra Government Health Spending Is Necessary but Not Sufficient-- Health Sector Strengthening Is Also Required, and Spending Needs to Be BetterTargeted Views diverge widely on government health spending. One government health programs has an impact on Millennium view holds that low government health spending in the devel- Development Goal outcomes in well-governed countries. But oping world is the cause of poor health indicators.According to the impact of such across-the-board increases on health outcomes this view, government health spending needs to be increased would be modest, so substantial increases in spending would be substantially and financed largely through development assis- needed. Carefully targeted spending can improve Millennium tance.1 At the other extreme is the view that government Development Goal outcomes even in poorly governed countries. health spending has little impact on health outcomes at the But policies and institutions need to be strengthened if across- margin,reflecting the multitude of weak links in the chain run- the-board increases in spending are to be productive. ning from government spending to health outcomes.2­4 Shaping strong policies and institutions for achieving the According to this view, more government health spending goals requires broad thinking, across a wide variety of policy would be unproductive; the focus should instead be on instruments and areas.They include those not normally consid- strengthening the weak links. ered integral parts of the health sector, such as households and This chapter argues for a more nuanced view. Adding to communities, and those outside the health sector, such as water, government health budgets and scaling up proportionally all sanitation, transport, and education. 55 consistent with recent World Bank studies on aid effec- The limits and returns to government tiveness.10,11 In countries with "good" policies and institu- health spending tions (strong property rights, absence of corruption, a The impact of government spending on mortality depends good bureaucracy), an extra 1 percent of GDP in aid has in part on its impact on coverage rates--and in part on the been estimated to lead to a decline in infant mortality of impact of coverage rates on mortality.The case for substan- 0.9 percent. By contrast, in countries in which policies are tial increases in government health spending has been only average, the impact is just 0.4 percent.Where policies weakened by the absence of firm evidence on the first of are "bad," aid has no statistically significant effect on infant these links.It is also weakened by the failure to factor in the mortality. contributions and resource requirements associated with These findings are also borne out in the results of a changes in other sectors (water, sanitation, roads, female study undertaken for this report (see table 4.1 and appen- education).5 dix A),in which the quality of policies and institutions was The case against extra government health spending is measured by the World Bank's Country Policy and built on empirical evidence that shows that additional gov- Institutional Assessment (CPIA) Index. As policies and ernment health spending has no perceptible impact on institutions improve, this index increases from 1 to a maxi- infant and under-five mortality, once other influences are mum of 5 (figure 4.1). As the second panel of table 4.1 held constant.2 But this analysis has limitations. It refers shows, the elasticity also increases in absolute size--gov- simply to child mortality, not to health outcomes in gen- ernment spending has a larger impact on health outcomes eral. Child mortality differs from, say, maternal mortality in at the margin in better-governed countries. For example, that several of the interventions aimed at children are deliv- at a CPIA score of 4 (one standard deviation above the ered by the household, while almost all interventions mean), a 10 percent increase in the share of GDP devoted aimed at mothers are delivered by professional providers. It to government health spending results in a 7.2 percent is unlikely that, say, maternal mortality is as unresponsive decline in the maternal mortality ratio. (or inelastic) to government health spending as child mor- The elasticities in the second panel of table 4.1 reinforce tality is.The results are also very general:they indicate what the finding that across-the-board increases in government happens to child mortality among the population as a health spending have no perceptible impact on under-five whole, in an average country.As a result, they hide signifi- mortality in poorly governed countries.But they also show cant spending effects among specific subpopulations and something new: the same is true of other Millennium specific types of countries. Finally, the results indicate what Development Goal outcomes. Whether the outcome is would happen to child mortality if additional government underweight, maternal mortality, or tuberculosis mortality, spending were to take the form of a proportional scaling- in countries one standard deviation below the mean CPIA up of all government health programs.They do not show score, across-the-board additions to government health what would happen if extra spending focused on specific spending have no significant effect. Indeed, for malnutri- subpopulations or specific programs. tion and tuberculosis mortality, the CPIA score has to get above the population-weighted average of 3.5 for addi- tional government health spending to have any bite. The importance of good The message, then, is clear. In countries with poor gov- policies and institutions ernance--and weak policies and institutions--improve- For the developing world as a whole,then,and for popula- ments need to occur in these areas if additions to tions as a whole, proportional scaling-up of government government health budgets are to have any impact.These health programs has no perceptible impact on child mor- improvements may require reforms that entail transition tality, once other influences are held constant.2 But recent costs, which may need to be covered by government evidence suggests that scaling up does make a difference spending.And it may be possible to phase in the additional for specific subpopulations and specific types of countries. government health spending as these reforms are being Child mortality among the poor--whether the poorest implemented. But the fact remains: in these countries, 20 percent of the population or those living under $1 (or extra government health spending by itself will do little, if $2) a day--does improve with additional government anything, to accelerate movement toward the Millennium health spending.6­8 And among countries with good gov- Development Targets.This is why the rest of part II--the ernance (measured by the quality of the bureaucracy), bulk of the report--is devoted to the difficult but often additional government health spending reduces child overlooked question of how to shape effective policies and mortality (see the first panel of table 4.1).9 This finding is institutions in the health sector. 56 | The Millennium Development Goals for Health Table 4.1 Elasticities of Millennium Development Goal outcomes to government health spending, as a percent of GDP, by quality of policies and institutions Maternal mortality Underweight among Tuberculosis Governance score Under-five mortality ratio children under five mortality Quality of bureaucracy index 2.2 ­0.350 -- -- -- 3.7 ­0.450* -- -- -- 5.2 ­0.560* -- -- -- CPIA index 1.0 0.799 ­0.622 0.130 0.651 2.0 0.507 ­0.654 ­0.087 0.276 3.0 0.215 ­0.687 ­0.305 ­0.098 3.25 0.142 ­0.695* ­0.360 ­0.192 3.5 0.069 ­0.703* ­0.414 ­0.285 4.0 ­0.077 ­0.720* ­0.523* ­0.472 4.5 ­0.223 ­0.736* ­0.632* ­0.659* 5.0 ­0.369 ­0.752* ­0.740* ­0.847* -- Not available. Note: An elasticity of ­0.622 means that a 10 percent increase in spending reduces mortality by 6.22 percent. The values 2.2, 3.7, and 5.2 are, respectively, one standard deviation below the mean, the mean, and one standard deviation above the mean. The means are based on a sample that includes industrial and develop- ing countries. The population-weighted mean CPIA score is 3.5, with a standard deviation of 0.4. These figures are based on developing countries only, though industrial countries were also included in the regression (with CPIA values set at 5). *Significantly different from zero at the 90 percent confidence level or above. Source: Elasticities for the quality of bureaucracy index are from reference 9. Elasticities for the CPIA Index are based on a study undertaken for this report (see appendix A for details). More government spending in well- governed countries--how large are the returns relative to the Millennium Development Goal challenge? Figure 4.1 CPIA scores across Bank regions In well-governed countries, and ones with good policies 50 1 and institutions, there is scope for improving health out- 45 2 3 comes through a proportional scaling-up of government 40 4 health programs.According to the upper panel of table 4.1, 35 extra government health spending would reduce child mor- 30 tality in countries with above-average governance condi- countries of 25 tions. According to the lower panel, it would reduce 20 maternal mortality ratio in countries with above-average Number 15 CPIA scores and reduce malnutrition and tuberculosis mor- 10 tality in countries with CPIA scores that are respectively 5 1.00 and 1.75 standard deviations above the mean. So well- governed countries in the developing world could, in prin- 0 East Europe Latin Middle South Asia Sub- ciple, accelerate progress toward the health goals simply by Asia and America East Saharan and Central and and North Africa adding to government health budgets. But how much extra Pacific Asia the Africa momentum could they achieve by doing so? Or to put the Caribbean question the other way around, how much extra would Source: World Bank staff estimates. they need to spend to hit the Millennium Development Targets? Extra Government Health Spending Is Necessary but Not Sufficient |57 What would extra government health spending in share of GDP, assuming the country in question has a well-governed countries actually achieve? CPIA score of 4 (one standard deviation above the mean). Figure 4.2 shows how much faster child malnutrition, Also shown are the current rate of reduction of the under-five mortality,and maternal mortality would fall per Millennium Development Indicator, the estimated contri- year if an additional 2.5 percentage points were added to butions from faster economic growth and quicker growth the annual growth rate of government health spending as a in girls' secondary school completion and access to drink- Figure 4.2 The contributions of faster growth of government health spending in countries with good policies and institutions Underweight Under-five mortality (2) 4 0 rate 2 ­1 rate ­2 0 growth ­3 growth ­2 ­4 ­4 ­5 percentage ­6 ­6 percentage ­7 ­8 Annual ­8 East Europe Latin Middle South Sub-Saharan Annual ­10 Asia and America East Asia Africa and Central and the and ­12 Pacific Asia Caribbean North East Europe Latin Middle South Sub-Saharan Africa Asia and America East Asia Africa and Central and the and Extra government Current Pacific Asia Caribbean North health spending Required 2000­15 Africa Extrasectoral Target Extra government Current contributions health spending Required 2000­15 Extrasectoral Target contributions Maternal mortality Under-five mortality (1) 0 0 rate rate ­1 ­2 ­2 growth ­4 growth ­3 ­6 ­4 ­5 ­8 percentage percentage ­6 ­10 ­7 Annual ­12 Annual ­8 East Europe Latin Middle South Sub-Saharan East Europe Latin Middle South Sub-Saharan Asia and America East Asia Africa Asia and America East Asia Africa and Central and the and and Central and the and Pacific Asia Caribbean North Pacific Asia Caribbean North Africa Africa Extra government Current Extra government Current health spending Required 2000­15 health spending Required 2000­15 Extrasectoral Target Extrasectoral Target contributions contributions Note: Panels b and c are based on the elasticities in the first and second panels of table 4.1. For government health spending, the figure shows the effects on annual reductions in mortality and malnutrition of adding 2.5 percentage points to the growth of the share of GDP devoted to government health spending. In panel a the elasticity used corresponds to a quality-of-bureaucracy score of 3.7, the mean for all countries in the sample, including developing countries. The actual score is probably somewhat closer to one standard deviation above the mean for developing countries. In panel b the elasticity corresponds to a CPIA score of 4, roughly one standard deviation above the mean. The extrasectoral contributions are from table 2.1. Source: World Bank staff estimates. 58 | The Millennium Development Goals for Health ing water, the target rate of reduction, and the required an additional 2.5 percentage points were added to the rate rate of reduction for the period 2000­15 given progress in of growth of government health spending as a share of the 1990s. For example, in East Asia and the Pacific under- GDP. If the second set is used, the extra government five mortality fell by only 2.7 percent a year in the 1990s, health spending would have little--if any--impact. so it would need to fall by 5.3 percent a year in 2000­15 How much would reaching the Millennium for the region to achieve the targeted two-thirds reduction Development Goals cost in well-governed countries? between 1990 and 2015. For the maternal mortality ratio, the contribution from Another way of looking at the spending question is to ask an additional 2.5 percentage points of growth in govern- how much faster government health spending as a share of ment health spending as a share of GDP would be sub- GDP would need to grow to hit the Millennium Develo- stantial, pushing two regions (Europe and Central Asia and pment Targets in well-governed countries. Would such South Asia) up to their required rates of reduction for extra growth be reasonable? Table 4.2 presents estimates 2000­15. For malnutrition the contribution would also be for two scenarios--one in which the expected economic appreciable but unnecessary given trends to date, except in growth materializes and the other relevant Millennium Sub-Saharan Africa, where it would be insufficient to get Development Goals are met, another in which growth the region to the target. For under-five mortality the esti- fails to materialize and there is no acceleration in the rate mated contribution of extra health spending depends on of progress toward these other goals. Obviously in the sec- the set of elasticity estimates used. If the first set is used, ond case, the costs of reaching the health goals will be Latin America and the Caribbean and the Middle East and higher, because the health sector cannot ride as much on North Africa would hit the under-five mortality targets if the coattails of other sectors. Table 4.2 How far would the share of GDP devoted to government health spending need to rise in well-governed countries to meet the Millennium Development Goals? Assumes that economic growth Assumes that economic growth continues at current rate and no accelerates and other acceleration occurs in progress Millennium Development toward other Millennium Goals are met Development Goals Annual Assumes percentage that 1990s change trend Under- Under-five Maternal Under- Under-five Maternal Region 2000 values in 1990s continuesa weight mortality mortality weight mortality mortality East Asia and 1.7 ­0.2 1.7 1.7 3.7 1.7 1.7 3.9 2.2 the Pacific Europe and 3.5 0.5 3.8 3.8 6.3 4.5 3.8 9.9 5.5 Central Asia Latin America and the 3.4 4.6 6.6 6.6 9.0 22.3 6.6 9.1 22.3 Caribbean Middle East and North 2.3 6.2 5.7 5.7 8.1 5.7 5.7 8.3 5.7 Africa South Asia 0.9 0.7 1.0 1.0 2.1 1.4 1.0 2.6 2.4 Sub-Saharan 1.8 1.2 2.2 6.4 12.2 4.6 6.8 16.5 7.4 Africa Developing 1.6 1.0 2.1 2.6 4.6 3.8 2.6 5.5 4.6 countries Note: Averages are population weighted. Under-five mortality estimates based on elasticity in first panel of table 4.1 corresponding to a quality-of-bureaucracy score of 3.7. This is the mean for all countries in the sample, including developing countries. The figure for developing countries is probably closer to one standard deviation above the mean. For underweight and maternal mortality, elasticities correspond to a CPIA score of 4, roughly one standard deviation above the mean. Developing world averages are weighted by regional populations. aAssumes compound growth. Source: Baseline data are from reference 30. Estimated 2015 values are derived using regressions in appendix A. Extrasectoral contributions from table 2.1. Extra Government Health Spending Is Necessary but Not Sufficient |59 In East Asia and the Pacific the share of GDP devoted to the share of GDP to be devoted to government health government health spending was a little over 1.7 percent spending are impressively large. In the few well-governed in 2000, and the share shrank at ­0.2 percent a year in the countries in Sub-Saharan Africa, the share might need to 1990s. If this trend were maintained, the share would be be nearly eight times what it is predicted to be in 2015 about 1.67 percent in 2015. If economic growth is as based on current trends. In poorly governed countries, expected in the new millennium (a slight reduction is then, extra spending alone will not result in faster progress anticipated for East Asia and the Pacific),and the other rel- toward the goals. In well-governed countries it will, but evant goals are reached, East Asia and the Pacific would hit the implied resource increases are unrealistically high.This the underweight and maternal mortality ratio targets even second group of countries faces two options: raising the without additional health spending. It would, however, productivity of existing spending and targeting additional miss the under-five mortality target--an additional five spending carefully. percentage points would need to be added to the rate of growth of the government health share of GDP to hit the Improving expenditure target, taking it to 3.7 percent in 2015. If there were no allocation and targeting change in economic growth in the new millennium and no acceleration in progress toward the other goals, the rate It is important to be clear about what the elasticities in of growth of the government health share of GDP would table 4.1 show.They are estimates of the effects at the mar- have to rise to 3.9 percent, more than twice the projected gin of adding to the government's health budget on the 2015 value based on 1990s trends. assumption that the additional spending will be allocated For all regions except Latin America and the Caribbean across programs and institutions in proportion to current the binding constraint is under-five mortality--govern- allocations. From the viewpoint of expanding health ment spending has to rise faster to hit this target than it budgets--through, say, budgetary support by external does to hit the others.And the results for under-five mor- donors--these calculations are clearly highly relevant. But tality may be overly optimistic, based on the elasticities in even here they may be misleading if budget support is the first panel rather than the second. Since the second set linked to policy changes that entail shifts in spending pat- of estimated elasticities is not statistically significant, the terns or a different spending pattern for the extra spend- extra spending required could be infinite. Looking across ing.The elasticities in table 4.1 do not tell us what health the developing world, well-governed countries could impacts will be achieved if, rather than making additions achieve the goals if they raised the share of GDP devoted to a ministry of health budget without changing the com- to government health spending to 4.6­5.5 percent in position of government health spending, additional 2015, depending on whether anticipated economic resources were instead spent in very specific and targeted growth materializes and the other relevant targets are hit. ways. This is two to three times the GDP share projected for Geographic targeting 2015 on the basis of 1990s trends. In Sub-Saharan Africa the implied increases in the GDP The poorest district in Lesotho receives only 20 percent share going to government health programs are even less of the amount the capital city receives in per capita allo- palatable. Even in the optimistic first scenario, the share cations of public expenditures on health--inequality that would need to rise to 12.2 percent by 2015. In the second is not resolved when accounting for nongovernmental scenario, the required share in 2015 would be even more services. In Peru per capita spending allocated through unrealistic--16.5 percent. And this applies only to coun- the regional budget (which excludes teaching hospital tries whose quality of policies and institutions is above allocations) is 66 percent higher in the Lima region than average. Sub-Saharan Africa has few of them: only five of in the very poor regions.12 In Bangladesh more devel- 47 Sub-Saharan countries had a CPIA in 2000 of at least oped districts receive more per capita than less developed 4. For the other Sub-Saharan countries, accelerated districts.13 progress toward the goals will not happen through extra Resource-allocation formulas can reduce government spending alone, however large the increases. spending gaps across regions and ensure that the poor also benefit from government spending. But they have an effi- The bottom line ciency angle too--resources can be diverted from areas So, yes, in well-governed countries, additions to govern- where the marginal benefit is probably fairly low (high-tech ment health budgets will by themselves lead to reductions hospitals in the capital city) to those where the marginal in malnutrition and mortality. But the implied increases in benefit is likely to be high (immunization in rural areas). 60 | The Millennium Development Goals for Health Such formulas have narrowed regional gaps in industrial early results are encouraging.16,17 A recent World Bank countries.14 In developing countries, too, they have begun study18 of the Millennium Development Goals in India to be used. Bolivia has used these formulas since 1994 as provides further support for the idea that the way govern- part of its decentralization efforts--and in its allocation of ment spending is allocated across programs affects its newly available Heavily Indebted Poor Country (HIPC) impact on Millennium Development Indicators. The Initiative resources. Funds went to municipalities on a per report finds that a one percentage point increase in the capita basis based on poverty indicators (the poorer the share of total government health spending devoted to pub- indicator, the larger the per capita amount allocated), with lic health and family welfare activities (prevention and con- the mandate that municipalities spend such resources on trol of diseases, population and family planning services, prespecified health, education, and other social programs. and maternal and child health programs, including immu- The program has been associated with some fairly large-- nization) leads to a reduction of about 0.8 infant deaths per and propoor--improvements in maternal and child health 1,000 live births. indicators (chapter 5). This effect--estimated holding constant the amount spent per capita on all medical, public health, and family Changing the allocation of spending across levels of care welfare activities--is more pronounced for poor states and Spending on health in developing countries is character- not statistically significant for nonpoor states. For the poor ized by a surprisingly high concentration of spending on states, a one percentage point increase in the share of total secondary and tertiary infrastructure and on personnel, government health expenditure on public health and fam- despite low bed-occupancy rates. Armenian hospitals, for ily welfare is associated with a decrease of 1 infant death example, receive more than 50 percent of the government per 1,000 live births. budget for health. Health clinics and ambulatory facili- Targeting specific population groups ties--according to household surveys, the preferred service providers for sick people in the poorest quintile--received Many countries subsidize all government health services just over 20 percent of expenditures.15 This pattern is also for everyone. These blanket schemes fail to target inter- seen in lower-income countries. In Tanzania government ventions that give rise to externalities, but they also fail to spending in hospitals accounted for about 60 percent of the benefit the poor disproportionately (box 5.8 in chapter 5). budget in 2000, compared with only 34 percent of spend- This is so despite the stronger equity case for subsidizing ing on preventive and primary care facilities. Recent gov- care for the poor--and the fact that the poor tend to bear ernment efforts to change this brought the respective a disproportionate burden of child and maternal mortality proportions to 43 percent and 48 percent in 2002. and malnutrition. (Chapter 5 provides a variety of exam- Simply reallocating the budget toward primary care ples of programs that have successfully targeted govern- need not result in higher payoffs to government health ment spending on poor and vulnerable groups, some of spending in lowering child and maternal mortality and which have been evaluated and found to have had an malnutrition.3,4 In many instances service providers have impact.) failed to deliver quality care or to use resources efficiently Using social investment funds to reach the poor (chapter 6). So even though many key Millennium Development Goal interventions can be and are delivered Targeting resources to poor regions and provinces may at lower levels of care, simply redirecting money toward require nontraditional mechanisms for setting priorities these facilities will not necessarily yield higher returns. and implementing programs. Social investment funds, The trick is to couple expenditure reallocations with mea- which have achieved important impacts on such health sures to improve the performance of primary care facilities outcomes as infant mortality rates (box 4.1), may fulfill and district hospitals (chapter 6) and measures to ensure such a role. that households actually demand the relevant interven- The returns to targeted spending tions (chapter 5). from removing bottlenecks Targeting specific programs Another approach is to assess--at the country level--health Programs such as Integrated Management for Childhood sector impediments to faster progress, identify ways to Illness and directly observed treatment, short-course remove them,and estimate both the costs of removing them (DOTS) therapy for tuberculosis are good examples of and the likely impacts of their removal on Millennium programs that may yield high returns to government Development Goal outcomes (box 4.2).22Work along these spending. Both are the subject of ongoing evaluation, and lines has begun in several African countries and in India. Extra Government Health Spending Is Necessary but Not Sufficient | 61 Box 4.1 Coupling targeting with institutional innovation through social investment funds Social funds can be defined as"agencies that finance small projects A recent evaluation20 of social funds was carried out in six in several sectors targeted to benefit a country's poor and vulnera- countries: Armenia, Bolivia, Honduras, Nicaragua, Peru, and ble groups based on...demand generated by local groups and Zambia.Among other specific objectives, the study evaluated the screened against a set of eligibility criteria."19 Bolivia introduced degree to which social fund interventions reach poor areas and the first such fund in 1986, when it established the Emergency poor households and affect living standards (as measured by edu- Social Fund.International donors soon recognized the potential of cation and health outcomes).The study used panel data on pro- the social fund as a channel for social investments in rural Bolivia ject beneficiaries and comparison groups, applying several and as an international model for community-led development. evaluation methodologies. In 1991 the Bolivian social fund began concentrating on The results of the study are interesting because they reveal delivering social infrastructure to historically underserved areas, better targeting (especially in Peru) and positive health impacts moving away from emergency-driven employment generation (especially in Bolivia).The geographic distribution of social fund projects. It proved that this type of institution could operate to expenditures was propoor in all countries studied, with poor dis- scale, bringing small infrastructure investments to vast areas of tricts receiving more per capita than wealthier districts and the rural Bolivia that line ministries had been unable to reach because very poorest districts receiving shares exceeding their shares of of their weak capacity to execute projects. Providing financing to the population. The high levels of investment in some of the communities rather than implementing projects itself, the social poorest areas refute the idea that such areas are systematically fund rapidly absorbed a large share of public investment. incapable of accessing resources from demand-driven programs. The concept spread quickly, as other countries sought to ease But in most cases the overall distribution of resources at the the social impact of economic crises and increase investment in household level was only mildly progressive. Positive discrimina- underserved areas. Now established in most countries in Latin tion toward poor households was best reached by latrine and America, social funds have spread to Africa, the Middle East, health projects, with sewage projects benefiting the better off. Eastern Europe, and Asia. By May 2001 the World Bank had For health the research found that social fund health interven- invested about $3.5 billion in social funds through more than 98 tions had a positive impact on infrastructure quality and the avail- investment operations in 58 countries. These funds had also ability of medical equipment and furniture. Essential drugs and attracted more than $4.5 billion from other international agencies replaceable medical supplies were generally more available in as well as from domestic financing from governments.Despite the social fund facilities, though all facilities had difficulties securing sizable investments, however, the funds remain a small part of adequate supply of essential drugs. Social fund facilities were also poverty and social protection activities in most countries, with as well or better staffed than comparators. In Bolivia, perhaps total expenditures typically equal to less than 1 percent of GDP. because they went beyond simply improving infrastructure, inter- Social funds have the following common characteristics: ventions in health clinics raised utilization rates and were associ- ated with substantial declines in under-five mortality rates.20 · Second-tier agencies. Social funds do not execute projects The results in under-five mortality rates in Bolivia are robust, directly. Instead, they appraise, finance, and supervise based on three alternative methodologies: propensity score investments carried out by other agencies, such as local matching;changes in mortality using household surveys between representatives of line ministries, local governments, and treatment and comparison groups at two different dates (1993 nongovernmental organizations. and 1997), also known as double differences or difference in dif- · Multisectoral choice of investments. Social funds typically offer ferences methodology; and econometric analysis estimation implementing agencies a wide range of choices of invest- (Cox proportional hazard function). The changes in mortality ments to be financed in different sectors. using double differences show a statistically significant decline in the percentage of children dying, from 10.3 percent in 1993 to 6 · Demand-driven investments. Social funds rely on project pro- percent in 1997 in the treatment group relative to an increase posals submitted by a variety of local actors, including local from 10.3 percent to 10.7 percent in the comparison group.The governments, nongovernmental organizations, community econometric estimates show a statistically significant decline in groups, and others. under-five mortality, from 88.5 to 65.6 deaths per 1,000 among · Operational autonomy and modern management practices. Social children living in the service area of a health center that received funds reside in the public sector but operate like private social fund investment. firms.They are usually granted exceptional status, either as Source: Reference 21. autonomous institutions or with operational autonomy under existing ministries. 62| The Millennium Development Goals for Health Box 4.2 Marginal budgeting for bottlenecks Despite extensive sector reforms, the health systems in many in physical accessibility, human resources, supplies and logistics, developing countries still fail to reach large numbers of women demand and use, and technical and organizational quality. and children--particularly the poor and most vulnerable of these Proxy indicators are used so that each intervention package can populations--with interventions that could significantly reduce be adequately represented by a proxy intervention, which in turn morbidity and mortality. Marginal budgeting for bottlenecks, has a proxy indicator for each corresponding constraint (or bottle- recently developed by the United Nations Children's Fund, the neck). Mali used the proportion of deliveries attended by trained World Bank, and the World Health Organization, has been tested staff as a proxy for use of institutional professional deliveries (ideally in several countries. The tool helps policymakers and program measured by proportion of deliveries attended by a professional in a managers improve the delivery of health services and interven- health center or hospital).The average of the values of the proxy tions. It can help formulate medium-term national or provincial indicators for a bottleneck--across an intervention package--rep- expenditure plans as well as Poverty Reduction Strategy Credits resents the extent of the bottleneck at a particular level of service that explicitly link expenditure to health Millennium Develop- delivery.Based on an analysis of the realistic possibility that the bot- ment Goals and optimally allocate newly available resources to tleneck can be reduced in a certain time period, marginal budget- achieve such health outcome targets. It also allows the likely ing for bottlenecks then sets a new target for the proxy indicator impact of alternative options on health outcomes to be assessed to corresponding to the bottleneck--the"new performance frontier." improve the allocative efficiency of government health budgets. This distinguishes it from traditional approaches of programming Module 2: Costing and budgeting and budgeting of health interventions. The costing and budgeting module estimates the volume of addi- Marginal budgeting for bottlenecks helps answer the follow- tional resources (the incremental cost) required to overcome the ing questions: bottlenecks and achieving the new performance frontiers set in · What are the major health systems bottlenecks--"the Module 1. It does so by costing a set of strategies or programs to weakest links in the chain"--that hamper the delivery of address the bottlenecks in each intervention package.The costing health services, and what can be done to address them? module has six components,five corresponding to the categories of bottlenecks defined above.The sixth component adds the expected · How much money is needed to achieve the expected results? costs of needed stewardship to steer the increase in coverage levels.It · How much can health outcomes be improved by removing includes costs for capacity strengthening, central supervision, the bottlenecks? research,and assistance with planning.The module can be applied to different regions (urban or rural,poor or rich,low or high morbidity Marginal budgeting for bottlenecks consists of three modules: and mortality). Such estimates can be used to assess resource alloca- bottleneck identification, costing and budgeting, and expected tion among regions.Once the marginal costs of addressing the bot- impact. tlenecks within a certain time period have been identified,they can be spread over the time period using different budget formats. Module 1: Bottleneck identification The first module defines intervention packages according to ser- Module 3: Expected impact vice delivery mode: family and community-based care, popula- The expected impact module uses an epidemiometric model to tion-oriented services, and individual-oriented clinical care. calculate the potential impact on the Millennium Development Current levels of coverage of these packages are estimated based Goals if the new performance frontiers for coverage levels of on available country data.The module then uses proxy indicators intervention packages are achieved. Providing policymakers with from data sources such as Demographic and Health Surveys or an idea of the consequences of their choices, it can serve as a Multiple Indicator Cluster Surveys to identify bottlenecks that basis for policy dialogue. hamper expansion of coverage levels.The identification of bot- tlenecks is performed in the following five broad categories: gaps Source: Reference 22. In Mali key bottlenecks to supporting home-based supplementation,and antenatal care);shortages of qualified practices and delivering both periodic and continual pro- nurses and midwives; and the lack of effective third-party fessional care were identified.These included poor access payment mechanisms for the poor for continual profes- to affordable commodities; the need for community-based sional care. Removing these impediments at a cost of $12 support for home-based care; poor geographical access to per capita between 2002 and 2007 could reduce under- preventive professional care (immunization, vitamin A five mortality by 20­40 percent and maternal mortality by Extra Government Health Spending Is Necessary but Not Sufficient |63 Table 4.3 Marginal budgeting for bottlenecks--how targeting can raise government health spending elasticities Current Change in Implied elasticity Under-five government government of under-five Baseline mortality spending Extra spending spending mortality rate to under-five reduction (U.S. dollars (U.S. dollars per capita government Country or state mortality rate (percent) per capita) per capita) (percent) health spending Benin 158 59 12.00 7.32 61 ­0.967 Ethiopia 185 31 4.00 3.77 94 ­0.329 Madagascar 142 42 5.00 3.15 63 ­0.667 Mali 231 57 10.00 8.13 81 ­0.701 Gujarat (India) 85 16 4.69 1.53 33 ­0.490 Orissa (India) 104 29 3.15 1.80 57 ­0.508 Rajasthan (India) 115 29 4.61 1.77 38 ­0.755 Source: Reference 31. 40­80 percent, depending on the poverty level of the Households are pivotal but often overlooked region. Cost and mortality estimates from studies such as this Households play two roles in the health system (chapter can be used to compute an implied elasticity of govern- 5).They are demanders of community-based and facility- ment health spending--the percentage change in mortal- based interventions, and they are deliverers of home-based ity divided by the percentage change in spending. The interventions. For interventions that are--or ought to implied elasticities (table 4.3) from several recent and be--delivered by a professional (antenatal care, the safe ongoing exercises in marginal budgeting for bottlenecks delivery of a newborn, immunization), the patient (or the are considerably larger than those in table 4.1, particularly caregiver for a child) is crucial. He or she makes the initial because the poor quality of policies and institutions in contact and plays a key part in what follows, in compli- these countries would mean elasticities in table 4.1 in the ance, follow-up, referral, and so on. If a mother fails to rec- low and possibly insignificant range. The results suggest ognize the signs of a sick child and does not take the child that in Ethiopia, spending an additional $7.32 per capita to a provider,the consequences can be fatal.A recent study on removing the bottlenecks identified in the marginal in Bolivia found that 60 percent of children who died budgeting for bottlenecks exercise would yield a return in during the period covered by the study had never been lower child mortality considerably larger than what would taken to a formal health care provider in the sickness be achieved if the extra money were simply added to the episode culminating in their deaths.23 government's health budget. Only time--and careful The household's role as a deliverer of care is also cru- monitoring and evaluation--will tell whether these esti- cial. It is mothers who breastfeed their children, and it is mates turn out to be accurate. they and other household members who purchase, treat, and use bednets. It is in the household that timely and appropriate complementary feeding gets provided to the Improving policies and institutions-- growing child. It is the caregiver who administers oral and the rest of the report rehydration therapy to a child with diarrhea and anti- Without better policies and institutions, countries with low malarials to a child with fever. Much of a child's health CPIA scores will not progress faster toward the Millennium is--or at least can be--"produced" in the household. Development Goals through extra government health spend- Policymakers can influence households through the ing. What do better policies and institutions entail in the resources households have at their disposal, including their health sector? Health systems are very broad--far broader knowledge of health matters, and through their influence than many people think.Weak policies and institutions can on community factors, by affecting both social norms and arise at several different points along the path from govern- infrastructure.All too often, however, health policy toward ment health spending to health outcomes (figure 4.3).Part II households is piecemeal and half-hearted, even failing to of the report uses this framework to tackle the difficult ques- acknowledge that households are deliverers of health tion of how to build stronger policies and institutions. interventions. 64 | The Millennium Development Goals for Health Figure 4.3 Paths to better health, nutrition, and population outcomes Millennium Health Sector and Government Development Households/Communities Related Sectors Policies and Actions Goals Health providers ·Private: for-profit, Health Household nonprofit actions and risk ·Public: hospitals, ·Underweight factors Household assets primary care, ·Under-five informal mortality rate ·Health sector policies ·Human ·Maternal ·Use of health at macro, health system mortality ratio services ·Physical and micro levels ·Communicable ·Dietary, sanitary, ·Financial ·Government spending diseases and sexual Finance and inputs practices ·Lifestyle ·Public and private insurance; financing and coverage ·Drugs Other MDG outcomes · Poverty Other policies Community Suppliers in factors related sectors ·Infrastructure ·Cultural norms ·Food ·Transport ·Community ·Energy ·Energy institutions ·Roads ·Agriculture ·Social capital ·Water and ·Water and ·Environment sanitation sanitation Source: Reference 29. Improving provider accountability Strengthening input markets Health providers in the public and private sectors play a key Provider organizations do not operate in a vacuum--they role in delivering interventions of relevance to the Millen- rely on human resources and drugs to deliver services nium Development Goals.Yet they are often highly ineffi- (chapter 7). In both "input markets," problems arise. cient, deliver poor-quality care, and are unresponsive to Human resource stocks are often inadequate,and drugs are patients, reflecting in part poor management and a lack of often unavailable. But there are other issues, too. Providers accountability to the public (chapter 6). Management often often are not as competent as they need to be. In Burundi lacks an outcome focus, a clear delineation of responsibilities in 1992 only 2 percent of children with diarrhea taken to and accountabilities within provider organizations, and a health facilities were correctly diagnosed.24 In the same proper link between performance, compensation, and pro- facilities only 13 percent of children correctly diagnosed as motion prospects.Citizens,in their capacities as patients,often having diarrheal disease were correctly rehydrated.24 Even exert less influence over providers than they could. In their if patients are initiated in correct treatment, there is no capacities as voters, they often exert too little influence over assurance that it will be administered successfully: in policymakers, who in turn exert too limited influence over India's public sector in the early 1990s less than 45 percent providers.The short route of accountability--from client to of patients diagnosed with tuberculosis were successfully provider--can be strengthened through such schemes as treated.25 vouchers, report cards, and citizen management groups.The Mistakes and poor quality care are not inevitable. One long route--from citizen to policymaker to provider organi- strategy for preventing such problems is training (chapter zation to frontline provider--can be strengthened through 7). Another is improving program design--one of the such schemes as contracting and public-private partnerships. issues of core public health functions (chapter 8). Both Extra Government Health Spending Is Necessary but Not Sufficient |65 have proved effective. In Tanzania providers trained in the 8. Wagstaff,A. 2003."Child Health on a Dollar a Day: Some Integrated Management of Childhood Illness were twice Tentative Cross-Country Comparisons." Social Science Medicine 57 (9): 1529­1538. as likely as other providers to prescribe antibiotics appro- 9. Rajkumar,A., andV. SwaroopV. 2002."Public Spending and priately; in Bolivia they were 10 times more likely to rec- Outcomes: Does Governance Matter?" Policy Research ognize the danger signs of a sick child.26,27 In India in Working Paper 2840,World Bank,Washington, DC. 2002 the tuberculosis treatment success rate had risen to 10. Feyzioglu,T.N.,V. Swaroop, and M. Zhu. 1996."Foreign Aid's 84 percent where DOTS was implemented.25 In Thailand Impact on Public Spending." Policy ResearchWorking Paper 1610,World Bank,Washington DC. the HIV infection rate among 21-year-old military con- 11. Burnside, C., and D. Dollar. 2000."Aid, Growth, the Incentive scripts dropped from 4 percent in 1993 to less than 1.5 Regime and Poverty Reduction." In TheWorld Bank: Structure percent in 1997 through active, effective, and open pre- and Policies, eds. C.L. Gilbert and D.Vines, 210­227. Cambridge: vention efforts launched in 1991.28 Cambridge University Press. For drugs it is not just that providers lack them in their 12. World Bank. 1999. Peru: Improving Health Care for the Poor. Washington, DC:World Bank. facilities--often they prescribe inappropriate ones, often 13. Ensor,T.,A. Hossain, Q.Ali, S. Begum, and A. Moral. 2001. knowingly. Even if drugs are available and prescribed "Geographic Resource Allocation in Bangladesh." Research appropriately, poor households and poor countries may Paper 21, Ministry of Health and FamilyWelfare, Health not be able to afford them (chapter 7). Economics Unit, Dhaka. 14. Diderichsen, F., E.Varde, and M.Whitehead. 1997."Resource Strengthening core public health functions Allocation to Health Authorities:The Quest for an Equitable Formula in Britain and Sweden." British Medical Journal 315 Strengthening core public health functions--monitoring, (7112): 875­878. evaluation, surveillance, regulation, social mobilization, and 15. World Bank. 2002. Armenia Public Expenditure Review.World intersectoral action--are also vital elements of a strategy for Bank,Washington, DC. strengthening the health sector (chapter 8). Governments 16. Lambrechts,T., J. Bryce, andV. Orinda. 1999."Integrated Management of Childhood Illness:A Summary of First have the responsibility for achieving and measuring progress Experiences." Bulletin of theWorld Health Organization 77 (7): toward the health goals.They have to formulate the strate- 582­594. gies and norms for disease prevention and control; establish 17. Santos, I., C.G.Victora, J. Martines, H. Goncalves, D.P. Gigante, the mechanisms for monitoring and evaluation; build insti- N.J.Valle, and G. Pelto. 2001."Nutrition Counseling Increases tutions and the capacity to monitor, educate, and mobilize Weight Gain among Brazilian Children." Journal of Nutrition 131 (11): 2866­2873. communities; and regulate and steer other sectors toward 18. World Bank. 2003. Attaining the Millennium Development Goals in the health goals.The recent SARS epidemic drew attention India: How Likely andWhatWill ItTake? Washington, DC: to well-performing public health functions in preventing, Washington, DC. controlling,and responding to new health challenges. 19. Jorgensen, S., and J.V. Domelen, 2001."Helping the Poor Manage Risk Better:The Role of Social Funds." In Shielding the Poor: Social Protection in the DevelopingWorld, ed. N. Lustig. References Washington, DC: Brookings Institution. 20. Newman, J., M. Pradhan, L. Rawlings, G. Ridder, R. Coa, and 1. World Health Organization. 2001. Macroeconomics and Health: J.L. Evia. 2002."An Impact Evaluation of Education, Health, Investing in Health for Economic Development. Report of the andWater Supply Investments by the Bolivian Social Commission on Macroeconomics and Health. Geneva:WHO. Investment Fund." World Bank Economic Review 16 (2): 241­274. 2. Filmer, D., and L. Pritchett. 1999."The Impact of Public 21. Rawlings L, L. Sherberburne-Benz, and J. van Domelen. 2004. Spending on Health: Does Money Matter?" Social Science and "Evaluating Social Fund Performance:A Cross-Country Medicine 49 (10): 1309­1323. Analysis of Community Investments."World Bank, 3. Filmer, D., J. Hammer, and L. Pritchett. 2000."Weak Links in Washington DC. the Chain:A Diagnosis of Health Policy in Poor Countries." 22. Soucat,A.,W.Van Lerberghe, F. Diop, S. Nguyen, and R. World Bank Research Observer 15 (2): 199­224. Knippenberg. 2002."Marginal Budgeting for Bottlenecks:A 4. Filmer, D., J. Hammer, and L. Pritchett. 2002."Weak Links in New Costing and Resource-Allocation Practice to Buy Health the Chain II:A Prescription for Health Policy in Poor Results."World Bank,Washington, DC. Countries." World Bank Research Observer 17 (1): 47­66. 23. Aguilar,A.M,. R.Alvarado, D. Cordero, P. Kelly,A. Zamora, and 5. World Bank. 2004. World Development Report 2004: Making R. Salgado. 1998. Mortality Survey in Bolivia:The Final Report. ServicesWork for Poor People.Washington, DC:World Bank. Basic Support for Institutionalizing Child Survival (BASICS) 6. Bidani, B., and M. Ravallion. 1997."Decomposing Social Project,Arlington,VA. Indicators Using Distributional Data." Journal of Econometrics 24. World Health Organization. 1998. CHD 1996­97 Report. 77 (1): 125­139. Geneva:WHO. 7. Gupta, S., M.Verhoeven, and E.R.Tiongson. 2003."Public 25. Ministry of Health and SocialWelfare. 2002. RNTCP Spending on Health Care and the Poor." Health Economics 12 Performance Report, India, 3rd Quarter. 2002. CentralTuberculosis (8): 685­696. Division, Delhi. 66 | The Millennium Development Goals for Health 26. Schellenberg, J. 2001. The MCE-TanzaniaWorking Group on the Strategy Framework. Report of aWorld Bank Consultation, 2001 Health Facility Survey. Health Facility Survey Submitted to November 28­29.World Bank, Health, Nutrition and the Department of Child and Adolescent Health and Population Department,Washington, DC. Development. Geneva,World Health Organization. 30. World Bank. 2003. World Development Indicators 2003. 27. MOH BW BASICS/USAID and Sociedad Boliviana De Washington, DC:World Bank. Pediatria. 1999. Report of the Health Facility Survey in Bolivia. 31. UNICEF andWorld Bank. 2003. Marginal Budgeting for 28. World Health Organization. 2000. Health a Key to Prosperity: Bottlenecks: How to Reach the Impact Frontier of Health and Success Stories in Developing Countries. Geneva:WHO. Nutrition Services and Accelerate Progress towards the MDGS:A 29. Claeson, M., and E. Bos. 2002. Health, Nutrition and Population Budgeting Model and Application to Low-Income Countries. New Development Goals: Measuring Progress Using the Poverty Reduction York: UNICEF;Washington, DC:World Bank. Extra Government Health Spending Is Necessary but Not Sufficient |67 CHAPTER 5 Households--Key but Underrated Actors in the Health Sector When physician Truong Cong Thang started practicing at the have come for examinations, and all the commune's women De Ar Commune health clinic in the Central Highlands, he have learned about family planning. had to take drastic action to win locals' trust.A belief exists in VIETNAM NEWS, October 19, 2003 the commune that when childbirth begins a woman must go to the forest by herself, and unaccompanied, deliver her baby, Households matter in the health sector--more than bringing the child home when it is over. A woman following most policymakers acknowledge. Improving the health this custom gave birth to a baby girl in the forest. But several of households is what the health sector is all about. days later, she developed a high fever as parts of her placenta People rely on their health in their everyday lives, and were still in her womb. Her desperate parents had asked a tra- for poor households, health is one of their major assets. ditional healer to pray over her and chase away the forest Households are also key actors in the "production" of ghosts, to no avail. health. Indeed, they play a dual role--as users of health When Thang heard about the woman's predicament, he services delivered by professional providers and as pro- issued a bold challenge to her family. If he couldn't save her ducers of health through the delivery of home-based life, he would submit to their punishment, as it is forbidden interventions and in their everyday health behaviors. under local laws for a strange man to see the body of a married In these roles, households are influenced by several woman. Only when commune leaders witnessed and sealed factors. The economic resources at their disposal. The Thang's letter of guarantee did her family let him treat the woman's control over these resources. The knowledge dying woman.Thirty-year-oldThang's gambit paid off after he household members have of health matters.The numer- saved her life.After this incident, villagers believed their physi- acy and literacy skills that enable them to acquire new cian was much better than the traditional healers. knowledge.The accessibility of health facilities.The water, When they started working in the commune, the young sanitation,and electricity in and around the house.And so doctor and nurse Sep went from house to house trying to per- on. This chapter shows how communities shape these suade pregnant women to visit the clinic for check-ups but met influences, and how policymakers can modify them.Ten a lot of resistance.However,after word spread ofThang's success case studies--of child health, maternal mortality, and in treating the woman who nearly died, gradually more women communicable diseases--illustrate the policy options. 69 Ten case studies on child health, maternal mortality, and The dual role of households in health communicable diseases illustrate the policy options (tables Households play a dual role in health.* They are users of 5.1,5.2,and 5.3);these are located at the end of this chapter. health services delivered by professional providers, and they are producers of health through the delivery of home- Households are not islands--health based interventions and in their everyday health behaviors. providers and communities help Health is one of the engines of economic growth and shape health outcomes poverty reduction. It is also one of the fruits of economic Health providers and communities influence the health- growth--possibly one of the most highly valued. People related behaviors of households. They, in turn, are influ- rely on their health in their everyday lives.As an Egyptian enced--often unwittingly--by policymakers. Health woman put it in Voices of the Poor, "We face a calamity providers influence households' decisions about use of ser- when my husband gets ill. Our life comes to a halt until he recovers and goes back to work."1 And households are vices--providers who deliver care that is perceived as good quality encourage return visits and build faith in the health where the activities of the health sector are ultimately system. Providers also influence decisions about home- directed. based interventions, providing knowledge about why If, as in the story fromVietnam, pregnant women do not nutrition matters and how to prepare nutritious meals on a seek assistance during delivery or do not seek antenatal very limited budget. care during pregnancy, health providers have their hands Through social norms and values, informal networks, and tied: they cannot deliver care without a client.The same is local governance structures, communities influence both true when caregivers do not seek assistance when their household practices and the delivery of services (see chapter child falls sick, when patients fail to show up for follow-up 6 on the delivery of services). In some settings the bound- visits or do not adhere to the provider's instructions.Weak aries between households and communities are vague, espe- demand for professional services would appear to be a cially where traditional household structures have broken major obstacle to achieving the Millennium Development down,due to civil strife,HIV/AIDS,or other circumstances. Goals. A recent study in Bolivia found that 60 percent of AIDS orphans are a case in point.Throughout many coun- children who died had not been taken to a formal provider in the sickness episode culminating in their death.2 And yet tries in Africa the community acts as a safety net that keeps families and households from destitution, providing material health planners rarely concern themselves with the factors relief,labor,and emotional support that would otherwise not that hold down demand.The assumption is that if patients be available.4 need health care they will demand it and the professional Policymakers can also influence providers and communi- provider will then deliver it. ties.They can affect what providers do, wittingly or unwit- Households are also deliverers of care--especially care tingly,to influence household decisions in their dual roles.At directed at children. Mothers breastfeed their children. the community level, policymakers can encourage or dis- They and other household members purchase, treat, and courage informal networks.Social norms are not immutable: use insecticide-treated nets. Caregivers provide timely governments can help make communities strong, resilient, and appropriate complementary feeding to the growing and well informed if they choose to do so. Community- child, wash their hands before preparing food, safely dis- based health projects can help, improving household health pose of feces, administer oral rehydration therapy to chil- practices and the delivery of services by professionals. dren with diarrhea and antimalarials to children with Consider theWorld Bank's Comprehensive Rural Health fever. These are not incidental interventions--they are Project in India's Maharashtra State.This project improved highly effective.The fact is that much of a child's health the health of more than 250,000 people by helping villagers is--or could be--produced in and by the household organize to address health needs and by training local village through the delivery of interventions or improved house- health care workers. The village health workers in turn hold or family practices (box 5.1). The same is true of organized women's development associations, which initi- nutrition outcomes, maternal and reproductive health ated credit circles to fund cooperative business enterprises. outcomes, and the prevention of other communicable (The role of the community in service delivery is examined diseases. in chapter 6.) *The term "household" is used here to refer to whatever grouping of peo- Raising incomes ple share responsibility for health. It is not necessarily limited to biological relatives (stepparents can be caregivers) and can encompass the broad array Low income is a barrier to using most interventions--pre- of kinship and household patters around the world. ventive and curative. Holding constant other influences on 70 | The Millennium Development Goals for Health Box 5.1 Key family practices for the production of child health and nutrition Families play a crucial role in keeping their children healthy and Households also interact with health care providers,both in seeking well nourished. A recent review summarizing the potential care for their children once they fall sick and in following through impact of family practices on child mortality, morbidity, growth, on care and advice delivered at health facilities. Households can and development highlighted the key practices or behaviors that have a crucial impact on child health outcomes in several ways. have a significant impact on child health outcomes. Many of · Take children as scheduled to complete a full course of immuniza- them can be delivered by households. tions (BCG, diphtheria, tetanus, pertussis, oral polio vaccine, and · Breastfeed infants exclusively for six months. Breastfeeding is measles) before their first birthday. Immunizations can prevent associated with reduced child mortality and morbidity and an estimated 3 million child deaths each year. Measles vac- improved development. Evidence suggests that breastfed cination is particularly important. infants under two months of age are six times less likely to · Recognize when sick children need treatment outside the home and die of infectious diseases than infants not breastfed. And a seek care from appropriate providers. Studies examining factors protective effect against diarrhea and pneumonia has been contributing to child deaths have found that poor care- observed both in industrial and developing countries. seeking is implicated in 6­70 percent of deaths. A high · Starting at six months of age, feed children freshly prepared energy- number of deaths has also been attributed to delays in care- dense and nutrient-rich foods while continuing to breastfeed for two seeking.A median of 23 percent of all fatally ill children are years or longer. Breast milk continues to be a source of key never taken to a health facility. nutrients and to confer protection against infectious diseases · Continue to feed and offer more fluids, including breast milk, to chil- throughout the second year of life. Beginning at six months, dren when they are sick. Children require more food and fluids however, it is not sufficient to meet nutritional requirements during illness, but 16­65 percent of caregivers withhold alone. Observational studies indicate that improving feeding food, breastmilk, or fluids during illness. Randomized con- practices could save 800,000 lives a year. trolled trials have found that feeding nutritionally complete · Ensure that children receive adequate amounts of micronutrients. diets to children with diarrhea increases net energy and Improving the intake of vitamin A, iron, and zinc will have a nutrient absorption.Contrary to many beliefs,feeding locally substantial impact on mortality, morbidity, and development, available foods does not increase the duration of diarrhea. particularly among poor or micronutrient-deficient popula- · Give sick children appropriate home treatment for infection. tions.Food supplementation or fortification may be necessary. Uncomplicated diarrhea, malaria, and local infections can · Safely dispose of feces,including children's feces,and wash hands after be managed at home with efficacious treatments. Oral defecating. Hand-washing after defecating and before prepar- rehydration therapy can prevent death from watery diarrhea ing meals and feeding children can reduce diarrheal diseases in all but the most severe cases. Home treatment of malaria- in children under five by a median of 33 percent (range related fevers by training mothers and increasing access to 11­89 percent). And better access to sanitation is associated treatment can have a large impact. with a reduction in all causes of child mortality. · Follow health workers' advice about treatment, follow-up, and referral. · Protect children in malaria-endemic areas by ensuring that they sleep Not adhering to treatment and referral recommendations by under insecticide-treated nets. A meta-analysis of four African ran- health workers leads to incomplete treatment, therapy failure, domized controlled trials showed that treated bednets are asso- drug resistance,and the later misuse of the left-over medicines. ciated with a 17 percent reduction in child mortality compared with control populations with no nets or untreated nets. Source: Reference 3. demand, higher incomes have been found to increase the Options for raising incomes likelihood of a pregnant woman receiving antenatal care and skilled care during delivery.5­9 Higher incomes are Given the link between income and the use of key health also associated with a higher probability of a child being interventions, economic growth will be an important immunized, sleeping under a treated bednet, being given source of progress toward the Millennium Development oral rehydration therapy when sick with diarrhea, and Goals.There is no doubt that economic growth has been a being taken to a formal provider when it has fever.10­13 major factor underlying the long-term improvements in Income--or rather the ability to buy food with money health outcomes. And there can be little doubt that slow earned from selling commodities such as crafts--is a key economic growth has meant slow progress on health out- factor in determining whether households have enough comes.It has been estimated,for example,that half a million food to eat (box 5.2). child deaths would have been averted in Africa in 1990 Households--Key but Underrated Actors in the Health Sector | 71 Box 5.2 Why people can go hungry when food is plentiful Traditionally,famines have been attributed to a sudden decline in laborers--who saw no appreciable rise in the price of the things the availability of food. On the face of it this explanation seems they sold--with vastly worse "terms of trade."They were unable plausible--it seems natural to assume that food supply must have to buy enough food to survive. By contrast, peasants and share- declined sharply if people start to starve. The official Famine croppers (tenant farmers paying their rent with a part of their Inquiry Commission set up by the Indian government following crop) were hardly affected. What caused the rapid rise in the the Bengal famine of 1943, which killed 3 million people, con- price of rice in 1943? Part of the explanation lies in the fall of cluded just this--that there had been a sudden decline in the Burma to the Japanese, which brought Bengal close to the war supply of rice, the staple food of the Bengali. front, causing rapid growth in military and civil construction. Amartya Sen has shown why a focus on food availability can The moral of Sen's argument is not complicated,but it is often be misleading. In 1943 there was actually more food available in overlooked: food production and availability are important, but Bengal than there had been in 1941, when no one went hungry. the key issue from a hunger perspective is the capacity of different Sen argues that a household can be plunged into starvation if groups to afford whatever food is available. The use of public there is a sudden fall in the price of the commodities it sells rela- employment programs to increase the purchasing power of vul- tive to the price of food.What marked 1943 as special was a sud- nerable populations is a key part of a strategy against hunger. den rise in the price of rice. The increase left groups such as barbers, craftspeople, fishers, transport workers, and agricultural Source: References 14­16. alone if the continent's economic growth in the 1980s had can take advantage of social networks to access otherwise been 1.5 percentage points higher.17 But economic growth, inaccessible health services.22 while important, will not be fast enough to achieve the Millennium Development Goals. Although projected rates Increasing affordability-- of economic growth will leave all but one Bank region especially for the poor (Sub-Saharan Africa) on target for reducing the number of The demand for health care, like the demand for almost people living on $1 a day, not one Bank region will achieve everything else, depends on price. Out-of-pocket expendi- the required reduction in under-five mortality through eco- nomic growth alone.18 Income transfer schemes provide tures for professionally delivered health services can be sub- stantial. A normal hospital delivery in Dhaka, Bangladesh, policymakers with a means of raising the incomes of certain groups more quickly than they would through economic growth. Such schemes are rarely rationalized in terms of improving health outcomes. There is, however, clear evi- Box 5.3 Social solidarity in Côte d'Ivoire dence that they yield health payoffs. South Africa's old-age pension provides an example.The program was originally a Formal government mechanisms do not provide adequate safety net for whites who reached retirement age without access to health services for all people in many countries. In an adequate employment-based pension. By the end of Côte d'Ivoire the imposition of user fees for public services in 1993 the program had been extended to all racial groups, the 1990s effectively made services inaccessible to the poorer segment of the population.Yet many poor people use modern and it has been an important source of income for non- health care services that have become quite expensive. whites.Recent research finds that the pension improved not Solidarity among parents, friends, and members of community only the health of pension recipients but also the health of social networks allows them to do so. other members of households in which resources are Many factors at the household level are significantly associ- pooled.19,20 Among black children under the age of five,the ated with using financial solidarity to gain access to health ser- pension is estimated to have led to an eight-centimeter vices. The number of children born to the head of the increase in height--equivalent to half a year's growth.And household to which the ill person belongs was found to be the South Africa's pension program has been found to dispro- strongest. Households with no children were less likely to ben- portionately benefit poorer households and households efit from financial solidarity than households with children. with children.21 Gender is also a determining factor: women have a better Informal solidarity schemes at the community level chance of receiving financial support for illness than men. often substitute for formal social protection programs (box Source: Reference 22. 