Assessing Costs and Benefits of Sexual and Reproductive Health Interventions

Publication date: 2004

Assessing Costs and Benefits Of Sexual and Reproductive Health Interventions Michael Vlassoff, Susheela Singh, Jacqueline E. Darroch, Erin Carbone and Stan Bernstein Occasional Report No. 11 December 2004 This report was written by Michael Vlassoff, inde- pendent consultant, Susheela Singh, vice president for research at the Alan Guttmacher Institute (AGI), Jacqueline E. Darroch, former senior vice president and vice president for science, Erin Carbone, research associate and Stan Bernstein, United Nations Popula- tion Fund. Akinrinola Bankole, associate director for international research, with the assistance of Rubina Hussain, senior research assistant, contributed to data analyses; Kathleen Randall, with the assistance of Michael Greelish and Judith Rothman, was responsi- ble for production of the report and Rose MacLean pro- vided editorial support. The authors are grateful to the following colleagues who reviewed an earlier draft of the report and made many helpful suggestions: John Stover, Futures Group International; John Bongaarts, Population Council; Rudolfo Bulatao, independent consultant; and John Ross, Futures Group International, who deserves spe- cial thanks for reviewing this report and for providing advice, data and information for the estimates present- ed in Chapter 3. The final version of this report benefits from com- ments provided by our colleagues on AGI’s recently published report, Adding It Up: The Benefits of Invest- ing in Sexual and Reproductive Health Care, which presents the main findings from this Occasional Re- port. Special thanks go to the following individuals: Sara Seims, who inspired and guided the development of the report, and AGI colleagues Ann Biddlecom, Susan Cohen, Beth Fredrick, Dore Hollander and Cory Richards, who reviewed and provided comments on drafts. The authors are also grateful to the following colleagues who made many helpful suggestions: Arnab J. Acharya, Research Triangle Institute; David Bloom, Harvard University; Kwesi Botchwey, Harvard Uni- versity; Lynn Freedman, Columbia University; Barbara Janowitz, Family Health International; and Thomas Merrick, the World Bank. In addition, the authors appreciate the advice, infor- mation and materials provided by the following indi- viduals: Lori Bollinger, Futures Group International; John Cleland, London School of Hygiene and Tropical Medicine; Helga Fogstadt, World Health Organization; Gaverick Matheny, The Johns Hopkins University; Bill McGreevy, Futures Group International; Anthony Measham, the World Bank; Anne Mills, London School of Economics; Philip Musgrove, George Wash- ington University; Malcom Potts, University of Cali- fornia, Berkeley; Iqbal Shah, World Health Organiza- tion; Eva Weissman, UNFPA; George Zeidenstein, Harvard University; and Hania Zlotnick, United Na- tions Population Division. Finally, special thanks are due to Stirling Scruggs, UNFPA, whose initiative and vision were essential to the creation of this report. The research for and prepa- ration of this report were supported by a grant from UNFPA. The views expressed in this publication are those of the authors and do not necessarily represent the views of UNFPA, the United Nations or any of its affiliated organizations. Suggested citation: Vlassoff M, et al., Assessing Costs and Benefits of Sexual and Reproductive Health Interventions: Occasional Report, New York: The Alan Guttmacher Institute, 2004, No. 11. Acknowledgments Chapter I: Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Benefits of Interventions: Medical versus Nonmedical Perspectives . . . . . . . . . . . . . . . . . 5 Cost-benefit Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Plan of This Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Chapter 2: Major Approaches to Evaluating the Costs and Benefits of Sexual and Reproductive Health Programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 The Global Burden of Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Disease Control Priorities in Developing Countries (DCP-1) . . .13 World Development Report 1993: Investing in Health . . . . . . . .16 Macroeconomics and Health . . . . . . . . . . . . . . . . . . . . . . . . . . . .17 Current and Ongoing Research . . . . . . . . . . . . . . . . . . . . . . . . . . . .19 Cairo 1994: The Cost of Reproductive Health . . . . . . . . . . . . . . . .20 UNFPA Costing Initiative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21 The Effect of Family Planning Programs . . . . . . . . . . . . . . . . . . . .21 UNAIDS Cost Estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21 WHO Mother-Baby Package . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22 Cost-Savings Analyses by The Alan Guttmacher Institute . . . .22 Cost-Benefit and Cost-Effectiveness Models in Reproductive Health . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23 Selected Country-Level Cost-Benefit Studies of Family Planning Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24 Tables: 2.1 Burden of Disease related to Reproductive Health . . . . . . . . . . . . . 27 2.2 Percentage of DALYs, YLLs and YLDs by WHO Regions, 2001 . . . . . 27 2.3 HIV/AIDS has become a major cause of sexual and reproductive ill health worldwide . . . . . . . . . . . . . . . . . . . . . . . . . . . .28 2.4 Deaths and DALYs in Developing Countries Due to Unsafe Sex and Lack of Contraception, 2000 . . . . . . . . . . . . . . . . . . . . . . . . .28 2.5 Savings per Birth Averted . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .28 2.6 Costs of Averting a Birth through Family Planning, 1980 . . . . . . . .29 2.7 Benefits verses Costs in Three Hypothetical Countries . . . . . . . . .29 2.8 Components of Two Obstetric Care Models . . . . . . . . . . . . . . . . . . .30 2.9 Costs and benefits of five hypothetical programs in “Himort”, a country with a population of one million and high mortality . . . . 31 2.10 Estimates of the cost effectiveness of selected interventions, World Health Report 1993 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .32 2.11 Elements of Public Health Component of Essential Package . . . .32 2.12 Elements of Clinical Services Component of Essential Package .32 2.13 Resource Requirements for the ICPD Program of Action . . . . . . . .33 2.14 Estimated Average Costs per contraceptive user by major region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33 2.15 Resource Requirements by Major Region . . . . . . . . . . . . . . . . . . . . .34 2.16 Resource Requirements for Global HIV/AIDs Program . . . . . . . . . .34 2.17 Interventions Included in the Mother-Baby Package . . . . . . . . . . .34 2.18 Recent Cost-benefit and Cost-effectiveness Models in Reproductive Health . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .35 2.19 Savings in Government expenditures in Thailand . . . . . . . . . . . . . .35 Chapter 3: A New Look at the Benefits And Costs of Contraceptive Services in Developing Countries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37 Estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .43 Tables: 3.1 Estimated number and percent of women with unmet need and using no contraceptive, 2003 . . . . . . . . . . . . . . . . . . . . . . .49 3.2 Married women of reproductive age . . . . . . . . . . . . . . . . . . . . . . . . . .50 3.3 Number of unintended pregnancies by region . . . . . . . . . . . . . . . . . 51 3.4 Estimated distribution of unplanned pregnancies according to outcome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .52 Chapter 4: A Broader Approach to Measuring Benefits and Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .53 Benefits of Sexual and Reproductive Health Services Measurement of costs: how can it be improved? . . . . . . . . . .54 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60 Key Points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60 Chapter 5. Summary and Conclusions . . . . . . . . . . . . . . . 61 Appendix: Definitions, Methodology and Data Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .65 Appendix Tables: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .68 1.1. Methods of economic evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . .68 2.1. Diseases/conditions included in the Global Burden of Disease (1990) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .70 Table of Contents 2.2. Codes from the International Classification of Disease 9th edition used in the 1996 Global Burden of Disease . . . . . . . . . . 71 2.3. World Health Organization regional classification . . . . . . . . . . . . . 72 2.4. Summary tables of the UNFPA Costing Initiative . . . . . . . . . . . . . . . 74 3.1a.List of countries by geographic regions and subregions . . . . . . . .82 3.1b.List of countries according to income level. . . . . . . . . . . . . . . . . . . .84 3.2. Est. total population and no. and percent distribution of females (15-49) by union status, risk for unintended pregnancy and fertility-prefererence . . . . . . . . . . . . . . . . . . . . . . . .86 3.3. Est. annual no. of pregnancies by outcome and intention status, and % of all unintended pregnancies and births that are unintended . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .87 3.4 Est. annual no. of maternal deaths, % distribution and mortality rates per 100,000 events, by pregnancy intention status and non-abortion or abortion-related cause, deaths to infants by pregnancy intention status and rate per 1,000 births . . . . . . . . . .88 3.5. Est. no. of DALYs, YLLs and YLDs from maternal causes, by non-abortion and abortion-related pregnancies and perinatal causes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .89 3.6. Est. no of women at risk of unintended pregnancy, by contraceptive method use and fertility-preference status . . . . . .90 3.7. Percentage distribution of women (15-49) at risk for unintendedpregnancy by contraceptive method and fertility- preference status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .93 3.8. Est. annual service costs for current modern contraceptive users, by type of cost and average total cost per user . . . . . . . . . .94 3.9. Est. no. of unintended pregnancies to current modern contraceptive users, by type of outcome, % they represent of all unintended pregnancies, births and abortions; no. of unintended pregnancies occuring if they were not using any contraceptive method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .95 3.10.Est. annual no. of unintended pregnancies averted by current use of modern contraceptive methods, type of pregnancy outcome, est. no. of deaths, children losing their mothers and DALYs averted . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .96 3.11. Est. no. of women with unmet need and % they represent of all women (15-49) with unmet need; % distribution of women with unmet need by fertility-preference status, union status and method use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .97 3.12.Est. annual pregnancies to women with unmet need, by fertility-preference status and pregnancy outcome. . . . . . . . . . . .98 3.13.Est. annual maternal and infant deaths from unintended pregnan- cies to women with unmet need, no. of children who lose their mothers through these deaths and the no. of DALYs lost from these unintended pregnancies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .99 3.14.Est. no. of new users if all women with unmet need used contraceptives, by method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .100 3.15.Est. total annual costs to provide contraceptive services to women currently with unmet need, by type of cost and women’s fertility-preference status and average cost per user . . . . . . . . . .101 3.16.Est. total annual costs to provide contraceptive services to women currently with unmet need, by type of method . . . . . . . . .102 3.17. Est. annual unintended pregnancies to women with unmet need, no. estimated to occur if they all used modern contraceptive methods and no. averted if all used modern contraceptive methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .103 3.18.Est. no. of unintended pregnancies averted if women with unmet need all used modern contraceptive methods, by outcome and fertility-preference status . . . . . . . . . . . . . . . . . . . . . .104 3.19.Est. maternal and infant deaths averted if women with unmet need used modern contraceptive methods; % distribution of deaths averted; children who would not lose their mothers; no. of maternal and infant DALYs averted . . . . . . . . . . . . . . . . . . . . .105 4.1 Interventions: information and services . . . . . . . . . . . . . . . . . . . . . .106 4.2. Examples of medical benefits of sexual and reproductive health services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .107 4.3 Examples of nonmedical benefits of contraceptive services . . . .108 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .109 In this current climate of financial constraints coupled with competing priorities among developmental goals, it becomes ever more critical for policymakers and oth- ers responsible for allocating resources to have first- rate tools available as a guide for effective decision making. The overall aim of this report is to inform such decision makers about the key findings of existing studies about the costs and benefits of investments in sexual and reproductive health, to identify what factors the studies encompass and what they leave out, and to provide a complete picture of what the costs and bene- fits would look like, including benefits that are hard to measure. This report is a technical companion to a shorter monograph.1 It has three parts: (1) a review and synthesis of what is known about the costs and benefits of investments in sexual and reproductive health; (2) a comprehensive outline that can be used by researchers and policymak- ers to view the gamut of costs and benefits, which, it is hoped, will lead to improvement in the measurement of costs and benefits of sexual and reproductive health in- vestments; and (3) in order to demonstrate the advan- tages of taking a more comprehensive approach to measuring costs and benefits, a partial application of the framework in the reproductive and maternal health field, namely in the area of contraceptive services and supplies. Benefits of Interventions: Medical versus Nonmedical Perspectives A medical perspective regarding the benefits of health interventions has attained a predominant position in pol- icy analysis over the last decade. Starting with the World Bank’s World Development Report 1993: Investing in Health2 and continuing with the Disease Control Prior- ities project,3 the Global Burden of Disease project,4 the World Health Organization’s Commission on Macro- economics and Health (CMH)5 and ultimately the Unit- ed Nations Millennium Development Goals (MDGs),6 a major policy thrust has concentrated on the medical benefits of more effective health care services and sys- tems. The CMH report, however, broadened the scope of benefits beyond direct medical ones (deaths and dis- ability averted) to economic gains from better health. While it is only natural that the effectiveness of health interventions be primarily measured against gains in the health status of the affected population, there are important aspects of sexual and reproductive health, the focus of this project, that do not fit into this medical perspective. Bearing children is not an illness. Ideally, it is a healthy reproductive act voluntarily un- dertaken by women in their desire to build a family. Nonetheless—and aside from the health risks associat- ed with pregnancy, delivery and the postpartum peri- od—important and significant costs result from un- planned pregnancies. Unplanned pregnancies are those that occur to women who wish to have no more children or who wish to postpone childbearing to some future time. Not only do unplanned pregnancies lead to im- portant negative medical outcomes in the form of death and disability, they also generate other important nega- tive consequences in economic, social and psychologi- cal areas; ultimately, these outcomes negatively impact socioeconomic development and poverty reduction. The consensus reached at the 1994 International Conference on Population and Development (ICPD) which resulted in the ICPD Program of Action,7 called for concrete and quantitative improvement in many key components of sexual and reproductive health, includ- ing safe motherhood. At Cairo, governments agreed to a 20-year program to improve sexual and reproductive health and set intermediate benchmarks, which the global community committed itself to achieve. For ex- ample, by 2005 the unmet need for contraceptives is to be reduced by 50% by expanding access to a broad range of contraceptive methods; prenatal and basic ob- stetric care, as well as testing and treatment for sexual- ly transmitted infections (STIs), are to be available in 60% of all primary health care facilities. Concrete tar- gets were set to reduce maternal mortality, combat HIV Chapter 1 Introduction 5 and increase child survival. The donor community agreed to provide $5.7 billion annually by 2000 and $6.1 billion by 2005 to achieve these goals, and devel- oping countries themselves were to contribute $11.3 billion by 2000 and $12.4 billion by 2005.8 In 2000, however, the global expenditure on repro- ductive health was only $10.9 billion—$6.1 billion short of the total amount committed. Donor countries contributed $2.6 billion annually—less than half of their commitment. Developing countries contributed $8.3 billion—about 73% of what they committed at ICPD.9 Despite the agreement at Cairo, competing de- mands for resources for development have led to fi- nancial shortfalls, which threaten the goals set at Cairo. Cost-benefit Methodology Several techniques are available for the economic eval- uation of a proposed action – be it a project, interven- tion, investment or budget allocation.10 Appendix 1.1 is adapted from a useful summary of different tech- niques.11 As can be seen in the appendix, each analyt- ical method has advantages and disadvantages, and each is best suited to particular situations. The evaluative technique that is especially useful and that is a focus of this study is cost-benefit analysis (CBA). CBA can be a very useful evaluation technique in the area of sexual and reproductive health interven- tions. Together with cost-saving analysis, which can be considered an abbreviated form of CBA, CBA is capa- ble of demonstrating both the intrinsic worth of a proj- ect or intervention and the relative benefits of the proj- ect or intervention vis-à-vis some other project or intervention. Because both costs and benefits are ex- pressed in monetary units—usually dollars—it is pos- sible to construct cost-benefit ratios, which can be used by themselves (any project whose cost-benefit ratio is greater than unity yields a net gain) or comparatively (whichever project has the highest cost-benefit ratio should have the highest priority). Despite these advantages, several difficulties have been noted in CBA.12 First of all, it is difficult, if not impossible, to include all benefits in the analysis. To the extent that important benefits are omitted, the result of the CBA may lead to an erroneous decision about undertaking the project. It is also possible that negative benefits (for example, unintended consequences) may also have been omitted from the analysis; in this case, the net value of undertaking the project may be over- stated. Secondly, some benefits may be hard to quantify monetarily, and those that lack adequate measurement methodologies may be omitted. If CBA is being used to compare projects from different sectors, this diffi- culty may bias results. For example, one researcher has pointed out that health benefits are “notoriously diffi- cult to estimate,”13 so that a CBA comparison between a health-sector and a nonhealth-sector project may eas- ily be biased in favor of the nonhealth project. A third problem encountered in CBA studies is the present valuation of future benefits.14 The solution usu- ally adopted is to apply an annual rate of discount. The choice of any rate of discount is to a certain extent ar- bitrary (rates from 0–15% have been noted in the liter- ature), although attempts have been made to study pub- lic preferences for future benefits.15 Plan of This Report This report consists of three major components: • First, the report critically reviews and synthesizes re- sults from the major approaches to measuring costs and benefits as currently applied to the three main areas of sexual and reproductive health: contraceptive services; STIs including HIV/AIDS (prevention, diagnosis and treatment); and maternal health (prenatal care, treat- ment for unsafe abortion and obstetric services). The report reviews the strengths and weaknesses of the major methodological approaches which have been developed. This component includes an overview of the different measures through which costs and bene- fits are estimated. One common shortcoming of these estimates is the inclusion of only the medical benefits, and the exclusion of nonmedical benefits—for exam- ple, a medical benefit of increased contraceptive serv- ices is the prevention of unplanned births, whereas nonmedical benefits would include increased propor- tions of women completing their education, improve- ment in women’s health by wider spacing of births and improved survival of infants, as well as better outcomes for other family members and for society at large. There are also many disparities in the methodolo- gies applied to assess costs. Some estimates include only direct economic costs—such as the fees or salary of the physician or midwife providing a specific serv- ice or the cost of purchasing drugs—and do not include indirect economic costs—such as the use of facilities and the recurring costs of clinic staff. When very dif- ferent measures are employed by different studies to estimate the costs and benefits of interventions in re- productive and maternal health, results are not compa- rable and it is difficult to reach firm conclusions. Con- sequently, it is difficult to apply results from these analyses to decision making, and the usefulness of The Alan Guttmacher Institute 6 these tools becomes limited. • Secondly, the report presents a comprehensive outline of the benefits of investing in each of the three main areas of sexual and reproductive health mentioned above. Some of the benefits of these investments have not previously been measured and some are not meas- urable in quantifiable terms. Nevertheless, it is impor- tant to specify all aspects of costs and benefits so that, even as measurable and quantified estimates are calcu- lated and utilized, policymakers are aware of gaps, weaknesses and the likelihood of under- or over-esti- mation. This comprehensive approach will illuminate areas for further research and improvement in cost- benefit methodologies, as well as ways to enhance the comparability of studies. Through this framework, the report also addresses ways to overcome one of the major shortcomings of current methodologies—the lack of consideration given to how investments in one area of sexual and re- productive health can have a positive impact on anoth- er area. For example, family planning services have benefits not only for the health and social status of women but also for the health and survival of children. Similarly, the costs of investments in contraceptive services and supplies can have benefits for preventing STIs (by increasing condom use) and managing infec- tions (through integration of STI and family planning services). The report identifies these spillover benefits in order to increase awareness of these benefits and of the need to measure them. • Thirdly, the report provides new estimates—to the ex- tent possible with available data—of the costs and ben- efits related to investing in one area of reproductive health, contraceptive services and supplies, to demon- strate the advantages of taking a more comprehensive approach to measuring costs and benefits. The report estimates the costs of increasing contraceptive servic- es and supplies to meet the needs of both married cou- ples and of unmarried, sexually active women. The benefits include births averted; total pregnancies avert- ed; unsafe abortions averted; improvements in mater- nal health and survival achieved through the prevention of unsafe abortion and of unplanned and unwanted births and pregnancies; improvements