47980 W O R L D B A N K W O R K I N G P A P E R N O . 1 6 6 Clearing the Global Health Fog A Systematic Review of the Evidence on Integration of Health Systems and Targeted Interventions Rifat Atun Thyra de Jongh Federica V. Secci Kelechi Ohiri Olusoji Adeyi V H P R I O R I T I E S T O O L S S Y S T E M S N R E C A I N T E R V E N T I O N R I L E I U V E T G L C T I N H H O R I Z O N T A L R C C I A C L I E E V I D E N C E T D A T A S T U D I E S E I L P I P O P U L A T I O N P D O G M A F L N R N R A O G L O B A L T D I A G O N A L A S Y C O V E R A G E H THE WORLD BANK W O R L D B A N K W O R K I N G P A P E R N O . 1 6 6 Clearing the Global Health Fog A Systematic Review of the Evidence on Integration of Health Systems and Targeted Interventions Rifat Atun Thyra de Jongh Federica V. Secci Kelechi Ohiri Olusoji Adeyi THE WORLD BANK Washington, D.C. Copyright © 2009 The International Bank for Reconstruction and Development/The World Bank 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First Printing: March 2009 printed on recycled paper 1 2 3 4 5 12 11 10 09 World Bank Working Papers are published to communicate the results of the Bank's work to the development community with the least possible delay. The manuscript of this paper therefore has not been prepared in accordance with the procedures appropriate to formally-edited texts. Some sources cited in this paper may be informal documents that are not readily available. The findings, interpretations, and conclusions expressed herein are those of the author(s) and do not necessarily reflect the views of the International Bank for Reconstruction and Development/TheWorld Bank and its affiliated organizations,or those of the Executive Directors of The World Bank or the governments they represent. The World 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 The World Bank of the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law.The International Bank for Reconstruction and Development/The World Bank encourages dissemination of its work and will normally grant permission promptly to reproduce portions of the work. 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, Tel: 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, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, Fax: 202-522-2422, email: pubrights@worldbank.org. ISBN-13: 978-0-8213-7818-2 eISBN: 978-0-8213-7936-3 ISSN: 1726-5878 DOI: 10.1596/978-0-8213-7818-2 Library of Congress Cataloging-in-Publication Data has been requested. Contents Acknowledgments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v Acronyms and Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2. Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 The Conceptual Framework for Analysing Integration of Targeted Health Interventions into Health Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 3. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Outcomes Reported in the Studies Analyzed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 The Extent and Nature of Integration of Health Interventions into Critical Health Systems Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 The Extent of Intervention Success in the Studies Analyzed . . . . . . . . . . . . . . . . . . . 18 How the Context Influences the Extent and Nature of Integration . . . . . . . . . . . . . 19 4. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 5. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Appendixes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 A. Search Strategy and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 B. Summary of the Included Studies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 C. Contextual Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 LIST OF TABLES 1. Critical Health Systems Functions and Elements of Integration . . . . . . . . . . . . . . . . . . 5 LIST OF FIGURES 1. Flow Chart Representing the Selection Process for Studies Included in the Review. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 2. The Extent and Nature of Integration by Targeted Health Intervention and Intervention Success as Reported in the Study. . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 iii Acknowledgments he work program underpinning this paper was initiated and funded by theWorld Bank T and entitled"Integration of Health Systems and Priority Health, Nutrition and Popu- lation Interventions."Preparation of the paper was co-funded in a collaborative exercise by the Center for Health Management at Imperial College London and the World Bank. In that context the authors acknowledge consultations with a number of colleagues during an international consultation hosted by the World Bank in May 2008, and comments from several reviewers who helped to improve the manuscript.The reviewers were Tom Novotny (Professor, University of California, San Francisco) and Richard Coker (Reader, London School of Hygiene & Tropical Medicine). Sylvia Robles reviewed the initial version of a Cochrane Review, which was an earlier step in this exercise. Logan Brenzel and Peter Berman commented on the conceptual framework on which the paper is based. Sonalini Khetrapal provided comments on an earlier version of the paper and Inas Ellaham helped with formatting. The authors alone are responsible for the contents of the paper. About the authors: Rifat Atun, MBBS, MBA, FRCGP, FFPH. Professor of International Health Manage- ment, Imperial College, London. E-mail: r.atun@imperial.ac.uk Thyra de Jongh, MSc, DIC, PhD. Researcher, Centre for Health Management, Impe- rial College London. E-mail: thyra.de-jongh06@imperial.ac.uk FedericaV.Secci, MSc. Doctoral Researcher, Centre for Health Management, Imperial College London. E-mail: f.secci07@imperial.ac.uk Kelechi Ohiri, MD, MPH, MS. Health Specialist in the Human Development Network of the World Bank. E-mail: kohiri@worldbank.org OlusojiAdeyi, MD,DrPH,MBA.Coordinator of Public Health Programs in the Human Development Network of the World Bank. E-mail: oadeyi@worldbank.org Cover design: concept by Olusoji Adeyi, text by Olusoji Adeyi and Sonalini Khetrapal, graphics by Stuart K. Tucker. v Acronyms and Abbreviations AIDS Acquired Immune Deficiency Syndrome BINP Bangladesh Integrated Nutrition Project CWG Community Working Groups EPI Expanded Program for Immunization EPOC Cochrane Effective Practice and Organisation of Care Group FP Family Planning HIV Human Immunodeficiency Virus ICDDR,B International Center for Diarrheal Disease Research, Bangladesh IDCS India Integrated Child Development Services IMCI Integrated Management of Childhood Illness LHWP Lady Health Worker Program M&E Monitoring and Evaluation MCH Maternal and Child Health PHC Primary Health Care PPI Pulse Polio Immunization TB Tuberculosis UNICEF United Nations Children's Fund USAID United States Agency for International Development VCT Voluntary Counselling and Testing WHO World Health Organization vii Summary longstanding debate on health systems organization relates to benefits of integrating A health programs that emphasize specific interventions into mainstream health systems to increase access and improve health outcomes. This debate has long been characterized by polarization of views and ideologies,with protagonists for and against integration argu- ing the relative merits of each approach.Recently,the debate has been rekindled due to sub- stantial increases in externally funded programs for priority health, nutrition and population interventions and enhanced international efforts aimed at health systems strengthening. However, all too frequently these arguments have not been based on hard evidence. In this paper we present findings of a systematic review that explores a broad range of evidence on: (i) the extent and nature of integration of targeted health programs that emphasize specific interventions into critical health systems functions (defined in the Methodology Section),(ii) how the integration or non-integration of health programs into critical health systems functions in different contexts have influenced program success, (iii) how contextual factors have affected the extent to which these programs were integrated into critical health systems functions. We use a new conceptual framework to guide the analysis.The review evaluates peer-reviewed studies that focus on health interventions,and which have been introduced on a regional or national scale. The debate on health interventions has tended to narrowly focus on vertical or inte- grated descriptors. However, our analysis shows this to be a false dichotomy as few inter- ventions are purely vertical (single-disease oriented) or horizontal (fully integrated into mainstream functions) health system. As this review shows,in practice the nature of the problem,the interventions to address these and the adoption and assimilation of health interventions in health systems vary greatly in different contexts, as does success. The purpose, nature, speed and the extent of integration also vary--in part, dependent on the intervention complexity, the health sys- tem characteristics and the contextual factors. There are few instances where there is full integration of a health intervention or where an intervention is completely non-integrated. Instead, there exists a highly heterogeneous picture both for the nature and also for the extent of integration. The review suggests that the evidence base for integration versus targeted health pro- gramming is very limited.As success was measured in different ways (for example,in terms of varied programmatic goals related to efficiency, effectiveness, or equity), analysis and drawing lessons from this is further complicated. Perhaps the modesty of evidence creates the context for strong opinions for or against integration in global health. We have attempted to show that in practice the dichotomy between integrated and non-integrated (traditionally described as vertical and horizontal) is not rigid. Health systems combine both non-integrated and integrated interventions, but the balance of these interventions varies considerably. Our findings suggest that the purpose, nature and extent of integra- tion vary enormously between interventions and in countries, creating a rich mosaic of local solutions to address emergent problems. We found rare instances of full integration but a wide range of instances where health interventions are integrated into one or more critical health system functions.Further,it is important to note that in many countries that ix x Summary simultaneously have applied multiple health interventions, the nature of integration for these interventions varies and different degrees of integration for the same critical health system function co-exist. The findings provide new synthesis of evidence to inform the debate on health systems and targeted interventions. Given the highly varied contexts and adoption systems that reflect local nuances, different health system capacities and the range of problems being addressed, it is not surprising that in practice a rich mix of solutions exist. While the dis- cussion on the relative merits of integrating health interventions will no doubt continue, discussions should move away from the highly reductionist approach that has polarized this debate. Future efforts are best spent on generating and learning from useful evidence. CHAPTER 1 Introduction A longstanding debate on health systems organization relates to benefits of integrat- ing health programs that emphasize specific interventions into mainstream health systems to increase access and improve health outcomes. This debate has long been characterized by polarization of views and ideologies, with protagonists for and against integration arguing the relative merits of each approach (Cueto 2004; Magnussen, Ehiri, and Jolly 2004; Newell 1988; Walsh and Warren 1979; Warren 1988; Wisner 1988). Recently, the debate has been rekindled due to substantial increases in externally funded programs for priority health, nutrition and population interventions (such as those targeting communicable diseases, reproductive health interventions, and nutrition programmes--hereafter referred to as health interventions) and enhanced international efforts aimed at health systems strengthening (World Bank and World Health Organiza- tion 2006). However, all too frequently these arguments have not been based on hard evidence (Atun, Bennett, and Duran 2008). The presence of both integrated and non-integrated programs in many countries suggest there may be benefits to either approach, but the rel- ative merits of integration in various contexts and for different interventions have not been systematically analyzed and documented. Such an analysis is complicated as the term `inte- gration' is used to describe a variety of organizational arrangements in different settings (Atun, Bennett, and Duran 2008). Further, as the nature and extent of integration varies, there are methodological challenges to comparing various interventions. In this paper we present findings of a systematic review that explores a broad range of evidence on three critical aspects of integration: (i) the extent and nature of integration of targeted health programs that emphasize specific interventions into critical health systems functions (defined in the Methodology Section), (ii) how the integration or non-integration 1 2 World Bank Working Paper of health programs into critical health systems functions in different contexts have influenced program success, (iii) how contextual factors have affected the extent to which these pro- grams were integrated into critical health systems functions. To guide our analysis, we have used a new conceptual framework, discussed in detail in a complementary paper (Atun and others 2009), to guide our analysis. Our review evaluates peer-reviewed studies that focus on health interventions, and which have been introduced on a regional or national scale. This paper is organized in five chapters. This Introduction is followed by the method- ology chapter, which includes a brief section on the conceptual framework used to analyze the studies retrieved and the programs presented within these to map the nature and extent of integration into critical health system functions. The Results chapter includes: a sum- mary of the outcomes for each study grouped by the disease area or the clinical problem the intervention seeks to address, including the reported success; for each program, analy- sis and mapping of the nature and extent of integration into critical health system functions; and an analysis of how contextual factors either created opportunities for introducing or integrating a program or influenced the desirability or feasibility of program integration. The Discussion chapter provides an overview of the implication of findings for policy mak- ers, practitioners and researchers. The final chapter draws conclusions. CHAPTER 2 Methodology W e developed a search strategy based on the use of exploded-MeSH terms, sup- plemented with a broad search for keywords in the titles or abstracts for which no appropriate MeSH terms exist. In all cases, MeSH terms were chosen to rep- resent the highest order of relevance within each MeSH-tree. A systematic review by Briggs and Garner, which used Cochrane Systematic Review methodology, served as the basis for the development of the search strategy (Briggs and Garner 2006). The strategy combined two parts: the first was designed to identify articles related to organizational arrangements for health care delivery, and the second was designed to limit the search to specific areas identified as relevant to the World Bank health programs. The search strategy is shown in detail in Appendix A. We reviewed 8,274 potential articles, which yielded 55 articles that were included in this review (Figure 1). These papers are summarized in Appendix B. The Conceptual Framework for Analysing Integration of Targeted Health Interventions into Health Systems A separate paper (Atun and others 2009) provides details of the analytic framework1 on which this systematic review is based. Here, we provide a summary of the key features of the framework. 