5.3).Although a lack of money is an obstacle,communities 72 | The Millennium Development Goals for Health costs the equivalent of one-quarter of the average monthly who do not receive the intervention. For example, an income.23 A single hospital visit for a Vietnamese in the individual can reduce the probability of contracting poorest fifth of the population in 1998 entailed fees and drug malaria through a variety of preventive interventions, costs equivalent to 22 percent of that person's annual discre- including the use of insecticide-treated nets. Once a per- tionary income (income after deducting food expenses).24 son is infected, the use of an antimalarial will eliminate the In many cases out-of-pocket payments are informal malaria in the blood. By reducing the reservoir of infected rather than formal. In Armenia in 1999, 91 percent of people, the use of both preventive and treatment interven- health service users made an informal payment to a health tions reduces the probability of mosquitoes becoming care provider.25 In Azerbaijan and Poland the figure was infected, which in turn reduces the probability of other 78 percent. In nearly all developing countries tuberculosis people being bitten by an infected mosquito. treatment is officially provided free of charge, but indirect Immunization is another intervention that generates a costs and informal illegal payments mean that the real positive externality. If person A gets vaccinated against a prices of care are much higher, and can diminish early and particular disease, persons B and C benefit through the full help-seeking, and increase transmission of disease. lower risk of transmission from person A.The benefits of It is not just the fees paid to health providers that mat- vaccination to person A spill over to other people. Subsi- ter--households incur other money costs in using health dizing the use of such interventions makes economic sense. facilities, including the cost of transportation. A study of In the absence of subsidies the use of immunization and referral and follow-up recommendations based on the preventive and treatment interventions against malaria Integrated Management of Childhood Illness strategy in would be at inefficiently low levels. Sudan found that only about half the children judged in The same case can be made for the interventions listed need of urgent referral were taken for such care within 24 in figure 3.2 in chapter 3 for all but two of the other lead- hours.The cost of getting to the hospital was the reason ing causes of childhood deaths (the exceptions are birth most frequently cited for not getting the needed care.26 asphyxia and preterm delivery).And it can be made for the Voices of the Poor identifies other money costs.27 A young interventions for tuberculosis, malaria, and HIV/AIDS woman in Muynak, Uzbekistan, said: "We do not go to and for several interventions to reduce maternal mortality. hospital because it is necessary to bring your own bed Even where there are no spillovers, an equity case can linen, dishes, sometimes even a bed." be made for reducing prices.The equity concern is likely Higher money prices tend to reduce demand--especially to relate to poorer groups, who cannot afford care at the among the poor--unless accompanied by improvements in regular price or would risk being plunged into--or fur- service quality.28 Cost is one of the main reasons sick people ther into--poverty if they paid for it. Price reductions give for not seeking care. In Georgia in 1997, 94 percent of should, in principle, target such groups. those who did not seek care when ill said it was the high Reductions in the price of care at the point of use may cost that prevented them doing so.25 In China,when tuber- also come about as part of an expansion in insurance cover- culosis treatment was made free,demand for it increased sig- age. Risk aversion coupled with the unpredictability of ill- nificantly. Informal payments deter as much as formal ness provides a motivation for pooling risks through an payments. Informal payments were identified as one of the insurance scheme. Revenues are collected from those at factors underlying the low use of community hospitals and risk--or at least from a large number of them--and then health centers inThailand in the 1970s.29 used to pay the costs incurred by people if and when they The demand for preventive home-delivered interven- fall ill.To counter moral hazard--the tendency for insured tions also depends on price. High food prices, for example, people to increase use beyond an efficient level when they tend to be associated with low nutrient intake and low lev- get sick--insurance programs often require a copayment els of child survival and nutrition.30­33The nutritional status by the insured at the time services are used. Such insurance and nutrient intake of the poor are particularly sensitive to schemes range from private insurance to community-based changes in the price of food.31,33 In a recent study, afford- schemes to social security schemes and tax-financed ability was the most frequently cited reason for not owning programs. a mosquito net,especially among the poorest group.34 From blanket to targeted subsidies Reasons to reduce the price of health care General revenues are often used to finance all government Many interventions relevant to the health Millennium health services for everyone.These blanket subsidy schemes Development Goals are activities that generate externali- suffer from two drawbacks: they are not targeted on inter- ties--the benefits of the intervention"spill over"to people ventions that produce externalities, and for the most part Households--Key but Underrated Actors in the Health Sector |73 they are not well targeted on the poor.Although the poor attractive approach. Subsidies for specific interventions can appear to benefit disproportionately in cost terms from be implemented through differential charging (setting a subsidies to the health sector in some countries and states zero fee for, say, immunizations in government clinics) or in the developing world (Costa Rica, Malaysia, Kerala), the through a system of vouchers that allows the holder to better off generally benefit most (box 5.4,figure 5.1).35­42 obtain the intervention free of charge from any approved Focusing subsidies on specific interventions (immuniza- facility. tions) or on certain groups (the poor) is in principle a more As part of a deliberate policy to improve maternal and child health, especially among the poor, the government of Malaysia introduced free antenatal care and home deliv- eries by government midwives.43 Bolivia's Maternal and Box 5.4 Blanket subsidies often benefit the Child Health Insurance program,introduced in 1996,cov- better off most ers the cost of antenatal care, care received during labor Benefit-incidence analysis provides a means of seeing how dif- and delivery, postpartum and newborn care, and manage- ferent income groups benefit from public subsidies to a partic- ment of all obstetric and newborn complications.44 The ular sector. For each income group the analysis examines the program appears to have increased the use of antenatal use of different types of facilities and the amount of subsidy per care and skilled birth attendants, especially among the visit (or inpatient day). The subsidy may vary across income poorest fifth of the population, which saw nearly a dou- groups if, say, the better off are charged higher fees. bling in pregnant women receiving two or more antenatal In most countries the better off benefit most from govern- visits (figure 5.2). In Ifakara,Tanzania, vouchers were used ment subsidies to the health sector (see figure 5.1). (Colombia, to provide a 17 percent subsidy on the cost of a mosquito Costa Rica, Malaysia, and Sri Lanka are exceptions.) In Guinea net for pregnant women and children under five. In com- the richest 20 percent of the population receives nearly 50 per- bination with social marketing to increase the knowledge cent of the government's subsidy.The failure of blanket subsidies and availability of bednets, these subsidies led to increased to benefit the poor disproportionately stems from the upper use of bednets.45­47 One attraction of an explicit voucher income groups' above-average use of government-subsidized services, especially services provided by hospitals, where much scheme is that it ensures that the provider gets paid for of the typical government subsidy is spent.This skewed pattern treating the patient in question (box 5.5). A zero fee may of use could in principle be offset by linking user fees to simply leave providers out of pocket, creating incentives income. Although this happens in some countries, the results for them to deliver unsubsidized interventions. make it clear that if it does happen, the user fee gradient is not The use of entitlement cards provides a means of focus- steep enough to offset the higher use rates among the better off. ing government subsidies on specific groups.A health card Source: World Bank staff. scheme introduced in Indonesia during the crisis of the Figure 5.2 Under Bolivia's National Maternal and Figure 5.1 Who gets subsidies? Child Insurance Program key maternal health 50 interventions rose fastest among the poor Poorest 20% 45 Richest 20% 100 40 Percentage with one or 90 more antenatal visits to 35 medically trained person 98­ 80 subsidies 30 Percentage with two or of 70 more antenatal visits 25 1994 Percentage of births 60 20 attended by medically trained person 15 50 increase Percentage 10 40 5 30 0 20 Percentage 10 1992 1995 1998 1992 1994 1987 (rural) 1993 1998 1994 1989 1993 1992/93 Rica 0 d'IvoireEcuadorGhanaGuinea 1992 Africa First Second Third Fourth Fifth UruguayVietnam CostaCôte Indonesia Nicaragua Kenya Madagascar South Tanzania Wealth quintile Source: Reference 95. Source: Reference 96. 74 | The Millennium Development Goals for Health Some recent programs--especially in Latin America-- Box 5.5 Vouchers for sex workers in have pushed these ideas even further. Rather than simply Nicaragua reducing the cost of using specific interventions, these pro- grams provide cash payments to users. Receipt of cash pay- In Managua, Nicaragua, a voucher scheme has been used to ments is linked to the use of specific interventions, and increase the uptake of reproductive health services among participation is restricted to certain groups--often poor female sex workers.50 Every three to five months, fieldworkers mothers and their children.The experience with these pro- and nongovernmental organizations (NGOs) at prostitution grams, in targeting and impact, is encouraging (box 5.6). sites distribute about 1,200 vouchers, depending on the esti- mated number of sex workers operating in the city at any given Financial inducements--including negative copayments-- time.The vouchers entitle the sex workers to free services at have also been used to induce people to seek treatment of one of 8­10 private, NGO, and public clinics, contracted to the particular illnesses,such as tuberculosis. voucher agency by competitive tender. Approved providers must follow a set treatment protocol and receive training. Empowering women Contracts are reviewed after each round of voucher distribu- It is not just a household's total income that matters-- tion and renewed subject to an assessment of quality of care. The clinics return the vouchers to the voucher agency, which women's control over its use also makes a difference to reimburses the provider an agreed fee per voucher. health outcomes. In some countries the proportion of Sex workers were involved in the design of the program and women who can decide how to spend their own money is have opportunities to express their preferences and complaints. staggeringly small (figure 5.3). Poor women are likely to be In each round 10 percent of recipients are interviewed about particularly disadvantaged. Research shows that women their experience. Initially, sex workers reported that the gate- who exert relatively little control over household financial keepers to care (nurses and receptionists) lacked sensitivity:train- resources are, all else constant, less likely to receive antenatal ing and sensitization of this group helped improve their attitudes care,less likely to have antenatal visits,and less likely to have toward these clients.The technical quality of care (as assessed by visits in the first trimester of pregnancy.57 It seems likely that an examination at the outset of the project) was lower than part of the effect of the South African pension program on expected,so training and treatment protocols were introduced. child health stems from the fact that it is paid to women Although the prevalence of sexually transmitted infections is (grandmothers) rather than to men. Britain decided to only slightly lower than at the beginning of the project (possibly due to the high turnover of female sex workers), incidence make child benefit payments to women rather than to men among women who used vouchers more than once dropped by for similar reasons. 65 percent in the first three years of the program. Following a Microcredit programs aimed at poor women are thought recommendation by the sex workers, they now receive vouch- to be one way of increasing women's financial autonomy, ers to give to their regular partners and clients as well. Sex and they may have increased the use of maternal health ser- workers appreciated the fact that they could choose which vices.The evidence is somewhat mixed, but it seems likely clinic to attend and made their choices on the basis of distance that the impact varies from one scheme to the next.58 Box and staff friendliness.The clinics reported that their main bene- 5.7 provides an encouraging case of women's empower- fit was improvement in the technical quality of their services ment for better health. and that the lessons learned were applied to all of their clients. It is not just women's financial independence that is They felt their reputation was enhanced by being contracted by important.Their ability to take decisions more generally-- a prestigious public health agency (the Central American including those that do not have major financial conse- Health Institute). quences--also matters. For example, in India, although Source: World Bank staff. contraceptives are readily available in retail shops, commu- nity pressure or the disapproval of husbands often prevents women from using them. late 1990s was targeted toward the poor.48 In Egypt a school health insurance program was introduced covering Improving information and knowledge all school-attending children. Although coverage rates are higher among the better off (simply because better-off Affordability--the price paid relative to discretionary children are more likely to be enrolled in school), the pro- income--is undoubtedly one important barrier prevent- gram is likely to have increased coverage more among the ing the use of health services. But it is not the only one. poor than among the better off (many of whom were Knowledge--or a lack of it--is another. already covered under other schemes).49 In addition to A lack of knowledge can result in people not seeking being fairly well targeted, both the Indonesia and Egypt care when they need it, despite the absence of price barri- programs have had a considerable impact.48,49 ers. In Bolivia a large fraction of poor babies are not deliv- Households--Key but Underrated Actors in the Health Sector | 75 Box 5.6 Increasing coverage of key interventions through demand-side incentives A convincing body of evidence is beginning to accumulate on the of readily verifiable indicators. Despite these differences, PRAF, potential role of cash transfer programs in bringing about the PROGRESA, and Nicaragua's Social Protection Network behavior changes necessary to achieve the Millennium Develop- (RPS) pilot program each managed to ensure that more than 50 ment Goals. Used as a means of combating short-term poverty, percent of beneficiary households (more than two-thirds for cash transfer programs have been seen as vulnerable to political Honduras and Nicaragua) were from the poorest 30 percent of manipulation and ill suited for promoting longer-term develop- all households. ment. But a new generation of programs being implemented in These programs have in most cases invested heavily in impact Brazil, Colombia, Honduras, Jamaica, Mexico, and Nicaragua evaluation, providing important lessons about their strengths and makes receipt of the cash payments conditional on household weaknesses. Most of the programs have conducted thorough behaviors that foster the development of human capital, such as assessments of levels of development in their areas of influence using preventive health care services or keeping young children in before phasing in the program benefits, allowing investigators to school. track changes over time. These programs tend to be well targeted to the poor.51 The In Nicaragua vaccination rates (complete vaccination in one- mechanisms for achieving this propoor targeting vary greatly year-old children) increased by 18 percentage points.52 In Honduras from program to program.Honduras'Family Allowance Program opportune delivery of early infancy immunization increased by (PRAF-II) used a height census of all first-grade school children 7­10 percentage points,but later vaccinations,already at high cover- in the country to identify municipalities at risk of malnutrition. age levels, were not affected.53 In Brazil,54 Honduras,53 Mexico,55 Mexico's PROGRESA program (now Oportunidades) supple- and Nicaragua,52 health service use increased substantially. In mented geographic targeting with a household screening tool to Honduras the proportion of women receiving antenatal care five or estimate each family's standard of living from a limited number more times during their pregnancy increased by 18­20 percentage Operational characteristics of conditional cash transfer programs in Latin America and the Caribbean Payment method Condition for payment Enforcement Brazil, Bolsa Alimentação Magnetic debit card can be used to Six antenatal care visits during pregnancy; Compliance with preventive health care withdraw cash at an automatic teller monthly well-child checkups, including undertakings supposed to be reviewed machine or from a lottery ticket seller complete immunization coverage and by local health team after six months; no growth monitoring information on number of beneficiaries suspended Colombia, Familias en Acción Cash payment by bank Regular attendance of child at growth- Attendance monitored; no information monitoring sessions on number of beneficiaries suspended Honduras, Family Allowance Program (PRAF) Phase II Freely exchangeable voucher distributed Five antenatal care visits during pregnancy, Beneficiaries deposit bar-coded coupons through primary schools or by perinatal checkup within 10 days of at health center on attendance; no program directly delivery, monthly well-child checkups exclusions implemented as of mid-2003 Jamaica, PATH Cash payment at "pay agency" Regular preventive health care checkups for Intended; no information about current children under six years, pregnant and operation lactating women, elderly, disabled, and destitute adults Mexico, PROGRESA/Oportunidades Cash distributed directly Regular preventive health care checkups for Conditions monitored and apparently the entire family, attendance at health enforced education sessions Nicaragua, Social Protection Network (RPS) Cash distributed directly Growth promotion for young children, All conditions rigorously monitored and complete immunization coverage, no more enforced than two consecutive months of inadequate growth, attendance at health education sessions 76| The Millennium Development Goals for Health Box 5.6 Increasing coverage of key interventions through demand-side incentives (continued) points.In Nicaragua the poorest residents of the program area bene- eral, household-level incentives are probably best combined with fited from the intervention more than the less poor.Similar findings measures that improve the quality of services offered by schools were reported in Honduras.56 and clinics in poor communities.The Honduras program explic- Although conditional cash transfers are a powerful way of itly aims to evaluate the synergies between these two approaches, changing the behaviors of poor families, thought needs to be but the impact of the supply-side interventions will not be given to identifying the precise behavioral responses that need to known until the final evaluation round planned for 2004. be fostered in order to achieve life-transforming impacts. In gen- Source: Saul Morris. Figure 5.3 Poor women have less of a say in Box 5.7 Helping poor women protect spending their own money themselves--India's SEWA Percentage of women who can decide to spend their own money The Self-Employed Women's Association (SEWA), an organi- 0 20 40 60 80 100 zation of poor self-employed women in Gujarat, India, helps women attain full employment and makes them self-reliant.An Nicaragua 2001 Integrated Social Security Scheme provides life insurance, Haiti 2000 medical insurance, and asset insurance. The premium is the Colombia 2000 equivalent of $1.70, just under half of which is earmarked for medical insurance that covers hospital care up to a ceiling of Peru 2000 $28 a year. Several preexisting conditions, and diseases caused Mali 2001 by addiction, are excluded from coverage.There is no restric- Uganda 2000/01 tion on the choice of provider--members can use public, pri- vate, and nonprofit providers. Members make claims for Guatemala 1998 Population reimbursement after discharge, on presentation of specified Egypt, Arab Rep. 2000 Poorest 20% documentation. In 1999­2000, 23,214 women participated in India 1998/99 SEWA's medical insurance fund, in 10 districts of Gujarat. A recent study found that women making claims were Zimbabwe 1996 poorer than the general population, suggesting that the scheme Kazakhstan 1999 was including the poor.59 This may have been due to the embedding of the insurance plan within a scheme that targeted Zambia 2001 poor employed women and the fact that the premium was rela- Malawi 2000 tively low (0.4 percent of median income among claimants). Turkey 1998 Insurance significantly reduced the financial cost of hospitaliza- tion. On average, 76 percent of hospital costs were reimbursed. Nepal 2001 Without reimbursement, hospital costs would have been "cata- Source: World Bank tabulations and reference 97. strophic" (more than10 percent of annual household income) for 36 percent of claimants. It was estimated that 15 percent of claimants faced expenses that were catastrophic even after reim- bursement. The poorest groups were found more likely to be ered by a trained attendant even though the mothers are protected by the fund from catastrophic expenses.The fairly low eligible for free care under the Maternal and Child Health reimbursement ceiling limited the degree of coverage but also Insurance program.44 In India, where immunization is free, reduced "adverse selection" (the tendency of below-average 60 percent of children have not been fully immunized. risks to self-insure because premiums reflect average risks). Asked why they had not immunized their child,30 percent Source: World Bank staff. of mothers said it was because they were not aware of the benefits.Another 30 percent said they did not know where to get their child vaccinated.60 the poor--like the better off--report a large number of A lack of knowledge can also result in people seeking and short-duration illnesses,but unlike the rich they tend to visit receiving inappropriate care--and paying for it.The infor- a doctor immediately, usually a private one.61 They end up mation asymmetry between patient and provider gives spending a large proportion of their income on these short- providers scope to induce demand for their services.In India duration illnesses,often without getting the cause diagnosed. Households--Key but Underrated Actors in the Health Sector | 77 Furthermore, the medication they receive is often inappro- that piped water in India has a substantially greater impact priate or contraindicated. A poor child in Indonesia gets on the prevalence of diarrhea among the better-off and more than four (often useless) drugs per diarrhea attack-- better-educated.67 Likewise, malnutrition is often caused-- instead of oral rehydration therapy. Many studies find that at least in part--by the lack of knowledge about how and female education is an important determinant of coverage of how much to feed infants and young children.68 key Millennium Development Goal interventions.The use In many societies, inaccurate knowledge based on tradi- of almost all child health interventions is higher in house- tional beliefs or popular notions of modernization--often holds with better-educated mothers, typically the better off influenced by mass media advertising--can have a signifi- (figure 5.4).The list includes intake of complementary foods cant influence on health behavior. In The Gambia poor among infants, hand-washing, appropriate disposal of exc- breastfeeding practices, including prelacteal feeding and reta,the likelihood of receiving antenatal care,the likelihood failure to practice exclusive breastfeeding,were found to be of choosing formal rather than traditional care,the timing of due to traditional beliefs and perceptions--among women antenatal consultations, the likelihood of a baby being deliv- and men--that bottle feeding was part of modernization.69 ered away from home and by a trained person, the use of Some women do not use chloroquine tablets to prevent well-baby clinics,the likelihood of a child being immunized, malaria because they believe the pills induce abortion.70 the use of oral rehydration therapy, and the likelihood of a caregiver seeking care for a child with fever.5­11,13,62­66 Girls'--and boys'--education A lack of knowledge and skills may also result in people Increasing education--especially among girls--is likely to not getting the full health gain from inputs they have avail- yield health payoffs. But there are two caveats. First, the able to them and use. Many women, for example, do not payoffs will be felt on the health of the next generation of know that piped water often requires further purification children: efforts under way today to increase primary and that hand-washing confers much of the health benefit school completion rates among girls will not yield health of piped water. It is scarcely surprising in the light of this payoffs in lower rates of under-five mortality and child malnutrition for at least another 5­10 years. Second, the health benefits to increased primary education for girls--in Figure 5.4 Fewer poor women have completed better health outcomes for their children--may not be as the fifth grade large as previously thought.A recent study found that once Percentage of women age 15­49 confounding factors were taken into account, primary who have completed fifth grade education of the mother had a statistically significant 0 20 40 60 80 100 120 impact on child survival in only 3 of 23 countries stud- Kazakhstan 1999 ied.71 The effect appears to be especially muted in Sub- Zimbabwe 1996 Saharan Africa.71,72 Turkey 1998 Poorest 20% Population Namibia 1992 Another recent study73 sheds light on how mother's Peru 2000 education influences child health and suggests why the Tanzania 1999 effect appears to be small. It finds that it is not schooling, Colombia 2000 literacy, or numeracy that leads to better child health but Ghana 1998 the fact that better-educated women tend, on average, to Zambia 2001 have more health knowledge.The study also finds that girls India 1998/99 acquire this knowledge after leaving school, using the gen- Bolivia 1998 eral literacy and numeracy skills they acquired at school. Malawi 2000 Cameroon 1998 This points to the need to ensure high-quality education Uganda 2000/01 at the primary level. If the quality of primary education is Egypt, Arab Rep. 2000 low, girls will fail to acquire sufficient numeracy and liter- Nicaragua 2001 acy skills to enable them to develop the health knowledge Haiti 2000 later in life that will help their children remain healthy. Bangladesh 2000 Low school quality, coupled with a more hostile health Nepal 2001 environment, may be part of the reason for the small and Guatemala 1998 largely insignificant effects of primary schooling on child Mali 2001 survival in Africa.Where school quality is low, it may make Source: World Bank tabulations and reference 97. sense for girls and boys to be taught health education at the primary level. 78 | The Millennium Development Goals for Health Conveying health knowledge through the health sector Box 5.8 Working with the private sector to Conveying health knowledge through public facilities has a improve hygiene behaviors long history.A recent example is the lactation management clinic opened five years ago at the Children's Hospital in A recent review found that 12 hand-washing interventions in Islamabad.The program's aims were to promote exclusive nine countries achieved a median reduction in diarrhea inci- breastfeeding--rare in Pakistan--until four to six months dence of 35 percent.77 Many of the most successful interven- of age and to try to solve mothers' breastfeeding difficul- tions provided soap to mothers, explained the oral-fecal route for disease transmission, and asked mothers to wash their hands ties.Mothers receive one-on-one consultation with trained before preparing food and after defecating. health professionals on the benefits and proper techniques One promising approach to the delivery of these interven- of exclusive breastfeeding. So far, more than 4,000 mothers tions is social marketing, using private-public partnerships. Such have been counseled at the clinic. Preliminary results show an initiative was implemented in 1996­99 in Costa Rica, El that 67 percent of mothers had solved their breastfeeding Salvador, and Guatemala.78The objective was to improve hand- difficulties. More than 60 percent were exclusively breast- washing habits in order to reduce diarrheal disease among chil- feeding, and another 25 percent were breastfeeding and dren under five.The approach was to combine the expertise and giving other fluids. Only 9 percent of those followed were resources of the soap industry with the facilities and resources of no longer breastfeeding.74 governments to promote hand-washing with soap. Four private This commonly used approach can be effective for those soap companies launched hand-washing campaigns in collabo- who participate, but its impact is limited by the fact that ration with the public sector. The key features of the project many people--including many poor people--do not use were behavioral research to identify consumers' baseline hand- washing habits; mass media coverage (both education and com- the public sector when they fall ill.And in many cases they mercial); community activities through the public sector, do not seek care from a formal provider at all. Programs NGOs,and foundations;and monitoring and evaluation. based in the community--where the conveyers of knowl- The private-public partnership succeeded in improving edge actively seek out target groups--seem likely to have a hand-washing behaviors and reducing the incidence of diar- better chance of reaching a broader catchment group and rheal disease. The results included a 30 percent increase in of reaching the poor. hygienic hand-washing behavior in mothers, with an estimated There are several success stories here. Health workers in 300,000 fewer cases of diarrhea a year in poor children under Brazil trained in the Integrated Management of Childhood five in Guatemala. Including USAID, the BASICS project, soap Illness (IMCI) provided information and counseling at manufacturers, the Pan American Health Organization, and the health facilities and in the community.75 Health knowledge World Bank,the initiative leveraged resources and sustained the among mothers improved, as did feeding practices. After involvement of the private sector in social programs. only 18 months the nutritional status of children in the area Source: World Bank staff. improved. In India a counseling strategy was devised jointly by government, NGOs, community volunteers, and health workers in eight villages in rural Haryana.76 Compared with tions, training, policy change, and product development the control areas (where no community volunteers or health and marketing into an overall strategy to achieve the workers were trained),the intervention areas achieved longer desired sustainable behavior change.When the focus is on breastfeeding durations, a higher quality of complementary product marketing,social marketing organizations are often feeding provided to children after six months, and increased nonprofit firms or associations, but the products tend to be hand-washing before feeding children. distributed through various for-profit outlets and NGOs. Social marketing has been applied to such diverse inter- Killing two birds with one stone--social marketing ventions as family planning, the treatment of sexually Social marketing is proving a promising approach to the transmitted infections, the use of insecticide-treated mos- design and delivery of both health messages and products quito nets, the use of soap for hand-washing, water purifi- (box 5.8). It aims at promoting new or modified behav- cation, improved weaning and breastfeeding practices, and iors that are acceptable and feasible to most people. It increased consumption of foods rich in vitamin A. A actively involves key actors in designing and implement- recent evaluation of Tanzania's social marketing of mos- ing activities.68 quito nets--which included an element of subsidy as well Social marketing refers not only to the marketing of as commercial marketing--found that coverage increased socially valuable products (contraceptives, vitamins) but to substantially in all project areas and that socioeconomic the promotion of key behaviors. It integrates communica- differentials in net ownership narrowed.45,46 Households--Key but Underrated Actors in the Health Sector | 79 Media and behavior change · A young woman in Muynak, Uzbekistan, said:"We do not go to hospital because it is necessary to bring your The media often play a role in communication efforts aimed own bed linen, dishes, sometimes even a bed." at behavior change. Under government leadership, messages on AIDS were broadcast frequently during the peak of the · In Tanzania, people complain of being treated "worse epidemic in Thailand, a key element in the campaign to than dogs" by rude health staff and of being "yelled at, reduce the spread of HIV/AIDS.79 The well-designed and told they smell bad, and called lazy and good for culturally sensitive radio communication project in Nepal nothing." 81 appears to have promoted positive health outcomes (box · People complain about corrupt health staff, who 5.9).80 But many of the poor do not have access to a radio.In demand unofficial "fees" or expect small "gifts." Bolivia, India, Morocco, and Mozambique less than 30 per- cent of the poorest fifth of the population has access to the · Indigenous people in Ecuador complain of staff who media. mistreat patients because of their lack of knowledge of Spanish, and women everywhere are reluctant to seek care for maternal health problems from male staff. The influence of providers on demand (Making provider organizations more responsive to The way health services are organized (hours of operation, patients' legitimate concerns is discussed in chapter 6.) waiting time), their amenities and facilities, the disposition People's perceptions of the technical quality of the ser- and integrity of staff, and the cultural appropriateness of vice they receive also matters.They may, of course, not be services all make a difference to the household's demand the best judges of technical quality, but the fact is that per- for professionally delivered health care. Voices of the Poor is ceptions of quality influence demand. From a policy per- full of complaints by poor people about the lack of friend- spective the challenge is not just to improve the quality of liness of facilities and health service staff.27 care but to provide methods for patients to better judge · A woman in Los Juries,Argentina, remarked:"You go whether specific providers are delivering quality care. to the hospital, you have to get a number, you go to (These issues are also taken up in chapter 6.) the guard, the nurses are chatting. You have to wait until they fancy giving you a number . . .`Is the doctor Roads, transport, and accessibility here?'`No, the doctor isn't here,' they lie." Transportation systems, road infrastructure, and geography influence the demand for care delivered by formal providers--through their impact on time costs, which can be substantial.6,30­32,82 In rural Senegal only 42 percent of Box 5.9 Radio dramas to promote the population live within 30 minutes of a health facility, contraception and 43 percent live more than an hour away.83 In rural The Radio Communication Project uses the radio to satisfy communities, where the roads are poor and transportation the large unmet need for contraception in Nepal, strengthen unreliable, the time spent waiting for transportation may the quality of services and service delivery, and increase client be as great as--if not greater than--the time spent travel- demand for quality reproductive health services. It produced ing to the facility.Time costs tend to be a major issue for two complementary radio serial dramas, one for service maternal mortality--health centers cannot provide essen- providers and one for community members. Education tial obstetrical care for a complicated delivery, and women through radio had a positive impact on both groups. would have to travel to distant hospitals to get such care.In On the client side 81 percent of the women who listened Voices of the Poor aTogo resident observed:"If a woman has to the drama were using modern family planning, compared a difficult delivery, a traditional cloth is tied between two with 33 percent of the women who did not listen. Exposure to sticks and we carry her for seven kilometers to the health radio was also linked to greater spousal communication about center.You know how long it takes to walk like that?" 84 family planning: 87 percent of men who listened to the serial drama discussed family planning with their wives, compared Road and transportation improvements with the 64 percent of men who did not listen.The program support the health sector linked positive attitudes toward family planning, discussion of family planning with health workers, and perceived policy sup- Better access to health facilities is rarely the primary goal of port for family planning. road rehabilitation projects--or even an explicit objective Source: Reference 80. at all. But a recent road rehabilitation project in Vietnam cut the average walking time to the local hospital by 17 80 | The Millennium Development Goals for Health Box 5.10 Argentina's water privatization program saved young lives Public water companies serving about one-third of Argentina's ment in new equipment. Connections increased and quality municipalities and covering almost 60 percent of the population improved on a number of fronts, including the speed of repairs, were transferred to private control between 1991 and 1999.94The water leakages, sewerage blockages repaired, and percentage of largest privatization was the 1993 transfer of the federal company clients with appropriate water pressure. Obras Sanitarias de la Nácion (OSN) to the private consortium A recent study assesses the impact of privatization.94 In the Aguas Argentinas, led by the French company Lyonnaise des country as a whole and especially in Buenos Aires, access grew at Eaux.The concession required that 100 percent of households be a faster rate between 1991 and 1997 in municipalities in which connected to piped water by the end of the 35-year concession water had been privatized than in municipalities in which it had and 95 percent of households be connected to the sewerage sys- not (see figure). The expansion of coverage was pronounced tem. Quality standards were also imposed for service and waste among poorer households.The study also finds a beneficial effect treatment.Use fees and connection charges were regulated,and-- of privatization on child mortality, especially for poorer munici- in response to protests over a large increase in the connection palities. The study finds that mortality decline was faster for fee--a fixed fee was introduced for all customers.The new rev- causes of death that are common among the poor (infectious, enues subsidized the cost of new connections. Aguas Argentinas parasitic, and perinatal disease). cut the old OSN labor force and dramatically increased invest- Source: World Bank staff. Water privatization in Argentina reduced child mortality for the extremely poor by more than 25 percent Impact of privatization on Impact of privatization on access to piped water child mortality 8 0 in 97­ 6 in rate 1991 ­10 connected change 4 mortality change ­20 2 households Percentage under-five Percentage 0 ­30 Not Privatized Difference Extremely Poor Non-poor privatized poor Argentina Municipalities Buenos Aires Source: Reference 94. minutes and by 22 minutes for the poorest 40 percent of complications arriving at the hospital from the project area the population.85 The benefits of using existing roads to increased from 0.9 to 2.6 a month, and the case fatality rate improve transportation to and between health facilities are dropped from 20 percent to 10 percent. also evident. Malaysia and Sri Lanka provide free or subsi- Other options exist for increasing accessibility dized transportation to hospitals in the case of emergen- cies.43 Brazil recently made major investments in Upgrading and improving roads and improving trans- ambulances to facilitate rapid referral from a primary care portation between existing facilities are options for facility to the hospital. In Sierra Leone focus group discus- improving accessibility to health facilities. But other sions made it clear that poor transportation was a major options may be cheaper and more effective. Providers impediment to emergency obstetric care.So the city hospi- could take services to clients through an outreach program tal in Bo acquired a four-wheel drive vehicle and two-way using existing facilities. Governments could construct new radios to link the hospital, the primary care units, and the public facilities where access to existing facilities is limited. vehicle.86 The number of women with major obstetric They could form partnerships with NGOs, private Households--Key but Underrated Actors in the Health Sector |81 providers, or community organizations to provide addi- Changes that improve accountability along one or more tional services in underserved areas. Or they could con- of the lines suggested by the World Development Report can tract with NGOs, private providers, and community improve access--for everyone.In Argentina the provision of organizations to provide publicly financed services to water services was privatized under a regulatory regime that underserved communities. (These service delivery options laid down strict rules for quality, access, and price. The are discussed in chapter 6. The role of roads and trans- regime was revised after pressure from the community, portation is discussed in chapter 8.) which saw the price rises authorized as harmful to the interests of the poor.The World Development Report observes that "community involvement is essential in the regulatory Water and sanitation process but has not been sufficiently encouraged." It notes Hygiene and the quality and quantity of drinking water with approval the consultations that preceded the conces- depend on infrastructure. Hand-washing is easier if the sion processes in Manila and SouthAfrica.And it reports the household has piped water that provides readily available results of a survey showing how a majority of Peruvians quantities of safe water.The safe disposal of feces is easier if would have favored the privatization of the electric utility if the household has an improved form of sanitation (a con- the process had been transparent and tariff increases had nection to a public sewer or septic system, a pour-flush, been under the control of the regulator. (The intersectoral ventilated-improved, or a simple pit latrine).The hygienic synergies of water and health are discussed in chapter 8.) storage of food is easier if the household has electricity. And so on. Not surprisingly, therefore, the availability of References plenty of water and improved sanitation are associated with better maternal and child health outcomes, at least 1. Narayan, D., R. Patel, K. Schafft,A. Rademacher, and S. Koch­Schulte. 2000. Voices of the Poor: Can Anyone Hear Us? among the better educated, even after controlling for NewYork: Oxford University Press. other influences.30,32,67,87­92 2. Aguilar,A.M., R.Alvarado, D. Cordero, P. Kelly,A. Zamora, and The developing world lags well behind the industrial R. Salgado. 1998. Mortality Survey in Bolivia:The Final Report. world in water and sanitation: 1.1 billion people live with- Basic Support for Institutionalizing Child Survival (BASICS) out access to safe drinking water, and 2.4 billion live with- Project,Arlington,VA. out access to improved sanitation.93 In the 1990s the least 3. Hill, Z., B. Kirkwood, and K. Edmond. 2003. Family and Community Practices that Promote Child Survival, Growth, and developed countries saw a decline in safe water coverage, Development:A Review of the Evidence. London School of and in Sub-Saharan Africa there was no increase in the cov- Hygiene, Department of Epidemiology and Population Health, erage of improved sanitation.Asia lags behind in improved Public Health Intervention Research Unit. sanitation, with coverage rates of 34 percent in South Asia 4. Foster, G., and J.Williamson. 2000."A Review of Current Literature on the Impact of HIV/AIDS on Children in Sub- and 48 percent in East Asia and the Pacific. Although East Saharan Africa." AIDS 14 (Suppl. 3): S275­S284. Asia and the Pacific has a higher coverage rate for safe water 5. Barbhuiya, M.A., S. Hossain, M.M. Hakim, and S.M. Rahman. than Sub-Saharan Africa,it has the largest number of people 2001."Prevalence of Home Deliveries and Antenatal Care who lack access. Poor people fare especially badly: they are Coverage in Some SelectedVillages." Bangladesh Medical less likely to be connected to a network, and the sources Research Council Bulletin 27 (1): 19­22. they rely on tend to be more costly per liter than the net- 6. Wong, E., B. Popkin, D. Guilkey, and J.Akin. 1987. "Accessibility, Quality of Care and Prenatal Care Use in the worked services used by the better off.18 Philippines." Social Science and Medicine 24 (11): 927­944. The issues of increasing access to water and sanitation 7. Panis, C., and L. Lillard. 1994."Health Inputs and Child have recently been examined in the World Bank's World Mortality." Journal of Health Economics 13 (4): 455­489. Development Report 2004: Making Services Work for Poor 8. Guilkey, D., and R. Riphahn. 1998."The Determinants of People.18 The report notes the need for provider organiza- Child Mortality in the Philippines: Estimation of a Structural Model." Journal of Development Economics 56 (2): 281­305. tions to be accountable to policymakers,for policymakers to 9. Schwartz, J., J.Akin, and B. Popkin. 1988."Price and Income be accountable to households, and for front-line providers Elasticities of Demand for Modern Health Care:The Case of to be accountable directly to households through user Infant Delivery in the Philippines." World Bank Economic Review charges and competition. Publicly managed water and sani- 2 (1): 49­76. tation networks sometimes seem more directed at serving 10. Steele, F., I. Diamond, and S.Amin. 1996."Immunization the interests of politicians than households, while unregu- Uptake in Rural Bangladesh:A Multilevel Analysis." Journal of the Royal Statistical Society. Series A 159 (2): 289­299. lated private providers have sometimes been criticized for 11. Gage,A.J.,A.E. Sommerfelt, and A.L. Piani. 1997."Household delivering low-quality services and ignoring the poor. Structure and Childhood Immunization in Niger and Nigeria." Demography 34 (2): 295­309. 82 | The Millennium Development Goals for Health 12. Rashed, S., H. Johnson, P. Dongier, R. Moreau, C. Lee, 32. Benefo, K., andT. Schultz. 1996."Fertility and Child Mortality R. Crepeau, J. Lambert,V. Jefremovas, and C. Schaffer. 1999. in Côte d'Ivoire and Ghana." World Bank Economic Review 10 "Determinants of the Permethrin Impregnated Bednets (PIB) (1): 123­158. in the Republic of Benin:The Role of Women in the 33. Alderman, H. 1993."New Research on Poverty and Acquisition and Utilization of PIBs." Social Science and Medicine Malnutrition." In Including the Poor, eds. M. Lipton and J.Van 49 (8): 993­1005. Der Gaag.Washington, DC:World Bank. 13. Filmer,D.2000."Fever and ItsTreatment among the Poor and 34. Hanson, K., and E.Worrall. 2002."Report on the Analysis of Less Poor in Sub-Saharan Africa."World Bank,Washington,DC. SMITN2 End-of-Project Survey." London School of Hygiene 14. Sen,A.K. 1981. Poverty and Famines:An Essay on Entitlement and andTropical Medicine. Deprivations. Oxford: Clarendon Press. 35. Peters,D.H.,A.S.Yazbeck,R.Sharma,G.Ramana,L.H.Pritchett, 15. Drèze, J., and A.K. Sen. 1989. Hunger and Public Action. Oxford: and A.Wagstaff.2002.A Better Health Systems for India's Poor: Clarendon Press. Findings,Analysis,and Options.Washington,DC:World Bank. 16. Sen,A.K. 1994."The Political Economy of Hunger." In 36. Wagstaff,A. 2001."Inequalities in Health in Developing Overcoming Global Hunger: Proceedings of a Conference on Actions to Countries: Swimming against theTide?" Policy Research Reduce HungerWorldwide, eds. I. Serageldin and P. Landell­Mills. Working Paper 2795,World Bank,Washington, DC. Washington, DC:World Bank. 37. Castro-Leal, F., J. Dayton, L. Demery, and K. Mehra. 1999. 17. Pritchett, L., and L.H. Summers. 1996."Wealthier Is Healthier." "Public Social Spending in Africa: Do the Poor Benefit?" World Journal of Human Resources 31 (4): 841­868. Bank Research Observer 14 (1): 49­72. 18. World Bank. 2004. World Development Report 2004: Making 38. Castro-Leal, F., J. Dayton, L. Demery, and K. Mehra 2000. ServicesWork for Poor People.Washington, DC:World Bank. "Public Spending on Health Care in Africa: Do the Poor 19. Case,A. 2001."Does Money Protect Health Status? Evidence Benefit?" Bulletin of theWorld Health Organization 78 (1): 66­74. from South African Pensions." Center for Health andWellbeing 39. Filmer, D., J. Hammer, and L. Pritchett. 2000."Weak Links in Working Paper,WoodrowWilson School, Princeton University, the Chain:A Diagnosis of Health Policy in Poor Countries." Princeton, NJ. World Bank Research Observer 15 (2): 199­224. 20. Case,A. 2001."Health, Income, and Economic Development." 40. Filmer, D., J. Hammer, and L. Pritchett. 2002."Weak Links in Paper presented at the AnnualWorld Bank Conference on the Chain II:A Prescription for Health Policy in Poor Development Economics, Washington, DC. Countries." World Bank Research Observer 17 (1): 47­66. 21. Case,A., and A. Deaton. 1998."Large CashTransfers to the 41. Sahn,D.,and S.Younger.2000."Expenditure Incidence in Africa: Elderly in South Africa." Economic Journal 108 (450): Microeconomic Evidence."Fiscal Studies 21 (3):329­348. 1330­1361. 42. Yaqub,S.1999."How Equitable Is Public Spending on Health and 22. Ayé, M, F. Champagne, and A.P. Contandriopoulos. 2002. Education?"Background Paper for the World Development Report "Economic Role of Solidarity and Social Capital in Accessing 2000/2001:Attacking PovertyWorld Bank,Washington,DC. Modern Health Care Services in the Ivory Coast." Social Science 43. Pathmanathan, I., J. Liljestrand, J.M. Martins, L.C. Rajapaksa, and Medicine 55 (11): 1929­1946. C. Lissner,A. De Silva, S. Selvaraju, and P.J. Singh. 2003. 23. Nahar, S., and A. Costello. 1998."The Hidden Cost of `Free' "Investing in Maternal Health: Learning From Malaysia and Sri Maternity Care in Dhaka, Bangladesh." Health Policy Planning Lanka."World Bank, Health, Nutrition, and Population 13 (4): 417­422. Department,Washington, DC. 24. World Bank, SIDA,AusAID, Royal Netherlands Embassy, 44. Koblinsky, M.A., and O. Campbell 2003."Factors Affecting the Ministry of Health of Vietnam. 2001. Vietnam. Growing Healthy: Reduction of Maternal Mortality." In Reducing Maternal A Review ofVietnam's Health Sector. Hanoi:World Bank. Mortality: Learning from Bolivia, China, Egypt, Honduras, Indonesia, 25. Lewis, M. 2000."Who Is Paying for Health Care in Eastern Jamaica, and Zimbabwe, ed. M.A. Koblinsky.Washington, DC: Europe and Central Asia?"World Bank,Washington, DC. World Bank. 26. Al Fadil,S.M.,S.H.A.Alrahman,S.Cousens,F.Bustreo, 45. Schellenberg,J.R.,S.Abdulla,R.Nathan,O.Mukasa,T.J. A.Shadoul,S.Farhoud,and S.M.El Hassan.2003."Integrated Marchant,N.Kikumbih,A.K.Mushi,H.Mponda,H.Minja, Management of Childhood Illnesses Strategy:Compliance with H.Mshinda,M.Tanner.,and C.Lengeler.2001."Effect of Large- Referral and Follow-Up Recommendations in Gezira State, Scale Social Marketing of Insecticide-Treated Nets on Child Sudan."Bulletin of theWorld Health Organization 81 (10):708­716. Survival in RuralTanzania."Lancet 357 (9264):1241­1247. 27. Dodd, R., and L. Munck. 2001. Dying For Change: Poor People's 46. Abdulla, S., J.A. Schellenberg, R Nathan., O. Mukasa, Experience of Health and Ill-Health. Geneva:World Health T. Marchant,T. Smith, M.Tanner, and C. Lengeler. 2001. Organization;Washington, DC:World Bank. "Impact on Malaria Morbidity of a Programme Supplying Insecticide-Treated Nets in Children Aged under 2Years in 28. Alderman, H., andV. Lavy. 1996."Household Responses to Tanzania: Community Cross-Sectional Study." British Medical Public Health Services: Cost and QualityTradeoffs." World Bank Journal 322 (7281): 270­273. Research Observer 11 (1): 3­22. 47. Armstrong Schellenberg, J.,A. Mushi, H. Mponda, and 29. Roemer, M.I. 1981."The Health Care System ofThailand:A C. Lengeler. 2002."DiscountVouchers forTreated Nets in Case Study." Report No. 11,World Health Organization Tanzania:Targeted Subsidy for Malaria Control." Paper Regional Office for Southeast Asia, Manila. presented at a meeting on malaria and equity organized by the 30. Lavy,V., J. Strauss, D.Thomas, and P. DeVreyer. 1996."Quality of World Bank and the London School of Hygiene andTropical Care, Survival and Health Outcomes in Ghana." Journal of Medicine, London. Health Economics 15 (3): 333­357. 48. Saadah, F., M. Pradhan, and R. Sparrow. 2001."The 31. Thomas, D.,V. Lavy, and D. Strauss. 1996."Public Policy and Effectiveness of the Health Card as an Instrument to Ensure Anthropometric Outcomes in the Côte d'Ivoire." Journal of Access to Medical Care for the Poor During the Crisis."World Public Economics 61 (2): 155­192. Bank,Washington, DC. Households--Key but Underrated Actors in the Health Sector | 83 49. Yip,W., and P. Berman. 2001."Targeted Health Insurance in a Therapy for Acute Diarrhea in a Region of Northeastern Low Income Country and Its Impact on Access and Equity in Brazil, 1986­1989." [In Portugese.] Revista de Saude Publica 28 Access: Egypt's School Health Insurance." Health Economics 10 (6): 416­422. (3): 207­220. 66. Coreil, J., and E. Genece. 1988."Adoption of Oral Rehydration 50. Gorter,A.,P.Sandiford,Z.Segura,and C.Villabella.1999. Therapy among Haitian Mothers." Social Science and Medicine 27 "Improved Health Care for SexWorkers:AVoucher Program for (1): 87­96. Female SexWorkers in Nicaragua."Research for SexWork 2:11­14. 67. Jalan, J., and M. Ravallion 2001."Does PipedWater Reduce 51. Mesoamerica Nutrition ProgramTargeting Study Group. 2002. Diarrhea for Children in Rural India?" Policy Research "Targeting Performance ofThree Large-Scale, Nutrition- Working Paper 2664,World Bank,Washington, DC. Oriented Social Programs in Central America and Mexico." 68. Favin, M., and M. Griffiths. 1999. Communication for Behavior Food Nutrition Bulletin 23 (2): 162­174. Change in Nutrition Projects.Washington, DC:World Bank. 52. International Food Policy Research Institute. 2002. Final 69. Semega-Janneh, I.J., E. Bohler, H. Holm, I. Matheson, and Report: Nicaragua Social Protection Network, Pilot Evaluation G. Holmboe-Ottesen 2001."Promoting Breastfeeding in Rural System, Impact Evaluation.Washington, DC: IFPRI. Gambia: CombiningTraditional and Modern Knowledge." 53. International Food Policy Research Institute. 2003. Sexto Health Policy Plan 16 (2): 199­205. informe. Proyecto PRAF/BID Fase II: Impacto intermedio. 70. Ndyomugyenyi, R., S. Neema, and P. Magnussen. 1998."The Washington, DC: IFPRI. Use of Formal and Informal Services for Antenatal Care and 54. International Food Policy Research Institute. 2003. Estudo de MalariaTreatment in Rural Uganda." Health Policy Plan 13 (1): avaliação de impacto para o Programa Bolsa Alimentação. Relatório 3: 94­102. ánalise de impacto final.Washington, DC: IFPRI. 71. Desai, S., and S.Alva. 1998."Maternal Education and Child 55. Gertler, P., and S. Boyce. 2001."An Experiment in Incentive- Health: IsThere a Strong Causal Relationship?" Demography 35 BasedWelfare:The Impact of PROGRESA on Health in (1): 71­81. Mexico." University of California, Berkeley. 72. Hobcraft, J. 1993."Women's Education, ChildWelfare and 56. Morris, S., R. Flores, P. Olinto, and J. Medina. 2003."A Child Survival:A Review of the Evidence." HealthTransition RandomizedTrial of Conditional CashTransfers to Households Review 3 (2): 159­175. and Peripheral Health Centers: Impact on Child Health and 73. Glewwe, P. 1999."Why Does Mother's Schooling Raise Child Demand for Health Services." Paper presented at the Fourth Health in Developing Countries?" Journal of Human Resources International Health Economics AssociationWorld Congress, 34 (1): 124­59. San Francisco. 74. Abbas, K.A. 1995."Counselling in a Hospital Setting. 57. Beegle, K., E. Frankenberg, and D.Thomas. 2001."Bargaining Breastfeeding." Dialogue Diarrhoea (59): 3. Power within Couples and Use of Prenatal and Delivery Care in Indonesia." Studies in Family Planning 32 (2): 130­146. 75. Santos, I., C.G.Victora, J. Martines, H. Goncalves, D.P. Gigante, N.J.Valle, and G. Pelto. 2001."Nutrition Counseling Increases 58. Pitt, M.M., S.R. Khandker, S.M. Mckernan, and M Abdul Latif. Weight Gain among Brazilian Children." Journal of Nutrition 1999."Credit Programs for the Poor and Reproductive 131 (11): 2866­2873. Behavior in Low-Income Countries:Are the Reported Causal Relationships the Result of Heterogeneity Bias?" Demography 76. Bhandari, N., R Bahl, J. Martines, and M.K. Bhan. 2002. 36 (1): 1­21. "A Community-Based Randomized Controlled 59. Ranson, M. 2002."Reduction of Catastrophic Health Care Intervention to Improve Breast Feeding and Complementary Expenditures by a Community-Based Health Insurance Feeding Practices through Nutritional Counselling in Rural Scheme in Gujarat, India: Current Experiences and Haryana, India." Presentation at theWHO Regional Advisers Challenges." Bulletin of theWorld Health Organization 80 (8): Meeting. 613­621. 77. Hill, Z., B. Kirkwood, and K. Edmond. 2001."Family and 60. Pande, R.P., and A.S.Yazbeck. 2003.What's in a Country Community Practices that Promote Child Survival, Growth, Average?Wealth, Gender, and Regional Inequalities in and Development:A Review of the Evidence." London School Immunization in India." Social Science and Medicine 57 (11): of Hygiene, Department of Epidemiology and Population 2075­2088. Health, Public Health Intervention Research Unit. 61. Das,J.,and C.Sánchez-Párano.2003."Short But Not Sweet:New 78. Saade,C.,M.Batemen,and D.B.Bendahmane.2001.The Story of Evidence on Short Duration Morbidities from India."Policy a Successful Public-Private Partnership in Central America. Arlington, ResearchWorking Paper 2971,World Bank,Washington,DC. VA:BASICS II,EHP,UNICEF,USAID,andWorld Bank. 62. Cebu StudyTeam. 1991."Underlying and Proximate Determi- 79. Ainsworth, M., C. Beyrer, and A. Soucat. 2003."AIDS and nants of Child Health:The Cebu Longitudinal Health and Public Policy:The Lessons and Challenges of `Success' in Nutrition Study." American Journal of Epidemiology 33 (2): Thailand." Health Policy 64 (1): 13­37. 185­201. 80. Heerey, M., and M.A. Kols. 2003."Improving the Quality of 63. Pebley,A.R., N. Goldman, and G. Rodriguez. 1996."Prenatal Care: Quality Improvement Projects from the Johns Hopkins and Delivery Care and Childhood Immunization in University Bloomberg School of Public Health Center for Guatemala: Do Family and Community Matter?" Demography Communication Programs." Center Publication No. 101, Johns 33 (2): 231­247. Hopkins University, Baltimore.. 64. Gertler, P., O. Rahman, C. Feifer, and D.Ashley. 1993. 81. Alubo, O. 2001."The Promise and Limits of Private Medicine: "Determinants of Pregnancy Outcomes andTargeting of Health Policy Dilemmas in Nigeria." Health Policy Plan 16 (3): Maternal Health Services in Jamaica." Social Science and Medicine 313­321. 37 (2): 199­211. 82. Mwabu, G., M.Ainsworth, and A. Nyamete. 1993."Quality of 65. Galvao, C.E.,A.A. Da Silva, R.A. Da Silva, S.A. Dos Reis Filho, Medical Care and Choice of MedicalTreatment in Kenya:An M.A. Novochadlo, and G.J. Campos. 1994."Oral Rehydration Empirical Analysis." Journal of Human Resources 28 (4): 838­862. 84 | The Millennium Development Goals for Health 83. Diop, F.,A. Soucat,A.Wagstaff, and F. Zhao. 2002."Health and 91. Wolfe, B., and J. Behrman. 1982."Determinants of Child the Poor in Senegal." Draft chapter for Senegal Country Mortality, Health and Nutrition in a Developing Country." Economic Memorandum.World Bank,Washington, DC. Journal of Development Economics 11 (2): 163­93. 84. Narayan-Parker, D., and P.L. Petesch. 2002. From Many Lands: 92. Behrman, J., and B.Wolfe. 1987."How Does Mother's Voices of the Poor.Washington, DC: Oxford University Press and Schooling Affect Family Health, Nutrition, Medical Care World Bank. Usage and Household Sanitation? Journal of Econometrics 36 (1): 85. Van DeWalle, D., and D. Cratty. 2002."Impact Evaluation of a 185­204. Rural Road Rehabilitation Project."World Bank,Washington, 93. UNICEF. 2001. Progress Since theWorld Summit for Children:A DC. Statistical Review. NewYork: UNICEF. 86. Samai, O., and P. Sengeh. 1997."Facilitating Emergency 94. Galiani, S., P. Gertler, and E. Schargrodsky. 2002."Water for Obstetric Care throughTransportation and Communication, Life:The Impact of the Privatization ofWater Services on Child Bo, Sierra Leone." Bo PMMTeam. International Journal of Mortality."Working Paper No. 154, Stanford University Center Gynaecology and Obstetrics 59 (Suppl. 2): S157­S164. for Research on Economic Development and Policy Forum, 87. Merrick,T.W. 1985."The Effect of PipedWater on Early Stanford, CA. Childhood Mortality in Urban Brazil, 1970 to 1976." 95. Filmer D. 2003."The Incidence of Public Expenditures on Demography 22 (1): 1­24. Health and Education." Background note for the World 88. Ridder, G., and I.Tunali. 1999."Stratified Partial Likelihood Development Report 2004: Making ServicesWork for Poor People. Estimation." Journal of Econometrics 92 (2): 193­232. World Bank,Washington, DC. 89. Lee, L.­F., M. Rosenzweig, and M. Pitt. 1997."The Effects of 96. Gwatkin,D.,S.Rutstein,K.Johnson,R.Pande,and A.Wagstaff. Improved Nutrition, Sanitation, andWater Quality on Child 2000."Socioeconomic Differences in Health,Nutrition and Health in High-Mortality Populations." Journal of Econometrics Population."Health,Nutrition and Population Discussion Paper. 77 (1): 209­235. World Bank,Washington,DC.http://www.Worldbank.Org/ Poverty/Health/Data/index.Htm. 90. Esrey, S.A., and J.P. Habicht. 1988."Maternal Literacy Modifies the Effect ofToilets and PipedWater on Infant Survival in 97. ORC Macro. 2003. Demographic and Health Surveys. Malaysia." 1988. American Journal of Epidemiology 127 (5): http://www.Measuredhs.Com/. 1079­1087. Households--Key but Underrated Actors in the Health Sector |85 Table 5.1 Reducing barriers facing households in the use of effective child health interventions Brazil, Egypt, Mexico, Philippines: Promotion of oral rehydration Costa Rica, Bolivia: National Maternal therapy in National El Salvador, Guatemala: and Child Insurance Control of Diarrheal Central American Honduras: Integrated Item Program Diseases Programs Hand-Washing Initiative Child Care Program Background to High maternal and child mortal- In 1980 diarrhea was the lead- In 1995 diarrhea was the In the late 1980s international policy shift ity have plagued Bolivia for ing cause of child mortality in cause of 45 percent of under- and local research prompted decades. In 1994 the under-five the world, accounting for 4.6 five mortality in Guatemala key actors in the Ministry of mortality rate was 132 per million deaths a year. Diarrhea and 20 percent in El Salvador. Health to emphasize that mal- 1,000 live births and the mater- accounted for 41 percent of Hand-washing with soap can nutrition imposes a severe nal mortality rate 390 per infant mortality in Brazil and have an important impact on burden on the population, 100,000 live births. Both indica- was the second-leading cause the incidence of diarrhea. But that malnutrition is the under- tors were among the highest in of under-five mortality, 78 percent of mothers in lying cause of more than half the region. Use of health accounting for 8.5 deaths per Guatemala used inadequate of the deaths in children services was very low in the 1,000 population. Oral rehy- hand-washing practices in under five, and that mid-1990s. In 1996 only 53 dration therapy was 1996. Lack of knowledge con- inadequate weight gain rather percent of pregnancies introduced in 1979 as an inter- stituted an important barrier to than nutrition was the best received medical attention, less vention to treat diarrhea. utilization. early warning sign of a sick than 45 percent of births Consisting of the oral adminis- child. occurred in health facilities, and tration of sodium, a carbohy- only 36 percent of women in drate, and water, oral labor were attended by skilled rehydration therapy has been medical personnel. The cost of described as potentially the health services was considered most significant medical one of the most important advance of the twentieth cen- barriers to utilization. tury. It was recognized in 1980 that barriers to utilization included lack of knowledge and availability and accessibility concerns. Policy shift Decentralization of the govern- In recognition of the severe The Central American Hand- Honduras began to reform its ment structure in 1994 burden of disease imposed by Washing Initiative was imple- health sector in the early changed the financing and diarrhea and the fact that mented in Costa Rica, El 1990s, with decentralization a delivery of health services. effective interventions such as Salvador, and Guatemala main goal. The strategy Twenty percent of government oral rehydration therapy had between 1996 and 1999. The emphasized the critical role of revenues were shifted to the become available, a number of initiative's goal was to reduce communities in the health country's municipalities, and a countries introduced National morbidity and mortality among sector. As part of these reform certain proportion of those Control of Diarrheal Diseases children under five by promot- efforts, in 1994 the Ministry funds was to be earmarked Programs in the early 1980s. ing proper hand-washing with of Health established the for the health sector. This shift The programs aimed to reduce soap. Integrated Child Care significantly improved equity, morbidity and mortality from Program as a critical compo- since it was the first time that dehydration associated with nent of its child health strat- rural municipalities were given acute diarrhea in children egy. The program had been access to these funds. Owner- under five. National Control of successfully pilot tested in ship of health facilities and Diarrheal Diseases Programs 1992­93. By 1994 the focus responsibility for financing of were operational in 35 coun- of the program had shifted equipment and basic inputs tries by 1983 and in 80 coun- from interventions at the facil- were also decentralized to the tries by 1990. The experience ity level to interventions at the municipalities. Management of of Brazil, Egypt, Mexico, and household and family level human resources was decen- the Philippines is highlighted through increased community tralized to the Ministry of here because the impact in participation. At the same Health's regional adminis- these countries has been evalu- time, the malnutrition para- trations. In 1996 it was deter- ated in detail. digm changed from disease mined that a portion of the treatment to health promo- central government's funds tion. should be targeted to a new special program, aimed exclu- sively at pregnant women and children under five. The new program, called the National Maternal and Child Insurance Program, aimed to reduce maternal and child mortality rates by removing fees for key health interventions. Because it allocated funds on a per capita basis, it constituted a propoor financing mechanism. Historically, financing had been biased toward richer urban municipalities. 