in child health and survival achieved through better spacing of births and the prevention of high-risk pregnancies; and the number of children who would not lose their mothers. In conclusion, the report points out that important work has been done to evaluate health interventions, but current approaches fail to account for all the bene- fits of sexual and reproductive health interventions and therefore underestimate the impact these interventions could have. The report offers recommendations for fu- ture research to improve existing methodologies. Assessing Costs and Benefits 7 8 This chapter reviews and synthesizes results from major research endeavors that have focused on the measurement of the costs and benefits of sexual and re- productive healthcare. The first objective of the chap- ter is to summarize the key findings of the major stud- ies in this area that have been carried out over the last decade or so. A second objective is to discuss the range of methodologies used in these studies to make clear the advantages as well as the assumptions and limita- tions of each of the methods employed. The goal is to stimulate more work and improve current methodolo- gies by broadening the range of costs and benefits in- cluded in evaluating sexual and reproductive health interventions. The review covers two large-scale, on-going re- search projects—the Global Burden of Disease (GBD) and Disease Control Priorities (DCP)—that have been sponsored by the World Bank and the World Health Or- ganization (WHO).16 First-phase results from both of these projects fed into a very influential report by the World Bank in 1993 on the economics of health strate- gies. The WHO’s Commission on Macroeconomics and Health (CMH) carried the economic analysis of health one step further by examining nonmedical ben- efits of health programs in addition to medical ones. Meanwhile, WHO, the United Nations Population Fund (UNFPA) and the United Nations Programme on HIV/AIDS (UNAIDS) initiated costing studies of re- productive health interventions, including contracep- tive services, safe motherhood and the prevention and treatment of HIV/AIDS.17 Moreover, several comput- er-based models have been recently developed to help countries perform economic analyses of health systems more easily,18 and several country-specific cost-bene- fit analyses of family planning programs have been car- ried out in the 1980s and 1990s.19 All of these studies are reviewed in this chapter. Studies fall into a number of different categories, and specific labels or terms are used to describe them (Cost Effectiveness Analysis, Cost-Consequence Analysis, Cost Utility Analysis, Cost Benefit Analysis; Cost Savings Analysis). Definitions and examples of each of these types of studies are given in Appendix 1.1. Finally, this chapter also reviews certain major studies examining only costs since they form an im- portant source of costing information, a necessary building block in the analysis of costs and benefits. The Global Burden of Disease GBD20 is an ongoing international initiative cospon- sored by WHO and the World Bank. It sprang into prominence when the World Bank’s influential World Development Report 199321 used the GBD approach to extensively evaluate priorities for resource allocation in the health sector. GBD represents a large and important effort to reli- ably assess epidemiologic conditions and the burden of disease on the most detailed format possible. As Dean Jamison, editor of Disease Control Priorities in De- veloping Countries has stated: “Publication of the Global Burden of Disease and Injury Series marks the transition to a new era of health outcome account- ing–an era for which these volumes establish vastly higher standards for rigor, comprehensiveness and in- ternal consistency.”22 GBD provides detailed estimates of the burden of death and disability resulting from all major diseases and risk factors measured using the same metric, thereby providing a wealth of compara- ble policy-relevant data. The World Development Report 1993: Investing in Health identified a number of priority health interven- tions based on GBD data and assessed the cost and ef- fectiveness of curative and preventive health interven- tions known to reduce this burden. Both were measured in terms of disability-adjusted life years (DALYs), the burden of disease in terms of DALYs lost, and the cost- effectiveness of interventions in terms of cost per DALY gained. Interventions were classified as high Chapter 2 Major Approaches to Evaluating the Benefits and Costs of Sexual and Reproductive Health Programs 9 priority if the burden of disease was large and the cost- effectiveness of interventions high. The results of the GBD study were published by WHO in 1996,23 and the cost-effectiveness information was reported in another volume.24 What are DALYs? Although from a methodological point of view the GBD enterprise included several innovative approach- es towards systematizing available data, especially where data were patchy or of uneven quality, the study has become best known for its promotion of the DALY as a standard indicator of disease burden. A discussion of the DALY and its component measures, years of life lost (YLLs) and years lived with disability (YLDs), is therefore in order. The DALY is designed to quantify the burden of dis- eases by taking into account not only mortality but also morbidity. Based on the classifications of the Interna- tional Classification of Diseases (9th revision), also known as ICD-9,25 107 diseases and 483 disabling se- quelae were chosen in an effort to cover all possible causes of mortality and about 95% of the possible caus- es of disability. (The complete list of diseases included is shown in Appendix 2.1.) For all those diseases and their sequelae, the number of healthy life years lost due to premature mortality and morbidity were calculat- ed.26 DALYs have two components: an estimation of years of life lost by death and an estimation of years lived with disability after contracting a disease or de- veloping a disabling condition. To determine the number of YLLs due to premature mortality, the GBD study assigned each death to a par- ticular disease category and grouped all deaths by age, sex and region. This exercise was based on death records where available and “expert judgment” where no records were available. The number of YLLs was es- timated, by evaluating the differences between the ac- tual age at death and an ideal standard life expectancy at that age. In the interest of equity, the same ideal life expectancy was used for all countries. Life expectancy at birth was assumed to be 82.5 years for females, and 80 for males. For YLDs, the GBD study estimated the incidence of cases by age, sex and region on the basis of community surveys, or where those were not available, again on “ex- pert opinion.” YLDs were then obtained by multiplying the expected duration of the disability (to recovery or death) by a disability weight that measured the severity of the disease-induced disability compared with death. • Disability weighting. In order to compare YLLs and YLDs, severity weights had to be assigned to years lived with particular disease sequelae. In fact, severity weights were assigned to all of the 483 disabling se- quelae considered in the GBD study. Disability was considered in six broad classes, each with a severity weight between 0 (perfect health) and 1 (equivalent to death). The classification was carried out by an inter- national panel of health experts who were asked to focus solely on functional disability. Thus, social, cul- tural or economic factors which might impact the over- all “burden” or the ability of people to cope were ex- plicitly excluded. For example, rectovaginal fistula was assigned a weight of 0.43 and infertility a weight of 0.18. By contrast, a leg amputation carried a weight of 0.30 and cretinism a weight of 0.80. Once the YLL and YLD for a particular disease or condition were estimated, the corresponding DALY was calulated, simply: DALY = YLL + YLD. Two further methodological considerations should be mentioned vis-à-vis the construction of DALYs: ad- justments for age and time discounting. • Age Weighting. In computing DALYs, time lived at different ages was valued differently; a year of life lived by a young or middle-aged adult was given a greater weight than a year of life lived by a child or an elderly person. Due to the lack of empirical data on age preferences, a formula for weighting life years lost at different ages was chosen somewhat arbitrarily by a group of health experts.* For example, one healthy life- year lost by a five-year-old was worth only 0.66 years, but a healthy life-year lost by a 25-year-old was count- ed as almost 1.5 years. • Time Preference. DALYS were also adjusted for time preference. Similar to financial discounting, fu- ture life saved or improved by health interventions was given a lesser value than life saved today (using an an- nual discount rate of 3%). This follows the general ob- servations (1) that people value, say, a $100 paid at a certain time in the future less than they value $100 paid today and (2) that the further into the future the pay- ment is to be made, the less present value it has. Dis- counting future DALYs using a rate of 3% meant, for instance, that one year of healthy life was counted as approximately half a year if it occurred 22.5 years from now and as just three months if it occurred some 45 years into the future. *The weighting frequency distribution resembles a Poisson distribution. The Alan Guttmacher Institute 10 Global burden of disease estimates The GBD study estimated that the number of deaths worldwide was 50.5 million* in 1990 and 56.6 million in 2001† A variety of sources, from vital statistics to cause-of-death modeling, were used to estimate deaths by approximately 100 causes of death in eight regions, among sexes, and both by seven age-groups. Appendix 2.1 lists the diseases and conditions used in the GBD study. A series of steps were followed to adjust the dis- aggregated estimates to be consistent with cause-of- death models and with existing epidemiologic analy- ses. The authors acknowledged that, because data from some regions and about some causes of death are more complete and more accurate than from other regions and about other causes, the estimates had narrower or wider confidence intervals associated with them.27 For instance, estimates from Sub-Saharan Africa are seen as the least reliable of all the regions. Appendix 2.2 shows the exact definitions used for identifying repro- ductive health conditions. Burden of disease related to reproductive health in 1990 and 2001. For 1990, the worldwide estimate of YLLs was 907 million, the estimate of YLDs was 473 million, and the worldwide estimate of DALYs lost was 1.38 billion. By 2001, total DALYs lost had grown to 1.47 billion.28 For the diseases and conditions related to reproductive health estimated by the GBD study see Table 2. Further analyzing the GBD results for 1990, the death and disability components of DALYs show dis- tinct patterns between regions (see Table 2.2) and among age groups. Worldwide, about two– thirds of lost DALYs are due to premature deaths (i.e., the YLLs due to deaths from disease) and only one–third are due to disability. For STIs and abortion, however, these pro- portions are roughly reversed, and for maternal condi- tions as a whole, including abortion, only 45% of lost DALYs are due to death. In the case of HIV/AIDS, however, more than 70% of the total burden is con- tributed by premature deaths. There are notable differences in the composition of DALYs between regions. Only 1% of STI-related DALYs are caused by premature death in the Americ- as and Europe, whereas 37% of DALYs in Africa, East- ern Mediterranean, South East Asia and the Western Pacfic regions combined are due to such deaths. Simi- larly, YLLs related to maternal conditions are a much larger factor in developing countries—contributing al- most half of all DALYs—than in developed countries, where YLLs contribute only 10%. Looking at the disease burden by age group, there are some striking developing-developed world differ- ences. The burden of premature death (YLLs) due to STIs is felt at much earlier ages in developing coun- tries: 94% of YLLs among males and 74% among fe- males happen to those younger than 15. In developed countries, the corresponding figures are 20% among males and 0% among females. In the case of the bur- den from HIV/AIDs, a similar pattern is found, al- though the size of the differentials is decreased. With regard to maternal conditions, on the other hand, the age patterns in the two regions are broadly equal. According to WHO’s 2001 estimates, sexual and re- productive health problems account for 18% of the total global burden of disease and 32% of the burden among women of reproductive age (15–44) world- wide:29 • Maternal conditions (hemorrhage or sepsis result- ing from childbirth, obstructed labor, pregnancy-relat- ed hypertensive disorders and unsafe abortion) account for 2% of all DALYs lost (13% of all DALYs lost among women of reproductive age). • Perinatal conditions (low birth weight, birth as- phyxia and birth trauma) account for 7% of all DALYs lost. • HIV/AIDS accounts for 6% (14% among women of reproductive age). Other sexual and reproductive health conditions ac- count for 3% (5% among women of reproductive age).‡ Burden of disease due to unsafe sex The GBD initiative also reported on the burden of dis- ease resulting from the practice of unsafe sex based on 1990 DALYs.30 The diseases and conditions consid- ered in this 1998 report include HIV/AIDS, STIs (gonorrhea, syphilis and chlamydia only), human pa- pilloma virus (HPV), hepatitis B, complications in pregnancy and abortion. The WHO World Health Report 2002 has presented *The study did not refer to any other source for death estimates. Nev- ertheless, the current estimate by the United Nations Population Divi- sion of 50.35 million is quite close. †In this section, both the original Global Burden of Disease study (see reference 22) and the most recent update of GBD (see reference 28) are referred to. Revisions to estimates of DALYS are in progress. Preliminary results show a small decline in DALYs due to HIV/AIDS/AIDS, overall and among women of reproductive age. ‡These include STIs other than HIV/AIDS, iron-deficiency anemia among women aged 15–44, breast cancer, ovarian, cervical and uterine cancer, and genito urinary diseases, excluding nephritis and nephrosis. Assessing Costs and Benefits 11 a more recent estimate updated to 2000 for deaths and disability due to sexual health risks, which were de- fined differently from the 1998 estimates. The 2002 re- port included HIV/AIDS, STIs and cervix uteri cancer as unsafe sex and lack of contraception.* Table 2.4 shows the estimates for 2000. Two notable changes can be observed from the 1990 estimates to the 2000 estimates. First, due to the enor- mous increase in the burden caused by HIV/AIDS, the magnitude of the burden from the (revised) definition of unsafe sex more than doubles (from 40 million to 91 million DALYs) and because the HIV/AIDS burden is shared fairly equally between men and women, the proportionate burden of unsafe sex borne by women decreases substantially—women bore 80% of the bur- den in 1990 but only a little more than half in 2000. Second, the burden attributed to lack of contraception (“complications in pregnancy” in the earlier publica- tion† increases significantly—by more than one-quar- ter, from 6.8 million DALYs in 1990 to 8.8 million DALYs in 2000. Probably most of this increase is due to different methodologies used in the two studies. Critique of DALYs Despite the wide acceptance of the DALY approach as an important advance in the health policy arena, there have been many criticisms, some methodological, some more conceptual.31 The following is a summary of concerns expressed by other researchers. Methodological criticisms: • Many reproductive health complications were not ad- equately considered in the GBD DALY methodology. Approximately 100 major diseases and conditions that contribute to the vast majority of the health burden were studied, but many, less prevalent diseases were simply ignored and subsumed into the more important ones. For example, only three STIs were studied (syphilis, gonorrhea and chlamydia). • Death (or disability) was always assigned to just one cause, but the interaction between two or more conditions is often encountered in reality. For instance, indirect obstetric complications (e.g., malaria or ane- mia), gynecological morbidity (e.g., herpes or vagi- nosis), female genital cutting, rape and sexual abuse, puerperal psychosis, infertility and stillbirths did not get recorded as DALYs caused by reproductive risks, but rather as DALYs caused by malaria, anemia, etc. Thus, many DALYs were ascribed to other causes when the underlying contributing cause was related to reproduction.32 • To measure disability, several methodological ap- proximations had to be made given the paucity of data. One of these was the use of a panel of experts to deter- mine the severity rating of functional disability due to specific diseases. Severity weights for seven classes of disease were developed, and all diseases and sequelae were categorized into these seven classes. Critics have argued that people themselves should have rated sever- ity because there are many social, cultural and eco- nomic discomforts associated with particular condi- tions that may make the real severity greater or lesser than the mere functional aspect of the disease or con- dition.33 It has been suggested that this may be partic- ularly true in the case of reproductive health condi- tions.34 The example of fistulas has often cited, because the social stigma may be far worse than the functional disability itself.35 • In determining YLLs, the GBD study very high life expectancies as the standard—the highest ever ob- served (Japanese females). Thus, in countries with low life expectancies, the number of years lost by prema- ture death was estimated to be substantially higher than it would have been if the actual life expectancy of those countries were used as the standard. The criticism has therefore been made that the choice of a high standard life expectancy means that the DALY measures not only disease burden, but also the burden of “underde- velopment.”36 A counter argument to this criticism would be, however, that if most YLLs were eliminated by health interventions the underlying life expectancy would probably rise to a high level anyway. • The GBD methodology also assumed a 2.5-year difference between male and female standard life ex- pectancies. In fact, the difference in high-income pop- ulations is substantially more than 2.5 years. It has been argued, therefore, that this produces DALY estimates that are biased in favor of men.37 • Further criticism of the DALY methodology has questioned the study’s choice of an age weighting *Note that the definitions of “unsafe sex” vary from the 1998 GBD pub- lication to the WHO 2002 publication. For the WHO publication, unsafe sex is a risk factor for the following three conditions: HIV, STIs and cervix uteri cancer. In the 1998 GBD study, 90% of cervical cancer was attrib- uted to Human Papillomavirus. *The component “complications in pregnancy” (Berkley, 1998) refers to women who expressed an unmet need for contraceptives, but were not contracepting although sexually active. All DALYs caused by conditions related to pregnancy are multiplied by the proportion of women with unmet need, calculated from existing survey data, to produce estimates of DALYs due to unwanted or unplanned pregnancy. The Alan Guttmacher Institute 12 scheme and a discount rate.38 Basically, years lost by young children and adults older than 55 were given rel- atively less weight in the GBD study than those of other adults, because economic productivity and human cap- ital investment should be reflected in the cumulative ef- fect of disease. Also, a 3% discount rate was used. Re- search has indicated, however, that the way people discount the future value of human life is considerably more complex.39 Conceptual criticisms: • A widespread criticism of the DALY measurement system is that distributional and equity concerns were not built into the measure but should have been. DALYs demonstrate the level of disease burden but do not indicate how the burden is shared among different economic groups within a country. For example, the difference between life expectancies of the upper and lower income groups in the United States is 15 years, but DALYs do not reflect this.40 It has also been noted that people with higher incomes have 2–4 times the ac- cess to healthcare as do the poor, so estimates of costs of interventions should take this into account by look- ing at marginal costs, which might be substantially higher than average costs. The authors of the GBD agreed with the importance of equity and distribution- al issues, but wonder if it is perhaps better to have two measures and keep the two issues separate.41 a) The 1999 GBD results were also used by WHO to produce “league tables” that ranked the overall per- formance of national health systems. A country’s rank- ing was determined by taking each measure of attain- ment and performance—disability-adjusted life expectancy, health equality in terms of child survival, responsiveness level, responsiveness distribution, fair- ness of financial contribution, performance on the level of health and overall health system performance—and assigning a ranking.42 This approach has been criti- cized43 by some who have said that there is nothing to be gained from the GBD efforts to quantify the burden of disease and the resultant “league tables”. What is im- portant is to focus on the most cost-effective methods in each health system, whatever the position of a coun- try’s health system in the league table.44* This criticism bears on how the World Bank’s essential package was fashioned: The interventions in the package were se- lected on the basis of cost-effectiveness and magni- tudes of disease burdens. • Self-critique of methodology. The authors of the GBD study recognized that cause-of-death estimates mixed together data of very different qualities and degrees of completeness: “Substantial uncertainty will remain for many years about the precise distribution of mortality by cause for most of the developing world.”45 The ben- efits of having a complete and disaggregated set of cause-specific mortality estimates from the most com- plete compilation of available information were thought to outweigh these shortcomings. With respect to disability estimates, the authors rec- ognized that the uncertainties of data and conceptual- ization were much greater. The justification in this case, besides repeating the great need for comparable estimates, stressed the methodological refinements and advances that will be stimulated by the GBD’s enor- mous efforts to compile and standardize such a large dataset on disabilities. They concluded that “research is required to improve the basic disease model used in this study; furthermore, extensive empirical work is necessary to create and field-test new instruments for collecting data and information on disability.”46 Disease Control Priorities in Developing Countries (DCP-1) This large review of the cost-effectiveness of health in- terventions was a project of the World Bank and served as a major source of information for the 1993 World Development Report. DCP-2, a new project that will be a complete revision of DCP-1 is currently underway and is scheduled to be published in 2005. The three chapters from DCP-1 on excess fertility, maternal and perinatal health, and STIs and HIV/AIDS are of par- ticular interest to this report, and we summarize each here. Note that the methodology of each chapter is different. Disease Control Priorities: Chapter on Excess Fertility As opposed to the rest of DCP-1, this chapter did not focus on any one disease, but rather on “excess fertili- ty,” a condition that has direct negative consequences *A simple example can illustrate how looking at the magnitude of disease burden, instead of cost-effectiveness, could distort health policy. Suppose there are only two illnesses, A and B, for a health system to confront. Ill- ness A contributes 2,000 DALYs to the overall disease burden and an ef- fective intervention to prevent/treat it costs $20/DALY. Illness B con- tributes 200 DALYs and an effective intervention to counteract it costs $10/DALY. Focusing only on the relative sizes of the burdens —and given a budget of $5,000—one policy option might be to allocate $4,500 to A and $500 to B (since illness A causes more than 90% of all DALYs). This policy would save 275 DALYs (225 + 50). Basing policy purely on cost-effective- ness, however, would save 350 DALYs (150 + 200) with the same budget. Assessing Costs and Benefits 13 for infants and children as well as for maternal health.47 More significantly, however, excess fertility negative- ly impacts the health and social and economic well- being of families, community and society. The first part of this chapter of DCP-1 estimated total excess fertility in developing countries.