1. A longer version of the conceptual framework is available as a World Bank Health Nutrition and Population Discussion Paper. (Atun, R., Ohiri, K., Adeyi, O. Integration of Health Systems and Priority Health, Nutrition and Population Interventions: A Framework for Analysis and Policy Choices. August 2008. Available online at: http://siteresources.worldbank.org/HEALTHNUTRITIONANDPOPULATION/ Resources/281627-1095698140167/IntegrationHealthSystemsandInterventionsFramework2008.pdf 3 4 World Bank Working Paper Figure 1. Flow Chart Representing the Selection Process for Studies Included in the Review Database search 8,274 I. Title scanning 6,723 excluded 1,551 II. Abstract scanning 1,046 excluded 200 III. Full text scanning 28 not available 172 117 excluded 55 Included studies In this framework, we define integration as the extent, pattern, and rate of adop- tion and eventual assimilation of health interventions into each of the critical health system functions, which include inter alia: (i) stewardship and governance, (ii) financing, (iii) planning, (iv) service delivery, (v) monitoring and evaluation (M&E), and (vi) demand generation. An "intervention" in this context refers to combinations of technologies (for example, vaccines, drugs), inputs into service delivery, organizational changes and modifications in processes related to decision making, planning, and service delivery. We view a health intervention using a "diffusion of innovation" lens, and consider this as an idea, practice, or object that is perceived as new by an individual or a unit of adoption, while recognising that in some cases the interventions that have previously been implemented on a small scale are scaled up and increased in intensity. In such instances, the "newness" relates less to the technical element of the intervention itself but the organizational changes, new financing schemes and novel processes that accom- pany scaling up, intensification, integration and eventual institutionalization of the intervention. Drawing on relevant empirical evidence and theory (Atun and others 2009), we propose that the adoption and diffusion of new health interventions and the extent to which they are "assimilated" or "integrated" into the "general" health system will be influenced by five factors. These include the nature of the problem being addressed, the intervention, the adoption system which adopts and assimilates the intervention (key actors and institutions), the health system characteristics such as the absorptive capac- ity, and the broader context of the health system setting. Clearing the Global Health Fog 5 Table 1. Critical Health Systems Functions and Elements of Integration Critical Health System Function Elements Stewardship and governance ­ Accountability function ­ Reporting ­ Performance management Financing ­ Pooling of funds ­ Provider payment methods Planning ­ Needs assessment ­ Priority setting ­ Resource allocation Service Delivery ­ Structural ­ Human resources ­ Infrastructure ­ Operational integration ­ Referral and counter-referral systems ­ Guidelines or care pathways ­ Procurement ­ Supply chain management Monitoring and Evaluation ­ Information management and technology ­ Data collection and analysis Demand Generation ­ Incentives--financial (e.g. conditional cash transfers, health or social insurance) and non- financial (peer recognition, non-monetary awards) ­ Population interventions--e.g. education and promotion The conceptual framework provides a basis for evaluating these five constituents with respect to the purpose, extent and nature of integration of the health intervention(s) under study into critical health system functions. We consider integration of elements of a health intervention or program with crit- ical health system functions, to include, inter alia, stewardship and governance, financ- ing, planning, service delivery, M&E, and demand generation (Table 1). With respect to extent, we identify whether the integration is full, partial or non-existent, and by level we refer to integration of these functions at local (provider unit), district, regional or national tiers of the health system. Given the broad mix of outcome measures used in different studies it was difficult to directly compare the relative success of programs and interventions. For this analysis, we attempted to ascertain the "success" for each program by analyzing the reported achievements against the outcome measures and indicators predefined at the start of the study. We assigned a "high success" rating to a program if all or most of the explicitly stated objectives were achieved in the study, even if the number of reported outcome measures was limited. CHAPTER 3 Results Outcomes Reported in the Studies Analyzed In this section we present all interventions grouped by the disease area or outcomes they seek to address. Data underpinning the assertions and narrative in this section are pre- sented in Appendix B, which includes a summary of the studies evaluated. Neglected Tropical Diseases The control of neglected diseases in developing countries has been predominantly carried out through specific interventions that have targeted one or more of these diseases with limited integration into mainstream health system functions. Dengue. To build a sustainable national program for dengue control, areas of Cuba integrated community working groups (CWG), which included Primary Health Care (PHC) workers, into the existing program that had limited integration into the mainstream health system (Toledo Romani and others 2007). Following integration, areas that had CWG involvement reported greater improvements in entomological indices and were able to maintain these effects with community level support for ongoing control though no data was provided on incidence. Malaria. Certain areas in Colombia strengthened community participation and net- works to improve malaria control (Rojas, Botero, and Garcia 2001). The program, which was managed by a Central Coordinating Committee and which was centered on educa- tion, provision of diagnosis and treatment, as well as vector control at the local level, suc- ceeded in decreasing malaria incidence by 45 percent. However, there were no control sites 7 8 World Bank Working Paper for comparison or figures on changes in malaria incidence in the periods that preceded the introduction of the program. Schistosomiasis. Various strategies have been implemented to control schistosomia- sis. For example Brazil, Burundi, Cameroon, and Saudi Arabia have integrated some tar- geted interventions into primary care structures or community centers (Coura Filho and others 1992; Engels, Ndoricimpa, and Gryseels 1993; Engels, Sindayigaya, Gryseels 1995; Bausch and Cline 1995; Cline and Hewlett 1996; Ageel and Amin 1997; Jarallah and oth- ers 1993). In contrast, Uganda and China have adopted essentially vertical structures (Kabatereine and others 2006; Sleigh and others 1998a, 1998b, 1998c). In the case of Brazil (Coura Filho and others 1992), integrated strategies helped to suc- cessfully reduce the incidence and prevalence of schistosomiasis infections at rates similar to those reported in targeted interventions without health system integration. In Saudi Arabia (Ageel and Amin 1997; Jarallah and others 1993) integration of tar- geted interventions into PHC led to a greater decline in the incidence, prevalence and intensity of infections than that achieved by targeted interventions with no integration. In Cameroon population knowledge about schistosomiasis and the utilization of health centers increased after introduction of a program that integrated targeted interven- tions into PHC (Bausch and Cline 1995; Cline and Hewlett 1996). In Burundi, an integrated model of control resulted in an annual yield of 60 percent of cases detected by the PHC services. The participation rate in the integrated model (60 percent) was lower than in the targeted interventions (80 percent), with no integration aimed at selective population treatment. Under the integrated approach, the proportion of cases treated declined from 16 to 10 percent. However, these outcomes were achieved at an eighth of the cost of targeted interventions with no integration (Engels, Ndoricimpa, and Gryseels 1993). In China, programs that were predominantly non-integrated in nature but which enjoyed strong community participation were successful in eradicating schistosomiasis in all endemic areas at an annual cost of less than US$ 0.50 per capita for the protected pop- ulation (Sleigh and others 1998a, 1998b, 1998c). In Uganda (Kabatereine and others 2006) a large pilot project demonstrated the viability of using schools as channels for drug distribution and the feasibility of using community-directed treatment for mass drug distribution to reach poor communities in remote areas. Onchocerciasis. A study in Uganda demonstrated that provision of additional PHC tasks by community health workers, such as malaria or tuberculosis (TB) control, family planning, or immunization, was positively correlated with an increase in the quality of treatment for onchocerciasis (Katabarwa and others 2005). Leprosy. In many countries, leprosy services have historically been delivered as targeted interventions with no integration. An independent evaluation (data not pre- sented) is reported to have found that in Tamil Nadu, India, following integration, the number of detected cases increased slightly, but also that, as follow-up and treatment completion were no longer monitored, completion rates declined (Rao and others 2002). Clearing the Global Health Fog 9 A pilot project in Andhra Pradesh, India showed negligible differences in leprosy prevalence and newly detected cases before and after integration. After integration, there was an increase in detection of previously hidden cases, but the rate of case-holding (adherence to treatment in accordance with program guidelines) declined (Parkash and Rao 2003). Whereas, a comparison of leprosy programs in the area of Jamkhed (integrated) and in the neighboring district of Osmanabad (targeted without integration) in Maharashtra, India showed that, in communities with an integrated approach to leprosy care, social stigma experienced by leprosy patients (reduction of which was a program objective) was greatly diminished (Arole and others. 2002). In Sri Lanka, integration of targeted interventions for leprosy into PHC was associated with higher case detection reflected by a 36 percent increase in prevalence and 41 percent increase in new cases (Kasturiaratchi, Settinayake, and Grewal 2002). Nutrition The nutrition programs studied included general programs to address undernutrition and programs providing micronutrient supplementation. The Bangladesh Integrated Nutrition Project (BINP) was designed to provide education, basic care and nutritional services to children and mothers. An evaluation concluded that the BINP had failed to achieve its objectives (reducing prevalence of severe undernutrition by 40 percent and moderate undernutrition by 25 percent). Although the mothers in the intervention area reported better care practices than those in the control area, there was no substantial difference in the rate of malnutrition between the two groups of children (Hossain, Duffield, and Taylor 2005). Multi-country studies on micronutrient supplementation (including vitamin A, iodine and iron) demonstrate that in many countries these projects were successfully launched, implemented and sustained, but that campaign-based interventions (e.g. vita- min A supplementation during immunization days) were more successful in achieving their objectives than facility-based models (e.g. iron distribution at antenatal care clinics) (Deitchler and others 2004). In Peru, compared to a non-integrated food-distribution program, a weekly multi- micronutrient supplementation program with an integral communication campaign and community involvement in education and distribution targeting children, women and adolescent girls of childbearing age, improved coverage and led to an increase in the knowl- edge on the beneficial effects of supplementation (Gross, Mamani Diaz, and Valle 2006a; Gross, Valle, Mamani Diaz 2006b; Lechtig and others 2006a, 2006b; Lopez de Romana and others 2006). In addition, it had a protective effect on the haemoglobin levels of the three target groups, and was more cost-effective and better targeted to the population needs. Immunization Evaluation of mass immunization campaigns (targeted interventions with no integration) in three different contexts revealed contrasting results. In rural India, the Pulse Polio Immunization (PPI) campaign, which was launched in 1995, aimed at promoting social mobilization and immunizing the population in remote areas of the country during National Immunization Days (NID). This program led to a significant increase in immunization 10 World Bank Working Paper coverage for the first dose of polio vaccine. However, the coverage declined between the first and the third dose, and there was no change in inequities related to gender, caste, wealth status, religion, and geography for polio immunization or for other non-polio Expanded Program for Immunization (EPI) immunizations before and after the PPI cam- paigns (Bonu, Rani, and Baker 2003). In Sierra Leone, mass immunization campaigns led to high coverage rates for those vaccines requiring only one dose (90 percent BCG, 62 percent measles), with lower coverage rates for the vaccinations requiring three doses (59 percent DPT, 58 percent polio, 76 percent tetanus). With each full dose of vaccines there was a significant decline in infant and child mortality across all the socio-economic groups (Amin 1996). Two mass immunization campaigns in seven refugee camps in the Former Yugoslav Republic of Macedonia during the military conflict in 1998, achieved coverage rates of around 90 percent of children, but no other outcomes were reported (Koop, Jackson, and Nestel 2001). Child Health and Development Integrated Management of Childhood Illness. Studies that compared the effectiveness of the Integrated Management of Childhood Illness (IMCI) strategy with routine care com- prising a series of non-integrated programs, have reported positive impacts on the quality of care delivered in a range of settings. In Bangladesh, the quality of care (i.e. adherence to IMCI guidelines), care-seeking behavior and utilization of government health facilities significantly improved in facilities providing IMCI when compared with 20 paired facilities that had not implemented IMCI guidelines (El Arifeen and others 2004). In South Africa, introduction of IMCI in selected districts led to improved process and health outcomes (as measured by a quality index based on the quality indicators developed by WHO) (Chopra and others 2005), while in Tanzania it led to significantly improved case management and prescribing and a reduction of under-5 mortality rates as compared with those in districts where IMCI was not introduced (Armstrong Schellenberg and others 2004; Masanja and others 2005)--as well as improvements in equity, immunization cov- erage, caretaker knowledge and behavior, nutritional status and morbidity (Masanja and others 2005). Additional studies, which examined the