WORLD SANK TECHNICAL PAPER NO s33 WVf 333 Work In progress 'l t 9 for public discussion i Measuring the Burden of Disease and the Cost-Effectiveness of Health Interventions ,A Case Study in Guinea _ -, A01, , 4 Puzabhatjba Kent Ranson Josd Luis Bobadilla RECENT WORLD BANK TECHNICAL PAPERS No. 256 Campbell, Design andt Operation of Smalllholder Irrigation in Souith Asia No. 258 De Geyndt, Managing the Quality of Health Care in Developing Colunitries No. 259 Chaudry, Reid, and Malik, editors, Civil Service Reform in Latin America and the Caribbean: Proceedings of a Conference No. 260 Humphrey, Paymenit Systems: Principles, Practice, and Improvements No. 261 Lynch, Provisionifor Chiildrenz wvitlh Special Eduz cational Needs in the Asia Region No. 262 Lee and Bobadilla, Health Statistics for the Americas No. 263 Le Moigne, Subramanian, Xie, and Giltner, editors, A Guide to the Fornmu lation of Water Resources Strategy No. 264 Miller and Jones, Organic and Compost-Based Growing Media for Tree Seedling Nurseries No. 265 Viswanath, Bu ilding Partnerslhipsfor Poverty Reduction: The Participatori Project Planniing Approaclh of the Women's Enterprise Management Training Ouitreaclh Program (WEMTOP) No. 266 Hill and Bender, Developing tlhe Reguilatory Enivironmentfor Competitive Agrictultural Markets No. 267 Valdes and Schaeffer, Surveillance of Agricuiltu ral Prices and Trade: A Handbook for the Dominic an Republic No. 268 Valdes and Schaeffer, Suirveillance of Agricultural Prices and Trade: A Hlndbookfor Colomnbia No. 269 Scheierling, Overcominig Agricuiltuiral Polluition of Water: The Clhallenge of Integrating Agricultural and Environniental Policies in the European Union No. 270 Banerjee, Rehabilitation of Degradtedi Forests in Asita No. 271 Ahmed, Technological Development antd Polluition Abatement: A Stu dy of HoL7w Enterprises Are Finding Alternatives to Chllorofluorocarbonis No. 272 Greaney and Kellaghan, Equiity Iss,ues in Pu blic Exanlinations in Developing Countries No. 273 Grimshaw and Helfer, editors, Vetiver Grassfor Soil andi Water Conservation, Lanid Rehlabilitation, and EmbankAment Stabilization: A Collection of Papers atnd Newsletters Comipiled by the Vetiver Network No. 274 Govindaraj, Murray, and Chellaraj, Healtht Expendituires in Latin Anmerica No. 275 Heggie, Managemenit and Financing of Roads: An Agenda for RefoLrm No. 276 Johnson, Quality Review Sch)emes.for Auditors: Their Potenitallfor Sub-SahUran Africa No. 277 Convery, Applying Environmental Economnics in Africa No. 278 Wijetilleke and Karunaratne, Air Quality Mvlanagemnent: Consider7tions for Developing COUntries No. 279 Anderson and Ahmed, The Case for Solar E nergtj Investments No. 280 Rowat, Malik, and Dakolias, Judiiciazl Refornm in Latin Amnerica and the Caribbean: Proceedings of a World Bank Conference No. 281 Shen and Contreras-Hermosilla, Environnmental and Economonic Issues in Forestry: Selecte d Case Studies in Asia No. 282 Kim and Benton, Cost-Benefit AnaltYsis of the Oniclhocercinsis Control Progranm (OCP) No. 283 Jacobsen, Scobie and Duncan, Statuitory Intervention in AgricUltuiral Marketing: A New 2 Zeailtand Perspective No. 284 Valdes and Schaeffer in collaboration with Roldos and Chiara, Survneillance of Agricultural Price and Trade Policies: A Handbookfor Ulruguai! No. 285 Brehm and Castro, Thle Market for Water Righits in Chlile: Major Issues No. 286 Tavoulareas and Charpentier, Clean Coal Technologiesbor Developing Couzntries No. 287 Gillham, Bell, Arin, Nlatthews, Rumeur, and Hearn, Cotton Production Prospects for the Next Decade No. 288 Biggs, Shaw, and Srivastiva, Teclhtnologicall Capabilities andti Learning in] African Enterprises No. 289 Dinar, Seidl, Olem, Jorden, Duda, and Johnson, Restoring and Protecting the World's Lakes andii Reservoirs No. 290 Weijenberg, Dagg, Kampen Kalunda, Nlailu, Ketema, Navarro, and Abdli Noor, Strengtlheninlg National Agricultual Researclh Systemiis in Eastern and Centmrl Africa: A Franmewvorkfor Action No. 291 Vald6s and Schaeffer in collaboration with Errazuriz and Francisco, Surueillance of Agricuiltuiral Price and Trade Policies: A Handlbookfor Clhile No. 292 Gorriz, Subramanian, and Simas, Irrigation Managenment Transfer in Mevexico: Process and Progress No. 293 Preker and Feachem, Market Mechanisnms and tlhe Health Sector in Centrazl andi Eastern Eiurope No. 294 Valdes and Schaeffer in collaboration with Sturzenegger and Bebczuk, Suirvneillantce of Agricultural Price anId Trade Policies: A Handbookfor Argentintn (List continues on the inside back cover) WORLD BANK TECHNICAL PAPER NO. 333 Measuring the Burden of Disease and the Cost-Effectiveness of Health Interventions A Case Study in Guinea PrabhatJha Kent Ranson JostLuis Bobadilla The World Bank Washington, D. C. Copyright @ 1996 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing September 1996 Technical Papers are published to communicate the results of the Bank's work to the development community with the least possible delay. The typescript of this paper therefore has not been prepared in accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibility for errors. Some sources cited in this paper may be informal documents that are not readily available. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility whatsoever for any consequence of their use. The boundaries, colors, denominations, and other information shown on any map in this volume do not imply on the part of the World Bank Group any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. Requests for permission to reproduce portions of it should be sent to the Office of the Publisher at the address shown in the copyright notice above. The World Bank encourages dissemination of its work and will normally give permission promptly and, when the reproduction is for noncommercial purposes, without asking a fee. Permission to copy portions for classroom use is granted through the Copyright Clearance Center, Inc., Suite 910, 222 Rosewood Drive, Danvers, Massachusetts 01923, U.S.A. The complete backlist of publications from the World Bank is shown in the annual Index of Publications, which contains an alphabetical title list (with full ordering information) and indexes of subjects, authors, and countries and regions. The latest edition is available free of charge from the Distribution Unit, Office of the Publisher, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433, U.S.A., or from Publica- tions, The World Bank, 66, avenue d'Ina, 75116 Paris, France. ISSN: 0253-7494 Cover photo: Health worker and baby. Credit: World Bank/Curt Carnemark. Prabhat Jha is a public health specialist and Kent Ranson, a consultant in the World Bank's Human Development Department. Jose Luis Bobadilla is a senior public health specialist in the Inter-American Development Bank and was formerly a senior public health specialist in the World Bank. Library of Congress Cataloging-in-Publication Data Jha, Prabhat, 1965- Measuring the burden of disease and the cost-effectiveness of health interventions : a case study in Guinea / Prabhat Jha, Kent Ranson, and Jose Luis Bobadilla. p. cm. - (World Bank technical paper ; no. 333) Includes bibliographical references. ISBN 0-8213-3727-0 1. Medical economics-Guinea. 2. Medical care-Guinea-Cost effectiveness. I. Ranson, Kent, 1970- . II. Bobadilla, Jose Luis. 1I. Title. IV. Series. RA410.55.G92J48 1996 338.4'33621'096652-dc2O 96-31866 CIP TABLE OF CONTENTS Foreword . . . . . . . . v Abstract . . . . . . . . . . vii Acknowledgments . . . . . . . . ix PURPOSE OF STUDY . . . . . 1 DEMOGRAPHIC AND HEALTH SYSTEM PROFILE OF GUINEA. . . 3 MEASURING MORTALITY AND DISABILITY IN GUINEA . 7 DATA SOURCES . . . . . . . 7 METHOD FOR DERIVING MORTALITY AND MORBIDITY ESTIMATES . 8 RESULTS FOR BURDEN OF DISEASE . . . . . . 10 CONCLUSIONS AND POLICY IMPLICATIONS OF DISEASE BURDEN . 18 COST-EFFECTIVENESS OF HEALTH INTERVENTIONS IN GUINEA . 21 METHODOLOGY FOR COST-EFFECTIVENESS ANALYSIS . . . 21 RESULTS OF COST-EFFECTIVENESS ANALYSES OF INTERVENTIONS . 23 A MINIMUM PACKAGE OF HEALTH SERVICES . 25 POLICY IMPLICATIONS OF COST-EFFECTIVENESS ANALYSIS . . 29 BIBLIOGRAPHY . . . . . . . . . 33 . . TABLES Table Number Page 1 Health expenditures in Guinea from 1988 to 1994 (in millions of Guinean francs), and by level of care (as a percentage of total health expenditures) 6 2 Distribution of population and number of deaths, in Guinea by sex and age, 1992 . . . . . 11 3 Distribution of deaths by cause in Guinea, by sex 1992 . . 11 4 Top ten causes of death in Guinea, by age, 1992 (ranked in order and percentage except as noted) . . . 13 5 Distribution of years of life lost in Guinea, by sex and cause, 1992 . . 14 6 Top ten causes of years of life lost in Guinea, by age, 1992 (ranked in order and percentage except as noted) . . . . 14 7 Distribution of years of life lived with a disability in Guinea, 1992 by cause and sex . . . . . . 15 8 Relative importance of causes of death in Guinea and Ghana. . . 20 9 Summary of population, coverage, and cost-effectiveness for forty health interventions in Guinea, 1994 . . . 24 10 Health interventions in Guinea, ranked by cost-effectiveness . . 26 11 Cost-effectiveness of health services in Guinea, by level of care and future costs from expanded coverage of health services (assuming no improvement in quality of care) . . . 