* Broadly, “ex- cess fertility” can be equated with unplanned pregnan- cies (i.e, those not wanted at all or not wanted at the time they occurred). Three different estimations of excess fertility were made, which range from 12.9 million to 39 million births per year. The third method, using three “model” countries, yielded estimates of excess fertility in the range of 14–22% for women wanting to limit family size and 26–39% for those wanting to limit plus those wanting to increase spacing between births. The study looked at both the health gains from re- ducing excess fertility and at some specific social-sec- tor benefits. The health benefits included the survival of offspring, measured as the deaths and DALYs avert- ed by eliminating excess fertility. The social benefits examined by the study were reductions in educational and health expenditures from births averted. The study included a detailed discussion of several other indirect benefits that would flow from the reduc- tion or elimination of excess fertility that were not in- cluded in the study (and for which data for quantitative estimation may not be available). “High fertility and close child spacing are a significant determinant of poor health of mothers and infants in the first week of life…. [They] also have consequences beyond the first week of life, at least up to age five, and have negative consequences beyond those immediate health conse- quences….on the health and economic and social well- being of the family by diluting resources available for each child and putting pressure on parents to work harder and save less….[It] may also have negative con- sequences to society as a whole.”48 The negative consequences of excess or unwanted births, which would be mitigated by reducing or elim- inating this excess fertility, include: • lower rate of economic development (via reduced savings and investment, less technological change and changes in efficiency); • greater resource depletion and pollution; • in households, additional costs of food, clothing, medical care, schooling and housing; • additional time spent caring for children but less time for each child; • reduced expenditure per child leading to poorer health and reduced school participation; • additional efforts to increase family income which may lead to child labor, added labor of parents or re- duced household savings; if a child is “unwanted,” the negative effects are probably even larger (impaired child development, infanticide, abandonment, neglect, less antenatal care, selective nutrition and medical care have been reported); • if women are unmarried, having an unwanted or mistimed birth or pregnancy may result in less educa- tional and employment benefits, and increased chances of abortion or fostering out; and • greater societal burden, including higher expendi- tures on education, health, food subsidies, shelter and safe water. Total global estimates of benefits In the DCP-1 study, health benefits were estimated in terms of deaths and DALYs to both mothers and children.† The study’s analysis of the socioeconomic benefits focused on savings of public expenditures in primary education, secondary education and health. Three hy- pothetical countries with differing regional and mor- tality characteristics were presented for analytical pur- poses: Libana, high-mortality African, Banglapal, high-mortality non-African and Colexico, low-mortal- ity. Using a 5% discount rate, benefits were estimated as shown in Table 2.5. The basis for these estimations was not discussed at length, but footnote 3 of the study referred to World Bank “internal documents” used for estimating educa- tional costs. Other World Bank sources provided per capita public health expenditure estimates ($6 for the high-mortality countries and $28 for the low-mortality country).49 Estimates of Costs The study estimated “cost per birth averted” for a num- ber of countries. A summary of the cost estimates is shown in Table 2.6. Note that the difference between the “low” and “high” estimates was not discussed in the chapter. *The study did not give a single definition of “excess fertility” but rather offered three different definitions: (1) a societal one —population growth above 2% can be considered harmful to economic development; (2) a medical one—too young, too old, too many or too close births increase mortality risks of women and their offspring; and (3) an individual one— self-reported excess fertility, from data either on actual fertility in ex- cess of desired fertility or on desires to stop or postpone future births. †The table in Cochrane and Sai (1993) giving the DALYs saved by reduc- ing excess fertility is difficult to interpret. Repeated inquiries were un- able to clarify the numbers found in the table. The Alan Guttmacher Institute 14 Combining the above analyses, the costs and bene- fits of a family planning program to reduce excess fer- tility, using a 5% discount rate, is summarized using 1987 U.S. dollars in Table 2.7. At a 10% discount rate, family planning programs would not be justified in Libana if the only benefits ac- cruing to family planning programs were government savings in education and health. Critique of methodology In general, the methodology of the study was not ex- plained in detail and used internal World Bank data and estimates that cannot be independently verified. In the study, the two parts of the cost-benefit analy- sis were disjointed because socioeconomic benefits are calculated for three hypothetical countries, while pro- gram costs were presented for 16 actual countries. The authors noted that “the conclusions apply only to the economic benefits of family planning, and the health benefits, which are substantial, as shown above, would be additional.”50 The report gave no reason why a fiscal year of 1987 is used in a 1993 analysis. For instance, it stated that “the costs as collected refer to 1980, but they have been inflated to 1987 in Table 16-14.”51 More detail is needed to be able to evaluate which costs were included in the delivery of family planning services and which were omitted. Were ancillary costs (e.g., information, education and counseling activities) included? Were economies of scale taken into account? The study did not discuss these issues. Despite these shortcomings, this study is important because, in the context of the large DCP-1 initiative, which is primarily devoted to investigating cost-effec- tiveness in terms of illness avoided (i.e., DALYs), this study focused on nonhealth-related benefits, while still dealing with the benefits of preventing illnesses and health conditions. Disease Control Priorities: Chapter on Maternal and Perinatal Health This chapter of the DCP-1 volume attempts to relate the costs of providing maternal and perinatal care to several specific health benefits.52 The health benefits are reductions in maternal and perinatal mortality, in maternal morbidity and in the incidence of low-birth- weight babies. The study is divided into two parts. The first consists of an introductory section discussing the risk factors that lead to increased maternal and perina- tal mortality, detailed estimates of the extent of such mortality, and risk reduction strategies, including spe- cific interventions at each stage (preconception, preg- nancy, delivery and neonatal). The second part of the study presents a brief cost-effectiveness exercise using a hypothetical country called Himort. Another case is presented—“Lomort”—but no estimates of benefits are provided. The Himort cost-effectiveness analysis does not present results in terms of DALYs, but rather in terms of births averted, maternal deaths averted, maternal morbidity averted, perinatal infant deaths averted and low-birth-weight babies averted. For Himort—a ficti- tious country with a population of one million, a con- traceptive prevalence rate of 0%, and a maternal mor- tality ratio of 1,000—five scenarios for investments in women’s health are presented. Three of these scenarios concentrate on one intervention alone, namely, family planning. The fourth and fifth scenarios combine a low increase in family planning (equal to the first scenario) with different levels of obstetric care. The expenditure items that make up the two levels of obstetric care are described in detail in the study. These are summarized in Table 2.8. With regard to the three family planning scenarios, a program in South Korea, which resulted in a “20% in- crease in women who accepted contraceptives”53 and cost $0.47 per capita, was used as the basis for the first scenario. No details were given, for the other two fam- ily planning scenarios. The results of the cost-effectiveness exercise are summarized in Table 2.9. Cost per death averted refers to both maternal and perinatal deaths. Cost per event averted refers to deaths, morbidity cases and low-birth-weight babies. One im- plication of these findings is that family planning in- terventions become increasingly costly as the level of contraceptive use rises. Another conclusion seems to be that a moderate family planning program and a pro- gram of moderate obstetric improvement would be about equally cost-effective. The authors note that up to the time of this study (the early 1990s) there had been little empirical research on the outcomes of maternal health interventions. “We have found virtually no data on the effect of maternal health programs on maternal health…. The declines in adverse outcomes that we have suggested are no more than best estimates of the likely effect … based on the limited ev- idence available from the literature.”54 However, a recent study investigates the relationship between maternal health programs and maternal mortality, finding a sig- nificant inverse relationship.55 The study compares pro- grams across countries and uses a rating system devised by the authors and based on judgments of experts. Assessing Costs and Benefits 15 Disease Control Priorities: Chapter on HIV and STIs This DCP-1 study is a detailed analysis of the preva- lence of HIV and STIs, their effects in terms of DALYs and cost information on corresponding interventions.56 As discussion concerning effective prevention and treatment strategies, particularly with regard to HIV/AIDS, has advanced since this article was pub- lished in 1993, only findings regarding cost-effective- ness will be summarized here. • STIs and information, education and communication (IEC) programs. This section of the study summarizes information about cost-effectiveness according to type of STI, core or noncore group, and cost per year of pro- tection of the program. The results point toward focus- ing IEC efforts on core groups. For instance, in the case of syphilis, an increase of one person-year of protec- tion to the core group leads to averting 385 DALYs, but only 10 DALYs if the protection focuses on the non- core group. Similarly, the cost per year of protection for the noncore group is typically 4–7 times more expen- sive as for the core group. Again in the case of syphilis, the cost per year of protection for core groups ranges from $0.13 to $1.17, but for noncore groups the range is $0.52–4.64.57 • Mother-to-infant transmission. Gonococcal oph- thalmia neonatorum infections can be prevented. One estimate in the study is that one DALY may be averted for the cost of a $6 silver nitrate treatment. Congenital syphilis can also be prevented, but no costs are sug- gested in the study.58 • Infection through blood transfusion. Estimates are presented for cost per discounted DALY* by blood screening. Again, the results are far more cost-effective for core groups than for noncore groups. The cost of a blood test which ranges from $2–10, is also a factor, as is the HIV prevalence in the population. If the HIV prevalence rate is 5%, for example, the cost per dis- counted DALY ranges from $0.74 to $3.71.59 • STI treatment. Estimates are presented according to the following variable factors: presence or absence of an HIV epidemic, cost per clinic hour ($2 to $30), type of STI, prevalence of the STI (1–25%), and core or noncore group. The range of costs per DALY saved is $0.02 (treatment of syphilis at 25% prevalence, core group, $2 per clinic hour, HIV epidemic present) to $2,460 (treatment of chancroid at 1% prevalence, non- core group, $30 per clinic hour, HIV epidemic absent). In general, treating core groups is the cost-effective op- tion, as is treatment in the presence of an HIV epidem- ic, because STI infections increase the risk of HIV transmission, an interaction that leads to program syn- ergy. High STI prevalence also increases cost efficien- cy. Treatments for syphilis and chlamydia (male) are most cost-effective, treatments for chancroid, chlamy- dia (female) and gonorrhea (female) are least cost- effective.60 World Development Report 1993: Investing in Health The World Bank report elaborates three broad policy recommendations for improving health: 1. Foster an environment that enables households to improve health by pursuing economic policies that benefit the poor, investing in education and promoting women’s empowerment. 2. Improve public spending on health by reducing spending on tertiary facilities and cost-inefficient in- terventions, implementing a package of interventions aimed at health “externalities,” and improving man- agement. 3. Promote competition by encouraging health insur- ance schemes, competition among suppliers and infor- mation dissemination.61 For the purposes of this report, we summarize those parts of the 1993 publication that deal with the devel- opment of a cost-effective package of interventions. The key inputs in this regard are (1) estimates of the magnitudes of various diseases and conditions taken from initial findings of the GBD and (2) estimates of the cost-effectiveness of interventions that prevent or treat specific diseases and conditions. The latter esti- mates were derived from the DCP-1 report and internal World Bank documents. The 1993 report does not attempt to systematically present cost-effectiveness estimates for all health in- terventions. It states: “Only a small share of the thou- sands of known medical procedures has been analyzed, but the approximately fifty studied would be able to deal with more than half the world’s disease burden.”62 Regarding its estimates of the costs of interventions, the report makes these clarifications: 1) Costs are based on actual conditions; 2) some fixed costs of health systems are omitted because they are general, but costs related to interven- tion-specific capacity are included; 3) costs are assessed at market prices; 4) indirect costs, being difficult to valuate, are “largely ignored”; 5) the unit of study is sometimes a “package” rather than individual interventions; The Alan Guttmacher Institute 16 6) data on real-life outcomes are used, taking into account actual levels of coverage and compliance; and 7) a 3% discount rate is used to evaluate future gains.63 Although the report does not systematically identify the 47 analyzed interventions, it does quote estimates of the cost-effectiveness of selected interventions.64 These are listed in Table 2.10. The “package of public health and essential clinical services” for low-income countries has an average cost-effectiveness of around $97 per DALY. For mid- dle-income countries this rises to around $580 per DALY.65 The public health part of the package66 and the selected clinical services part of the package67 are shown, along with cost-effectiveness estimates, in Tables 2.11 and 2.12. Additional estimates are available; they are based on country-specific work and thus show a wide range of costs.68 These costs were drawn from the DCP report, discussed earlier in this chapter, which also provided average costs for a wider range of preventive and treat- ment interventions. The report estimated that prenatal and delivery care costs $30–250 per DALY saved and prevention of breast and cervical cancer costs $50–100 per DALY saved. By comparison, the cost per DALY saved by other health interventions ranged from $5–20 for preventing deficiencies in iron, vitamin A or iodine, to $5–250 for prevention of malaria to $1,600–3,500 for environmental control of dengue.69 Treatment gen- erally costs much more than prevention—for example, prevention of cervical cancer costs $100 per DALY saved but treatment costs $2,500 per DALY saved. Pre- venting cardiovascular conditions costs $150 per DALY saved, treatment $2,000–30,000.70 While both prevention and treatment are necessary, in Sub-Saha- ran Africa, interventions to prevent HIV are at least 28 times as cost-effective as antiretroviral therapy.71 Note that cost-effectiveness estimates for STI treat- ment assume the presence of an AIDS epidemic simi- lar to the actual situation in Sub-Saharan Africa. “Lim- ited care includes assessment, advice, alleviation of pain, treatment of infection and minor trauma, and treatment of more complicated conditions as resources permit.”72 It is worth noting that no methodological discussion is provided that would explain the develop- ment of these estimates of cost-effectiveness and the source cited is generally “World Bank calculations.” One critique of the essential package approach73 lists the following methodological shortcomings: (1) the package’s focus on average rather than marginal costs may bias the results; (2) the package focuses on potential rather than actual costs — although the report says that it does focus on real costs; (3) the approach is biased against the introduction of new technologies which are typically cost-inefficient at first; and (4) it fo- cuses on public expenditure, not total costs. The cri- tique also details several practical drawbacks to the package as a policy instrument which, while important, need not be mentioned here. Macroeconomics and Health This major study is the main report of the WHO Com- mission on Macroeconomics and Health, chaired by Professor Jeffrey Sachs.74 The report extends the World Bank’s World Development Report 1993 approach, ad- vocating a compact health agenda similar to the “es- sential package,” but with more emphasis on HIV (in- cluding an expensive component for treatment), malaria and nutrition. According to the analysis in this report, with a substantial increase in development as- sistance focusing on this health agenda, great progress would be achieved in lessening the burden of disease and reaping macroeconomic benefits. It is noteworthy that high fertility gets particular men- tion in the report even though family planning is neither included as a priority health intervention nor subjected to macroeconomic analysis to gauge the extent of its economic benefits. The following quotes, nevertheless, make the importance attached to family planning clear: “One of the most important health interventions is greater attention to reproductive health, not only to con- trol STIs such as HIV, but also to limit fertility through family planning, including access to contraception.”75 “Although we did not ourselves make cost estimates of the increasing need for family planning services and an adequate supply of contraception,” a funding gap exists “though it represents only a modest proportion of total funding needs.”76 “If more individuals are saved through health inter- ventions, for what kind of life are they being saved? The answer, fortunately, is an optimistic one. Health in- terventions…will contribute to slower, not faster, pop- ulation growth, but for this to occur it is important to combine health interventions with intensified efforts to offer family planning services and increased access to contraception.”77 This can cut the time of the demo- graphic transition—the report cites the examples of Bangladesh, Tamil Nadu and Andhra Pradesh. According to the framework adopted in the report, Assessing Costs and Benefits 17 there are three main ways that disease impedes eco- nomic well-being: 1. Healthy years lost because of disease cause direct economic loss, “a significant percentage of the nation- al incomes of the low-income countries.”78 2. Parental investments in children are lower in high-mortality settings. High infant mortality leads to high fertility. This, in turn, leads to less investment in the health and education of children (the quality-quan- tity trade-off). 3. Generally, there is a depressing effect on returns to business and infrastructure investment. “Whole in- dustries…are undermined by a high prevalence of dis- ease.”79 This framework is succinctly summarized in the re- port as follows: “The cost-of-illness literature probably dramatically understates the costs of nonfatal chronic conditions at all stages of the life cycle. Healthier workers are physically and mentally more energetic and robust, more productive, and earn higher wages. Their productivity makes companies more profitable, and a healthy workforce is important when attracting foreign direct investment.”80 The study concentrates on the poorest billion of the world’s population and sets the following health agen- da of priority interventions:81 • HIV prevention and treatment services;82 • malaria; • tuberculosis; • maternal and perinatal health;83 • causes of child mortality such as measles; • malnutrition; • other vaccine-preventable illness; • tobacco-related disease. The basis for selecting this set of priorities is the GBD initiative, which described the extent of the bur- den caused by specific diseases and conditions and the World Bank’s World Development Report 1993, which estimated the cost-effectiveness of specific health in- terventions. Both the extent of burdens and the degree of cost-effectiveness were compared using a common currency, the DALY. The macroeconomic part of the report attempts to quantify the economic gains from implementing the proposed health agenda. The approach taken does not attempt precise estimates of such gains but rather rough estimates based on a few broad assumptions and pa- rameters. First, each “life year” saved by implementa- tion of the health agenda is valued at 1–3 times the av- erage annual earnings. For the purpose of estimation, the conservative estimate is used: One “life year” equals one year of average earnings. Second, the effect on the economy is viewed in terms of the total size of the economy—i.e., (Gross Na- tional Product). The report uses the example of malar- ia in Sub-Saharan Africa to illustrate the macroeco- nomic gains possible through good health. In 1999, malaria accounted for an estimated 36 million DALYs. Valuing each DALY at the average per capita income for the region, an immediate economic effect is a loss of 5.8% of total GNP (36/616 = 5.8%).