86 | The Millennium Development Goals for Health Table 5.1 Reducing barriers facing households in the use of effective child health interventions (continued) Brazil, Egypt, Mexico, Philippines: Promotion of oral rehydration Costa Rica, Bolivia: National Maternal therapy in National El Salvador, Guatemala: and Child Insurance Control of Diarrheal Central American Honduras: Integrated Item Program Diseases Programs Hand-Washing Initiative Child Care Program Programmatic The program provided key The program focused on mak- This public-private partnership The program aims to action health interventions, such as ing oral rehydration therapy included the ministries of health improve health by promoting antenatal and neonatal care available on a large scale, and education, soap producers, the healthy growth of chil- and treatment of acute respi- through primary care facilities, NGOs, the media, and the pro- dren. The main focus is on ratory infections and diarrhea for example. This effort was ject facilitator, the U.S. Agency monthly growth-monitoring free of charge to pregnant combined with extensive train- for International Development and promotion sessions. mothers and children under ing of health workers and (through Basic Support for During the sessions, commu- five on a universal basis. Both information campaigns, includ- Institution Child Survival and nity volunteers (monitoras) public and private institutions ing mass media, to inform the the Environmental Health weigh each child under two, were included in the program. general public about the avail- Project). The initiative began by assess the growth rate rela- The participating institutions ability and benefits of oral analyzing the hand-washing tive to the expected weight included Ministry of Health rehydration therapy. behavior of the targeted popu- gain, and provide counseling facilities, social security hospi- lation. Public health messages to the caregiver. Seriously ill tals, and private nonprofit about hand-washing (when children are referred to organizations. and how to wash) were then health centers, where they delivered through the mass also receive immunizations. media and NGOs, while soap Training is provided to nurses producers used their marketing of the health centers and the skills to sell soap by promoting monitoras. hand-washing with soap. Affordability The program explicitly aimed Mobile units distributed free The government subsidizes to promote demand by reduc- soap samples to households. the main components of ing the cost barriers for users the program, though com- and guaranteeing free access munity leaders and mothers to services. incur some costs. A recent cost analysis of the program found that it is both inexpen- sive and effective. Women's Women's autonomy was The program empowers autonomy increased, as the program women to prevent malnu- enabled them to treat many trition by making small episodes of diarrhea them- changes in behavior using selves, thereby avoiding costly existing resources within visits to health centers and the household. It builds admissions to hospitals. women's confidence because monthly feedback on the changes in child growth show them that they can make a difference by themselves. Knowledge Through promotional efforts, Efforts to educate the public The main barrier to utilization During the counseling such as public announcements significantly increased knowl- addressed by the initiative was sessions, caregivers are on radio and television and in edge of the availability and knowledge. To increase trained in how to maintain or community charlas (informal appropriate use of oral demand for soap, the initiative improve the growth of their discussions), the government rehydration therapy. An evalua- stressed the importance of children. Key messages succeeded in informing the tion of the National Control of washing hands with soap at include breastfeeding, child public about the program Diarrheal Diseases Program in critical times and in appropriate feeding, care for sick children, and, therefore, about health Egypt showed that most moth- ways and stressed that hand- and hygiene. services in general. ers were able to mix oral rehy- washing reduces the incidence dration therapy correctly. of diarrhea. Accessibility Decentralization and the pro- Accessibility was affected by Accessibility was increased by Coverage of growth monitor- gram increased accessibility increasing the supply of oral mobile units distributing free ing and promotion was by bringing health services rehydration therapy and bring- soap samples and by school deemed critical to the closer to the people. ing it closer to clients. Local programs. program's success. As the Increased coverage reached production and imports of sup- government realized that not only the urban rich but plies increased substantially. In adequate coverage could not also rural and poor Mexico the availability of oral be achieved if these services populations. In addition, since rehydration salts increased were available only at the primary-level facilities had from 7.6 in 1984 to 79.7 mil- health-facility level, the pro- more control over funds, they lion packets in 1993. Extensive gram focuses on the house- were better able to stock the promotion of oral rehydration hold and community level, facilities with drugs and supplies. therapy in primary care facilities bringing services closer to the made therapy available closer people. The monitoras make to the population. home visits to children who were not present during the monthly sessions and to sick children or children with inad- equate weight gain. Households--Key but Underrated Actors in the Health Sector | 87 Table 5.1 Reducing barriers facing households in the use of effective child health interventions (continued) Brazil, Egypt, Mexico, Costa Rica, Philippines: Promotion El Salvador, Bolivia: of oral rehydration Guatemala: Central National Maternal therapy in National American and Child Insurance Control of Diarrheal Hand-Washing Honduras: Integrated Item Program Diseases Programs Initiative Child Care Program User-friendliness User-friendliness is built into the basic design of the program, since the development of counseling messages is conducted through trials of improved practices in the homes of beneficiaries. Water and Water and sanitation is sanitation addressed indirectly through quarterly community meetings, at which community problems and solutions, including water and sanitation issues, are discussed. Impact of Introduction of the program Utilization of oral rehydration ther- Hand-washing behavior A recent mid-term evaluation of policies on resulted in substantial apy increased in all four countries. improved as a result of the the program found significant utilization increases in the utilization of In Mexico use of oral rehydration initiative, both in terms of improvements in the knowledge key maternal and child therapy increased from 47.5 per- hand-washing at critical and behavior (caregiving and feed- health interventions, partic- cent of all diarrhea episodes in times and hand-washing ing practices) of caregivers in the ularly among the poor. 1986 to 80.7 percent in 1993. In technique. A recent evalua- program. Program coverage is Skilled birth attendance the Philippines the access rate tion of the initiative in almost universal (92 percent) in the increased from 37 percent increased from 60 percent to 85 Guatemala found that communities surveyed by the eval- to 60 percent in nonindige- percent between 1989 and 1992. inadequate hand-washing uation. Compared with control nous municipalities between practices of mothers communities, program caregivers 1996 and 2001; in indige- decreased from 78 percent were significantly more likely to nous municipalities, it to 68 percent between have received counseling on doubled, rising from 18 1996 and 1999. breastfeeding, care of sick children, percent to 36 percent. Intermediate and optimal hygiene, and iron and vitamin A Treatment of pneumonia in practices increased from 22 supplementation. Immunization children also increased percent to 32 percent. coverage, iron supplementation, significantly. Many of the use of oral rehydration therapy , people served under the and vitamin A intake were also program were people who much better in program communi- had previously not ties. The program has also used modern health services. contributed to more rational use of health facilities: nurses say that they no longer see patients whom they do not need to see, allowing them to focus more attention on people with severe disease. Impact of The under-five mortality Evaluations of the four countries In Guatemala the A case study found that malnutri- policies on rate was reduced from 132 show significant improvement in prevalence of diarrhea tion had been reduced by 10 per- health outcomes in 1994 to 99 in 1998, an health outcomes. In Brazil the pro- among children under five cent in most communities by average annual decrease of portion of infant mortality caused fell about 4.5 percent, 1999. about 6 percent. Maternal by diarrhea decreased from 41 per- affecting almost 1.6 million health also improved: by cent in 1980 to 25 percent in 1989. children under five. 2000 the maternal mortality In Egypt infant mortality caused by rate had dropped to 234. diarrhea fell by an annual average There is evidence that these of 15.9 percent between 1984 and achievements are linked 1990, a more rapid decline than for to the policy shift in 1994 mortality attributed to other causes. and the improved access In Mexico the proportion of under- made possible by public five deaths caused by diarrhea health insurance. A survey declined from 26.4 percent in 1983 on the use of public health to 11.0 percent in 1993. In the insurance revealed that the Philippines both infant and under- probability that women five mortality rates associated with receive skilled birth atten- diarrhea fell by about 5 percent dance (which increased annually between 1975 and 1993. during the time of the insur- Although other factors may explain ance) is linked to the these improved health outcomes, it maternal mortality rate. seems that the National Control of Diarrheal Diseases Program had an important impact. A simulation model performed in Brazil indicated that factors other than oral rehydration therapy explained only about a third of the actual decline. Source: World Bank staff. 88 | The Millennium Development Goals for Health Table 5.2 Reducing barriers to the use of effective maternal health interventions by households Item China, Yunnan Province Honduras Sri Lanka Background Although China has recorded impressive In 1990 the maternal mortality ratio In 1950, two years after independence, the to policy shift reductions in the maternal mortality ratio was 182 per 100,000 live births. maternal mortality ratio was as high as 555 since 1950 (when it was as high as 1,500 per 100,000 live births. per 100,000 live births), the ratio was still about 100­200 in rural areas in China in 1980. In 1989 it was 149 in Yunnan Province. Policy shift While policies are formulated at the The report of the high maternal Reducing maternal deaths was one of the national level, strategies to translate policies mortality ratio in 1990 alarmed the goals of the Ministry of Health of the new into action are determined at the provincial government and triggered a massive government that took office after indepen- level or below. Since 1980 Yunnan Province effort to reduce maternal mortality. The dence. Several decisions reveal the impor- has implemented several policy actions, government focused its efforts strate- tance it placed on maternal health. In the such as decentralization to increase local gically on areas with high maternal 1950s the government increased access ownership and accountability and skills mortality ratios. and coverage to basic maternal and child enhancement to reduce maternal deaths. health services by expanding the health unit system. The system emphasized pre- ventive and health promotion services at the community level, delivered by skilled attendants and field workers with guaran- teed professionalism. The management and infrastructure of maternal health pro- grams was improved with an effective monitoring and supervisory system. Transportation and communications were improved to provide safe transportation and referral of women with complications. Programmatic In the 1970s "barefoot doctors" staffed To lower the maternal mortality ratio, Efforts to reduce maternal mortality took action village clinics in the province. Since 1982 the government implemented a two- place in three phases. The first phase government regulations have required that pronged approach: ensuring that focused on making services accessible to doctors take an exam on applied aspects of women who develop obstetric emer- the majority of the population and improv- maternal health (attending clean delivery gencies are referred to the hospital and ing antenatal coverage, midwifery services, and identifying risk factors and complica- identifying women at high risk for com- and detection and early referral of compli- tions before and during delivery). If they plications and encouraging them to cations. A public health midwife is respon- pass, they become "village doctors" and deliver in a health facility. To support sible for all pregnant women in her receive on-site training. Career growth of this approach, the government built jurisdiction, covering a population of village barefoot doctors is thus linked to seven new rural area hospitals and, 4,000­5,000. The second phase involved demonstrated ability to provide maternal with community input, new maternity monitoring the field maternal and child health care. waiting homes and birthing centers. health program through improved report- Yunnan Province launched the Systematic The maternity waiting homes are fully ing systems. The third phase emphasized Management of Pregnant Women (SMPW) managed by the local communities. The quality-of-care considerations, particularly in 1985. The program provides every number of health staff, especially auxil- by using the findings of maternal death woman with a management booklet, at iary nurses, was increased. Skills of clini- inquiries to identify deficiencies in the sys- least five antenatal care checkups, a cal staff and community health tem and address those deficiencies through minimum of three postnatal visits in her workers, including traditional birth new initiatives. Provider accountability and home, and modern delivery methods. attendants, were enhanced through skill enhancement are critical components training, which focused on recognizing in the efforts to improve maternal health in The Emergency Referring System for risks in pregnancy and danger signs in Sri Lanka. Pregnant Women was formed in Tonghai childbirth and make appropriate refer- County in Yunnan Province in 1990. It aims rals. The Ministry of Health also to ensure that women with complications produced new technical manuals in this are moved from the village clinic to the area. county maternal and child health station or hospital. The Emergency Referring System consists of eight strategies: coordination, ensuring functioning phones at all levels, equipping emergency facilities (including vehicles), improving skills, enhancing the referral network, supervising and evaluat- ing, and focusing on meeting targets. Provider accountability is built into the schemes; facility grants, continuation of employment, and promotion of managers are determined on the basis of achievement of set targets. Households--Key but Underrated Actors in the Health Sector | 89 Table 5.2 Reducing barriers to the use of effective maternal health interventions by households (continued) Item China, Yunnan Province Honduras Sri Lanka Affordability A maternal and child health prepayment Sri Lanka has a very high level of govern- scheme was piloted and is now available ment commitment to, and popular support throughout the province. Households pay for, the welfare state, including universal a modest fee and are covered if complica- free access to health care. Although user tions associated with delivery or diseases of fees were introduced early on, it appears the child arise. Those who have joined this that patients are rarely charged for health scheme have sought maternal and child services, including drugs and supplies. health services more actively, though whether this reflects adverse selection or moral hazard is unclear. The scheme does not include delivery charges. Women's The autonomy of women was enhanced autonomy through the promotion of women's rights and empowerment by cross-sector initiatives. Knowledge Knowledge as a barrier to utilization was In focus group sessions, women in areas Special efforts were made to reach Indian addressed by providing every woman with where traditional birth attendants were Tamils living and working on estates, who the SMPW booklet and at least five antena- trained acknowledged the risk of going had not benefited from social change, tal care checks and by creating community to an untrained traditional birth including improvements in education and awareness and demand for the SMPW attendant. health. Following nationalization of the package. The government also provided estates in 1972, medical officers with trans- health education to pregnant women. portation facilities were appointed to estab- lish a network of estate health clinics. The clinics provided maternal and child health and family planning services to the estate population. Through this program, knowl- edge and skills of the estate health staff were improved. This is a good example of a strategy for improving access to underserved and neglected population groups. Accessibility To improve accessibility, the SMPW provides Accessibility was explicitly addressed by Accessibility was addressed by expanding every woman with access to antenatal care building maternity waiting homes, with the health service infrastructure to cover the and at least three postnatal visits in her community assistance, alongside sev- whole country and by rapidly educating and home. Access to facilities offering emer- eral hospitals. It had been recognized training health staff. Transportation was gency services was improved by providing that it was sometimes difficult for supported through a system of ambulance vehicles. women living in remote rural areas to services. Most facilities had ambulances by get to a hospital when labor began. 1950. In addition, the government The waiting homes provided a way to reimburses health staff for the cost of alter- ensure that women were already near native modes of transportation, enabling the hospital when labor began. them to better reach the population. Construction of new birthing homes brought skilled attendance at birth closer to women in hard-to-reach areas. User-friendliness Water and Following the acceptance by estate man- sanitation and agement of the initiative to reach estate other cross- workers with health interventions, the two sector linkages government estate agencies implemented a series of interventions to improve water and sanitation on the estates. In Sri Lanka as a whole, efforts were made to improve sanitation. Impact of In Tonghai an already high rate of hospital Skilled attendance increased from 46 Attended deliveries by personnel in govern- policies on delivery in 1990 (70 percent) was increased percent in 1987­91 to 54 percent in ment facilities increased from 33 percent in utilization to 92 percent. In a poorer county, Huaning, 1997. Maternity waiting homes 1950 to 87 percent in 1995, while deliver- the rate increased from 49 percent in 1990 increased the likelihood that high-risk ies in the home decreased from 25 percent to 62 percent in1999. women (older women and women and to 2 percent. Today 95 percent of births are who had already given birth several attended by a skilled practitioner, with the times) delivered in a hospital. majority taking place in hospitals. For estate women, deliveries in government facilities increased from 20 percent in 1986 to 63 percent in 1997, deliveries in estate mater- nity units rose from 29 percent to 42 per- cent, and deliveries in the home decreased from 37 percent to 8 percent. 90| The Millennium Development Goals for Health Table 5.2 Reducing barriers to the use of effective maternal health interventions by households (continued) Item China, Yunnan Province Honduras Sri Lanka Impact of In Tonghai county in Yunnan, an already Skilled attendance increased from 46 The maternal mortality ratio in Sri Lanka policies on high rate of hospital delivery in 1990 (70 percent in 1987­91 to 54 percent by declined from 555 in 1950 to 24 in 1995. health percent) was increased to 92 percent. In a 1997. Maternity waiting homes outcomes poorer county, Huaning, the rate increased increased the likelihood that high-risk from 49 percent to 62 percent between women (older women and women who 1990 and 1999. Between 1989 and 1998, had had several births previously) the maternal mortality ratio in Yunnan delivered in a hospital. Between 1990 decreased from 149 to 101. and 1997, the maternal mortality ratio fell from 182 to 108. Source: World Bank staff. Table 5.3 Reducing barriers to the use of effective interventions for malaria, HIV/AIDS, and tuberculosis by households Home-based treatment of malaria in Home-based care of people with Strengthening national tuberculosis Item Tigray region of Ethiopia HIV/AIDS in northern Thailand control interventions in China Background to Malaria is a leading killer of children under Two-thirds of the approximately 1 mil- Unlike in most developing countries, policy shift five in Africa. Almost a million children die lion people living with HIV/AIDS reside tuberculosis treatment in China was not in Africa from malaria every year. Eighty in the northern part of Thailand. exempt from the shift to fee-for-service percent of childhood malaria cases are Historically, rates of infection have been care. Case detection and cure rates were treated outside of the public sector. Most high among women who are or were low, and drug-resistant disease was treatment is provided at home, with drugs sex workers, but in the late 1990s the emerging. purchased from informal drug sellers or highest rates of increase were among pharmacists. The drugs purchased through women in stable relationships. At that drug sellers are often of poor quality or time, HIV prevalence among women counterfeit. Dosages purchased are fre- attending antenatal clinics was 4­8 per- quently inadequate to fully treat infection. cent. The Sanpatong area was one of the hardest hit by the epidemic. Government hospitals were hard- pressed to provide care, with as many as 60 percent of hospital beds occupied by people with HIV/AIDS. Policy shift It was recognized in the Tigray region of Volunteers from the faculty of medicine The World Bank­financed DOTS program Ethiopia that an effective strategy for at Chiang Mai University, Chiang Mai has provided free tuberculosis treatment ensuring that children receive rapid and public health staff, and nurses from the for infectious tuberculosis patients in parts effective treatment for malaria would have Sanpatong Red Cross health center, all of 16 provinces since 1992. Village doctors to focus on improving treatment practices with extensive community experience, and tuberculosis dispensaries provide DOTS in the household. This would include decided to pool their practical and aca- care; hospitals are supposed to refer improving mothers' ability to detect fever demic experience to develop a new patients. Providers receive payment from and malaria symptoms and to correctly community-based holistic model of care the Ministry of Health for each infectious treat their children with antimalarial drugs. to provide "bio-psycho-social support" case detected and each case cured. to help people living with HIV/AIDS be treated at home. The bio component involved treatment of symptoms, the psycho component helped people deal with stress and problems, and the social component helped improve the ability to cope with, participate in, and be accepted by the community. Programmatic Selected "mother coordinators" from the The project provided education and The National TB Control Program within action community were trained to provide educa- skills training to family members, volun- the new Center for Disease Control was tion to other mothers in their community teers from the community, and village strengthened to support capacity-building, on the symptoms and sign of fever and leaders. It recruited volunteer medical monitoring, and supervision of providers in malaria. Mother coordinators were also personnel and people living with the provinces and counties participating in supplied with chloroquine, which they HIV/AIDS through the community. the DOTS program. The logistics system provided to mothers with sick children was overhauled, and an efficient national along with information on how to drug procurement and supply system was administer the drug. developed. Households--Key but Underrated Actors in the Health Sector | 91 Table 5.3 Reducing barriers to the use of effective interventions for malaria, HIV/AIDS, and tuberculosis by households (continued) Home-based treatment of malaria in Home-based care of people with Strengthening national tuberculosis Item Tigray region of Ethiopia HIV/AIDS in northern Thailand control interventions in China Affordability The program provided mothers with free With an orientation toward building With free drug provision and consults, chloroquine, which costs about $0.08 per partnerships and alliances, the project there was a dramatic improvement in the child treatment dose. has created strong referral services from affordability of six-month therapy-- the home to the health center to the especially important given evidence of the district hospital and vice versa. higher prevalence of tuberculosis in poorer Partnerships between the public health counties and marginalized communities. system and the private sector have reduced the cost of care to both indi- vidual families and the government health services. Material support for school fees, food, and clothing has been obtained from private donations for needy families. Women's Mothers are the primary caregivers of sick In northern Thailand, HIV infection has Tuberculosis case detection rose for both autonomy children, who suffer most from malaria in often resulted in the illness of women male and female patients by increasing Africa. This strategy empowered mothers who are wives and mothers. Caregiving access to care in local communities and to take actions to effectively treat their in these situations falls to the grandpar- improving support systems. children in the home. ents, particularly grandmothers. Although Thai grandmothers have always had an important role in the care and upbringing of children, this unexpected burden in their later years is a heavy one. The project has sought to improve the ability of grandmothers to function as home caregivers through training, financial support from the Social Welfare Ministry, and emotional support. Knowledge A key component of this intervention was Training and education for family and There were no mass campaigns about the improving mothers' knowledge and ability community members is conducted on new service, but information was rapidly to detect fever and malaria and to provide home care, ways to decrease stigma, disseminated, as increased demand their sick children with proper treatment. and spirituality, using local Buddhist revealed. monks to teach meditation. Accessibility This approach dramatically increased Services have been brought to the Access to tuberculosis care increased as a access to effective treatment in the home, household level for hundreds of result of the proximity of village doctors to where most malaria treatment is provided. infected people and their families. patients and the provision of free care. User- The approach used was very straight- Providers got whole villages to accept Ambulatory care provided as part of the friendliness forward, with simple messages and skills and participate in the care of people DOTS strategy offers a better alternative to promoted by the mother coordinators. living with HIV/AIDS and their families hospitalization. and to adopt a holistic approach to health care. Water and sanitation Impact of This approach can be expected to alleviate The model has relieved pressure on Utilization increased and more than 1 mil- policies on the burden of severe malaria cases on government hospital beds and strength- lion patients were treated. Patients were utilization hospitals. ened the referral between community still seeking diagnosis in hospitals that were health centers and central hospitals. It not referring patients for DOTS care. Some has also been replicated and recognized poorer provinces and counties had as a "model" community-based service difficulty covering costs to participate. Both throughout Asia. challenges are addressed in a follow-up project. Impact of The under-five mortality rate was reduced Most people living with HIV/AIDS More than 1 million patients were cured policies on by 40 percent in intervention areas. The experienced substantial weight gain between 1991 and 2000. DOTS areas expe- health approach was also expected to alleviate the after three months in the project. rienced a 35 percent reduction in tuberculo- outcomes burden of severe malaria cases on hospitals. Compared with the mid-1990s, fewer sis prevalence over 10 years, compared with families and people living with HIV/AIDS a 3 percent reduction in non­DOTS areas. asked that confidentiality be maintained. The case detection rate almost doubled between the first and fifth years of imple- mentation (from 14 to 26 per 100,000) in areas covered by the project, indicating sig- nificant demand among prevalent cases. Source: World Bank staff. 92 | The Millennium Development Goals for Health CHAPTER 6 Improving Service Delivery I was assigned to PuskesmasTanjungAru,East Kalimantan,five Health providers in both the public and private sectors, hours from the district capital by boat through river and sea. It and in both the formal and informal sectors, play a key was staffed by two nurses,one midwife,one immunization officer, role in delivering interventions of relevance to the and one sanitation overseer. Unexpected problems confronted me Millennium Development Goals. Many are efficient, on arrival.The 52-year-old senior nurse tried to persuade the deliver high-quality care, and are responsive to their other staff that I was unable to run the health center and that I patients. But many are not.As a result, resources--public would rigidly control their day-to-day activities. He might have and private--are wasted, and facilities sit underused. been correct. I was not at all confident about my management Patients often receive care that is inappropriate to their ability. Later, I learned he was worried my presence might disturb needs, paying for it out of very limited means.They may his private practice, which he had developed since the last doctor also receive care that is downright dangerous. left six years before.Nobody knew about the budget except him. Two things can make a difference.One is the quality of The health center's workload and achievements were baf- management. Better management means a clearer delin- fling.There were only three to five patients a day.The latest eation of responsibilities and accountabilities inside organi- data, aside from the registers, were from two years before. zations, a clearer link between performance and reward, Guidelines could not be found.They were taken by the last and so on.The first part of the chapter reviews these and doctor when he left. Four months later we received photocopies other elements of good management, providing examples of the guidelines from the District Health Office.They were of where better management has made a difference. not too helpful for managerial matters.There was no money left Something else can also make a difference. Manage- for operational tasks even though the year had three months to ment means getting accountabilities right within an go.The doctor's house was dirty and unfurnished.The previous organization. Equally important--if not more so--are doctor took the furniture when he left. A request for replace- accountabilities between the organization and the public. ment was sent to the District Health Office, but six years had These can be improved along one or both of two routes.1 passed with no response. Staff suggested I take a spare exami- The short route, leading directly from the patient to the nation bed from the health center. I did this because I could not provider, can be strengthened through a variety of afford to buy a bed.The staff lacked initiative. No one came to schemes, including vouchers, report cards, and citizen work before 10 a.m., and everyone left before 1 p.m.They said management groups.The long route,leading from the cit- that there was little to do. izen to the policymaker and then to the provider, can be Source:World Bank staff, Indonesia. strengthened at two points--the citizen-policymaker 93 relationship (making policymakers more responsive to cit- two relationships--the link between the patient and the izens) and the policymaker-provider relationship (making front-line provider and the link between the policymaker providers more responsive to policymakers). and the provider organization. The interested reader is The World Development Report 2004: Making Services referred to the World Development Report for ways of Work for Poor People1 provides a thorough discussion--for strengthening the citizen-policymaker link, a generic several sectors--of ways to strengthen both routes and rather than health-specific issue. both elements of the long route.The focus here is on just Another area of service delivery is only partially cov- one sector (health) and within it on services and service ered in this chapter. This is the delivery mode--where providers of special relevance to the Millennium different services should be delivered and by which med- Development Goals for health.The focus is also on just ical staff (see box 6.1 and chapter 7). · Ownership of the business or activity. Ownership refers not Who's who in the health sector-- to the ownership of the premises but to the ownership of and the challenges that face them any leftover funds (or debts) at the end of the year,after all Anyone attempting to improve service delivery has to costs have been paid.When clinics or hospitals are pub- wrestle with the fact that a variety of providers deliver care licly owned,leftover funds belong to the treasury or pub- of relevance to the Millennium Development Goals, often lic purse.When clinics or hospitals are privately owned, to the same person in the same illness episode.The array of the organization itself or a private person has a legal claim relevant providers is bewildering (box 6.2). to all leftover revenue. For a nonprofit entity, the profits The boundaries between the different sets of actors-- cannot be distributed outside the organization. especially between public and private--can seem confusing, · Ownership of premises. Public clinics are almost always because a provider may be, say, private on one dimension located in publicly owned buildings. Private clinics, and but public on another. even hospitals, however, often do not own their · Payment can be public or private, regardless of who premises but rather rent them.Sometimes private clinics provides the services. are even run in publicly owned buildings.This does not Box 6.1 The changing mix of cure and care: Who treats what--and where? Throughout the twentieth century, service institutions have evolve. Hospitals are being transformed into long-term care cen- responded--albeit slowly--to rapid changes in health technology. ters for the elderly,while more complex procedures are conducted Countries choose combinations of "delivery modes" based on in ambulatory clinics.Home-based nursing care is being revived. costs and international standards but also on country-specific char- Countries at similar levels of technology have opted for dif- acteristics, such as geographic and density constraints, transport ferent models with comparable success. Independent practition- and infrastructure capacity, existing health infrastructure inherited ers developed first in Western countries and have been the from previous technological innovations,labor market characteris- cornerstone of Western systems. The hegemony of hospitals in tics,training and orientation of providers,and so on.What is deliv- the Western world is no older than the twentieth century. In ered as inpatient treatment, outpatient hospital, health center or contrast, hospitals have played a much larger role in the provision home visits--and by whom--is far from standard across countries. of outpatient care in Eastern Europe and Central Asia, Latin Technological progress triggers modifications in the nature, America, Sri Lanka, andVietnam. In Africa health systems devel- type, and quantity of services required. Hospital tuberculosis treat- oped through hospitals and mobile clinics since the beginning of ment (the sanatorium) was replaced by outpatient clinical care the twentieth century, with primary health care emerging only thanks to antibiotics. Screening followed by treatment--DOTS in the 1980s. Different skill levels are also used for similar inter- (directly observed treatment, short-course)--were later standard- ventions. Health technicians and nurses have performed cae- ized to allow delivery through community outreach. Similarly, sarean sections in Mozambique, while other countries use new treatments for HIV and cancer cut long hospitalization general practitioners or skilled obstetricians. requirements. The care and cure functions of the hospital also Source: Reference 1. 94 | The Millennium Development Goals for Health make the business, or operation, any less "private," since standing the organization of health services and the incen- the leftover revenue still remains with the organization tives associated with the organization. or its owners. The public sector · Employment. Health professionals can work in the Many of the problems associated with the public sector are public sector, the private sector, or both. Sometimes rooted in the more general structural problems related to this is legal. Sometimes it is not. Sometimes people the broader public sector: employed in the public sector "steal" the time they are supposed to devote to public employment and sell it · service commitments greatly in excess of allocated privately, in much the same way that they sometimes funds,undermining the accountability relations between steal drugs and sell them privately. It is not difficult to policymakers and service providers distinguish what is public and private in this case. · weak public sector accountability arrangements and lim- When employees work for a public organization and ited capacity for core administrative and policymaking steal time, they are stealing from a public employer. tasks When employees sell what they have stolen--be it drugs or time--it is a private transaction (they are sell- · managers who are judged not on performance but on ing something they own, even if they obtained it by "political merit" stealing). · managers who are appointed not on their technical merits but through cronyism By being clear about which aspect of health services deliv- ery is being discussed, it is easy to distinguish whether it is · budgets that are unpredictable and funds that leak out public or private. Such distinctions are critical in under- of the public sector Box 6.2 Who delivers which care for the Millennium Development Goals for health? The answer to the question "Who delivers which care for the healers (ayurvedic healers, traditional Chinese medicine health Millennium Development Goals?" is simple--a bewilder- healers, midwives). ing variety. · Nonprofit formal providers or nongovernmental organiza- Public providers include primary health clinics, health cen- tions often operate primary health care facilities as well as ters, and hospitals. Public providers of potential relevance for district--and sometimes referral--hospitals. They often achieving the Millennium Development Goals include rural and include organizations undertaking outreach activities, such urban clinics and health posts, as well as first-level referral hospi- as information, education, and communication programs tals. Public facilities are almost always arranged as a hierarchy of and social marketing. facilities--of decreasing sophistication as one gets farther into rural areas. Facilities are formally staffed by qualified medical · For-profit informal providers include traditional healers, staff, who are salaried civil service employees (and managed drug sellers, pharmacists (who offer informal diagnosis and largely from outside the facility). recommendations on medications to take), unqualified Services are often officially free or subsidized. Public funds practitioners of allopathic medicine, and traditional birth flow to facilities based on installed capacity and centrally estab- attendants. lished norms.Accountability for performance is through admin- All play a role in delivering services and interventions relevant to istrative oversight,usually by the Ministry of Health,less often by the Millennium Development Goals. And there is a striking local government bodies. Oversight typically focuses on appro- degree of heterogeneity within a country for the same type of priate use of funds and adherence to rules related to personnel. intervention or service. One woman might have her baby deliv- The largest facilities--referral hospitals--are often integrated ered by a traditional birth attendant. Another might have hers into the political realm, in that senior management is appointed delivered by a midwife at home. A third might have hers deliv- and dismissed following changes in government. ered by a doctor at a hospital. One child with diarrhea might be Private providers come in three broad guises: taken to a public facility, while another might be taken to a pri- · For-profit formal providers include registered self- vate provider. To complicate matters further, the same person employed doctors,clinics,hospitals,and diagnostic clinics,as may visit several types of provider within the same illness well as registered, organized, formally trained traditional episode--some in the public sector, some in the private. Source: World Bank staff. Improving Service Delivery | 95 · self-selection into the public sector of staff who prefer or services, and the desire for return customers can lead a work environment not based on performance for-profit providers to deliver unneeded or inappropriate care when patients want it.In developing countries,mutual · an inappropriate wage structure distrust, and often animosity, between the government and While important,these problems are generally best addressed the private health care sector is common, meaning that through interventions that cover the entire public sector. interaction with the public system is minimal. Even where Other problems can be addressed within the health common interests exist, the lack of mechanisms or forums ministry or sector.These include: for interaction and coordination makes working together very difficult. · low wages The informal private sector can also be weakened by · leakages of funds from the ministry of health to facilities various underlying structural issues. In developing coun- tries,regulation usually has very little influence on informal · weak external accountability arrangements for service private providers.They tend to be weakly organized, mak- delivery tasks ing it hard for policymakers to reach them with any policy · a lack of management skills and training on the part of intervention, and there is little or no professional support health facility management; managers who do not for quality improvement. Informal providers are almost manage staff (civil service) and who cannot therefore always organized as solo practices, a setting known to be held accountable or make required cost savings or undermine the quality of care.They are also isolated from quality-enhancing changes the public sector and the formal private sector,further con- · significantly higher costs to deliver services in rural areas straining coordination and the quality of care. Animosity (people educated in cities prefer not to live in rural areas) from the formal allopathic medical profession keeps them isolated and excluded from most policy dialogue. · unclear objectives and responsibilities of provider organizations Better management--improving · the fact that the services produced are valued by indi- accountability within provider organizations viduals and can be sold privately, which, coupled with weak management, leads to dual practice and absen- The Malaysian state of Sarawak and the Indonesian province teeism of West Kalimantan face one another across the Malaysia- Indonesia border.They share language,ethnicity,and religious · the perception that the work (such as treating people characteristics (box 6.3).Yet the approach to management in with HIV/AIDS) is dangerous their public health facilities is markedly different. Sarawak · general skills that are valued elsewhere,leading to emi- health workers are highly committed to quality and client sat- gration and loss of staff to the private sector isfaction.They work in teams in a cooperative fashion, and they are encouraged to set priorities in the light of local cir- The private sector cumstances,to take initiative,and to be innovative.Their atti- Various structural features create service delivery problems tude reflects a deliberate policy in Malaysia to encourage local for the nonprofit sector as well. Its dependence on dona- initiative,right down to the individual staff member.The aim tions and other inputs tied to philanthropy or altruism is to instill, through training and the use of health systems limits its opportunity to expand. Many nonprofits operat- research as a management tool, strong problem-solving and ing in developing countries rely extensively on interna- decisionmaking skills in management teams--at the facility, tional nonprofits for funding, making their funding district,and provincial levels.Initiative is rewarded:each year a unpredictable. Some nonprofits undertake religious or health worker is chosen to receive a prestigious national political activities or are closely associated with organiza- award recognizing personal initiative. tions that do.This sometimes contributes to a lack of trust Across the border in West Kalimantan, things work dif- by governments and the population, hampering their ferently. The Indonesian government imposes detailed activities and growth prospects.2 The absence or weakness guidelines, activity schedules, and workload norms on staff, of mechanisms or forums for regular communication and leaving no room for local decisionmaking and initiative. coordination with other nonprofits and with policymakers These two sets of facilities provide examples of two reduces the efficiency of nonprofits. entirely different perspectives on management. The The formal for-profit sector also exhibits certain struc- Indonesian approach has its roots in the hierarchical tural features that can cause problems. The bottom-line command-and-control model.Initiative and decisionmaking focus discourages the sector from cross-subsidizing patients are exercised only at the highest level.Problems encountered 96 | The Millennium Development Goals for Health Box 6.3 Different management styles, different countries--but just miles apart In March 1999 a team that included Indonesian and Malaysian villages with low health status, a key step in Sarawak's initiative to experts and an expatriate medical anthropologist spent a month decentralize priority-setting to clinics. visiting health facilities and assessing service delivery in Sarawak (Malaysia) and the adjacent Kalbar Province of West Kalimantan Quality issues (Indonesia), areas that share some language, ethnicity, and reli- Quality-related slogans and pictures appear on office walls in gious characteristics. Both Indonesia and Malaysia are commit- facilities throughout Sarawak.Almost everyone,from the director ted to the Health for All vision. But discussions with staff, in Kuching (the state capital) to medical assistants at Klinik Desa, patients, community leaders, and others revealed important dif- talks about quality. Quality is their culture, influencing the way ferences in approach--differences that probably account for they think and deliver services. Starting with top managers, much of the strikingly different results observed in these adjacent everyone in the system is trained in quality.This contrasts starkly provinces. For example, the estimated infant mortality rate in with the situation in Indonesia, where efforts to improve quality Sarawak, 9 deaths per 1,000 live births, is roughly a fifth of the started at the grass roots without preparing managers. The rate thought to prevail in Kalbar. Sarawak Health Department's vision, mission, and client charter are displayed to staff and the public. Every service and support Facilities unit must develop and exhibit its client charter. In Sarawak clinics are spacious and clean, and they provide a healthy environment in a well-maintained setting.Local attendants Competence of staff are hired to maintain each facility. Every clinic has a delivery Medical assistants, who perform the same tasks as nurses in room. Clean water is available, and the latrines and furniture are Indonesia's Puskesmas, were trained to diagnose and treat certain superior to those in Indonesia. In Kalbar the team encountered diseases. Their training--senior high school plus three years of filthy unused rooms with equipment in disrepair and patients who nursing school--is equivalent to Akademi Perawat in Indonesia. needed to be referred to other facilities.But it also found a pristine New staff also receive training on quality assurance and corporate clinic with a healing herb garden,patients happy with services,and culture before they are deployed. Staff receive briefs about field staff whose only complaint was that they weren't paid enough. activities and the obstacles they may encounter. Motivational and corporate culture courses are held annually.The Sarawak Health Uniformity or innovation Department also provides staff with technical guidelines.All staff Conditions and services of the same level in Sarawak are stan- concerned use the guidelines, unlike Kalbar, where most staff said dardized and predictable regardless of where facilities are located. they had never read or referred to the guidelines. Sarawak's tech- Only three staff are posted in each rural clinic, with consistency nical and administrative guidelines are readable, clear, well- of effort and skill maintained.The health care provided may not structured, and comprehensive. And implementation of the stray from that guaranteed by the system.The medical assistant is guidelines is closely monitored through quarterly quality checks. permitted to follow up a complex case only after the doctor Staff performance,linked to salary increments,is assessed yearly. receiving the referral has established the treatment regimen. But the state health department encourages personnel to be innova- Teamwork and management tive. A prestigious award goes every year to staff who develop new ways to accomplish their tasks. Teamwork is demonstrated at all levels in Sarawak.Two key ele- ments are dedication to quality and customer value and an environ- Program activities ment of cooperation with rewards for the success of teams rather than individuals. An informal and professional atmosphere exists, Sarawak's clinics, unlike Kalbar's, provide only outpatient and with staff throughout the state having met each other or even mother and child health services. Health education is integrated worked together. In Kalbar interpersonal and interorganizational into every activity. Outreach activities are done by village health communication is poor, community acceptance of services is lim- teams and the Flying Doctor Service, each based in the divisional ited,and effort and imagination in training personnel is minimal. health office.The only outreach handled by clinics is the updating of the village health survey every two years.This census identifies Source: Reference 6. at lower levels are passed up to higher levels for a decision. The Malaysian approach, by contrast, has more in com- There are no managers in the true sense of the word. mon with the new public sector management philosophy.3,4 Instead, administrators execute decisions according to previ- Responsibility for tasks and decisionmaking is delegated to ously agreed protocols and rules, with little or no scope for specific parts of the organization and to specific individuals. autonomous decisionmaking at the facility level. Individual accountability is emphasized, and there is a focus Improving Service Delivery |97 on performance--not inputs and processes but outputs and cantly increased provider compliance with clinical stan- outcomes.5 Good performance is rewarded, financially or in dards for maternal health, even without additional some other way.There is a focus on clients and a belief that resources to implement the initiative.12 In Ghana and an organization is ultimately accountable to its clients. A Jamaica the introduction of audits as a monitoring and client-oriented strategy emphasizes customer choice and sat- nonpunitive educational tool was associated with better isfaction. Business techniques enhance performance and are processes for obstetric care (more use of protocols), with a standard part of strategic planning. feedback identified as crucial to successes.13 In South Elements of the philosophy are evident in successful Africa an adolescent services program links achievement nutrition and child health programs. In Tamil Nadu's of national standards and criteria to accreditation of public Integrated Nutrition Program (box 6.4), community sector "adolescent-friendly" clinics. Using a combination nutrition workers were given clearly defined duties. of quality improvement methods, external assessment, and Information on outputs enabled the community to keep a rating system, the program has been integrated into a workers accountable, and it enabled the workers to see broader sexual health program, with promising results.14,15 how their programs were working.In Céara's Programa de Agentes de Saúde (box 6.5), credited with a substantial Institutional reform--strengthening the reduction in child mortality, health agents and nurse- accountability of provider organizations supervisors were assigned clear tasks and given clear responsibilities,and the intended outcomes of the program Public providers were emphasized throughout to both health workers and Attempts to improve the performance of public providers-- members of the public. Good performance by a team most often focusing on productivity and quality--are the (high immunization rates) was rewarded with a prize.And most common service delivery reforms in developing health agents were held accountable for their performance countries.The mechanisms applied in the health sector typ- through a community-based monitoring process. ically seek to alter institutional arrangements to create better World Health Organization guidelines in developing incentives for public providers--either by increasing the national Integrated Management of Childhood Illness leverage of patients (the short route of accountability) or by (IMCI) programs stress clear definitions of the roles of improving the effectiveness of supervision (part of the long central and district levels, strengthening existing manage- route). Mechanisms to increase patient leverage and super- ment structures, involving staff with good clinical skills in visory effectiveness often rely on financial incentives by supervision, and using management information to solve problems and improve planning.7 Performance feedback, seeking to tie payments to provider organizations (not nec- essarily individual providers) more closely to performance. which is also emphasized, has been found in Niger to have Some initiatives also seek to reduce organizational con- a "significant and consistent" impact on provider perfor- mance among those with previously low compliance.8 straints associated with public sector ownership that block responses to existing incentives. Some are in the realm of In maternal health, too, successful programs include ele- public sector reform more broadly (civil service reform, ments of the new public sector management philosophy. budget management and execution reform, improved China has a multilevel management strategy to achieve human resource policies) and are not reviewed here. maternal health goals. Management plans for maternal and reproductive health care are developed at the lower levels, PERFORMANCE-BASED PAYMENT (INTERNAL and those managers are rated on specific tasks, which also CONTRACTS) Public sector organizations in develop- serve as a basis for promotion. Greater autonomy and ing countries are typically neither recognized nor accountability through this system have helped promote rewarded for better performance. So performance-based preventive measures and increase referral rates among payment is a common strategy to improve performance. It high-risk pregnant women.9 consists of formal agreements between a government In other settings, audits and evidence-based standards have supervisory agency and public providers linking remuner- improved the quality of obstetric care. In Indonesia the gov- ation to specific aspects of performance. This funding ernment committed itself to systematized,district-level audits arrangement for public services, widespread in industrial to monitor performance and ensure quality of care. Malaysia countries, is now more common in middle-income coun- intensifies management attention to underperforming dis- tries. The agreements typically focus on increasing pro- tricts identified through such procedures, revising protocols ductivity (and sometimes quality) by clarifying related for personnel and facilities on the basis of data it collects.10,11 organizational objectives and responsibilities, thereby Performance-oriented measures have been shown to be strengthening policymakers' accountability relationship effective. In Ecuador quality assurance measures signifi- with the provider(s). When applied to primary care, the 98 | The Millennium Development Goals for Health Box 6.4 Management in India's Tamil Nadu Integrated Nutrition Program TheTamil Nadu Integrated Nutrition Program,started in 1980,was Training and supervision converted to the Integrated Child Development Services Program The community nutrition workers received training and super- in 1997. Core features of the program were growth monitoring, vision support in two main areas: nutrition education, and food supplementation programs, imple- mented in community nutrition centers by community nutrition · Preservice training. Community nutrition workers and com- workers.The program was successful in several respects. It reduced munity nutrition supervisors received training that lasted severe malnutrition in the program areas. It has been sustained for a two months, a significant amount of time for this kind of long period.And it operates on a large scale. Reductions in moder- program.Training groups were kept small, enabling a strong ate malnutrition were lower than expected,but the results have been emphasis on role-playing exercises. attributed to unrealistic targets rather than the level of achievement. · Supervision and in-service training. The high supervisor-to- How did the management of human resources contribute to worker ratio (one community nutrition supervisor super- the positive results of the program? Changes in staffing and job vised 10­15 community nutrition workers) meant that design,training and supervision,and monitoring made a difference. community nutrition supervisors had time to adequately supervise the community nutrition workers. In addition to Staffing and job design routine supervision, joint home visits were made to families The program implemented well-designed procedures for staffing who failed to bring their children to weighing and feeding and job design: or whose children did not gain weight after 90 days of food supplementation. The community nutrition supervisors · Recruitment. Individuals were eligible to become commu- were in turn supervised by community nutrition instructors. nity nutrition workers if they resided in the village, had at least eight years of schooling, and were acceptable to the Monitoring community. Special efforts were made to recruit women who were both poor and had healthy and well-nourished The monitoring system generated timely and good-quality data, children. The rationale for this approach was that there although too much information was sometimes collected. would be no social barrier between the community nutri- Monitoring information was used for two purposes: tion workers and their poor clients and that they would be · Information for workers and clients. Special efforts were made credible, since they had already managed to raise well- to collect data not only for managers but for frontline nourished children despite their low socioeconomic status. workers and their clients in the communities as well. Every · Duties and work routines.To help community nutrition work- month data collected by the community nutrition workers ers focus on priority issues, their job description included a were displayed outside the community nutrition centers. limited number of clearly defined duties, to be performed The data included information on the number of children according to a specific work schedule.Village registration of who had been weighed, the number of children of different women and children was to be conducted once every three nutritional status, and the number of children who received months. Growth monitoring took place during three days a supplementary feeding. These data provided community month. The supplementary feeding program was imple- members with an idea of how the program was doing. mented between set hours in the morning. Home visits · Information for management. Community nutrition supervi- were conducted in the afternoon. sors and instructors collated information from the commu- · Rewards. The community nutrition workers received a low nity nutrition workers during monthly meetings.The data monthly wage, but their compensation was still higher than were sent to the project coordination office, to be con- they would have earned as agricultural workers. Community verted into key performance indicators and used for overall nutrition workers were also motivated by factors other than management of the program. income: the satisfaction of helping others and the higher sta- tus that came from being associated with a program that improved the lives of their peers in the community. Source: Reference 16. payment often flows to providers based on patient choice. sides and the fact that at least part of the funding flows are In these cases, the strategy also relies on increasing the tied to some measurable aspect of performance. leverage of patients to improve accountability.Another key element of the improved incentives is the enhanced Does it work? Evidence on impact in developing enforceability of the agreement--that is, the ability to ver- countries is mixed. Performance-based payment appears to ify when responsibilities have been fulfilled or not on both work fairly well in middle-income countries and less well Improving Service Delivery | 99 Box 6.5 Innovative and effective management of public health workers in Céara, Brazil Although it is one of the poorest states in Brazil, the northeastern relatives and friends in return for political loyalty. The new state of Céara has achieved impressive improvements in child health. Brazilian constitution of 1988 increased mayors' access to rev- The infant mortality rate fell from about 100 per 1,000 live births in enues for health expenditures, as it increased federal transfers to the mid-1980s to 65 in 1992,a 36 percent reduction.A rural preven- the municipalities.The planners of the PAS had to work within tive health program, Programa de Agentes de Saúde (health agent this context of decentralization.They feared that hiring a large program,PAS),has been credited with much of the improvement. number of health agents and nurse-supervisors would be vulner- A central feature of the program was the use of health agents, able to clientelism.The state addressed the problem in three main who visited households to provide advice and assistance on oral ways. First, it hired workers based on merit and offered them rehydration therapy, vaccination, antenatal care, breastfeeding, temporary contracts, not job tenure. Second, the health agents' and growth monitoring.The agents also collected data for health salaries were paid directly from the governor's office.Third, the monitoring purposes. By 1993 the health agents were making state health department appropriated the responsibility for hiring monthly home visits to 850,000 families--roughly 65 percent of the health agents, while leaving the hiring of nurses to supervise the state's population. Even though health agents did not earn the health agents to each municipality. more than the minimum wage and worked under temporary contracts without job security or fringe benefits, they performed Hiring well and helped achieve important health gains. The state-level coordinating committee placed strong emphasis How did Céara achieve this? Effective public management of on hiring people based on merit. It required all applicants to sub- human resources played a large role, in a variety of areas, includ- mit written applications, from which it selected candidates to ing addressing concerns about clientelism, the hiring process, interview. Two members of the selection team (usually a nurse motivation, supervision, and community monitoring. and a social worker) then traveled to each town for interviews, followed by a group meeting.The group meeting was often fol- Addressing concerns about clientelism lowed by a second round of individual interviews with candidates Before PAS was initiated in 1987, most municipalities had no likely to be selected.The visits of the hiring team created a sense public health program.At most, the mayor had an ambulance at of excitement in the towns and added to the prestige of being his disposal and kept a small stock of prescription medicines at selected. Although the jobs paid only the minimum wages and his home. Medicines and ambulance rides were usually given to came with no fringe benefits, many people found the position in low-income countries.18­23 In both settings the evidence of inpatient and ambulatory services.27 In Central Europe, is relatively strong that the strategy leads to increases in out- where capped, case-based payments cover inpatient ser- put, access, productivity, and responsiveness.18­24 Quanti- vices, increases in the volume of services and declining fiable positive impact on health care quality is less easy to average lengths of stay--in Bulgaria, the Czech Republic, identify,partly because it is so rarely measured. Estonia, Hungary, Latvia, Lithuania, Poland, Romania, and Performance-based payment appears easier to apply for the Russian Federation--suggest increases in productiv- primary care services than for acute care, probably because ity.18* Deterioration in quality (due to skimping on costly it is more feasible to tie the payment allocation to patient nonobservable aspects of quality) has not arisen as a serious choice of provider. (It is widely accepted that patients can concern. Elsewhere in the former Soviet republics, and in judge the quality of primary care services, and hence Latin America, the results are more ambiguous.This lack of providers, better than they can judge hospitals and hospital response is typically attributed to insufficient provider service quality.) autonomy and a weak link between performance and pay There is more evidence from middle-income countries, (insufficient funds relative to service commitments). especially from Latin America and Central and Eastern When does it work best? Key influences on the Europe. In both regions there is stronger evidence of posi- success of the approach include whether the public tive impact for performance-based payments at the pri- provider has the ability to respond (see the discussion of mary care level than at higher levels of care. Such evidence comes from Croatia, Estonia, Latvia, Lithuania, Peru, Poland, and Romania.18,23,25,26 In Argentina and Nicaragua * Since other systemic changes have been implemented along with social security institutes have increased productivity by performance-based payment in Central Europe,it is not possible to attribute establishing capitation payments for an integrated package the results to any single aspect of the reforms (such as the payment mechanism). 