total and incremental start up and implementation costs of IMCI (Adam and others 2005) and the costs borne by various health system levels in relation to a composite measure for correct management of child- hood illness (Bryce and others 2005a), demonstrated that implementation of IMCI was not associated with higher costs than routine care, but led to significant improvement in case management compared to controls that used routine training approaches. In Uganda introduction of IMCI led to improvements in the performance of trained health workers and the quality of care delivered to children aged less than five years (Pariyo and others 2005), while in China, implementation of IMCI helped to strengthen the local health system and improved communications between physicians and patients (Zhang, Dai, and Zhang 2007). Similarly, in Brazil (Amaral and others 2004) and Morroco (Naimoli and others 2006), introduction of IMCI led to significant improvements in quality of care (e.g. weight checked against growth chart; checking for at least three danger signs; and eval- uation of feeding practices in Brazil and adherence to guidelines and correct prescription Clearing the Global Health Fog 11 of antibiotics in Morroco ) as compared with routine care provided in facilities where IMCI was not introduced. By contrast, in Peru, implementation of IMCI did not positively affect the quality of care and utilization of services in facilities where it was implemented (Huicho and others 2005a, 2005b). Integrated Child Development Services. The Indian Integrated Child Development Services (ICDS) program (which comprises a package of education services, nutritional support for pregnant and nursing women and children, and links to other primary care based health services for children) has been evaluated in a number of studies. Districts that adopted the ICDS scheme demonstrated improvements in the quality and quantity of services offered, improved coverage of immunization and a marked reduction in mal- nutrition in children aged under six years (Lal 1980), as well as higher uptake of health- care services such as immunization and antenatal care, and significant improvement in feeding practices and nutritional and health status of children (Gupta and others 1984). Addition of therapeutic food supplementation and nutritional support to ICDS led to declines in malnutrition levels in severely malnourished children (Kapil and others 1999), improvements in weight gain for pregnant women and reductions in the number of pre-term and low birth weight deliveries in ICDS districts compared to non-inter- vention sites (Agarwal and others 2000). ICDS services, when delivered in a coordinated manner with EPI, led to significant increases in overall immunization coverage with greater improvement in knowledge, attitudes and practices of carers with regard to immunization in intervention districts compared with districts where ICDS and EPI were delivered separately (Tandon and others 1992; Tandon and Sahai 1988). One study demonstrated a strong correlation between the degree of utilization of all ICDS services and nutritional status and morbidity of children between 0­36 months of age (Saiyed and Seshadri 2000). In contrast, one study found no statistically significant impact of ICDS on malnutrition levels in ICDS interventions sites but observed higher rates of cov- erage for DPT and measles immunization in non-ICDS districts (Trivedi, Chapararwal, and Thora 1995). Reaching intended target groups for ICDS has proved difficult. In response, the Dular program developed by UNICEF-India has built on the existing ICDS infrastructure with an enhanced package of education, training of health workers and active outreach to the community. This program was successful in enhancing the effectiveness of ICDS and led to statistically significant improvements in all measured outcomes (for example, iodized salt use, prenatal care, colostrum feeding, and use of a delivery kit for childbirth) between Dular and non-Dular villages (Dubowitz and others 2007). Family Planning Services Matlab Family Planning (FP) and Health Services Project. The "Matlab FP and Health Services Project" (FPHSP), initiated by the International Center for Diarrhoeal Disease Research, Bangladesh (ICDDR,B), which combined FP services with a basic package of Maternal and Child Healthcare (MCH) was more cost-effective in stimulating uptake of modern contraceptive services than the pre-existing governmental program which pro- vided these services in a non-integrated manner (Simmons, Balk, and Faiz 1991). How- ever, extension of the original program with provision of a wider range of MCH services 12 World Bank Working Paper did not increase uptake of contraceptive services when compared with the original pro- gram (DeGraff and others 1986; Phillips and others 1984). Lady Health Worker Program. The Pakistan National Program of Family Planning and Primary Health Care, more commonly known as the Lady Health Worker Program (LHWP), delivers a bundle of integrated services related to MCH and FP door-to-door in rural areas. It has achieved significantly higher use of reversible, modern contraceptive methods in areas served compared to routine practice where these services are provided in primary care centers (Douthwaite and Ward 2005). Combined Family Planning and MCH Programs with Basic Health Services. A study in Nepal compared the effectiveness of two FP/MCH programs; one providing only FP/MCH services and the other (the `integrated' program) combining the range of FP/MCH services with other basic health services that included immunizations and case finding and treat- ment for TB and leprosy. While the non-integrated FP/MCH services achieved greater impact on knowledge of FP, intention to use FP services and infant mortality, the integrated program had a relatively greater impact on child survival (Tuladhar and Stoeckel 1982). HIV/AIDS The GHESKIO clinic in Haiti was set up to provide Voluntary Testing and Counselling (VCT) and care for patients with HIV/AIDS (Peck and others 2003). It has gradually added other primary care services for communicable diseases and reproductive health. The inte- gration of on-site primary care services was accompanied by an increased demand for VCT. Patients attending the clinic could receive coordinated care for co-morbidities. The Extent and Nature of Integration of Health Interventions into Critical Health Systems Functions We present below an analysis that describes, for each intervention included in the study, how the intervention elements are integrated into the six critical health system functions. Further, using published data from the studies, we map the nature and extent of integra- tion of intervention elements into critical health system functions. These are further dis- cussed below and summarized in Figure 2. Our analysis draws on the data available in the published papers. The mapping of the extent of integration reflects the actual situation as reported in the published paper in the period under concern rather than what was intended. Stewardship and Governance This dimension captures aspects related to distribution of governance responsibilities, accountability and performance management for an intervention. We considered full inte- gration to have occurred when the governance arrangements for the targeted health interven- tion were the same as those for the general health services or the local/national administrative structures. For example, in a number of settings interventions for schistosomiasis control Clearing the Global Health Fog 13 Figure 2. The Extent and Nature of Integration by Targeted Health Intervention and Intervention Success as Reported in the Study Level of Integration Degree of Success Fully integrated Most to all outcomes nance Partially integrated Mixed outcomes ver aluation Not integrated Few to no outcomes Go y ation & er Ev & Unknown Unknown deliv gener ing wardship vice Dengue Ste FinancingPlanningSer MonitorDemand Success Cuba (Toledo Romani 2007) Malaria Colombia (Rojas 2001) Schistosomiasis control Brazil (Filho 1992) Burundi (Engels 1993, 1995) Cameroon (Clino 1996) China (Sleigh 1998) Saudi Arabia (Ageel 1997, Jarallah 1993) Uganda (Kabatoroine 2006) Leprosy India (Rao 2002, Thakar 2003) Sri Lanka (Kasturiaratchi 2002) Nutrition Peru (Gross2006ab, Lachtig2006ab, Lopez de Romana 2006) Bangladesh (Hossain 2005) Various (Doitchler 2004) Immunisation Sierra Loone (Amin 1996) India (Bonu 2003) Macedonia (Koop 2001) Child health & development IMCI* ICDS (Agarwal 2000, Gupta 1984, Kapil 1999, Lal 1980, Trivedi 1995, Saiyed 2000) ICDS + Dolar (Dubowitz 2007) Family planning services Bangladesh - FPHSP (Philips 1984, de Graff 1986) Pakistan - LHWP (Douthwaite 2005) Nepal (Tuladhar 1982) HIV/AIDS & STD services Haiti (Pack 2003) 14 World Bank Working Paper adopted a fully integrated governance structure--such as that in Brazil, where the inter- vention was managed by the municipality (Coura Filho and others 1992); in Cameroon, where the PHC service was accountable for schistosomiasis control (Cline and Hewlett 1996) and Saudi Arabia, where the PHC team cooperated and shared responsibility with the district supervisor and regional health authorities (Ageel and Amin 1997; Jarallah and others 1993). In India the governance structures for interventions aimed at leprosy con- trol were successfully integrated with the PHC services (Rao and others 2002), while in Nepal governance structures for the FP/MCH intervention were integrated with services provided in the district health offices (Tuladhar and Stoeckel 1982). Partial integration occurs where responsibility is shared by the existing general health- care system and a specific structure created purposely for the intervention. For instance, in Cuba composite entities comprising PHC level and vector control team workers jointly managed the interventions for dengue control (Toledo Romani and others 2007), whereas in Colombia a Central Coordinating Committee consisting of representatives of the regional health services, NGOs, scientific centers, and local politicians was formed to man- age and oversee the interventions for malaria control (Rojas, Botero, and Garcia 2001). In Sri Lanka, the responsibility for managing interventions for leprosy control moved between the local health authorities and the Central Leprosy Clinic personnel (Kasturiaratchi, Settinayake, and Grewal 2002). In India anganwadi centers, created as part of a self-managed system for child health and development, were also responsible for organization of a wide range of primary care services (Agarwal and others 2000; Gupta and others 1984; Kapil and others 1999; Lal 1980; Saiyed and Seshadri 2000; Tandon and others 1992; Tandon and Sahai 1988; Trivedi, Chaparwal, and Thora 1995). This review considers the governance role to be non-integrated when accountability remains exclusively with dedicated specialist entities that are charged with implementation and management of the targeted health interventions, without involvement of the general healthcare system--for example, interventions directly managed by dedicated units within national or regional governments without integration into main health system functions as with the interventions for schistosomiasis control in China (Sleigh and others 1998a, 1998b, 1998c) and Uganda (Kabatereine and others 2006), nutrition-related campaigns in Peru (Gross, Mamani Diaz, and Valle 2006a; Gross, Valle, and Mamani Diaz 2006b; Lechtig and others 2006a, 2006b; Lopez de Romana and others 2006) and various Asian countries (Deitchler and others 2004; Hossain, Duffield, and Taylor 2005); the LHWP in Pakistan (Douthwaite and Ward 2005); and the FPHSP in Bangladesh (DeGraff and oth- ers 1986; Phillips and others 1984). While non-governmental organizations (NGOs) and external donors work with regional or national health services, they often retain direct governance roles beyond financing or service delivery, as with the stand-alone HIV/AIDS clinic in Haiti offering vol- untary counseling and testing with additional services for HIV/AIDS STI and reproductive health services (Peck and others 2003); the nutrition interventions in Peru (Gross, Mamani Diaz, and Valle 2006a; Gross, Valle, and Mamani Diaz 2006b; Lechtig and others 2006a, 2006b; Lopez de Romana and others 2006) (PISA); the mass immunization campaigns in Sierra Leone implemented by UNICEF (Amin 1996) and in Macedonia implemented by International Medical Corps (Koop, Jackson, and Nestel 2001); or the Dular program in India managed by UNICEF (Dubowitz and others 2007). Clearing the Global Health Fog 15 Financing For the purposes of this review, financing refers to both the pooling of financial resources and the provider-payment methods used to allocate these. Revenue generation, a critical function of financing, is beyond the scope of the review. We considered an intervention to be fully integrated into the financing function if it was funded entirely through the national or regional general healthcare budget. Examples of full integration of financing include interventions for schistosomiasis control in Brazil (Coura Filho and others 1992) and interventions for leprosy control in India (Rao and oth- ers 2002) and Sri Lanka (Kasturiaratchi, Settinayake, and Grewal 2002). Partial integration into the financing function was achieved with interventions for schis- tosomiasis control in Cameroon, where earmarked funding was provided by the United States Agency for International Development (USAID) but channeled through the PHC system (Cline and Hewlett 1996). Under the ICDS scheme in India resources were pro- vided by the national government directly to the intervention which comprised a range of essential health services to supplement other general services locally provided and funded by local and national governments (Agarwal and others 2000; Gupta and others 1984; Kapil and others 1999; Lal 1980; Saiyed and Seshadri 2000; Tandon and others 1992; Tandon and Sahai 1988; Trivedi, Chaparwal, and Thora 1995). When financing was provided directly to an intervention and addressed only a partic- ular disease or problem, the function was considered to be non-integrated. In some instances interventions were directly funded by the government, either at local level, for example dengue control in Cuba (Toledo Romani and others 2007), or at national level, for example malaria control in Colombia (Rojas, Botero, and Garcia 2001) and the LWHP in Pakistan (Douthwaite and Ward 2005). Many interventions were directly funded by external donors: such as the BINP funded by the International Development Association of the World Bank (Hossain, Duffield, and Taylor 2005); the Ugandan National Schistosomiasis Control Program funded from the Bill & Melinda Gates Foundation through its collaboration with the international Schistosomiasis Control Initiative (Kabatereine and others 2006); the Dular ICDS financed by UNICEF-India; the micronutrient supplementation programs in many Asian countries supported by the Canadian International Development Agency and UNICEF (Deitchler and others 2004), and in Peru by the NGO PROMESA (Gross, Mamani Diaz, and Valle 2006a; Gross, Valle, and Mamani Diaz 2006b; Lechtig and others 2006a, 2006b; Lopez de Romana and others 2006); and the immunization program in Sierra Leone funded by a range of external donors (Amin 1996). Planning In our framework, the planning function includes activities, processes and systems for needs assessment, priority setting, and resource allocation. We consider planning for an intervention to be fully integrated if the decision-making in relation to the above three areas is undertaken by institutions/stakeholders who are involved in the same tasks for the general health system. Examples of full integration include community leadership in schis- tosomiasis control in Cameroon (Cline and Hewlett 1996) and in India integration of deci- sion making for leprosy control with the PHC system (Rao and others 2002). 