* Third, with regard to per capita effects, the report cites econometric estimates that in economies where populations experience “high malaria risk,” economic growth is about one percentage point less than other- wise.84 The cumulative effect is an eventual per capita income only half of what it would be in a zero risk en- vironment. Combining the total and per capita income effects, “dozens of percent of GNP” are lost to malar- ia, according to the report. In the report, the phrase “scaling up” refers to the additional investments needed for the health agenda advocated by the CMH to be implemented. A rough cost-benefit analysis for low-income countries is at- tempted. On the one hand, the additional cost of scal- ing up is put at $66 billion in 2015. On the other hand, the economic benefits† are calculated at $186 billion in 2015 (330 million DALYs x $563 = $186 billion). This represents the total GNP benefit, yielding a cost-bene- fit ratio of about three to one. However, improved health would also spur eco- nomic growth, known as the per capita income benefit. Faster growth, as mentioned, would occur due to a faster demographic transition, higher human capital growth, increased household savings, increased foreign investment and greater social and macroeconomic sta- bility. At lower bound this can be estimated at an added $180 billion per year by 2020 (based on assuming an extra one-half percentage point of growth each year). Combining this benefit with the one based on total GNP, the economic benefits would grow to $360 bil- lion by 2015–2020 “and possibly much larger,”85 re- sulting in a cost-benefit ratio of 6 to 1. At a microlevel, the report suggests—without pro- viding quantitative findings—that for individual house- holds a single disease episode can lead to asset deple- *Updating the analysis with 2001 data (42 million DALYs lost to malaria, 669 million population), the percent of total GNP lost becomes 42/669 = 6.3 percent. †Assuming 2 percent growth for the period 2000-2015, per capita income of $563 in 2015 for that region and one DALY being the equivalent of one year of per capita income. The Alan Guttmacher Institute 18 tion and, consequently, the household falling into a per- manent poverty trap. Brazilian data are quoted that show a clear link between income and educational at- tainment (including cognitive ability), using height as a proxy.86 Evidence also shows that good health and nu- trition are precursors to educational attainment, includ- ing both attendance and cognitive ability.87 With regard to business, high labor turnover lowers the profitability of companies, depresses tourism and may prevent eco- nomic use of land. The report cites the experience of businesses in high HIV prevalence areas that have to hire and train more than one person per position.88 The report advocates that AIDS treatment be given high priority in the health agenda, in addition to HIV preventive interventions. Estimating treatment costs at $500 to $1000 per year, the report suggests that around five million infected individuals could be treated by the end of 2006.89 • Comment. The CMH study sponsored a large num- ber of background studies that fed into the final report. The report, however, does not make close references to these studies which makes it difficult to trace the spe- cific assumptions behind various assertions. Therefore, many assertions in the report do not stand on their own. While it is likely that most findings are supported in these background papers, without specific references it is difficult to verify. The following are specific areas where the report could have usefully provided further details and clarification on the cost benefit analysis to make clear how rigorously it was done. First, the criteria for selection of the diseases and conditions for priority action are not clearly spelled out in the report. We are only told that the report focused on diseases with “the greatest excess mortality in the poor countries relative to the rich countries.”90 There is no appeal to cost-effectiveness made in the report. Second, the issue of what would happen to GNP per capita is not considered in the report. The report shows that total GNP would be bigger, but so would the pop- ulation because of reduced mortality. A better approach would be to separate the YLL component from the YLD component of DALYs. The YLL component ba- sically reduces both the numerator and the denomina- tor of GNP per capita (and so its overall effect on GNP per capita is unclear), while the YLD component, it could be argued, reduces only the numerator. Third, further explanation and discussion is needed, especially in terms of cost-efficiency, for the inclusion in the report of AIDS treatment interventions and sup- port for a large investment in AIDS treatment, over and above HIV prevention interventions. A number of stud- ies point out that prevention interventions are several times more cost-effective as treatment, although there is an ongoing debate on this point.91 As alluded to above, the report mentions the impor- tance of, but nonetheless excludes from its analysis, the positive economic effects of reduced fertility through family planning programs. Separating family planning in practice from other health investments, though, may be neither feasible nor advisable. In many sociopoliti- cal situations, family planning continues to be accept- ed largely within a health context not only as a ration- ale for policies but also as a conduit for services. From the perspective of health policy also, the CMH report has been questioned.92 The report advocates the influx of a large amount of mostly external donor funds. This could well distort the structure of the exist- ing health system and work against its long-term sus- tainability. Focusing on a narrow range of diseases and corresponding interventions could also lead to a duali- ty of the health system and the neglect of the ongoing delivery of care for diseases that are not in the priority agenda. A heavy reliance on donor funding could also lead to a bias toward vertical programming and com- modity procurement instead of capacity building. Current and Ongoing Research In addition to the major works summarized above, sev- eral current research activities will produce findings in the future that will be of interest to the topic being re- viewed here. The following lists briefly these ongoing efforts: • Disease Control Priorities Project— second phase (DCP-2). DCP-2 began in 2002 as a joint initiative of WHO, the World Bank and the National Institutes of Health (Fogarty International Center), with funding from the Bill and Melinda Gates Foundation. The DCP-2 study will be a complete revision of the work done in the original project (DCP-1) reviewed above. A limited number of working papers are now available, and the full report is expected to be published in 2005. • Global Burden of Disease. This is an ongoing proj- ect of WHO. A major updating of the methodology, particularly with regard to the estimation of YLLs is ongoing. More information can be found on the WHO Website: http://www3.who.int/whosis/ menu.cfm?path =evidence,burden. • Maternal-Newborn Health and Poverty. WHO has begun an initiative to provide “an actualized overview of current knowledge and experience regarding the re- lationship between maternal health and poverty”.93 The scope of this research effort will include costs to indi- Assessing Costs and Benefits 19 viduals and families, macrolevel costs, cost-effective- ness of interventions, and strategies for benefits of in- vestment in maternal health. Work is ongoing in 2004. • Millenium Development Goals—Task Force on Child and Maternal Health. A multi-agency task force, under the auspices of the United Nations, has been formed to develop a strategy for implementing the MDGs covering child and maternal health. A final re- port is expected by June 2005. The following are tar- gets and indicators of the task force:94 • Reduce child mortality: • Reduce by two-thirds, between 1990 and 2015, the under-five mortality rate • Indicators: Under five mortality rate, infant mor- tality rate and proportion of 1-year-old children immu- nized against measles • Improve maternal health. • Reduce by three-quarters, between 1990 and 2015, the maternal mortality ratio. • Indicators: Maternal mortality ratio, proportion of births attended by skilled health personnel • Improve maternal health. • Reduce by three-quarters, between 1990 and 2015, the maternal mortality ratio. • Indicators: Maternal mortality ratio, proportion of births attended by skilled health personnel Cairo 1994: The Cost of Reproductive Health The International Conference on Population and De- velopment (ICPD) held in 1994 adopted a 20 year Pro- gram of Action (1994–2015) in the areas of population and reproductive health, and developed estimates of re- source requirements for the success of this program.95 The program had four elements: family planning serv- ices, other reproductive health services, HIV/AIDS prevention and basic research in population. Resource requirements totaled (in 1993 U.S. dollars): $17.0 bil- lion in 2000; $18.5 billion in 2005; $20.5 billion in 2010; and $21.7 billion in 2015. Of this overall total, around 65% represents inputs to the service delivery system for reproductive health and family planning services. Table 2.13 summarizes the projected global resource requirements for the ICPD program. Family planning and expanded reproductive health For the estimation of the first component, family plan- ning, costs per user of contraception are estimated ac- cording to the level of contraceptive prevalence. Con- traceptive prevalence is projected in subregions separately and depends on levels of unmet demand for family planning services. Current contraceptive preva- lence rates and estimates of regional levels of unmet demand for family planning are used to project how prevalence will increase and further demand will be generated as access to reproductive services increases and becomes universal. The assumption is made that the historical pattern of reductions in annual unit costs per contraceptive user will continue due to economies of scale, improved technical performance and specialization of the rele- vant institutions (governmental, nongovernmental and private sector) in particular areas, thereby maximizing efficiency. Other underlying assumptions are that spe- cialized providers will also ensure universal access to services and that specialization will be the outcome of increased competition. The second component of the integrated program aims at further improving the quality of care and pro- viding family planning as part of a broader package of reproductive health services and referrals. It is esti- mated to cost an additional $1.03 per capita per year. Total resources needed amount to (in 1993 U.S. dol- lars): $5.0 billion in 2000, $5.4 billion in 2005, $5.7 billion in 2010 and $6.1 billion in 2015. Roughly 65% of this component consists of additional inputs to basic service delivery systems; the remaining 35% represents specialized inputs particular to reproductive health pro- grams.96 HIV/STI prevention A draft document prepared by the WHO Global Pro- gram on AIDS presents estimates of global resource re- quirements for HIV prevention in developing coun- tries.97 Three of the program’s seven components— mass media, school education and condom distribution —are elements of an integrated population program with service delivery based at the primary-health-care level. WHO estimates the annual resource require- ments for these components to be between $1.1 billion and $1.6 billion (between $0.26 and $0.43 per capita) depending on the chosen scenario for intensity of the HIV epidemic. The ICPD estimation adopts the lower per capita estimate. The lower per capita calculations adopted leads to the estimations of the ICPD Program of Action, namely that resources required annually for HIV/STI prevention programs will increase from $1.3 billion to $1.6 billion between 2000 and 2015. The Alan Guttmacher Institute 20 Additional population data, policy and analysis requirements Pre-ICPD estimated resource requirements, based on earlier time periods when the coverage of censuses was not as complete as it has become and before the devel- opment of the numerous additional needs for decen- tralized, regional and local population databases, were considered too low. Additional population and devel- opment policy analysis needs existed. One such need is the building up of national capacity for data collec- tion and analysis, research, policy development and training in demographic as well as program-relevant areas. Moreover, census costs, which are a sizeable component of population data and population and de- velopment research expenditures, are not directly re- lated to the number of users of contraception but de- pend on the size of the total population. The ICPD calculation conservatively assumes that demographic and other population and policy-relevant research activities will cost $1 per capita in many de- veloping countries, spread over a multiyear period. The proportion of these costs will be highest during years of census collection and analysis. Using these assump- tions, this component’s total cost is estimated at be- tween $260 million and $670 million per year, depend- ing on where the year is in the decennial census cycle. UNFPA Costing Initiative In 2002, UNFPA undertook an extensive review of costing information from empirical studies on repro- ductive health services.98 This work built on previous reviews in the same or allied areas.99 Over 500 pub- lished and unpublished reports were reviewed, mainly from 1990–2002. The reproductive components cov- ered included family planning, maternal health, postabortion care, STI prevention and treatment, HIV prevention, behavior change communication and other reproductive health components. In turn, each of these components was broken down into several subcompo- nents. The resulting body of information probably rep- resents the most current and complete synthesis of data on the costs of reproductive health interventions avail- able at one source. This resource currently resides on UNFPA’s intranet site. The site is accessible to all UNFPA offices and can be made available to researchers upon request. Sum- mary tables from the costing initiative are given in Ap- pendix 2.3. In developing this site, attention was given to systematizing the data and condensing it to increase its usability. To the extent possible, costs were broken down into subcosts including costs of drugs and sup- plies, staff costs, overhead costs and capital costs. At the subcomponent level, tables and summaries were made available that synthesize and standardize the data from all related studies, thus increasing the usability of the information. In terms of content, the most comprehensive costing data are available for family planning, followed by ma- ternal health and STIs. Almost no costing studies were found in the area of behavior change communication. The HIV/AIDS subcomponent is relatively less devel- oped, given the vast amount of information being col- lected by UNFPA’s sister agency, UNAIDS. The Effect of Family Planning Programs An ongoing debate contrasts the effect of family plan- ning programs on fertility versus the effect of socioe- conomic development. The question is to what extent publicly supplied contraceptive services merely sub- stitute for private consumption of contraceptives in populations where socioeconomic progress has made contraceptives affordable as well as making smaller families more desirable. Answers to this question have relied on cross-national comparisons using program ef- fort scores.100 Although it is found that in relatively well developed countries program effort seems to be more correlated with fertility decline, family planning has a strong and independent effect. For example, “Bangladesh, one of the world’s 20 poorest countries, has a program rated among the best in the developing world and has seen a substantial decline in fertility over the last decade.”101 The evidence from an intensive mi- crostudy in Matlab, Bangladesh102 confirms this inde- pendent effect. Although some skeptics have argued that the pro- gram effect is less than most research has indicated,103 a careful analysis subsequently put the reduction in fer- tility in developing countries due to family planning programs at more than 40 percent of the total from the 1960s to the end of the 1980s.104 The consensus view at present is that, while socioeconomic progress does lead to reduced fertility, a very substantial part of the reduction, independent of socioeconomic factors, is due to public efforts in family planning.105 UNAIDS Cost Estimates A recent UNAIDS report106 provides, inter alia, a global estimate of costs of interventions needed to con- front the HIV epidemic. Twenty-five key interventions “required to achieve the overall goals laid out in the De- Assessing Costs and Benefits 21 claration of Commitment on HIV/AIDS, which was signed by 189 countries at the United Nations in June 2001”107 are given cost estimates. The estimates cover 135 low- and middle-income countries for the 2001– 2007 and take into account the “maximum feasible coverage” as governed by existing physical infrastruc- ture and human resources.108 Globally, resource requirements rise from $3.2 bil- lion in 2001 to $10.5 billion in 2005 and $15 billion in 2007. Table 2.16 summarizes the distribution of need- ed resources by type of intervention over the period, antiretroviral treatment costs are estimated to increase from 14% to 25% of total cost. The report does not discuss methodology, but it ap- parently uses a straightforward estimation of costs of providing services calculated by multiplying popula- tion in need of coverage by unit costs.109 WHO Mother-Baby Package The Mother-Baby Package (MBP)110 is a model that can be used to assess and analyze the costs associated with implementing maternal health care interventions at the district level. Such interventions include: antena- tal care, delivery care, treatment of obstetric complica- tions (e.g., hemorrhage, sepsis, eclampsia, Caesarean section, and family planning. The model can be applied to locally collected data to estimate the actual cost of current services as well as the cost of upgrading the dis- trict health system to meet MBP standards. The model calculates total, per capita and per birth costs. Estimates are presented by intervention, by input (e.g., drugs, vac- cines, salaries and infrastructure) and by service loca- tion or level (hospital, health center and health post). The model estimates the cost of providing to a given target population a package of maternal and newborn interventions (Table 2.17). The information that needs to be collected can be grouped into three main categories: 1) demographic and epidemiologic information (such as population, birth-rate, incidence of pregnancy and delivery-related complications, contraceptive prevalence and mix); 2) Costs of inputs (drug prices, salaries of medical and support personnel, building, equipment and supply costs); and 3) Information about current treatment practice (if current treatment cost is to be assessed). In terms of the model’s outputs, the package pro- vides the user with a cost estimate for the implementa- tion of the whole MBP, but also breaks down the total cost in a variety of ways. For example, cost estimates are given by individual interventions, different types of input (drugs, salaries, etc.), different types of costs (re- current and capital costs), various rates (per facility, per capita and per birth) and local cost and imported cost (foreign exchange requirements). The MBP is basically a costing tool that helps health system managers organize costs in a comprehensive and coherent way so that recurrent costs, capital costs, over- heads and so on are all entered into the costing calculus. The package also provides for cost breakdowns by level of facility. The MBP spreadsheets are particularly use- ful in cases where costing data are deficient, because they contain a variety of default values for specific costs. An example of an application of the MBP in Ugan- da is illustrative. The Ugandan government wanted to implement a comprehensive safe motherhood program in an effort to reduce high levels of maternal and neona- tal morbidity and mortality in the country. The MBP was used to set standards regarding the scope and qual- ity of the health care provided to pregnant women and newborn babies. To provide program planners with a better appreciation of the costs entailed in implement- ing the MBP, a costing study was undertaken. In the two districts studied, it was found that the Ugandan govern- ment spent about $0.50 per capita on maternal and new- born health care. To upgrade this care to conform to MBP standards and guidelines would cost approxi- mately $1.40 per capita, representing an incremental cost of $0.90. The inclusion of capital and overhead costs would raise the cost to approximately $1.80 per capita, bringing the incremental cost up to $1.30. Cost-Savings Analyses by The Alan Guttmacher Institute In the 1990s, The Alan Guttmacher Institute (AGI) un- dertook a number of studies on the costs and benefits of publicly funded family planning services.111 These studies are of particular interest because the methodol- ogy employed serves as a starting point for the cost- benefit analysis of contraceptive services in the devel- oping world presented in Chapter 3 of this report. Forrest and Singh in 1990112 examined federal and state expenditures in family planning in the United States for fiscal year 1987. The study found that if the $412 million spent on contraceptive services had not been spent, other public programs (medical care, wel- fare and supplementary nutrition) would have expend- ed an additional $1.2–2.6 billion—on average $4.40 ex- pended for each dollar saved—to cover additional The Alan Guttmacher Institute 22 demand generated by the extra pregnancies and births by women who would be denied access to contracep- tives. The methodology used in the study consisted of four steps: (1) determining the number of women using publicly-funded contraception; (2) estimating the addi- tional unintended pregnancies and their consequences if public funding were cut using four different scenar- ios depicting possible behavior; (3) calculating the ad- ditional expenditures in social services as a result; and (4) comparing costs from step three to savings from step two. The four behavioral patterns referred to possible contraceptive use following cuts in public funding and were based on differing evidence of past behavior. A later study by Forrest and Samara in 1996113 re- estimated the earlier work using newly available con- traceptive use data to refine the methodology. An im- portant improvement was the availability of use data disaggregated by specific contraceptive method. Again, four scenarios were explored: (1) the use pattern of women affected by funding cuts would resemble that of nonsubsidized women; (2) it would resemble that of women who discontinue pill use; (3) the pattern would revert to the behavior prior to first clinic visit; and (4) no method at all would be used.