100 | The Millennium Development Goals for Health desirable, particularly as it was a year-round rather than a seasonal they had been paid in their previous jobs.These factors helped job.The newly hired workers began their jobs with a strong sense the state ensure good supervision of its health agents. of prestige for having been selected. Community monitoring Motivation In addition to effective supervision by nurses, the community The positive effects of the hiring process were enhanced during made sure that health agents performed well. During the hiring the implementation period.The program received much publi- process, the program linked prestige not just to those who passed city, in large part thanks to funding from private firms for radio the rigorous selection process but to the whole program and its and television campaigns. Municipalities that achieved the high- mission of reducing infant mortality and morbidity. A special est immunization coverage were awarded prizes, further adding message was given to applicants who were not chosen for the to the prestige of agents in the communities.The program pro- job.They were encouraged to make sure that the health agents vided three months of full-time training to the health agents, a followed the rules of the program.The rules, constantly repeated, strong motivating factor. Access to this kind of training was were that the health agents had to live in the area in which they unimaginable to most people in the interior of Céara. Health worked, work eight hours a day, visit each household at least agents were also motivated by the fact that their jobs were more once a month, and attend all training and review sessions. diverse and satisfying than most other jobs in the area. To avoid turning disgruntled job-seekers into overzealous watchdogs, the selection committee conveyed the message to the Supervision applicants in a way that made them feel involved in the program. Indeed, community members reported to the nurse-supervisors Supervision of the health agents by nurses was a critical part to when they felt satisfied about the job a particular agent was the program's success. In urban clinics and hospitals, where many doing, not only when they perceived that the agent had done of the nurse-supervisors had previously worked, they had a sub- something wrong.This socialization of all job applicants to the ordinate status and performed much more administrative work program's mission created an informal but powerful monitoring rather than actual nursing duties.They often felt alienated from presence in the communities, adding to a sense of collective their work and ignored as professionals. In the program in Céara, responsibility for the program. their status changed dramatically. Each nurse supervised and trained 30 health agents and was considered a very important person in the community.The nurses were also paid more than Source: Reference 17. autonomization below),whether service commitments are AUTONOMIZATION The weak impact of performance- congruent with funding levels, whether output and key based payment strategies in public hospitals is frequently components of performance expectations are easily mea- attributed to the lack of decisionmaking authority and flex- surable (as in primary healthcare), and how far capacity ibility to respond to incentives.That is why another strat- strengthening of the payer or funder is addressed as a cen- egy--autonomization--is used, often in conjunction with tral part of the initiative. or following those just described. Autonomization consists Other factors are important, too. Positive impacts are of delegating additional decisionmaking authority to hospi- more likely when the potential bonuses to be gained by tals. It aims to improve productivity by giving providers the better performance are sufficient to justify substantial flexibility to respond to performance incentives, especially effort by the provider.And responses are greater when the to manage labor and production, to pay staff based on per- provider organization can retain and allocate the extra rev- formance, and to undertake cost-saving or quality-enhanc- enues. Schemes in which this is not the case have little or ing changes.28 Autonomization also often seeks to create no impact. Can the patient choose the provider and in the some risk-bearing for the provider to motivate performance process direct funding flows? For hospitals it appears that (residual claimant status). In low-income countries autono- the threat to take away funding is usually perceived as hol- mization is most often driven by fiscal crises and a desire to low. So reforms are more successful when performance shift more care toward the primary level.There is usually a pressure relies not on pressure from competition for mar- focus on mobilizing resources in the autonomized facilities ket share but on the possibility of bonuses. from user fees. Autonomization does not work through improving the accountability relations for policymakers or Public providers rarely perceive the threat of not receiving a contract as patients--it seeks to increase the ability of providers to credible, even where alternative providers exist.28 respond to whatever incentives--created primarily by the Improving Service Delivery | 101 public payment system and the choices of private payers-- ber of visits per doctor and the number of visits per inhabi- exist. tant rose. Immunization coverage, which is paid for sepa- rately,rose from 74 percent to 88 percent.20 In Romania the Does it work? No rigorous evaluation of autonomiza- use of primary healthcare--a key goal of the initiative-- tion has been conducted in low-income settings.The assess- increased significantly.23 Efficiency also increased.And as in ments that have been done have produced mixed results.29 Estonia, significant increases in immunization rates were Efficiency and sometimes quality have improved when labor observed.In Croatia efficiency was not assessed in the priva- management is delegated--as they did in India and tized practices, but physicians were more accessible and Kenya.30,31 But the emphasis on revenue generation from fees patient satisfaction had increased.22 has undermined access for the poor in a number of countries, The reform appears to generate two sources of including Indonesia and Malaysia.32,33 Results in middle- accountability: pressure to please patients enough to main- income countries are better,with evidence of efficiency gains tain registered patients and requirements for pricing, pro- in Colombia34 and efficiency gains combined with quality fessional licensing, facility adequacy, and other staffing to improvements inTunisia.32The greater capacity of the funder maintain the contract with the insurer.But privatization in or payer in middle-income countries allows the use of pay- these three countries, and others in the region, has been ments as performance leverage for providers. associated with fragmentation (such as reduced practice When does it work best? A variety of factors influ- size), raising concerns about reduced efficiency and a pos- ence the success of autonomization.Are services for the poor sible decline in clinical quality. paid for or otherwise supported and monitored? Are agree- When does it work best? Privatization of health ser- ments about maintenance of important but money-losing vices in developing countries has been assessed only at the services explicit and monitored? Is there explicit support for primary healthcare level.It appears to work well when there central government and other administrative officials to shift is an established, solvent purchaser in place and capitation the nature of their oversight? The reform time horizon for payments are combined with patient choice.The tendency autonomization is long (three to five years). Management toward fragmentation may be reduced by contract processes capacity needs to be in place in the provider unit before the that promote group practice (price-setting) and by provi- reform is imitated--or at least strengthened as a central part sions to promote access to capital and premises. of reform.There has to be an explicit and focused effort to create accountability mechanisms that are not based on day- DECENTRALIZATION Decentralization is intended to to-day administrative control (oversight boards, regulation, improve the quality of services and the productivity of accreditation).The organization's mandate has to be narrow providers by strengthening the policymaker's accountabil- and clear.And output-based or performance-based funding ity relations with providers (the local government supervi- has to be in place before the provider reforms. sor is closer). Indirectly, decentralization may increase patient leverage, as patients are likely to be able to exert PRIVATIZATION Like autonomization for hospitals,pri- more pressure on local governments to improve health vatization has been applied at the primary level to give pub- services than they could on the central government, partly lic providers the freedom and the high-powered incentives because information on preferences and performance is that come from being the owner (residual claimant). easier to obtain at the local level. Privatization at the primary level converts a salaried publicly Does it work? The impact of decentralization has employed physician into a self-employed independent con- been mixed in low-income countries.The health services tractor,with a service contract with a public payer (such as a element of broad decentralization initiatives has been par- social insurance organization). The strategy was imple- ticularly poorly designed and implemented, often driven mented to different degrees in almost all Central European by political or fiscal considerations. In Tanzania decentral- countries following the establishment of social insurance ization improved the efficiency of and access to primary systems (95 percent of primary healthcare providers in the healthcare services, but in Ghana efficiency declined.35 Czech Republic and 80 percent in Estonia have been priva- There have been notable examples of deterioration in the tized). In Croatia, Estonia, and Romania the capitation pay- provision of public goods, such as disease surveillance and ment is allocated based on patients'choice of provider. preventative services (Ghana, Indonesia,Tanzania), and in Does it work? Privatization of primary healthcare, in the equity of access (Uganda). There is less evidence on conjunction with capitation-based payment and training in the impact of decentralization in middle-income coun- family medicine, is associated with positive results in three tries, though it has generated improvements in responsive- assessed cases in Central Europe.In Estonia the annual num- ness and equity in Colombia.36 102 | The Millennium Development Goals for Health When does it work best? Adequate resources to greater complexity of the services and organization of operate services have to be allocated along with new service hospitals appears to undermine the influence of commu- responsibilities. Local political institutions have to be rela- nity representatives in oversight processes. And strong tively functional (voters need to have influence--another technical and advisory support has to be provided to com- element of the long route of accountability).Health author- munity representative bodies. ities have to be integrally involved in designing and imple- ACCREDITATION OF HOSPITALS Accreditation ini- menting the decentralization initiatives for health services. tiatives are becoming more common in developing coun- Concerted efforts to build local government capacity-- tries. Under this strategy, public hospitals are motivated to planning, supervision, budgeting, and expenditure and improve quality standards as assessed by external reviewers financial management--have to be part of the initiative. in order to receive "accredited" status.This approach does Mechanisms also need to be in place to ensure capacity to not directly use either policymaker oversight or patient cover the core public functions outlined in chapter 8 (dis- leverage, though it often relies on those forces indirectly ease surveillance and control, funding for and attention to by linking accreditation to eligibility for public funds or preventive services and services that promote health,human by directing patients to accredited facilities. resource planning, interregional coordination). Adequate time has to be allocated to design and implementation, Does it work? Only two developing countries that especially for building the capacity of local authorities to have implemented accreditation for health services perform new oversight tasks. providers have been assessed.The accreditation initiative in South Africa has improved process quality, though there is GOVERNANCE PARTICIPATION Participation is no evidence on the impact on outcomes.39 In Zambia intended to improve productivity and the quality of ser- there has been no demonstrable impact.40 vices by involving representatives of community interests in the governance of local facilities. It relies on directly When does it work best? Accreditation works best strengthening the leverage of patients (the short route of when it is linked to financial leverage--when it adds to accountability), or at least their representatives, in provider the provider's "bottom line." It is most successful when oversight. Patients' first-hand knowledge and influence provider associations are productively engaged and shorten the "feedback" loop of supervisory accountability. involved in developing standards. Substantial facilitation or consulting services are required to help hospitals imple- Does it work? Some governments have had difficulty ment needed improvements.41,42 establishing meaningful participation, but there have been positive changes where it has been established. In Burkino Nonprofit providers Faso participation of community representatives in public When governments work with nonprofit providers to primary healthcare clinics increased immunization coverage achieve sector objectives, their efforts usually focus on and the availability of essential drugs.37 The percentage of expanding services or increasing efficiency through better women with two or more antenatal visits also increased.But coordination. Several approaches are commonly used, the resources spent increased as well, so it is not possible to including informal arrangements for in-kind subsidies and assess the impact on efficiency. In Peru participation is asso- public-private partnerships. But the only strategy that evi- ciated with reduced absenteeism of staff, reduced waiting dence suggests may work is service contracting. times,and the perception of higher quality by patients.26,38 Contracting initiatives for health and nutrition services When does it work best? No formal evaluation of have been implemented in several countries. Most initiatives participation has been conducted, but several characteris- work through nonprofit organizations seeking to build on tics seem to contribute to success. Better results are their comparative advantage (efficiency, quality, willingness achieved where the selection process for representatives is to cross-subsidize) and focus (slum dwellers, malnourished transparent and involves community members rather than children, poor socially excluded groups). Implicitly, the ini- appointed representatives. Governance participation is tiatives also build on the close alignment of the providers' more likely to improve service provision where reasonably organizational objectives and the policymakers' goals. This strong community organizations already exist and moti- alignment is believed to reduce the need for oversight and vated community members are present. Supervisory general capacity of the government in interacting with responsibilities need to be well defined and focused (as in them.Such initiatives are rarely rigorously evaluated,because the allocation of revolving drug funds in the Bamako ini- little emphasis is placed on collecting information or moni- tiative and elsewhere). Governance participation probably toring providers,whom policymakers tend to trust.This lack works better for primary care than for hospitals--the of rigorous evaluation makes evaluating the impact difficult. Improving Service Delivery |103 HEALTH SERVICE CONTRACTING Government, it advantage than they would by providing services directly. is argued, gets more value for money from public spend- Services for which governments commonly contract ing by paying a nonprofit provider to provide priority include outreach for health promotion or education in services, such as basic services for maternal and child urban slums, outreach to stigmatized or hard-to-reach health, and communicable diseases, or by serving priority groups, and social marketing of priority health goods or population groups in which they have a comparative services (box 6.6). Box 6.6 Contracting health services in Cambodia, Guatemala, and Pakistan Cambodia NGOs considerable autonomy to run existing clinics (similar to contracting out in Cambodia); a mixed model, which required the In Cambodia the government contracted with nongovernmental NGOs to keep the same staff and so gave them less autonomy organizations (NGOs) in two ways.43 The first was a contracting- (closely resembling contracting in);and the traditional model,which out model, in which contractors had responsibility for delivering was essentially the usual way the government delivered services. specified services, directly employed their own staff, and had full Municipalities covering some 190,000 people were assigned management control.The second was a contracting-in model, in to one of the approaches, and a household survey was under- which contractors managed the district health care system within taken after about two years. No baseline survey was available. the Ministry of Health.The government met recurrent operating The household survey found that the mixed approach costs through normal channels, although a small supplement was achieved the best results in antenatal care, immunization, and provided,which the contractors could spend as they saw best.In a receipt of oral rehydration salts by young children with diarrhea. control group,services were delivered under the existing Ministry But the areas in which the direct model was implemented had of Health system and the same small supplement was provided. more isolated households with considerably less access to health Twelve districts, covering about 1.5 million people, were ran- facilities.The traditional and mixed samples appeared quite similar. domly assigned to one of the three groups (contracting-out, The mixed model appears more successful than the tradi- contracting-in,or control),and baseline household and health facil- tional model, with a difference in coverage of services of 5­16 ity surveys were administered before the contracts started.The sur- percentage points.The absence of a baseline makes it difficult to veys were repeated two and a half years after the contracts began. be sure that this is the true size of the effect, but it appears that What were the results? Much larger improvements in immu- contracting with NGOs had a real impact. It is also difficult to nization coverage,antenatal care,and other indicators were observed conclude much about the direct approach because it was imple- in the contracting-out and contracting-in districts than in the con- mented in more remote and difficult areas. trol districts, although all districts were quite similar at the outset. The increases were especially pronounced in the contracting-out Pakistan districts. The poor appear to have benefited disproportionately. Vitamin A supplementation increased faster among the poorest half In a poorly performing district of Punjab, an NGO was allowed to of the population.And treatment of illness among the poorest half of run all the primary health care facilities and implement changes in the population increased several times faster in contracted districts organization and management.The NGO was given the same bud- than in the control districts.The results show that contracting with get previously provided to the district.The NGO nearly tripled the NGOs can improve service delivery in a short period. salaries of selected doctors and had them cover three different basic What about cost? Cost was considerably higher for contract- health units instead of the usual one.The NGO also improved the ing out than for contracting in or the control groups, but con- supply of drugs available without increasing the budget. tracting out led to a considerable savings in out-of-pocket An interrupted time series was used to assess the interven- expenditures for people in the communities.The cost difference tion. Information from the routine reporting system on the between the contracting-in and the control districts ($0.96 per number of outpatient visits in the district was tracked over time. capita per year) was smaller, reflecting almost entirely the cost of No data from other nearby districts are available yet.The district the contract with the NGO. Because the two groups thus had comprises 104 basic health units for about 2 million people. the same amount of resources to spend on actual service delivery, How were patients affected? Once the NGO started running extra resources do not explain the performance difference. the system, outpatient visits to the basic health units enjoyed Overall, contracting was considerably more successful than gov- nearly a fourfold increase. ernment delivery of the same services. The intervention took place in one medium-size district. No household survey data are available, nor are data available on any other aspect of health service delivery,such as immunization.But Guatemala the sudden dramatic increase in outpatient visits suggests that the The government of Guatemala contracted with NGOs to deliver NGO was more successful than the government in operating the primary healthcare services to indigenous populations in mountain- services with managerial autonomy. ous areas using three models: a direct model, which involved giving Source: World Bank staff and references 43­45. 104 | The Millennium Development Goals for Health Does it work? Contracting with NGOs is used most lacks capacity and equipment, such as diagnostic services often in low-income countries.In most cases that have been and high-tech services. evaluated,positive impacts were observed on target outcome Does it work? The results are mixed. Contracting or output variables. In Bangladesh, for example, contracts with formal for-profit providers is frequently effective in with nonprofit providers for the planning and implementa- increasing access, use, and responsiveness. In many cases it tion of an expanded immunization program were credited appears that for-profit providers can also produce contracted with increasing coverage from 25 percent in 1985 to 80 per- cent in 1990.46 Contracting for a primary healthcare services at lower cost.What is not clear is whether the gov- ernment captures any of the efficiency gains or whether the package in Haiti also generated significant increases in immunization coverage.47 Nutrition services also appear to cost savings come at the expense of clinical quality or access by high-cost patients. Efficiency increased through be amenable to contracting in low-income countries. In contracting for high-tech diagnostic services in Thailand.50 Bangladesh, Madagascar, and Senegal, significant reductions in nutrition rates were attributed to contracting initiatives.48 Contracts for high-tech services in India improved some aspects of quality (availability of services, frequency of func- Only a small number of cases assess efficiency. One tioning equipment),though private payment for the services found that the contracted provider (Prosalud) in Santa almost certainly constrained access for poor patients. In Cruz, Bolivia, outperformed the Ministry of Health facili- Zimbabwe inpatient and outpatient services were tendered, ties on productivity, use, efficacy, and cost-effectiveness in outpatient care.49 Many social marketing initiatives rely on and the cost per service decreased.51 But the lack of volume control led to an increase in total cost. In South Africa the contracted nonprofit organizations to market bednets, oral cost of hospital services declined, but the cost to govern- rehydration salts, fortified foods, and the like. ment did not.52 When does it work best? Nonprofit providers need A weak contracting capacity on the part of the govern- to be strong, with well-functioning accountability arrange- ment often allows the provider to capture any efficiency ments and internal motivation to perform--as those in gains or to expand volume to generate more income. In Bangladesh and Bolivia are. The government needs to be Brazil contracting with for-profit hospitals led to increases capable of assessing, selecting, and managing the ongoing in access.53 But it also led to cream-skimming to avoid relationship with the providers. It must also fulfill its end of costly patients, as well as fraud (false billing for services). the deal (contractual agreements on funding are genuine) Contracting with formal for-profit providers for pri- and not interfere with the running of the nonprofit mary healthcare services appears to succeed more often. In providers,which need full control of the operation (box 6.6). El Salvador and Peru, social insurance institutes contracted with private primary healthcare providers to serve their For-profit formal providers clients, which increased access, choice, and consumer satis- The past 10 years have seen greater efforts to harness the faction.54 Contracting with (privatized) self-employed formal for-profit sector to improve the quality of care it physicians in Croatia, Estonia, and Romania has improved provides, to use it to serve certain populations in exchange efficiency, quality, and use.20,22,23 for public funding, and to reduce the cost to poor people When does it work best? Investment in developing for for-profit services. capacities for managing the contracting process clearly HEALTH SERVICE CONTRACTING Contracting in needs to be a central part of such initiatives.55 And quality developing countries is most often used to try to mobilize in the private sector has to be at least as good as in the for-profit providers, usually to increase access or effi- public sector (since ability to monitor quality is usually ciency. In exchange for public funds, private providers low). Such contracting probably works best for primary deliver services of a specified type and quality, in agreed care or other relatively observable services (diagnosis). quantities, to agreed recipients, and over agreed periods. ENHANCING THE REGULATORY FRAMEWORK Like performance-based payment of public providers, TO IMPROVE QUALITY Given widespread concern contracting strengthens accountability relations between with the quality of health care services in developing coun- the policymaker and the provider and emphasizes finan- tries, there have been many attempts to improve the effec- cial incentives. But the for-profit orientation of these tiveness of the regulatory framework for health services and providers means that more rigorous tendering and con- tract management process are required to bring about Contracting in these cases was introduced at the same time as the conver- desired improvements. Contracting with for-profit sion from public to private practice and the introduction of the "family doc- providers often focuses on primary care services and on tor" model of practice. Hence it is not possible to attribute these results to procuring services in areas in which the public sector any one factor. Improving Service Delivery | 105 facilities.Strategies typically consist of enacting or updating health goods (condoms, oral rehydration salts, fortified laws or regulations related to minimum standards, espe- foods): conducting a marketing campaign for the branded cially for inputs and starting a profession or business.These good and commercializing the sale of the branded good efforts are sometimes accompanied by efforts to strengthen for target providers. Sometimes the strategy is applied to the capacity of the relevant regulatory agency. shift consumption from lower-quality goods to better ones. It uses the (existing) market-based accountability of Does it work? There is no documented case in a sellers to purchasers to influence retailers and producers of developing country of a successful initiative to improve the goods. It has been widely successful in increasing the use effectiveness of the regulatory framework for health ser- of important public health goods.60­62 vices broadly. Targeted regulatory initiatives (for pharma- More recently, it has been applied to health services, cies) have, however, improved targeted practices.56,57 The such as reproductive health services and treatment of sexu- role of professional and provider organizations in such ini- ally transmitted diseases (STDs). tiatives appears to have been very small--in marked con- trast to industrial countries. Does it work? In Uganda a social marketing- commercialization initiative improved the quality of STD When does it work best? Regulatory initiatives have treatment offered by formal for-profit providers in three to target a specific group of providers and a well-defined rural districts. Targeted retailers stocked and prescribed set of behaviors (overprescribing antibiotics, mixing phar- prepackaged STD drugs, increasing treatment compliance maceuticals). And they have to be accompanied by an from 87 percent to 93 percent, cure rates from 47 percent effective information dissemination strategy and enhanced to 84 percent, and condom use during treatment from 17 inspection (the Lao People's Democratic Republic),57 rein- percent to 36 percent.62 Another such initiative created a forced by education and peer influence (Vietnam)56 and by "branded" reproductive health service, which attracted the dissemination of information (on, say, the use of oral 11,000 providers. The use of the branded providers rehydration therapy rather than antimotility drugs for chil- increased significantly, providing contraceptive services to dren's diarrhea [Pakistan]).58 one Pakistani couple in five at a cost of less than $4 a year DISSEMINATING INFORMATION TO PATIENTS per person.63 As with health goods, social marketing and AND CAREGIVERS One of the biggest problems with commercialization of services is more sustainable than formal for-profit providers is their willingness to accede to contracting because it relies on patient payment and is patients' requests for inappropriate treatment, including implemented through commercial organizations. But the prescriptions. Some countries have tried to address this focus on cost recovery makes this strategy less effective in problem by improving patient knowledge of appropriate reaching the very poor. treatment and thus changing their demand in targeted When does it work best? Adequate attention needs areas. Easy-to-understand information is disseminated on to be paid to market segmentation to reduce "crowding appropriate treatment and care-seeking. The intention is out."§ Goods or services need to be homogeneous, so that that providers will then deliver better-quality care because consumers can easily distinguish the targeted goods, ser- the demand for inappropriate care declines. vices,or providers.Other provisions such as targeted subsi- Does it work? Some positive results have been dies and vouchers are needed to promote use by the poor. reported, though only from a small number of experi- Informal private providers ences. In Bangladesh a 10-year campaign to inform moth- ers about oral rehydration therapy led to a sharp increase Given the significance of informal private providers for peo- in the number of rural drug sellers prescribing it. ple in developing countries,especially the poor,it is unfortu- Physicians also increased their recommendation of oral nate (to say the least) that very few attempts have been made rehydration therapy, albeit by less.59 to improve the quality of their services. It is widely accepted that the quality of care they provide is low,but little has been When does it work best? A key issue is whether the done to improve it. In the few instances in which initiatives problem is grounded in inappropriate patient demand or have been taken to improve the quality of care of informal in financial benefit to the provider.The information pro- vided needs to be easy to understand. § Subsidizing consumption has no effect if buyers shift from commercial to SOCIAL MARKETING AND COMMERCIALIZA- subsidized sources of supply.This is the case when subsidized suppliers TION Social marketing and commercialization consists "crowd out" existing commercial suppliers, so that there is no overall of a two-part initiative to increase the use of important increase in purchase (or use) of the good. 106 | The Millennium Development Goals for Health providers, the focus has been mostly on training. Several of References these efforts have been successful (chapter 7). Some limited attempts have been made to regulate infor- 1. World Bank. 2003. World Development Report 2004: Making ServicesWork for Poor People.Washington, DC:World Bank. mal providers. One such initiative, which focused on 2. Dejong, J. 1991. Non-Governmental Organizations and Health improving the quality of services offered by drug vendors, Delivery in Sub-Saharan Africa.WPS 708.World Bank, was assessed in the Lao People's Democratic Republic.57 Washington, DC. Changes in regulations were reinforced by more frequent 3. Flanagan, H. 1996. Public Sector Management Effectiveness:Theory inspections,with feedback provided to vendors.Information and Practice in the National Health Service. Buckingham, United Kingdom: Open University Press. about the regulations was also disseminated.Results included 4. Mintzberg,H.1994.The Rise and Fall of Strategic Planning: greater availability of essential drugs,more appropriate infor- Reconceiving Roles for Planning,Plans,Planners.NewYork:Free Press. mation given to customers, and less mixing of drugs in the 5. Shaw,P.2002."New Public Sector Management in Health."HNP same package. Discussion Paper, ed.A.S.Preker.World Bank,Washington,DC. 6. World Bank. 2000."Health Strategy in a Post-Crisis, Decentralizing Indonesia." Report No. 21318-IND,World What works? What doesn't? Why? Bank,Washington, DC. Six key lessons emerge from the assessment of service 7. Lambrechts,T., J. Bryce, andV. Orinda. 1999."Integrated Management of Childhood Illness:A Summary of First delivery: Experiences." Bulletin of theWorld Health Organization 77 (7): 582­594. · The evidence on management and broader institu- 8. Kelley, E., C. Geslin, S. Djibrina, and M. Boucar. 2001. tional reforms is not strong. Debate is often guided "Improving Performance with Clinical Standards:The Impact more by political prejudice than by hard evidence. In of Feedback on Compliance with the Integrated Management part, this is because too little effort has been invested of Childhood Illness Algorithm in Niger,West Africa." International Journal of Health Planning and Management 16 (3): in strengthening the evidence base. 195­205. · Limited but encouraging evidence suggests that appli- 9. Koblinsky, M.A. 2003."Reducing Maternal Mortality: Learning from Bolivia, China, Egypt, Honduras, Indonesia, Jamaica, and cation of some of the key ingredients of the new public Zimbabwe."World Bank,Washington, DC. sector management philosophy seems to improve per- 10. Pathmanathan, I. 2003." Investing in Maternal Health in formance. Going down the Sarawak road (see intro- Malaysia and Sri Lanka."World Bank,Washington, DC. duction) can,in the right circumstances,lead to gains. 11. Ronsmans, C. 2001."What Is the Evidence for the Role of Audits to Improve the Quality of Obstetric Care?" In Safe · Institutional reforms may be most effective when Motherhood Strategies:A Review of the Evidence, ed.W.Van more than one approach is used--coupling autono- Leberghe, 203­223. London and Antwerp: Centre for Sexual and Reproductive Health; UON Network Unmet Need for mization with performance-based payment, for exam- Major Obstetric Interventions. ple, or regulation with information for providers or 12. Hermida,J.,and M.E.Robalino.2002."Increasing Compliance patients, or social marketing with commercialization. with Maternal and Child Care Quality Standards in Ecuador." International Journal for Quality in Health Care 14 (Suppl.1):25­34. · Efforts need to focus on the most relevant providers, 13. Wagaarachchi, P.T.,W.J. Graham, G.C. Penney,A. Mccaw-Binns, not the ones policymakers may prefer or the ones easi- K.Yeboah Antwi, and M.H. Hall. 2001."Holding Up a Mirror: est to reach.This implies a need to focus reform efforts Changing Obstetric Practice through Criterion-Based Clinical Audit in Developing Countries." International Journal of much more on informal providers,who continue to be Gynaecology and Obstetrics 74 (2): 119­130. a major supplier of services in the developing world. 14. Harrison, D. 2002."Lovelife Appears to Be an Effective Prevention Program for South AfricanYouth." Press release, · The effectiveness of contracting with for-profit providers American Medical Association, Barcelona, July 7. differs across types of services and organizations. Most 15. Dickson-Tetteh, K.,A. Pettifor, andW. Moleko. 2001."Working success has been with primary healthcare providers. with Public Sector Clinics to Provide Adolescent-Friendly Success has been limited with specific high-tech diagnos- Services in South Africa." Reproductive Health Matters 9 (17): tic services.And success has so far been very limited with 160­169. hospitals. 16. Heaver R. India'sTamil Nadu nutrition program: lessons and issues in management and capacity development. HNP · Despite the lack of broader regulatory initiatives in Discussion Paper.Washington, DC:World Bank, 2002. developing countries, there is no alternative to govern- 17. Tendler, J., and S. Freedheim.1994."Trust in a Rent-Seeking World: Health and GovernmentTransformed in Northeast ments wrestling more successfully with this critical role Brazil." World Development 22 (12): 1771­1791. in the health sector. A concerted effort to help devel- 18. Langenbrunner, J. 2003."Resource Allocation and Purchasing oping country governments in this effort is warranted. in ECA Region:A Review."World Bank,Washington, DC. Improving Service Delivery | 107 19. Naimoli, J.F. 2003."Performance-Based Management in an 37. Eichler, R. 2001."Improving Immunization Coverage in an Evolving Decentralized Public Health System inWest Africa: Innovative Primary Health Care Delivery Model: Lessons from The Case of Burkina Faso." Background paper prepared for the Burkina Faso's Bottom Up Planning, Oversight and Resource World Development Report 2004.World Bank,Washington, DC. Control Approach that Holds Providers Accountable for 20. Koppel,A., K. Meiesaar, H.Valtonen,A. Metsa, and M. Lember. Results."World Bank,Washington, DC. 2003."Evaluation of Primary Health Care Reform in Estonia." 38. Ewig, C. 2003."The Contributions of Community-Based Social Science and Medicine 56 (12): 2461­2466. Decentralization to Democracy: Peru's Local Health 21. Howard, K. 2002."Lessons from Poland on the Private Administration Committees." Paper presented at the annual Provision of Primary Health Care." London, London School of meeting of the American Political Science Association. Economics. 39. Bukonda, N., P.Tavrow, H.Abdallah, K. Hoffner, and J.Tembo. 22. Hebrang,A., N. Henigsberg,V. Erdeljic, S. Foro, S.Turek, and M. 2002."Implementing a National Hospital Accreditation Zlatar M. 2002."Privatization of the Croatian Health Care Program:The Zambian Experience." International Journal for System: Effect on Indicators of Health Care Accessibility in Quality in Health Care14 (Suppl. 1): 7­16. General Medicine." LijecVjesn 124 (8­9): 239-243. 40. Tavrow, P., J. Heavens, J.Warren Salmon, and C. Lombard. 2003. 23. Vladescu, C., and S. Radulescu. 2001."Primary Health Services: The Impact of Accreditation on the Quality of Hospital Care: Output-Based Contracting to Lift Performance in Romania." Kwazulu-Natal Province, Republic of South Africa Quality Public Policy for the Private Sector 239. Assurance Project, USAID, Bethesda, MD. 24. Eichler, R., P.Auxilia, and J. Pollock. 2001."Promoting 41. Whittaker, S., R.W. Green-Thompson, I. Mccusker, and B. Preventive Health Care: Paying for Performance in Haiti." In Nyembezi. 2000."Status of a Health Care Quality Review Contracting for Public Services: Output-Based Aid and its Programme in South Africa." International Journal of Quality in Applications, eds. P.J. Brook and S. Smith.Washington, DC: Health Care 12 (3): 247­250. World Bank International Finance Corporation. 42. Whittaker, S., C.D. Shaw,A. Bruwer, D. Green, J.J.Taljaard, and 25. Chawla, M., P. Berman,A.Windak, and M. Kulis. 1999. A. Skibbe. 1994."The South African Pilot Hospital "Provision of Ambulatory Health Services in Poland:A Case Accreditation Programme. Part II.The Development of Study from Krakow." Harvard School of Public Health, Standards." South African Medical Journal 84 (4): 193­194. International Health Systems Group, Boston. 43. Bhushan, I., S. Keller, and B. Schwartz. 2002."Achieving the 26. Cotlear, D. 1999."Peru: Improving Health Care for the Poor." Twin Objectives of Efficiency and Equity: Contracting Health Human Development Department (LCSHD) Paper Series 57, Services in Cambodia." ERD Policy Brief Series 6,Asian World Bank,Washington, DC. Development Bank, Manila. 27. Bitran, R. 2001. Paying Health Providers through Capitation in 44. Loevinsohn, B. 2000."Contracting for the Delivery of Primary Argentina, Nicaragua andThailand: Output, Spending, Health Care in Cambodia: Design and Initial Experience of a Organizational Impact and Market Structure. PHR Project, Large PilotTest." WBI Online Journal.Washington, DC,World USAID,Washington, DC. Bank. 28. Preker,A.S., and A. Harding. 2003."Innovations in Health 45. Loevinsohn, B, and A. Harding. 2004. Contracting for Health Service Delivery."World Bank, Human Development Network, Service Delivery:A Review of Developing Country Experience and Washington, DC. Results. Health Nutrition, Population Discussion Paper,World Bank,Washington, DC. 29. Pearson, M. 2000. International Experience of Hospital Autonomy. London: Institute for Health Sector Development. 46. Loevinsohn, B. 2002."Practical Issues in Contracting for Primary Health Care Delivery: Lessons FromTwo Large 30. Chawla, M., and A. George. 1996."Hospital Autonomy in Projects in Bangladesh." WBI Online Journal.World Bank, India:The Experience of APVVP Hospitals." Project supported Washington, DC. by USAID. Harvard University, Cambridge, MA. 47. Eichler,R.,P.Auxilia,and J.Pollock.2001.Output-Based Health 31. Collins, D., G. Njeru, J. Meme, andW. Newbrander. 1999. Care:Paying for Performance in Haiti.World Bank,Washington,DC. "Hospital Autonomy:The Experience of Kenyatta National Hospital." International Journal of Health Planning and Management 48. Marek,T., I. Diallo, B. Ndiaye, and J. Rakotosalama. 1999. 4 (2): 129­153. "Successful Contracting of Prevention Services: Fighting Malnutrition in Senegal and Madagascar." Health Policy Planning 32. Harding,A., and A. Preker. 2003."Private Participation in 14 (4): 382­389. Health Services."World Bank, Human Development Network, Washington, DC. 49. Richardson, P. 1992. Quality, Costs and Cost Recovery:A Comparative Study of the Unidad Sanitaria of the Ministry of Health 33. Bossert,T., S. Kosen, B. Harsono, and A. Gani. 1997."Hospital (MOH) and PROSALUD in Santa Cruz, Bolivia. University Autonomy in Indonesia." Project supported by USAID. Research Corporation and International Science and Harvard University, Cambridge, MA. Technology Institute,Washington, DC. 34. Mcpake, B., F.J.Yepes, S. Lake, and L.H. Sanchez. 2003."Is the 50. Tangcharoensathien,V., and N. Khongsawatt Nas."Private- Colombian Health System Reform Improving the Sector Involvement in Public Hospitals: Case Studies in Performance of Public Hospitals in Bogota?" Health Policy Bangkok." In Private Health Providers in Developing Countries: Planning 18 (2): 182­194. Serving the Public Interest? ed.A. Mills. London: Zed Books. 35. Gilson, L., and A. Mills. 1995."Health Sector Reforms in Sub- 51. Mcpake, B., and C. Hongoro. 1995."Contracting Out of Saharan Africa: Lessons of the Last 10Years." Health Policy 32 Clinical Services in Zimbabwe." Social Science and Medicine 41 (1­3): 215­243. (1): 13­24. 36. Londono, B., I. Jarmillo, and J. Uribe. 1999. Decentralizaiton and 52. Broomberg J., P. Masobe, and A. Mills. 1977."To Purchase or to Reforms in Health Services:The Colombian Case.World Bank, Provide?The Relative Efficiency of Contracting Out versus Washington, DC. Direct Public Provision of Hospital Services in South Africa." 108 | The Millennium Development Goals for Health In Private Health Providers in Developing Countries, ed.A. Mills. 59. Chowdhury,A.M., F. Karim, S.K. Sarkar, R.A. Cash, and A. London: Zed Books, Bhuiya. 1997."The Status of ORT (Oral RehydrationTherapy) 53. Slack K., andW. Savedoff. 2001."Public Purchaser-Private in Bangladesh: HowWidely Is It Used?" Health Policy Planning Provider Contracting for Health Services: Examples from Latin 12 (1): 58­66. America and the Caribbean." Sustainable Development 60. Schellenberg, J.R., S.Abdulla, H. Minja, and others. 1999. DepartmentTechnical Paper Series 111, Inter-American "KINET:A Social Marketing Programme Of Treated Nets And Development Bank,Washington, DC. NetTreatment For Malaria Control InTanzania,With 54. Fiedler, J.L. 1996."The Privatization of Health Care inThree Evaluation Of Child Health And Long-Term Survival." Trans R Latin American Social Security Systems." Health Policy Planning SocTrop Med Hyg 93 (3): 225-231. 11 (4): 406­417. 61. Abdulla, S., J.A. Schellenberg, R. Nathan, and others. 2001. 55. Mills,A. 1998."To Contract or Not to Contract? Issues for Low "Impact on Malaria Morbidity of a Programme Supplying and Middle Income Countries." Health Policy Planning 13 (1): InsecticideTreated Nets in Children Aged under 2Years in 32­40. Tanzania: Community Cross-Sectional Study." British Medical Journal 322 (7281): 270-273. 56. Chuc, N.T., M. Larsson, N.T. Do,V.K. Diwan, G.B.Tomson, and T. Falkenberg. 2002."Improving Private Pharmacy Practice:A 62. Jacobs, B., F.S. Kambugu, J.A.Whitworth, M. Ochwo, R. Pool, Multi-Intervention Experiment in Hanoi,Vietnam." Journal of A. Lwanga, S.Tifft, J. Lule, and J.R. Cutler. 2003."Social Clinical Epidemiology 55 (11): 1148­1155. Marketing of Pre-PackagedTreatment for Men with Urethral Discharge (Clear Semen) in Uganda." International Journal of 57. Stenson, B., L. Syhakhang, C.S. Lundborg, B. Eriksson, and STD and AIDS 14 (3): 216­221. G.Tomson. 2001."Private Pharmacy Practice and Regulation: A RandomizedTrial in Lao P.D.R." International Journal of 63. Agha, S., C. Squire, and R.Ahmed. 1997. Evaluation of the Green Technology Assessment in Healthcare 17 (4): 579­589. Star Pilot Project. Population Services International,Washington, DC. 58. Bhutta,T.I., and C. Balchin 1996."Assessing the Impact of a Regulatory Intervention in Pakistan." Social Science and Medicine 42 (8):1195­1202. Improving Service Delivery | 109 CHAPTER 7 Tackling Human Resource and Pharmaceutical Constraints Joytsna Neopane, an anesthesiologist from Nepal based in New nium Development Goals. Stocks of human resources York City,has just completed her medical residency and expects to for health systems in low-income countries are small-- make $225,000­$250,000 a year once she is hired."Compare and in some countries emigration and HIV/AIDS are that with less than $100 a month I used to make at a government making them smaller. Poor management and inappro- hospital in Kathmandu,and you have the answer why thousands priate human resources policies often cause flows from of doctors from the Indian Subcontinent end up here,"she says. the public sector to the private, and from rural areas to "We observe it--we cannot lie about it. Not respecting urban, increasing internal imbalances. The motivation working hours is a problem that we got used to, and this has and productivity of health workers are often low. created problems for the patients. During teatime, for example, Problems with drugs include national shortages, lack of people go to other places--to work in the private sector, for per- affordability for the poor,counterfeiting and substandard sonal commitments, etc. Even that small amount of time we quality, inappropriate prescribing and use, and inefficient have in the facility we do not use effectively, and patients com- logistics systems. plain about that," says a health assistant in Ethiopia. Policies need to be broad-based.For human resources "The effect of HIV/AIDS on health workers has been they need to address compensation issues but also nonpe- totally ignored. It is difficult to differentiate between patients cuniary factors that motivate staff.Training needs to focus who are infected with the virus and those who are not. Health on skills of relevance to the Millennium Development workers are exposed to difficult situations. It is like sacrificing Goals,in part to reduce skill marketability.Skills also need your life.This is serious at lower levels and in rural areas.There realignment. Relocating low-skilled staff willing to work are several emergencies that do not allow much time to look for in rural areas offers considerable scope for expanding the gloves, and this forces health workers to treat a patient with bare coverage of key Millennium Development Goal inter- hands.And we know the risks we are taking.A lot of colleagues ventions. Public policy also needs to come to grips with have died of AIDS,"reports a health assistant in Ethiopia. issues related to drugs and consumables. A combination Source: References 1 and 2. of sound policies on pricing, procurement, and logistics management and better information for consumers and Health systems rely on two key sets of inputs--human healthcare providers could significantly increase stocks of resources and drugs.Weaknesses in policies and practices drugs, vaccines, and consumables, accelerating progress in both areas undermine progress toward the Millen- toward the Millennium Development Goals. 111 This chapter sets out the key issues and pulls together what is known about the causes of the problems and the Figure 7.2 Doctors across Sub-Saharan Africa in effectiveness of policies to deal with them. the 1990s--how many and how much change 25 Côte d'Ivoire Human resources for health 20 Eritrea Nigeria doctors An increasingly common lament in international health is in Seychelles 15 Zambia that faster progress toward the health-related Millennium 10 Liberia Development Goals is being impeded by a variety of change people Cameroon human resources problems.A recent United States Agency 5 Lesotho 1,000 Botswana São Tomé and for International Development (USAID) report argued that Senegal Principe per 0 Namibia Uganda there is already a "grave and complex human resource cri- percentage Equatorial Guinea Congo, Rep. Sierra Leone Cape Verde ­5 Congo, Dem. Rep. sis"in the health sector in Sub-Saharan Africa that threatens Togo Sudan Mauritania Guinea-Bissau achievement of the child survival and other Millennium Annual ­10 Development Goals.3A recentWorld Bank study found that 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 Average number of doctors per 1,000 people 26 of 28 Poverty Reduction Strategy Papers (PRSPs) from Sub-Saharan Africa identify the performance of the health Source: Reference 96. workforce as an important issue. Of the 22 high-burden countries that account for 80 percent of the world's tuber- culosis cases, 17 have reported that staffing problems are hampering their efforts to reach the 2005 targets.4 increase considerably, while South Asia saw its stock decline (see figure 7.1).WithinAfrica,changes in stocks varied,with What are the issues? Côte d'Ivoire and Nigeria seeing their stocks increase and Sierra Leone and Sudan seeing their stocks fall. Hard data STOCKS OF MEDICAL HUMAN RESOURCES ARE are difficult to come by, but the situation in Sub-Saharan LOW, AND IN SOME COUNTRIES THEY ARE Africa seems to have worsened considerably at the end of FALLING In Europe and Central Asia there are an average the 1990s (e.g., see box 7.1).Tanzania's health workforce is of 3.1 physicians per 1,000 people (figure 7.1). In Sub- projected to fall from 49,000 in 1994 to 36,000 in 2015.5 Saharan Africa there is just 0.1.Within Africa are large dif- Between 1998 and 2001 the number of nurses working in ferences (figure 7.2), with Namibia and Nigeria relatively the public sector inThe Gambia fell from 784 to 655.6 well endowed with doctors, and Eritrea and Liberia having Do low stocks matter for Millennium Development very few. Goal outcomes? Do falling stocks jeopardize progress?The Human resource stocks are not constant.In the 1990s the answer is presumed to be "yes" to both questions, but the Middle East and North Africa saw its physician stock evidence is scant. A study of infant mortality in Malaysia found that, holding other factors constant, a lower popula- Figure 7.1 Doctors across the world in the tion per doctor ratio was associated with a lower infant 1990s--how many and how much change mortality rate.8 But the population per public doctor was not found to be significantly related to the infant mortality 20 Middle East and North rate, leading the authors to conclude that increases in the Africa 15 doctors number of public doctors is partly offset by decreases in in the number of doctors working in private practice. 10 Two other studies--one of Ghana9 and one of Côte East Asia and Pacific change people 5 Industrial d'Ivoire10--found ambiguous evidence. Having more doc- Sub-Saharan Africa Latin America 0 tors in local facilities was associated, other things equal, with 1,000 and the Caribbean Europe and Central Asia taller children,but the estimated effect was very small:raising per ­5 percentage the average number of doctors in a facility to the national ­10 South Asia average would raise the mean height-for-age by only 0.04 Annual ­15 (the national average is ­1.32).Further,the estimated impact 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 of human resource stocks on Millennium Development Average number of doctors per 1,000 people Goal outcomes was highly sensitive to the focus--urban or Note: Data are weighted by population size. rural,nurse or doctor,malnutrition or child survival. Source: Reference 96. Such ambiguity and sensitivity of results is not reassuring. This evidence--together with the widespread evidence on 112 | The Millennium Development Goals for Health the peripheral region of Kolda has just one physician per Box 7.1 Ghana's loss of health sector 86,000 people.The problem is replicated around the world. workers THE MIX OF SKILLS, NOT JUST THE SKILLS The State of Ghanaian Economy Report 2002 reports that 31 per- THEMSELVES, MATTER The skills of the health work- cent of trained health personnel, including doctors, nurses, force are often woefully inadequate. Misdiagnosis is com- midwives, and pharmacists, left the country between 1993 and monplace: in Burundi in 1992 only 2 percent of children 2002, leaving approximately 1.48 physicians per 100,000 peo- with diarrhea taken to health facilities were correctly diag- ple. Bleak as these figures are, they may underestimate the loss. nosed.12 Mistreatment is also commonplace: in the same A report on human resources by the Government of Ghana facilities in Burundi only 13 percent of children correctly based on biannual data (table) shows significantly greater num- diagnosed as having diarrheal disease were correctly rehy- bers of health workers lost from 1996 to 2002.The University drated.12 Even if the correct treatment is administered, of Ghana Medical School, the School of Medical Sciences of there is no assurance that it will be administered success- the Kwame Nkrumah University of Science and Technology, fully: in India in the early 1990s less than 45 percent of and the University for Development Studies Medical School train about 150 medical officers annually. But half of every patients in public health facilities diagnosed with tubercu- graduating class leaves the country within the second year, and losis were successfully treated.13 80 percent leaves by the fifth year. It is not just the skills that matter, it is also the skills mix. This exodus of medical officers is mirrored in other health Often, unskilled labor substitutes for skilled labor.A recent professions. Of 944 pharmacists trained between 1995 and study in Tanzania estimated the required full-time equiva- 2002, 410 had left the country by the end of 2002.The prob- lents for service provision according to national staffing lem is most severe among nurses and midwives: of the 10,145 norms for unskilled and skilled labor and compared these nurses trained during that same period, 1,996 had left Ghana estimates with employed unskilled and skilled staff.5 The by the end of 2002.And only 12 medical laboratory technolo- study found an excess of unskilled labor of 5,000 full-time gists were being produced annually as of 2002,with no guaran- equivalents and a shortage of skilled labor of 8,000 full- tee of their remaining in Ghana after graduation. time equivalents. One interpretation of this finding is that every fourth task that requires a skilled health professional Loss of trained public sector health staff in Ghana, is performed by an unskilled worker. The results of the 1996­2002 ongoing human resources census in Tanzania and time- and-motion studies performed in public health facilities Category 1996 1998 2000 2002 seem to bear this interpretation out. Doctors 1,154 1,132 1,015 964 APPLICATION TO THE JOB IS LOW IN MANY Nurses (including 14,932 15,046 13,742 11,325 DEVELOPING COUNTRIES One obvious indicator of auxiliaries) low application to the job is absenteeism. Recent random Pharmacists n.a. n.a. 230 200 surveys of primary health facilities in six developing coun- n.a. Not available tries found absenteeism rates between 19 percent (Papua New Guinea) and 43 percent (India).14 Low application Source: Reference 7. manifests itself in other ways, too. In Tanzania time-and- motion studies showed overall staff productivity in public facilities as low as 57 percent.5 On average, staff was absent in 7 percent of all observations. Only 37 percent of staff time was spent on patient care.An additional 10 percent of underuse of facilities--cautions against jumping to the con- working time was spent on irregular breaks and social clusion that the size of health workforce stocks alone is the contacts. Staff productivity was found to be related to the major human resources problem facing the health sector in demand for services, which in turn depended on the avail- developing countries. ability of drugs in rural facilities. GEOGRAPHIC IMBALANCES IN STOCKS MATTER Low levels of application reduce technical quality and Another concern is the geographic imbalance of stocks weaken patient confidence in the health system, leading to across rural and urban areas, poor and less poor regions, and declining demand. A recent study in five West African types of institution. Dakar, the richest and most urban countries identified impoliteness, lack of attention to region of Senegal, is home to only 24 percent of Senegal's patient needs, verbal and physical violence, corruption and people, yet 65 percent of the country's physicians work nepotism, and informal fees as key causes for the poor rep- there:there is one physician per 5,000 people.11 By contrast, utation of public health services.15 Tackling Human Resource and Pharmaceutical Constraints | 113 Tackling health workforce issues--stocks and flows Box 7.2 The devastating impact of The stock of human resources in health in a country at a HIV/AIDS on the health workforce particular time reflects the previous period's stock and the inflows and outflows that took place during the previous The HIV/AIDS epidemic is having a devastating impact on period (figure 7.3).The outflow includes migration abroad the stock of health workers, particularly in Africa. It has been (see box 7.1 on the brain drain in Ghana), deaths of med- estimated that HIV/AIDS accounts for 19­53 percent of all ical staff (see box 7.2 on the impact of HIV/AIDS on deaths among government staff in a typical African country. In medical staff), exit from the medical profession for other several countries in Africa 25 percent of nurses are HIV posi- occupations (ask a taxi driver in Havana about his or her tive. Studies of HIV prevalence among African healthcare workers suggest that doctors and nurses are at least as likely to former profession and chances are it will be medicine), become infected as other people. Health workers in Africa are and retirement (many countries in which the health significantly more likely to become infected with HIV than workforce has been aging rapidly use early retirement their peers in the industrial world: the HIV seroconversion risk inducements as a policy lever). For the public sector, out- among surgeons in Sub-Saharan Africa may be 15 times higher flows to the private sector are a major factor. Inflows than in industrial countries. reflect the numbers of newly trained staff being produced, Several country-specific examples confirm this bleak pic- the number of foreigners entering the health sector, and ture. In Botswana the share of health workers with AIDS rose the number of health care workers returning from abroad from 0.3 percent in 1991 to 4 percent in 2003. By 2005 it is or re-entering the profession. projected that 5 percent of healthcare workers will have While hard data are virtually nonexistent, many devel- HIV/AIDS and that 16 percent of healthcare workers will have oping countries appear to face the double burden of low died between 2003 and 2005. In Malawi deaths among nurses inflows and high outflows.1,16­19The Philippines loses three are equivalent to 40 percent of the annual output from nursing times more licensed nurses each year than it produces.20­22 schools. In Mozambique the death rate of nurses tripled during the 1990s. In 2000, 20 percent of student nurses died from In The Gambia, the number of newly trained nurses and AIDS. In Zambia a study in two hospitals found that mortality midwives is insufficient even to fill vacancies in the nation's rates among nurses rose from 2 per 1,000 deaths in 1980 to main hospital.6 A United Nations Conference on Trade almost 27 per 1,000, largely as a result of AIDS. and Development study estimated that 56 percent of all Source: References 3, 24, and 25. migrating physicians move from developing to industrial countries, while only 11 percent move in the opposite direction.The imbalance is even greater for nurses.23 people from entering training institutions, from complet- TACKLING COMPENSATION Differentials in compen- ing their studies, and from joining the profession if they sation--both wages and benefits such as housing--have graduate.They encourage people to think about leaving-- repercussions for inflows and outflows of human resources by exiting the labor force, joining another profession, or in the health sector (box 7.3). Low wages in the medical leaving for another country where compensation is bet- professions relative to wages in other professions discourage ter.26 Low levels of compensation encourage absenteeism, as health workers seek other earnings to supplement the earnings in their regular health sector job. This creates Figure 7.3 Stocks and flows in human resources heavier workloads for those left behind and reduces moti- vation,prompting further absenteeism and exodus. Inflows Outflows Compensation differentials across specialties influence spe- cialization decisions during and after training.27 Low rates of Death Employment compensation in rural settings help explain rural-urban Immigration Disability imbalances. And compensation differentials between the public and private sectors influence transitions from the pub- Training Retirement lic sector to the private sector--and occasionally back again. Emigration Narrowing and even reversing compensation differentials is an important policy tool. Narrowing these differences-- Unemployed Inactive between healthcare and other professions,between the pub- lic and private sectors, between urban and rural areas, across countries--provides an opportunity to shift to a virtuous Source: World Bank staff. circle, by encouraging inflows and reducing outflows, absenteeism, and shirking.Thailand has attracted back med- 114 | The Millennium Development Goals for Health good income. But having a good income was not what Box 7.3 Competing with the private sector health workers said they aspired to most--especially work- in Bolivia ers in the private sector.Training opportunities; challenging work; relations with colleagues; a desirable location (includ- In Bolivia the basic pay of general physicians working in the ing proximity to a good school, for example); and good public health system is $440 a month. Over the years, their physical working conditions all ranked higher than pay in salary increases by a maximum of $65 a month.Working in a both the public and private sectors.And many workers felt deprived border area add another $85 to the monthly paycheck. that their jobs did not meet expectations, especially for At the same time, a single medical consultation in the private sector typically costs $10­$15 in major cities. With limited training opportunities. competition among private providers and the low costs of run- People's job decisions do appear to reflect nonpecuniary ning a medical office, the large private-public salary differentials factors.26,27 A recent study in Uganda found that medical attract physicians into private practice, nullifying the monetary staff in religious nonprofit institutions are hired at below- incentive to work in deprived border areas. market wages.36 The authors suggested that this was prob- Source: Reference 35. ably due to the fact that these institutions provide more services to the poor than do other providers and more ser- vices with a public good element--dimensions of the job that presumably give health workers some satisfaction. ical professionals through a reverse brain drain program Pay, in short, isn't everything--what people can do in offering generous research funding and monetary incen- their jobs also matters. Countries could gain from better tives.28,29 In Zambia nursing salaries were more than dou- understanding the expectations that health workers have bled through the support of 16 development partners, of their jobs and the degree to which these expectations including the European Union and the Danish and Swedish are met.A recent six-country World Health Organization international development agencies. Several countries have (WHO) study of factors that influence migration is a good experimented with bonus schemes for health workers in example of a move in this direction.38 rural areas, some successfully.29,30 Other countries allow There are also encouraging signs that building policies public providers in rural areas to supplement their income around a better understanding of nonpecuniary motiva- through private practice during official working hours.31,32 tions may work. A major reason why health professionals This can have disadvantages, as it is in the financial interest leave rural areas is the perceived unmet need for continu- of public providers to lengthen waiting times and advise ing education.39 Retraining nurse practitioners turned out patients who are able to pay fees to seek private services. not to be very successful in Thailand, while on-the-job Industrial countries can also help.They can raise wages at training in basic medical care for district hospital nurses home to encourage the recruitment and retention of non- proved to be both effective and sustainable. 28,29 immigrant personnel, and they can provide development assistance to help developing countries raise wages to attract WHO YOU RECRUIT MATTERS People raised in rural and retain staff.There are some signs that industrial countries areas are more likely to practice in rural locations and to are taking such steps.With growing evidence that low com- choose family medicine.40 Changing the admission crite- pensation makes recruitment and retention difficult,26,33 the ria of training institutions can have a powerful impact on British government recently awarded nurses pay increases the number of physicians practicing in rural and under- that were much greater than the rate of inflation.34 served areas. Programs that targeted students based on There are limits to what can be achieved through their rural background and commitment to practice rural changes in compensation alone, however. Raising salaries family medicine were able to successfully increase the in the developing world to the levels in industrial coun- physician workforce in rural areas. Family physicians seem tries (even at purchasing power parity exchange rates) is more likely to practice in small and isolated areas.41 Rural simply unfeasible.Neither is using compensating wage dif- recruitment and training yielded some success inThailand. ferentials to keep staff in the public sector from moving to Evidence suggests that the rural service of graduates the private sector (box 7.3). lengthened, with two-thirds of graduates continuing to work in rural areas after their compulsory years of ser- LOOKING BEYOND COMPENSATION Research vice.28,29 suggests that nonpecuniary aspects of jobs matter to people. A recent study in India shows what health workers hope to TRAINING HEALTHCARE WORKERS IN THE get from their job and what they actually get (box 7.4). In RIGHT SKILLS CAN HELP RETAIN THEM IN THE Andhra Pradesh's public and private sectors, less than 40 PUBLIC SECTOR Training opportunities are clearly percent of respondents felt that their current jobs provided a valued by health workers. Training increases skills and Tackling Human Resource and Pharmaceutical Constraints | 115 Box 7.4 What do health workers in India want most? A recent study of motivation among health workers in the Indian ing opportunities to improve or learn new skills. On these issues, state of Andhra Pradesh sheds light on the issue of staff morale (see Indian health workers are less satisfied with their current jobs, figure).37What health workers in both the public and private sec- especially with training. Although health workers in both states tors value most is a good working relationship with colleagues. and in both sectors report earning less than they would like,a high And in both public and private sectors and in both states, a very income is not considered one of the most valued job attributes. high percentage (80 percent or so) of health workers feel that their Frustration over the discrepancy between actual and desired current job provides this. Number two and three on the list of income is felt equally in both the public and private sectors. most valued job attributes are good working conditions and train- What health workers in Andhra Pradesh want from their job--and whether they are getting it Andhra Pradesh: Public Sector Health Workers Good physical working conditions Good income 100 Knowing what is expected and being able to achieve it Good employment benefits 80 Freedom from political interference 60 Time for personal or family life Not needing to pay bribes 40 20 Desirable location Security of job Important to health workers? 