16 World Bank Working Paper Partial integration occurs when the decision-making responsibility for planning is retained by those managing the health intervention but involve a range of stakeholders (such as civil society representatives, PHC level, or local/regional/national government) through inclusive groupings as illustrated by the Community Working Groups for dengue control in Cuba (Toledo Romani and others 2007), the Central Coordinating Committee for malaria control in Colombia (Rojas, Botero, and Garcia 2001) and interventions for leprosy control in Sri Lanka where the responsibility for planning rested with the PHC ser- vices but relied heavily on the workers from the program (Kasturiaratchi, Settinayake, and Grewal 2002). Planning was considered to be non-integrated when the decision-making focused solely on the intervention without consideration of general healthcare activities. This may include specific national government units at national level, as in the schistosomiasis con- trol project in China (Sleigh and others 1998a, 1998b, 1998c) or the LHWP in Pakistan (Douthwaite and Ward 2005); dedicated units at national and local levels, as with the micronutrient supplementation programs in several Asian countries (Deitchler and oth- ers 2004); or NGOs and external donors--for example, PISA in the micronutrient sup- plementation interventions in Peru, UNICEF in the Dular program in India (Dubowitz and others 2007) and the International Medical Corps in Macedonia (Koop, Jackson, and Nestel 2001). Service Delivery Service delivery relates to structural and organizational dimensions of a particular inter- vention, either at or close to the interface with the customer. In our analysis services within a health intervention are considered to be fully integrated if their provision is the respon- sibility of general or multi-purpose health workers, as with the schistosomiasis control in Burundi and Cameroon (Cline and Hewlett 1996; Engels, Ndoricimpa, and Gryseels 1993), leprosy care in Sri Lanka (Kasturiaratchi, Settinayake, and Grewal 2002) and ICDS services in India (Agarwal and others 2000; Gupta and others 1984; Kapil and others 1999; Lal 1980; Saiyed and Seshadri 2000; Tandon and others 1992; Tandon and Sahai 1988; Trivedi, Chaparwal, and Thora 1995). Partial integration refers to instances where there is shared responsibility for the pro- vision of services between general health workers and the staff of the targeted health inter- vention; as with interventions for dengue control in Cuba (Toledo Romani and others 2007), schistosomiasis control in Brazil (Coura Filho and others 1992) and Saudi Arabia (Ageel and Amin 1997; Jarallah and others 1993), and the interventions for leprosy con- trol in India which after integration with PHC services were provided by the former staff of leprosy centers as the general purpose PHC workers were insufficiently trained to take over this responsibility (Rao and others 2002). Partial integration was also achieved in the ICDS intervention in India through collaboration of anganwadi workers and purpose- trained volunteers (Agarwal and others 2000; Kapil and others 1999; Tandon and others 1992; Tandon and Sahai 1988). Partial integration also occurred when service delivery for a number of interventions were linked; for example family planning and maternal and child health services (DeGraff and others 1986; Phillips and others 1984; Tuladhar and Stoeckel 1982) or integration of general health services with HIV/AIDS VCT services in Haiti (Peck and others 2003). In Uganda, National Immunization Days were used as a Clearing the Global Health Fog 17 vehicle to deliver interventions for schistosomiasis control (Kabatereine and others 2006). The IMCI strategy is a good example of partial integration where interventions for man- agement of a number of childhood illnesses are bundled together (Bryce and others 2005b), but can be administered either separately through community workers trained exclusively in IMCI, or through PHC facilities using general staff. A number of interventions rely solely on single purpose workers and have no integra- tion with other interventions or general health services; such as the interventions for malaria control in Colombia (Rojas, Botero, and Garcia 2001), interventions for schisto- somiasis control in China (Sleigh and others 1998a, 1998b, 1998c), or the immunization for polio in India (Bonu, Rani, and Baker 2003) and other childhood illnesses in Macedonia (Koop, Jackson, and Nestel 2001). Monitoring & Evaluation The M&E function of a targeted health intervention was considered to be fully integrated if the responsibility for this rested with institutions that retained overall responsibility for M&E in the health system--for example, with interventions for schistosomiasis control in Cameroon and Saudi Arabia, monitored respectively by the Ministry of Public Health and the district supervisor of the local PHC services (Ageel and Amin 1997; Cline and Hewlett 1996; Jarallah and others 1993), or the interventions for leprosy control in Sri Lanka mon- itored by the regional PHC services (Kasturiaratchi, Settinayake, and Grewal 2002). In con- trast, M&E of interventions for malaria control in Colombia was partially integrated as it was undertaken jointly by staff from the regional health services and the malaria control program (Rojas, Botero, and Garcia 2001). For many health interventions M&E is not integrated with mainstream M&E systems. The interventions for schistosomiasis control in China used a dedicated parallel system; the international sponsor for schistosomiasis control in Uganda employed its own systems for M&E (Sleigh and others 1998a, 1998b, 1998c); the nutrition interventions in Peru man- aged by an international organization, PISA undertook its own M&E with the assistance of independent institutions (Gross, Mamani Diaz, and Valle 2006a; Gross, Valle, and Mamani Diaz 2006b; Lechtig and others 2006a, 2006b; Lopez de Romana and others 2006); and the immunization program in Macedonia was monitored directly by volunteers of the implementing organization (Koop, Jackson, and Nestel 2001). Demand Generation Demand generation is increasingly recognized as an important health system function. It relies on a number of interventions such as the use of appropriate financial incentives and monetary support such as insurance, or of information, education and communication (IEC) activities designed to change behavior. Demand generation was considered to be fully integrated if mechanisms used to cre- ate financial incentives or IEC activities were provided jointly with the general services or were delivered by PHC workers. For example, in Burundi, Cameroon and Saudi Arabia, IEC for schistosomiasis control was the responsibility of staff of the public health centers (Ageel and Amin 1997; Cline and Hewlett 1996; Engels, Ndoricimpa, and Gryseels 1993; Jarallah and others 1993) and in Sri Lanka IEC activities were considered a fundamental 18 World Bank Working Paper part of the integration strategy (Kasturiaratchi, Settinayake, and Grewal 2002). A particu- larly interesting example of integration of demand generation was in Haiti, where VCT activities for HIV/AIDS at a clinic were used to generate demand for related services for STIs, reproductive health and basic PHC (Peck and others 2003). In Colombia, education on the prevention, diagnosis and treatment of malaria was provided both by community volunteers working in malaria control and staff from the general health services (Rojas, Botero, and Garcia 2001). In Peru and Bangladesh health education for nutrition interventions was jointly provided by staff working for the targeted intervention and regional health workers (Gross, Mamani Diaz, and Valle 2006a; Gross, Valle, and Mamani Diaz 2006b; Hossain, Duffield, and Taylor 2005; Lechtig and others 2006a, 2006b; Lopez de Romana and others 2006). In China, information and education related to schistosomiasis control was provided through a partnership between staff work- ing on schistosomiasis control, general health workers, schoolteachers and community health workers (Sleigh and others 1998a, 1998b, 1998c). Similarly, staff involved in the LHWP in Pakistan and the ICDS in India provided a wide range of IEC activities on gen- eral health, women's health, family planning, maternal care and child health in addition to those related to their specific interventions (Agarwal and others 2000; Douthwaite and Ward 2005; Gupta and others 1984; Kapil and others 1999; Lal 1980; Saiyed and Seshadri 2000; Tandon and others 1992; Tandon and Sahai 1988; Trivedi, Chaparwal, and Thora 1995). In contrast, in a number of countries, information campaigns related to health interventions tended to be stand-alone activities, focusing solely on a single problem or disease, and delivered by single-purpose health workers or volunteers: for example for dengue control in Cuba (Rojas, Botero, and Garcia 2001), schistosomiasis control in the Ugandan National Control Program (Kabatereine and others 2006) and for childhood immunization in Macedonian refugee camps (Koop, Jackson, and Nestel 2001). The Extent of Intervention Success in the Studies Analyzed We have summarized in Figure 2 the extent of outcome success for each study included in the review, and relate this to the nature and level of integration. We assigned a "high success" rating to a program if all or most of the explicitly stated objectives were achieved in the study, even if the number of reported outcome measures was limited. A number of programs was given a `high success' rating, namely the Cuban dengue control program (Toledo Romani and others 2007); the malaria control program in Colombia (Rojas, Botero, and Garcia 2001), the schistosomiasis programs in Brazil (Coura Filho and others 1992), Cameroon (Cline and Hewlett 1996), China (Sleigh and others 1998a, 1998b, 1998c), Saudi Arabia (Ageel and Amin 1997; Jarallah and others 1993) and Uganda (Kabatereine and others 2006); the immunization campaigns in Sierra Leone (Amin 1996) and Macedonia (Koop, Jackson, and Nestel 2001); the Indian ICDS and Dular programs (Agarwal and others 2000; Dubowitz and others 2007; Gupta and others 1984; Kapil and others 1999; Lal 1980; Saiyed and Seshadri 2000); the LHWP in Pakistan (Douthwaite and Ward 2005) and the HIV/AIDS VCT clinic in Haiti (Peck and others 2003). We assigned a "moderate success" rating for programs that reported success against a number of outcome measures identified at the outset of the study but failure against others. Clearing the Global Health Fog 19 Examples of these are the schistosomiasis program in Burundi (Engels, Ndoricimpa, and Gryseels 1993), the integrated leprosy services in Sri Lanka (Kasturiaratchi, Settinayake, and Grewal 2002) and India (Rao and others 2002), the immunization program in India (Bonu, Rani, and Baker 2003) and the Peruvian nutrition program (Gross, Mamani Diaz, and Valle 2006a; Gross, Valle, and Mamani Diaz 2006b; Lechtig and others 2006a, 2006b; Lopez de Romana and others 2006). The nutrition programs in a number of Asian coun- tries and South Africa were not individually reported and were collectively assigned a "moderate success" rating to reflect varying levels of success achieved in different settings (Deitchler and others 2004). Programs that failed to achieve most or all of the intended outcomes were assigned a "low success" rating. These included the Bangladesh Integrated Nutrition Project (Hossain, Duffield, and Taylor 2005), the extended form of the--otherwise successful--FPHSP in Bangladesh (DeGraff and others 1986; Phillips and others 1984) and the integrated strategy for FP/MCH services in Nepal (Tuladhar and Stoeckel 1982). How the Context Influences the Extent and Nature of Integration To better understand which factors have contributed to success or failure of health inter- ventions in a number of settings, we analyzed the context in which these interventions were implemented, in particular factors pertaining to the adoption system, the health system and those relating to the wider politico-economic and socio-cultural context. Collectively, these factors interact to either produce opportunities, or create barriers, or generate recep- tive contexts thereby influencing the desirability and sustainability of an intervention. These are summarized in Appendix C. Sustainability A changing socio-demographic and economic context impacts on population needs and affects sustainability of programs. For example, in Peru rapid urbanization and arrival of large numbers of indigent slum-dwellers significantly altered the morbidity profile of the local population with a substantial rise in micronutrient deficiencies. This created a sig- nificant new burden that had to be addressed by targeted interventions (Gross, Mamani Diaz, and Valle 2006a; Gross, Valle, and Mamani Diaz 2006b; Lechtig and others 2006a, 2006b; Lopez de Romana and others 2006). In contrast, a rise in living standards in Saudi Arabia which led to improvements in general hygiene and sanitation helped reduce the burden of schistosomiasis (Ageel and Amin 1997; Jarallah and others 1993). Understanding local socio-cultural context is critical to the design of an appropriate intervention as illustrated by the schistosomiasis control in Uganda. This intervention ini- tially targeted school children, but, as up to 30 percent or more of children were not enrolled in schools in certain endemic areas, the intervention failed to reach large popu- lation groups (Kabatereine and others 2006). The intervention was then modified to use a wider range of opportunities, such as National Immunization Days, for deworming activities. While the disease-specific focus of interventions for the control and eradication of lep- rosy (India, Sri Lanka) or schistosomiasis (such as in Burundi and Saudi Arabia) was 20 World Bank Working Paper appropriate when these diseases were highly endemic, success of these interventions in reducing the disease burden meant that strategies which were previously cost-effective became financially unsustainable in the new epidemiological context. Given the low endemicity and prevalence, the more financially sustainable option was to integrate these interventions into mainstream PHC services. In Brazil, integration of the intervention for schistosomiasis control into PHC was pos- sible as municipalities possessed enough absorptive capacity to assume responsibility for planning, financing and delivery of the intervention when external funding ceased (Coura Filho and others 1992). Similarly, the presence of a well-developed PHC infrastructure with high population coverage and utilization rates enabled Saudi Arabia to embed schis- tosomiasis control into PHC (Ageel and Amin 1997; Jarallah and others 1993), and allowed integration of IMCI into the general health services in Peru (Huicho and others 2005a; Huicho and others 2005b), South Africa (Chopra and others 2005) and Tanzania (Armstrong Schellenberg and others 2004; Masanja and others 2005). The FPHSP intervention in Bangladesh was able to provide efficient family planning services for many years thanks to its dedicated human and physical resources (Simmons, Balk, and Faiz 1991). On the other hand, the LHWP program in Pakistan struggled to find a place in a fragile health system which was already overstretched and experienced signifi- cant funding shortfalls (Douthwaite