* For 1988, an estimat- ed additional 1.3 million unplanned pregnancies would occur if funds were cut, resulting in 0.6 million induced abortions and 0.5 million unintended births. For every dollar saved through defunding, an average of $3.00 would need to be spent in Medicaid services. Another study by AGI in 2000114 examined the non- monetary costs of a potential reduction in U.S. funding for family planning aid to developing countries. The methodology of the study was similar to the U.S. stud- ies described above. Based on indirect evidence, fund- ing cuts, which would hit poor women already using subsidized public clinics, would change modern con- traceptive users into either users of traditional methods with high failure rates or nonusers. Unintended preg- nancies, unplanned births, induced abortions, and in- fant and maternal mortality were the physical costs es- timated in this cost-benefit exercise. A further study measured the benefits of an increase in family planning assistance from the U.S. Agency for International Development in terms of lives saved. The study estimated that a $169 million increase in family planning funding in 2001 would save the lives of 15,000 women (8,000 who would have died as a result of unsafe abortion and 7,000 who would have died from other pregnancy-related causes) as well as the lives of 92,000 infants.115 Cost-Benefit and Cost-Effectiveness Models in Reproductive Health A number of cost-benefit and cost-effectiveness mod- els have been developed to examine aspects of repro- ductive care. These are computer-based simulation models that incorporate certain assumptions and em- pirical relationships. Intended for health-policy devel- opment work, they generally require users to enter country or program-specific input information before generating a series of different output scenarios. This section lists these models and briefly describes a set of representative ones (Table 2.18). The objective is to give a comprehensive overview of recent methodolo- gies that are at least partially related to the determina- tion of cost-benefit in the area of sexual and reproduc- tive health. • GOALS Model. This model116 is associated with attempts to cost the global resources that would be needed to achieve the goals of the Declaration of Com- mitment on HIV/AIDS (June 2001).117 The model is very detailed, covering five care and treatment inter- ventions, 14 prevention interventions and seven other interventions. The basic question that planners can ad- dress with GOALS is what level of funding is required to achieve the goals of the national strategic plan for combating HIV/AIDS. An alternative question is how far the goals can be achieved with a given amount of re- sources. The effectiveness of different interventions can also be assessed by the model so that a budget can be devised that achieves an optimum allocation of re- sources. The costs in the model are the costs of a set of inter- ventions. The benefits are the number of potential HIV infections averted, future health expenditures averted and years of life gained as a result of the set of inter- ventions. One application of GOALS has been reported for Lesotho.118 Lesotho had a strategic plan and estimated a budget to fulfill the plan. The GOALS model showed that the three-year budget for HIV ($1 billion) was grossly overestimated. GOALS showed that $40.5 mil- lion per year would lead to a drop in prevalence from 35% to 30% over three years. • BenCost Model. The main purpose of this model119 is to conduct a public-sector cost-benefit analysis of family planning programs in order to eval- uate the financial savings to governments as a result of *This last scenario, no contraceptive use, was used for comparative pur- poses only and was not used to arrive at average estimates. Assessing Costs and Benefits 23 providing the same level of services to a smaller group of people. Public cost-benefit analysis looks at such public-sector services as primary, secondary and terti- ary education, health, food subsidies, social welfare, housing, utilities and infrastructure. BenCost is capa- ble of examining the financial savings generated in all or any subset of these services as a result of expendi- tures in a family planning program. Three main issues in cost-benefit analysis are: (1) what the proper enumeration of costs and benefits are, (2) how benefits are valued and (3) what discount rate is appropriate. BenCost avoids some of these difficul- ties by restricting its analysis to public financial sav- ings rather than trying to explore all possible benefits. The study notes that noneconomic benefits such as ma- ternal and child health are “in practice very difficult to measure. Even if they can be measured correctly, it is still difficult to convert them into monetary units so that they can be combined with other economic bene- fits.”120 • The Injecting Drug User Model. IDU model is one of five simulation models developed by the London School of Hygiene and Tropical Medicine (LSHTM) in a consortium led by UNAIDS to show the impacts of different HIV preventive interventions.121 The model can show the number of HIV infections averted by a particular intervention. Neither the IDU model nor any of the other models in the suite is designed to link ben- efits—infections averted—to intervention costs. • Costing the Essential Health Package Spreadsheet. This training-oriented, spreadsheet-based model is still under development by the World Bank.122 The data built into the model come from an amalgam of cost and output data from Zambia and Bangladesh, and thus represent a low-income setting.* The essential servic- es consist of family planning, HIV/AIDS prevention and management, antenatal care, nutrition, delivery care, postnatal care, reproductive tract infections and STIs control and management, immunization (EPI plus), management of childhood diseases, tuberculosis control, malaria control and curative care. The hypothetical country in the spreadsheet has a population of 10 million. The health system comprises community level facilities, dispensaries, clinics and district level hospitals. The model places specific in- terventions at each level of service, fixes the coverage for every intervention-service-level point and assigns costs in a similarly disaggregated fashion. The embed- ded data yield a total annual cost of $11.3 million or a per capita cost of $11.30. This amount approximates the estimated $12 per capita cost for low-income coun- tries for the essential care package proposed in the World Bank’s World Development Report 1993.123 Summary. Each of the models discussed covers one or more aspects of reproductive health. Several models cover both costs and benefits—sometimes in physical terms, such as numbers of infections averted, and sometimes in monetary terms. The LSHTM models only look at benefits. One thing all the models have in common, however, is that they are a rich source of de- tailed data on costs and parameters and linking inter- ventions to results. Selected Country-Level Cost-Benefit Studies of Family Planning Services Several cost-benefit analyses have been carried out on family planning programs at the country level. Policy decisions to invest in one type of intervention rather than another are commonly taken at the country level, and cost-benefit studies of national programs have often proved to be useful inputs to these decisions. Al- though such studies are primarily meant to influence national policies and their application outside the na- tional context must be approached cautiously, exami- nation of a representative set of them is an important means of demonstrating the methodologies involved and how they might be adapted to other countries. This section reviews a few of the most significant national cost-benefit studies with these objectives in mind. Mexico 1972–1984 This study124 evaluated whether the costs of the Mexi- can Social Security System (IMSS) family planning services yielded a net savings to IMSS by reducing the load in its maternal and child health services. Based on the average cost per case, the analysis disclosed that for every peso that IMSS spent in family planning servic- es to its urban population between 1972 and 1984, the agency saved nine pesos. The analysis focused on the urban clientele of the IMSS system, some 5.3 million women. The evaluation looked at the following: • Cost for contraceptive recruitment and supply; • Annual number of births averted; • Estimate of annual number of IMSS treatments for incomplete abortions; and • IMSS expenditures per pregnant and postpartum *Personal communication from Tom Merrick, reproductive health ad- viser, the World Bank. The Alan Guttmacher Institute 24 woman, per incomplete abortion treated, and per child cared for during the first year of life. Of the 539 billion pesos spent in 1984, 70 billion went to family planning and maternal and infant care. Had the agency not instituted family planning services in 1972, an additional 51 billion pesos would have been expended on maternal and infant care in 1984. As a con- sequence of its family planning program, IMSS has been able to divert a total of 318 billion pesos (1983 pesos) during 1972–1985 from maternal and infant care to payments for pensions and general health services. • Comment. This study looks at a very specific sub- set of all the possible benefits flowing from family planning programs, namely, savings in maternal and child health services due to avoidance of unwanted or unplanned births. Presumably, the total benefits would be far larger, but by concentrating on maternal and child health services covered within the IMSS insur- ance system itself, the study avoided the need to ascribe monetary amounts to non-valued benefits. The study is notable, too, for the completeness of the data used and the large size of the population analyzed. Thailand 1970–2010 This study125 found that the average return on each dol- lar invested in Thailand’s family planning program is more than $7 for the first nine years of the program’s existence (1970–1980) and more than $16 over a 40 year period (1972–2010). The cost-benefit estimates are derived from a pro- jection over 40 years of government expenditures for family planning and government expenditures averted as a result of births averted by the family planning pro- gram. It is projected that between 1972 and 2010, 16.1 million births will be averted by the program; this num- ber of births averted will be achieved at a total estimat- ed cost of $536 million in family planning expendi- tures, but will in turn yield estimated cumulative savings of $11.8 billion in government expenditures in social services. Using a discount rate of 13.5% the total cost of the program becomes $68 million, and the total benefit generated becomes $1.1 billion. The savings for specific years of the project period are shown in Table 2.19. The effect of the program is to lower spending on social services by 13% in 1980, by 25% in 1990, by 17% in 2000 and by 23% at the end of the projection period. Egypt 1992–2015 This study126 combined a cost-savings analysis of the public sector with an econometric model of the Egypt- ian economy. Cost-benefit analysis. The usual cost-benefit analy- sis approach is used in this study to compare the costs of the family planning state program in Egypt to savings from births averted in other sectors (e.g., food subsidies, education, water, sewage, housing and health). A table summarizing the findings shows an overall cost-benefit ratio of 31 to one. In other words, one dollar invested in the family planning program saves the government from spending $31 in the six other sectors analyzed. This ratio is higher than any found in similar studies in other countries. The relative contributions to the accu- mulated public savings from the six sectors considered are as follows: education (31%); food (7%); health (7%); housing (21%); water (19%); and sewage (15%). The projection period used in the study was 1992–2015, and the discount rate used was 0%. How- ever, even if a 15% discount rate were used, the cost- benefit ratio would be reduced only from 31 to one to 25 to one. The study includes a good comparison of these results with several other similar studies and an- alyzes the reasons that the greatest cost-benefit ratios were found in Egypt. The greater number of sectors and the 0% discount rate are two simple reasons for the high cost-benefit ratio in Egypt, but sector coverage and relative sector costs also contribute to the result. Macroeconomic model.A model of the same sectors was also developed in this study. The model is a Com- putable General Equilibrium model based on a social accounting matrix. The model was run for the period 1992–2010. At the end of the period, Gross Domestic Product increased by 4.5% percent under the lower fer- tility simulation, GDP per capita increased by 8% and investment increased by 20%.127 Vietnam 1979–2010 This study128 analyzed Vietnam’s national population and family planning program using data from 1979 to 1996 and projecting costs and benefits through 2010. The study compared the impact of a strongly funded family planning program with a weak program with lit- tle government support. In 1979, the combined total family planning expen- ditures by the ministry of health and donor agencies in Vietnam was Vietnam Dong (VND) 47 billion (in con- stant 1995 VND). By 1989, national program expendi- Assessing Costs and Benefits 25 tures had risen to VND 79 billion and reached VND 406 billion by 1996. The benefits analyzed were savings in the provision of health, education and other social services due to prevented births. Future costs and benefits were esti- mated using a 10% discount rate. The study showed that for every one VND invested at the national level in population and family planning, by 2010 VND 7.6 in social sector spending would be saved. However, the initial start-up phase of the pro- gram did not show immediate net benefits. For the years 1979 through 1995, the benefit-cost ratio was below the break-even level of 1. Only after 1995 did the annual ratio rise above this threshold. Cumulating the benefits over the 31-year study pe- riod, it was found that 90% of total social sector sav- ings resulting from averted births and slower popula- tion growth will accrue in the education sector. Nearly 5% of social sector savings will be from lower mater- nal child health expenditures, another 3% from popu- lation-based services and 2% from avoided social se- curity expenditures. India 1956–1987 This study129 evaluated the family welfare program of the Tata Iron and Steel Company in Jamshedpur, India. The benefit-cost ratio was calculated by reviewing the total program of benefits provided to employees by TISCO, isolating those benefits which could be im- pacted by fertility, computing the financial impact of one birth and then calculating the total financial impact of the birth averted. TISCO provides a comprehensive service package to its employee and their families, in- cluding maternity care, inpatient hospitalizations for dependents, outpatient hospitalization for dependents and education for dependent children. The study found that for each rupee spent on family planning, 2.39 rupees were returned in cost savings over the life of the program. For 1987–1988 each rupee expended resulted in 3.50 rupees saved, showing an in- creasing trend of the benefit-cost ratio. Overall, by 1987, 43,872 births had been averted since the program began in 1956. The Alan Guttmacher Institute 26 Table 2.1 Burden of disease related to reproductive health YLLs (m.) YLDs (m.) DALYs (m.) % of all DALYs 1990 2001 1990 2001 1990 2001 1990 2001 STIs 6.5 5.4 12.0 7.0 18.5 12.4 1.3% 0.85% HIV 8.8 80.0 2.3 8.4 11.2 88.0 0.8% 6.0% Maternal conditions 13.3 15.0 16.5 15.9 29.8 30.8 2.2% 2.1% Total 28.6 100.4 30.8 31.3 59.5 131.3 4.3% 9.0% Table 2.2. Percentage of DALYs, YLLs and YLDs lost by cause, according to region*, 2001 World Africa Americas Europe Eastern Mediteranean South East Asia Western Pacific Region DALYs STIs excluding HIV/AIDS 0.8% 1.4% 0.4% 0.2% 1.0% 1.0% 0.2% HIV/AIDS 6.0% 18.8% 1.9% 0.6% 1.3% 3.2% 0.8% Maternal conditions 2.1% 3.2% 1.3% 0.5% 3.0% 2.4% 1.1% Perinatal conditions 6.7% 6.1% 4.9% 1.9% 9.1% 9.4% 5.7% Other sexual and reproductive health conditions 2.7% 1.7% 3.5% 3.7% 2.5% 2.8% 3.0% Percent 18.4% 31.3% 12.0% 6.9% 16.9% 18.9% 10.8% Total DALYs (000’s) 270 112 17 10 23 79 28 YLLs STIs excluding HIV/AIDS 0.6% 1.2% 0.1% 0.0% 0.7% 0.5% 0.1% HIV/AIDS 8.7% 22.2% 3.2% 0.8% 1.8% 4.6% 1.1% Maternal conditions 1.6% 2.5% 1.0% 0.1% 2.3% 1.7% 0.5% Perinatal conditions 9.0% 6.9% 8.1% 2.8% 11.7% 12.8% 8.5% Other sexual and reproductive health conditions 2.5% 1.3% 4.4% 4.2% 2.1% 2.6% 3.0% Percent 22.4% 34.1% 16.9% 8.0% 18.5% 22.2% 13.1% Total YLLs (000’s) 207 95 11 7 17 60 18 YLDs STIs excluding HIV/AIDS 1.3% 2.4% 0.7% 0.5% 1.6% 1.9% 0.4% HIV/AIDS 1.6% 7.3% 0.7% 0.4% 0.2% 0.8% 0.4% Maternal conditions 2.9% 5.3% 1.5% 1.0% 4.5% 3.7% 1.9% Perinatal conditions 2.7% 3.4% 2.0% 0.8% 4.2% 3.5% 2.4% Other sexual and reproductive health conditions 3.1% 3.2% 2.6% 3.0% 3.3% 3.2% 3.0% Percent 11.6% 21.6% 7.7% 5.7% 13.9% 13.0% 8.1% Total YLDs 63 17 6 4 6 20 10 Assessing Costs and Benefits 27 Table 2.3 HIV/AIDS has become a major cause of sexual and reproductive ill-health worldwide. 1990 2001 HIV/AIDS 11.2 88.0 STIs, excluding HIV/AIDS 18.5 12.4 Maternal conditions 29.8 30.8 Perinatal conditions 92.3 98.3 Other sexual and reproductive health conditions 42.5 23.0 Total 194.3 252.5 *See Appendix Table 2.3 for a list of countries included in each region. Table 2.4 Deaths and DALYs in developing countries due to unsafe sex and lack of contraception, 2000 Risk Factor Deaths (000s) DALYs (millions) Unsafe sex 2,830 90.0 Lack of contraception 149 8.7 Total 2,978 99.0 % of burden borne by women 55% 58% Table 2.5 Savings per birth averted (US$ 1987) Libana Banglapal Colexico $440 $480 $1,600 Note: Estimates are government savings in primary and secondary education and in health services. The Alan Guttmacher Institute 28 Table 2.6 Costs of averting a birth through family planning, 1980 (US$ 1987) Country Cost per user Cost per averted birth (low) Cost per averted birth (high) Sri Lanka $8 $31 $41 Colombia $7 $21 $29 Peru $10 $34 $38 Panama $36 $136 $231 Nepal $80 $330 $364 Kenya $100 $350 $386 Libana $238 $259 Banglapal $191 $213 Colexico $121 $144 Note: Out of 16 countries, only those with the three lowest and three highest costs are shown here. Table 2.7 Benefits verses costs in three hypothetical countries Country Benefits Costs Ratio Libana $440 $248 1.8 to 1.0 Banglapal $480 $202 2.4 to 1.0 Colexico $1,600 $133 12.0 to 1.0 Assessing Costs and Benefits 29 Table 2.8 Components of two obstetric care models Limited effort model Moderate effort model Upgrading existing facilities for maternal health care; four centers with surgical capacity Establishment of community outreach system Emergency transportation: one vehicle for each center More health posts (one per 10,000 population); training traditional birth attendants Risk screening; three mobile units; maternity villages Five new referral centers Training traditional birth attendants; provisioning and fees for family planning services Ten maternity beds in district hospital; 1 maternity operating room Coordination of outreach services Training for regional network of maternity services Operational research Development of emergency transportation system Other components The Alan Guttmacher Institute 30 Table 2.9 Costs and benefits of five hypothetical programs in “Himort,” a country with a population of one million and a high rate of mortality No Inter- vention Family planning only (20% cost per user) Family planning only (40% cost per user) Family planning only (60% cost per user) Family planning and limited obstetric Family planning and moderate obstetric Maternal mortality ratio 1,000 1,000 1,000 1,000 800 600 Perinatal mortality rate (%) 52 49 46 44 48 37 Maternal deaths 495 412 342 284 329 247 Maternal morbidity 7,900 6,600 5,500 4,500 5,300 4,000 Perinatal infant deaths 5,600 2,000 1,600 1,200 2,000 1,500 Low-birth-weight babies 7,400 5,800 4,400 3,400 5,400 4,900 Births averted n/a 8,300 15,000 21,000 8,300 8,300 Maternal deaths averted n/a 83 153 211 166 248 Perinatal infant deaths averted n/a 540 970 1,300 590 1,000 Total program cost (US$) $0 $500,000 $1,500,000 $4,500,000 $980,000 $2,000,000 Cost per capita (US$) n/a $0.50 $1.50 $4.50 $0.98 $2.00 Cost per death averted (US$) n/a $810 $1,300 $3,000 $1,300 $1,600 Cost per event averted (US$) n/a $140 $230 $510 $180 $260 Assessing Costs and Benefits 31 Table 2.10 Estimates of the cost-effectiveness of selected interventions, World Health Report 1993 Vitamin A supplementation $1–2/DALY Chemotherapy for tuberculosis $1–3/DALY Family planning, community based distribution, Mali $4–5/DALY Iodization, entire population $8/DALY Iron supplementation, pregnant women $13/DALY Measles immunization $15–19/DALY Food supplementation, pregnant women $24/DALY Family planning, community based distribution, Colombia, Thailand $25+/DALY Management of diabetes $250/DALY Treatment of leukemia $1,000–2,000/ DALY Environmental control of dengue $4,000–5,000/DALY Source: see reference 64. Table 2.11 Elements of public health component of essential package Low-income Middle-income Expanded Program on Immunization Plus $12-17/DALY $25-30/DALY School health program $20-25/DALY $38-43/DALY Tobacco and alcohol control program $35-55/DALY $45-55/DALY AIDS prevention program $3-5/DALY $13-18/DALY Source: see reference 66. Table 2.12 Elements of Clinical Services Component of Essential Package Low-income Middle-income Chemotherapy for tuberculosis $3-5/DALY $5-7/DALY Management of sick child $30-50/DALY $50-100/DALY Prenatal and delivery care $30-50/DALY $60-110/DALY Family planning $20-30/DALY $100-150/DALY Treatment of STIs $1-3/DALY $10-15/DALY Limited care $200-350/DALY $400-600/DALY Source: see reference 67. The Alan Guttmacher Institute 32 Table 2.13 Resource requirements for the ICPD Program of Action (billions of US$) 2000 2005 2010 2015 Population and family planning programs 10.2 11.5 12.6 13.8 Reproductive health 5.0 5.4 5.7 6.1 HIV/STI prevention 1.3 1.4 1.5 1.5 Data/policy/analysis .6 .3 .7 .3 Total 17.0 18.5 20.5 21.7 Table 2.14 Estimated average costs per contraceptive user by major region (US$/user) Region 2000 2005 2010 2015 Sub-Saharan Africa 28.33 24.65 20.61 19.57 Latin America 14.43 14.15 13.97 13.85 North Africa/Western Asia 14.21 13.52 12.98 12.58 East Asia 12.07 12.02 12.00 11.98 South-East Asia 8.37 8.19 8.10 8.05 Southern Asia 13.18 12.80 12.59 12.47 Former Soviet Union and Eastern Europe 12.04 11.99 11.96 11.93 Assessing Costs and Benefits 33 Table 2.15 Resource requirements by major region (billions of US$) Region 2000 2005 2010 2015 Sub-Saharan Africa 1.2 1.7 2.1 2.7 Latin America 1.1 1.3 1.4 1.4 North Africa/Western Asia .4 .5 .6 .7 East Asia 3.5 3.6 3.7 3.7 South-East Asia .7 .7 .8 .8 Southern Asia 2.5 3.0 3.3 3.7 Former Soviet Union and Eastern Europe .7 .8 .8 .7 Total 10.2 11.5 12.6 13.8 Table 2.17 Interventions included in the Mother-Baby Package Care during pregnancy Care during and after delivery Postpartum family Planning Antenatal care Treatment of severe anemia Treatment of syphilis Treatment of other STIs such as gonorrhea and chlamydia Delivery by a skilled birth attendant, including clean and safe delivery and routine newborn care Management of eclampsia Management of postpartum hemorrhage Management of obstructed labor/caesarean delivery Management of sepsis Management of basic newborn complications Postpartum care Management of abortion complications Condom Depo-Provera IUD Norplant Oral contraceptives Sterilization Table 2.16 Resource requirements for global HIV/AIDS program (billions of US$) 2001 2007 Prevention interventions 1.4 6.6 Care/treatment interventions 1.7 7.5 Orphan care interventions 0.1 0.9 Total 3.2 15 The Alan Guttmacher Institute 34 Table 2.18 Recent cost-benefit and cost-effectiveness models in reproductive health Title Date Developed by: Description GOALS Model 2003 Futures Group International (FGI) Estimates the effects of resource allocation decisions on achieving the goals of an HIV/AIDS strategic plan. Resource Needs for HIV/AIDS 2002 FGI A simplified version of GOALS, estimates resource needs for prevention, care and mitigation of HIV/AIDS. MBP Package (MBP) 1999 WHO Relates costs to outputs for a maternal health delivery system. Safe Motherhood Model Under develop- ment FGI Builds on MBP to examine the cost-effectiveness of interventions and the resources required to attain certain levels of maternal mortality. Cost-Estimate Strategy (CES) 1999 Management Sciences for Health A planning, budgeting and management tool to help decision makers in reproductive health commodity management. BenCost, Version 4 1999 FGI Estimates the financial benefits and costs of family planning programs. Prevention of Mother-to- child transmission, Version 1 2002 TFGI Evaluates strategies to prevent mother-to-child transmission of HIV. IDU: Injecting drug user intervention impact model, Version 2.0 2000 London School of Hygiene and Tropical Medicine (LSHTM) Models benefits (reduction in infections)of strategies to reduce HIV transmission among injecting drug users; does not measure costs. SexWork: Sex worker intervention impact model, Version 3.0 1999 LSHTM Models the impact of HIV prevention interventions (condom use and improved STI treatment) on sex workers and their clients. School: School intervention impact model, Version 2.0 1999 LSHTM Models the impact of school-based education projects, simulating the patterns of HIV and STI transmission between in-school youth and older age groups. Blood: Blood transfusion impact model, Version 3.0 1999 LSHTM Models the impact of interventions to strengthen blood transfusion services for HIV prevention. Costing the Essential Health Package Spreadsheet Under development World Bank Extends MBP to include all interventions in the World Bank’s “Essential Package”.