0 Present in workplace? No interference by superiors Training opportunities Respected and trusted by clients Challenging work Superior recognizes good work Tools and materials to use skills fully on the job Good working relationship Good opportunities with colleagues to advance Andhra Pradesh: Private Sector Health Workers Good physical working conditions Good income 100 Knowing what is expected and being able to achieve it Good employment benefits 80 Freedom from political interference 60 Time for personal or family life Not needing to pay bribes 40 20 Desirable location Security of job Important to health workers? 0 Present in workplace? No interference by superiors Training opportunities Respected and trusted by clients Challenging work Superior recognizes good work Tools and materials to use skills fully on the job Good working relationship Good opportunities with colleagues to advance Source: Reference 37. 116| The Millennium Development Goals for Health motivation. It is, however, a double-edged sword--train- away by the private sector, urban provider organizations, ing makes health workers more marketable and more and foreign health sectors. likely to leave the public sector for the private sector or for other countries. This is an old problem in economics.42 Medicines and other health supplies The standard way around it is to focus government spend- Medicines and other health commodities are key elements ing on developing specific skills and to let health workers of effective interventions against child and maternal mortal- themselves pay--in the form of lower wages--for any ity and deaths from communicable diseases (see figures 3.2 general training. Such a policy seems sensible in the light and 3.3). Increased use in the developing world of good- of the emerging evidence on poaching and international quality vaccines, vitamin and mineral supplements, antibi- migration of health workers.The countries that have emu- otics, antimalarials, and tuberculosis and HIV/AIDS drugs lated the training standards of industrial countries (such as Ghana) have been most vulnerable to poaching by them.14 would substantially reduce mortality if used rationally. So would use of insecticide-treated bednets, condoms, and In Ethiopia and The Gambia community nurses and other health supplies.The successful release and widespread health officers who lack internationally recognized quali- use of medicines still under development could reduce fications are less likely to migrate than those who have them.14 mortality rates still further.Given this potential contribution to human health, it would be surprising if medicines were Training in areas of relevance to the Millennium not the subject of an important array of policy issues. Development Goals is a good example of specialized training that is unlikely to be especially valued in industrial countries. What are the policy and implementation issues? The competent training of the Integrated Management of DRUGS AND OTHER SUPPLIES ARE ABSENT IN Childhood Illness (IMCI) initiative is a case in point.It seeks MANY FRONTLINE FACILITIES It is not just health to improve provider skills in the management of childhood workers who are often absent from facilities. So are drugs illness, notably in the treatment of the major killers of chil- and other supplies (box 7.5).This is problematic--studies dren under five; in the delivery of preventive measures, such have shown repeatedly that when drugs are available, as immunization;and in imparting advice and counseling on feeding practices and protective behaviors.43,44The treatment patients have more confidence in the public health system guidelines and training methodology were developed and refined through research and field testing in numerous countries, including Bangladesh, The Gambia, Ethiopia, Kenya,Tanzania, and Uganda.45­48 The quality of care pro- Box 7.5 Lack of drugs threatens the vided by health workers trained in this methodology has Millennium Development Goals been found to be significantly better than that provided by nontrained health workers in several settings. In Tanzania Acute respiratory infections are one of the leading killers of trained providers were twice as likely to prescribe antibiotics children under five, accounting for about one in five deaths among children in this age group.53 The survival of children appropriately. In Bolivia trained providers were 10 times with acute respiratory infections depends to a large extent on more likely to recognize the danger signs of a sick child.And the availability and appropriate use of effective antibiotics. Data in Niger health workers' performance (tested against the from health facility surveys conducted in 21 countries during IMCI algorithm) increased 34­85 percent after training.49­51 1992­97 showed that in some countries less than 30 percent of health facilities had first-line antibiotics for acute respiratory REALIGNING THE SKILLS MIX CAN ALSO HELP infections.The median for all 21 countries was 79 percent.12 Many Millennium Development Goal interventions can In Côte d'Ivoire in 1987­88,24 percent of rural facilities sur- be delivered by lower-skill providers. Community health veyed did not have antibiotics in stock, and 42 percent did not workers, for example, are providing more maternal and have vaccines in stock.10 A report for the Burkina Faso Poverty child health and nutrition services. (Box 6.8 on Céara Reduction Strategy Paper found that nearly 20 percent of facili- describes an example of successful skill mix realignment.) ties had run out of essential vaccines and that 24 percent of cen- A recent study estimates that at high levels of coverage, ters had refrigerators for storing the vaccines that did not some 90 percent of full-time equivalent staff required to function.InTanzania,with its donor-funded vertically organized provide interventions related to childhood diseases in essential drugs program,the picture looks fairly good:89 percent Tanzania fall into the categories of nursing and midwifery of facilities had measles vaccines in stock, 88 percent had tetanus skills and unskilled labor.52 An additional attraction of toxoid vaccines, 86 percent had BCG vaccines for tuberculosis, lower-skill health workers is that they are more likely to and 84 percent had diphtheria,pertussis,and tetanus vaccines.54 be willing to work in rural areas and less likely to be lured Tackling Human Resource and Pharmaceutical Constraints | 117 and in government.The availability of pharmaceuticals has also been shown to motivate health professionals. Box 7.7 The problem of counterfeit drugs INFORMATION ASYMMETRIES LEAVE PATIENTS Counterfeit drugs are a widespread global problem and a seri- VULNERABLE Providers and drug vendors have a huge ous threat to health in developing countries. The spread of information advantage over patients, who are vulnerable to counterfeit drugs is facilitated by weak institutional regulatory prescriptions of inappropriate, substandard, and counterfeit structures in many developing countries and the ease of copy- drugs (boxes 7.6 and 7.7).This asymmetry makes reliance ing many drugs. According to the International Federation of on financial incentives to get drugs into facilities highly Pharmaceutical Manufacturers Association (IFPMA), 7 percent problematic. Informational asymmetries pose multiple of drugs sold globally are counterfeit. Counterfeiting drugs can challenges and suggest the need for regulation (of involve changes in the active ingredients, dosage, package providers, retailers, distributors, and manufacturers) and inserts, packaging, manufacturers' names, batch numbers, expi- behavior change (for patients but also for providers and ration dates, and documentation related to quality controls. Counterfeit drugs are associated with a spectrum of medical retailers). risks (table). HIGH DRUG COSTS DETER EFFECTIVE USE OF A recent editorial in the British Medical Journal57 reviewed DRUGS The cost of medicines can be a large part of the the evidence on counterfeit drugs in developing countries. A overall costs of treatment.As with other medical expenses, bleak picture emerged. A recent survey in the Philippines the uncertainty surrounding the need to incur drug showed that 8 percent of drugs were fake.57 In Cambodia a study found that 60 percent of drug vendors sold antimalarial expenses makes insurance coverage attractive. The high tablets from stock that should have been destroyed or fakes with out-of-pocket spending on drugs in many developing no active ingredient.A recent survey of mainland Southeast Asia countries reflects a lack of effective insurance arrange- reported that 38 percent of tablets sold as the new antimalarial ments or the lack of coverage of drugs in insurance drug artesunate were fakes.57 And the list goes on. schemes that are operating. The high cost of drugs creates another problem as well. The medicines used to prevent and treat communicable Medical risks associated with counterfeit drugs diseases are associated with positive externalities--their Type of counterfeit drug Associated medical risk Perfect imitation--the same Limited, assuming that the active ingredients and quality is good packaging as the real drug Box 7.6 Inappropriate or "irrational" Inadequate imitation--the Reduced efficacy and, in the drug use same active ingredient but case of antibiotics, of insufficient quality and development of pathogen quantity resistance A potentially large--and perhaps increasing--share of house- hold spending on drugs is likely to be at best wasteful and at "Placebo"--looks like a real Lack of efficacy drug, but contains no active worst hazardous to health. Studies have indicated that 30­60 ingredient percent of patients in primary healthcare centers receive Poisonous--contains Physical injury or death antibiotics--perhaps twice what is clinically needed. Fifteen harmful or poisonous billion injections are administered each year. Half of them are substances not sterile, and a large share are unnecessary.55 It has been esti- mated that if Integrated Management of Childhood Illness Source: Reference 57. guidelines were followed in Uganda, the number of drugs pre- scribed per consultation would fall sharply and the cost per consultation would drop from $0.82 to $0.17. The consequences of inappropriate drug use are not just consumption by one person benefits other people as well, economic. They include increased drug resistance. A recent by reducing their risk of infection.This makes subsidiza- study reported significant increases in penicillin-resistant Streptococcus pneumonia, one of the most important pathogens tion appropriate, since the price set by the market would associated with acute respiratory infections. Rates of resistance be too high and use rates of externality-generating medi- range from zero to 60 percent in developing countries, with cines too low. the highest rates in South America, Sub-Saharan Africa, and Beyond insurance and externality arguments is an parts of Asia, such as Hong Kong (China), China, and the equity argument. Universal access to medicines--espe- Republic of Korea.56 cially life-saving ones--is a principle that commands much support (it was endorsed in the 2001 Doha 118 | The Millennium Development Goals for Health Declaration on Trade-Related Aspects of Intellectual Property Rights, or TRIPS, and Public Health). Universal Box 7.9 High drug costs in Vietnam deter access might be interpreted to mean that poor and near- use and cause impoverishment poor households ought to pay a low price even for medi- InVietnam in 1993 a single visit to a public hospital by a mem- cines that involve no externality, so that the cost does not ber of a household in the poorest fifth of the population deter them from getting necessary medicines or push resulted in outlays on medicines equivalent to 40 percent of them into (or further into) poverty (boxes 7.8 and 7.9). the household's annual nonfood consumption (discretionary This provides a rationale for governments seeking to influ- income) (see figure).A single visit to a commune health center ence--at least for poor and near-poor households--the resulted in a bill for drugs equal to 11 percent of the house- retail price of all drugs, not just those used to prevent and hold's annual nonfood consumption. treat communicable diseases.They have a variety of instru- Not surprisingly, a large number ofVietnamese--3 million ments at their disposal. people according to one estimate61--were pushed into poverty Because governments end up footing a sizable share of a as a result of high out-of-pocket payments for healthcare,much country's drug bill, they have an interest in the price they of it attributable to drug expenditures. Households inVietnam pay for drugs, as well as in the quality and quantity of appear to have been deterred from using health services drugs dispensed. Through strategic purchasing, govern- because of high drug costs:having insurance coverage for drugs and inpatient care has a substantial positive effect on the use of ments can exert a substantial influence on the price they hospital care.62,63 pay for drugs. Governments also have several means to influence prescribing behavior and use patterns. The high cost of drugs inVietnam, by source, PATENTS AND THE PAUCITY OF R&D ON DIS- 1993 and 1998 EASES AFFECTING PEOPLE IN DEVELOPING 45 COUNTRIES LIMIT TREATMENT OPTIONS The Poorest quintile 40 use of patents to provide incentives for research and devel- Richest quintile drug (%) nonfood 35 by 30 annual contact 25 absorbed per 20 capita Box 7.8 Do we know how affordable drugs 15 really are? per of 10 expenditure, consumption 5 Despite the importance of drug prices, little is known about Share 0 the prices people pay for medicines in developing counties. A 1993 1998 1993 1998 1993 1998 new approach to measuring drug prices focuses on a range of Public Commune Drug 30 key medicines that address the global disease burden, partic- hospital health center vendor ularly for low- and middle-income countries.58 Several of these drugs--including antibacterial medicines, such as co- trimoxazole pediatric suspension for treatment of acute respi- ratory infection in children under five, and antimalarials, such opment (R&D) leaves many key drugs beyond the financial as artesunate and pyrimethamine with sulfadoxine--are rele- reach of poor countries.The extension of the patent period vant to the Millennium Development Goals. to 20 years under the TRIPS agreement (even the United The preliminary results of surveys based on this new States, with its robust patent system, previously had only a approach show that in South Africa the price that the lowest- 17-year period) made this problem worse.These concerns paid government worker needs to pay to afford a course of have led to heated debate, brinkmanship by some develop- treatment with amoxicillin (an antibacterial drug) in the pri- ing country governments, and exemptions for public health vate sector is between half a day's and a day and a half's wages, depending on the type of amoxicillin used.59 In Armenia the crises. The concerns,it should be noted,currently apply to rela- lowest-paid government worker had to work for 148 days to pay for a course of treatment with branded aciclovir, an antivi- tively few drugs, since most drugs on WHO's Essential ral drug.The same money would be enough to pay for enough Drug List (the most cost-effective therapies for a broad list rice or sugar for 10 years.The use of a generic drug reduced of diseases) are off-patent in all countries. But as resistance the price by 40 percent but still left it unaffordable for most to some essential medicines increases, replacements are people. In Kazakhstan surveys found that the prices of only 4 being developed.These new treatments will be unaffordable of 85 drugs were lower than the international median.60 to developing countries. In 2002, for example, the cost for the antimalarial drug chloroquine was $0.10 per treatment. Tackling Human Resource and Pharmaceutical Constraints | 119 But resistance to chloroquine is growing in many parts of 12 percent of government spending allocated to the pur- the world, and alternative treatments are far more expen- chase of drugs in Africa is used effectively.69 sive--30 times more for coartem and 400 times for The management chain can be improved, however. Key malarone. Similarly the six-month treatment under the measures include: directly observed treatment, short-course (DOTS) regimen · moving the logistics functions from the backroom to for people with tuberculosis is $20, while the newer mul- tidrug therapy treatment is $400.64 the boardroom, recognizing that it is a core function that needs to be strategically managed and invested Another concern is the low levels of R&D for drug treatment of "neglected" diseases--diseases that dispropor- · putting computerized information management sys- tionately or exclusively affect poor countries, with effec- tems in place that gather timely consumption and tive demand that is too limited to make R&D profitable. stock data and using information to drive the system The facts are striking. Global expenditure on health R&D has risen significantly and continues to rise. Yet of the · preparing annual forecasts and revising them semian- 1,393 new drugs approved between 1975 and 1999, only nually 16 (just over 1 percent) were specifically developed for · negotiating better purchase prices through group tropical diseases and tuberculosis--diseases that account purchasing organizations for 11.4 percent of the global disease burden but which are disproportionately concentrated among poorer popu- · strategically managing distribution by reducing interme- lations.65 By contrast, 179 new drugs were developed for diate distribution points and outsourcing the distribution cardiovascular diseases.65 · establishing key performance measures to continually assess the performance of the system What are the options for getting drugs and other supplies to the frontline and enhancing treatment practices? · focusing on driving down the total delivered cost, rather than the unit cost70 CHANGING BEHAVIOR AMONG CONSUMERS AND PROVIDERS Providing health education (infor- All of this is doable. In Ghana, as part of the Strategies for mation on dosage and mode of administration) to patients Enhancing Access to Medicines (SEAM) initiative funded and building trust between patients (and their caregivers) by the Bill and Melinda Gates Foundation, a portion of and prescribers (particularly as part of DOTS) have the drug distribution system was extensively reformed to proved successful in addressing low patient compliance as improve the use of medicines by mission hospitals. The well as high levels of self-medication.66 Group discussions effort has included dovetailing pharmaceutical distribu- with mothers, training seminars for providers at the com- tion with other proven logistics systems (such as social munity level, and districtwide monitoring have reduced marketing). the use of "irrational" injections in children.67 Treatment guidelines and training and continuing education for CREATING INCENTIVES TO ENCOURAGE health providers on the use of drugs are also potentially RESPONSIBLE PRESCRIBING Behavior change pro- grams for providers and retailers will have only limited useful tools. effects if there are strong economic incentives not to be IMPROVING THE MANAGEMENT CHAIN TO GET responsible in prescribing drugs. An extreme example is THE DRUGS TO THE FRONT LINE In the retail sec- that of doctors in Georgia, who contributed to the spread tor in the industrial world, supply-chain costs are as low as of drug-resistant tuberculosis (box 7.10). Prescribing 3­6 percent of the cost of goods sold in the grocery sector DOTS drugs to tuberculosis patients offers the doctors, and 6­8 percent in the retail sector. In healthcare, supply- who are paid just $50 a month, no additional income, chain costs are as high as 38 percent of the cost of goods. since the drugs are provided free by the German govern- Some of this reflects a lack of continued investment in ment. Instead, the doctors prescribe a cocktail of other healthcare supply chains, resulting in manually controlled drugs, at a cost of $300­400 a month, causing patients to logistics systems with considerable paper shuffling, lengthy develop multidrug-resistant tuberculosis. This is a classic order and delivery cycle times, excessive inventory, multi- example of a well-intentioned policy--to protect the poor ple product handling activities,lack of information sharing from drug costs--that has not been thought through. If among trading partners, and an operational rather than a irresponsible prescribing is to be deterred, providers have customer focus.68 Few of these problems are limited to the to have a financial incentive to prescribe responsibly.This logistics sector. One study estimates that because of waste can be done through a voucher scheme in which patients and inefficiencies in drug selection and procurement, only present the provider with a voucher (usable only for 120 | The Millennium Development Goals for Health Box 7.10 Perverse provider incentives and tuberculosis drugs in Georgia Twenty-three-year-old Gia is thin, and his eyes shine with fever. Recognizing tuberculosis as a disease of poverty, the DOTS "Please tell me, isn't there anything to get rid of this heat?" he strategy requires the treatment to be free.The German govern- asks,sitting hunched on a bench in a damp hall in the Abastumani ment has been donating antituberculosis drugs to Georgia since hospital, 1,500 meters high in Georgia's forested mountains. "A 1995. Dr. Endeladze said patients did not pay for their treatment thousand paracetamol won't get rid of this temperature." at Abastumani, but nearly all the patients interviewed said they Gia and his fellow patients are hoping to be cured of tubercu- were paying far more than the $50 it costs to treat a patient with losis, but they are at the center of a public health disaster that a standard course of DOTS. Gia and others said they paid threatens Europe with a deadly form of tuberculosis resistant to $300­$400 a month for treatment, a small fortune in a country standard drug regimens, the so-called multidrug-resistant tuber- where a doctor's monthly salary--if it is paid--is $50. culosis.To blame is the treatment these patients receive in places With cash rather than cure on many impoverished doctors' like Abastumani, 125 miles west of the capitalTbilisi. minds, the uncontrolled distribution of antituberculosis drugs has Gia had to sell his apartment,leaving his wife and two-year-old created the perfect breeding ground for multidrug-resistant tuber- son homeless,to pay for a cure at the hospital after more than three culosis, dubbed "ebola with wings" by a Harvard medical school years of treatment elsewhere failed to rid him of his cough, night study. Annabel Baddeley, manager in Georgia of the British aid sweats, and crushing fatigue. He now rests in a wing of the run- agency Merlin,which is working to adapt DOTS to Georgia's rural down hospital.The myth of fresh,mountain air as a cure for tuber- areas, said that unpaid doctors were one of the greatest threats to culosis still holds among Georgia's 4 million people. The Soviet health there."Doctors will charge tuberculosis patients who come system once supported a large network of such hospitals, which to them without referring them to the specialized tuberculosis allowed patients to convalesce for two years or more.The system facilities and treat them for `pneumonia' or `bronchitis' with one or was rich in time but poor in medication, giving only a small frac- two tuberculosis drugs when they should be using four or five.The tion of patients the antituberculosis drugs with proven efficacy. patients will then eventually find their way to the tuberculosis facil- In the chaos that followed Georgia's independence and the ities with advanced tuberculosis, having developed drug resistance conflicts that racked the country between 1991 and 1994, the as part of the unnecessarily expensive package they've paid for." tuberculosis notification rate tripled, from 29 new cases for every Patients like Gia are also being treated incorrectly atAbastumani 100,000 people in 1988 to 89 in 2001.To combat the disease, in with drugs usually reserved for treatment of multidrug-resistant 1995 Georgia adopted theWorld Health Organization's standard tuberculosis, worsening the resistance problem and condemning strategy--the directly observed treatment,short-course (DOTS). those infected with tuberculosis to years of debilitating illness or Widely promoted in the developing world, DOTS is a simple death as second-line drugs such as kanamycin lose their potency. regimen of four to five drugs taken three times a week for at least Treatment for multidrug-resistant tuberculosis is about 100 times six months under the observation of trained staff who make sure more expensive than standard DOTS. The World Health that the medicine is taken properly. Finishing the course is essen- Organization will not sanction the use of second-line drugs until tial, as patients often feel better after a month or so, stop taking the standard strategy is up and running, leaving patients like Gia their drugs, and promptly relapse with a tougher strain of tuber- without treatment.They could go on to infect 15 people in a year. culosis that is resistant to standard DOTS medicines. Reliable statistics are hard to come by in Georgia, but tuber- The hospital's director, Tariel Endeladze, is an enthusiastic culosis specialists estimate that as many as 10 percent of new convert to the DOTS regimen. But the collapse of Georgia's patients with tuberculosis and up to 25 percent of patients being healthcare system means that the misuse of antituberculosis drugs retreated for the disease could have multidrug-resistant tubercu- is widespread.The government has no money to pay its nurses losis. The structural health problems that created this crisis in and doctors, it cannot afford health education, and the market in Georgia are repeated in almost all the former Soviet states. pharmaceuticals is unregulated. Source: Reference 74. approved drugs), which the provider takes to a payer (pos- absent because they have been "borrowed" by providers: sur- sibly a donor) to convert into cash. veys in Uganda suggest that an average of 70 percent of med- ical supplies and drugs in public facilities were appropriated OTHER INCENTIVES MATTER, TOO Better prescrip- by staff for use in their private work.71­73 Where providers tion and management are important. But without institu- have an incentive to have drugs available, they often are. In tional arrangements that give the right incentives to all parties China, where providers and retailers make much of their involved in drug distribution and prescription,availability will income from selling drugs, drugs are readily available--the improve only marginally. Part of the reason--sometimes a public policy challenges are to reduce the prescription of major part--why facilities lack drugs is that providers have inappropriate and unsafe drugs and to make drugs more little incentive to have drugs there. Sometimes drugs are affordable.The same story emerges in Georgia (box 7.10). Tackling Human Resource and Pharmaceutical Constraints | 121 The drawback to creating incentives to prescribe drugs, tion between the two organizations.An autonomous entity of course,is that inappropriate or poor-quality drugs will be might not be an efficient option for small and very poor prescribed. Creating regulation and behavior change pro- countries. In such cases, countries could rely on interna- grams to limit this response is not impossible, but it is not tional or regional mechanisms, such as the WHO easy. It is almost certainly unwise to try to solve the drug Certification Scheme, created in 1975 to help small availability problem by providing financial incentives to pre- importing countries gain access to relevant information scribe drugs.There may,however,be scope for changing the and establish minimum safeguards. More recently the incentive arrangements in the distribution of drugs. Some WHO, supported by several UN organizations and the countries have contracted out the distribution of medicines World Bank, launched a pilot procurement initiative to to the private sector. Doing so may yield efficiency gains prequalify manufacturers of HIV-AIDS drugs. (evaluations do not appear to have been undertaken). INSURANCE POLICIES ON DRUG PRICES AND REGULATING DRUGS According to the WHO, an DRUG SPENDING Drug costs are often not covered or effective drug regulatory system ought to include a full only partly covered by insurance (box 7.11).Whether the drug registration process, unbiased drug information, over- exclusion of outpatient drug costs makes sense is debat- sight of promotional activities, post-approval drug safety able. It makes sense for insurance to focus on high-cost, monitoring,quality-control testing,pharmaceutical inspec- low-frequency costs, such as expensive inpatient drug tion services, and certified compliance with good manu- treatment. But not covering outpatient drug costs may facturing practices.75 Just one in three drug regulatory deter people from taking preventive measures and seeking agencies works effectively in developing countries, as care from a low-level provider as soon as they fall ill.They became clear in a recent WHO study of pharmaceutical may well get even sicker later and end up in a hospital, regulation in Australia,Cuba,Cyprus,Estonia,Malaysia,the incurring large bills against which they are fully insured. Netherlands,Tunisia,Uganda,Venezuela,and Zimbabwe.76 The result may be higher costs for the entire insurance Some countries, such as the Lao People's Democratic scheme.The fact that private insurers in countries like the Republic, have no formal drug regulatory agency. Others United States cover drug expenses likely reflects a belief require only a notification from the manufacturer or that covering certain preventive measures and outpatient- importer in order to market the product, not authorization administered medicines lowers overall costs. from a regulatory agency. On paper Nigeria has an appro- priate formal system: market authorization requires evi- DRUG PRICES AND THE ROLE OF GOVERNMENT dence of safety and efficacy, and samples of the product Governments have ways of influencing drug prices--both must be regularly analyzed. But implementation is usually the prices they (and ultimately taxpayers) pay when they weak because of lack of human and financial resources. subsidize the cost of drugs and the prices consumers pay out Guyana applies a reasonable, pragmatic approach, making of pocket.Governments can exert a direct influence by reg- authorization dependent on proof of registration by a drug ulating retail prices through fixed prices,risk-sharing agree- regulatory agency in Australia, Canada, the United States, ments, and reference-based pricing schemes. In China, or the United Kingdom. where the majority of drugs are purchased from retailers, a Developing countries will not be able to set up a drug sizable fraction of drugs are sold at a regulated price. regulatory agency with the resources and capacity of the Governments can influence retail prices indirectly U.S. Food and Drug Administration or the European through policies affecting the domestic pharmaceutical Medicines Agency. But almost all countries should be able industry (antitrust regulation, profit controls, incentives to to set up a drug regulatory agency that can handle the most encourage market entry by new competitors) and foreign essential functions in an efficient, transparent, and afford- trade. The large increase in affordability of drugs in able way.One important policy issue is whether the agency Vietnam between 1993 and 1998 (see box 7.9) was due should be a department of the ministry of health or an largely to deregulating the pharmaceutical industry and independent agency. An independent status is assumed to opening the sector to international trade.These two mea- allow the drug regulatory agency to perform its functions sures resulted in a 33 percent decline in real terms in the without political interference.There can, however, be dis- medical price index.63 Governments can also influence advantages in having an autonomous drug regulatory retail prices indirectly by making large bulk purchases agency. Badan POM, Indonesia's drug regulatory agency, from manufacturers and then selling drugs at wholesale once a department of the Ministry of Health, is now prices to the private nonprofit sector, as the governments directly accountable to the President's Office. But lines of of Ethiopia, Ghana, Malawi, Tanzania, Uganda, Zambia, authority are not clear, resulting in dysfunctional competi- and Zimbabwe do.81 122 | The Millennium Development Goals for Health Box 7.11 Drug subsidies and drug insurance: Different countries, different policies In Hungary drugs are free of charge or reimbursed at 50 percent, In Indonesia drugs are provided free of charge to specific 70 percent, or 90 percent, depending on patient characteristics. groups--poor families, schoolchildren, people with disabilities, Generic products on the pharmaceutical benefits list are reim- the elderly, government officials, and public enterprise employees bursed at 90 percent of the price. People on public assistance and and their families--through primary health centers and public those with chronic diseases receive free drugs, or they are reim- hospitals.78 People not included in any medical insurance system bursed at 90 percent.77 have to pay for drugs obtained from public or private facilities.79 Estonia has had a drug reimbursement system since 1993 for In Nepal drugs are distributed free of charge in public health pharmaceuticals purchased from outpatient pharmacies.There is facilities. Large hospitals, managed by semi-autonomous hospital a mandatory patient copayment for each purchase of drugs. All management committees, can sell drugs at either subsidized or prescription drugs exceeding the patient's copayment are cov- full price. ered by at least a 50 percent reimbursement.The reimbursement In Myanmar cost-sharing drug shops have opened in hospi- rate increases up to 90 percent or 100 percent for medicines on a tals, with 43 drug items supplied by the Central Medical Stores list of more serious diseases, in which case the copayment is also Depot (CMSD). Drugs are sold at a maximum 15 percent profit lower.77 margin over the CMSD price.The CMSD cost of drugs has to be Since July 1997 in Albania only children under the age of returned to the government budget from the proceeds of sales.79 one, war veterans, and people with disabilities receive a full sub- In Guatemala the Ministry of Health provides drugs free of sidy for essential drugs.77 charge in primary or secondary health centers and hospitals. In In Lithuania drug costs are fully covered for inpatient care 2001 almost 70 percent of the population had access to essential but are not covered for the majority of the population for outpa- drugs.80 tient care.77 Ghana is currently scaling up for a national health insurance In Azerbaijan drugs are free of charge for inpatient care but scheme that will include limited coverage for pharmaceuticals. not for outpatient care, except for people with cancer and some psychiatric diseases. Source: World Bank staff. Whether governments sell such drugs to the private the government (box 7.12).Governments can increase their sector or keep them for use or sale in their own facilities, bargaining power still further by joining forces with other they can exert a major influence over drug prices by governments to set up a multinational purchasing pool.The engaging in strategic purchasing. One issue is whom to scope for putting downward pressure on drug prices is very purchase from. In the past many governments relied heav- large. ily on their own domestic industry. Considerations of The third issue is what governments should purchase.A economies of scale, however, point to purchasing from major choice is between brand name drugs and generics. multinational companies--or at least from domestic affili- Since generics are sold under an international nonpropri- ates of multinational companies.Today developing coun- etary name, there is limited scope for product differentia- tries seem to be relying less and less on domestic tion and much greater scope for competition. This puts manufacturers, whose share of world pharmaceutical pro- downward pressure on the prices not only of generic duction decreased from 11 percent in 1985 to 7 percent in drugs but also of brand drugs. If the generic and the brand 1999.A few cases buck this trend.Cuba's industry has been drug are therapeutically equivalent, quality should not highly innovative, bringing $100 million a year in export need to be compromised. Drugs not subject to any exclu- earnings. And Brazil recently fell back successfully on its sive marketing rights, patents or otherwise, such as most of domestic industry to make credible a threat to invoke a the drugs on theWHO's Essential Drug List, are the obvi- compulsory license to produce its own antiretrovirals ous target of a generics policy. unless significant price concessions were made by the research-based global pharmaceutical industry (they SUBSIDIZING DRUGS Many drugs--even those for were). But there are also plenty of failures, with govern- communicable diseases--are not subsidized, in part ments wasting scarce revenues supporting uncompetitive because of budget constraints in developing countries. domestic industries.82 According to the WHO, a regular supply of "essential" The second issue is whom to purchase with. Govern- drugs would cost at least $5 per capita per year.79 This is ments can increase their monopsony (single buyer) power beyond the means of many developing countries--many by getting donors to engage in pooled procurement with spend this much on all healthcare costs. Donors can help Tackling Human Resource and Pharmaceutical Constraints | 123 Box 7.12 The attractions of pooled procurement In Mozambique drug imports are financed largely by donor Analysis of the negotiations for HIV/AIDS drugs suggests that assistance. In the past a lack of coordination in drug procurement regional pooling procurement benefits not only countries with resulted in huge inefficiencies.83 The government determined small markets, high need, lack of private insurance, and lack of a which types and quantities of drugs to purchase only after a pharmaceutical industry, such as countries in Central America donor had allocated a specific amount of finance.This resulted in and the Caribbean, but also countries with large markets and an unpredictable funding, erratic purchasing cycles, difficulties with established pharmaceutical industry, such as Argentina, Colombia, long-term planning, and tied donations that resulted in frequent and Mexico.The use of generic manufacturers prequalified by the stockouts and expiring drugs. Decisions on what and how much WHO allowed for greater price competition and further reduc- to buy were offer driven, and drug imports expanded or shrank tion of prices. according to donor financing. After several agencies pushed for The Rockefeller Foundation, Management Sciences for restructuring based on the Ministry of Health's specification of Health, and the South Center recently identified lessons from needs, in 1997/98 donors pooled their funds and responded to pooled procurement initiatives in Latin America and the the government's priority. Currently, seven donors are making Caribbean, the Middle East and North Africa, and Sub-Saharan multiyear commitments to a common pool for drugs. Africa.86 The analysis identified several success factors: political The benefits of pooled purchasing are also illustrated by the will and organizational commitment, adherence to a single pur- use of multinational purchasing pools for HIV drugs in the chasing agreement, secure and trustworthy finance and payment Americas. Almost 2 million people in Latin America and the mechanisms, a permanent and autonomous procurement secre- Caribbean are living with HIV/AIDS, according to Joint tariat, harmonization and standardization, good pharmaceutical United Nations Program on HIV/AIDS (UNAIDS) figures.84 procurement practices, and effective quality assurance. The Caribbean, where about 500,000 people are living with Pooled procurement is more likely to succeed if pool mem- HIV/AIDS, has the second-highest HIV prevalence rate in the bers share similar economic,regulatory,and cultural backgrounds. world after Sub-Saharan Africa. The high cost of antiretroviral Currency, language, and a harmonized drug registration proce- drugs to treat HIV/AIDS makes it difficult for patients to pay dure seem to be key factors of success.Two regional groups were for the drugs and for public and private insurance to make treat- identified with significant potential for expansion:the Association ment affordable. Negotiated agreements between ministries of Africaine des Centrals d'Achats de Médicaments Essentiels and health and pharmaceutical companies reduced the prices of the Commonwealth Regional Health Community and Southern antiretroviral drugs in Latin America and the Caribbean,85 but African Development Community. Each organization has 16 the prices were still high compared with generic prices in coun- member countries, with 12 countries belonging to both organi- tries like India. There were also wide differences between zations.The expected next steps are to agree and formally request countries, with some countries paying up to 10 times more for assistance for funding proposal development. the same treatment. Source: World Bank staff. governments meet the cost of providing free or subsidized Health Sector Reform project set out to do precisely this: drugs. Between 1999 and 2002, $400 million in World accompanying the $6.9 million four-year loan were a series Bank loans was spent on pharmaceuticals and other med- of promised policy reforms, including strengthening the ical goods. In Tajikistan external donations accounted for government's drug regulatory capabilities and improving more than 40 percent of pharmaceutical expenditures in the cost-effectiveness of public financing.77 1998.77 This level of support is clearly not sustainable (see chapter 4 on financing issues). PATENTS, TRIPS, AND THE AFFORDABILITY OF The long-term solution is to set priorities for govern- NEW DRUGS The strengthening of patent rights under ment health spending so that sufficient funds are available TRIPS led to concerns that new drugs would become even to subsidize activities that genuinely warrant a subsidy less affordable to developing countries. In November 2001 (such as medicines against communicable diseases) and to WTO ministers announced the Doha Declaration on scale back spending on programs of lower priority or pro- TRIPS and Public Health, stating, "We affirm that the grams for which government spending is not indicated or Agreement can and should be interpreted and implemented is poorly targeted. Donor finance to cover recurrent drug in a manner supportive of a WTO member's right to pro- costs should be time limited, tied to policies that support tect public health and, in particular, to promote access to the development of a robust and sustainable pharmaceuti- medicines for all." Under this interpretation of article 31(f ) cal system, and linked to support for the reallocation of of the TRIPS Agreement, countries with a public health government spending.TheWorld Bank's recent Kyrgyzstan crisis can forgo patent law and issue a compulsory license to 124 | The Millennium Development Goals for Health a local manufacturer. But since most developing countries PUSH AND PULL FOR R&D TO DEAL WITH lack the domestic capacity and technical expertise to manu- NEGLECTED DISEASES One way out of this impasse is facture on-patent pharmaceuticals, interpretation of what to separate R&D from drug manufacturing and sales.89 this meant became the subject of a global pharmaceutical Industrial countries, donors, or foundations would commit policy debate. In August 2003 it was agreed that countries to purchase--for a sizable fee--any patents resulting from without the capacity to manufacture medicines could still the development of a major new vaccine or drug and to use compulsory licensing by contracting-out agreements make the patent freely available to drug manufacturers.This with firms in other countries. So far the provision has been would increase the incentive to engage in R&D on diseases invoked only to increase the supply of antiretroviral therapy for which effective demand for drugs is limited (push), globally. while allowing the drug to be sold at marginal cost imme- The August 2003 provision should encourage competi- diately after approval (pull).Another approach is the global tion and make antiretrovirals more affordable to the develop- public-private partnership known as the International ing world,but the administrative intricacies of implementing AIDS Vaccine Initiative (box 7.14). And Médecins sans the compulsory licensing provision are likely to prove cum- Frontières is leading a global initiative that would encour- bersome for most developing countries.An additional route age research and development of drug therapies for most to making antiretrovirals affordable is theAcceleratingAccess neglected diseases by developing countries themselves. Initiative (box 7.13). References 1. Upadhyay,A. 2003."Nursing ExodusWeakens Developing Box 7.13 The Accelerating Access Initiative World." Inter Press Service News Agency. www.Ipsnews.Net/ for antiretroviral drugs Migration/Stories/Exodus.Html. 2. Lindelow, M., P. Serneels, andT. Lemma. 2003."Synthesis of The Accelerating Access Initiative, launched in May 2000 by Focus Group Discussions with HealthWorkers in Ethiopia." five UN agencies and five pharmaceutical companies,87 aims to World Bank,Washington, DC. address the high cost of antiretroviral drugs and to increase 3. U.S.Agency for International Development. 2003. The Health access to HIV/AIDS care and treatment. The initiative was Sector Human Resource Crisis in Africa:An Issue Paper. structured into working groups for country support, commu- Washington, DC. nications, and procurement. UN agencies offered technical 4. World Health Organization. 2003. GlobalTB Control. 2003 support on planning for the care and support of people living Report. Geneva. with HIV/AIDS, particularly on how to increase access to 5. Kurowski, C., S.Abdulla, and A. Mills. 2003. Human Resources for antiretroviral drugs.The countries then negotiated with phar- Health: Requirements and Availability in the Context of Scaling-Up Priority Interventions.A Case Study fromTanzania. London School maceutical companies in discussions facilitated by the UN of Hygiene &Tropical Medicine. agencies. 6. World Bank. 2002. The Gambia:A Country Status Report on By May 2002, 80 countries had expressed an interest in the Health and Poverty.Washington, DC. initiative, and 39 countries had developed or were developing 7. Blanchet, N., G. Dussault, and B. Liese. 2003."The Human national plans to improve the care of people living with Resource Crisis in Health Services." Background Paper to the HIV/AIDS. Also by May 2002, 19 countries had reached 2004 World Development Report: Making ServicesWork for Poor agreements with the pharmaceutical companies to supply their People.World Bank,Washington, DC. antiretroviral drugs at significantly reduced cost.The countries 8. Hammer, J., I. Bnabi, and J. Cercone. 1995."Distributional agreed to waive import taxes and duties on the drugs. Some Effects of Social Sector Expenditures in Malaysia 1974­89." In Public Spending and the Poor:Theory and Evidence, ed. K. Nead. countries have also introduced generic antiretroviral drugs at Baltimore: Johns Hopkins University Press. competitive prices. 9. Lavy,V., J. Strauss, D.Thomas, and P. DeVreyer. 1996."Quality of By December 2002 the cost of the drugs offered by the Care, Survival and Health Outcomes in Ghana." Journal of pharmaceutical companies in the Accelerating Access Initiative Health Economics 15 (3): 333­357. had decreased substantially, in some cases to 10­20 percent of 10. Thomas, D.,V. Lavy, and D. Strauss. 1996."Public Policy and their price in industrial countries. In the 19 countries that Anthropometric Outcomes in the Côte d'Ivoire." Journal of reached agreements with the pharmaceutical companies, Public Economics 61 (2): 155­192. 27,000 people had gained access to antiretroviral treatment, 11. Diop, F.,A. Soucat,A.Wagstaff , and F. Zhao. 2002."Health and almost 10 times the number of patients treated before the the Poor in Senegal." Draft chapter for Senegal Country agreements. Despite these efforts, however, less than 6 percent Economic Memorandum.World Bank,Washington, DC. of people in developing countries who could benefit from 12. World Health Organization. 1998. CHD 1996­97 Report. antiretroviral treatment are receiving it.88 Geneva. 13. Ministry of Health and SocialWelfare. 2002. RNTCP Source: World Bank staff. Performance Report, India, 3rd Quarter. CentralTB Division, Delhi. Tackling Human Resource and Pharmaceutical Constraints | 125 Box 7.14 Global public-private partnerships to accelerate the introduction of new vaccines A company's decision to invest in developing and commercializ- development of specific vaccines (HIV/AIDS, malaria, acute res- ing a vaccine is based largely on an evaluation of economic fac- piratory infection, and diarrheal disease).To move forward with tors: costs, risks, and timing of investments, and expected returns. credible partnerships, both partners must understand the costs, To influence these economic factors, governments are seeking risks, and benefits in order to identify the costs and risks that are new ways to assess and share risk and accelerate the development sensitive to public sector support. And both public and private and introduction of priority vaccines. They are using "push" partners must be confident that the agreements defining the mechanisms to reduce the risks and costs of investments and partnership protect both of their interests--more rapid develop- "pull" mechanisms to ensure future returns. The benefits and ment, expanded capacity, and lower prices on one side and real constraints of each mechanism, and examples of how they have financial commitments that cover investments on the other. been used, are shown in the table. The impact of partnerships The role of partnerships Public-private partnerships are expanding the number and quality It is unlikely that adequate resources can be raised to finance all of vaccine candidates in the pipeline and accelerating late-stage the costs of vaccine development and production (100 percent development activities (ensuring clinical trials in developing coun- push) or to purchase vaccines at prices equivalent to those paid in tries,influencing production capacity decisions,working with gov- the United States or Europe (100 percent pull).So the public sec- ernments to collect data for national decisionmaking). The tor must leverage its resources by targeting as directly as possible International AIDS Vaccine Initiative (IAVI) and the Malaria the obstacles inhibiting a vaccine's progress. If the risks are linked Vaccine Initiative (MVI) have supported these efforts. More than to scientific uncertainty, as for an HIV/AIDS vaccine, push 75 percent of IAVI's resources are slotted for vaccine development mechanisms may prove more valuable than pull mechanisms. If and a number of vaccines are targeted for trials in 2003 in such the risks stem from the market, as for the meningococcal A con- countries as China, India, Kenya, South Africa, and Uganda.The jugate vaccine,pull mechanisms become more important. MVI currently has nine vaccine development projects in its portfo- Public-private partnerships allow for sharing the risks and lio,with clinical trials under way inAfrica for two of those projects. costs of developing and introducing priority vaccines in novel ways. Several public-private partnerships are accelerating the Source: References 90­95. Push and pull mechanisms to accelerate the introduction of new vaccines Mechanism Benefits Constraints Examples Push · Reduces risk and cost of invest- · Requires "picking a horse" early · Direct financing: provides funds to ment and thus may spur or in the process implement activities critical to vaccine accelerate product development · Difficult to estimate value and development (clinical trials) · Well-known policy tool with negotiate some return from · Facilitating environment builds clini- proven track record industry cal trial capacity and helps gain gov- · Credible to industry since · No promise of a successful ernment support money is offered upfront outcome · Tax credits on R&D Pull · Reduces risk to manufacturer, · Locked in even if results are unfa- · Increased uptake of existing vaccines thus creating incentives for vorable (if commitments are made (Global Alliance for Vaccines and private investment early, the public sector may be Immunization and the Vaccine Fund): · Public sector not forced to locked in to a suboptimal provides five-year commitments to select one product; instead, outcome) governments for purchase of vaccines competition encouraged · Mechanisms untested and risky · Copayments: fixed copayment (less · Public funds not committed · Promised return may be too than price) guaranteed unless a vaccine is developed distant and risky to spur · Market guarantee: commits to pur- · If structured correctly, highly investment chase a vaccine if it is developed credible to industry 126| The Millennium Development Goals for Health 14. World Bank. 2003. World Development Report 2004: Making 36. Reinikka, S. 2003."Working for God? Evaluating Service ServicesWork for Poor People.Washington, DC:World Bank. Delivery of Religious Not-for-Profit Health Care Providers in 15. World Bank. 2003."Briefing Note:A GlobalTrust for Health: Uganda." Policy ResearchWorking Paper 3058,World Bank, Revitalizing theWorkforce." Health, Nutrition, and Population Washington, DC. Department,Washington, DC. 37. Peters,D.H.,A.S.Yazbeck,R.Sharma,G.Ramana,L.H.Pritchett, 16. Raufu,A. 2002."Nigeria Concerned over Exodus of Doctors and A.Wagstaff.2002.Better Health Systems for India's Poor: and Nurses." British Medical Journal 325: 65. Findings,Analysis and Options.Washington,DC:World Bank. 17. Federation for American Immigration Reform. 2002."Brain 38. Awases, M.,A. Gbary, and R. Chagtora. 2003. Migration of Health Drain." www.Fairus.Org/Immigrationissuecenters/ Professionals in Six Countries:A Synthesis Report. Geneva:World Immigrationissuecenters.Cfm?ID=1242&C=17. Health Organization. 18. Marcelo, R. 2003."Hospital `Angels' Look for Heaven 39. Srinavichakron, S. 1998."Conditions, Constraints, and Elsewhere." Indian FinancialTimes, October 13. Strategies Increased Contribution of General Practitioners to the Health System inThailand." Human Resources for Health 19. Padarath,A., C. Chamberlain, D. Mccoy,A. Ntuli, M. Rowson, Journal 2 (1) 48­59. and R. Loewenson. 2003."Health Personnel in Southern Africa: Confronting Maldistribution and Brain Drain." 40. Rabinowitz, R. 1999."A Program to Increase the Number of EQUINET Discussion Paper 3. Family Physicians in Rural and Underserved Areas: Impact after 22Years." Journal of the American Medical Association 281 (3): 20. Adversario, P. 2003."Philippines Suffer from Hemorrhage of 255­260. Nurses." ManilaTimes, April 21. 41. Rosenblatt, R.A., M.E.Whitcomb,T.J. Cullen, D.M. Lishner, 21. Adversario, P. 2003."Quality of Nursing Education and L.G. Hart. 1992."Which Medical Schools Produce Rural Deteriorating." ManilaTimes, April 22. Physicians?" Journal of the American Medical Association 268 (12): 22. Adversario, P. 2003."Confusing PoliciesWorsen Outflow of 1559­1565. Nurses." ManilaTimes, April 23. 42. Becker, G.S. 1993. Human Capital:ATheoretical and Empirical 23. Zarilli, S., and C. Kinnon. 1998. InternationalTrade in Health Analysis, with Special Reference to Education. 3rd ed. Chicago: Services:A Development Perspective. Geneva: United Nations University of Chicago Press. Conference onTrade and Development/World Health 43. Tulloch, J. 1999."Integrated Approach to Child Health in Organization. Developing Countries." Lancet 354 (Suppl. 2): SII16­ SII20. 24. Buve,A., S.D. Foaster, C. Mbwili, E. Mungo, N.Tollenare, and 44. Gove, S. 1997. Integrated Management of Childhood Illness by M. Zeko. 1994."Mortality among Female Nurses in the Face of Outpatient HealthWorkers:Technical Basis and Overview.WHO the AIDS Epidemic:A Pilot Study in Zambia." AIDS 8 (3): 396. Working Group on Guidelines for Integrated Management of 25. Tawfik, L., and S. Kinoti. 2001. The Impact of HIV/AIDS on the the Sick Child. Bulletin of theWorld Health Organization 75 Health Sector in Sub-Saharan Africa:The Issue of Human Resources. (Suppl. 1): 7­24. The HRM Response to the Impact of HIV/AIDS on the Health 45. Weber, M.W., E.K. Mulholland, S. Jaffar, H.Troedsson, S. Gove, Sector. USAID,Washington, DC. and B.M. Greenwood. 1997."Evaluation of an Algorithm for 26. Holmas,T.H. 2002."Keeping Nurses atWork:A Duration the Integrated Management of Childhood Illness in an Area Analysis." Health Economics 11 (6): 493­503. with Seasonal Malaria inThe Gambia." Bulletin of theWorld 27. Thornton, J., and F. Esposto. 2003."How Important Are Health Organization 75 (Suppl. 1): 25­32. Economic Factors in Choice of Medical Specialty?" Health 46. Perkins B.A., J.R. Zucker, J. Otieno, H.S. Jafari, L. Paxton, S.C. Economics 12 (1): 67­73. Redd, B.L. Nahlen, B. Schwartz,A.J. Oloo, C. Olango, S. Gove, 28. Wilbulpolprasert, S. 1999."Inequitable Distribution of Doctors: and C.C. Campbell. 1997."Evaluation of an Algorithm for Can It Be Solved?" Human Resources Development Journal 3 (1): Integrated Management of Childhood Illness in an Area of 2­22. Kenya with High MalariaTransmission." Bulletin of theWorld Health Organization 75 (Suppl. 1): 33­42. 29. Wilbulpolprasert, S. 2002."Integrated Strategies toTackle Inequitable Distribution of Doctors inThailand: Four Decades 47. Simoes, E.A.,T. Desta,T.Tessema,T. Gerbresellassie, M. Dagnew, of Experience." and S. Gove. 1997."Performance of HealthWorkers after Training in Integrated Management of Childhood Illness in 30. Organisation for Economic Co-operation and Development. Gondar, Ethiopia." Bulletin of theWorld Health Organization 75 2002. Geographical Imbalances of HRHC: Size, Determinants, and (Suppl. 1): 43­53. Policy Responses.Working Party on Social Policy, Human Resources for Health Care Systems, Paris. 48. World Health Organization, Division of Child Health and Development and Regional Office for Africa. 1997."Integrated 31. Setiadi, G. 1999. Discussion of S.Wibulpolprasert,"Inequitable Management of Childhood Illness: FieldTest of the Distribution of Doctors: Can It Be Solved?" Human Resources WHO/UNICEFTraining Course in Arusha, United Republic for Health Development Journal 3 (1): 2­22. of Tanzania." Bulletin of theWorld Health Organization 75 (Suppl. 32. Zurn, P. 2002."Imbalances in the HealthWorkforce: Briefing 1): 55­64. Paper."World Health Organization, Geneva. 49. Schellenberg J., and the MCETanzaniaWorking Group on the 33. "Nursing Nursing Back to Health," Lancet 352, no. 9124 2001 Health Facility Survey. 2001. Report of the Health Facility (1998): 249. Survey Submitted to the Department of Child and Adolescent Health 34. Carvel, J. 2001."Inflation: Busting Rise for GPs and Nurses." and Development.World Health Organization, Geneva. Guardian, December 18. 50. Ministry of Health, Bolivia,World Health Organization, 35. World Bank. 2003."Bolivia: Health Sector Reforms in the BASICS/USAID, and Sociedad Boliviana De Pediatria. 1999. Context of Decentralization." 26140­BO.Washington, DC. Report of the Health Facility Survey in Bolivia. Geneva. Tackling Human Resource and Pharmaceutical Constraints | 127 51. Kelley E.,C.Geslin,S.Djibrina,and M.Boucar.2001. 71. Reinikka, R. 1999."Using Surveys for Public Sector Reform." "Improving Performance with Clinical Standards:The Impact of PREM Notes.World Bank,Washington, DC. Feedback on Compliance with the Integrated Management of 72. Mcpake,B.,D.Asiimwe,F.Mwesigye,M.Ofumbi,L.Orthenblad, Childhood Illness Algorithm in Niger,West Africa."International P.Streefland,andA.Turinde.1999."Informal EconomicActivities Journal of Health Planning and Management 16 (3):195­205. of Public HealthWorkers in Uganda:Implications for Quality and 52. Oliveira-Cruz,V., C. Kurowski, and A. Mills. 2003."Delivery of Accessibility of Care."Social Science and Medicine 49 (7):849­865. Priority Health Services: Searching for Synergies within the 73. Ablo, E., and R. Reinikka. 1998." Do Budgets Really Matter? Vertical versus Horizontal Debate." Journal of International Evidence from Public Spending on Education and Health in Development 15 (1): 67­86. Uganda." Policy ResearchWorking Paper 1926,World Bank, 53. Black, R.E., S.S. Morris, and J. Bryce. 2003."Where andWhy Washington, DC. Are 10 Million Children Dying EveryYear?" Lancet 361 (9376): 74. Bird, C. 2003."Impoverished Georgia Harbours Fatal Form of 2226­2234. TB." Guardian August 9. 54. World Bank. 2002. Tanzania:A Country Status Report on Health 75. World Health Organization. 2001. The Impact of Implementation and Poverty.Washington, DC. of ICH Guidelines in Non­ICH Countries. Geneva. 55. Quick, J. 2002."Essential Medicines:Twenty-FiveYears on 76. Ratanawijitrasin, S., and E.Wondemagegnehu. 2002. Effective Closing the Access Gap." Bulletin of theWorld Health Organization Drug Regulation:A Multicountry Study. Malta. Geneva:World 80 (11): 913­914. Health Organization. 56. Schrag, S., B. Beall, SF. D. 2001."Resistant Pneumococcal Infections: 77. World Health Organization. 2004. Country Profiles.WHO The Burden of Disease and Challenges in Monitoring and Controlling Regional Office for Europe (Hungary). www.Euro.Who.Int/ Antimicrobial Resistance.WHO/CDS/CSR/DRS/2001.6. ed. Pharmaceuticals/Topics/20020226_1. World Health Organization, Geneva. 78. Andajaningsih. 1997."Implementation of Essential Drug 57. Newton, P.N., N.J.White, J.A. Rozendaal, and M.D. Green. Concept and Drug Financing Strategies."World Health 2002."Murder by Fake Drugs." British Medical Journal 324 Organization, Geneva. (7341): 800­801. 79. Santoso, B., H. Shein, and S. Suryati. 1997."Financing Drugs in 58. World Health Organization and Health Action International. South East Asia." Health Economics and Drugs DAP Series(8). 2003. Medicine Prices:A New Approach to Measurement. Geneva: 80. Government of Guatemala. 2002. Evaluación del sector farmacéu- WHO. tico en Guatemala. Ministerio de Salud Pública y Asistencia 59. Kishuna,A. 2003."Drug Pricing Survey in Kwazulu-Natal." Social, Dirección de Regulación y Control de Programas de la Essential Drugs Monitor (32): 4­7. Salud. Report is in Spanish. 60. World Health Organization. 2003. Essential Drugs Monitor. 81. Bennett, K., and E. Ngalande-Banda. 1994."Public and Private Geneva:WHO. Roles in Health:A Review and Analysis of Experience in Sub- 61. Wagstaff,A., and E.Van Doorslaer. 2003."Catastrophe and Saharan Africa." SHS Paper 6.World Health Organization, Impoverishment in Paying for Health Care:With Applications Division of Strengthening Health Systems, Geneva. toVietnam 1993­98." Health Economics 12 (11): 921­933. 82. Kaplan,W. 2003."Local Production: Industrial Policy and 62. Wagstaff,A., and M. Pradhan. 2003."Evaluating the Impacts of Access to Medicine:An Overview of Key Concepts, Issues, and Health Insurance: Looking beyond the Negative."World Bank, Opportunities for Future Research." Prepared forWorld Bank Washington, DC. meeting on the role of generics and local industry in attaining 63. World Bank, SIDA,AusAID, Royal Netherlands Embassy, and the MDGs in pharmaceuticals and vaccines (June 24, 2003). Ministry of Health ofVietnam. 2001. Vietnam. Growing Healthy: 83. Pavignani, D. 1999."Managing External Resources in A Review ofVietnam's Health Sector. Hanoi:World Bank. Mozambique: Building New Relationships on Shifting Sands?" 64. Hogerzeil, H. 2002."WHO Department of Essential Drugs and Health Policy and Planning 14 (3): 243­253. Medicines Policy:The Concept of Essential Drugs and the 84. UNAIDS andWorld Health Organization. 2002. AIDS WHO Model List of Essential Medicines." Presentation to the Epidemic Update 2003. Geneva. www.unaids.org/en/ Expert Committee on the Use of Essential Drugs,April,World default.asp/. Health Organization, Geneva. 85. Pan American Health Organization. 2002."AIDS Dugs Prices 65. Médecins Sans Frontières. 2001. Fatal Imbalance:The Crisis in Drop 54 Percent in Latin America." www.Paho.Org/ Research and Development for Drugs for Neglected Diseases. English/DPI/Pr020718.Htm. 66. Olivera-Cruz,V., K. Hanson, and A. Mills. 2001."Approaches to 86. Rockefeller Foundation, Center for Pharmaceutical Overcoming Health Systems Constraints at the Peripheral Management, and Management Sciences for Health. 2002. Level:A Review of the Evidence."WG5 15. CMHWorking Regional Pooled Procurement of Drugs: Evaluation of Programs. Paper Series, Geneva. Arlington,VA.. 67. Santoso,B.,S.Suryawati,and J.E.1996."Prawaitasari Small Group 87. World Health Organization and UNAIDS. 2002. Accelerating Intervention vs.Formal Seminar for Improving Appropriate Access Initiative:Widening Access to Care and Support for People Drug Use."Social Science Medicine 42 (8):1163­1168. Living with HIV/AIDS. Progress Report, June. Geneva. 68. Brumburgh, S., and S. Raja. 2001. Ghana: Process Mapping: First 88. UNAIDS. 2003."Access to HIVTreatment and Care Fact Step to Reengineering the Health Supply Chain of the Public Sector Sheet." Geneva. System. Arlington,VA: Deliver/John Snow, Inc. for USAID. 89. Weisbrod, B. 2003."Solving the Drug Dilemma." Washington 69. World Bank. 1994. Better Health in Africa : Experience and Lessons Post, August 22. Learned.Washington, DC:World Bank. 90. International AIDSVaccine Initiative. 2003. Improving and 70. John Snow, Inc. 2000. Programs that Deliver: Contribution to Better Accelerating the Clinical Pipeline of AIDSVaccine Candidates for Use Health in Developing Countries. Arlington,VA: Family Planning Worldwide:The IAVI Research and Development Agenda Logistics Management/John Snow, Inc. for USAID. 2002­2004. NewYork. 128 | The Millennium Development Goals for Health 91. MalariaVaccine Initiative. 2003. MVIVaccine Development 94. Performance Innovation Unit, UK Cabinet. 