and Ward 2005). In contrast, the Dular strategy for health education and child health services was successfully sustained as it managed to build on the infrastructure of the ICDS at low cost (Dubowitz and others 2007). Opportunity and Necessity Critical events create windows of opportunity or a necessity for action, by mobilizing civil society and other key actors (such as health professionals and policy makers) to introduce new systems and designs for finance and delivery of health interventions. For example, an outbreak of dengue in a non-endemic area of Cuba motivated the local community to strengthen existing interventions for dengue control (Toledo Romani and others 2007). In Cameroon, a number of development projects raised concerns about the possible expansion of snail habitats with a concomitant increase in schistosomiasis infection, prompting the government to establish a dedicated intervention for schistosomiasis con- trol with strong community involvement and integration into PHC. This intervention was in line with the Government's commitment to strengthen the national PHC system (Cline and Hewlett 1996). Confronted with unsustainably high rates of population growth, the Government of Bangladesh introduced the FPHSP program to provide family planning education and contraceptive services (DeGraff and others 1986; Simmons, Balk, and Faiz 1991). Over- crowding in refugee camps that followed the military conflict in Macedonia increased the risk of rapid transmission of vaccine preventable diseases, necessitating urgent implemen- tation of an EPI for children (Koop, Jackson, and Nestel 2001). The Peruvian nutrition supplementation program was set up in response to evidence which showed a high preva- lence of iron-deficiency in the newly urbanized population (Gross, Mamani Diaz, and Valle 2006a; Gross, Valle, and Mamani Diaz 2006b; Lechtig and others 2006a, 2006b; Lopez de Romana and others 2006). Clearing the Global Health Fog 21 Synergies between targeted health interventions can create opportunities for integra- tion and can positively influence the ultimate success of these interventions. For example, positive synergies between FP and MCH services in Bangladesh (DeGraff and others 1986; Phillips and others 1984), HIV/AIDS VCT, STI and reproductive health services in Haiti (Peck and others 2003), and the ICDS and EPI in India helped to mutually reinforce effec- tiveness of each intervention with improved results. In contrast, poor coordination between related interventions can be detrimental to effectiveness and sustainability. For example, in Morocco multiple and contrasting guidelines for managing child health acted as a bar- rier to effective implementation of IMCI (Naimoli and others 2006). Similarly, in Peru the presence of a number of child health programs with overlapping remits led to inefficien- cies in delivery and destructive competition for resources for IMCI (Huicho and others 2005a; Huicho and others 2005b). In contrast, successful introduction of IMCI in Tanzania coincided with other measures to improve the management of the district health systems, creating an enabling environment for the implementation of IMCI strategy (Armstrong Schellenberg and others 2004; Masanja and others 2005). Desirability A desire to implement and sustain an intervention could emanate not just from the lead- ership but also from the health workers. Government commitment was critical to the suc- cess of the interventions aimed at schistosomiasis control in China (Sleigh and others 1998a, 1998b, 1998c) and successful integration of leprosy services with PHC in Sri Lanka (Kasturiaratchi, Settinayake, and Grewal 2002). In contrast, a lack of government com- mitment contributed directly to the poor IMCI results in Peru (Huicho and others 2005a, 2005b). In India, strong opposition from the staff involved in delivering the National Lep- rosy Program hindered for a considerable time integration of the program with PHC (Rao and others 2002). Similarly, in Cuba, reluctance of health workers to relinquish responsi- bilities delayed integration of interventions for malaria control into mainstream health ser- vices (Rojas, Botero, and Garcia 2001). CHAPTER 4 Discussion T he debate on health systems and targeted interventions has tended to narrowly focus on vertical or integrated descriptors. However, our analysis shows this to be a false dichotomy as few interventions are purely vertical (single-disease oriented) or horizontal (fully integrated into mainstream health system functions). As this review shows, in practice the nature of the problem, the interventions to address these and the adoption and assimilation of health interventions in health systems vary greatly in differ- ent contexts, as does success. The purpose, nature, speed, and the extent of integration also vary--in part, dependent on the intervention complexity, the health system charac- teristics and the contextual factors. There are few instances where there is full integration of a health intervention or where an intervention is completely non-integrated. Instead, there exists a highly heterogeneous picture both for the nature and also for the extent of integration. The review suggests that the evidence base for integration versus specific health pro- gramming is very limited. As success was measured in different ways (for example, in terms of varied programmatic goals related to efficiency, effectiveness, or equity), analysis and drawing lessons from this is further complicated. Perhaps the modesty of evidence creates the context for strong opinions for or against integration in global health. We have attempted to show that in practice the dichotomy between integrated and non-integrated (traditionally described as vertical and horizontal) is not rigid. Health systems combine both non-integrated and integrated interventions, but the balance of these interventions varies considerably. Our findings suggest that the purpose, nature and extent of integra- tion vary enormously between interventions and in countries, creating a rich mosaic of local solutions to address emergent problems. We found rare instances of full integration of health intervention into all critical health system functions, but a wide range of instances 23 24 World Bank Working Paper where health interventions are integrated into one or more critical health system functions. Further, it is important to note that in many countries that simultaneously have multiple health interventions, the nature of integration for these interventions varies and different degrees of integration for the same critical health system function co-exist. This study has a number of limitations. Firstly, program evaluations reported in the peer-reviewed literature used in this study provide limited and varied detail on the orga- nizational structures surrounding the intervention and the health systems within which these interventions were implemented. However, the framework presented (Atun and oth- ers 2009) has enabled us to consistently analyze the data on health system elements and contextual factors. Secondly, our assessment of the `success' of interventions was based on that reported by the authors against a range of measures specified in the studies. This assessment could potentially overlook other important but unreported measures of suc- cess or failure. Thirdly, in our review, we selected predominantly "program evaluations"-- not a well-defined study design. Unlike the Cochrane criteria for appraising quality of randomized controlled studies there is no consensus on the criteria that could be used to appraise the quality of program evaluations. Hence, we selected these studies on the basis of their relevance to our research question but their quality was not rated. Finally, the inherent heterogeneity of the included studies, both in their clinical focus and in the set- ting, makes it difficult to generalize findings without a better understanding of how con- textual factors have shaped the organizational structures of health interventions. We intend to address this issue through a more comprehensive and in-depth approach, using a multi- country case study design. CHAPTER 5 Conclusions I n spite of the study limitations that we tried to address, the findings provide new syn- thesis of evidence to further the debate on health systems and targeted interventions which has ossified in a binary mode. Given the highly varied contexts and adoption systems that reflect local nuances, different health system capacities and the range of prob- lems being addressed, it is not surprising that in practice a rich mix of solutions exist. While the discussion on the relative merits of integrating health interventions will no doubt continue, discussion should move away from the highly reductionist approach that has polarized this debate. Future efforts are best spent on generating and learning from useful evidence. What is needed is a series of robust country case studies of health interventions that use replication logic with common methodology and shared methods, informed by appropriate theoret- ical frameworks (such as the one used in this study). The case studies would allow com- parisons among countries and programs, generating evidence that has relevance beyond a country case. 25 Appendixes 27 APPENDIX A Search Strategy and Methods 1. ((vertical[Tiab] OR horizontal[Tiab] OR integrat*[Tiab] OR coordinat*[Tiab] OR co-ordinat*[Tiab] OR link*[Tiab]) AND (program*[Tiab] OR care[Tiab] OR service*[Tiab]) OR "Delivery of Health Care, Integrated"[MeSH]) AND 2. (("child health services"[MeSH] NOT "early intervention (education)"[MeSH]) OR "immunization programs" [MeSH] OR "family planning services"[MeSH] OR ("maternal health services"[MeSH] NOT "preconception care"[MeSH]) OR "maternal-child health centers"[MeSH] OR "community health centers"[MeSH] OR immunization[MeSH] OR "reproductive medicine"[MeSH] OR "adolescent health services"[MeSH] OR cholera[MeSH] OR dengue[MeSH] OR "Dengue Hemorrhagic Fever"[MeSH] OR Fascioliasis[MeSH] OR Trypanosomiasis[MeSH] OR Leishmaniasis[MeSH] OR Ele- phantiasis[MeSH] OR yaws[Mesh] OR "buruli ulcer"[MeSH] OR dracunculiasis[MeSH] OR lep- rosy[MeSH] OR schistosomiasis[MeSH] OR helminths[MeSH] OR "chagas disease"[MeSH] OR onchocerciasis[MeSH] OR "alcohol-related disorders"[MeSH] OR "cocaine-related disorders"[MeSH] OR "opioid-related disorders"[MeSH] OR "substance abuse, intravenous"[MeSH] OR "anxiety dis- orders"[MeSH] OR "depressive disorder"[MeSH] OR schizophrenia[MeSH] OR "eating disorders"[MeSH] OR dementia[MeSH] OR HIV/AIDS[MeSH] OR tuberculosis[MeSH] OR malaria[MeSH] OR "sexually transmitted diseases"[MeSH] OR "mental health"[MeSH] OR "Anemia, Iron-Deficiency"[MeSH] OR "Vitamin A Deficiency"[MeSH] OR "Food, Fortified"[MeSH] OR micronutrients[MeSH] OR zinc[MeSH]) Search results were limited using the PubMed filters "English" and "humans". We searched the following electronic databases: PubMed; the Cochrane Central Register of Con- trolled Trials (CENTRAL); and the Cochrane EPOC specialized register and Database of Abstracts of Reviews of Effectiveness. Supplementary searches were conducted through reference and citation tracking of the key articles retrieved during the search. Review Methods The search strategy retrieved 8,274 potentially relevant articles from PubMed. This was followed by selection of articles deemed relevant for further analysis by two independent reviewers on the (continued) 29 30 World Bank Working Paper basis of the titles retrieved. In order to establish a common set of inclusion criteria the first 100 titles were assessed jointly. The criteria for inclusion at the first stage were based on the potential relevance of each article to the research question for the review, independent of study design. Once a common approach and understanding was developed each reviewer independently assessed one half of the remaining 8,274 titles (i.e. 4,137 each). In addition, to verify sufficient inter-rater agreement each reviewer evaluated 20 percent of titles assessed and rated by the other reviewer. Each study deemed relevant by either of the reviewers was further considered for inclusion in the second stage of analysis. This yielded a total of 1,551 titles, which were retained for the second stage of selection. In this stage, the two reviewers independently assessed the abstracts of each of the 1,551 remaining studies for relevance to the review. This exercise showed the inter-rater agree- ment (Cohen's coefficient) to be equal to 0.78. The studies selected by each of the reviewers were then compared. Where there was concordance to include or exclude a study they were retained or discarded. Where there was disagreement, the article in question was retained for full text analysis. A total of 200 studies were retained for the third stage of the review at which the full-text was ana- lyzed. The full-text of 172 of these identified papers was retrieved, whereas the remaining 28 stud- ies could not be accessed. Studies were then evaluated based on study design as well as relevance. The considered studies were primarily program evaluations, as well as studies that assessed the relative performance of different models of care after a change in organizational structure. For inclusion, the study had to present data on outcome measures such as health outcomes, quality of care, access to care and service utilization, patient satisfaction and cost or cost-effectiveness. Papers that merely described the process of development of a care model, without discussion of any of the relevant process or outcome measures, were therefore excluded. Programs, further- more, needed to have been implemented on a regional or national scale (i.e. beyond a trial set- ting) to be eligible for inclusion. Data Extraction and Management Data extraction was done independently by the two reviewers, each using a common checklist pur- posefully developed for this review and based on the analytical framework outlined in detail below. Extracted data included: (i) General information: title, authors, and year of publication; (ii) Setting: clinical area, country, setting/care delivery system, and target groups; (iii) Intervention: description, duration, comparisons, and co-interventions; (iv) Study characteristics: study design, duration; (v) Participants: unit of analysis, number of participants in intervention and comparison groups; (vi) Outcomes: outcome measures, results of the intervention; (vii) Extent and nature of integration: stewardship & governance, financing, planning, service delivery, monitoring & evaluation, demand generation; (viii) Contextual factors: sustainability, opportunities, desirability APPENDIX B Summary of the Included Studies 31 Time Description Description Intervention Control Outcome Name Setting Frame of Intervention of Control Group Group Measure Results Dengue ToledoRomani Cuba 2001­2004 CWG integrated Strengthened General; 3 General; 3 Maintenance of Correct water 2007 into vertical vertical health zones health zones effects at 2 years storage cover in program program within a within the after the end of 87.5% of the district same district the intervention; surveyed houses in institutionaliza- intervention area tion; capacity and 21.5% in building in the control area; main community breeding sites: communal areas in the intervention area (75.8%) and domestic environ- ment in the con- 32 trol area (86.4%) Malaria Rojas 2001 Colombia not clear Education, n/a General; 23 n/a Incidence; length Incidence of diagnosis and communities of sick leave malaria decreased treatment within in 2 areas by 45.4%; length communities, of sick leave vector control decreased from 7.5 to 3.7 days Schistosomiasis