(see World Development Report, 1993). CET, cost-effectiveness tool for Mother-to-child transmission interventions, Version 1.0 1999 Health Strategies International (for UNAIDS) Evaluates the cost-effectiveness of interventions to prevent mother-to-child transmission of HIV. Table 2.19 Savings in government expenditures in Thailand (millions of US$) Year Education Health Housing Other Total 1980 129 23 1 12 165 1990 361 38 3 25 426 2000 210 54 33 38 336 Assessing Costs and Benefits 35 36 This chapter uses the best and most up-to-date data available from numerous sources on the current pat- terns of reproductive behavior, services and outcomes across all areas of the developing world. Over the past three decades, surveys of women in developing coun- tries and projects that monitor and analyze reproduc- tive outcomes and health status have yielded valuable information on key aspects of sexual and reproductive health. These include union status and childbearing preferences; contraceptive use patterns, effectiveness and service costs; pregnancy outcomes; and the preg- nancy-related health status of women and infants. We use these data to put together information on re- productive behavior and care for countries where data are available and by making estimates for countries without data by using information from similar coun- tries. In this chapter, we assess the benefits and costs of reproductive health services from two perspectives. First, we estimate the current contribution of contra- ceptive use toward preventing unintended pregnancies, including abortions, and the extent to which mortality and morbidity thus avoided. Next, we focus on women and their partners who have unmet need for contracep- tive care (i.e., are using no method or a traditional method even though they are at risk for unintended pregnancy). Using current patterns of reproductive be- haviors and outcomes and current contraceptive serv- ice costs, we estimate the benefits and costs that could be achieved if all these couples had access to and used modern contraceptive methods. These estimates are limited to the impact of contracep- tive services and supplies, in large part because relevant data for this aspect of sexual and reproductive health care are most readily available. They are conservative figures: As discussed in earlier chapters, these esti- mates do not encompass all of the impacts of inade- quate contraceptive services and supplies. They are limited by available data to those that are most direct- ly tied to unintended pregnancies and their immediate health outcomes. The estimates are further limited in that they do not include impacts from services that often accompany contraceptive care, such as the pre- vention of HIV and other STIs through use of con- doms, or improved infant health from birth spacing. The focus on contraceptive use and service costs is not meant to imply that these are the sole determinants of pregnancy levels, outcomes or health effects. For ex- ample, education of girls and their future life prospects help determine their childbearing goals and reproduc- tive behavior. Effective contraception is necessary, but not sufficient, to allow couples who want to delay or stop future childbearing to do so. Similarly, preventing unintended pregnancies can improve the health of some women and children, but other steps are also needed: for example, providing access to emergency care to women giving birth and improving the nutri- tional status of infants. Thus, the approach and esti- mates presented here provide first steps toward more comprehensive identification of the benefits and costs of investing in sexual and reproductive health care. Methodology • Selection and classification of countries: The esti- mates encompass all developing countries. The “more developed countries” of Europe, Northern America, Australia, New Zealand, Japan and the former Soviet Union were excluded, except for countries in Central Asia. Countries were classified by region and subre- gion according to the schemas used by the United Na- tions Population Division (Appendix Table 3.1a).130,131 To classify countries by economic status, we used cat- egories defined by the World Bank for 2001, which are based on per capita gross national income (Appendix Table 3.1b).132 The groups are: low income, $745 or less; lower middle income, $746-2,975; upper middle income, $2,976-9,205; and high income, more than $9,205.133 When using regional and subregional data from other sources-for example the World Health Organiza- tion (WHO)-values for the appropriate regional and Chapter 3 New Look at the Benefits and Costs of Contraceptive Services in Developing Countries 37 subregional groupings were applied to countries with- in those groupings. While the calculations were made at the country level, data are aggregated and presented here for regional and income groups of countries to lessen the potential variation in measures that can re- sult from small numbers and estimation. In the tables, developing countries in Oceania and Micronesia are in- cluded in the Southeast Asia subregion (see Tables 3.1a and 3.1b for details on the classification of countries according to both geographic region and income groups). • Total population and number of women of reproduc- tive age. Population numbers for each country were es- timated as of July 1, 2003, by straight-line interpola- tion between data for 2000 and 2005.134 Women aged 15-49 were defined to be women of reproductive age. In the source data, some countries with very small pop- ulations are sometimes missing from detailed country listings, but included in the regional totals. Populations in these countries were assigned the relevant subre- gional average for purposes of estimation. • Marital status of women aged 15-49. Women were classified according to whether they were currently married,* formerly married or never married using sev- eral sources, listed here in order of priority: 1. The most recent Demographic and Health Survey (DHS) or other national survey for a country.135 2. Proportions of women aged 15-49 who were cur- rently married, compiled in a recent report that draws from data from the United Nations Population Divi- sion and various national surveys.136 The proportion of unmarried women were distributed into formerly married and never married according to the distribu- tion of unmarried women aged 15-49 in the United Nations marriage database (see below, item 3), if available. Otherwise, they were distributed based on the unweighted average distribution from countries in the subregion with DHS surveys or from a similar country. 3. United Nations Population Division, Database on Marriage Patterns, an unpublished compilation of census and survey data on marital status by age and sex over the past 40 years, provided June 5, 2002. For our estimates, distributions of women by marital sta- tus for the most recent available year (1990 or later) were used. 4. Estimates based on the unweighted average per- centage distribution by marital status of countries with DHS data in the relevant subregion. 5. Estimates based on the DHS data available for a country in the region that has similar marriage patterns. • Women at risk for unintended pregnancy. This was de- fined to be all women using modern contraceptive methods (including sterilization), as well as those with unmet need for effective contraception-i.e., women using a traditional method and those using no contra- ceptive method who are sexually active, able to become pregnant and who do not want more children (“lim- iters”) or do not want a child in the next two years (“spacers”). • Contraceptive method use. Categories for contracep- tive method use were sterilization, male or female, modern reversible methods-IUD, long-acting hormon- al methods (injectable and implant), the pill, the con- dom, vaginal barrier methods and spermicides-and tra- ditional methods, including periodic abstinence, withdrawal and other nonmodern methods. Most women using long-acting hormonal methods use in- jectable contraceptives. • Distribution of women aged 15-49 by risk for unin- tended pregnancy, contraceptive method use and fertil- ity-preference status (spacing or limiting), according to marital status. 1. Several sources were used to estimate the per- centage distributions of women aged 15-49 in each marital status category by risk for unintended preg- nancy, contraceptive method use and fertility-prefer- ence status. The type of source and estimation method- ology varied according to what data were available: a. For all countries with a DHS survey from 1990 or later that was available as a public-use file, the most re- cent DHS was used. These percentages were obtained by special tabulations of DHS surveys.137 In some cases, the percentages of women with unmet need and using no method differ slightly from those published by Westoff.138 For one country, this is due to an error in the DHS report. For the others, dif- ferences apparently are due to revisions in the data files after the DHS reports were prepared. All formerly married women using no method who had had sex in the last month and half of those who had had sex in the last year but not in the last month were considered to be sexually active. It was assumed that the other half of formerly married women who had unmet need, were not using contraceptives and had had* In this report, “married” includes women in consensual unions. The Alan Guttmacher Institute 38 sex in the last year but not in the last month were not at risk for unintended pregnancy. b. For countries with no recent DHS survey, but for which published information was available by marital status on the percentage of women not in need, the per- centage with unmet need using no method and the per- centage using contraceptive methods, the published in- formation was used and, if necessary, the distribution of users according to method was estimated by apply- ing the unweighted average distribution of countries in the subregion with DHS data.139 Information for all unmarried women was applied to never married and formerly married women if data for these two categories were not separately available. Since estimates published in DHS reports exclude un- married women who had not had sex in the month be- fore the interview from the category of women with unmet need using no contraceptive method, the report- ed percentage of women in this category was inflated by the ratio of the percentage of all never married women (or formerly married women) with unmet need using no method to the percentage of never married women (or formerly married women) with unmet need using no method who had had sex in the last month, ac- cording to spacing or limiting status, based on the un- weighted average of countries for which special tabu- lations were done, in the relevant subregion (see above, item 1.a.). c. For countries with no recent DHS survey avail- able, but with information from a published source for women aged 15-49 by marital status on the percentage with unmet need and using no method, the percentage not in need and the distribution of method use, by spac- ing and limiting, was estimated from unweighted aver- age distributions for countries in the subregion with DHS data.140 If information by spacing and limiting status was not available, it was assumed that all women using sterilization were seeking to limit births and that half of women using other methods were spacing and half were limiting. d. For China, it was estimated that 2% of currently married women aged 15-49 had unmet need for spac- ing and were not using any method and 2% had unmet need for limiting and were not using a method. Based on the percentage distribution of need and method use among all married women from the 1992 National Fer- tility and Family Planning Survey,141 it was estimated that 12.6% of married women aged 15-49 were not in need of contraceptives. Further, it was assumed that all couples using sterilization were limiting further child- bearing, that half of IUD users were limiting and half spacing and that all users of pill, injection, condom and traditional methods were spacing. e. For each country that did not have a nationally representative fertility survey or for which data were not available for a particular marital status, either the unweighted average distribution of its subregion based on countries in the subregion that had surveys, or the distribution from a country at a similar level of demo- graphic transition in the same subregion or region was used. 2. Based on the above assumptions, the estimated numbers and percentages of women with unmet need and using no contraceptive for the developing world as a whole (minus China) in 2003 are shown below. They are close to the estimates made for 2000 by Ross and Winfrey (see Table 3.1).142 3. The estimated percentage distribution of current- ly married women according to contraceptive method used is similar to the proportions estimated among cou- ples of reproductive age in less developed regions for 1998 by the United Nations Population Division.143 However, the actual number of married women of re- productive age using some method has grown substan- tially, by about 11 million, between 2000 and 2003 (see Table 3.2). • The cost of contraceptive services. The cost of con- traceptives ranges widely across available studies, even within the same country, often reflecting different serv- ice settings and differing cost components. The esti- mates in this report use the average costs available from the UNFPA Costing Initiative database to represent an- nual cost across all regions.144 These average costs summarize results from a large number of studies, sep- arating costs for each method into components of drugs and supplies, labor, overhead (including capital costs, although these are likely to be incompletely reported) and other costs such as hospitalization for tubal liga- tion. Summary tables of this information are given in Appendix 2.3 of this report. Given the variation seen in costing studies, even within the same country, the small numbers of studies in some regions and the roughness of the cost estimates, we did not adjust the UNFPA Costing Initiative figures by region. Costs of long-term methods were annualized using standard assumptions: 10 years for sterilization and three years for the IUD to take into account the average length of coverage from these methods.145 For other methods, the estimates are based on supplying 13 cy- Assessing Costs and Benefits 39 cles of oral contraceptives, 96 condoms or four injec- tions per year. The UNFPA Costing Initiative estimat- ed average costs in 2001 dollars. These were projected to 2003 dollars using an inflation factor of 4%. The annual method-specific cost estimates used (in 2001 dollars) were as follows: 1. IUD: Average total cost per user was $26.43. a. Drugs and Supplies: An average cost of $4.05 was used. This was based on visit costs for insertion, follow-up and removal. The average drug and supply cost for IUD insertion is adjusted from $1.37 to $2.50, based on footnote 1 of UNFPA Costing Initiative not- ing that the $1.37 average cost was low, in large part be- cause of very low drug and supply costs of $0.21 and $0.72 in a Turkish study, but that the median drug and supply cost would be $2.50 without the Turkish study. The UNFPA Costing Initiative averages of $1.02 for follow-up visit and $0.53 for removal visit were used. b. Labor: $3.35 ($1.46 for insertion, $1.30 for fol- low-up and $0.59 for removal visit). c. Overhead: $19.03, based on available studies showing that drugs, supplies and labor accounted for 28% of total costs and overhead for 72% of total costs. 2. Injectables cost an average of $30.35 per user per year. a. Drugs and supplies: $1.41 for acceptance visit and $1.21 for follow-up visit. b. Labor: $0.65 for acceptance visit and $0.43 for follow-up visit. c. Overhead: $6.90 for acceptance visit and $5.49 for follow-up visit, based on assumption that drugs, supplies and labor accounted for 23 percent of total costs and overhead for 77%. These proportions were estimated as the average of percentages from available studies for IUDs and for pills. d. For total cost, this study assumed one acceptance visit ($8.96) and three follow-up visits (3x $7.13). 3. Oral contraceptives cost an average of $35.70 per user per year. a. Drugs and supplies: $0.84 for acceptance visit (on average 2 cycles and other materials) and $0.78 for follow-up supply visits (2-3 cycles). b. Labor: $0.52 for acceptance visit and $0.36 for follow-up supply visits. c. Overhead: $6.42 for acceptance visit and $5.29 for follow-up visits, based on available studies show- ing that drugs, supplies and labor accounted for 18% and overhead for 82% of total costs. d. For total cost, we assumed one acceptance visit ($7.78) and 4.4 follow-up visits ($6.43) for a total of 13 cycles. 4. Condoms were estimated at an average annual cost per user of $13.56. a. Drugs and supplies: $0.79 per visit (12-20 con- doms and other materials; the assumed average num- ber of condoms dispensed was 16 per visit). b. Labor: $0.34 per visit. c. Overhead: Estimated at $1.13, assuming that overhead accounted for half of total cost. d. Total cost: We assumed six visits, for a total of 96 condoms per year (6x$2.26). 5. Female sterilization: Average total cost was estimat- ed at $88.70. This includes $80.10 for surgery, $5.73 for an evaluation visit and $2.87 for a follow-up visit. a. Drugs and supplies: $20.39 for surgery visit. b. Labor: $22.21 for surgery visit, $3.26 for evalua- tion visit and $1.63 for follow-up visit. Estimated av- erage costs for evaluation visit and follow-up visit as- sumed the same distribution between labor and overhead as for surgery visit. c. Overhead: $16.82 for surgery visit; estimated $2.47 for evaluation visit and $1.24 for follow-up visit, assuming the same distribution between labor and overhead as for surgery visit. d. Hospitalization: $20.68 for “bed” or hospitaliza- tion costs for the overnight stay required for some procedures. 6. Vasectomy: Total cost averaged $59.42. For vasec- tomy, we assumed the same drug and supply, labor and overhead costs as for tubal ligation, but no other or hos- pitalization costs. The UNFPA Costing Initiative gives information for only one study, a Brazilian vasectomy campaign, at $9.30 per couple year of protection. Acharya cites Janowitz, Bratt and Fried in estimating a unit cost of $100 for both female sterilization and vasectomy.146 • Pregnancies averted. The number of pregnancies averted by current use of modern contraceptive meth- ods was estimated by subtracting the number of preg- nancies occurring to current users of modern contra- ceptives from the number that would occur if they used no method. The number of pregnancies that would be averted by serving all those with unmet need for contraceptives (i.e., those using no method or those using a tradition- al method) was estimated as the difference between the The Alan Guttmacher Institute 40 number of pregnancies currently occurring to women with unmet need and the number that would occur if they used modern contraceptives in the same distribu- tion as women in their country who are current users, by fertility-preference status and marital status. • Pregnancy rates for women using each method and for women at risk of unintended pregnancy using no method were estimated from method-specific use- failure rates, which were adjusted to be consistent with estimates of the number of unintended pregnancies in 2003 in each major region. 1. There were an estimated 75.9 million unintended pregnancies in developing countries in 2003. To calcu- late this, the estimated number of unintended pregnan- cies in developing countries circa 1999 (74.7 mil- lion)147 was projected to 2003 by multiplying the number in 1999 by the ratio of 2003 births (120.6 mil- lion) to 1999 births (118.7 million).148 Similar calcu- lations were done by major region to estimate the num- ber of unintended pregnancies in each region in 2003. A further proportional adjustment was made so that the estimated regional figures totaled to 75.9 million. 2. Annual pregnancy rates for users of each method and for women at risk of unintended pregnancy using no method were estimated in multiple steps. Base use- failure rates were estimated for each method. Reversible methods: We used median method-specif- ic cumulative probabilities of failure per 12 months of use provided by Cleland.149 These were calculated from DHS data from married women in 13-18 coun- tries, depending on the method. Across the surveys, the total number of abortions (induced and spontaneous) was assumed to be underreported. The base failure rates are: • Pill: 6.9% • IUD: 1.8% • Injectables (also used for implants): 2.9% • Condom (also used for other supply methods): 9.8% • Periodic abstinence: 21.6% • Withdrawal: 15.1% • Other non-supply/traditional methods: Used average of rates for periodic abstinence and withdrawal: 18.35% Sterilization: We used pregnancy rates from Trussell et al.:150 • Tubal ligation: 0.5% • Vasectomy: 0.2% No method: an initial annual pregnancy rate of 40% was assumed. This 40% estimate is much lower than the 85% an- nual pregnancy rate that Trussell et al. estimate for cou- ples continually sexually active throughout a year's time. Some studies have suggested, however, that couples at risk of unintended pregnancy who are using no contra- ceptive method are not continually sexually active.151 3. The base annual pregnancy rates were adjusted so that the total number of unplanned pregnancies to women at risk of unintended pregnancy in developing countries equaled the estimated total of 75.9 million. First, the numbers of nonusers and users of each method were multiplied by the base relevant pregnan- cy rates, with no adjustment for differences in failure rates by union status, intention for future pregnancy or age. These calculations yielded a total of 82.1 million unintended pregnancies, distributed across regions as shown in Table 3.3. Adjustment factors were calculat- ed for each region as the ratio of the expected number of unintended pregnancies to the estimated number be- fore adjustment. These regional adjustment factors were applied to the method-specific pregnancy rates for each country in the region. Pregnancy Outcomes: It was assumed that all preg- nancies to women at risk for unintended pregnancy would be unplanned pregnancies. Pregnancies were distributed according to outcome (unplanned births, in- duced abortions and spontaneous abortions or miscar- riages) based on the estimated distribution of outcomes of unplanned pregnancies for subregions.152 Subre- gional averages were applied to all countries within that subregion. In these calculations, it was estimated that the num- ber of induced abortions in Southeast Asia outside of China was 0.5 million, that there were 10.6 million abortions in China and that there were 50,000 abortions in Oceania. The total number of induced abortions in North Africa was estimated in 1995153 as 600,000, but the number of unsafe abortions in the region was esti- mated by WHO as 700,000. These estimates assumed that the total number of abortions in the region was 750,000.154 • Maternal deaths: The numbers of maternal deaths due to abortion and to all other pregnancy-related causes were estimated by drawing on data from several sources: 1. The number of maternal deaths due to all preg- nancy-related causes for each country in 2003 was es- Assessing Costs and Benefits 41 timated by multiplying the number of deaths in 2000 (526,000)156 by the ratio of the number of births in 2003 to the number of births in 2000.157 The resulting total for 2003 was 530,000. 2. Maternal mortality due to abortion: The estimated number of unsafe abortions and associ- ated maternal mortality includes abortions provided in countries where the procedure is highly restricted, and those provided under unsafe conditions in countries where abortion is permitted under broad legal grounds. a. Numbers of abortions in unsafe settings and in legal medical settings: Estimates of the number of un- safe abortions, by subregion, were taken from WHO estimates for 2000 (18.33 million unsafe or nonlegal abortions in 2000 in developing countries) because the 2000 estimates were very similar to estimates for 1995, they were used as the 2003 estimates without change.158 The number of safe abortions in legal set- tings was estimated for each subregion by subtracting the estimated number of unsafe abortions159 from the estimated total number of abortions in 2003.160 Abor- tions in each setting were distributed across countries in each subregion based on the legal status of abortion so that the total number equaled the total unintended pregnancies ending in induced abortion in each coun- try. The country-level estimates of abortions by setting are consistent with the estimated total number of in- duced abortions in each country and with the regional totals from the original sources of data. However, they are still rough estimates. b. Maternal mortality from abortion: Estimates of the number of maternal deaths from unsafe abortion per 100,000 unsafe abortions, by subregion, were taken from WHO estimates for 2000 and applied to all unsafe abortions in countries in each subregion.161 Mortality rates per 100,000 abortions in legal and medical set- tings were based on experience in developed coun- tries.162 We used the average rate of 0.5 deaths per 100,000 legal abortions from seven Western European countries 1976-1995 for East Asia, Israel, Singapore, Tunisia, Turkey and Vietnam. For other countries, we assumed that the mortality rate was 1 death per 100,000 abortions in legal and medical settings. 