2001. Tackling the Projects. http://www.malariavaccine.org/ab-current_projects. Disease of Poverty. London. www.pm.gov.uk/output/ htm. page4162.asp. 92. Kremer, M. 2000. Creating Markets forVaccines. Part II: Design 95. Batson,A., and M.Ainsworth M. 2001."Obstacles and Issues. Cambridge, MA: National Bureau of Economic Solutions: Understanding Private Investments in HIV/AIDS Research. Vaccine Development." Bulletin of theWorld Health Organization 93. Madrid,Y. 2001. A New Access Paradigm: Public Access to Assure 79 (8): 721­728. Swift Global Access to AIDSVaccines. NewYork: International 96. World Bank. 2003. World Development Indicators 2003. AIDSVaccine Initiative. Washington, DC. Tackling Human Resource and Pharmaceutical Constraints | 129 CHAPTER 8 Strengthening Core Public Health Functions Chang Sun's wife is HIV positive. So is his mother. So is his tions. By the peak--around 1995--Henan had become the aunt. So are his cousin and his cousin's wife. So is the woman nation's blood farm.The system had been adapted so that vil- next door and, probably, so is her husband. In fact, it is quite lagers could give such huge amounts of blood without suffering possible that almost every adult and many of the children in anemia.After extracting plasma from each 800 cc donation, the his small, remote village are infected.Among them is Chang's collectors would pump 400 cc back into the arms of the donors. father, who died of AIDS last year, and his three-year-old It is believed that people's blood often got mixed up in this way, daughter, who succumbed the year before that. His first wife is spreading HIV to almost everyone involved. there too--she threw herself down the village well in 2000 "Almost everybody did it," said Chang's cousin, Ming. after a doctor told her she was no longer worth treating because "We would sell extra if there was a marriage ceremony coming she had the virus. up or if we wanted to build a house.The most I ever did was This is Xiongqiao village in Henan Province, the ground four donations in a single day." zero of arguably the world's worst HIV/AIDS epidemic, with The consequences for China will be devastating as many up to a million people infected in this single province through a infected villagers are migrating to work in Beijing and other big vast, largely unregulated blood-selling operation.The situation cities. is already a catastrophe, but the risks are growing. GUARDIAN, October 25, 2003 It was almost inevitable that the outbreak occurred in The Chinese government has started providing free treatment Henan. Here in the most populous and impoverished of for poor people with HIV and AIDS and plans to expand the China's provinces, life is cheaper than almost anywhere else in program next year until every poor person who has tested posi- the world.The average Henan farmer survives on 80¢ per day. tive is receiving medical help, a top health ministry official said Henan's officials turned to almost their only untapped resource: in a speech this week. the blood of the province's 90 million population. Vans were NEWYORK TIMES, November 8, 2003 converted into mini-clinics and driven out into the countryside. Ambitious peasants established themselves as "bloodheads" This account from ground zero of one of the world's (brokers) to meet the demand among both buyers and sellers. major AIDS epidemics and the 2002 outbreak in China For an 800 cc donation,villagers were paid 45 renminbi (about of severe acute respiratory syndrome (SARS) serves as a $5.50),enough to feed a family for a week.Realizing that they forceful reminder of the importance of strong public could get far more for milking their veins than for tending the health systems. Vulnerable populations need to be land, they lined up day in and day out for years to make dona- empowered,protected from risks,informed and educated, 131 and encouraged to participate in health activities. Public sectoral actions are government responsibilities that must health regulations need to be established and enforced. be discharged as part of the Millennium Development Infrastructure needs to be in place to reduce the health Goal effort. Good practices in these areas need to be impact of emergencies and disasters.All of this needs to be carefully analyzed and multiplied. It is important for done through a public health system that is transparent national leaders and donor agencies to appreciate the and accountable. necessity of investing in these core public health func- Thus governments have responsibilities that go tions, which are often overlooked as politicians and fun- beyond providing and financing health services in attain- ders focus exclusively on service delivery. It will not be ing the Millennium Development Goals. Strengthening possible to go to scale with basic services for HIV/AIDS, core public health functions such as policy formulation, tuberculosis, malaria, or maternal and child health with monitoring and evaluation, disease surveillance, provider faltering or absent public health infrastructure and weak and insurance regulation, social mobilization, and cross- public health functions. managing health services (chapter 6), pharmaceutical pol- What is public health? icy (chapter 7), and health promotion and behavior change The previous chapters emphasize the importance of gov- (chapter 5).Several others are discussed in this chapter. ernments building strong policies and institutions across the Market failure and public health health sector.Without such improvements,increases in gov- ernment health budgets in most countries will have little In each of these areas of public health is at least one element impact on Millennium Development Goal outcomes. of "market failure."This means that without some form of Governments play a role in developing and implementing government involvement, a free market would produce an policies to lower the barriers that households--especially outcome that is inefficient. In many of these areas are ele- poor ones--face.Households,as users of health services and ments of "externalities" and "public goods." Immunizations, producers of health, are key actors in the health system. for example,benefit not only the immunized person but also Governments also play a role in improving the quality and whole communities, by reducing the risk of others getting efficiency of service providers, which are also key actors. infected.The same is true of the use of insecticide-treated They can do so through a combination of measures to nets (ITNs) and antimalarials. Externalities also provide an improve management in provider organizations and important rationale for the public health agenda to concern improve the accountability of provider organizations to the itself with intersectoral issues. For example, investments in public, whether directly to patients and community organi- water and sanitation infrastructure enable households to zations or indirectly through interactions with policymak- improve their hygiene, reducing their risk of contracting ers. Because provider organizations rely on inputs--human communicable diseases and also the risk to others. resources and medicines--government policies aimed at Communicable disease programs--surveillance, preven- increasing stocks of human resources and improving their tion, treatment and control--are classic examples of public distribution and quality are crucial.Policies to get medicines goods:benefits to some do not diminish the benefits to oth- to providers and ensure that providers face the right incen- ers, and it is typically not feasible to exclude specific people tives for prescribing medicines are also important, as are from benefiting from such programs. Data collection and global policies to ensure appropriate levels of research and analysis and monitoring and evaluation are other examples development in "neglected" diseases and affordable medi- of public goods. In some cases the rationale for government cines for poor countries. involvements stems from informational asymmetries: the These responsibilities are vital elements of the govern- patient, having less knowledge than the provider, can be ment's stewardship role in the health sector. But beyond vulnerable to exploitation by unscrupulous providers (chap- them lies another set of essential responsibilities of the pub- ter 7).This asymmetry goes beyond medicines. It arises, for lic sector: those in public health (table 8.1).1,2 Many of example,in blood contamination,as in Henan Province. these responsibilities are examined in other chapters of the Government involvement is required to deal with these report, including financing health services (chapter 9), market failures. Externalities call for government subsidies, 132 | The Millennium Development Goals for Health Table 8.1 Public health responsibilities and functions Collection and dissemination of evidence for public Human resource health policies, Prevention and Intersectoral action development and capacity Policy development strategies, and actions control of disease for better health building for public health Public health regulation Health situation moni- Surveillance and control Environmental protec- Development of policy, plan- and enforcementa toring and analysisa of risks and damages in tion and health, includ- ning, and managerial Evaluation and Research, development, public healtha ing road safety, indoor capacitya promotion of equitable and implementation of Management of commu- air pollution, water and Human resources access to necessary innovative public health nicable and noncommu- sanitation and disease development and training in health servicesa solutionsa nicable diseases vector control in infra- public healtha structure, management Assurance of the quality Provision of information Health promotiona of medical wastes, Community capacity building of personal and to consumers, providers, Behavior change tobacco legislation, population-based health policymakers, and interventions for disease school health, and servicesa financiers prevention and control education Health policy formulation Development of health Social participation and and planning information and empowerment of Financing and management systems citizens in healtha management Research and evaluation Reducing the impact of of health services emergencies and disas- Pharmaceutical policy, ters on healtha regulation, and enforcement a. Pan American Health Organization, World Health Organization, and the U.S. Centers for Disease Control "essential" public health function.2 Source: Reference 1. since without them,caregivers fail to take into account the they lack the immediate urgency and concrete appeal of dis- external benefits (the benefits to others) of their actions. ease-specific programs and thus tend to be neglected.That Too few people would be immunized. Surveillance activi- their influence on health outcomes is difficult to measure ties, research and development, and monitoring and evalu- also contributes to their frequent neglect by governments ation would be inadequate. Indeed, in some cases, without and donors.* It is difficult,for example,to say just how much government intervention none would be produced at all. a strong disease surveillance and reporting system con- For activities involving externalities and public goods, tributes to reducing the incidence of tuberculosis, whereas governments need not undertake the activities themselves, the benefits of distributing and consistently using drugs for but at a minimum they need to provide some financing treating tuberculosis are obvious. Still, where these core and close oversight,to ensure an acceptable level of perfor- functions are well performed, countries tend to have greater mance, given the information asymmetries. Monitoring a success in achieving their health targets. nongovernment supplier can be difficult, so governments often undertake these activities themselves in the belief-- Why do core public health functions matter sometimes mistaken--that government officials are less for the Millennium Development Goals? likely to exploit their informational advantages over The core public health functions complement vertical pro- patients. grams for achieving disease-specific goals (see table 8.3 at Obstacles to performing public health functions the end of this chapter). Investing in these functions helps build capacity across all of the Millennium Development Developing country governments generally recognize that Goals,whether the challenge is HIV/AIDS,malaria, tuber- these public health functions are important--for making culosis, or communicable diseases in childhood. progress toward the health- and nutrition-related Millen- nium Development Goals and for health policy in general-- *It has also been difficult to define and estimate the cost of implementing but they often lack the capacity and financial resources to core public health functions, although the Bank and its partners (the Pan implement them. Indeed, few low-income countries invest American Health Organization, the U.S. Centers for Disease Control and in these public health functions.Because these functions deal Prevention) are trying to do so as part of national health accounts and public with broad health system issues rather than specific diseases, expenditure reviews. Improving Service Delivery |133 Of the many potential public health functions, four are key: Box 8.1 Increasing the supply of and demand for insecticide-treated nets · establishing national strategies for disease prevention, treatment, and control; Several large-scale efficacy trials have demonstrated that the reg- ular use of insecticide-treated bednets reduces child deaths by · installing government-led monitoring and evaluation some 25 percent.The Roll Back Malaria Global Partnership has systems through integrated disease surveillance, pro- adopted the widespread use of insecticide-treated nets (ITNs) by gram assessment, and collection and analysis of demo- children and pregnant women as a core strategy for reaching its graphic and vital registration data; goal of halving malaria illness and death by 2010. · establishing and strengthening national institutions In many countries, insecticide-treated nets are available only in a small number of commercial outlets in major cities. The and local capacities; and nets are often of poor quality, and the cost is high ($5­$15 · taking intersectoral actions that go beyond the remit each). Insecticide retreatment kits are rarely available. Even of ministries of health. where activities to boost demand have been successful, com- mercial markets have not always responded by scaling up supply. Establishing national strategies for Governments and their partners have used several approaches disease prevention, treatment, and control to increase demand and supply.The most common have been social marketing and public-private partnerships. In both, gov- By employing skilled public health professionals, govern- ernments and donors have supported demand-creation activities ments can effectively monitor the health of communities and provided subsidies for the purchase of nets,through voucher and broader populations, develop and enforce standards, schemes or sales of subsidized products. Eighteen countries have and emphasize health education, public information, also waived or significantly reduced taxes and tariffs on nets, health promotion, and disease prevention. Public action insecticides,and treated bednets. can improve consumers' knowledge and change attitudes These approaches have increased the accessibility and use of so that private markets can operate effectively to meet the ITNs.InTanzania a social marketing scheme stimulated local tex- needs of the poor. Examples are social marketing of ITNs tile manufacturers to produce high-quality bednets, while the to reduce malaria transmission (box 8.1) and social mar- government drastically reduced taxes and tariffs on these prod- keting of condoms for protection against HIV/AIDS. ucts. The resulting competition in the marketplace caused the price of treated bednets to drop from almost $15 to $2, making Installing government-led monitoring the treated bednets accessible to a much larger segment of the and evaluation systems population. Use of ITNs has increased in some districts from less than 2 percent to more than 25 percent. Integrated disease surveillance, program assessment, and Source: World Bank staff. collection and analysis of demographic and vital registra- tion data are essential if governments and donors are to determine whether policies and programs are having an impact on the health Millennium Development Goals. Development Goal. Cuba has made the Millennium Box 2.1 in chapter 2 lists intermediate indicators and Development Goal for reducing child mortality a national proxy indicators for the Millennium Development Goals priority and established a surveillance system to monitor and that can help monitor progress, assess the impact of poli- follow up each child death. Some governments, including cies, and adjust programs going forward. Much greater the Dominican Republican, are developing or modifying investments are needed in systems to monitor these inter- their monitoring and evaluation framework to focus on the mediate indicators. Disease surveillance helps determine Millennium Development Goals (box 8.3). whether health outcomes are improving.Some good prac- Establishing and strengthening national tices in surveillance are being developed (box 8.2). institutions and local capacities Not all developing countries can afford to invest in the infrastructure required for strong surveillance systems. Most Ideally, national institutions that implement core public rely on alternative short- to medium-term solutions for data health functions include a national center for disease sur- gathering, such as intermittent household surveys, health veillance, associations of health professionals with a strong facility surveys, and simplified facility-based routine report- licensing arm, associations of health providers with strong ing systems.In a few cases countries have made special efforts licensing and accreditation arms, a drug safety and regula- to improve surveillance for specific interventions, such as tory agency, a health insurance regulatory body, interagency tuberculosis treatment or immunization, while others task forces for such multisectoral issues as road safety and attempt to monitor progress toward a specific Millennium tobacco control,and others.These institutions and capacities 134 | The Millennium Development Goals for Health Box 8.2 Reducing communicable diseases Box 8.3 Monitoring progress toward the through disease surveillance in Brazil Millennium Development Goals in the Dominican Republic The Brazil Disease Surveillance and Control Project exhibits a best practice in surveillance and public health partnership. It is Improvements in health failed to keep up with economic devel- implemented by the Ministry of Health in collaboration with opment in the Dominican Republic over the past decade. In the U.S. Centers for Disease Control and Prevention, the CDC 2000 maternal and infant mortality rates were higher than in Foundation,the United Nations Development Programme,the other countries in the Caribbean Region at similar levels of Pan American Health Organization, and the World Bank. A development. Coinciding with the endorsement of the Millen- main objective is to strengthen the national public health sur- nium Development Goals,the Dominican legislature passed leg- veillance system. islation that paved the way for substantial reforms of the health The project trains staff at the municipal, state, and federal system. Within the framework of a provider-purchaser split, a levels in epidemiology,disease prevention and control,environ- single-source National Health Insurance Plan will remove finan- mental health surveillance, laboratory reporting and safety, and cial barriers to access, while a reimbursement system with con- management. International technical assistance focuses on tracted autonomous provider networks provides competition improving field investigations of epidemic outbreaks and pro- that empowers consumers and should lead to improved service viding support for operational research in public health surveil- quality. Reforms will be rolled out gradually, beginning with the lance and use of surveillance data for decisionmaking.A public poorest province. The objective is to reduce infant, child, and health laboratory network is being rehabilitated, expanded, and maternal mortality in line with the Millennium Development equipped, while the telecommunications system is being Targets. upgraded for rapid and effective data reporting and use.Across In an Adaptable Program Loan to support implementation sectors environmental health is supported by the legal system. of the reform program, the government and the World Bank The water quality surveillance system has been implemented in agreed on the pivotal role of monitoring and evaluation. more than half the states. Analytical work established chains of cause and effect, linking Public health surveillance and disease control activities con- reforms with intended outcomes. The exercise resulted in a tributed to reductions in such communicable diseases as rabies, comprehensive monitoring framework that disentangled inputs, Chagas disease, and cholera and to dramatic declines in malaria outputs, and outcomes. Key outputs and outcomes were incor- between 2000 and 2002. porated as triggers and performance indicators in the loan framework. For example, the government committed to a 30 Source: References 1 and 3. percent reduction in infant, child, and maternal mortality in designated provinces by the end of the first four years of pro- gram implementation, when 30 percent of health expenditure are entirely lacking or in short supply in many low-income will be channeled through the new financing arrangements. countries. The government and the World Bank agreed that all efforts With these institutions and capacities in place,a country to monitor and evaluate progress will be financed under the can more readily adapt to changes in its health profile and Health Reform Support Project. Monitoring and evaluation deal with new challenges as they arise, including the range capacity in the Executive Commission for Health Reform will of health challenges related to the Millennium Develop- be strengthened. A computerized health management and ment Goals.The recent rapid response to severe acute res- information system will be established during phase one.At the piratory syndrome (SARS) in Vietnam demonstrates the end of this phase, loan proceeds will cofinance the 2005 Demographic and Health Survey to ensure the timely avail- critical role of a strong public health system. Successful ability of high-quality data. Annual evaluations of the project management of the disease outbreak in Vietnam was due will be conducted by an impartial outside entity. not only to high-level political commitment but also to strong surveillance, enforcement of regulations to isolate Source: World Bank staff. infected individuals and protect the public from further exposure to the SARS virus, and cross-sector collabora- tion with officials responsible for water and sanitation, mechanisms to address public health crises and lower the education, and local government services. Drawing on this fatality rate should the disease reemerge. experience,China's strategy following the SARS epidemic Additional lessons on how countries can strengthen the will be to improve SARS­related diagnosis, clinical man- institutions responsible for core public health functions agement, and infection control. It also aims to strengthen can be drawn from a review of public management of the capacity of the public health system to prevent and these functions. Box 8.4 identifies some of the key lessons. control infectious disease--improving surveillance and One of the best examples of institutional changes result- case-reporting systems and setting up alert and response ing in improved health outcomes is the often mentioned Improving Service Delivery | 135 Box 8.4 Core public health functions and the case for public management A recent review of the main themes of public management literature and because local governments have little incentive to invest drew lessons on how public management relates to the core public in public goods and may neglect them. Core public health health functions.Some of the key lessons include the following: functions should either remain under central control--with managerial autonomy or other strategies to permit local · Curative services, preventive services, and core public health adaptation and responsiveness--or if already decentralized functions have distinct properties requiring different policy should be subject to alternative forms of central oversight prescriptions. and control,such as grants-in-aid or earmarking. · Because core public health functions have characteristics of · Since core public health functions are heavily influenced by public goods, user fees are not appropriate. rules and norms in the broader institutional environment that · Promoting competition among agencies responsible for cannot be addressed through training alone, efforts to build public health functions does not improve efficiency. On the management capacity should include components in addition contrary, it may hamper collaboration and technical assis- to training. Public sector norms and rules that impede effec- tance, compromising the effectiveness of such activities as tive administration should be changed whenever possible. surveillance and health promotion. · Provider incentives are difficult to design for core public · Contracting works for some services but not for others. For health functions. For incentives to be useful, measurement preventive services that are measurable and discrete, such as indicators should be chosen carefully. Incentives should be immunization or campaign-based programs, contracting team based or network based rather than individualized and can be effective. But since measurement of core public should include nonfinancial benefits. Performance improve- health functions is complex, expensive, and requires strong ment can also be achieved in more traditional ways--say, by information systems, contracting imposes transactions and implementing merit-based selection and promotion criteria monitoring costs that make efficiency gains unlikely and and clear job descriptions. can reduce effectiveness. · Increasing hierarchical accountability within the public · If introduced cautiously and with instruments to ensure con- health system is critical for strengthening core public health sistency across units and jurisdictions, managerial autonomy functions. Doing so requires changes in the capacity, auton- can be an important way of promoting adaptation and inno- omy, and behavior of service managers, and it requires vation in the core public health functions. monitoring systems and instruments. Monitoring instru- ments need to strike the right balance between simplicity · Decentralizing the core public health functions can be a and complexity and should be designed for operational risky strategy, because of the importance of central coordi- rather than research use. nation, oversight, and technical assistance for these functions Source: Reference 4. case of Céara, Brazil (see box 6.5). Several government Roads and transport, water, hygiene, and sanitation, indoor actions there were important. Substantial investment in air pollution, and agriculture require multisectoral activities public information about public health services led com- with the potential to contribute to the Millennium munities to hold elected mayors accountable for these ser- Development Goals for health and nutrition. vices in their area. Careful control over selected human resource issues reduced patronage at the local level.A phased ROADS AND TRANSPORT Improving transport of approach to implementing programs allowed early adopters intermediate services reduces poverty not by directly increas- to be nurtured to success and news about "success stories" ing consumption of transport but by improving the quality to spread widely, influencing nonadopting areas.And there and security of access to work, markets, and services, accord- was strong state government involvement in motivating ing to a recentWorld Bank report.5 Better transport and roads frontline health workers. can reduce delays that contribute to maternal deaths. In Tanzania 63 percent of women who died after reaching a Taking intersectoral actions--going beyond hospital had traveled 10 kilometers or more for treatment.6 In the remit of ministries of health India a study found that half of maternal deaths occurred A review of the evidence for key determinants of the health before the women reached a treatment facility.7 and nutrition Millennium Development Goals identifies Several developing countries have improved access to significant potential for intersectoral synergies (table 8.2). health services using locally available resources for emer- 136 | The Millennium Development Goals for Health Table 8.2 Potential for intersectoral synergies to achieve the Millennium Development Goals for health and nutrition Millennium Development Goal Target Multisectoral inputs Country examples Key sectors Reduce Improve access to emer- Availability of transport, Bangladesh, Tanzania, Transport, road maternal gency obstetric care roads, and referral facilities Vietnam, infrastructure, health mortality Reduce indoor air pollution Improved cooking practices, China, Guatemala, India, Energy, housing, fuel, and ventilation Kenya health Reduce child Reduce diarrheal diseases Improved hand-washing Costa Rica, El Salvador, Water and sanitation, health mortality in children through hand- practices, using soap and Guatemala, Ghana, India, private sector (soap washing, use of latrines, plenty of water Nepal, Peru, Senegal manufacturers) and proper disposal of young children's stools Reduce indoor air pollution Improved fuel, ventilation, China, Guatemala, India, Energy, housing, private and childplay practices Kenya, Mongolia, health, private sector Nicaragua (improved stove production) Reduce hunger Regulate food prices, raise Improved agricultural prac- Bangladesh, India, Kenya, Agriculture, rural and improve women's income, and tices, tariffs, and trade; Tanzania, Vietnam development, gender, trade nutrition promote dietary diversity reduced women's workload; and food security at the better gender relations and household level intrahousehold decisionmaking Source: World Bank staff. gency transport and communication. Better transportation nated by handling. And constructing water supply and can increase the ratio of health facilities to population.8 sanitation facilities is not enough to improve health out- Several community-driven development programs have comes--sustained human behavior change must accom- demonstrated that women's access to health services can pany the infrastructure investment. be improved by making arrangements with local trans- So what can the public health sector do? In collabora- porters and by organizing emergency interest-free loans tion with other sectors, it can develop public health pro- financed and managed by the communities.9 Preliminary motion and education strategies. It can work with agencies evidence from Mali, where referral funds managed by that plan, develop, and manage water resources and those local health communities financed a system of radio calls responsible for monitoring water quality and sanitation, and ambulances, shows an increase in emergency referral and it can provide leadership for action on hygiene educa- rates from 1 percent to 3 percent.10 tion. It can also provide other sectors with reliable data on A 10-year study in Rajasthan, India, showed that better water-associated diseases and the effectiveness of interven- roads and transport helped women reach referral facilities tions--and advocate for water, sanitation, and hygiene but that many women continued to die because there interventions in poverty reduction strategies. And it can were no corresponding improvements at the household work with the private sector to manufacture,distribute,and and facility levels.7 This illustrates the need to improve promote affordable in-home water purification solutions performance in all relevant sectors to achieve results on and safe storage vessels. the Millennium Development Goals. Hand-washing is one of the most effective interventions for reducing diarrhea. Measurable reductions in diarrhea- WATER, HYGIENE, AND SANITATION Better associated child mortality have been achieved through hygiene (hand-washing) and sanitation (use of latrines, safe public-private partnerships to promote hand-washing.12The disposal of children's stools) are at least as important as Bank's water sector is working with a consortium to drinking water quality to health outcomes, especially the improve the water supply questions on standardized health reduction in diarrhea and associated child mortality.11 surveys used in the developing world (such as the Increased quantity of water has been shown to have greater Demographic Health Surveys and Multiple Indicator impact than improved quality of water,11 possibly because Cluster Surveys) to gather more accurate information on the an adequate supply of water increases the feasibility of links between water and health.The Bank's health,nutrition, adopting safe hygiene behaviors and reduces the length of and population and water sectors are implementing a joint time that water must be stored and may become contami- work program with the World Health Organization to Improving Service Delivery | 137 update scientific evidence on water, sanitation, hygiene, and low-cost technologies,modern fuel alternatives,and renew- health links to inform policies at national level. able energy solutions.14 In China the health sector initiated a large community-based project to reduce indoor air pol- INDOOR AIR POLLUTION Most indoor air pollution lution in rural areas, after policymakers became troubled by in developing countries is caused by the use of low-cost, the leveling off in child mortality rates among the rural traditional energy sources, such as coal and biomass poor. Having already achieved high immunization rates and (wood, cow dung, crop residues), in primitive stoves for other child health interventions, China is now seeking ways cooking and heating--the main source of energy for some to reduce risk factors beyond the health sector.The health 3.5 billion people.The health burden from indoor air pol- sector is generating the data, assessing the impact of inter- lution is greatest in high-altitude rural areas among poor ventions, and promoting behavior change that will reduce families who use biomass in primitive stoves without exposure to indoor air pollution and complement the pro- proper ventilation. Indoor air pollution is a major risk fac- vision of hardware,such as improved stoves and ventilation. tor for pneumonia and associated deaths in children and for lung cancer in women who are at risk of exposure AGRICULTURAL POLICIES AND PRACTICES Agri- during cooking. cultural policies and practices that affect food prices, farm Eventually, most developing countries will move up the incomes, diet diversity and quality, and household food energy ladder, but this move is delayed by low income and security also require government leadership to achieve bet- limited access to high-quality fuel. Improved biomass stoves ter Millennium Development Goal outcomes.Agricultural have been effective in improving health outcomes in India policies that focus on women's access to resources (land, and elsewhere.13 Large community-based intervention trials training, agricultural inputs); their role in production; and are documenting the affordability, cost-effectiveness, feasi- their income from agriculture are likely to have a greater bility, and sustainability of multisectoral interventions. impact on nutrition than policies that do not focus on Studies in China, Guatemala, and India are under way to women, particularly if combined with other strategies, such improve access to efficient and affordable energy sources as strategies for improving women's education and effecting through local design, manufacturing, and dissemination of behavior change communication.15,16 Box 8.5 Strengthening the health system and core public health functions to combat HIV/AIDS To achieve the Millennium Development Target for HIV/AIDS, · ensured access to voluntary counseling and testing massive efforts are needed to strengthen health systems.In the most · capacity to recognize and manage common HIV­related severely affected countries, the health sector suffers from severe illnesses and opportunistic infections shortages of human and financial resources.Many health sector ser- vices are struggling to cope with the growing impact of · reliable laboratory monitoring services, including routine HIV/AIDS. In Sub-Saharan Africa, people with HIV­related ill- hematological and biochemical tests for the detection of nesses occupy more than half the hospital beds, overwhelming drug toxicity as well as access to facilities for monitoring the health services and other organizations providing care and support. immunologic and virologic parameters of HIV infection And just as demand for health services increases, more and more · ensured supply of good-quality drugs healthcare personnel are becoming infected with HIV/AIDS and unable to work.Added to this is the ongoing attrition of healthcare · sufficient resources to pay for treatment on a long-term basis workers for other reasons and weakened infrastructure. · information and training on safe and effective use of anti- In many African countries healthcare systems also suffer from retroviral drugs for health professionals in a position to pre- chronic shortages of drugs, infrequent equipment maintenance, scribe antiretroviral therapy inadequate logistical support,and weak supervision.Also lacking are procedures or systems to monitor and evaluate the quality of health- · establishment of reliable regulatory mechanisms to prevent care and to ensure that providers are accountable to clients.All this misuse and misappropriation of antiretroviral drugs needs to change, not only to meet the HIV/AIDS treatment target Only through investments in strengthening of health systems and but to achieve the health Millennium Development Goals. core public health functions will these conditions be met. Because of the high cost of antiretroviral drugs, the complex- ity of the regimens, and the need for careful monitoring, specific Source: References 17­19. services and facilities must be in place in order to introduce anti- retroviral therapy: 138 | The Millennium Development Goals for Health 4. Khaleghian, P., and M. Das Gupta. 2004."Public Management What do governments and donors need to and the Essential Public Health Functions."World Bank Policy do next to improve public health? ResearchWorking Paper,Washington, DC. 5. World Bank. 2002."Transport." In A Sourcebook for Poverty To accelerate progress toward the Millennium Development Reduction Strategies, 325­362.Washington, DC. Goals, most countries need to significantly increase invest- 6. Biego, G., ed. 1995. Survey on Adult and Childhood Mortality, ments in the core public health functions,in addition to pro- Tanzania. Calverton, MD: Macro International. viding and financing health services. Strengthening these 7. Pendse,V. 1999."Maternal Deaths in an Indian Hospital:A functions should be included in the Poverty Reduction Decade of (No) Change?" Reproductive Health Matters (Special Issue on Safe Motherhood Initiatives). Strategy Paper process and public expenditure framework 8. Samai, O., and P. Sengeh. 1997."Facilitating Emergency exercises. Many existing health partnerships and initiatives Obstetric Care throughTransport and Communication, Bo, could contribute more to strengthening public health func- Sierra Leone." International Journal of Gynecology and Obstetrics 59 tions by helping build surveillance capacity. In India aWorld (Suppl. 2): 157­64. Bank­supported $146 million immunization project was 9. Eissen, E., D. Efenne, and K. Sabitu. 1997."Community Loan Funds andTransport Services for Obstetric Emergencies in instrumental in strengthening disease surveillance over the Northern Nigeria." International Journal of Gynecology and succeeding two years and beyond.The project spurred devel- Obstetrics 59 (Suppl. 2): S37­S46. opment of a new national surveillance strengthening pro- 10. Debrouwere,V., R.Tonglett, andV.W. Lerbergh. 1998. gram, now being implemented with additional World Bank "Strategies for Reducing Maternal Mortality in Developing financing. Countries:What CanWe Learn from the History of the IndustrializedWest?" Tropical Medicine and Hygiene 100: Unless low-income countries strengthen such core 771­782. public health functions as surveillance, policy formulation, 11. Esry, S., J. Potash, and L. Roberts, and C. Schiff. 1991 "Effects of and program monitoring and evaluation, they will find it ImprovedWater Supply and Sanitation on Ascariasis, Diarrhea, difficult to deliver the basic health interventions to fight Dracunculiasis, Hookworm Infection, Schistosomiasis, and Trachoma." Bulletin of theWorld Health Organization 69 (5): diseases and achieve the Millennium Development Goals 609­621. for health and nutrition.The key constraints to progress on 12. Saade, C., M. Batemen, and D.B. Bendahmane. 2001."The HIV/AIDS,for example,are weak government systems for Story of a Successful Public-Private Partnership in Central policy analysis, disease surveillance, and monitoring and America."Arlington,VA: BASICS II, EHP, UNICEF, USAID, andWorld Bank. ineffective policies for human resources in health. To 13. Hughes, G., K. Lvovsky, and M. Dunleavy. 2000. achieve their goals, national HIV/AIDS programs need to "Environmental Health in India: Priorities in Andhra Pradesh." strengthen these core public health functions rather than World Bank, Environment and Social Development Unit, focus exclusively on prevention, condoms, testing and Washington DC. counseling services, and treatment for HIV and oppor- 14. World Bank. 2003."Public Health at a Glance Fact Sheet: tunistic infections (box 8.5). Indoor Air Pollution." Health, Nutrition, and Population Department,Washington, DC. 15. Quisumbing,A.R. 1995."Gender Differences in Agricultural References Productivity:A Survey of Empirical Evidence." FCND Discussion Paper-IFPRI 15. 1. World Bank. 2002. Public Health andWorld Bank Operations. 16. Johnson-Welch, C. 1999."Focusing onWomenWorks: Human Development Network,Washington, DC. Research on Improving Micronutrient Status through Food 2. Pan American Health Organization. 2000. National Level Based Interventions." International Center for Research on Instruments for Measuring Essential Public Health Functions. Public Women (ICRW)/Opportunities for Micronutrient Health in the Americas. PAHO/Centers for Disease Interventions (OMNI),Washington, DC. Control/Centro Latino Americano de Investigaciones en 17. World Health Organization. 2000. Use of AntiretroviralTreatments Sistemas de Salud,Washington, DC. in Adults with Particular Reference to Resource Limited Settings. 3. Khaleghian, Peyvand, and Monica Das Gupta. 2004."Public Geneva:WHO. Management and Essential Public Health Functions." Policy 18. World Health Organization. 2003. Global Health Sector Strategy ResearchWorking Paper 3220,World Bank Development, for HIV/AIDS 2003­2007: Providing a Framework for Partnership Research Group,Washington, DC. econ.worldbank.org/ and Action. Geneva:WHO. working_papers/33192/. 19. World Bank. 1994. Better Health in Africa: Experience and Lessons Learned.Washington, DC:World Bank. Improving Service Delivery |139 Table 8.3 Examples of public health functions and infrastructure requirements for preventing and controlling communicable diseases Acute respiratory infection, diarrhea, Public health function HIV/AIDS Tuberculosis Malaria and measles Developing policy and strategy Adopting and reinforcing Providing HIV testing, Controlling drug quality Removing taxes and tar- Regulating antidiarrheal legislation ensuring confidentiality, and the monitoring of iffs on essential drugs use and protecting the rights private sales of drugs commodities (ITNs, of people living with Notifying people with insecticides, drugs) HIV/AIDS tuberculosis when their cases are detected Promoting equitable Targeting vulnerable Providing free diagnosis Authorizing mid-level and Encouraging community- access to health services groups (often poor or and treatment community workers to based prevention and marginalized) with Increasing community- provide effective treatment HIV/AIDS services based treatment options treatment in the Monitoring the cost of Addressing gender community Addressing gender services and establishing barriers barriers Adjusting or removing protective mechanisms Contracting NGOs and user fees for people at for poor families community-based high risk (such as children organizations (which are under five) often more effective in reaching high-risk groups) Developing quality Implementing protocols Ensuring adequate Balancing the needs of Establishing guidelines standards and norms in voluntary counseling supervision and the health sector in and standards for clinical and testing, treating accountability based on discussions on civil service care for pneumonia and opportunistic infections, monitoring and evaluation reform diarrhea reducing mother to child and performance-based Contracting out health Evaluating clinical perfor- transmission, and contracting services mance providing antiretrovirals Crafting malaria Providing training and treatment policies incentives to service providers Strengthening supervision and monitoring and evaluation Developing national Developing strategies that Establishing disease Incorporating malaria Adopting integrated con- strategies for disease take into account the control within sector control policies in strate- trol of acute respiratory control specifics of the situation programs gic plans for the health infections, diarrhea, and (stage of the epidemic, sector measles and incorporating high-risk groups, capacity, it into the overall child resources) health policy and plan (Integrated Management of Childhood Illness) Providing finance and Calculating the cost of a Ensuring financing for Including malaria control Including child health pre- resource management package of key HIV/AIDS drug supplies; account- in sector budgets ventive and curative ser- interventions and ensur- ability of health personnel; Expanding allocation for vices in the overall sector ing its financing strong central normative, nonsalary recurrent budget surveillance; and a moni- expenditures commensu- toring and evaluation unit rate with need Developing policies for Adopting a policy on Developing efficient plan- Monitoring the availability Monitoring drug rapid approval of drugs antiretrovirals ning, procurement, and of insecticides and effec- resistance, particularly for and monitoring the supply of quality-assured tive antimalaria drugs pneumonia and dysentery quality of drugs drugs and monitoring pri- in the marketplace vate sales, where feasible 140 | The Millennium Development Goals for Health Table 8.3 Examples of public health functions and infrastructure requirements for preventing and controlling communicable diseases (continued) Acute respiratory infection, diarrhea, Public health function HIV/AIDS Tuberculosis Malaria and measles Measuring results: Collecting and disseminating evidence for policies, strategies, and actions Conducting health Implementing second- Applying a standardized Using existing survey tools Conducting routine moni- situation monitoring generation surveillance quarterly system of (such as the Demographic toring of health facility and analysis, including recording and reporting and Health Survey and the data surveillance on case detection and Multiple Indicator Cluster Conducting periodic outcomes Survey) to monitor household surveys of Conducting periodic sur- outcomes and impacts immunization coverage, veys, where feasible Conducting population- oral rehydration therapy based surveillance coverage, and acute respi- Making decisions based ratory therapy treatment on data Conducting disease surveillance Conducting research and Fostering partnerships to Fostering partnerships to Strengthening partnerships Conducting operational development of conduct operational pursue operational with research and technical research on new delivery innovative solutions research on AIDS vaccines, research and extending partners to conduct mechanisms, such as com- new drugs, and new access to new technolo- operational research on munity management of preventive tools gies and strategies efficacy and effectiveness acute respiratory of new drugs, insecticides, infections and preventive tools Strengthening partnerships for new pneumonia vaccine devel- opment and distribution Providing information to Providing information to Expanding information, Disseminating best Using information, educa- consumers, providers, policymakers on the education, and communi- practices tion, and communication policymakers, and socioeconomic costs of cation programs and Providing information, programs and social mar- financiers HIV/AIDS, the costs of social mobilization efforts education, and communi- keting techniques to pro- scaling up interventions, to increase case detection cation that promote mote appropriate and the cost-effectiveness and demand for services behavior change behaviors, such as hand- of different interventions within communities and washing using key stakeholders Strengthening health Strengthening health Integrating standard TB Using health information Providing health workers information and manage- information systems, reporting into regular systems data to monitor at the most peripheral ment systems surveillance, HIV/AIDS health systems without performance of health level with feedback of program monitoring and losing required depth of workers and facilities results of health informa- evaluation, and their data for decision making tion systems linkages Conducting research Evaluating HIV/AIDS pro- Ensuring periodic evalua- Conducting operations Developing standardized and evaluation grams and conducting tion at the local and research on innovative indicators for program intervention research on national levels that methods to improve treat- monitoring and evaluation interrupting modes of increases accountability ment and prevention transmission, epidemio- and motivation to perform practices at the commu- logical risk and burden nity level research, and health sys- Monitoring and evaluat- tems and operational ing programs using stan- research dardized indicators Improving Service Delivery |141 Table 8.3 Examples of public health functions and infrastructure requirements for preventing and controlling communicable diseases (continued) Acute respiratory infection, diarrhea, Public health function HIV/AIDS Tuberculosis Malaria and measles Preventing and controlling diseases Managing communicable Implementing effective Adapting design of case Collaborating with public Managing supervision, diseases interventions with ade- detection, diagnostic, health centers on imple- supply chain, immuniza- quate coverage to make and treatment delivery menting Integrated tion and other child health an impact schemes to local condi- Management of programs at the central, Strengthening links tions and resources Childhood Illness regional, and district levels between HIV/AIDS and Establishing good-quality other diseases control community-based treat- programs (sexually ment capacity through transmitted diseases, NGOs and the private tuberculosis, maternal and sector child health) Supervising and monitor- Ensuring a reliable supply ing public and private of good-quality condoms sector health workers and drugs for treating sexually transmitted infec- tions and opportunistic infections Providing health Providing information, Expanding information, Promoting use of Promoting key family and promotion information, education, and communi- education, and communi- insecticide-treated nets community practices for education, and cation programs for cation programs and social and improving home treat- child health communication high-risk groups as well mobilization efforts to ment practices by as the general population expand case detection and sponsoring behavior demand for services within change campaigns communities using key stakeholders Promoting social Adopting a multisectoral Expanding information, Promoting use of insecti- Promoting social participation and approach to HIV/AIDS education, and commu- cide-treated nets and mobilization for national empowerment of citizens Involving people living nication programs and improving home immunization days and with HIV/AIDS and high- social mobilization efforts treatment practices by community-based child risk groups in HIV/AIDS to expand case detection sponsoring behavior health and nutrition activities and demand for services change campaigns interventions within communities using key stakeholders Reducing the impact of Establishing social support Undertaking routine Rapidly providing preven- Establishing effective sur- emergencies and disasters networks for people living reporting and conduct- tive tools (insecticide- veillance for detection of with HIV/AIDS and AIDS ing periodic surveys to treated nets) outbreaks of measles or orphans identify special risks Establishing readily acces- diarrheal diseases Establishing social safety (emergence of sible treatment capacity common in population nets for poor households multidrug-resistant dis- displacement affected by HIV/AIDS and ease or outbreaks in pris- Establishing effective, AIDS orphans (if possible) ons or health facilities) rapid access to immuniza- tion and treatment Undertaking intersectoral action to improve health Improving environmental Treating medical waste Encouraging improvement Using proven, safe, and Reducing indoor air health generated by HIV/AIDS in conditions (ventilation, cost-effective environ- pollution services space) of households, mental management Forging public-private worksites, prisons, and techniques partnerships to improve refugee camps in order to water and sanitation reduce exposure (hand-washing) and other hygiene behaviors Improving school health Supporting school-based Increasing awareness to Facilitating rapid treat- Promoting water and san- HIV/AIDS education improve case detection ment and access to itation and life skills programs and youth participation preventive strategies through Focusing in treatment supervision through schools Resource(s) on Effective School Health (FRESH) 142 | The Millennium Development Goals for Health Table 8.3 Examples of public health functions and infrastructure requirements for preventing and controlling communicable diseases (continued) Acute respiratory infection, diarrhea, Public health function HIV/AIDS Tuberculosis Malaria and measles Taking intersectoral actions Supporting HIV/AIDS Supporting public-private Institutionalizing health Collaborating with educa- activities in other sectors: collaboration on drug impact assessment for tion, water and sanitation, tourism, transport, labor, supply and drug quality new initiatives and pro- and other sectors on key education control and stimulating jects in other sectors family practices research and development Using appropriate envi- ronmental management and vector control strate- gies in agricultural, water, and infrastructure projects Developing human resources and building capacity for public health Developing policy, Sharing the experience of Ensuring that policy- Establishing capacity for Establishing a comprehen- planning, and managerial success stories makers and donors short-, medium-, and sive human resources capacity Maximizing the role of in- understand the impact long-term planning and development plan with country HIV/AIDS Theme improved distribution, budgeting to support capacity building, follow- Groups in capacity build- retention, and quality of malaria control programs up, retention, and incen- ing for policymakers and primary healthcare work- Integrating planning for tives to work in rural areas planners ers has on outcomes malaria control in health sector planning activities Providing human resource Training health profes- Integrating technical and Ensuring adequate Training epidemiologists, development and training sionals, social workers, managerial training needs numbers of trained health and program managers counselors, peer educa- within larger preservice workers at facilities and in Providing preservice and tors, program managers, and in-service training and communities, including in-service Integrated and epidemiologists supervision efforts through use of contracting Management of with NGOs and private Childhood Illness skills- sector based training Building capacity in Training and funding Ensuring that the benefits Engaging NGOs and Engaging and training communities NGOs and community- to the community of private sector in com- NGO staff and other pri- based organizations in engagement in disease munity-based prevention vate sector actors in deliv- HIV/AIDS programming control efforts are known and treatment activities ering services and and management, and barriers to participa- preventive messages at prevention, care, tion overcome the community level and support Improving Service Delivery | 143 CHAPTER 9 Financing Additional Spending for the Millennium Development Goals--In a SustainableWay Nha's family has 12 members.They used to be one of the richest Mobilizing additional domestic resources--by raising families in the village (in Lao Cai Province,Vietnam).Now they the share of government spending on the health sector or are one of the poorest.They have suffered two shocks in recent raising the share of GDP that is taxed--is worth explor- years.First Nha's father died two years ago.So there are now only ing.But raising additional domestic resources for govern- two main laborers in the family--Nha, 26, and his mother, 40. ment health spending takes time. Even on optimistic Nha has two young children.Two years ago,his daughter Lu Seo assumptions about the growth of the health share of gov- Pao also had a serious illness and had to be operated on in the ernment spending, the tax share of GDP, and economic district and province hospital. Nha's family had to sell four buf- growth, health spending per capita might easily take 10 faloes, one horse, and two pigs to cover the expenses of treatment. years to double in real terms. The operation cost several millionVND, but Lu Seo Pao is still Official development assistance for health can play a not cured.All the people in Nha's community helped,but no one temporary role in financing additional expenditures. It can contribute more than 20,000VND (a little over $1). Nha's increased in the 1990s, but overall development assis- younger brother, Lu Seo Seng, who was studying in grade 6, had tance did not, so development assistance for health to leave school to help his family. If Lu Seo Pao was not ill, says probably will not increase indefinitely. Development Nha, his family would still have many buffaloes, he could have a assistance also entails difficulties, including its volatility. house for his younger brother,and Seng could stay in school.1 That volatility makes it an unreliable source of funding for permanent increases in recurrent expenditures.And Additional health spending is necessary in at least some increased expenditures financed through development countries. How should the extra spending be financed? assistance eventually have to be accommodated within How should the cost be divided between governments each country's budget. The sustainability of increased and donors? Many developing countries spend less than expenditures therefore requires efforts on both domes- they appear to be able to afford. Other countries with tic and external fronts. similar per capita incomes devote larger shares of their Private health spending, whether out-of-pocket or GDP to health spending. Countries that spend less do from private insurance, is another source of funding for so either because they place a lower priority on health health.The importance of private spending as a share of in their public expenditure allocations or because they GDP varies across country income groups, but it is raise smaller shares of GDP in tax and nontax revenues. higher in low-income countries (figure 9.1). 145 In India a study estimated that nearly a quarter of people Figure 9.1 Government health spending is higher admitted to the hospital were above the poverty line when in richer countries, but private health spending is they went in and below it when they came out.6 InVietnam higher in low-income countries health expenses are estimated to have pushed about 3.5 per- 10 30,000 cent of the population into absolute poverty in both 1993 9 and 1998.7 In Cambodia a single hospital stay is estimated to 25,000 8 have absorbed 88 percent of an average household's non- dollars) 7 (%) 20,000 food consumption in 1997; for a household in the poorest 6 (PPP GDP 20 percent, the cost was greater than its entire annual non- 5 15,000 of food consumption.8The risk of large-scale impoverishment 4 income 10,000 3 is clearly greater the poorer the country, since poorer coun- Share 2 tries tend to have larger shares of poor people.9 A country's 5,000 capita 1 Per private share of health spending in GDP may not in practice 0 0 be related to its per capita income. But on poverty-reduc- Low Lower Upper High High income middle middle income income tion grounds,there are good reasons to wish that it were. income income non-OECD OECD Government has a clear role. It must not end up paying Official development assistance the medical bills of people who can afford to spend their to health sector own resources. Instead, it needs to concentrate financing Domestically financed government health spending on essential public goods and other areas where private Private health spending spending is inefficient, target its limited resources on the GDP per capita poor, and use its stewardship capacity and resources to Note: The data for each income group are weighted by population size. leverage private spending at lower levels of care to protect Source: Reference 14. against catastrophic risks through some type of pooling mechanism.Vietnam--a country with one of the highest private spending shares in the world--has recently started doing just this,by providing central government support to Health spending--government the provinces to allow them to enroll poor and other disad- isn't the only player vantaged groups in the national social insurance scheme. Private spending absorbs a larger share of income in poorer countries. In fact, in low-income countries, private health Government spending--what's affordable? spending absorbs a larger share of GDP, on average, than domestically financed public spending. If private spending Government spending is an important part of the picture, were private insurance, this might not matter. But in low- but how much can governments afford to spend? In con- income and lower middle-income countries, private trast to private spending, government spending as a share spending almost invariably means out-of-pocket expendi- of GDP is higher in richer countries (figure 9.1). But at tures, not private insurance.2 In Cambodia, where the pri- any given per capita income, there is a surprising amount vate health spending share of GDP is 6 percent, or in India of variation across countries in the share of GDP allocated and Vietnam, where the share is 4 percent, private health to government health programs. Countries that appear spending is almost entirely out-of-pocket.The same is true able to spend similar shares of GDP on government health of many lower middle-income countries,such as China. programs end up spending quite different amounts. The large share of private spending leaves many near- A regression of the share of GDP absorbed by government poor households heavily exposed to the risk of impover- health spending on per capita income provides some indica- ishing health expenses. In China evidence suggests that tion of what a country could afford to spend on government exposure to the risk of medical expenses3 causes rural health programs (see figure 9.2). Bolivia, whose domestically households to hold more wealth--and to hold more of it financed government health spending accounts for 4.2 per- in liquid form.4This self-insurance is only partly successful cent of its GDP,is considerably above the 2 percent predicted at smoothing consumption when income shocks occur by the regression line. By contrast, Uganda's domestically (due to a variety of factors, including illness).The problem financed government health spending accounts for barely 0.5 is most pronounced for the poor: for the poorest 20 per- percent of GDP,well below its predicted share of 1.4 percent. cent,40 percent of an income shock is passed on to current Countries above the regression line have a stronger case consumption, while the richest 20 percent is protected for additional development assistance than countries from almost 90 percent of an income shock.5 below it. For countries below the regression line, the focus 146 | The Millennium Development Goals for Health Figure 9.2 Some countries spend considerably less than expected on government health programs--and some spend more 8 Croatia 7 GDP) Czech Republic of 6 (percentage Tunisia Colombia Slovak Republic Panama 5 Uruguay Lesotho FYR Macedonia Belarus Argentina spending Solomon Turkmenistan Estonia Lithuania Islands St. Vincent and the Costa Rica Bolivia Grenadines Dominica Poland Saudi Arabia health 4 Guyana Jordan Namibia South Africa Honduras Botswana Mongolia El Salvador Turkey Latvia Samoa Grenada Brazil Lebanon 3 Armenia SwazilandAlgeria Chile government Tonga Papua New Guinea Kazakhstan St. Kitts and Nevis Zambia Mexico Moldova Zimbabwe Jamaica Ukraine Bulgaria Trinidad and Tobago Senegal Oman Thailand financed 2 Burkina Faso GuatemalaFiji Belize MadagascarAngola Gabon Albania Romania Mauritius Comoros Egypt, Arab Rep. Dominican Republic Congo Rwanda Vanuatu Philippines Malaysia Kenya Vietnam Morocco 1 Ecuador Sudan Domestically Azerbaijan Haiti Burundi India Georgia Malawi Indonesia Nigeria Uganda 0 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 Per capita income (PPP dollars) Source: Reference 11. of the health financing debate should initially be on why revenues collected by the government (the general revenue health spending appears to be less than what the country share) and the share of general revenues allocated to the could afford and on finding ways of remedying the situa- health sector (the health share of government spending).10 tion. External assistance in such a setting is not precluded, Low government health spending could be due to either or however. Indeed, countries in that setting could benefit both being low. In poorer countries, both shares are typically from technical assistance and are likely to require financial lower than they are in richer countries (figure 9.3).But there assistance to meet the adjustment costs of mobilizing addi- are differences across countries that cannot be explained by tional resources to move to affordable spending (box 9.1). per capita income alone. Figure 9.4 divides low-income countries into those Reasons for underspending below the regression line in figure 9.2 (those spending less Domestically financed government health spending comes than appears affordable by international standards) and those from general revenues and social insurance contributions.The above the regression line. Countries in quadrant I have amount of general revenues flowing into the health sector above-average values of both government health spending depends on both the amount of general (tax and nontax*) as a share of total government spending and total govern- ment spending as a share of GDP. Not surprisingly, these * Some countries with large public sector companies, especially in the countries are well above the regression line in figure 9.2. Middle East and North Africa, South Asia, and Sub-Saharan Africa, have Countries in quadrant I place a strong emphasis on the important nontax revenues, which can represent as much as 9 percent of health sector in government spending allocations, along GDP. Increasing revenues from this source would require consideration of the competitive environment in which the companies operate and the need with a strong emphasis on government finance in the econ- for reinvestment in the companies. omy in general. By contrast, countries in quadrant III have Financing Additional Spending for the Millennium Development Goals--In a SustainableWay | 147 Box 9.1 Raising tax levels in the developing world--hard but doable Direct taxes include taxes on personal income, corporate profits, the tax burden on a few firms makes raising taxes difficult, espe- payrolls,property,and wealth.Indirect taxes (taxes on transactions cially if the owners are politically influential. If extractive and and commodities) include general sales taxes, value-added taxes, service sectors have been privatized, international corporations excise taxes, turnover taxes, import duties, and export taxes.The are likely to be involved (especially in petroleum and mineral ratio of tax revenues to GDP increases with GDP. Low-income industries), and a high corporate income tax rate risks driving countries raise an average of 14 percent of their GDP through them away. taxation, while lower middle-income countries raise 19 percent Import taxes, especially on luxury goods (tobacco, alcohol, (see table).These averages conceal large variations: a low-income cars), are a politically attractive option. But import taxes can cre- country may raise as little as 4 percent of GDP in taxation ate significant economic distortions, especially if imports have a (Myanmar) or as much as 36 percent (Lesotho), while a lower high capital content. Such taxes may also be hard to collect. middle-income group may raise as little as 9 percent of GDP Bolivia has more than 6,000 kilometers of borders with its five (Guatemala) or as much as 30 percent (Ecuador). neighbors, making it impossible to control illegal imports. The small size of the formal economy is a major constraint to Thousands of people carry all types of products--from alcoholic the use of payroll taxes to collect personal income taxes.In Bolivia, beverages to flour and rice--on their backs through the borders for example,45 percent of the population lives in rural areas,and a each day, depending on the relative prices on each side of the large part of the urban population works in the informal economy. border. As a consequence, only about 30 percent of household income is Increasing revenues through tax reforms requires political paid through payroll taxes.Increasing personal income taxes would commitment and administrative will, but it is possible. Bolivia be a burden only for civil servants and employees of a few compa- raised tax revenues from 3 percent of GDP in 1983 to 17 percent nies and would increase the incentives for transferring activities to by the end of 1987. Facing hyperinflation of 25,000 percent in the informal sector. 1985, Bolivia, with World Bank technical assistance and financ- Increasing tax revenues by raising corporate income taxes ing for adjustment costs, undertook major structural reforms, may also be difficult when there are only a handful of companies including tax reform. It replaced a complex tax system of more reporting income, as is the case when extractive and service sec- than 100 tax rates (differential import tariffs, earmarked taxes, tors are in the hands of the government.A high concentration of progressive personal income taxes, and corporate taxes with Taxes as a percentage of GDP, by income group Total tax Taxes on General Country income group revenue international trade Excise duties sales tax Social security Low-income (less than 14.0 4.5 1.6 2.7 1.1 $760 per capita) Lower middle-income 19.4 4.2 2.3 4.8 4.0 ($761­$3,030 per capita) Upper middle-income ($3,031­$9,360 22.3 3.7 2.0 5.7 5.6 per capita) High-income (more 30.9 0.3 3.1 6.2 8.8 than $9,360 per capita) below-average health shares in government spending and government spending allocations,but they tax relatively large below-average general revenue shares. By international shares of their GDPs. In some cases, the general revenue standards these countries place comparatively little emphasis share is sufficiently large to offset the relatively small health on health and on public finance in the economy in general. share and the country spends more than is expected on the Countries in quadrant II attach importance to health in basis of its per capita income (Angola and Zimbabwe are government spending allocations,but they tax comparatively examples).In other cases--such as Burundi and Eritrea--the small shares of their GDP. As a result, these countries fall relatively high general revenue share is more than offset by below the regression line in figure 9.2. Countries in quad- the relatively small health share and the country spends less rant IV attach relatively little priority to the health sector in than expected on the basis of its per capita income. 