Ageel 1997 Saudi Arabia 1990­1996 PHC provided n/a Population of n/a Coverage; Better coverage of education, high-risk prevalence of the population detection and villages, who infection (60%­90%) than the treatment. attended the vertical program Schoolchildren health centers (6%­50%); overall helped in prevalence of diagnosis within infection the community maintained at <1% for >5 years Bausch 1995 Cameroon n/a Local health No schistoso- All Takelale 3 non- Prevalence; Major reduction in centers and miasis control quarters Takelale disease-specific prevalence (7% vs. elementary activities quarters knowledge 71%) and intensity teachers accidentally (1% vs. 26%) of performed health excluded from schistosomiasis education, the control infection; screening and activities improved knowl- drug therapy edge about the disease (sources, risk factors) Cline 1996 Cameroon 1991­1993 Local health Same as the 4 selected 1 village in Prevalence of Reduction in centers and intervention, villages another infection; prevalence (from teachers but without within Kaele subdivision disease-specific 21% to 7%) and in performed health health subdivision knowledge; the number of education, education health service heavy infections treatment, snail utilization (from 23 to 8); control increased student knowledge on the 33 role of snails and water contact in infection (respectively, from 16% to 76% and from 11% to 72%); increased utilization of health centers (20%) Coura Filho Brazil 1974­1987 CPqRR and, later, n/a Peri Peri n/a Prevalence; Prevalence 1992 local staff district incidence decreased from performed population 43.5% to 15.2% demographic under CPqRR and census, further to 4.4% examination and under municipal treatment management; incidence decreased from 19% to 10.9% and further to 2.9% (continued) Time Description Description Intervention Control Outcome Name Setting Frame of Intervention of Control Group Group Measure Results Engels 1993 Burundi 1985­1990 Selective n/a General; n/a Number of 10% of all cases population based population detected cases received treatment chemotherapy by every year; annual mobile control yield of cases teams, later detected by basic integrated into health services was PHC 60% at an eighth of the cost per inhabitant Engels 1995 Burundi 1993­1994 Case detection n/a General; n/a Number of 15 months after the based on primary population detected cases civil unrest, the care service output has recovered in most endemic provinces and the program output has contin- 34 ued to increase after an initial depression in the least affected provinces Jarallah 1993 Saudi Arabia 1984­1989 Integration of the n/a Riyadh n/a Population Six-fold increment existing population coverage; in population schistosomiasis and areas prevalence; coverage (up to vertical program treatment and more than 95% in into primary care follow-up; snail some health control centers); significant reduction in preva- lence (from 13.2% to 0.2%); reduction in dropout rate (from 54.4% to 22%); significant reduction in the number of surveyed water sources that resulted positive for snails (from 1.8% to 0.1%) Kabatereine Uganda n/a Health education, n/a Communities n/a Coverage; costs Coverage was 2006 drug distribution, at high risk in 91.4% in schools and deworming schistosomiasis- and 64.7% in by school teachers endemic communities; and community areas estimated drug distributors financial costs per person equal to US$ 0.34 Sleigh 1998a China 1953­1992 Schistosomiasis n/a Guangxi pop- n/a Snail distribution Eradication of Control Program: ulation and and control; schistosomiasis in case detection, areas (>11 water supplies the 19 endemic treatment, snail million, of and sanitation; counties­reduction eradication and which 1.1 detection, control in the number of health education million at and treatment; infected snail foci risk) benefits of from 4,716 in 1958 eradication to 0 after 1992; people and animals found positive at 35 any test were treated (96,537 people received drug therapy) Sleigh 1998b China 1953­1992 Schistosomiasis n/a Guangxi n/a Program costs by Annual variable Control Program: population category and costs of 60%; tetanus, measles, children in coverage; infant decline in infant pertussis, diph- two rural and child and child mortality theria, polio, districts morbidity and rate from 162/1,000 whooping cough mortality to 77/1,000; the and tuberculosis) Immunization most frequently mentioned disease symptoms at the time of death of an infant were­in order­fever, cough, troubled breathing and diarrhea 42 Bonu 2003 India 1993­1999 Pulse Polio n/a Sample of n/a Completion rates Increase in coverage Immunization approximately for polio immu- of polio vaccine (I (PPI) campaign 23,000 nization; coverage dose: from 48% to women and of non-polio EPI 73%, III doses: from their surviving vaccines; equity 34% to 45%, indicat- children aged in coverage of ing a dropout rate of 12­35 months polio and non- 28%); no significant (around 6,400) polio EPI change in immu- vaccination nization coverage of all 5 doses of non- polio EPI (from 17.7% to 18.5%) Koop 2001 Macedonia April­May Two mass EPI n/a All <4 years n/a Immunization Immunization cov- 1999 campaign in children living coverage erage rates in 4/5 refugee camps in 7 refugee camps were >89% and weekly camps who during the first mobile immuniza- didn't have campaign; cover- tion clinics for new an up-to-date age rates of weekly arrivals and chil- vaccination clinics was >90% dren who missed card the first campaign Child Health & Development­IMCI Adam 2005 Tanzania 1999 IMCI Usual care 2 rural IMCI 2 neighboring Total economic Lower costs associ- districts non-IMCI costs of start up ated with IMCI districts and implementa- (US$ 11.19) tion of IMCI; compared with incremental costs routine care of introducing (US$ 16.09)­44% and running IMCI difference, of which 6% could be attributed to IMCI; no additional staff required; higher cost-effectiveness of IMCI compared with standard care Amaral 2004 Brazil 2002 IMCI standard 20 selected Neighboring Quality of care Quality of care low-income municipali- provided to under IMCI signifi- 43 families in ties in which children <5 cantly improved municipali- there were no compared to stan- ties with a IMCI-trained dard care (e.g. population workers weight checked 5,000­50,000 against growth chart 77.3% in IMCI vs. 36.3% in control facilities; checking for at least 3 danger signs 38.3% in IMCI vs. 0.6% in non- IMCI; evaluation of feeding practices 35.4% in IMCI vs. 10.1% in non-IMCI) (continued) Time Description Description Intervention Control Outcome Name Setting Frame of Intervention of Control Group Group Measure Results Armstrong Tanzania 1999­2002 IMCI adapted to Usual care 2 districts in 2 districts Children's survival Improved case man- Schellenberg reflect national which IMCI where IMCI and health; agement (95% of chil- 2004 child-health had been had not yet economic cost dren were checked policies and local implemented been imple- for cough, diarrhea terminology for 2 years mented and fever in IMCI vs. 36% in control facili- ties); 13% lower mor- tality rate in children <5yr in IMCI districts; similar costs of care under IMCI and conventional case- management (US$ 11.19 vs. US$ 16.09­US$ 8.30 vs. US$ 8.76, excluding 44 hospital-level costs); high cost-effectiveness of IMCI Bryce 2005a Tanzania 1999 IMCI Usual care 2 districts in 2 districts Quality of care; IMCI was associated Tanzania where IMCI economic costs with significant where IMCI had not yet improvements in was imple- been imple- quality of care (e.g. mented in mented proportion of chil- 1997 (i.e. 2 dren who were man- years before aged correctly was the study was 65% under IMCI and conducted) 16% under usual care); lower costs (total cost per child in 1999 was US$7.86 in IMCI vs. US$8.34 in non-IMCI); improved efficiency (cost per child visit managed correctly was US$4.02 in IMCI vs. US$25.70 in non-IMCI) Bryce 2005b General n/a IMCI n/a n/a n/a Implementation Overextension of the of guidelines, guidelines to settings quality of care other than the origi- at first-level nal ones (developing facilities, health countries with infant system supports, mortality >40/1000 utilization of live births); improving health services, the quality of care in child mortality, first-level facilities coordination alone was not suffi- cient to increase low utilization levels; wide variability in IMCI implementation among the MCE-IMCI sites Chopra 2005 South Africa 2001­2002 IMCI adapted to n/a General; n/a Index of inte- Large improvement in local epidemio- randomly grated assess- the overall assess- 45 logical profile selected ment; primary ment of children (e.g. (­measles and clinics within and secondary assessment of danger malaria; + IMCI district indicators as signs in children: 7% HIV/AIDS and identified by before vs. 72% after asthma) WHO analysis IMCI); increased effi- plan ciency of care (e.g. rational prescribing: 62% before vs. 84% after IMCI); no change in level of counseling and absence of improvement in care- giver understanding of key practices (e.g. advice to return immediately if the child was not able to drink/breast feed: 32.5% before vs. 35.5% after) (continued) Time Description Description Intervention Control Outcome Name Setting Frame of Intervention of Control Group Group Measure Results El Arifeen Bangladesh 2001­2004 IMCI with Usual care 20 randomly 20 paired Quality of care Improvements in 2004 adapted case selected facilities in (adherence to quality of health care management facilities in area without IMCI guidelines); (mean index of cor- guidelines IMCI district IMCI care seeking rect treatment for behavior; sick children was 54 utilization of in IMCI and 9 in con- governmental trol facilities); health facilities increase in use of first-level facilities (0.6 visits per child at baseline and 1.9 after 21 months of IMCI); 19% of sick children in IMCI and 9% in control facilities were taken 46 to a health worker Huicho 2005a Peru 1996­2001 IMCI training for n/a All 34 districts n/a IMCI health Increase in the over- doctors and worker training; all number of health nurses, adapted supervision workers trained in to epidemio- activities; IMCI until 1999­ logical/cultural health facility decrease in 2000 and context (from 11 support; utiliza- 2001; overall training to 7 days and tion of child coverage of 10.3% for reduced health doctors and nurses; supervision) services average of 4.7 annual attendances per child in 2000; improved performance (e.g. children checked for 3 general danger signs: 35­40% vs. 5­4%); expected syn- ergies between health facility and commu- nity interventions were not realized Huicho 2005b Peru n/a IMCI n/a All 34 districts n/a Health services Average of 4.8 visits utilization; per child/year; 2/3 of coverage; all children had all impact basic vaccines (mortality and recorded on their nutritional cards; IMCI training status) coverage was 10.2% in 2000; in 2000, the average cumulative reduction of under-5 mortality rate was 40.3% of the 1996 level; non-significant correlations between IMCI training coverage and outpatient utiliza- tion, vaccine coverage, mortality or malnutri- tion indicators 47 Masanja 2005 Tanzania 1999­2002 IMCI Usual care 2 district with 2 districts Equity: morbid- Improvements in IMCI without IMCI ity, nutritional child health, not at status, immu- the expense of equity: nization cover- equity differentials for age, caretaker 6/10 child health indi- knowledge and cators (underweight, behavior, home stunting, measles management of immunization, access illness (house- to treated and hold wealth untreated nets, treat- index) ment of fever with antimalarials) signifi- cantly improved in IMCI (e.g. concentra- tion index for stunting improved from -0.102 to -0.032 in IMCI vs. ­0.122 to ­0.133 in non-IMCI) (continued) Time Description Description Intervention Control Outcome Name Setting Frame of Intervention of Control Group Group Measure Results Naimoli 2006 Morocco 2000 Outpatient Usual care Ill children Ill children Index of overall Improved quality of facilities in IMCI attending a attending a guideline care: adherence to where training of facility where facility where adherence; guidelines was 79.7% HC workers lasted workers workers appropriateness in IMCI vs. 19.5% in 12 days (instead received IMCI didn't receive of antibiotic non-IMCI facilities, of 11) and training IMCI training prescription and correct prescrip- guidelines were tion of antibiotics adapted was 60.8% in IMCI vs. 31.3% in non-IMCI facilities Pariyo 2005 Uganda July-- IMCI case man- No training in Children Health work- Factors mediat- Increase in IMCI December agement training IMCI provided <5 yr; Health ers managing ing the effect of training coverage 2000, 2001 to qualified and workers children who IMCI on the from 7.7% to 74.1%; and 2002 (later) auxiliary managing did not quality of care health workers 48 staff children who receive the to children: checked the vaccina- received training performance of tion status of nearly training in health workers; 50% of sick children; IMCI training cover- the index of availabil- age; health ity of the 4 essential facility support vaccines was >70%; indicators <20% of children needing immuniza- tions were vaccinated Zhang 2007 China 2004­2006 IMCI training in n/a 146 sick chil- Pre- Presence of Significant improve- township hospi- dren, 136 intervention: basic equipment ments in: availability tals, village mothers and 550 sick and supplies; of basic equipment health centers 74 health children, 542 inappropriate and drug supplies (no adaptations facilities from mothers and prescribing of (e.g. children's scales: specified) townships of 3345 health drugs; presence from 27.8% to 90.5%), the pre- facilities in of basic drug rational use of drugs intervention various supplies; satis- (inappropriate use of set townships faction of care antibiotics: from givers 59.3% to 6.2%), satis- faction of care givers (from 83.4% to 95.6%) Child Health & Development­ICDS Agarwal 2000 India 1987­1993 ICDS services plus ICDS without Pregnant Pregnant Anthropometric Improvements in nutritional food supple- malnourished malnourished measurements: the supplemented supplementation mentation women in 28 women in 21 maternal height, ICDS group com- provided to provided to villages in the villages in the weight and mid- pared with the malnourished malnourished ICDS block adjacent arm circumfer- unsupplemented pregnant women pregnant Harahua block ence, birth ICDS group: per- women weight; hemoglo- centage of low birth bin level weight deliveries was 14.4% vs. 20.4%; rate of newborns weighted >3 kg was 16.2% vs. 11%; 100 g extra gain in maternal weight; 0.3 weeks increase in gestational age Dubowitz India n/a Dular' Program: Usual ICDS Women and Women and Use of iodized Statistically signifi- 49 2007 ICDS plus training children (0­36 children salt; children's cant differences in of local resource months); 200 (0­36 immunization all measured out- groups and from Dular- months); 315 status; antenatal comes between Anganwadi intensive vil- from non- care in pregnancy; Dular and non-Dular workers lages and 229 Dular villages feeding of villages (e.g. iodized from Dular- (receiving colostrums; salt use: 95.8% vs. regular ones regular ICDS feeding of breast 53.2%; prenatal only) milk; use of care: 91.1% vs. delivery kit for 70.7%; colostrum childbirth fed: 82% vs. 20.7%; use of delivery kit for childbirth 48.2% vs. 6.5%) and, although less pronounced, in a number of measures between Dular-intensive and Dular-regular vil- lages (e.g. colostrum fed: 95% vs. 82%) (continued) Time Description Description Intervention Control Outcome Name Setting Frame