3. Maternal mortality from causes other than induced abortion. The estimated number of maternal deaths from induced abortion were subtracted from the total number of maternal deaths from all pregnancy-related causes in each country estimated by WHO163 to esti- mate the number of maternal deaths from causes other than induced abortion. Maternal mortality ratios from causes other than induced abortion were calculated for each country as the number of maternal deaths from causes other than induced abortion per 100,000 live births. • Infant deaths: The infant mortality rate (deaths under age one per 1,000 live births) for 2000-2005, by coun- try, was applied to the relevant number of births to cal- culate the number of infant deaths.164 • Children who would not lose their mothers: The num- ber of maternal deaths was multiplied by the average number of living children women have had to estimate the number of children impacted by maternal deaths. Estimates are based on the average number of living children women have had, according to whether they are spacers or limiters and according to type of method use (sterilization, reversible modern, traditional or no method), by union status. DHS data were used when available. When DHS data were not available, subre- gional unweighted averages were used. When DHS data were not available for computing subregional av- erages, estimates of the mean number of living children were based on results for similar countries in the sub- region or in a similar region. For China, we assumed that currently and formerly married women delaying or spacing future births have an average of 0.25 children, that currently and formerly married limiters average 1.0 child and that never-married women at risk for un- intended pregnancy average zero children. • Disability-Adjusted Life Years (DALYs): The number of DALYs lost among infants and children was esti- mated by using the number of DALYs lost per 1,000 births due to perinatal conditions, by subregion, in 2001. These rates were then applied to unintended births in 2003 in each country in the subregion, ac- cording to subgroups of women (e.g., current contra- ceptive users and nonusers). The number of DALYs lost among women because of maternal conditions other than induced abortion was estimated by using the number of DALYs lost per 1,000 births from all maternal conditions except in- duced abortion, by subregion, in 2001. These rates were then applied to unintended births in 2003 in each country in the subregion, according to subgroups of women (e.g., current contraceptive users and nonusers). The number of DALYs in 2003 lost among women The Alan Guttmacher Institute 42 because of induced abortions was estimated from the number of DALYs lost due to induced abortion per 1,000 births in 2001, by subregion, multiplied by the ratio of 2003 births to 2001 births. DALYs lost in 2001 by specific cause or condition, by subregion, were obtained from WHO.165 Births in 2001 and 2003 were from United Nations Population Division, 2003.166 Abortions in 2003 were from the Alan Guttmacher Institute estimates.167 Years of Life Lost due to premature mortality (YLLs) were calculated from the same sources and in the same manner as DALYs. Years Lost due to Disabil- ity (YLDs) were calculated by subtracting YLLs from DALYs. • Notes and limitations. These estimates present a vari- ety of measures of outcomes of contraceptive use and are not necessarily additive. DALYs, for example, in- corporate estimates of maternal and infant mortality. Costs to provide contraceptive services and supplies were average costs and were not increased for the pre- sumed higher costs associated with setting up new services or serving rural acceptors who may become clients at later stages of a delivery program, nor de- creased for economics of scale that may result from in- creased numbers of users relative to service infrastruc- ture and staffing or increasing cost-efficiency resulting from competition among service providers. The outcomes and costs estimated here are one-year measures and are not discounted for future value. No adjustments were made to take into account benefits or costs of contraceptive use in terms of longer birth in- tervals or births occurring more in line with couples' preferences. Rows and columns may not sum to totals because of rounding. In most tables, numbers are shown in thou- sands. However, estimates of maternal and infant deaths and numbers of children losing their mothers are shown without rounding to make them useful for pos- sible calculations combining numbers. When present- ed as absolute numbers, they should be rounded as ap- propriate. Estimates In 2003, 5.1 billion people were living in developing countries, 3.8 billion of them outside of China (see Appendix Table 3.2). One in four people in develop- ing countries-1.3 billion-were women aged 15-49. • Almost half (46%) of these women lived in the low- est income countries, where average per capita annual income was $745 or less.168 One-fourth (27%) live in China, 46% lived in other Asian countries, 15% in Africa and 11% in Latin America and the Caribbean. • Roughly two-thirds of women of reproductive age in developing countries were in a union, ranging from 72% of women in low-income countries to 60-61% in upper-middle and high-income countries. Twenty- seven percent had never married and 5% were former- ly married. • The proportions of women aged 15-49 currently in union were highest (71-74%) in low-income countries and in countries in Middle and Western Africa and in China and South Central Asia. They were lowest in Southern Africa (43%) and in Northern Africa and Latin America and the Caribbean (58-60%). Among all women aged 15-49 in developing countries in 2003, more than half were at risk for unintended pregnancy because they were sexually active, able to become pregnant if they and their partner used no contraceptive method and they did not want more children (“limiters”) or did not want a child in the next two years (“spacers”). • Forty-seven percent of all women aged 15-49 were not at risk because they were not sexually active, they or their partners were infertile or they wanted a child within the next two years. More than half of women in the lowest-income countries were not at risk for unin- tended pregnancy, as were 60% or more of all women aged 15-49 in all of Sub-Saharan Africa except South- ern Africa. • Eighteen percent of all women were at risk and seek- ing to delay or space future births. Roughly one quar- ter of women in Middle, Southern and Western Africa were at risk and trying to delay or space future births. • Thirty-six percent of all women aged 15-49 in devel- oping countries were at risk of unintended pregnancy and seeking to limit future childbearing. This propor- tion rose to more than 40% of women in lower-middle income countries, in China and in South America. An estimated 183 million women in developing coun- tries were pregnant in 2003, and 76 million of these women were pregnant even though they had wanted a birth at a later time or not at all (Appendix Table 3.3). • There were a total of 121 million live births, 35 mil- lion induced abortions and 28 million miscarriages in developing countries in 2003. • Some 107 million women had intended pregnancies (including a small proportion who had no preference Assessing Costs and Benefits 43 regarding the timing of the pregnancy). Of these, 89 million women gave birth and 18 million had miscar- riages. Sixty-four percent of the 76 million unintended preg- nancies were to women in Asia, 21% to women in Africa and 15% to women living in Latin America and the Caribbean. • An estimated 108 out of every 1,000 women at risk for unintended pregnancy in 2003 became pregnant when they had not wanted to. More than 20% of women at risk became pregnant unintentionally in Sub-Saharan African countries other than those in Southern Africa. Rates of unintended pregnancy were lowest in East Asia, including China, and in high-income countries. • Among women with unintended pregnancies, 31 mil- lion gave birth, 35 million had induced abortions and 10 million miscarried. Forty-two percent of all unin- tended pregnancies ended in birth, 46% were terminat- ed by induced abortion and 13% were miscarried. • Unintended pregnancies accounted for 41% of all pregnancies to women in developing countries. More than half of all pregnancies were unintended among women in Southern Africa, and in Latin America and the Caribbean. Only 34% of pregnancies in Middle and Western Africa were unintended. More than half a million women (530,000) in devel- oping countries died in 2003 from causes related to pregnancy (Appendix Table 3.4). About one-third of all maternal deaths were to women who had become pregnant when they did not want to be. • Of the 184,000 women with unintended pregnancies who died from pregnancy-related causes, 69,000 died from complications of induced abortion. Almost all of these women had abortions in unsafe or nonlegal condi- tions. The other 115,000 women died from pregnancy- related causes other than induced abortion. • Across all developing countries, an average of 382 women died from pregnancy-related causes other than induced abortion for every 100,000 births that oc- curred. The highest rates of maternal mortality from nonabortion causes were in low-income countries (570 deaths per 100,000 births) and in Sub-Saharan Africa outside of Southern Africa. In these African countries, 754-870 women died for every 100,000 births. • Mortality among all women who had induced abor- tions was 199 deaths per 100,000 abortions. It was highest in subregions of Sub-Saharan Africa, with 583- 834 women dying per 100,000 induced abortions. Again, rates of abortion-related death rates were high- est in low-income countries. More than 7 million of the 121 million babies born in developing countries died before their first birthday, a rate of 61 infant deaths per 1,000 live births. • 5.5 million infants who died in their first year had been intended conceptions, while 1.8 million had been born from unintended pregnancies. • Infant mortality was highest in Middle Africa, where 116 of every 1,000 babies died before their first birth- day. Infant death rates were lowest in East Asia, in- cluding China, and in Latin America and the Caribbean. • The risk of infant death was more than 10 times as high in low-income countries (78 deaths per 1,000 births) as in high-income areas (6 deaths per 1,000 births). In 2003, pregnancy and childbirth in developing countries resulted in the loss of an estimated 126 mil- lion DALYs, 30 million among women and 96 million among infants (Appendix Table 3.5). • Almost three-quarters of all maternal DALYs and two-thirds of all infant DALYs are accounted for by women and children in Sub-Saharan Africa and South Central Asia. • Among women, mortality and morbidity from causes other than abortion accounted for 25 million DALYs. These included 4 million DALYs from maternal hem- orrhage and 4 million from maternal sepsis, 3 million from obstructed labor, 2 million from hypertensive dis- orders of pregnancy and 12 million DALYs from other maternal conditions (data not shown). • Maternal DALYs from causes other than abortion are almost evenly split between 13 million YLLs and 12 million YLDs. • Causes related to abortion result in 5 million DALYs among women in developing countries. DALYs from abortion-related causes are more likely than those from other pregnancy-related causes to be linked to disabil- ity. These include 2 million YLLs (36%) and 3 million YLDs (64%). • The major causes of the 96 million perinatal DALYs are low birth weight (48 million DALYs), birth as- phyxia and trauma (33 million), and other perinatal conditions (15 million). • Rates of DALYs among women from causes related to pregnancy and infant DALYs were highest in low- income countries. The highest rate of DALYs from nonabortion maternal causes was in Southern Africa (382 DALYs per 1,000 births), while the highest abor- tion-related rate was in Middle Africa (589 DALYs per 1,000 abortions). The rate of perinatal DALYs was The Alan Guttmacher Institute 44 highest in South Central Asia (1,028 DALYs per 1,000 births). Among the 705 million women in developing coun- tries in 2003 at risk of unintended pregnancy, 504 million used modern contraceptives and 201 million used either a traditional method or no method at all (Appendix Table 3.6). • Almost half (47%) of all women using modern con- traceptives relied on sterilization (female, 204 million; male, 32 million). One-third were using long-acting methods, predominantly the IUD (137 million) or in- jections or implants (32 million). Twelve percent used oral contraceptives (62 million), and 7% relied on con- doms or other supply methods (37 million). • More than two-thirds of the 201 million women with unmet need for contraceptive services were using no contraceptive (137 million). Some 13% of these women were using periodic abstinence (27 million), 14% withdrawal (28 million) and 4% other nonsupply methods (9 million). Patterns of method choice vary across regions of the world and according to women's fertility-preference status. • Female sterilization is the most commonly used method in developing countries, with 204 million users, followed by the IUD, with 137 million. • Female sterilization is the most common method in East Asia, including China, as well as in South Central Asia and Latin America and the Caribbean. The IUD is the predominant method in Northern Africa. Long-act- ing hormonal methods are the most commonly used contraceptives in Eastern and Southern Africa and in South East Asia. Periodic abstinence is the predomi- nant method in Middle and Western Africa, while with- drawal is most common in Western Asia. • Among women and their partners who are trying to space future pregnancies, the most commonly used method is the IUD, while female sterilization predom- inates among those trying to limit future births. • The IUD is the most commonly used method among spacers in Northern Africa and in East Asia, including China. Long-acting hormonal methods are the most common in Southern Africa and in Southeast Asia, while oral contraceptives are predominant in Eastern Africa and Latin America and the Caribbean. Couples in Middle Africa and Western Africa are most likely to use periodic abstinence, and those in Western Asia are most likely to use withdrawal. • Among women at risk who do not want to have any more children, female sterilization is the most com- monly used method overall and in East Asia, including China, in South Central Asia and in Latin America and the Caribbean. Women in Northern Africa and South- east Asia are most likely to use IUDs, while long-act- ing hormonal methods are the most commonly used contraceptives among limiters in Eastern and Southern Africa. In contrast, traditional methods are the most commonly used contraceptives among couples in Northern Africa (periodic abstinence) and in Western Asia (withdrawal) who do not want to have any more children. Seventy one percent of all women in developing coun- tries at risk for unintended pregnancy and their part- ners are using a modern contraceptive method; the remainder (29%) have unmet need for modern fami- ly planning services. Forty-four percent of women trying to space future births and 21% of those seek- ing to limit childbearing have unmet need (Appendix Table 3.7). • More than four in 10 women at risk for unintended pregnancy in low-income countries have unmet need for modern contraceptive services. • Almost two-thirds (63%) of women at risk in Sub-Sa- haran Africa are using no contraceptive (47%) or a tra- ditional method (16%). More than half of women at risk in Eastern and Western Africa are using no method. • Roughly two-thirds of women at risk in low-income countries who are trying to space future births have unmet need and almost a third of those trying to pre- vent all future births have unmet need for modern con- traception. • In Sub-Saharan Africa, 70% of women at risk who are spacing have unmet need for modern contraceptives, as do 51% of those at risk who are trying to limit all future births. Levels of unmet need are also especially high among women who do not want a child soon in South Central and Western Asia (61-69%) and in Central America (54%) and among women in Western Asia seeking to limit all future childbearing (56%). Services and supplies for the 504 million women and their partners in the developing world currently using modern contraceptives cost an estimated $7.1 billion annually, averaging $14 per user (Appendix Table 3.8). • Drugs and supplies for contraceptive services cost an estimated $1.3 billion annually; labor to provide serv- ices costs $1.1 billion; and hospitalization costs asso- ciated with female sterilization cost an estimated $435 Assessing Costs and Benefits 45 million, prorated to take into account the typical length of contraceptive coverage from sterilization. • An estimated $4.3 billion of the service costs are for the overhead and capital needed to provide buildings, infrastructure and other staff necessary for service pro- vision. (Overhead costs, especially capital costs, are most likely underestimated since not all costing stud- ies in the UNFPA database that underlies these esti- mates, measure them fully.) • Costs for current services total an estimated $850 mil- lion in Africa, $5.2 billion in Asia ($3.0 billion in Asia outside of China) and $1.0 billion in Latin America and the Caribbean. • Apparent differences in the average cost per user ($13 in Asia, $18 in Latin America and the Caribbean and $24 in Africa) reflect different patterns of method use because available data did not support making country- or region-specific cost estimates for each method. While the aggregate estimates are broadly representa- tive because they are based on studies in many coun- tries, local-level costs will vary due to other factors, such as start-up costs and economies of scale. If new cost data become available, these estimates can be im- proved. Because contraceptive methods can be difficult to use perfectly, women relying on modern contraceptives experience almost 16 million unintended pregnancies annually, but if they were using no contraceptive this number would rise to 203 million (Appendix Table 3.9). • An estimated 8.8 million women who become preg- nant while using contraceptives (because of incorrect use or method failure) have induced abortions, 5.0 mil- lion give birth and 1.9 million miscarry. • While women using modern methods represent 71% of all women in developing countries at risk for unin- tended pregnancy (Table 3.7), they account for only 21% of all unintended pregnancies, 16% of all unin- tended births and 25% of all abortions (Table 3.9). Use of modern contraceptives annually averts 187 million unintended pregnancies, 215,000 maternal and 2.7 million infant deaths. These averted preg- nancies prevent 685,000 children from having to grow up without their mothers and avert 60 million DALYs (Appendix Table 3.10). • The difference between the number of unintended pregnancies that would occur if modern contraceptive users were using no method (203 million) and the num- ber that now occur (16 million) represents the number of unintended pregnancies that are prevented annually by use of modern contraceptives-187 million. • By averting 187 million unintended pregnancies, con- traceptive use in developing countries is preventing an estimated 60 million unintended births, 105 million in- duced abortions and 22 million women from having miscarriages. • Contraceptive use is averting 2.9 million deaths each year, 215,000 pregnancy-related deaths of women and 2.7 million infant deaths. Another 685,000 children are spared having to grow up without their mother. • An estimated 60 million DALYs are saved by avert- ing these unintended pregnancies, 16 million among women and 44 million among infants and children. The 201 million women at risk for unintended preg- nancy but with unmet need for modern contraceptive services represent 15% of all women aged 15-49 in developing countries (Appendix Table 3.11). Most are married and using no contraceptive method, and they are fairly equally split between women who want to delay or space future births and those who want no more children (Appendix Table 3.11). • Women with unmet need account for 20% of all women aged 15-49 in low-income countries, 22% of women in Africa, 14% of those in Asia and 16% of women in Latin America and the Caribbean. • Women in low-income countries are more likely to be in need of contraception to delay or space births than to limit them, while the opposite is true in higher-income countries. • Almost two-thirds of women in Africa with unmet need want a child in the future, compared with 47% of women in Asia and 46% of those in Latin America and the Caribbean. • While 88% of all women with unmet need in the de- veloping world are currently in union, 9% have never been married; however, 13% of women with unmet need in Africa and 13% in South America are unmarried. • More than two-thirds of all women with unmet need are using no contraceptive method, ranging from 74% of those in low-income countries to 55% of women with unmet need in high-income areas; the rest are using traditional methods. • More than 75% of women with unmet need are cur- rently using no contraceptive in Africa other than Mid- dle and Western Africa, in East Asia, including China, and in the Caribbean and Central America. An estimated 60 million unintended pregnancies occur annually to women with unmet need for con- The Alan Guttmacher Institute 46 traceptive services (Appendix Table 3.12). Women with unmet need account for 79% of all unintended preg- nancies that occur each year in developing countries. • An estimated 32 million unintended pregnancies occur annually to women with unmet need who want- ed to have children at a later time, and 28 million are to women who had not wanted any more children at all. • The 137 million women at risk of unintended preg- nancy but using no contraceptive method account for 50 million unintended pregnancies annually (two- thirds of the 76 million total unintended pregnancies) and the 64 million women at risk of unintended preg- nancy using traditional methods have 10 million unin- tended pregnancies each year. • Unintended pregnancies to women with unmet need annually result in an estimated 26 million births, 26 million induced abortions and 8 million miscarriages. Unintended pregnancies to women with unmet need annually cause 1.8 million deaths, including 166,000 women who die from pregnancy-related causes, leav- ing 578,000 children to grow up without their moth- ers. These pregnancies also lead to 1.6 million infant deaths. A total of 32 million DALYs result annually from unintended pregnancies to women with unmet need (Appendix Table 3.13). • Each year, 166,000 women in developing countries die because of unintended pregnancies while they were using no contraceptive or a traditional method, 104,000 women who were carrying their pregnancy to term or who had a spontaneous abortion and 63,000 women who sought induced abortions, mostly under unsafe conditions. • Some 1.6 million women with unintended pregnan- cies give birth to infants who die before their first birth- day, and mothers dying from pregnancies that resulted from their unmet need leave behind an estimated 578,000 children. • Eighty percent (1.4 million) of deaths caused by preg- nancies to women with unmet need occur in low-in- come countries. • Pregnancies to women with unmet need result in an estimated 31.5 million DALYs lost, 10.3 million among women and 21.2 million among infants and children. If all 201 million women in developing countries with unmet need adopted modern contraceptive methods, there would be substantial numbers of new users to be served (Appendix Table 3.14). • The numbers of additional women and men needing contraceptive services represents a significant chal- lenge to current providers of family planning care and supplies. • If those with unmet need adopted modern methods in proportions similar to those among women in their country or subregion who are using modern contra- ceptives to space or limit childbearing, there would be over 42 million new users of female sterilization, 43 million of the IUD, 48 