148 | The Millennium Development Goals for Health Box 9.1 Raising tax levels in the developing world--hard but doable (continued) many exceptions and loopholes) with a simple structure of six All other taxes, including personal and corporate income taxes: taxes, were eliminated. At the same time, Bolivia established a strong and credible tax administrative structure. · a uniform import tariff rate of 20 percent, reduced to 10 Although far from perfect,the tax reform dramatically increased percent in 1988 revenues and equilibrated oil prices to international standards.The · a 10 percent value-added tax on a broad base, excluding reform has endured for more than 12 years, and taxes now repre- only housing and financial services sent 19 percent of GDP.The government is currently considering a reform that would introduce a progressive income tax. · taxes on consumption of luxury goods (alcoholic beverages, Social expenditures must be accommodated within each perfumes, cosmetics, tobacco, and jewelry) country's budget constraint over a reasonable time span.Although · a 1 percent transaction tax (a cumulative turnover tax) difficult, such accommodation can take place if government com- · a progressive tax on vehicles (1.5­5 percent) and urban real mitment is strong and reforms are well planned, preferably within estate (1.5­3 percent),which was shared with municipalities a Poverty Reduction Strategy Paper and medium-term expendi- ture framework structure. · a 2 percent tax, increased to 2.5 percent in 1988, on the net worth of public and private enterprises (small informal enterprises paid a fixed lump sum instead) Source: World Bank staff. For countries in quadrant II the challenge is to push for Figure 9.3 Health absorbs a higher share of a higher share of general revenues in GDP rather than to government spending, and general revenues raise the health share of government spending, which is absorb a higher share of GDP in richer countries already relatively high by international standards.This can 50 be done. Some countries have achieved high average 45 annual growth of the share of general revenue in GDP, 40 though others have slipped backwards (figure 9.5). Bolivia 35 increased the tax share of GDP at an annual average rate of 30 5 percent during the 1990s, the result of a sustained 25 reform process that boosted the tax share of GDP from 3 Percent 20 percent in 1983 to 17 percent by the end of 1987 (see box 15 9.1).The health sector in Bolivia has benefited from this 10 growth of tax revenues, with government health spending 5 as a share of GDP growing at an annual rate of nearly 10 0 Low Lower Upper High High percent during the 1990s (figure 9.5). Ministries of health income middle- middle- income income rarely interest themselves in tax reform. But Bolivia shows income income non-OECD OECD that this indifference is probably misplaced. Government health expenditures as a share of government expenditures For countries in quadrant IV the challenge is to push for a Government expenditures as a higher share of health in government spending allocations, percentage of GDP since general revenues already represent a relatively high share of GDP by international standards. Several countries Source: Reference 14. raised their government health spending share in the 1990s, often considerably (figure 9.6).A challenge is to demonstrate to an often skeptical ministry of finance that additional Mobilizing extra domestic resources--where you resources in the government health budget are warranted. push depends on where you're coming from One cause of skepticism may be a misplaced perception Countries spending less than seems affordable by interna- that the health sector is necessarily an "unproductive" sec- tional standards and wanting to mobilize additional domestic tor.The perception is misplaced because--as emphasized resources for government health programs face different in chapter 1 and as the recent SARS episodes in Canada, challenges depending on whether they are in quadrants II, China, andVietnam illustrate--better health outcomes are III,or IV. closely linked to economic growth and poverty reduction. Financing Additional Spending for the Millennium Development Goals--In a SustainableWay |149 Figure 9.4 Small health shares and low government revenues cause some low-income countries to spend less than they can afford on health 80 Above regression line Quadrant IV Quadrant I Below regression line 70 Eritrea GDP 60 of Croatia Angola Cape Verde Botswana 50 Zimbabwe Solomon Is. Lesotho Czech Republic percentage St. Vincent and the Grenadines a Bulgaria Hungary as Poland Namibia 40 Brazil Djibouti Belize Estonia Romania Oman Swaziland Bosnia and Herzegovina Latvia Belarus Morocco Bolivia St. Lucia Mongolia Uruguay Central African Rep. AlbaniaFiji Zambia Dominica Jamaica Jordan Tunisia Lithuania FYR Macedonia expenditures 30 Togo Burundi Sierra Leone Saudi Arabia Colombia Niger Russian Federation Panama Slovak Republic Tanzania Moldova Chile Malawi Benin Samoa Rica Azerbaijan Sudan Mauritius Peru Tonga Mauritania HondurasCosta Georgia Vietnam Rwanda Turkmenistan Tajikistan Senegal Kazakhstan Argentina 20 Indonesia Bangladesh Chad Guyana Thailand China Kyrgyz Republic Government Nepal Guinea Madagascar Paraguay India Uganda Nigeria Cameroon Equatorial Guinea Mexico Dominican Republic El Salvador Ecuador Guatemala Haiti Congo, Dem. Rep. 10 Cambodia Quadrant III Quadrant II 0 0 5 10 15 20 25 Government health expenditures as a share of government expenditures Source: Reference 14. Insofar as more health spending leads to better health out- Table 9.1 shows the implications for trends in govern- comes, the health sector may be as "productive" as, say, the ment health spending as a share of GDP and in dollar terms education sector. The issue is not whether better health per capita of different assumptions about the growth of carries economic benefits, which is indisputable--it is government health spending as a share of total government whether more spending will lead to better health out- spending, government revenues as a share of GDP, and comes. Both issues tend to get overlooked by ministries of GDP per capita.The figures in table 9.1 are hypothetical health, which often view the importance of their mission but realistic enough to be worrying. In the pessimistic sce- of improving health outcomes as self-evident and take for nario A, the share of government health spending grows at granted that government programs are achieving their 2 percent a year, government tax revenues at 0.5 percent, intended health outcomes. and GDP per capita at 1 percent.These figures are achiev- able, though many countries have failed to sustain them in Some uncomfortable arithmetic: the past. A country spending 1 percent of its GDP on Raising domestic resources takes time health in 2000 would end up spending 1.45 percent in There are, then, two ways to mobilize additional domestic 2015; if its 2000 level of spending per capita were $10 resources for tax-financed health spending--increase the (about that of Mauritania or Zambia), its 2015 level would share of government spending going to health and be $16.84. Scenarios B, C, and D are progressively more increase the share of GDP that is taxed.Pursuing these two optimistic about how quickly growth would proceed. But avenues, how quickly could a country reach the regression even in the most optimistic scenario, government health line? And what level of spending would it reach? spending in 2015 is only 2.8 percent of GDP, or $32 per 150 | The Millennium Development Goals for Health Figure 9.5 General revenues as a share of GDP Figure 9.6 The share of GDP going to government rose significantly during the 1990s in some health spending during the 1990s rose significantly countries and fell in others in some countries and fell in others Annual percentage change in share of GDP Annual percentage change in share of GDP absorbed by taxes and general revenues devoted to government health spending ­10 ­5 0 5 10 ­20 ­10 0 10 20 Turkey Colombia Dominican Republic Paraguay Argentina Senegal Maldives Syrian Arab Rep. Peru Burkina Faso Bolivia El Salvador Paraguay Uruguay Nepal Estonia Namibia Equatorial Guinea St. Vincent & the Grenadines Bolivia Uruguay Guinea-Bissau Iran, Islamic Rep. Thailand Bhutan Peru Chile Botswana South Africa Taxes Bhutan Madagascar Ethiopia India Haiti Panama Lesotho Mexico Bangladesh Jordan General Belarus Philippines revenues Maldives Tunisia Tanzania Mauritius Ghana Costa Rica Iran, Islamic Rep. Pakistan Ecuador Venezuela, RB Kenya Thailand Mozambique Romania Bulgaria Swaziland Mongolia Sri Lanka Côte d`Ivoire Côte d'Ivoire Romania Bulgaria Moldova Hungary Costa Rica Oman Albania Myanmar Iraq Uzbekistan Source: Reference 14. FYR Macedonia Kyrgyz Rep. Nigeria Marshall Is. Myanmar Nicaragua capita.That 2.8 percent could still be below what appears Georgia to be affordable by international standards.And $32 would Azerbaijan Tajikistan probably be too little to fund a basic health program. That the share of GDP flowing into the health sector Source: Reference 14. cannot be raised overnight10 does not mean that countries that can apparently afford to spend more out of their own resources should not be encouraged to start. Development important in Sub-Saharan Africa: all 12 countries with agencies have a role in helping them do so--through external funding exceeding 35 percent of total health technical support of tax reform, assistance in developing expenditures in 2000 were in Africa.11 government commitment to health in public expenditure Development assistance is not, however, without draw- allocations, and financial assistance to ease the adjustment backs.Although development assistance for health increased costs and provide support while the gap between current in the 1990s,total development assistance did not,and there and affordable spending is being closed. is no assurance that development assistance for health will continue to grow. Industrial countries show no signs of closing the gap between current levels of development assis- Development assistance--a mixed blessing tance and the proposed 0.5 percent of GDP. Judging by the Development assistance tends to account for a larger share past,relying on continually increasing levels of development of government health spending in poorer countries (see assistance for health would not be wise.Furthermore,donor figure 9.1). Development assistance for health is especially financing depends on donor budgets, which are subject to Financing Additional Spending for the Millennium Development Goals--In a SustainableWay |151 Table 9.1 Some uncomfortable arithmetic--how long to raise domestic resources for health? Scenario Item A B C D Assumed annual growth (percent) Government health expenditure as a percentage of 2.00 3.00 4.00 5.00 government expenditure Government expenditure as a percentage of GDP 0.50 1.00 1.50 2.00 GDP per capita 1.00 1.50 2.00 2.50 Implied annual growth (percent) Government health expenditure as a percentage of GDP 2.51 4.03 5.56 7.10 Government health expenditure per capita 3.54 5.59 7.67 9.78 Trend in government health expenditure (percent of GDP) 2000 1.00 1.00 1.00 1.00 2005 1.13 1.22 1.31 1.41 2010 1.28 1.48 1.72 1.99 2015 1.45 1.81 2.25 2.80 Trend in government health expenditure per capita (U.S. dollars) 2000 0.00 10.00 10.00 10.00 2005 11.90 13.13 14.47 15.94 2010 14.15 16.40 18.98 21.93 2015 16.84 21.00 26.14 32.48 Source: World Bank staff. the usual business and political cycles and may go up or Figure 9.7 Importance of external financing in down each year. total health expenditures in selected countries in From a country perspective, development assistance is a Sub-Saharan Africa very volatile source of funding for health.The ratio of exter- 60% nal financing to total health expenditures increased or of decreased yearly in selected countries in South and East 50% Africa that showed improvements in child mortality share between 1995 and 2000--in some cases sharply, as it did in as Somalia (figure 9.7). Obviously, then, it is not prudent for 40% health countries to commit to permanent expenditures for such expenditures for items as salaries for nurses and doctors on the basis of uncer- 30% tain financing flows from development assistance funds. health Donors often require that assistance be kept in parallel assistance 20% budgets outside the ministry of finance, which precludes appropriate planning and targeting of expenditures. In government 2000 off-budget spending was estimated to represent more 10% than 46 percent of health spending in Tanzania and more Development than 50 percent in Uganda.12,13 Although some off-budget 0% spending, such as the resources collected from user fees, is 1995 1996 1997 1998 1999 2000 domestically funded, most is funded by donors, who Burundi encourage this practice in order to facilitate accounting Comoros Eritrea for the direct impact of their resources. Ethiopia But because money is fungible, once external financing Somalia becomes available for health,ministries of finance may sub- Uganda Tanzania stitute donor funds for regular treasury financing of expen- ditures,resulting in only marginal increases in overall health Source: Reference 11. expenditures (chapter 10).This is especially so in countries 152 | The Millennium Development Goals for Health operating under strict budget constraints, where increased tures in health diminishes when economies are in a down- expenditures in many sectors have been suppressed for turn and facing decreased tax revenues (as many countries some time.As resources from abroad become available for in Latin America and the Caribbean are) or when they health, political pressures are likely to divert resources pre- have to reduce the overall nonfinancial public sector deficit viously available to health to other uses. So a simple assess- in order to be fiscally prudent,meet regional commitments ment of the impact of external resources on outputs or the (such as joining the European Union), meet commitments purchase of additional inputs does not take into account to International Monetary Fund programs, or maintain the impact of reduced resources from regular budgets. international credit ratings. Commitments outside the Off-budget expenditures in health may partly explain the health sector also need to be kept in mind. Expenditures low government expenditures in some countries, such as that improve health outcomes--such as infrastructure pro- Uganda.Moreover,off-budget expenditures make it impos- jects crucial for enabling people to get to health facilities-- sible to properly target resources to particular interventions, may even have to be increased. All expenditures must be geographic locations, or population groups. Such targeting considered to avoid excessive leveraging at the country may be essential for improving the impact of expenditures level that would threaten fiscal sustainability (maintaining a on outcomes and for reaching the health Millennium country's debt-to-GDP ratios and liquidity requirements Development Goals,as explained in chapter 4. for debt service). Most important, spending commitments in the health While difficult, the prospects for achieving sustainable sector must be permanent. This means that any external health expenditures are not necessarily bleak.Fiscal sustain- financing must eventually be replaced by additional domes- ability implies working on all fronts of the fiscal deficit: tic revenues or by reallocating expenditures from other sec- raising tax revenues, reallocating expenditures across the tors. Because both these policy measures are difficult to budget, and increasing donor funding in the form of grants implement, countries must carefully analyze the commit- or other forms that may be registered as part of the revenue ments they make to their populations on the basis of tempo- stream.The issues point to a range of tasks that require a rary external financing. joint effort by developing countries and donors. Sustainability of health What needs to be done domestically expenditures--a dim picture? Low spenders need to increase resources for health by rais- A dim picture of the sustainability of additional health ing taxes, raising allocations to health, or both. All coun- expenditures can emerge from some of the main messages tries need to strive for good-quality Poverty Reduction of this chapter and chapter 4: Strategy Papers and medium-term expenditure frame- works.They need to align targets and goals with resources. · Health outcomes are influenced by multiple factors, They need to manage resources through improved public many of them outside the health sector, such as access expenditure management systems, focus resources at the to roads and clean water, and requiring additional margin on programs that will have the largest impact on investment. outcomes, and improve the impact of expenditures on · Not all the Millennium Development Goals are for outcomes through appropriate targeting of interventions health, and countries have many other priorities. and population groups.As indicated in chapter 4, policies and institutions in the health sector need improving. · Raising additional domestic revenues requires increased Countries need to find mechanisms for pooling out-of- capacity and takes time. pocket expenditures and expand the financial envelope · Development assistance has not increased to expected beyond fiscal efforts. levels and may be volatile at the country level, so expenditures financed with such funding must even- What donors need to do tually be accommodated within each country's budget Donors need to make good on the promise of increased constraints. financing. Development assistance needs to be timely and In discussing resource mobilization--whether domestic predictable, so that it can be used to finance carefully or external--a sense of realism needs to be maintained.All planned recurrent expenditures that may eventually be countries face budget constraints.They are more difficult covered by domestic financing. Donors need to improve to live with in low-income countries with limited capacity coordination among themselves, eliminating off-budget to increase tax revenues and relatively inflexible expendi- financing, for example, which inhibits appropriate country tures.The possibility of accommodating increased expendi- budgeting and targeting. Financing Additional Spending for the Millennium Development Goals--In a SustainableWay | 153 7. Wagstaff,A., and E. van Doorslaer. 2003."Catastrophe and References Impoverishment in Paying for Health Care:With Applications toVietnam 1993­98." Health Economics 12 (11): 921­933. 1. Vietnam-Sweden Mountain Rural Development Programme, ActionAid, Save the Children Fund, and Oxfam. 1999. A 8. World Bank, SIDA,AusAID, Royal Netherlands Embassy, Synthesis of Participatory Poverty Assessments from Four Sites in Ministry of Health ofVietnam. Vietnam. Growing Healthy:A Vietnam. Hanoi. Review ofVietnam's Health Sector. Hanoi:World Bank. 2. Musgrove, P., R. Zeramdini, and G. Carrin. 2002."Basic 9. World Bank. 2000. World Development Report 2000/2001: Patterns in National Health Expenditure." Bulletin of theWorld Attacking Poverty.Washington, DC:World Bank. Health Organization 80 (2): 134­142. 10. Hay, R. 2003."The `Fiscal Space' for Publicly Financed Health 3. Akin, J.S.,W.H. Dow, and P.M. Lance. 2004."Did the Care." Oxford Policy Institute Policy Brief. Distribution of Health Insurance in China Continue to Grow 11. World Health Organization. 2002. TheWorld Health Report Less Equitable in the Nineties? Results from a Longitudinal 2002: Reducing Risks, Promoting Healthy Life. Geneva. Survey." Social Science and Medicine 58 (2): 293­304. 12. World Bank. 2003."Health Sector Public Expenditure Review 4. Jalan, J., and M. Ravallion. 2001."Behavioral Responses to Risk Update FY 03."Washington, DC. in Rural China." Journal of Development Economics 66 (1): 23­49. 13. Republic of Uganda. 2002."Public Expenditure Review: 5. Jalan, J., and M. Ravallion. 1999."Are the Poor LessWell Report on the Progress and Challenges of Budget Reform." Insured? Evidence onVulnerability to Income Risk in Rural Republic of Uganda. China." Journal of Development Economics 58 (1): 61­81. 14. World Bank. 2003.World Development Indicators 2003. 6. Peters,D.H.,A.S.Yazbeck,R.Sharma,G.Ramana,L.H.Pritchett, Washington, DC:World Bank. and A.Wagstaff.2002.Better Health Systems for India's Poor: Findings,Analysis,and Options.Washington,DC:World Bank. 154| The Millennium Development Goals for Health CHAPTER 10 Applying the Lessons of Development Assistance for Health Development assistance to the health sector has been procurement only from the country of origin of the increasing, both in real terms and as a proportion of offi- funding.Also needed is a common framework for report- cial development assistance. It can be effective in coun- ing and assessing progress, with a commitment to learn tries with sound policies and institutions. But in from it and make appropriate policy adjustments. countries with weak policies and institutions, it has little Are the lessons of development assistance being impact on health outcomes. Recent work on develop- learned and applied in practice? They appear to be. In ment effectiveness yields other important lessons, too. the World Bank's analytical work, policy dialogue, and Conditionality can work--but only if governments are financial operations over the past 18 months are many committed to the conditions they agree to. Donors can- positive signs of change. not force policies on governments, but they can help The previous chapters point to the many actions design them. And aid is at least partially fungible, with developing country governments can take to accelerate governments shifting their spending patterns in response the pace of progress on the Millennium Development to donor allocations. Goals for health while ensuring that much of the benefit These and other lessons have spurred interest at the accrues to the poorest and most disadvantaged house- World Bank and among other donors in broader devel- holds.At the same time, the many international partners opment assistance mechanisms, such as sectorwide pro- to health and nutrition--bilateral and multilateral agen- grams and poverty reduction support credits, which aim cies, philanthropic organizations, and transnational com- to have countries leading donor coordination, with panies--have a heavy responsibility to assist.This chapter strategic coherence ensured through a Poverty assesses the track record of development assistance to Reduction Strategy Paper. To be effective, the strategy health, outlines the lessons learned, and reviews some should contain a careful analysis of the Millennium recent attempts to improve the effectiveness of develop- Development Goals and be linked firmly to outcomes. ment assistance to health and its contribution to meeting The related financial coherence then needs to be the health goals. It also reviews the Bank's responses to expressed in the medium-term expenditure framework. the Millennium Development Goal challenge--how the A practical lesson for donors is to work more closely Bank's commitment to the goals has influenced its ana- with each other and with governments to ensure that the lytical work, its dialogue with and technical assistance to number of country coordination bodies for health is lim- countries, its country assistance strategies, its lending ited, that resources are pooled, and that aid is untied to activities, and its work on monitoring and evaluation. 155 The review highlights breakthroughs and promising new the Millennium Development Goals for health indeed approaches and identifies continuing constraints to better pose formidable challenges and that it is vital for all to performance. rise to those challenges.They agreed that a greater sense At the High-Level Forum on the Millennium Devel- of urgency is required if 2015 is not to pass by with a opment Goals for health held in Geneva in January 2004, large number of countries having missed the targets.And a wide variety of actors--donors, international technical they agreed on actions to mobilize resources for health, agencies, philanthropists, and developing countries-- to enhance aid effectiveness and harmonization, to agreed broadly on a number of points.They agreed that increase human resources,and to monitor performance. have a greater impact on health outcomes?What are some Development assistance to health-- of the lessons of experience that can be carried forward as what have we learned? the external financing envelope expands? Despite a decline in overall official development assistance, Development assistance to health works-- development assistance to health rose in real terms and as a in a good policy environment proportion of official development assistance in the 1990s (figure 10.1).1 With new funding sources for health in Development assistance to health supports a vast array of 2000­02--including the Bill and Melinda Gates Founda- activities and services, some focused on specific diseases tion; the Global Fund to Fight AIDS, Tuberculosis and (polio, tuberculosis, HIV/AIDS); some on strengthening Malaria; the special U.S. financing for HIV/AIDS; and health systems (disease surveillance, training nurses and mid- World Bank International Development Association (IDA) wives); and some on particular services (reproductive and grants--commitments from all external sources including child health services). Did this assistance actually change foundations rose from an average of $6.4 billion in health outcomes? Recent work from the World Bank sug- 1997­99 to about $8.1 billion in 2002.* And funding can gests that it did.3,4 But development assistance to health does be expected to continue to rise in the coming years.2 not improve health outcomes in countries where the policy What has development assistance to health achieved in environment is poor.4 With "good" policies and institutions recent years? Can this money be spent in a way that will (strong property rights,little corruption,an efficient bureau- cracy),an extra 1 percent of GDP in aid is estimated to have lead to a decline in infant mortality of 0.9 percent. By con- *The International Development Association provides "credits" to the trast, where policies were only average, the decline was only world's poorest countries--loans at zero interest with a 10 year grace period and maturities of 35­40 years. 0.4 percent.And where policies were poor, aid is estimated to have had no significant effect on infant mortality. Corroborating the finding that aid has little or no effect if policies and institutions are poor is theWorld Bank's experi- Figure 10.1 Official development assistance and ence with projects. Numerous evaluation studies, develop- other flows for health are rising ment effectiveness reports, implementation completion ODA and Other Official Flows For Health, reports, and quality reviews demonstrate that the major 1990­2000 (Billion US $) weaknesses in the use of development assistance to health 6.00 arise from institutional issues. Projects are more likely to 5.00 have a development impact if adequate institutional analysis $ 4.00 is undertaken before a health sector operation is approved,if US 3.00 clear institutional lines of responsibility are established, and if implementation is entrusted to a strong team. Billion 2.00 This finding is influencing IDA allocations, which are 1.00 no longer based only on per capita income but also on 0.00 Bank staff assessments of countries' policies and institu- 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 tions. A similar approach is being followed at the subna- Year tional level. The Bank's country assistance strategy for Source: Reference 2. India over the past five years has focused on a few states, such as Andra Pradesh and Karnataka. These states have 156 | The Millennium Development Goals for Health put fiscal reform,governance issues,and sectoral strategy at Aid can help improve health the top of the development agenda--and been rewarded policies--under certain conditions by development partners with new financing for health Conditionality can work--if country commitment is in and other sectors. place.Tying aid to policy changes is a common practice,but But the productivity of aid is not a black-and-white recent studies have cast doubt on the ability of conditions issue.There are graduations of good policy, and as policy to bring about reform.6 If governments are committed to gets better the productivity of aid increases. Improving reform, conditions can help--by enabling governments to governance can enable countries to get more out of health commit publicly to certain reforms and persuade private spending. In Bangladesh, for example, an additional dollar investors of their seriousness. But if governments are not of government health spending spent after improvements committed to reform, conditions will not make them in governance are made is estimated to reduce under-five reform,often because disbursements continue anyway. mortality by 14 percent; without such improvements, Donors cannot force policies on governments--but under-five mortality would fall just 9 percent. Bangladesh they can help design policy. Donors can alert governments has made large strides in reducing child mortality in recent to the reasons for reform and help nurture commitment. years,relying on nongovernmental organizations to deliver But at the end of the day, it is governments that have to many services. Raising the quality of governance from sustain it (box 10.1).Analytical work, training and techni- below average to above average (based on the internation- cal assistance, disseminating ideas about policy reform and ally agreed governance index) would allow it to reap gains more quickly at given levels of public spending.5 development,and stimulating debate in civil society can all be valuable activities for donors to support while the gov- There may be a tradeoff between targeting assistance on ernment's commitment to reform is growing. the neediest countries and achieving the greatest impact from Vietnam in the late 1980s and early 1990s is a good exam- development assistance to health. But if there is, it is not an ple. At a landmark meeting in 1986, the ruling communist all-or-nothing decision:countries need not move from"bad" party decided to break with the past and introduce sweeping policies to"excellent"policies in order for aid to be justified. economic reforms. In the health sector this included intro- And in countries with weak policies, a focus by donors on ducing user fees at public facilities, legalizing private medi- policy dialogue and technical assistance to improve the envi- cine, deregulating the pharmaceutical industry, and opening ronment for development assistance to health can set the the pharmaceuticals and medical equipment industries to stage for a larger infusion of financial support. Box 10.1 The importance of country commitment: How Bangladesh and Thailand have fared in reducing malnutrition Nutrition has not made it to the national political agenda in most approved, but is struggling in the challenging policy environ- countries, because nutrition advocates have not succeeded in ment for social sectors. Contradictory messages from donors and linking improved nutrition with political and economic goals or other partners, and frequent changes in governmental portfolios, creating popular demand to eliminate malnutrition. Can donors have added to the challenge. help? In Bangladesh, UNICEF and the Bank partnered with the InThailand, commitment-building for nutrition was achieved government to present the case that the country could not and nurtured with little external support. It generated commit- achieve its economic goals without reducing malnutrition.They ment for nutrition by building a wide consensus (among the gov- persuaded policymakers that funding a national nutrition pro- ernment, nongovernmental organizations--NGOs--and the gram was a good investment, and a new nutrition project was private sector) around the benefits of nutrition--not as a welfare approved in 1995. But commitment is often fragile, and careful issue but as a human development issue.This initial commitment and continuing efforts are needed to sustain the initial commit- was sustained by ensuring that policy statements were closely ment through changes in political scenarios. linked to national investment plans, building strong technical and The issue is not just how to build initial commitment, but managerial capacity for nutrition in the country (often through how to broaden and maintain commitment and complement it external aid), and linking these actions with a strong buy-in and with systematic investments in institutional capacity develop- demand from communities. Malnutrition rates in Thailand ment.The first nutrition investment in Bangladesh was wound declined from 51 percent in the early 1980s to 18 percent in up in 2001. Child care and health-seeking behaviors improved 1990. substantially in project areas, and malnutrition rates in the coun- try declined. A new follow-on nutrition investment was Source: World Bank staff. Applying the Lessons of Development Assistance for Health |157 international trade. Initially,Vietnam saw no increase in aid. The fungibility in development assistance But institutions such as the United Nations Development to health is substantial Programme and theWorld Bank helped facilitate the reform Much aid is earmarked--both across sectors and within process by organizing international workshops for the them. One part of a development agency gives a grant to Vietnamese to exchange ideas on policy with their neigh- the ministry of health for a health sector reform while bors.This set the stage for a large inflow of donor financing, another does the same for a primary education project. starting about 1995 and continuing to the present. One agency makes a loan to the ministry of health for a Donors can support innovative financing mechanisms to tuberculosis project while another makes a loan for a improve performance by linking disbursements to specific malaria control project. The intention of donors is that performance measures, including better policies.This is the these activities remain tightly sealed--the health sector tack taken by the Global Fund for AIDS,Tuberculosis and reform funds are to be kept separate from the primary Malaria.And it is one way the Global Alliance forVaccines education project funds, and the tuberculosis project funds and Immunization disburses its funds to countries, using a are to be kept separate from the malaria control project per capita payment for each additional child fully vacci- funds.The idea is to ensure that the government makes a nated against a target schedule.Recent programmatic social certain spending choice that it would not have made had adjustment loans to Peru and Brazil by theWorld Bank link it been handed a check for the same amount. disbursement of large single tranches of funding to policy One view of aid is what-you-see-is-what-you-get: a changes targeting public health spending toward the poor. government receives $1 million for a water project, and Performance-based lending is also at the heart of the new the net impact is $1 million worth of extra spending on scheme for IDA credit buy-downs for polio (box 10.2). the water sector.This view has recently been challenged, on the grounds that aid is fungible--at least partially. Where aid is not fungible, the recipient government Box 10.2 Helping eradicate polio through increases spending on the health sector by an amount IDA credit buy-downs equal to the aid. For each dollar of aid to the health sector, government health spending rises by a dollar. The other To ensure financing for the Millennium Development Goals, extreme is where aid is fully fungible.The recipient gov- governments, foundations, agencies, and development banks are ernment treats the extra dollar of aid as if it represented an all exploring new financing approaches that have the potential extra dollar of government revenue, and it increases its to increase resource flows, adjust the concessionality of funding spending in the health sector by whatever amount it where appropriate,and help focus more attention on impact. increases its health spending when its own revenues rise by The IDA credit buy-down mechanism was recently piloted in several projects supporting polio eradication,clearly a global pub- a dollar.The intermediate position is incomplete fungibil- lic good.The mechanism enhances the concessionality of IDA's ity--government spending on health rises by less than a assistance in priority areas, mobilizes additional resources from dollar but by more than it would if the government external partners,and focuses the attention of governments,part- received an extra dollar in its overall budget. ners,and Bank staff on clearly defined performance objectives. Assessing whether aid is indeed fungible is not straight- Two projects--in Nigeria and Pakistan--were implemented forward. The difficulty is knowing how the government in fiscal 2003, in partnership with the Bill and Melinda Gates would have responded if its own resources had increased by Foundation, Rotary International, and the United Nations the amount of the aid, or equivalently if it had received a Foundation. The partnerships will buy down a country's IDA check for the same amount. Recent research suggests that loans on successful completion of the country's polio eradica- while an extra dollar of development assistance does result in tion program. Because of the generous loan terms, each grant an extra dollar of government spending in some countries,in dollar unlocks roughly $2.50 for countries to fight polio. To many it does not.6 Where it does,only 29 cents of the addi- fund the buy-downs,the partnership has established a trust fund with $25 million from the Gates Foundation and $25 million tional dollar of aid goes into government development pro- from Rotary International and the United Nations Foundation. grams--the rest leaks out into nondevelopment programs, This $50 million investment has the potential to buy down though not apparently into reduced tax effort. So official roughly $125 million in World Bank IDA loans. In this way, development assistance at the level of overall government developing countries can mobilize what ultimately becomes spending is fungible--the average government spends an grant funding to eradicate polio--and contribute beyond their additional dollar in exactly the same way, irrespective of borders to the global campaign to eliminate polio. whether it comes from domestic resources or from a donor. Source: World Bank staff. Research also suggests that aid in many countries is fungi- ble across sectors within the government's development 158 | The Millennium Development Goals for Health program--aid intended for the health sector gets spent on identify what their money buys. This recognition has other development sectors,and aid intended for other devel- encouraged donors to search for broader development assis- opment sectors gets spent on health. Fungibility implies, for tance mechanisms that recognize the importance of the example, that when development assistance to health is ear- entire expenditure program.These mechanisms range from marked for primary health services and excludes tertiary the Multi-CountryAIDS Program inAfrica,which supports care,governments simply focus their resources on health ser- national HIV/AIDS strategies, to sectorwide approaches in vices for the population served by public hospitals--a health to poverty reduction support credits, which back a wealthier,urban population in many poor countries. broad public spending agenda.In several countries,including One important implication is that donors should not Ethiopia, Ghana, Mozambique, Senegal, Tanzania and spend time and effort trying to channel their external fund- Uganda, these principles have improved country-level coor- ing to specific programs for certain priority diseases and dination of development assistance to health. Some key ele- populations, without engaging in a dialogue with the gov- ments include the following: ernment on basic changes in the overall patterns of public · ensuring that countries, not donors, drive the coordi- spending for health--the total allocation and the amounts nation allocated to, say, child health and communicable disease control services or improving community and primary level · ensuring strategic coherence, as expressed in the health delivery systems. If these basic changes were enacted, poverty reduction strategy and the health, nutrition, donors would then transfer their financial assistance to the and population analysis that feeds into it health sector as a whole, knowing that they are likely to · ensuring financial coherence through a medium-term have a positive impact on Millennium Development Goal expenditure framework and an agreement that all outcomes. donor funding will respect the overall spending plans and limits of the government The transactions costs of aid are still too high · pooling donor funds in a single account and untying In a single low-income country, more than 20 donors-- aid to procurement only from the country of origin of including bilaterals, multilaterals, global programs, founda- the funding tions, and large NGOs--can be involved in health. The · limiting the number of country coordination bodies demands placed on recipient countries can be huge, as that can bring together national and international donors are starting to acknowledge. They are recognizing actors involved in health that their individual procedures for reporting, accounting, and managing funds--often encompassing different budget · establishing a common framework for reporting and structures,ways of measuring progress toward objectives,reg- assessing progress, with a strong focus on countries ulations for the procurement of goods, services and works, conducting the monitoring and evaluation, learning and approaches and cycles to disbursing funds--place heavy from it, and using the information to make their pro- and unreasonable demands on recipient countries. This is grams more effective particularly so in poor countries forced to allocate limited Some of the experiments in country-level coordination of human resources away from managing service delivery to development assistance to health reveal the difficulties in managing donors. implementing the principles of better donor harmoniza- Donors have also learned that individual project man- tion. In some instances, it has been hard to persuade donors agement units have not made sustainable contributions. to pool their funds and break the tie between funding and They have sometimes run parallel to local structures.They procurement. Monitoring and evaluation systems have have also fostered a sense that the project staff are more sometimes not been strong enough to yield timely and accountable to the financier than the government. And meaningful data on progress, a critically important feature if they have redirected the most qualified human resources disbursements are linked to performance. Multiple national away from government employment toward employment coordination bodies for government,donors,and NGOs for in development assistance agencies. different diseases and services also persist in many settings. There is a need to enhance coordination, explicitly pool Global partnerships add value but involve risks aid, and put countries in the driver's seat Chapter 1 provided a long list of diverse global initiatives There is a growing view that if aid is indeed fungible and and partnerships (box 1.2). Each responds to priorities in earmarking imposes transactions costs on recipient coun- public health in developing countries and to health prob- tries,donors should dispense with the fiction that donors can lems that represent global threats. Applying the Lessons of Development Assistance for Health |159 Are these efforts competing to garner attention on an countries and donors to learn how to capitalize on these already overwhelmed global development agenda? Are global partnerships at the country level to throw light on they fragmenting the complex health challenges facing health system issues and ultimately strengthen the under- societies? Possibly both. But they reflect genuine collective lying human and physical infrastructure. attempts to solve large common problems, after long peri- So far there has been little cross-fertilization across the ods of moribund, parallel, or conflicting efforts.The part- major health partnerships, although efforts are under way-- nerships speak different technical languages, have vastly through the High-Level Forum for the Millennium different resource bases, and target different risk groups. Development Goals for health and through development of But they all seek to add value through a common array of a global partnership for child survival--to bring together functions, including the following: different partners around a common goal or around all the Millennium Development Goals for health.It may be useful · National and global coordination. Committees, forums, for countries active in multiple partnerships to initiate cross- communication channels, and Web sites make resource partnership learning and collaboration.The small Initiative or knowledge flows more efficient,identify where cross- for Public-Private Partnerships for Health,based in Geneva, national efforts are best merited, and reduce burdens on is one example of a modest effort to share the lessons of the those on the"frontlines"at the country or local levels. roughly 70 active programs involving industry and public · Strategy development and evaluation. Partnerships such as institutions in research and development and new product the Roll Back Malaria core strategies,the Global Plan to development to solve the health problems of developing Stop TB, and the current search for a common position countries. on antiretroviral drug resistance strive for consensus on Getting funds to the frontline effective intervention packages, monitoring indicators, and replicable solutions to specific implementation Central government funds can easily leak as they move obstacles or opportunities. through the system to the periphery of the country.And in the absence of local initiative and the right incentives,service · Global financing and delivery mechanisms and new tools. provision can fail to reflect the views of local people.Effective Partnerships such as the Global Fund to Fight AIDS, development assistance to health needs to address these Tuberculosis and Malaria; the GlobalTB Drug Facility; impediments.It needs to channel technologies,ideas,finance, theVaccine Fund; and the Global Alliance for Improved and technical assistance closer to households and health Nutrition, as well as public-private partnerships for providers and supervisory officials in ways consistent with developing new diagnostics,drugs,and vaccines provide national policies and amenable to monitoring and reporting. global public goods and respond to common externali- Development assistance to health is likely to be more ties or market failures. successful with four things in place: · Resource mobilization, social mobilization, and advocacy. · a decentralized system of fiduciary and technical man- Nearly all partnerships grapple with resource mobi- agement in the public sector (financial procedures, lization, advocacy, and community mobilization. competent officials, open reporting) Recent external evaluations of some of the major health · strong financial capacity of NGOs and private partnerships, such as Stop TB and the Global Alliance for providers Vaccines and Immunization, suggest that these collabora- tions are indeed adding substantial value. But some partner- · a government body equipped and charged with ships lack strategic focus and try to do too much. Others regulating the quality of public and private providers committed to stimulating a scaling-up of disease control and · a balanced approach to community-driven develop- other interventions fail to engage enough with country- ment in health, to ensure that social fund­type financ- level processes to make things happen on the ground. ing for community health initiatives is sustainable The global partnerships can also have cumbersome and costly governance and management arrangements, with Examples of development assistance to health reaching large governing boards that fail to make strategic decisions frontline workers in an expeditious and sustainable way or micromanage the full-time staff of the partnership sec- include block grants for districts in Uganda, social funds in retariats. And some ignore, or exacerbate, the problems CentralAmerica,contracting with urban and rural NGOs in afflicting the entire health system, such as the lack of the Reproductive and Child Health Program in India, and human resources for service delivery and weaknesses in support to community-led initiatives in the Multi-Country integrated monitoring and evaluation.The challenge is for AIDS Program (MAP). MAP is supporting thousands of 160 | The Millennium Development Goals for Health community-led initiatives, including 2,500 subprojects in · integrating the Millennium Development Goals in Bank Ghana,1,200 community-based activities in Kenya,and sup- sectorwide and programmatic instruments port to more than 4,600 villages in Ethiopia. · reorienting lending and grants toward Millennium Development assistance to health is still unpredictable Development Goal outcome objectives Donor financing for health has not been as reliable or sus- · using the Millennium Development Goals to build mon- tained as is often claimed or hoped for, even under new itoring and evaluation capacity "long-term" arrangements. In some countries the cuts in development assistance to health have been sharp. Donor Conducting country analyses of Millennium budgets are subject to the usual business and political Development Goal trends, prospects, and challenges cycles--and may go up or down yearly during budgetary The worldwide focus on the health, nutrition, and popula- processes. For countries such as Comoros and Eritrea, tion goals has led to the production of several innovative where the year-to-year changes in external funding can analyses of Millennium Development Goal performance, amount to as much as a fifth of all public spending for trends, and determinants. In India the Bank recently com- health, the fluctuations are so great that they make it nearly pleted a multivariate analysis of the large disparities in impossible to plan and implement a coherent national pro- Millennium Development Goal performance across states gram (see figure 9.7). and an assessment of the impact of public expenditures and Further work is needed in designing mechanisms for other policies on Millennium Development Goal out- development assistance to health that provide greater comes.7 The study concludes that the impact of additional assurance of sustained financial support.The challenges are public spending for improved child health and nutrition to overcome the factors that result in interruptions in would be greatest in the poorer Indian states. The study long-term development assistance to health, including recommends steps to improve the efficiency of public changes in political leadership and aid agency manage- expenditures, by targeting certain services (such as immu- ment that lead to reneging on earlier agreements. nization and community-based nutrition activities) and villages and districts in the poorer states in which health Applying the lessons in and nutrition indicators are the worst. A similar approach World Bank programs to geographical targeting has been proposed by the Hunger Task Force under the Millennium Project, which has iden- What are governments and development agencies actually tified 75 nutrition "hot spots" in Africa where more than doing to align their efforts with the Millennium Develop- 40 percent of the continent's malnourished children live. ment Goals? What new approaches are they taking to In Egypt a recent Millennium Development Goal improve their contributions? Are they applying the lessons analysis found that targeting publicly financed services and from experience with development assistance for health? involving civil society in implementing programs and Within the World Bank, a review of recent trends and monitoring progress for the health, nutrition, and popula- changes in the Bank's principal instruments--country tion goals are critical for success.7 In Burkina Faso an in- analyses, policy dialogue with government, and financing depth review of the child mortality target identified key of country-based projects and programs--suggests some constraints at the household, health system, and policy lev- important and encouraging developments in the way the els.8 The review also analyzed policy reforms already institution is responding to the Millennium Development undertaken and recommended further actions. Goal challenge.While much more needs to be done in the The Bank has collaborated with local analysts and poli- coming years,there are some positive signs.What follows is cymakers in several African countries over the past two a brief description of activities in the following areas: years to conduct studies on health, nutrition, and popula- · conducting country analyses of Millennium Develop- tion trends and services nationally and as they relate to dif- ment Goal trends, prospects, and challenges to inform ferent socioeconomic groups, with a special emphasis on Poverty Reduction Strategies the poorest households.The studies,known as country sta- tus reports, have been completed for nine countries-- · incorporating the Millennium Development Goals in Burkina Faso, Chad, The Gambia, Guinea, Malawi, government policies and budgets, including medium- Mauritania, Mozambique, Niger, and Tanzania--and are term expenditure frameworks ongoing for another seven (e.g., box 10.3). The reports, · strengthening the Millennium Development Goal focus important inputs into the national dialogue on the in Bank assistance strategies Millennium Development Goals and development policy, Applying the Lessons of Development Assistance for Health |161 the Millennium Development Goals. In Mauritania the Box 10.3 Improving performance of the analysis shaped the Poverty Reduction Strategy Paper and health sector in Guinea resulted in salary incentives for health workers serving in remote areas and an increase in the health sector budget of The country status report for Guinea shows that the outlook for nearly 70 percent over two years (box 10.4).Mauritania used reaching the Millennium Development Goals is bleak. In 1999 "marginal budgeting of bottlenecks" tool to link the children born in the poorest 20 percent of the population were Millennium Development Goals for health outcomes with twice as likely to die before reaching their first birthday as chil- essential services and the marginal cost of providing these ser- dren born in the richest 20 percent. Service use is particularly vices on a sustainable basis.The tool enabled the Mauritanians low for children, as evidenced by low rates of vaccination and treatment for acute respiratory infections. Despite high use rates for antenatal care, the rates of assisted deliveries are extremely low in rural areas. Cost is a major barrier to use, particularly for Box 10.4 Increasing health sector spending poor households. About 10­15 percent of households are per- and immunization coverage by reforming manently unable to pay for health services, and few exemption the budget process in Mauritania schemes and subsidization mechanisms are in place.As a result, many poor households frequently resort to self-medication. Mauritania's Poverty Reduction Strategy, launched in 2000, The key health sector performance issues included the profoundly reexamined health sector strategies, assessing past following: successes and failures and reorienting resources toward poor regions, poor groups, and efforts on outcomes related to the · Mismatch between where money is spent and epidemiological Millennium Development Goals. To address demand- and needs. Public expenditure in the 1990s focused on services supply-side constraints, the strategy focused on reorienting in urban areas, particularly in the capital city of Conakry. public spending along three main lines: addressing gaps in ser- · Low availability of resources. More than 60 percent of health vice delivery to the poor,developing income protection mech- personnel work in Conakry, serving only 20 percent of anisms, and strengthening the voice of poor people and the country's population. Drugs and vaccines are not communities in service design and management.The strategy readily and consistently available. Spending on operations identified performance parameters and key inputs, such as and maintenance at the primary level is too low. incentives for rural practice, training and recruitment, facility support, subsidized essential supplies, vaccination costs, emer- · Low perceived quality of services. Shortages of vaccines and gency AIDS plan, subsidized third-party payer hospital funds, drugs continue to undermine service quality, both techni- and monitoring of priority programs. The reform issues cal and perceived. The perceived quality of services is addressed include human resources, drug purchasing, logistics, often low. For example, many households do not view autonomy of management committees, reform of budget pro- assisted delivery services as being of good quality. cedures, and performance contracts with regional directorates The report recommends providing a package of essential based on coverage objectives. interventions to the poor, strengthening institutional capacity The Ministry of Health developed a medium-term expendi- by developing a framework for decentralization, reallocating ture framework that showed explicitly how it would strengthen resources to primary and essential secondary care, and increas- three approaches to service delivery: family-oriented services, ing the government's commitment to sustainable financing. population-based services, and clinical care to "purchase results" in child and maternal mortality and malnutrition with invest- Source: References 9 and 10. ments in health services.The Minister of Health led the budget preparation process, consulting at each step with the Minister of Finance.Budgeting was done on the basis of the service delivery have, in some countries, helped in drafting poverty strat- bottlenecks identified during the assessment phase. egy papers, estimating the cost of policies and programs, Results of the budget reform process have been impressive. and designing medium-term public spending proposals. The health sector budget increased by more than 50 percent in 2002 and by 13 percent in 2003. An early evaluation of the Incorporating the Millennium Development Goals in impact of the changes in the budgeting process shows an government policies and budgets increase in immunization coverage from 32 percent in 2001 to 82 percent in 2002.The clear links in the national health bud- The main findings of country status reports from African get between the new activities to be funded and the objectives nations have been fed into country-led processes--the of the Poverty Reduction Strategy proved essential in the sub- Poverty Reduction Strategy Papers and the medium-term sequent discussions with the Ministry of Finance and the expenditure frameworks--changing key policies in the International Monetary Fund. health sector and reallocating public spending toward health, Source: Reference 11. nutrition,and population services likely to have an impact on 162 | The Millennium Development Goals for Health to estimate, for example, the cost of investments that would bringing the government and external partners together allow for wider immunization of children and a higher per- around a common set of outcome targets--tied to the centage of women delivering babies under good conditions. goals. The Ghana sectorwide approach, now entering its second five-year phase, has led to a substantial upswing in Using the Millennium Development Goals to assess the use of health services for better child and maternal World Bank country assistance strategies health, with the largest increases in the poorest areas of the To better align its country assistance strategies to the country. Success factors include joint appraisal by all part- Millennium Development Goals, the Bank has conducted ners, strong government leadership, and expanded capacity informal assessments in several regions. One example is the to manage donor funds. Some challenges to enhance the portfolio ranking in Benin, which found that the Bank's impact of the Ghana sectorwide approach on the health activities were weakly related to the goals: they had only goals will be to address persistent inequalities across regions low-to-moderate relevance for maternal and child mortality and households, the brain drain of Ghanaian health person- reduction and mixed relevance for communicable diseases. nel,and the limited use of NGOs to increase coverage. The ranking concluded that the Bank needed to do more, Reorienting and increasing Bank loans and grants to through the Benin poverty reduction credit and other pro- achieve Millennium Development Goal outcomes jects, to have a substantial impact on health outcomes.The Benin social fund was seen as a promising instrument for To underpin national efforts to improve health outcomes, increasing the Bank's impact. The country assessment also theWorld Bank has significantly expanded its financial com- highlighted the importance of tracking intermediate indica- mitments in health over the past four years, through projects tors to measure progress in the short and medium terms.12 focused on specific diseases (the tuberculosis and AIDS loan In the Europe and Central Asia region of the Bank, staff in the Russian Federation) and population groups (the Brazil drafted a Millennium Development Goal "business plan," Family Health Project) as well as through broader health including a country-by-country analysis showing that sectorwide approaches and other multisectoral operations. official data on child mortality and malnutrition were of This increase has been consistent with the stated goal of questionable accuracy--and that many countries in the boosting health lending from $1 billion in fiscal 2001 to $2.2 region were unlikely to meet the goals for 2015.The busi- billion in fiscal 2005.Mirroring this target,actual lending has ness plan called for special efforts to strengthen capacity grown from yearly commitments of $0.95 billion in fiscal for monitoring and evaluation, align Bank projects better 2000 to $1.7 billion in fiscal 2003. Projected lending for fis- to support achievement of the Millennium Development cal 2004 is expected to be more than $2 billion. And the Goals, and expand multisectoral linkages. lending target for fiscal 2005 is likely to be met. An important feature of health lending is that more of it Integrating the Millennium Development Goals in the is being incorporated in health components in other sec- Bank's sectorwide and programmatic instruments tors, such as transport, social protection, and water supply In many cases the goals provide the strategic underpinnings and sanitation. For example, of the $1.7 billion in fiscal of sectorwide programs in health and multisectoral budget 2003 commitments, about 44 percent of total health lend- support. The Dominican Republic health project uses ing was in projects and programs outside the health sector. Millennium Development Goal intermediate and outcome These changes in the World Bank's response to the indicators to monitor progress.The health information sys- Millennium Development Goal agenda create three prob- tem has been designed to capture data on these indicators, lems. First, the Bank's organizational structure and incen- augmented by the results of periodic household surveys. In tives make it difficult to bring health specialists into project Bolivia performance indicators for the Bank-financed health teams led by staff from other disciplines, such as water sup- project include coverage and quality of key child and mater- ply and transport. Unless this problem is solved, inadequate nal health and disease control services, as well as changes in attention may be devoted to the health dimensions of these the incidence of diseases and deaths. Monitoring at the multisectoral operations.To address the problem, the Bank municipal and national levels, with the involvement of civil has organized special training and coaching arrangements society,will increase local participation and make local politi- for multisectoral teams, and it is testing budgeting and per- cians and healthcare providers more accountable for results. formance assessment techniques to encourage more effec- Some sectorwide operations for health in Africa embody tive teamwork in a multisectoral environment. a sharp focus on the goals. Since sectorwide approaches Second, the new projects require effective monitoring combine donor resources in a single budget envelope and and evaluation systems to track progress on intermediate apply common procedures to their disbursement, they pro- indicators and Millennium Development Goal outcomes. vide a good vehicle for reducing transactions costs and Traditionally, too little effort has gone to establishing such Applying the Lessons of Development Assistance for Health |163 monitoring systems at the outset of Bank-supported pro- the Millennium Development Goals is for HIV/AIDS, grams.And most governments continue to be reluctant to where the Bank has been asked to take the lead role in invest in the kind of operations research needed to evalu- coordinating partners (box 10.5) ate the impact of these programs. Recent efforts to stan- dardize the use of intermediate indicators for the Toward coordinated donor actions to Millennium Development Goals for health--as well as accelerate country level progress initiatives to build strong monitoring into AIDS programs in Africa, tuberculosis control projects in China and else- To pull together the lessons of effective development assis- where, and child health programs in Africa and Latin tance for the health goals, the donor community, in con- America--are pointing in the direction of a solution to cert with leaders from developing countries, has begun to this problem. Third, the Bank needs to address the in-house mix of skills needed to manage relations with countries in an envi- Box 10.5 Harmonizing, monitoring, and ronment in which the focus is on the Millennium Develop- evaluating HIV/AIDS programs ment Goals and on a new set of financial instruments. Required are staff with knowledge and experience of health The success of the Multi-Country AIDS Program (MAP) and policy,institutional reform,and the economic,financial,and other investments in HIV/AIDS will be determined by the regulatory dimensions of health. Existing staff will need to degree to which implementing partners, countries, and the be retrained and additional staff with the right skill sets donor community can learn by doing. This requires better information on program reach and effectiveness.Yet few coun- recruited. Over the past two years the Bank has begun to tries have adequate capacity to collect and use basic coverage tackle this challenge by expanding its external recruitment data. Managerial and budgeting structures in most countries and introducing new training programs on scaling up pro- lack performance incentives, and motivation is undermined by grams related to the Millennium Development Goals and lack of feedback on whether programs are succeeding or failing. on using new lending and financing instruments. Recognizing this,and fueled by the intense demand for results at all levels, UNAIDS decided to form the Global HIV/AIDS Using the Millennium Development Goals to build Monitoring and Evaluation Support Team (GAMET) at the monitoring and evaluation capacity World Bank. Effective national programs to pursue the health, nutrition, GAMET made considerable progress in its first year of oper- ation.A GAMET advisory board was established to structure and and population Millennium Development Goals depend provide guidance across agencies. A country support team--a on the use of intermediate indicators to track progress-- network of consultant experts (the majority from Africa) in and on building national monitoring and evaluation sys- monitoring and evaluation capacity building--made more than tems with local capacity.The Bank is supporting efforts in 70 visits to more than 20 countries active in the MAP. GAMET both areas. It convened a meeting of technical experts in also supported the design and implementation of a new manage- November 2001 to review and agree on a framework of ment development intervention, the Rapid Results Initiative in intermediate determinants.Those determinants were later Eritrea. The initiative focuses on achieving highly visible and published in a booklet now used in setting Millennium rapid results in AIDS prevention and treatment--and puts in Development Goal targets in Poverty Reduction Strategy place an accountability framework. Papers and other national strategies and in developing Each of the agencies participating in the GAMET initiative monitoring and evaluation systems.13 faces tensions between its internal requirements for monitor- Strengthening monitoring and evaluation capacity is ing and evaluation and its desire to facilitate a coordinated approach at the country level. Donors are committed to reduc- taking different forms.In Albania,as part of a recent Bank- ing the burden of multiple monitoring and evaluation requests financed poverty reduction credit, a monitoring and eval- and policies, but each is also under pressure to show impact in uation template was prepared to help four ministries the near term, which can undercut even the best intentions to (including health) develop their own monitoring and rely on country-based systems. Spending staff time coordinat- evaluation systems, and a specialized advisory body on ing approaches and building country capacity rather than ful- monitoring was established. In Mali a health card has been filling fiduciary responsibilities to monitor their own programs developed to give government officials and the public a generates additional problems. Despite the obstacles, GAMET snapshot of policy actions, health service indicators, and shows promise, especially as governments and agencies see the health outcomes for the entire country and for different benefits of a common approach to monitoring and evaluating income groups. national HIV/AIDS programs. One of the most important global and national efforts Source: World Bank staff. to improve monitoring and evaluation of progress toward 164 | The Millennium Development Goals for Health put together a mechanism for continuing consultation and menting and sharing successful cases of national efforts to for monitoring progress. achieve the malnutrition and maternal and child health The Framework for Action to Accelerate Progress on the goals. Health, Nutrition, and Population Millennium Goals, endorsed At the first High-Level Forum on health MDGs, at the high-level policy meeting in May 2003 in Ottawa, cohosted by theWorld Health Organization and theWorld Canada, holds promise for all stakeholders.The focus is on Bank in Geneva, January 8­9, 2004, heads of development building stronger national health systems as a platform for agencies, bilateral agencies, global health initiatives, and delivering essential services to the poor. The framework ministers of finance and health met to informally discuss lays out commonly held principles and describes a process concrete actions to accelerate and monitor progress for countries and donors to work together in expanding toward the Millennium Development Goals for health and and improving the effectiveness of their investments in nutrition. The forum focused on major constraints to health systems (box 10.6). It requires country actions, such progress in three areas: financial resources and aid effec- as incorporating analysis of Millennium Development tiveness, human resources, and monitoring performance. It Goal challenges and policy and funding gaps in poverty highlighted the policy steps that developing countries strategy papers, and simpler arrangements for donor coor- need to implement to save lives--and the harmonized dination. It also calls for stronger global efforts to invest in actions that donors must take to increase the impact of key public goods with multicountry benefits, such as financial and technical support to developing countries. research and development for new drugs and vaccines for The key conclusions and actions agreed on at the High- AIDS, tuberculosis, and malaria. And it promotes docu- Level Forum are presented in box 10.7. Box 10.6 Key actions to accelerate progress on the health, nutrition, and population Millennium Goals: The Ottawa consensus What donors should do What countries should do Create donor buy-in and coherence by subscribing to a country- Develop a credible strategy. Prepare as part of the Poverty based Millennium Development Goals for health strategy, includ- Reduction Strategy Paper process a credible strategy and imple- ing goals and targets, policy actions, financing proposals, and mentation plan (or review gaps and opportunities in existing monitoring arrangements, within a framework guided by the strategies and plans) for achieving faster progress toward the Poverty Reduction Strategy Paper and associated sectoral Millennium Development Goals for health, based on a solid strategies. analysis of impediments to faster progress and a thorough Build on existing mechanisms at the country level, including assessment of policy options. Poverty Reduction Strategy Papers and sectorwide approaches, Tackle key constraints, including the need for better human bilateral and multilateral funding streams, and those emanating resources, safe and predictable supply of drugs, stronger man- from the global health initiatives. agement systems, and more effective public-private interactions. Harmonize efforts to overcome funding gaps, fungibility issues, Adopt a strong multisectoral framework within and beyond the rigidity in financing recurrent expenditures, and lack of health sector. This framework would be explicit about policy predictability of aid flows. changes beyond health, such as overall public administration Commit to providing additional long-term financial assistance reform, and about other multisectoral activities. disbursed through existing multilateral and bilateral channels Make a commitment to improve governance and policies, and instruments. No new funding body is envisaged. Donors including policies on decentralization, civil society participation would move toward long-term assistance in a reliable and timely and monitoring, public-private partnerships, and public expendi- manner, including support for recurrent expenditures. ture allocation decisions. Increase efforts to work with NGOs and communities. Strengthen transparency, monitoring and evaluation, through Reduce the transactions costs of development assistance to commitment to open and transparent reporting. health, by seeking to harmonize reporting requirements, Provide voice to citizens--especially poor people, in formulating procurement rules, and financial management systems. and implementing strategy. Consultations would be undertaken Support capacity building for results-oriented monitoring and as part of the Poverty Reduction Strategy Paper process and evaluation. Donors would also commit to an independent review beyond and include NGOs and communities. of their actions and to sharing the lessons of this review with other stakeholders. Source: Reference 14. As a result of the Ottawa meeting, donors and leaders of devel- progress on the health goals, monitor changes in donor commit- oping countries decided to create a High-Level Forum that ments and behaviors in moving toward stronger harmonization, would convene twice a year over a two-year period to review and discuss and act on issues holding back faster progress. Applying the Lessons of Development Assistance for Health |165 Box 10.7 Summary of recommendations for action from the High-Level Forum on the Millennium Development Goals for Health held January 8­9, 2004 Resources for health and Human resources in health Poverty Reduction Strategy Papers The Secretariat of the High-Level Forum will: Countries should have a single process leading to one "Millen- · Assess expenditures on human resources in health by devel- nium Development Goal­responsive"Poverty Reduction Strategy opment partners. Paper.To facilitate the process, countries should be encouraged to · Develop a series of in-depth human resources studies on prepare health sector strategies with investment plans based on selected developing countries addressing the current stock of well-documented needs and costing scenarios. health personnel, the requirements to meet the Millennium Members of the High-Level Forum will ask the next Development Goals, and deployment of health personnel. Development Committee to request that the World Bank and This work should link with related work in theWorld Bank, International Monetary Fund incorporate into their joint assess- WHO, and the International Labour Organization Joint ments of Poverty Reduction Strategies explicit reference to and Learning Initiative. review of progress toward the Millennium Development Goals. The Millennium Project will ask the UN secretary-general to · Link with the Global Commission on International endorse this recommendation to the Development Committee. Migration to examine the impact of migration on health. Aid effectiveness and harmonization Monitoring performance The Secretariat of the High-Level Forum will link with the The High-Level Forum welcomed the Health Metrics Network Organisation for Economic Co-operation and Development­ and encouraged its rapid launch, at or around the World Health Development Assistance Committee (OECD/DAC) work on Assembly in May 2004. The Secretariat of the Health Metrics harmonization and, with developing countries, draw out lessons Network, which will be located at WHO for the first 18 months, on common methods and instruments for health and review was encouraged to disseminate its draft business plan to forum achievements to date. Second, it will follow up on country pilots members. Its priority tasks are agreement on a set of intermediate in budget support for health. indicators; definition of process indicators, including measures of A paper on aid effectiveness, poor performers, and countries policy and institutional performance,to regularly gauge short-term in crisis will be presented to the next High-Level Forum. It will progress toward the Millennium Development Goals for health; draw on existing work by the World Bank, World Health and creation of a "report card" to illustrate progress. The Health Organization (WHO), United Nations Development Pro- Metrics Network will disaggregate country data, improve their gramme, and OECD/DAC. It will provide recommendations on quality,and increase the coherence of data collection platforms.The how to work with these countries to achieve the Millennium World Bank and the WHO, in collaboration with the OECD, will Development Goals for health. work to improve the tracking of national financial investment in the health sector, from domestic and external sources, using National HealthAccounts and other data on financial flows. Source: Reference 15. References 6. World Bank. 2003. Attaining the Millennium Development Goals in India: How Likely andWhatWill itTake?Washington, DC:World 1. OECD Development Assistance Committee. 