of Intervention of Control Group Group Measure Results Gupta 1984 India 1981­1982 Standard package Non ICDS care Children <6 Children <6 Impact on Breast feeding prac- of ICDS services years in ICDS years in non- feeding practices, tices were almost provided by block of Rae ICDS areas growth and equal (99.8% ICDS vs. Anganwadi Bareli, Uttar development; 99.4% non-ICDS), but centers Pradesh prevalence of ICDS performed bet- malnutrition; ter than non-ICDS on utilization of the other indicators: health care higher weight-for- services age (e.g. Kg of 2 yr male children: 9.8 vs. 8.1); timely mile- stones of develop- ment (e.g. teething: 7.2 months vs. 8.5); 50 lower prevalence of severe grade of pro- tein energy malnutri- tion (3.2% in ICDS vs. 14.8% in non-ICDS); higher (but not satis- factory) immuniza- tion rates (e.g. oral polio vaccine cover- age: 54.3% vs. 0.3%); use of antenatal care (61.3% vs. 19.9%) Kapil 1999 India 1996­1997 Anganwadi com- n/a 61 severely n/a Grade of malnu- Significant reduction munity centers malnourished trition in grades of provided thera- children visit- undernutrition from peutic supple- ing the cen- 7.7% to 4.5% after mentary feeds tre, followed 12 months of and advise on for 1 year of intervention breastfeeding ICDS and nutrition Lal 1980 India 1976­1979 Anganwadi com- n/a Children <6 n/a Nutritional and Reduction in preva- munity centers years and health status; lence of severe provided nutri- pregnant utilization of degrees of malnutri- tional and immu- women in 10 therapeutic and tion in <6 yr chil- nization services anganwadi prophylactic dren from 17.6% to centers in nutrients; 8.4%; increase in uti- ICDS block of utilization of lization of therapeu- Kathura immunopro- tic and prophylactic phylaxis nutrients (iron and folic acid: children <6yr­from 2.3% to 58%; pregnant women­from 9.2% to 35%); increase in immunization cover- age (e.g. all 3 doses of DPT: from 6.7% to 69.9%) 51 Saiyed 2000 India n/a ICDS services pro- n/a Children <3 n/a Nutritional sta- Children who uti- vided at Angan- years in 9 tus measured as lized all four services wadi centers: anganwadi height-for-age, (immunization, immunization, centers weight-for-age vitamin A, monthly vitamin A, iron located in and weight-for- weighing, and food and folic acid slum areas of height; episodes supplementation) supplementation, urban Baroda, and duration of fully rather than growth monitor- Gujarat illness partially showed ing, supplemen- improvements in tary food their nutritional sta- provision and tus and in the num- referral for seri- ber of episodes and ous illness duration of illness (no illnesses: 72% vs. 39%; mean days of illness: 1.8 vs. 3.9) (continued) Time Description Description Intervention Control Outcome Name Setting Frame of Intervention of Control Group Group Measure Results Tandon 1988 India 1976­1984 Anganwadi com- EPI without Children <6 yr Children <6 yr Immunization Significant improve- munity centers ICDS and pregnant/ and pregnant/ coverage ments in immuniza- coordinated EPI lactating lactating tion coverage for and ICDS services women in women in children by about to improve areas with areas without 100% in 3­5 yr and 52 immunization ICDS of 3­5 ICDS support 125% in 8 yr ICDS coverage or 8 years programs (e.g.: BCG: duration 22.6% in EPI, 39.6% in 3­5 yr ICDS, 50.5% in 8 yr ICDS) and pregnant women (TT: 24.5% in EPI, 44.4% in 3­5 yr ICDS, 32.1% in 8 yr ICDS) Tandon 1992 India n/a Anganwadi com- EPI without 5367 children 2018 children Immunization Increased immu- munity centers established (1­2 yr) and (1­2 yr) and coverage nization coverage in coordinated EPI ICDS 5111 mothers 1890 mothers ICDS projects com- and ICDS services in 36 ICDS in 18 newly pared with non-ICDS to improve projects in sanctioned ones for children immunization place for at ICDS projects (BCG: 65% vs. 22%, coverage least 5 years DPT: 63% vs. 28%, polio 64% vs. 27%) and pregnant women (TT: 68% vs. 40%) Trivedi 1995 India 1990­1993 Standard package Care-as-usual; 709 children 500 children Nutritional sta- No significant differ- of ICDS services no ICDS between 1­6 between 1­6 tus (grade of ence in nutritional provided by services years in 6 years in 5 malnutrition); status between ICDS Anganwadi anganwadi matched vil- immunization and non-ICDS areas centers centers lages in non- status (BCG, (e.g. normal and within ICDS ICDS area DPT, polio, grade I malnutrition: blocks in measles) 74.3% and 72.4%); rural area of lower immunization Indore, coverage in ICDS com- Madhya pared with non-ICDS Pradesh areas (BCG: 80.2% vs. 88.8%; DPT: 79.5% vs. 94.4%; polio: 88% vs. 95.3%; measles: 45.7% vs. 62%) Family Planning and Maternal and Child Health DeGraff 1986 Bangladesh 1982­1984 Extended analysis n/a General; 2 n/a Contraceptive The addition of 53 of Phillips1984 blocks of 20 prevalence rate 18 months to the time villages each as function of series reported by MCH services Phillips (1984) did not change substantive results: intensification of MCH services in 1982 did not have a positive incremental impact on the con- traceptive prevalence rate (measles/tetanus vaccine and antenatal care/TBA training vari- ables did not have sig- nificant coefficients for explaining contracep- tive prevalence rate) (continued) Time Description Description Intervention Control Outcome Name Setting Frame of Intervention of Control Group Group Measure Results Douthwaite Pakistan 1993­2002 Lady Health Usual care Women from Women from Current use of Significantly higher 2005 Worker Program randomly sampled reversible current use of (LHWP): selected communities modern family reversible modern promotion and households not served by planning contraceptive meth- provision of and LHWs in the LHWP methods ods under LHWP door-to-door FP LHWP area (women using and primary reversible modern health care methods: 13% in services LHWP, 7% in control areas­national aver- age of 10%), even after controlling for socio-economic factors Phillips 1984 Bangladesh 1982­1983 Stepwise intro- Usual care General; 2 General; 2 Contraceptive No significant differ- 54 duction of limited blocks of 20 blocks of 20 prevalence rate ence between an or intensive MCH villages each villages each as function of incremental addition services into exist- MCH services of MCH services and a ing FP by family basic package of com- welfare centers prehensive FP with and community minimal MCH services health workers on contraceptive prevalence rate Simmons Bangladesh 1978­1985 Matlab FP and Government A large, fairly A large, fairly Costs of the The Matlab Project is 1991 Health Services of Bangladesh homogeneous homogeneous service delivery more expensive than Project (FPHSP): FP Program: region divided region divided system the government's FP community- compartmen- into two parts into two parts program, but it is also based services talized service (intervention (intervention more effective (e.g. (focus only on the orientation and control) and control) cost-effectiveness limited package ratio under Model III: of services) average cost per birth prevented ranged from $171 to $220 in treatment areas and from $240 to $298 in control areas) Tuladhar Nepal 1975­1978 Village health FP/MCH and Pregnant/ Pregnant/ Knowledge of Compared to the 1982 workers provid- fewer com- lactating lactating FP; use of and vertical program, the ing FP/MCH and prehensive women in 2 women in 2 intention to use integrated model numerous other and preven- districts districts FP services; reported: smaller services more fre- tive services where ser- matched to child survivor- impact on knowledge quently (e.g. nutri- delivered less vices had the interven- ship; infant of FP (increase in the tion education, frequently been inte- tion districts mortality number of women immunization) grated in who had knowledge 1976 of FP: 90% in V vs. 22% in I); higher impact on intentions to use FP services (decrease of 13% vs. 35%); rela- tively greater impact on child survivorship (1% vs. 4%); signifi- cantly smaller impact on infant mortality (decrease 55 of 47% vs. 25%) HIV/AIDS/AIDS Peck 2003 Haiti 1985­2000 GHESKIO clinic n/a HIV/AIDS- n/a Demographics Increase in the providing volun- affected fami- and n. individu- number of new peo- tary counseling, lies, health als for ple seeking volun- testing and care care workers counseling/ tary counseling and to HIV/AIDS chil- after acciden- testing; n. testing (from 142 to dren and adults, tal HIV/AIDS clients receiving 8175); care was pro- and services for exposure, each integrated vided to 17% of the other communi- female rape service; effect new patients who cable diseases victims of HIV/AIDS asked for AIDS treat- prevention; ment, 6% of TB, 18% HIV/AIDS of STI, 19% of FP; transmission HIV/AIDS transmis- sion between dis- cordant couples was 0 infections/100 follow-up years; vertical transmission was 11 infections/ 100 live births APPENDIX C Contextual Factors 57 Opportunities Desirability Sustainability Critical Events, Synergy, Political Economy, Socio-Cultural Priority Intervention Fiscal Space, Frailty Technology/Innovation Factors Dengue Cuba (Toledo Romani and I Small dengue outbreaks in non- I Dengue control low priority for others 2007) endemic area communities I Resistance to change amongst vector control personnel I Provision of adequate services at municipal level required 58 Malaria Colombia (Rojas, Botero, and Garcia I Program decentralized 2001) Schistosomiasis Brazil (Coura Filho and others 1992) I Municipal capacity Burundi (Engels, Ndoricimpa, and I Reduced disease burden I Civil unrest following a military Gryseels 1993; Engels, Sindayigaya, coup attempt, disturbed Primary and Gryseels 1995) Health Care services Cameroon (Bausch and Cline 1995; I Development projects that might I Strong government commitment Cline and Hewlett 1996) expand snail habitats and transmission to national PHC system I Urinary schistosomiasis as an entry point into the PHC system I Easy to recognize, diagnosis and treat China (Sleigh and others 1998a, I High priority for all levels of 1998b, 1998c) government Saudi Arabia (Ageel and Amin 1997; I Reduced disease burden I General rise in living standards Jarallah and others 1993) I Advanced system of PHC and socio-economic conditions, with >90% population improved sanitation and water coverage supplies, and medical care Uganda (Kabatereine and others I In remote communities up to 30% 2006) of school-aged children not enrolled in school: not reached by school based approaches Leprosy Sri Lanka (Kasturiaratchi, I Reduced disease burden I High degree of political support Settinayake, and Grewal 2002) from all levels of health administration India (Rao and others 2002; Thakar I Reduced disease burden I Resistance to change amongst and Kumar 2003) NLEP personnel; prior attempt at integration failed 59 Nutrition Nutrition­Peru (Gross, Mamani I Evidence of high prevalence of iron- I Rapid urbanization Diaz, and Valle 2006a; Gross, Valle, deficiency anemia and Mamani Diaz 2006b; Lechtig and others 2006a, 2006b; Lopez de Romana and others 2006) Immunization India (Bonu, Rani, and Baker 2003) I High transmission of wild poliovirus I Ruralareaswithpersistent inequities in immunization coverage by caste, gender, wealth and religion I Priority for the international agenda (continued) Opportunities Desirability Sustainability Critical Events, Synergy, Political Economy, Socio-Cultural Priority Intervention Fiscal Space, Frailty Technology/Innovation Factors Macedonia (Koop, Jackson, and I Emergency situation with high risk of I Military conflict Nestel 2001) vaccine-preventable diseases' Sierra Leone (Amin 1996) I Endemicity of common childhood I Military conflict 60 diseases and inadequate rural health I Widespread poverty services I Priority for the international agenda Child Health & Development­IMCI Morocco (Naimoli and others 2006) I Possibleconflictbetweentheintegrated case management approach and other clinical rule systems already in use. Peru (Huicho and others 2005a, I Good network of health I Overlapping child health programs and I Period of rapid political change 2005b) facilities with high uti- competition for resources. I Poor governance and lack of polit- lization rates ical commitment South Africa (Chopra and I Good facility infrastruc- others 2005) ture and management support Tanzania (Armstrong Schellenberg I High utilization rates of I IMCI introduced concurrently with and others 2004; Masanja and public health facilities measures designed to improve district others 2005) health systems management Child Health & Development­ICDS India­ICDS (Agarwal and others I ICDS introduced in pursuance of 2000; Gupta and others 1984; Kapil National Policy for Children and others 1999; Lal 1980; Saiyed and Seshadri 2000) India­ICDS + EPI (Tandon and oth- I ICDS and EPI target same group of ers 1992; Tandon and Sahai 1988) children India­ICDS + "Dular" (Dubowitz and I Builds on existing ICDS I Strong leadership others 2007) infrastructure Family Planning Bangladesh­FPHSP (DeGraff and I Many years of experience I High population growth rates hindering others 1986; Phillips and others I Dedicated human and economic growth and development 1984; Simmons, Balk, and Faiz physical resources and I Presence of MCH contributes to 1991) organizational ability legitimacy of FP, activating latent demand. Reduction in child morbidity 61 and mortality leads to change in reproductive motives Pakistan­LHWP (Douthwaite and I Overburdened health Ward 2005) system I Significant funding shortfalls HIV/AIDS/AIDS Haiti­"GHESKIO" (Peck and others I HIV/AIDS VCT attracts a population with 2003) high rates of co-morbidity. Reciprocally, on-site services attract more people to VCT References Adam, T., F. Manzi, J.R. Armstrong Schellenberg, L. Mgalula, D. de Savigny, and D.B. Evans. 2005. "Does the Integrated Management of Childhood Illness Cost More than Routine Care? Results from the United Republic of Tanzania." Bulletin of the World Health Organization 83(5):369­77. Agarwal, K.N., D.K. Agarwal, A. Agarwal, S. Rai, R. Prasad, S. Agarwal and T.B. Singh. 2000. "Impact of the integrated child development services (ICDS) on maternal nutri- tion and birth weight in rural Varanasi." Indian Pediatrics 37(12):1321­27. Ageel, A.R.M., and M.A. Amin. 1997. "Integration of schistosomiasis-control activities into the primary-health-care system in the Gizan region, Saudi Arabia." Annals of Tropical Medicine and Parasitology 91(8):907­15. Amaral, J., E. Gouws, J. Bryce, A.J. Madeiro Leite, A.L.A. da Cunha and C.G. Victora. 2004. "Effect of Integrated Management of Childhood Illness (IMCI) on health worker per- formance in Northeast-Brazil." Cadernos de Saude Publica 20 Suppl 2:S209­19. Amin, R. 1996. "Immunization coverage and child mortality in two rural districts of Sierra Leone." Social Science and Medicine 42(11):1599­1604. Armstrong Schellenberg, J.R., T. Adam, H. Mshinda, H. Masanja, G. Kabadi, O. Mukasa, T. John, S. Charles, R. Nathan, K. Wilczynska, L. Mgalula, C. Mbuya, R. Mswia, F. Manzi, D. de Savigny, D. Schellenberg, and C. Victora. 2004. "Effectiveness and cost of facility-based Integrated Management of Childhood Illness (IMCI) in Tanzania." Lancet 364(9445):1583­94. Arole, S., R. Premkumar, R. Arole, M. Maury and P. Saunderson. 2002. "Social stigma: a comparative qualitative study of integrated and vertical care approaches to leprosy." Leprosy Review 73(2):186­96. Atun, R., S. Bennett and A. Duran. 2008. When do vertical (stand-alone) programmes have a place in health systems? Copenhagen: World Health Organization. Atun, R., T.E. de Jongh, F.V. Secci, K. Ohiri and O. Adeyi. 2009 (forthcoming). " Integra- tion of Targeted Health Interventions into Health Systems: A Conceptual Framework for Analysis." Health Policy and Planning. Bausch, D., and B.L. Cline. 