million of oral contraceptives, 43 million of the condom and 23 million of long-acting hormonal methods (injectables and implants). The service costs to provide women with unmet need with modern contraceptive methods would total an extra $3.9 billion annually (Appendix Table 3.15). • Drugs and other supplies would cost an estimated $696 million annually; labor would cost $413 million; and hospitalization for women being sterilized would cost $90 million. Overhead and capital costs would total $2.7 million, though some of this might be real- ized through more intensive use of existing buildings, personnel and service structures. • More than half ($2.4 billion) of the costs to serve those with unmet need would occur in the low-income countries least able to afford them. • If method-mix patterns observed among current users were followed by women with unmet need, almost half of the total costs would be for providing oral contra- ceptives (Appendix Table 3.16). If all 201 million women with unmet need received the services they need and used modern methods of contraception, instead of the 60 million pregnancies they currently have each year, they would have an es- timated 8 million pregnancies, averting 52 million unintended pregnancies annually (Appendix Table 3.17). • An estimated 26 million pregnancies that now occur at a time when women do not want to have a child would be averted to a later time and another 26 million would be averted among women who want no more children. More than half of all these unintended pregnancies that could be averted occur in low-income countries. Averting 52 million pregnancies to women who cur- rently have unmet need for contraceptive services would prevent 23 million unintended births, 22 mil- lion abortions and 7 million miscarriages (Appendix Table 3.18). • Preventing 52 million unintended pregnancies would Assessing Costs and Benefits 47 reduce the number now occurring to women in need of modern contraception from 76 million to 24 million. • The number of unintended births would be reduced from 31 million to 9 million and the number of induced abortions from 35 million to 12 million. By preventing those 52 million unintended pregnan- cies that now occur to women with unmet need, 1.5 million lives would be saved each year and 27 million DALYs would be saved (Appendix Table 3.19). • An estimated 142,000 pregnancy-related deaths to women would be prevented by these averted pregnan- cies: 89,000 maternal deaths not related to induced abortion and 53,000 deaths from unsafe abortion. Averting these deaths would prevent an estimated 505,000 children from losing their mothers. • There would be 1.4 million fewer infant deaths annu- ally if women with unmet need were all using modern contraceptive methods. • Eight in 10 of the lives saved would be in low-income countries. • Providing contraceptive services to all women with unmet need for modern methods would result in sav- ings of 27 million DALYs, 9 million among women and 18 million among infants and children. • Three-quarters of the DALYs saved would be to women and children in low-income countries. Providing contraceptives to all women with unmet need would cost an average of $19 per user per year, providing substantial benefits in terms of saving lives and preventing years of disability. It would cost an av- erage of $74 for each pregnancy averted. • Each life saved would cost an average of $2,500. • For every $144 invested in sexual and reproductive health services in developing countries, one DALY would be saved. The Alan Guttmacher Institute 48 Table 3.1 Estimated number and percent of women with unmet need and using no contraceptive, 2003 Estimates from Estimates from this report (2003) Ross & Winfrey (2000)* Currently Married Women Number of women 15–49 (000s) 649,000 615, 234 Unmet need, using no method 111,676 105,205 Spacing 57,962 55,402 Limiting 53,714 49,803 Percentage of women 15–49 Unmet need 17.2 17.1 Spacing 8.9 9.0 Limiting 8.3 8.1 Unmarried Women Number of women 15–49 (000s) 312,540 263,813 Unmet need, using no method 11,065 8,442 Spacing 8,763 na Limiting 2,272 na Percentage of women 15–49 Unmet need 3.5 3.2 Spacing 2.8 na Limiting 0.7 na ∗ Estimates for year 2000: distribution by use for 1998 applied to 2000 population for married women of reproductive age for developing countries as reported in United Nations, Levels and Trends of Contraceptive Use as Assessed in 1998, New York: UN, 2000. Assessing Costs and Benefits 49 Table 3.2 Married women of reproductive age Estimates for 2003 Estimates for year 2000* Total (millions) 905.4 873.2 Number using a method (millions) 530 519 Percentage using method: 58.5% 59.4% Female Sterilization 22% 23% Vasectomy 4% 3% Pill 6% 6% IUD 14% 15% Injectables/Implants 3% 3% Condom 3% 3% Other supply methods <0.5% <0.5% Traditional methods 6% 6% ∗ In this report, “married” includes women in consensual unions. The Alan Guttmacher Institute 50 Table 3.3 Number of unintended pregnancies by region Number of unintended pregnancies (000s) Region Before adjustment After adjustment Adjustment factor Total 82,063 75,922 0.925 Africa 17,224 15,781 0.916 China 13,040 14,965 1.148 Other East Asia 953 773 0.811 Rest of Asia 41,339 32,775 0.793 Latin American and the Caribbean 9,355 11,399 1.218 Oceania 153 230 1.503 Assessing Costs and Benefits 51 Table 3.4 Estimated distribution of unplanned pregnancies according to outcome155 Unplanned births Induced abortions Spontaneous abortions Total All developing countries 41% 46% 13% 100% Africa 54% 32% 14% 100% Eastern Africa 52% 34% 14% 100% Middle Africa 54% 32% 14% 100% Southern Africa 62% 23% 15% 100% Western Africa 52% 34% 14% 100% Northern Africa 59% 27% 14% 100% Asia 36% 52% 12% 100% East Asia-China 24% 65% 11% 100% China 18% 71% 11% 100% South Central Asia 48% 39% 13% 100% Southeast Asia 26% 62% 12% 100% Oceania 63% 22% 15% 100% Western Asia 51% 35% 14% 100% Latin America and the Caribbean 49% 38% 13% 100% Caribbean 37% 50% 13% 100% Central America 55% 31% 14% 100% South America 48% 39% 13% 100% The Alan Guttmacher Institute 52 Clearly, there already exists a substantial body of work on the costs and benefits of interventions to improve health conditions worldwide. Researchers across dis- ciplines have developed a number of different ap- proaches and methodological techniques to assess and quantify these costs and benefits. Within the particular perspective of each field and within the limits of what can be quantified and measured, this existing body of work has provided useful information and guidance to policymakers regarding the relative value of invest- ments in health care services. It has also demonstrated a high level of awareness that the merits and impor- tance of such services, when measured against their costs, must be proven and cannot be taken for granted. Many of the existing studies are deficient in one im- portant respect: They fail to acknowledge the nonmed- ical benefits of reproductive health care services, with very few exceptions.169 These nonmedical benefits may be usefully categorized into three aspects: per- sonal, social and economic. Though these benefits are often unacknowledged and extremely hard to quantify, they represent a large and important component of the potential gains from services to improve sexual and re- productive health; these gains are additional to the kind of benefits that are typically measured—reductions in mortality and morbidity. One of the best known of these cost-benefit ap- proaches—the Disability Adjusted Life Years (DALYs) quantification of the burden on societies and individu- als imposed by disease and ill-health coupled with the costing of specific health interventions—has come under some criticism.170 Suggestions have been made both by those who developed DALYs and by others that the definition of ill-health that underlies the esti- mation of DALYs—disability in terms of physical functioning and survival—should be substantially ex- panded. Moreover, various efforts have been made to do so, as summarized in Chapter 2. In addition, the World Health Organization (WHO) has proposed al- ternative and much broader classification systems. One involves the measurement of the contribution of health care services to improved well-being in eight areas of an individual’s life: physical functioning, physical roles, emotional roles, social functioning, mental health, general health perceptions, bodily pain, and vi- tality. Some researchers have pointed out that even this much broader WHO system needs further expansion to include nonmedical benefits in additional areas of ex- perience, such as reduction in an individual’s level of shame, embarrassment, stigma and fear as a result of preventing or reducing the severity of disease and ill- health.171 Whatever the limitations of current measurements and methodologies in this area, the earlier cost-benefit efforts provide a good starting point and a strong basis to build on. However, to advance cost-benefit analysis in the area of sexual and reproductive health, some fur- ther steps are necessary. One is the development of a comprehensive outline of the medical and nonmedical benefits of sexual and reproductive health interventions which would serve to clarify gaps in coverage of exist- ing studies and approaches in representing the benefits from sexual and reproductive health interventions, and also to make clear whether such approaches and stud- ies are comparable and which of them come closer to providing a more comprehensive assessment of bene- fits. At the same time, it may also stimulate and guide new research in this area. In this chapter we outline the expanded medical and nonmedical benefits—in some cases nonquantifiable or hard to quantify, direct and indirect—that could be hypothesized or expected to result from health inter- ventions in each of the three main areas of reproductive health care: contraceptive services; services related to sexually transmitted infections (STIs) and HIV/AIDS, and gynecologic and urologic services; and pregnancy- related care. To do so, we synthesize and expand upon the work of other researchers. We then comment briefly on the need for greater uniformity or compara- bility in approaches for estimating the costs involved in Chapter 4 A Broader Approach to Measuring Benefits and Costs 53 the provision of sexual and reproductive health care services in general. Benefits of Sexual and Reproductive Health Services For each of the three main areas, services are defined to include information, education and counseling, as well as medical or clinical services. Clinical services comprise preventive, diagnostic and treatment meas- ures. For convenience, we include gynecologic and urologic services with STI-related care, though it can arguably be in a category of its own. The principal components of care in each of these three groups are outlined in general terms in Table 4.1. The evidence base for demonstrating the effect of health interventions on medical or health outcomes varies across the different areas of sexual and repro- ductive health but is not extensive for any of the three main areas. Probably the most extensive effort to meas- ure the impact of health interventions on outcomes was carried out in the early 1990s by the World Bank, in which the cost-effectiveness of 47 health interventions, spanning all areas of health, was investigated.172 As mentioned in Chapter 2, this study found that interven- tions in all of the three main areas of sexual and repro- ductive health (family planning, maternal care and pre- vention and treatment of STIs and HIV/AIDS) were cost effective. A new assessment of effectiveness of in- terventions and priority needs in regard to disease con- trol is currently underway under the World Bank’s coordination.* Others have developed theoretical models and ap- proaches to measure the relationship between contra- ceptive use and nonmedical outcomes for women and families. These approaches have begun to be imple- mented with empirical research and have identified some significant effects of contraceptive use on a num- ber of aspects of women’s lives—physical health, edu- cation, work, income, self-esteem, decision making and role in the family and community.173 Examples from this body of work are cited below, as relevant. Health benefits from sexual and reproductive services Contraceptive use and family planning services. The use of contraception can affect the number, timing and spacing of pregnancies and births, and thereby may di- rectly benefit women’s and infants’health in a number of respects (see Table 4.2). • Contraception may be used to lengthen the inter- val between births and to prevent unwanted pregnan- cies and unwanted births. A longer interval between births (3–5 years) is associated with a number of health benefits for women and for infants. For example, stud- ies find a strong negative relationship between spacing of births and the infant mortality rate: the longer the av- erage birth interval, the lower the infant mortality rate.174 Longer intervals also decrease the impact of pregnancy on women’s health. • By enabling control of timing of pregnancy, con- traceptive use can prevent high risk births and improve timing to minimize risk (those at very young and older ages, those to women who have already had many births and those to women suffering from preexisting medical conditions). • The use of contraception can prevent unwanted pregnancies and abortions, including some that are un- safe, and thereby prevent short-and long-term health impacts on women. • Contraception also prevents unwanted births, which leads to direct health benefits for women and in- fants. Maternal and infant deaths and ill-health due to maternal causes are prevented in proportion to the prevalence of contraceptive use. STI-related and gynecologic and urological services. Services (including information and counseling) to prevent and treat STIs, including HIV/AIDS, and to monitor gynecologic and urologic health, can increase protective behaviors and thereby prevent infections, re- duce the duration of infection, cure bacterial STIs and reduce the severity of symptoms of viral STIs, includ- ing HIV, achieve early diagnosis and increase the like- lihood of curing various conditions and cancers (see Table 4.2). Specific health benefits are listed below include: • Prevention of STIs and HIV can bring large re- ductions in ill-health and in deaths (particularly due to HIV/AIDS). • Prevention and treatment reduce transmission of infections from mothers to infants and among sexual partners.175 • Prevention and treatment of gonorrhea would re- duce the prevalence of its consequences, which include septicemia, arthritis and endocarditis in men, and eye infections and possible blindness in newborns deliv- ered by women with the infection. • Prevention and treatment of STIs lower the preva- lence of pelvic inflammatory disease and reduce infer- tility among women. • Prevention and treatment of human papilloma viruses (HPV) reduce prevalence of genital warts and* See Chapter 2, section on DCP-2. The Alan Guttmacher Institute 54 cervical cancer. In fact, sexually transmitted strains of HPV have a role in the development of most of the half million cases of cervical cancer that occur each year— 65% of the cases in developed countries and 87% of those in developing countries.176 • Appropriate antiretroviral drug treatment of those who are HIV positive can yield significant gains in years of productive life. • Gynecologic and urologic health care can improve prevention and enable early diagnosis and treatment of several conditions and illnesses including: cervical can- cer, breast cancer, prostate cancer, endometriosis, fi- broids and ovarian tumors; reproductive tract infections (including pelvic inflammatory disease, urinary tract in- fections, genito-urinary tract infections and vaginal in- fections); disorders of the reproductive system (e.g., menstrual disorders); and sexual dysfunction (vaginis- mus, dyspariunia and erectile dysfunction).177 Maternal health services. Benefits from these services include those experienced by women themselves, those that apply to infants in the perinatal period, some that occur before and others that occur after the childbear- ing years (Table 4.2). Perinatal outcomes are those that occur in the late fetal period (28 weeks of gestation or later) and in the first month of life. The postpartum pe- riod is generally defined to be the first 42 days after de- livery. Pregnancy-related care yields health benefits in a number of ways: • Prenatal care provides education and counseling on healthy behaviors, especially with regard to diet and nutrition during pregnancy which can benefit women’s and infants’health, even in low-income settings. Mon- itoring health during pregnancy can lead to prompt in- tervention in case of complications and provide the op- portunity for ongoing management of such conditions as hypertension. • Obstetric care provides the means of treating seri- ous complications that occur during delivery and the postpartum period, thereby reducing long-term or chronic pregnancy-related sequelae. Obstetric care can reduce the probability and severity of delivery-related conditions, such as hemorrhage and sepsis, and also of the more chronic consequences, such as obstetric fis- tula, urinary or fecal incontinence, scarred uterus and pelvic inflammatory disease.178 • Pregnant women with certain preexisting condi- tions and diseases may experience increased risks of mortality and morbidity because these conditions are worsened due to the physiological effects of pregnancy. Examples of such conditions include anemia, malaria, hepatitis, tuberculosis and cardiovascular disease. With appropriate medical care during pregnancy, such health complications can be managed and minimized. • Medical care for complications due to unsafe abor- tion can reduce mortality and the extent and severity of morbidity. Unsafe abortion accounts for an estimated one-seventh of total maternal mortality in the develop- ing world.179 Several of the benefits described above involve in- tegration of services—for example, when regular ante- natal care service is combined with treatment of malar- ia or management of hypertension. The costing of the separate interventions would have to take into account the synergistic nature of such benefits to assure that costs are accurately assigned to the resulting benefits. Nonmedical benefits of sexual and reproductive health services There is also a range of nonmedical benefits—person- al, social and economic—that can result from use of contraceptive services, maternal health services, STI- related services and other gynecologic and urologic care. Many of these benefits have tremendous direct value in themselves, particularly for individuals and households. Even more compelling for policymakers are the benefits at the societal level and the contribu- tions of these interventions to a range of development goals. Contraceptive use and family planning services • Personal benefits. Women who give birth when they want and who have the number of children they want are likely to experience a number of personal benefits that can lead to gains for their household and for soci- ety as a whole (Table 4.3). A first birth at a young age (younger than 20) can limit a woman’s prospects for education,180 training and employment, and, in the longer term, her earning power and financial security. Although this has not been shown to be a causal relationship,181 the unifor- mity and strength of the association worldwide sug- gests that delaying motherhood through the use of con- traceptive services is likely to be an important factor contributing to women’s achieving higher education. More education and work experience would also in- crease women’s status and improve their decision mak- ing role within the family. For example contraceptive users in the Philippines are more likely than nonusers to join their husbands in making household deci- sions.182 A woman’s increased ability to take advan- tage of opportunities that can improve her life (school- ing, work, etc.) is also likely to result in increased personal income, and the household’s economic status is likely to improve. Assessing Costs and Benefits 55 Contraception enables couples to have fewer chil- dren. The benefits of small families in reducing pover- ty at the household level and in improving educational and health outcomes for children have been supported by empirical studies.183 In smaller as compared with larger families, more resources tend to be devoted to each child. There is less differential spending (often gender-based) among children and hence less gender discrimination within households. One study in Ghana found that children, particularly girls, in larger families were less likely to attend school and experienced greater inequality within the household than those with fewer siblings.184 By contributing to smaller families, contraception may contribute to female education and equality. Contraceptive use can improve women’s self-effi- cacy, confidence, satisfaction with life, self-esteem and decision making autonomy. Some of these benefits re- sult from the greater involvement of women in the labor force, their ability to earn an income and their fi- nancial contributions to the household, benefits that are partly a result of greater control over fertility. Contra- ception can also bring about improvements in the qual- ity of life of the family as a whole by reducing stress and worry about unwanted pregnancies. By giving cou- ples greater freedom from fear of unplanned pregnan- cy, family planning can also improve relationships be- tween spouses. Women of all ages report that using contraceptives to time births and avoid unintended pregnancies im- proves their personal well-being and status in the household. Qualitative research shows that contracep- tive use reduces stress about the risk of unplanned preg- nancies and improves relationships between partners. In Bolivia, for example, women using contraceptives demonstrate greater self-esteem than nonusers, and in the Philippines, contraceptive users have reported greater overall satisfaction with their lives than nonusers. Women point out that delayed childbearing and smaller families, which are achieved through con- traceptive use, allow more leisure time as well as edu- cational and economic opportunities.185 Men and women who are healthy and have fewer children to care for have more time for other things, in- cluding civic activities. In a survey of older married women in two urban areas of Indonesia, half the women reported that family planning enabled them to spend more time in community activities.186 An in- creased sense of power and confidence conferred by family planning may help women take a more active role in community and political life. Contraception and smaller families may have a cumulative effect: As women gain time and freedom to become involved in social and political issues, they increasingly advocate for and take advantage of contraceptive services.187 • Economic benefits at the personal and household levels. As family planning increases women’s partici- pation in work and other income-generating activities, it leads to increased financial security for women and increases household income. Women are able to work longer hours and for a greater proportion of their pro- ductive lives. With increased health and functioning (given the decrease in morbidity from prevented high- risk births and unsafe abortions and the increase in psy- chological well-being) women are likely to enjoy in- creased productivity. Benefits also accrue to children in terms of future productivity: In smaller families par- ents are better able to feed and house their children, provide health care and make more intensive inputs into their physical, social and intellectual development; school participation of children (male and female) is also likely to be higher when families are smaller. Also, as shown in Chapter 3, reducing unwanted pregnancies would substantially reduce the number of children who lose their mothers through pregnancy-related deaths. Better health and smaller household workloads may lead to new economic possibilities and allow more op- portunities for productive investments for poor women. For example, a study of women in Kenya found that the main reasons for late repayment of microcredit loans included expenses related to having a large family, in- cluding diversion of funds to pay for medicines, school fees or

View the publication