2000. RecentTrends Bank. in Official Development Assistance to Health. Paris: OECD. 7. El-Saharty, S., E. Richardson, and S. Chase. 2003. Egypt and the 2. Michaud, C. 2003. Development Assistance for Health (DAH): Millennium Development Goals.World Bank,Washington DC. RecentTrends and Resource Allocation. Harvard Center for 8. Naimoli,J.F.,T.Johnston,and M.Schneidman.2003."Reaching Population and Development Studies, Cambridge, MA. the MDGs in Burkina Faso.An Assessment of MDG #4:Reduce 3. Feyzioglu,T.N.,V. Swaroop, and M. Zhu. 1996."Foreign Aid's Child Mortality byTwo-Thirds by 2015."Background paper for Impact on Public Spending." Policy ResearchWorking Paper the MDG Economic SectorWork.World Bank,Health, 1610,World Bank,Washington, DC. Nutrition,and Population Department,Washington,DC. 4. Burnside, C., and D. Dollar. 2000."Aid, Growth, the Incentive 9. World Bank and Ministry of Health of Guinea. Guinea:A Regime and Poverty Reduction." In TheWorld Bank: Structure Country Status Report on Health and Poverty (Health, Nutrition and and Policies, eds. C.L. Gilbert and D.Vines, 210­27. Cambridge: Population Inputs for the PRSP and HIPC Process).World Bank, Cambridge University Press. Africa Region Human Development Unit,Washington, DC. 5. World Bank. 1998. Assessing Aid:WhatWorks,What Doesn't, and 10. Soucat,A., S. Bonu, andY. Camara. 2003."Guinea Case Study." Why. Oxford: Oxford University Press. Background paper for the Millennium Development Goal 166| The Millennium Development Goals for Health Economic SectorWork.World Bank,Africa Region Human 13. World Bank. 2002. Annual Review of Development Effectiveness. Development Unit,Washington, DC. Achieving Development Outcomes:The Millennium Challenge. 11. Soucat,A. 2003."Mauritania Case Study: Poverty Reduction Washington, DC:World Bank. Strategy Leads to Dramatic Increase of Resources for Health 14. "Framework for Accelerating Progress on Health, Nutrition, and Nutrition Services." Background paper for the Millennium and Population MDGs 2003." Presented at Ottawa Policy Development Goal Economic SectorWork.World Bank,Africa Meeting on HNP MDGs. Region Human Development Unit,Washington, DC. 15. High Level Forum on Health MDGs, Geneva. 2004. 12. Abrantes,A. 2003. Personal communication.World Bank. http://www.who.int/hdp/en/summary. Applying the Lessons of Development Assistance for Health | 167 APPENDIX A Data and Methods Rates of change Under-five mortality Unless stated otherwise, rates of change are computed on Data come from the United Nations Children's Fund the assumption that the indicator in question falls by the child mortality database, available at www.childinfo.org. same proportion each year. Rates are computed using dis- The data comprise mostly complete series for 1960, 1970, crete growth formulae. If m1990 and m2015 are the rates of 1980, 1990, 1995, and 2000.Average growth rates for, say, malnutrition in 1990 and 2015, the Millennium Develop- the 1990s were calculated assuming constant proportion- ment Target requires that (m2015 ­ m1990)/m1990 = 0.5, or ate growth using the formula r = (m2000/m1990)1/10 ­1. equivalently m2015 = 0.5m1990.If m grows at an average (neg- ative) rate r per year, m2015 = m1990(1 + r)25. Substituting Maternal mortality m2015 = 0.5m1990 gives r = ­0.27. See Box 2.1 and appendix B. Malnutrition Data on malnutrition (underweight among children) Communicable disease mortality come from the World Health Organization's malnutrition See appendix B. database, available at www.who.int/nutgrowthdb/.The data comprise very incomplete series for the 1990s for a Modeling prospects for future progress on large number of developing countries. Many countries do the health Millennium Development Goals not have data for adjacent years. For each country, a regression was estimated for this report for the 1990s,link- The modeling strategy is to ask how future changes will ing the natural logarithm of underweight to the year of affect the growth rate of the Millennium Development the survey, the coefficient being interpretable as the aver- Indicator in question. Mortality or malnutrition, m, is age annual rate of reduction (or increase) over the period assumed to be related to a vector of determinants, x, in the for which data are available. The coefficients were then following way: averaged--using population sizes as weights--to obtain ln m = a + blnx +e regional averages. This method is less restrictive than the method used in an earlier study,1 where regressions were where ln denotes natural logarithm, a and b are coeffi- cients, b is interpretable as an elasticity, and e is an error in effect estimated at the regional level. Although coun- term. Differentiating this expression with respect to time tries were allowed to have different intercepts, they were yields a growth equation of the form: assumed to have the same slopes as other countries in the region (that is, common rates of change). m = b x + e 169 where m is the rate of growth of m and so on.The change identified by Filmer and Pritchett). Also included in the in the growth rate can be written: instrument set was the average value of an index of "voice" m = bx + e among neighbors. In contrast to Filmer and Pritchett, but like Rajkumar and Swaroop,1 this study allows for the possi- so that if the growth rates of each of the elements of x bility that the elasticity of government health spending may remain the same over time, m will continue to fall at the depend on the quality of governance, the hypothesis being same rate. If, by contrast, there is a change in the growth that health spending will have a larger impact on health rate of, say, x1 (which might be per capita income) but all outcomes in better-governed countries. Governance is the other x's (x2, . . . , xK) continue to grow at the same measured using the Bank's Country Policy and Institutional rate, the change in the rate of growth of m is given by: m = b1x1. Assessment (CPIA) index, which is increasing in quality of governance and takes a minimum of 1 and a maximum of Given an estimate of the elasticity b1 and an estimate of 6. Part I countries (the industrial countries), for which no change in the growth rate of x1, one can estimate the size CPIA score is available, were assigned a CPIA score of 5. of the change in the growth rate of m likely to result from The government spending variable is interacted with the a change in the growth rate of x1.The same method can CPIA score, and elasticities and t-statistics are computed for be used to assess the effects of changes in growth rates of different values of the CPIA index.The interaction is also several x's simultaneously. treated as endogenous, with the three original instruments interacted with the CPIA score serving as instruments for the spending-CPIA interaction.3 Regression models for the growth simulations Per capita income is converted using purchasing power Elasticities for the various country-level determinants of parity (PPP) dollars (as in Filmer and Pritchett) and offi- each of the Millennium Development Indicators are cial exchange rates. Using PPP dollars may be more needed to operationalize this growth simulation approach. appropriate, but since PPP dollars are not specific to the For this report, the study by Filmer and Pritchett2 of health sector (or more accurately to the goods and services country-level influences on child mortality was extended that make for lower mortality and malnutrition), it is not and replicated on more recent data and on other obvious that the overall PPP is necessarily always a more Millennium Development Indicators. appropriate converter than the exchange rate. The broad Regressions were estimated linking the Millennium conclusions are unaffected, but some of the magnitudes Development Indicators to government health spending as a involved are sensitive to the choice. percent of GDP, income per capita, and variables capturing Female education is captured through variables that female education, access to drinking water, income inequal- indicate the percentage of women 15 and over who have ity, ethnolinguistic fractionalization, and regional dummies. completed primary education, the percentage of women Regressions were also run including a measure of infrastruc- 15 and over who have completed secondary education, ture in addition to these determinants.Regressions were run and the percentage of women 15 and over who have com- for under-five mortality (2000), maternal mortality (1995), pleted higher-level schooling. The "missing" category is underweight (2000),and tuberculosis mortality (2000) as the women who have not completed any level of education. dependent variable.All variables were entered in logarithmic Including female education in this way allows investiga- form to facilitate the growth accounting outlined above.As tion of the returns--in terms of better health outcomes-- in Filmer and Pritchett's study,countries with missing values of increasing primary completion rates and of narrowing for variables other than health spending and income were gender differentials in education enrolment,both of which retained in the sample by including in the regression a miss- are Millennium Development Goals in their own right. It ing value dummy variable flag and setting the variable with also allows for the possibility that primary and secondary the missing value equal to zero for the country in question. education have different effects on health outcomes.The TableA.1 provides the variable definitions and the sources of water variable proved to be highly collinear with a sanita- the data. tion variable originally included, so only the water vari- As in Filmer and Pritchett's study, government health able was retained. In addition, as in Filmer and Pritchett, spending is treated as endogenous to eliminate any reverse income inequality (measured through the Gini coeffi- causality--governments may choose their spending levels in cient) and an index of ethnolinguistic fractionalization the light of their mortality rates.As in Filmer and Pritchett's were included. The final variable--not included in the study, the instruments used were the average levels of Filmer and Pritchett specification--is the length of the defense and health spending by the country's neighbors (as country's paved road network in kilometers (the number 170 | The Millennium Development Goals for Health Table A.1 Variable Definitions Variable Definition Source Log government health expendi- Government health expenditures as proportion of GDP Reference 7 tures as a percentage of GDP Log government health expenditures as a The CPIA (Country Policy and Institutional Assessment) assesses the World Bank percentage of GDP interacted with CPIA quality of a country's policy and institutional framework. The index (not public contains 20 items grouped into four categories: economic manage- domain) ment, structural policies, policies for social inclusion and equity, and public sector management and institutions. The index ranges from a minimum of 1 (unsatisfactory for an extended period) to a maximum of 6 (good for an extended period). Log per capita income in U.S. dollars Per capita income measured in US dollars at official exchange rates Reference 7 Log per capita income in international Per capita income measured at PPP exchange rates Reference 7 dollars Log percentage of female population 15 Proportion of females that completed primary school Reference 5 and over that completed primary school Log percentage of female population 15 Proportion of females that completed secondary school Reference 5 and over that completed secondary school Log percentage of female population 15 Proportion of females that completed higher school Reference 5 and over that completed high school Log percentage of population with Proportion of population with improved drinking water sources Reference 6 access to safe water and sanitation Gini coefficient for income inequality The Gini coefficient measures the proportional deviation of income Reference 7 distribution from perfect equality. A Gini coefficient of 0 represents perfect equality; a Gini coefficient of 1 means that one person receives 100 percent of income. Index of ethnolinguistic fractionalization Probability that two people speak the same native language Reference 2 Kilometers of paved road as fraction of Length of paved road in kilometers (number of kilometers of roads Reference 7 land area times the fraction of roads that are paved) as a proportion of the country's area (in square kilometers) Regional dummies One for industrial countries and one for each of the following World n.a. Bank regions: Europe and Central Asia, Latin America and the Caribbean, Middle East and North Africa, South Asia, and Sub-Saharan Africa (East Asia and the Pacific is omitted) Defense spending by neighbor Average level of defense spending by the country's neighbors Reference 7 Definition of neighbor: Reference 2 Health spending by neighbor Average level of health spending by the country's neighbors Reference 7 Definition of neighbor: Reference 2 n.a. Not applicable. of kilometers of roads times the fraction of roads that are better-governed countries have larger (in absolute size) paved), expressed as a proportion of the country's area (in elasticities of mortality and malnutrition with respect to square kilometers).This variable captures infrastructure. government health spending.This result is consistent with Table A.2 shows the results of the regressions. Equation the findings of Rajkumar and Swaroop. For a given value 3 in each case is closest to the Filmer and Pritchett of the CPIA index,the spending elasticities are smallest for specification--per capita income is converted using PPPs, under-five mortality and become statistically significant and the infrastructure variable is excluded. All equations for under-five mortality only at very high CPIA scores. differ from the Filmer and Pritchett specification in that The low elasticity of under-five mortality with respect to the spending-CPIA interaction is included. In virtually all government health spending is consistent with Filmer and specifications, the interaction is negative, suggesting that Pritchett, though somewhat at odds with Rajkumar and Applying the Lessons of Development Assistance for Health | 171 Table A.2 Regression Results U5MR 2000 MMR 1995 [1] [2] [3] [4] [1] [2] [3] [4] Coef. Coef. Coef. Coef. Coef. Coef. Coef. Coef. (|t|) (|t|) (|t|) (|t|) (|t|) (|t|) (|t|) (|t|) log govt hlth exps as 1.091 1.38 1.083 1.306 ­0.589 0.064 ­0.86 0.059 % gdp (2.13) (2.70) (1.95) (2.39) (0.71) (0.08) (0.67) (0.05) log govt hlth exps as % gdp ­0.292 ­0.328 ­0.267 ­0.286 ­0.033 ­0.170 0.035 ­0.130 interacted with cpia (2.92) (3.11) (2.57) (2.64) (0.21) (1.05) (0.15) (0.63) log per capita income ­0.325 ­0.358 0.025 0.035 US dollars (3.63) (4.00) (0.62) (0.86) log per capita income ­0.484 ­0.547 0.034 0.055 internat'l dollars (3.77) (4.63) (0.48) (0.80) log % female pop aged 15+ 0.009 ­0.042 ­0.009 ­0.047 0.120 0.035 0.127 0.046 completed primary school (0.10) (0.44) (0.09) (0.50) (0.76) (0.23) (0.72) (0.28) log % female pop aged 15+ ­0.054 ­0.053 ­0.053 ­0.047 ­0.353 ­0.318 ­0.328 ­0.296 completed secondary school (0.71) (0.67) (0.70) (0.59) (3.13) (2.89) (2.71) (2.55) log % female pop aged 15+ 0.056 0.053 0.060 0.058 0.052 0.025 0.007 ­0.008 completed higher school (0.78) (0.71) (0.85) (0.79) (0.42) (0.21) (0.05) (0.07) log % pop with access ­0.333 ­0.473 ­0.262 ­0.339 ­0.669 ­0.969 ­0.782 ­1.144 to safe water (1.59) (2.16) (1.23) (1.48) (1.87) (3.00) (1.89) (3.26) Gini coefficient for 0.007 0.010 0.010 0.012 ­0.015 ­0.008 ­0.021 ­0.009 income inequality (1.10) (1.49) (1.50) (1.93) (1.40) (0.83) (1.54) (0.81) index of ethnolinguistic 0.274 0.425 0.298 0.431 ­0.240 ­0.018 ­0.374 0.072 fractionalization (0.97) (1.56) (1.01) (1.58) (0.58) (0.05) (0.63) (0.14) kilometers of paved road ­0.091 ­0.072 ­0.131 ­0.159 as fraction of land area (2.16) (1.58) (2.35) (2.07) Europe & Central Asia 0.082 ­0.072 ­0.010 ­0.139 ­0.613 ­0.895 ­0.511 ­0.910 (0.38) (0.34) (0.04) (0.66) (1.50) (2.33) (0.95) (1.98) Industrialized country 0.017 ­0.097 ­0.250 ­0.363 ­0.721 ­0.705 ­0.743 ­0.840 (0.06) (0.33) (0.94) (1.34) (1.43) (1.39) (1.27) (1.54) Latin America & Caribbean 0.020 0.008 ­0.192 ­0.232 1.076 1.051 1.218 1.078 (0.08) (0.03) (0.77) (0.88) (2.51) (2.45) (2.28) (2.11) Middle East & North Africa 0.352 0.332 0.218 0.187 0.886 0.821 0.788 0.707 (1.28) (1.13) (0.79) (0.64) (1.85) (1.72) (1.30) (1.18) South Asia 0.532 0.427 0.517 0.428 1.132 0.971 1.137 1.036 (2.11) (1.59) (2.08) (1.66) (2.76) (2.36) (2.55) (2.34) Sub­Saharan Africa 0.662 0.582 0.498 0.401 1.828 1.783 1.819 1.696 (2.89) (2.48) (1.97) (1.63) (4.87) (4.79) (4.05) (4.02) constant 6.731 7.709 8.083 8.983 8.047 9.418 8.608 9.823 (6.99) (8.16) (7.59) (9.50) (4.88) (6.21) (4.59) (5.44) Additional variables All equations include dummy variables for when schooling, water, Gini coefficient and road network are missing (in these cases the variable itself is set equal to zero). # countries 120 120 119 119 113 113 113 113 Adjusted R squared 0.912 0.898 0.913 0.903 0.883 0.883 0.866 0.872 Elasticity of dependent variable with respect to gov hlth exps as % gdp, evaluated at value of cpia indicated 1.00 0.799 1.053 0.816 1.020 ­0.622 ­0.106 ­0.824 ­0.071 (1.87) (2.51) (1.75) (2.26) (0.90) (0.16) (0.78) (0.08) 2.00 0.507 0.725 0.548 0.734 ­0.654 ­0.276 ­0.789 ­0.201 (1.46) (2.16) (1.43) (2.02) (1.18) (0.51) (0.93) (0.27) 3.00 0.215 0.397 0.281 0.447 ­0.687 ­0.446 ­0.754 ­0.331 (0.77) (1.49) (0.89) (1.54) (1.58) (1.06) (1.15) (0.58) 3.25 0.142 0.315 0.215 0.376 ­0.695 ­0.489 ­0.745 ­0.363 (0.53) (1.24) (0.71) (1.36) (1.70) (1.24) (1.22) (0.68) 3.50 0.069 0.233 0.148 0.304 ­0.703 ­0.531 ­0.736 ­0.396 (0.27) (0.96) (0.51) (1.15) (1.82) (1.42) (1.30) (0.79) 4.00 ­0.077 0.069 0.014 0.161 ­0.720 ­0.617 ­0.719 ­0.460 (0.32) (0.30) (0.05) (0.66) (2.07) (1.81) (1.46) (1.05) 4.50 ­0.223 ­0.095 ­0.119 0.018 ­0.736 ­0.702 ­0.701 ­0.525 (0.96) (0.42) (0.46) (0.08) (2.26) (2.16) (1.61) (1.32) 5.00 ­0.369 ­0.259 ­0.253 ­0.125 ­0.752 ­0.787 ­0.683 ­0.590 (1.56) (1.09) (0.97) (0.51) (2.34) (2.38) (1.69) (1.55) Table A.2 Regression Results (continued) Underweight 2000 TB mortality 2000 [1] [2] [3] [4] [1] [2] [3] [4] Coef. Coef. Coef. Coef. Coef. Coef. Coef. Coef. (|t|) (|t|) (|t|) (|t|) (|t|) (|t|) (|t|) (|t|) log govt hlth exps as 0.348 ­0.014 0.473 0.022 1.025 1.239 1.124 1.304 % gdp (0.44) (0.02) (0.51) (0.02) (1.34) (1.67) (1.33) (1.61) log govt hlth exps as % gdp ­0.218 ­0.154 ­0.245 ­0.166 ­0.374 ­0.412 ­0.400 ­0.421 interacted with cpia (1.26) (0.90) (1.24) (0.84) (2.42) (2.61) (2.43) (2.50) log per capita income ­0.271 ­0.232 ­0.325 ­0.322 US dollars (2.14) (1.85) (2.36) (2.40) log per capita income ­0.300 ­0.243 ­0.351 ­0.392 internat'l dollars (1.89) (1.59) (1.77) (2.18) log % female pop aged 15+ 0.318 0.332 0.269 0.298 0.123 0.040 0.087 0.016 completed primary school (1.98) (2.10) (1.65) (1.86) (0.79) (0.26) (0.53) (0.10) log % female pop aged 15+ ­0.046 ­0.099 ­0.062 ­0.112 ­0.035 ­0.083 ­0.067 ­0.098 completed secondary school (0.43) (0.90) (0.56) (0.99) (0.29) (0.69) (0.55) (0.81) log % female pop aged 15+ ­0.096 ­0.062 ­0.086 ­0.057 0.091 0.126 0.109 0.136 completed higher school (0.81) (0.51) (0.71) (0.46) (0.79) (1.08) (0.93) (1.16) log % pop with access ­0.262 ­0.197 ­0.273 ­0.221 ­0.162 ­0.378 ­0.178 ­0.346 to safe water (0.82) (0.62) (0.84) (0.64) (0.49) (1.14) (0.51) (0.97) Gini coefficient for 0.009 0.007 0.009 0.007 0.007 0.008 0.007 0.009 income inequality (0.94) (0.72) (0.95) (0.71) (0.65) (0.86) (0.62) (0.92) index of ethnolinguistic 0.111 ­0.079 0.086 ­0.127 0.129 0.179 0.080 0.156 fractionalization (0.24) (0.17) (0.18) (0.27) (0.30) (0.44) (0.17) (0.37) kilometers of paved road 0.030 0.041 ­0.134 ­0.126 as fraction of land area (0.43) (0.54) (2.05) (1.76) Europe & Central Asia ­1.246 ­1.206 ­1.334 ­1.253 ­0.456 ­0.692 ­0.549 ­0.765 (3.60) (3.77) (3.71) (3.79) (1.37) (2.17) (1.56) (2.33) Industrialized country 0.347 0.250 0.011 ­0.024 ­0.320 ­0.599 ­0.619 ­0.863 (0.65) (0.46) (0.02) (0.05) (0.73) (1.37) (1.44) (2.02) Latin America & Caribbean ­0.166 ­0.189 ­0.337 ­0.321 ­0.792 ­0.796 ­0.984 ­0.999 (0.48) (0.53) (0.96) (0.87) (2.09) (2.02) (2.41) (2.39) Middle East & North Africa ­0.460 ­0.465 ­0.598 ­0.570 ­1.915 ­1.951 ­2.060 ­2.091 (1.15) (1.13) (1.47) (1.35) (4.42) (4.36) (4.59) (4.56) South Asia 0.447 0.524 0.471 0.557 ­0.079 ­0.219 ­0.031 ­0.176 (1.24) (1.48) (1.29) (1.56) (0.20) (0.54) (0.08) (0.43) Sub­Saharan Africa ­0.053 ­0.006 ­0.125 ­0.052 0.330 0.315 0.275 0.221 (0.17) (0.02) (0.36) (0.15) (0.93) (0.89) (0.69) (0.58) constant 5.156 4.790 5.969 5.426 5.465 6.919 6.350 7.817 (3.39) (3.15) (3.55) (3.30) (3.57) (4.83) (3.65) (5.23) Additional variables All equations include dummy variables for when schooling, water, Gini coefficient and road network are missing (in these cases the variable itself is set equal to zero). # countries 85 85 85 85 118 118 117 117 Adjusted R squared 0.709 0.699 0.704 0.693 0.825 0.813 0.818 0.809 Elasticity of dependent variable with respect to gov hlth exps as % gdp, evaluated at value of cpia indicated 1.00 0.130 ­0.168 0.228 ­0.144 0.651 0.827 0.724 0.883 (0.20) (0.26) (0.30) (0.19) (1.03) (1.37) (1.03) (1.33) 2.00 ­0.087 ­0.321 ­0.017 ­0.310 0.276 0.415 0.323 0.463 ­0.305 ­0.475 ­0.262 ­0.476 ­0.098 0.003 ­0.077 0.042 3.00 (0.18) (0.64) (0.03) (0.52) (0.54) (0.87) (0.56) (0.87) (0.82) (1.23) (0.61) (1.06) (0.24) (0.01) (0.17) (0.10) 3,.25 ­0.360 ­0.514 ­0.324 ­0.518 ­0.192 ­0.100 ­0.177 ­0.063 (1.04) (1.41) (0.81) (1.24) (0.49) (0.28) (0.40) (0.16) 3.50 ­0.414 ­0.552 ­0.385 ­0.559 ­0.285 ­0.203 ­0.277 ­0.168 (1.27) (1.60) (1.04) (1.43) (0.76) (0.59) (0.65) (0.44) 4.00 ­0.523 ­0.629 ­0.508 ­0.643 ­0.472 ­0.409 ­0.478 ­0.378 (1.75) (1.95) (1.52) (1.82) (1.35) (1.25) (1.19) (1.06) 4.50 ­0.632 ­0.706 ­0.630 ­0.726 ­0.659 ­0.615 ­0.678 ­0.589 (2.13) (2.19) (1.96) (2.13) (1.91) (1.87) (1.74) (1.66) 5.00 ­0.740 ­0.783 ­0.753 ­0.809 ­0.847 ­0.821 ­0.878 ­0.799 (2.32) (2.29) (2.22) (2.28) (2.38) (2.36) (2.21) (2.16) Swaroop, who found significant elasticities of government For secondary school completion the extra growth spending at the sample median level of governance. (Their required to achieve gender parity in the proportion of the index of governance was, however, different.) population 15 and over that has completed secondary The health spending elasticities for the Millennium education is computed. In the absence of this extra Development Indicators other than under-five mortality are growth, completion rates are assumed to grow linearly not only larger absolutely than those for under-five mortal- between 2000 and 2015, following the same trend as ity, they also become statistically significant at lower CPIA between 1995 and 2000. This yields the 2015 value for scores. The results suggest that--whatever the Millennium males that females will also need to reach to achieve the Development Indicator--simply adding additional dollars to gender Millennium Development Goal. The forecast the budget of a ministry of health in a country with a CPIA assumes that achievement of the goal occurs through index of less than 3 will not yield improvements in any accelerated linear growth for females from the female Millennium Development Indicator. By contrast, in coun- 2000 value to the male 2015 value.This gives the growth tries with CPIA scores of 3.25 and higher, additional dollars rate needed between 2000 and 2015 to achieve the gender in the ministry of health budget could reduce rates of mater- goal.This rate is compared with the current growth rate to nal mortality (see equation 1 in table A.2).The cut-off for obtain the extra growth in secondary completion required malnutrition is somewhat higher--a CPIA of 3.5 or so to achieve the gender goal. (compare with equation 2), which given the insignificance For access to drinking water, the extra growth required of the road variable might be argued to be preferable to to halve the proportion of the population without access equation 1. In the case of tuberculosis mortality, the CPIA between 1990 and 2015 is computed. In the absence of cut-off is about 4.25 (see equation 1).These elasticities indi- this extra growth, the access rate is assumed to rise (or fall) cate the effects at the margin of adding dollars to the min- between 2000 and 2015 linearly at the same rate as in the istry of health budget on the assumption that the additional 1990s. The growth that would need to occur to achieve dollars are allocated across programs and institutions in pro- the goal is found by assuming access grows linearly portion to current allocations.Adding dollars specifically to a between 2000 and 2015, starting from the 2000 value and government tuberculosis program might yield a significant ending in 2015 at the value implied by the Millennium payoff in terms of reduced tuberculosis mortality at much Development Goal. The extra growth is the difference lower CPIA levels than 4.25.The results cannot,however,tell between the projected growth based on current trends and us whether or not this is the case. the growth required to achieve the goal. Under-five mortality appears to be more responsive to per capita income than the other Millennium Develop- References ment Indicators, with the possible exception of tuberculo- 1. Rajkumar,A., andV. Swaroop. 2002."Public Spending and sis mortality. Female education has a significantly negative Outcomes: Does Governance Matter?" Policy Research effect only on maternal mortality--and even then only at Working Paper 2840,World Bank,Washington, DC. the secondary level.The lack of any effect on under-five 2. Filmer, D., and L. Pritchett. 1999."The Impact of Public mortality is at odds with the Filmer and Pritchett results, Spending on Health: Does Money Matter?" Social Science and which report a significant effect of years of education. Medicine 49 (10): 1309­1323. Access to drinking water has a significant negative effect 3. Wooldridge, J.M. 2002. Econometric Analysis of Cross Section and Panel Data. Cambridge, MA: MIT Press. on maternal mortality in all specifications and on under- 4. World Bank. 2003. Global Economic Prospects and the Developing five mortality in the specifications in which exchange rates Countries.Washington, DC:World Bank. are used to convert per capita incomes. Income inequality 5. Barro, R.J., and J.W. Lee. 2000."International Data on has mostly an insignificant effect, though there is some Educational Attainment: Updates and Implications." CID suggestion it makes a difference to under-five mortality. Working Paper No. 42, Harvard University, Center for International Development, Cambridge, MA. The same is true of the index of ethnolinguistic fractional- 6. UNICEF.2003.StatisticsWater and Sanitation Databases. ization.The road infrastructure variable, by contrast, has a www.childinfo.org/eddb/water.htm and millenniumindicators. significant negative effect on under-five mortality, mater- un.org/unsd/mi/mi_indicator_xrxx.asp?ind_code=30. nal mortality, and tuberculosis mortality. 7. World Bank. 2003. World Development Indicators.Washington, DC:Washington, DC. Forecasting trends in economic growth and the nonhealth goals Economic growth forecasts are taken from the Bank's Global Economic Prospects.4 174 | The Millennium Development Goals for Health APPENDIX B WhyTracking ProgressToward the Health Goals Isn't Easy Assessments of progress toward the Millennium Develop- changes more difficult.TheWHO has assembled a database ment Goals are subject to uncertainties because of differ- of such data, available at www.who.int/nutgrowthdb/. ences in how the indicators are defined,which measurement For this report, a regression was estimated for each country instruments are used,how frequently the information is col- for the 1990s,linking the natural logarithm of underweight lected, and how much effort has been invested in the devel- children to the year of the survey.The coefficient can be opment of measurement systems. interpreted as the average annual rate of decrease--or increase--over the period for which data are available.The Child malnutrition coefficients were then averaged--using population sizes as weights--to obtain regional averages.This is less restrictive The prevalence of underweight children is defined as the than the method used in an earlier study,1 in which regres- proportion of children under five whose weight for their sions were in effect estimated at the regional level. age is less than two standard deviations below the median Although countries were allowed to have different inter- for an international reference population. Data collection cepts, it was assumed they had the same slopes as other is most often done with household surveys, which require countries in the region (that is,the same rates of change). that the interviewers measure children and obtain their age in months. Inaccuracies in this information, as well as Child mortality--infant and other errors due to sampling errors, are the main reasons under-five mortality for the uncertainties about the estimated prevalence of malnutrition. Confidence intervals at the 95 percent level Infant and child mortality are measured using vital regis- in the Demographic and Health Survey sample are typi- tration systems and household surveys. Vital registration cally about plus or minus two percentage points, but the systems can provide detailed and precise annual statistics interval varies depending on the size of the sample and the when all or nearly all births and deaths of children under prevalence of underweight children. five are registered. But many countries have vital registra- Household surveys that measure malnutrition indicators tion systems that do not capture all births or deaths, mak- are usually conducted several years apart,and several coun- ing such systems of little use for estimating mortality tries do not have baselines for the earlier years of the indicators. When compared with other sources of infor- Millennium Development Goal period. In some coun- mation, estimates based on vital registration systems are tries, surveys are not nationally representative and may frequently found to be too low. produce estimates that are not comparable with earlier or Household surveys measure infant and child mortality later estimates. Some countries have used different refer- by asking female respondents to report the number of chil- ence populations, which are not usable at all. dren they have given birth to and the survival status of the As with all indicators, measuring changes over time children.Some surveys collect detailed information to con- implies using at least two estimates that are subject to struct full birth histories for each respondent, while others unknown biases, which makes establishing significant collect only enough information to estimate childhood 175 mortality indicators with additional demographic models. viving household members or others (such as health care Where both types of data are available,they frequently pro- workers) who were close to the deceased are interviewed to duce different estimates and trends in mortality.As with all establish cause of death. The questionnaire used in such household surveys, sampling and other errors produce studies includes questions about signs and symptoms of ill- uncertainty about the exact level of the indicators. nesses at the time of death, medical history, and obstetrical Confidence intervals at the 95 percent level for under-five history.The verbal autopsy methodology often fails to iden- mortality in Demographic and Health Surveys vary from as tify deaths early in pregnancy due to ectopic pregnancy or low as plus or minus 5 per 1,000 (India 1998/99) to as high abortion or those occurring some time after delivery. as plus or minus 11 per 1,000 (Ethiopia 2000).At any level Household surveys are the main instrument for measur- of disaggregation,confidence intervals tend to be larger. ing maternal mortality indicators. The usual approach is Because of the retrospective reporting on childhood based on the so-called sisterhood methods,in which female deaths, survey-based estimates generally are not estimates respondents are asked to report on the survival of sisters. for the year of the survey but are commonly presented as These methods do not produce reliable results when fertil- averages for the five-year period before the survey.Annual ity is low (a total fertility rate of about 3 or lower) or where estimates are rarely available, as these would require very fertility has changed a great deal in recent years.Depending large sample sizes. on which questions are asked, the estimates generally refer One advantage of collecting childhood mortality estimates to mortality that occurred several years before the survey with surveys is that one obtains estimates for as long as 15 (10­12 years in some cases). Sisterhood methods cannot years in the past,allowing trends to be estimated from a single reliably report on early pregnancy-related deaths, as preg- survey. When multiple household surveys are available and nancy status will be less likely to be known to the respon- estimates have been made from both birth histories and dents. Induced abortion­related deaths are likely to be demographic models,additional analytical efforts are required underreported for the same reason. Even with large sample to produce a consistent series.The latest exercise (available at sizes,standard errors tend to be very large.For example,the www.childinfo.org) has produced data for 1960, 1970, 1980, Malawi Demographic and Health Survey in 1992 gener- 1990, 1995, and 2000.2,3 The data for 1990 and 2000 were ated a maternal mortality ratio of 752, with a 95 percent used to compute the average annual rate of decline (or confidence interval of 523­803, for the six years before the increase) using the formula r = (m2000/m1990)1/10 ­ 1, where survey. Because of such large confidence intervals and the m1990 and m2000 are the mortality rates for 1990 and 2000. number of years before the survey to which the estimate refers, household surveys are not suitable for monitoring Maternal health--the maternal short-term trends in maternal mortality or assessing the mortality ratio impact of programs. The cost and complexity of collecting data with these The maternal mortality ratio is one of the more problem- methods has led theWHO and UNICEF to develop models atic indicators to monitor. The indicator is a ratio of the that predict maternal mortality indicators.3 Such models use number of deaths directly or indirectly related to preg- variables such as the proportion of births with a skilled atten- nancy and the number of live births in a given time period. dant, fertility rates, variables indicating regions, and Because not all pregnancies result in live births, the ratio HIV/AIDS prevalence to model the maternal mortality always overestimates the true risk of dying of maternal ratio. These variables may not be good predictors of the causes.The use of births as the denominator makes it theo- maternal mortality indicators, and they may themselves be retically possible for the measure to be greater than one. poorly measured.For these reasons,maternal mortality ratios Maternal mortality ratios can be estimated with vital estimated in this way have wide margins of uncertainty.This registration data, but the incompleteness of vital registra- modeling approach was used to estimate changes in the tion systems is a major obstacle to their use in epidemiol- maternal mortality ratios at the regional level in the 1990s. ogy. Where vital registration is complete, both the numerator and denominator for the maternal mortality Incidence, prevalence, and death from ratio can be obtained.But even in countries with relatively HIV/AIDS, tuberculosis, and malaria good overall cause of death data, maternal causes are gen- erally not very accurate. Even in countries with complete Monitoring trends in communicable diseases requires sur- coverage of deaths in vital registration, maternal deaths veillance for correct diagnosis of a disease through clinical may be misclassified in as many as 50 percent of cases. diagnosis or laboratory confirmation. In other countries the cause of death is sometimes estab- Surveillance of HIV/AIDS prevalence has focused on lished through the "verbal autopsy" method, in which sur- testing pregnant women in antenatal clinics. Such surveil- 176 | The Millennium Development Goals for Health lance is subject to various biases: the sites are more likely to tion recorded as part of vital registration systems.Too few be located in or near urban areas, pregnant women may children with severe malaria are admitted to hospitals to have different risks of infection than the overall female rely on hospital records: an estimated 90 percent of child population of reproductive age, and pregnant women are deaths occur at home in Sub-Saharan countries. Given the not likely to be representative of the entire adult popula- difficulty of diagnosing malaria correctly in these circum- tion.A new approach to establishing adult HIV prevalence stances, the best approach is believed to be monitoring all is being implemented as part of Demographic and Health causes of under-five mortality in malarious areas. Surveys by serosurveys of household members. But high levels of nonresponse may limit the usefulness of this References approach. 1. de Onis, M., E.A. Frongillo, and M. Blossner. 2000."Is For malaria, incidence is particularly difficult to estab- Malnutrition Declining? An Analysis of Changes in Levels of lish, as the symptoms of malaria are similar to those of Child Malnutrition Since 1980." Bulletin of theWorld Health many other acute infectious diseases that affect children. Organization 78 (10): 1222­1233. Even when microscopy is available, the cause of fever in 2. Hill, K., and A.Yazbeck. 1994."Trends in Child Mortality, children will be difficult to determine, as malaria para- 1960­90: Estimates for 84 Developing Countries." Background Paper No. 6 to World Development Report 1993: Investing in sitaemia are frequently present in children in endemic Health.World Bank,Washington, DC. areas, making it difficult to determine which organism is 3. Hill, K., R. Pande, M. Mahy, and G. Jones. 1999. Trends in Child responsible for the fever.Measurement of malaria mortality Mortality in the DevelopingWorld: 1960 to 1996. NewYork: is equally difficult in the absence of cause-specific informa- UNICEF. Applying the Lessons of Development Assistance for Health |177 Index A AIDS.See HIV/AIDS Abortion services,52­53 Air quality,16,141­142 Access and utilization Albania,20,123f, 164 allocation of government health spending and,8­9 Argentina,12,88f among poor populations,6,26,49 Armenia,61,80 consumer education and,10,82­87 Asia.See East Asia and Pacific;Europe and Central Asia; drugs and health supplies,14­15,66 South Asia;specific country financial barriers,9­10,70­71,79­82 Autonomization,101­102 household-level factors,28­29 Azerbaijan,80,123f marginal budgeting for bottlenecks,9,61­64 mortality outcomes and,49­53,54f B progress toward Millennium Development Goals Bangladesh,79­80 and,6,26,47 health resource allocation,60 social fund investment and,62 immunization program,12 social solidarity and,79f malnutrition intervention,35,157f time costs and,10,87 mortality reduction,45 transportation infrastructure and,16,87­89,140 Bednets,insecticide-treated,137,138f trends,54f Benin,163 vaccination and immunization coverage,49,53f Bolivia,8,10,17,20,64,81,84 women's issues,10 child mortality,62 worldwide disparities,49 government health spending,146,148­149 Accountability health resource allocation,61 in delivery of health services,11­13,93­94,95 social investment funds,62 management philosophies,96­98 Botswana,37 in water access,89 Brazil,10,86,100­101f, 123,139 Accreditation of hospitals,103 child mortality,11,49­52 Acute respiratory infection,52f, 117f disease surveillance in,138f core public health functions in preventing/ health services contracting,12 treating, 134­137t Burkino Faso,12,117f, 161 Africa. See Sub-Saharan Africa; specific country Burundi,27,65 Agricultural policies and practices intersectoral synergies to improve,142 C women's issues,16,142 Cambodia,104f, 146 179 Cash transfer programs,82,83f autonomization,101­102 Child morbidity decentralization,12­13,102­103 acute respiratory infection,52f, 117f focus of Millennium Development Goals,45 health care access and utilization and,6,49 government contracting for,12,103­105 household health resources and,70 government role in,132 strategies for increasing use of household health household role,49,64,69,70 interventions to reduce,86­88t informal provider network,106­107 vaccination and immunization coverage,49,53f management issues,11,93 Child mortality management philosophies,96­98 among poor populations,36,44­45 marginal budgeting for bottlenecks,9,61­64 causes of death,48­49 opportunities for improving,93­94,107 in developing world,2 ownership in,94­95 educational attainment and,40 potential benefits of increased coverage of home- effective existing interventions,48­49,48t, 50f delivered interventions,49­52 future prospects,4,31,38,39f, 41,45 potential challenges in,93,95­96 government health spending and,7,56,57,58,59,60 privatization,102 health care access and utilization and,6,26,47 rationale for government involvement, human resources for health and,112 132­133,140f linkages with other Millennium Development Goals,32 reforms to increase provider accountability,98­107 management factors in prevention program regulation of private health care providers,105­106 effectiveness,100­101f See also Human resources for health;Public health Millennium Development Goals,35 functions potential benefits of increased coverage of home- Developing countries delivered interventions,49­52 disease surveillance and monitoring,138 progress toward Millennium Development health care access and utilization,6,26,49 Goals,2,31,35­37,45 international health initiatives in,1 social fund investment and,62 malnutrition patterns,25,26,31­35 utilization of existing interventions and,49­52,54f maternal mortality patterns,37 water access and,40­41 mortality and morbidity patterns,1,25,26 worldwide,1,25,26 policy reforms tied to,157­158 China,45,139,142,146 progress toward Millennium Development Goals,2,31, child mortality,49­52 32­38,45­46 HIV/AIDS in,131 water access,89 Colombia,35 See also specific country Community context,70,79 Development assistance Consumer information and education,10,82­87,106 continuity issues,19,145,152,161 medication use,14­15,118,120 coordination of donor actions in,164­165 Contraception,10,87f effectiveness,7,156­157 Contracting for health services,12,103­105 fungibility,158­159 Cost of care,79­80,146 government allocation and,18,152,158­159 medication costs,15,118­119,122­124 international health initiatives and global partnerships, Côte d'Ivoire,38,79f, 112,117f 17,18­19,28,29f, 159­160 Cuba,138 monitoring use of,20 obstacles to effectiveness,17­18,152­153,155 D performance-based,18 Debt relief,28 as share of government spending,151,152f Decentralization of health service delivery,12­13,102­103 strategies for improving effectiveness,18­19,153,155, Delivery of health services 159,160­161 accountability in,11­13,93­94,95 transaction costs,18,159 actors,94­95 trends,28,145,156 180| The Millennium Development Goals for Health Diarrhea benefits of maintaining effort toward Millennium core public health functions in preventing/treating, Development Goals,31,42­46 134­137t development assistance,151­152,155­156 effective existing interventions,49 government health spending,7,153 hygiene and sanitation practices to prevent,141 potential benefits of increased access and related mortality,48,52 utilization,49­53 Dominican Republic,20,139f, 163 progress toward Millennium Development Goals,4,31, 38­42,156 E socioeconomic indicators,4,40 East Asia and Pacific World Bank health development strategies,21 economic growth,4,7,40 effects of government health spending,60 G future prospects,7 Gambia,112,114 malnutrition in,32,35 Gates Foundation,14,28,120,156 mortality patterns,26,37,43,59,60 Gender issues water access and sanitation,89 educational attainment,4,31,40 Economic development progress toward Millennium Development Goals expected growth patterns,4,31,40 and,41­42 ill health and,1,26­27 See alsoWomen's issues income transfer programs,79 Georgia,80,120,121f per capita income and,40 Ghana,14,98,102,113f, 120,123f, 163 progress toward Millennium Development Global Alliance for Improved Nutrition,28 Goals and,5f, 38­40,41­42,58­59, Global Alliance forVaccines and Immunizations, 58f, 71­79 1,19,28,158 Ecuador,98 Global Fund to Fight AIDS,Tuberculosis and Malaria, Educational attainment 1,18,28,156 child mortality and,40 Government health spending future prospects,4,40 allocation across health system,8­9,18,19,60­64,146 gender differences,4,31,40 allocation of development aid,18,152,158­159 health outcomes and,10,41­42,85­86 cash transfer programs,82,83f Millennium Development Goals,40 challenges to increasing domestic revenue,149­151 progress toward Millennium Development Goals determinants of,145,146­148 and,58­59 development assistance and,151­153 Egypt,19,45,82,161 effectiveness,7,19,55,56 mortality reduction,45 externalities,80,132 El Salvador,12 future prospects,153 Eritrea,112 health services contracting,12 Estonia,102,123f on human resources,13­14 Ethiopia,36­37,64 increases necessary to reach Millennium Development Europe and Central Asia Goals,59­60 economic growth,4,40 marginal budgeting for bottlenecks,9,61­64 future prospects,4 mortality outcomes and,56,57,58,59,60 health services contracting,12 national debt payment and,28 human resources for health care,13,112 quality of institutions and policies and,56­60,57t, 157 mortality patterns,2,4,41,43 rationale for targeted subsidies,80­82 See also specific country to reduce cost of care,80 revenue intake and,17,146­149 F risk-pooling,16,80,146 Famine,79f as share of gross domestic product,146,147­151 Fertility rate,43,44 sources of additional funding,145,148­149 Future prospects strategies for improving,153 Rising to the Challenges |181 Guatemala,104f, 123f compensation,13,114­115 Guinea,162f cross-border flows,114 HIV/AIDS effects on,114 H influence on consumer health behavior,87 Haiti,12 patterns,13 Health care workers.See Human resources for health performance and productivity,13,113 Higher-income countries policy issues,111 child mortality reduction in,35­36 problems in,111 health care access and utilization,6,49,53f progress toward Millennium Development Goals prevention and treatment strategies,53f and,111,112 HIV/AIDS,49 public vs. private,93 in China,131 recruitment and training,13­14,65­66,115­117 core public health functions in preventing/treating, skills,13,14,27­28,65,111,113,117 134­137t supply,112­113,114 drug supply and distribution,15 within-country differences,113 economic development and,1 worker expectations,14,14f, 115,116f effective existing interventions,48t worldwide disparities,49,112 effects on health worker supply,114 Hungary,123f health care access and utilization,6 Hygiene and sanitation,16,89 maternal mortality and,43 intersectoral synergies to improve,141 medications,124f, 125f Millennium Development Goals,37 I prevalence trends,37­38 India,10,16,70,79f, 84 program monitoring and evaluation,164f human resources for health,13,115,116f progress toward Millennium Development immunization program,84 Goals,45­46 nutrition programs,11,99f public health functions to combat,142f, 143 private health care spending in,146 vaccine development,126f public health spending,19,61,161 worldwide mortality and morbidity,1,25,26,37­38 Self-EmployedWomen's Association,84f Honduras,45,83f tuberculosis intervention,65,66 Households Indonesia,9,27,45,82,93,96­98,123f bottlenecks in delivery of health interventions,63­64 Infectious disease child health interventions,70 core public health functions in preventing/treating, community context,70 134­137t, 137 definition,9n economic effects,1 in delivery of health interventions,49,64,69,70 linkages with other Millennium Development Goals,32 determinants of health behavior,28,69,70­71,80, poverty and,26­27 84­85,86,87 progress toward Millennium Development Goals,37­38 health care spending,16,26,79­80,146 public health system response to crisis outbreaks,139 health knowledge and beliefs,82­87 surveillance and monitoring activities,138 health resources of,69 utilization of current interventions to prevent,6,26,49 health system role of,9­11,27,45,64,69,70,132 worldwide patterns,1 help-seeking behaviors,70 See also specific disease hygiene and sanitation behavior,16 Initiative for Public-Private Partnerships for Health,160 indoor air pollution,16,141­142 Institutions and policies,government,55 potential benefits of increased coverage of home- accountability in,11­13,65,96­105 delivered interventions,49­52 accreditation of hospitals,103 reducing economic obstacles to health care,71­82 autonomization,101­102 strategies for increasing use of health community representation,11­12 interventions by,86­92t decentralization,12­13,102­103 Human resources for health development aid effectiveness and,17­18,19,155, birth attendants,43­44,49 156­157 182| The Millennium Development Goals for Health development assistance conditional upon Malaysia,45,81,96­98,112 reforms in,157­158 Mali,9,20,63,164 drug supply and distribution,14­15,120­125 Malnutrition focus of Millennium Development Goals and,45 among poor populations,35 human resources management,111 child mortality and,48­49 influence of,in community health behavior,70,87 effective existing interventions,48t management issues in health service delivery,11,96­98 food supply and access,79f, 80 Millennium Development Goals and health goals of, government health spending and,56,57,57t, 58,59 42­44,162­163 household-level interventions,71f privatization,102 intersectoral strategies to improve,16 progress toward Millennium Health Goals and quality linkages with other Millennium Development Goals,32 of,56­60,64 management factors in nutrition program regulation of private health care providers,105­106 effectiveness,99f strategies for improving,19,64­66 Millennium Development Goals,32 structural problems of health service delivery national commitment to programs against,157f system,95­96 progress toward Millennium Development Goals, in support of core public health functions,139,140f 2,31,32­35 See also Public health functions social costs,26­27 Insurance,9­10,80 utilization of current interventions to prevent,6 drug coverage,15,122,123f within-country differences,2 Integrated Management of Childhood Illness,8,10,14,49, worldwide,1,25,26 61,66,80,86,98,117 Management of health service institutions,93 Integrated Management of Pregnancy and Childbirth,49 accountability,11,65,93,96­98 Intellectual property regimes,drug development and,15, government role in,132 119­120,124­125 performance-based approaches,98­101 Intersectoral cooperation,16,139­142 philosophies and styles,96­98 strategies for improving drug availability and use,120 J structure of health delivery system,95­96 Jamaica,98 Marginal budgeting for bottlenecks,9,61­64 Mass media,87 L Maternal mortality Latin America and Caribbean determinants of,43 child mortality patterns,59 effective existing interventions,48,48t, 51f economic growth,4,40 future prospects,4,31,41­42,43­44 effects of government health spending,59 government health spending and,57,58,59 health services contracting,12 linkages with other Millennium Development Goals,32 HIV/AIDS in,124f Millennium Development Goals,31,37,43 maternal mortality patterns,37,43 progress toward Millennium Development Goals, mortality patterns,2 2,4,31,37 progress toward Millennium Development Goals,2 trend data,2,43 water access,41 utilization of existing interventions and,49,52­53 See specific country water access and,40­41 Lesotho,60 worldwide,1,25,26 Liberia,112 Mauritania,162f Lithuania,123f Measles,134­137t Medicines and health supplies,66 M consumer information,14­15,85,118,120 Madagascar,12 counterfeit drugs,118f Malaria,46,52,138f determinants of supply and distribution,14 core public health functions in preventing/treating, government regulation,15,122­125 134­137t inappropriate/irrational use,118f effective existing interventions,48t intellectual property issues,15,119­120,124­125 Rising to the Challenges |183 logistics management,120 children worldwide,1 obstacles to improving health,117­120 government spending and,56 pooled procurement,123,124f quality trend data,2 prescribing practices,120­122 utilization of current interventions to prevent,6,49­53,54f prices,15,118­119,119f, 122­124 See also Child mortality;Maternal mortality research on neglected diseases,120,125 Mozambique,124f strategies for improving availability and use,120­125 Myanmar,123f Middle East and North Africa child mortality patterns,59 N economic growth,4,40 Namibia,112 effects of government health spending,59 Nepal,87,123f human resources for health care,112 Nicaragua,12,82f, 83f maternal mortality patterns,37,43 Nigeria,112,122 water access,41 See specific country O Middle-income countries Osiga,George,25 child mortality reduction in,35­36 health care access and utilization,49 P health services contracting,12 Packard Foundation,28 malnutrition reduction goals,32 Pakistan,86,104f progress toward Millennium Development Goals,2 Performance-based evaluation,18 Millennium Development Goals,28­29 in internal contracts,98­101 benefits of maintaining effort toward achieving, Peru,12,45,60 4,31,42­46 Philippines,26,114 challenges for development community,17­21 Poland,80 country-level challenges,6­17 Poor populations future prospects,4,31,38­42 cash transfer programs for,82,83f government health spending and,7,55,57­60,57t effectiveness of increased government health health-related goals,1,2f, 31,32,33­34t spending,56,61 human resource issues,111,112 effects of illness,1,26­27 indicators of progress,32,33­34t, 138 financial barriers to health care,9­10 intersectoral synergies to achieve,141t health care spending,16,26,146 linkages among,32 health services access and utilization,6,26,49 measurement of within-country differences,4 malnutrition patterns,35 monitoring national progress toward,138,139f, measurement of Millennium DevelopmentTargets,4,32 161­162,163­164 mortality patterns,26,27f, 36,44,45,56 obstacles to success,4­6,53 poverty reduction linkages in Millennium origins,1,25,28 Development Goals,32 outcomes focus,45 progress toward Millennium Development Goals, poverty reduction goals and,32 35,44­45 progress to date,2,3t, 31,32­38,45 social investment funds for,61,62 quality of trend data,2,32 targeting of government health subsidies for,80­82 role of core public health functions in achieving,132, Poverty Reduction Strategy Papers,18,19,20,21,28,112, 133,134­137t, 137,142­143 142,155 strategies for improving progress toward,165,165f, 166f Pregnancy and perinatal care structure of health delivery system and,95f benefits of targeted health subsidies,81 utilization of existing preventive interventions birth attendance,43­44,49 and,6,26,47 health care access and utilization in,6,26,49,52­53 World Bank assistance strategies and,19­21,161­164 management factors in health care quality,98 See also specific goal strategies for increasing use of household health Mortality interventions in,74­78t among poor populations,26,27f See also Maternal mortality 184| The Millennium Development Goals for Health Primary care StopTB Partnership,1,19,28 government spending,8,61 Sub-Saharan Africa influence on consumer health behavior,87 child mortality,49­52 performance-based payment programs,100 development assistance for health,17 privatization,102 economic growth,4,7,40 Private health care spending,16,26,79­80,118­119,146 effects of government health spending,59,60 Private sector health system,96,103­107 future prospects,4,7 Privatization,88f, 89,102 HIV/AIDS mortality,26 Professional associations,139 human resources for health,112 Public health functions malnutrition,2,32,35 in achieving Millennium Development Goals,132,133, mortality patterns,1,4,25,26,31,36­37,41,43 137,140­143t progress toward Millennium Development consumer health knowledge,82­87 Goals, 2, 31 intersectoral efforts,16,137­139 water access,41,89 market failure rationale,132 See also specific country monitoring and evaluation,138 Sudan,80 national institutions in support of,139 Surveillance and monitoring,15­16,20,138 obstacles to performing,133 outcomes measurement,133 T resource allocations,61 Tanzania, 8, 9, 12­13, 35, 61, 81, 102, 112, 113, 117f, role of households,9­11,27,45,64,69,70 140, 152 scope of activities in,131­132,133t Tax system,17,147­151 strategies for strengthening,15­16,66,142­143 Thailand,12,13 See also Institutions and policies,government HIV/AIDS in,45,66 human resources for health,114­115 R malnutrition intervention,157f Research,medical,120,125,126f Transportation,10,80 Risk pooling,16,80,146 intersectoral collaborations for improving,140 Rockefeller Foundation,28 public health infrastructure,16,87­89 Roll Back Malaria,28 TRIPS Agreement,124­125 Romania,102 Tuberculosis core public health functions in preventing/ S treating,134­137t Sen,Amartya,1,26 DOTS,8,38,45,61,120,121f Senegal,12,45 drug-resistant,120,121f Severe acute respiratory syndrome,131,139 effective existing interventions,48t Sex workers,82f government health spending and,56,57,57t Sexually transmitted diseases,106 progress toward Millennium Development Goals,45 Sierra Leone,88 worldwide mortality,1,25,26 Social investment funds,61,62 Social marketing,86,106,137,138f U South Africa,9,79,98 Uganda,13,106 South Asia child mortality,36­37 economic growth,4 government health spending,146,152 malnutrition,1,25,26,32 HIV/AIDS in,38,45 mortality patterns,4,26,41,43 UNAIDS,1,28 prenatal care,6,49 User charges,9­10 water access and sanitation,89 See also specific country V Spending.See Development assistance;Government health Vaccination and immunization,6,26,49,53f, 80,84, spending;Private health care spending 126f, 162f Sri Lanka,28,45 Vietnam,26,27­28,69,80,119f, 139,145,146,157­158 Rising to the Challenges |185 W Women's issues Water access agricultural policies,16,142 accountability issues,89 educational attainment,4,31,40 future prospects,41 financial autonomy,82,84f health outcomes and,10­11,16,85,89 health care access and utilization,10 intersectoral synergies to improve,141 World Bank linkages with other Millennium Development Goals,32 Comprehensive Rural Health Project,70 Millennium Development Goals,40­41 health lending,20,156­157,163 privatization programs,88f, 89 International Development Assistance grants,17,18 progress toward Millennium Development Goals,4,31, International Development Association grants,156,158f 41­42,58­59 Millennium Development Goals and,19­21, Within-country differences 155,161­164 child mortality,36 Poverty Reduction Strategy Papers,18,19,20,21,28, health care resource allocations,8­9,60­61 112,142,155 health services access and utilization,6,49 social investment funds,62 human resources for health,113 staff,164 malnutrition,2 WorldTrade OrganizationTRIPS Agreement,124­125 measurement of Millennium Development Goals,4 progress toward Millennium Development Goals,44­45 Z Zambia,1,13,115 Zimbabwe,12 186| The Millennium Development Goals for Health THE MILLENNIUM DEVELOPMENT GOALS FOR HEALTH RISING TO THE CHALLENGES N early half of the Millennium Development Goals directly or indirectly target health and nutrition. However, despite the efforts of the 1990s, the prospect of achieving these goals looks bleak--the extent of premature death and ill health in the developing world continues to be staggering. In 2001, 3 million people died from HIV/AIDS, and tuberculosis claimed another 2 million lives. In 2000, almost 11 million children died before their fifth birthday, and an estimated 140 million children under age five were underweight. In 1995, over half a million women died during pregnancy or childbirth. This heavy burden of death and suffering is concentrated in the world's poorest countries and acts as a formidable obstacle to poverty reduction. The Millennium Development Goals for Health: Rising to the Challenges provides data on the progress to date on the five health-related Millennium Development Goals: I reducing extreme poverty and hunger; I reducing the mortality rate of children under age five; I reducing maternal deaths; I fighting HIV/AIDS, malaria, and other diseases; and I improving access to water and sanitation. The book also brings together what is known about the constraints to faster progress and offers policies to address them. Finally, it examines how the international community (and the World Bank in particular) might work better with countries to achieve the Millennium Development Goals. TMxHSKIMBy357675zv":;:=:':( ISBN 08213-5767-0