1995. "The impact of control measures on urinary schistoso- miasis in primary school children in northern Cameroon: A unique opportunity for controlled observations." American Journal of Tropical Medicine and Hygiene 53(6): 577­80. Bonu, S., M. Rani, and T.D. Baker. 2003. "The impact of the national polio immunization campaign on levels and equity in immunization coverage: evidence from rural North India." Social Science and Medicine 57(10):1807­19. Briggs, C.J., and P. Garner. 2006. "Strategies for integrating primary health services in middle- and low-income countries at the point of delivery." Cochrane Database of Systematic Reviews (2). 63 64 World Bank Working Paper Bryce, J., E. Gouws, T. Adam, R.E. Black, J.A. Schellenberg, F. Manzi, C.G. Victora and J.P. Habicht. 2005a. "Improving quality and efficiency of facility-based child health care through Integrated Management of Childhood Illness in Tanzania." Health Policy and Planning 20:169­76. Bryce, J., C.G. Victora, J.P. Habicht, R.E. Black, R.W. Scherpbier, and MCE IMCI Tech Advisors. 2005b. "Programmatic pathways to child survival: results of a multi-country evaluation of Integrated Management of Childhood Illness." Health Policy and Planning 20:15­117. Chopra, M., S. Patel, K. Cloete, D. Sanders, and S. Peterson. 2005. "Effect of an IMCI inter- vention on quality of care across four districts in Cape Town, South Africa." Archives of Disease in Childhood 90(4):397­401. Cline, B.L., and B.S. Hewlett. 1996. "Community-based approach to schistosomiasis con- trol." Acta Tropica 61(2):107­19. Coura Filho, P., R.S. Rocha, MFF de Lima e Costa, and N. Katz. 1992. "A municipal level approach to the management of schistosomiasis control in PeriPeri, MG, Brazil." Revista Do Instituto De Medicina Tropical De Sao Paulo 34(6):543­48. Cueto, M. 2004. "The origins of primary health care and selective primary health care." American Journal of Public Health 94(11):1864­74. DeGraff, D.S., J.F. Phillips, R. Simmons, and J. Chakraborty. 1986. "Integrating health- services into an MCH-FP program in Matlab, Bangladesh­an analytical update." Studies in Family Planning 17(5):228­34. Deitchler, M., E. Mathys, J. Mason, P. Winichagoon, and MA. Tuazon. 2004. "Lessons from successful micronutrient programs. Part II: program implementation." Food and Nutrition Bulletin 25(1):30­52. Douthwaite, M., and P. Ward. 2005. "Increasing contraceptive use in rural Pakistan: An evaluation of the Lady Health Worker Programme." Health Policy and Planning 20(2): 117­23. Dubowitz, T., D. Levinson, J.N. Peterman, G. Verma, S. Jacob, and W. Schultink. 2007. "Intensifying efforts to reduce child malnutrition in India: An evaluation of the Dular program in Jharkhand, India." Food and Nutrition Bulletin 28:266­73. El Arifeen, S., L.S. Blum, D.M. Hoque, E.K. Chowdhury, R. Khan, R.E. Black, C.G. Victora, and J. Bryce. 2004. "Integrated Management of Childhood Illness (IMCI) in Bangladesh: early findings from a cluster-randomised study." Lancet 364(9445): 1595­1602. Engels, D., J. Ndoricimpa, and B. Gryseels. 1993. "Schistosomiasis-mansoni in Burundi­ progress in its control since 1985." Bulletin of the World Health Organization 71(2): 207­14. Engels, D., B. Sindayigaya, and B. Gryseels. 1995. "Sustainability of schistosomiasis case detection based on primary health care." Transactions of the Royal Society of Tropical Medicine and Hygiene 89(6):599. Gross, U., M. Mamani Diaz, and C. Valle. 2006a. "Effectiveness of the communication pro- gram on compliance in a weekly multimicronutrient supplementation program in Chiclayo, Peru." Food and Nutrition Bulletin 27(4, Suppl. 1):130­42. Gross, U., C. Valle, and M. Mamani Diaz. 2006b. "Effectiveness of distribution of multi- micronutrient supplements in children and in women and adolescent girls of child- bearing age in Chiclayo, Peru." Food and Nutrition Bulletin 27(4, Suppl.1):122­29. Clearing the Global Health Fog 65 Gupta, S.B., B.C. Srivastava, V. Bhushan, and P. Sharma. 1984. "Impact of the integrated child-development services in Uttar-Pradesh." Indian Journal of Medical Research 79(March):363­72. Hossain, S.M.M., A. Duffield, and A. Taylor. 2005. "An evaluation of the impact of a US$60 million nutrition programme in Bangladesh." Health Policy and Planning 20(1):35­40. Huicho, L., M. Davila, M. Campos, C. Drasbek, J. Bryce, and C.G. Victora. 2005a. "Scal- ing up Integrated Management of Childhood Illness to the national level: achieve- ments and challenges in Peru." Health Policy and Planning 20(1):14­24. Huicho, L., M. Davila, F. Gonzales, C. Drasbek, J. Bryce, and C.G. Victora. 2005b. "Imple- mentation of the Integrated Management of Childhood Illness strategy in Peru and its association with health indicators: an ecological analysis." Health Policy and Planning 20:132­41. Jarallah, J.S., S.A. al-Shammari, T.A. Khoja, and M. al-Sheikh. 1993. "Role of primary health care in the control of schistosomiasis. The experience in Riyadh, Saudi Arabia." Tropical and Geographical Medicine 45(6):297­300. Kabatereine, N.B., E. Tukahebwa, F. Kazibwe, H. Namwangye, S. Zaramba, S. Brooker, J.R. Stothard,C.Kamenka,S.Whawell,J.P.Webster,andA.Fenwick.2006."Progresstowards countrywide control of schistosomiasis and soil-transmitted helminthiasis in Uganda." Transactions of the Royal Society of Tropical Medicine and Hygiene 100(3):208­15. Kapil, U., M. Tandon, D. Nayar, S. Ramachandron, R. Yadav and S.N. Dwivedi. 1999. "Nutritional rehabilitation of severely malnourished children at domiciliary level through the integrated child development services (ICDS) scheme: A perspective study." Journal of Tropical Pediatrics 45(2):124­25. Kasturiaratchi, N. D., S. Settinayake, and P. Grewal. 2002. "Processes and challenges: How the Sri Lankan health system managed the integration of Leprosy Services." Leprosy Review 73(2):177­85. Katabarwa, M.N., P. Habomugisha, F.O. Richards, and D. Hopkins. 2005. "Community- directed interventions strategy enhances efficient and effective integration of health care delivery and development activities in rural disadvantaged communities of Uganda." Tropical Medicine & International Health 10(4):312­21. Koop, D.G., B.M. Jackson, and G. Nestel. 2001. "Results of the expanded program on immunization in the Macedonian refugee camps." American Journal of Public Health 91(10):1656­59. Lal, S. 1980. "Better primary health care services utilization through integrated child devel- opment service scheme in Haryana." Indian Journal of Pediatrics 47(387):293­96. Lechtig, A., R. Gross, O. Aquino Vivanco, U. Gross, and D. Lopez de Romana. 2006a. "Lessons learned from the scaling-up of a weekly multimicronutrient supplementa- tion program in the Integrated Food Security Program (PISA)." Food and Nutrition Bulletin 27(4, Suppl.1):160­65. Lechtig, A., R. Gross, J. Paulini, and D. Lopez de Romana. 2006b. "Costs of the multimi- cronutrient supplementation program in Chiclayo, Peru." Food and Nutrition Bulletin 27(4, Suppl.1):151­59. Lopez de Romana, D., S. Verona, O. Aquino Vivanco, and R. Gross. 2006. "Protective effect of multimicronutrient supplementation against anemia among children, women, and adolescent girls in lower-income areas of Chiclayo, Peru." Food and Nutrition Bulletin 27(4, Suppl.1):143­50. 66 World Bank Working Paper Magnussen, L., J. Ehiri and P. Jolly. 2004. "Comprehensive versus selective primary health care: Lessons for global health policy." Health Affairs 23(3):167­76. Masanja, H., J.A. Schellenberg, D. De Savigny, H. Mshinda, and C.G. Victora. 2005. "Impact of Integrated Management of Childhood Illness on inequalities in child health in rural Tanzania." Health Policy and Planning 20:177­84. Naimoli, J. F., A. K. Rowe, A. Lyaghfouri, R. Larbi and L. A. Lamrani. 2006. "Effect of the Integrated Management of Childhood Illness strategy on health care quality in Morocco." International Journal for Quality in Health Care 18(2):134­44. Newell, K. W. 1988. "Selective primary health-care­the counter revolution." Social Science & Medicine 26(9):903­06. Pariyo, G.W., E. Gouws, J. Bryce, G. Burnham, and IMCI Impact Study Team Uganda. 2005. "Improving facility-based care for sick children in Uganda: training is not enough." Health Policy and Planning 20:158­68. Parkash, I., and P. S. Rao. 2003. "Study on integration of the National Leprosy Eradication Programme into primary health care services: a pilot project." Indian Journal of Lep- rosy 75(1):25­35. Peck, R., D.W. Fitzgerald, B. Liautaud, M.M. Deschamps, R.I. Verdier, M.E. Beaulieu, R. GrandPierre, P. Joseph, P. Severe, F. Noel, P. Wright, W.D. Johnson, and J.W. Pape. 2003. "The feasibility, demand, and effect of integrating primary care services with HIV voluntary counseling and testing­Evaluation of a 15-year experience in Haiti, 1985­2000." Journal of Acquired Immune Deficiency Syndromes 33(4):470­75. Phillips, J.F., R. Simmons, J. Chakraborty, and A.I. Chowdhury. 1984. "Integrating health- services into an MCH-FP program­lessons from Matlab, Bangladesh." Studies in Family Planning 15(4):153­61. Rao, P.S.S.S., N. Gift, G.S. Rao, P. Samuel, and R.S. Bushanam. 2002. "Elimination of lep- rosy: the integration of leprosy related activities into the general health services of Tamil Nadu." Leprosy Review 73(2):123­29. Rojas, W., S. Botero, and H.I. Garcia. 2001. "An integrated malaria control program with community participation on the Pacific Coast of Colombia." Cadernos de Saude Pub- lica 17(Suppl.):103­13. Saiyed, F., and S. Seshadri. 2000. "Impact of the integrated package of nutrition and health services." Indian Journal of Pediatrics 67(5):322­28. Simmons, G.B., D. Balk, and K.K. Faiz. 1991. "Cost-effectiveness analysis of family-planning programs in rural Bangladesh­evidence from Matlab." Studies in Family Planning 22(2):83­101. Sleigh, A., S. Jackson, X. Li, and K. Huang. 1998b. "Eradication of schistosomiasis in Guangxi, China. Part 1: Setting, strategies, operations, and outcomes, 1953­92." Bul- letin of the World Health Organization 76(4):361­72. ------. 1998a. "Eradication of schistosomiasis in Guangxi, China. Part 2: Political econ- omy, management strategy and costs, 1953­92." Bulletin of the World Health Organi- zation 76(5):497­508. ------. 1998c. "Eradication of schistosomiasis in Guangxi, China. Part 3. Community diagnosis of the worst-affected areas and maintenance strategies for the future." Bul- letin of the World Health Organization 76(6):581­90. Tandon, B.N., N. Gandhi, B. Kumar, R.K.P. Sinha, V.K. Karan, R.G. Bajaj, D.K. Yadava, A. Rouf, S.P. Sharma, M. Ahmed, C.K. Sasidharan, K. Rajan, K. Thomas, M. Ketkar, Clearing the Global Health Fog 67 K.K. Kaul, N.R. Bhandari, S. Upadhyay, T. Guha, M. Natu, D.M. Meshram, R.B. Patel, S.U. Warerkar, M.K. Paul, A.C. Patowary, G. Camotim, R. Thakur, D. Sharma, S.P. Mehta, E.Y. Singh, L.I. Devi, D. Roy, Rothangliana, D.K. Srivastava, K.K. Vadhera, V. Mohan, A. Chandrabhushanam, J. Subramaniyam, and B. Roy. 1992. "Immunization coverage in India for areas served by the integrated child-development services pro- gram." Bulletin of the World Health Organization 70(4):461­65. Tandon, B.N., and A. Sahai. 1988. "Immunization in India­contribution of integrated child-development services scheme to Expanded Program of Immunization." Journal of Tropical Pediatrics 34(6):309­12. Toledo Romani, M.E., V. Vanlerberghe, A. Baly Gil, E. Ceballos, L. Valdes, M. Searret, M. Boelaert, and P. van der Stuyft. 2007. "Towards active community participation in dengue vector control: results from action research in Santiago de Cuba, Cuba." Transactions of the Royal Society of Tropical Medicine and Hygiene 101(1):56­63. Trivedi, S., B.C. Chaparwal, and S. Thora. 1995. "Utilization of ICDS scheme in children one to six years of age in a rural block of central India." Indian Pediatrics 32(1):47­50. Tuladhar, J.M., and J. Stoeckel. 1982. "The relative impacts of vertical and integrated FP/MCH programs in rural Nepal." Studies in Family Planning 13(10):275­86. Walsh, J.A., and K.S. Warren. 1979. "Selective primary health-care­interim strategy for disease control in developing-countries." New England Journal of Medicine 301(18): 967­74. Warren, K.S. 1988. "The evolution of selective primary health-care." Social Science & Med- icine 26(9):891­98. Wisner, B. 1988. "GOBI versus PHC­some dangers of selective primary health-care." Social Science & Medicine 26(9):963­69. World Bank and World Health Organization. 2006. "High Level Forum on the Health Mil- lennium Development Goals: Selected Papers 2003­2005." Zhang, Y.F., Y.H. Dai, and S. M. Zhang. 2007. "Impact of implementation of Integrated Management of Childhood Illness on improvement of health system in China." Jour- nal of Paediatrics and Child Health 43:681­85. Eco-Audit Environmental Benefits Statement The World Bank is committed to preserving Endangered Forests and natural resources. We print World Bank Working Papers and Country Studies on 100 percent postconsumer recy- cled paper, processed chlorine free. The World Bank has formally agreed to follow the rec- ommended standards for paper usage set by Green Press Initiative--a nonprofit program supporting publishers in using fiber that is not sourced from Endangered Forests. For more information, visit www.greenpressinitiative.org. In 2008, the printing of these books on recycled paper saved the following: Trees* Solid Waste Water Net Greenhouse Gases Total Energy 355 16,663 129,550 31,256 247 mil. * 40' in height and Pounds Gallons Pounds CO2Equivalent BTUs 6­8" in diameter Clearing the Global Health Fog is part of the World Bank Working Paper series. These papers are published to communicate the re- sults of the Bank's ongoing research and to stimulate public discussion. A longstanding debate on health system organization relates to the benefits of integrating programs that emphasize specific interventions into mainstream health systems to increase access and improve health outcomes. This debate has long been characterized by polarization of views and ideologies, with protagonists for and against integration arguing relative merits of each approach. Recently, the debate has been rekindled due to substantial rises in externally-funded programs for priority health, nutrition, and population (HNP) interventions and an increase in in- ternational efforts aimed at health system strengthening. However, all too frequently these arguments have not been based on hard evidence. In this book we present findings of a systematic review that explores a broad range of evidence on: (i) the extent and nature of integration of targeted health programs that emphasize specific interventions into critical health systems functions; (ii) how the integration or non-integration of health programs into critical health systems functions in different contexts have influenced program success; and (iii) how contextual factors have affected the extent to which these programs were integrated into critical health systems functions. The findings provide a new synthesis of evidence to inform the de- bate on health systems and targeted interventions. In practice a rich mix of solutions exists. While the discussion on the relative merits of integrating health interventions will no doubt continue, discus- sions should move away from the highly reductionist approach that has polarized this debate. World Bank Working Papers are available individually or on standing order. Also available online through the World Bank e-Library (www.worldbank.org/elibrary). ISBN 978-0-8213-7818-2 THE WORLD BANK 1818 H Street, NW Washington, DC 20433 USA Telephone: 202 473-1000 Internet: www.worldbank.org SKU 17818 E-mail: feedback@worldbank.org