27 12 Total and per capita costs of a minimum package of health services in Guinea (assuming no improvement in quality of care) . . 28 13 Contributions of burden of disease and cost-effectiveness analyses to evaluation of health sector performance . . . . . 31 FIGURES Figure . . . . . . . . . Page 1 Rapid method to calculate mortality data for Guinea . . . 9 2 Distribution of deaths, by major cause and age in Guinea, 1992 . 13 3 Relative importance of causes of death in Guinea, sub-Saharan Africa, developing and developed countries . . . . . 16 4 Projections of number of deaths in Guinea, by cause . . . 17 5 Distribution of deaths by age group in Guinea, 1992, 2000, and 2010 . 18 iv FOREWORD Health care investments in low-income countries such as Guinea are crucial to development. The choice of allocating spending to various health interventions is a complex one. This report draws upon the methodology of the 1993 World Development Report: Investing in Health to offer a set of guidelines for the rational allocation of scarce health resources. Analytic efforts, such as this one, can help guide policy choices towards viable and cost-effective health care interventions. This study draws upon local data from Guinea, and represents a useful addition to our knowledge of disease burden and cost-effectiveness in low-income African countries. As a fully collaborative effort with the Ministry of Health in Guinea, this study represents an important aspect of the non-lending policy development work of the World Bank. David de Ferranti Director Human Development Department v ABSTRACT This study draws upon the methodology of the 1993 World Development Report. Investing in Health to provide analyses of the burden of disease and cost-effectiveness of health interventions. The analysis aims to provide a common framework for the activities of the government of Guinea in prioritizing health care services. The burden of disease analysis presents detailed estimates of mortality in Guinea, by cause of death in 1992, using a new, rapid methodology. The cost-effectiveness analysis compares the costs of forty interventions selected from treatment protocols at health centers, health posts and first referral hospitals' health care intervention with their health impact. These two methodologies complement other methods to measure health system performance. The results of the burden of disease study underscore the need for national health care planners and international donors to focus on diseases that cause the most premature deaths. In 1992, the largely avoidable or treatable communicable disease, maternal, and malnutrition diseases together accounted for 86,400 deaths or seventy percent of all deaths and, with perinatal disease, account for nine in ten deaths before age five. Injuries caused 9,000 deaths (seven percent) and non communicable diseases caused 28,700 deaths (twenty-three percent). Many deaths among the economically-productive adult age group could be also avoided. Detailed data on years of life lost, simplified estimates of years lived with a disability, a comparison to other regions, and estimates of the impact of population growth on burden of disease support these findings. The cost-effectiveness analysis suggests that several affordable interventions already exist for the major causes of death in Guinea. Communicable, maternal, and perinatal diseases, the largest causes of death in Guinea, are subject to several highly cost-effective curative interventions. Those costing less than about one hundred dollars per year of life saved include treatment of childhood pneumonia, rehydration therapy for diarrhea at health centers and at health posts, treatnent of childhood malaria at health centers, childhood vaccination at health centers, short- course tuberculosis treatment at health centers, treatment of sexually-transmitted diseases at health centers, and impregnated bed nets against malaria. Maternal and perinatal diseases have slightly less cost-effective interventions, involving prenatal and delivery care at health centers or programs using out reach trained birth attendants. The report concludes with one possible minimum package of health services that includes outreach preventive programs and a package of curative treatments at health center and hospital levels. The total package of services would cost about thirteen dollars per capita and would address about sixty percent of total years of life lost. The analysis confirms that provision of a basic package of health services will require substantial increases in funding from the government of Guinea and donors. The minimum package is about three times the public per capita expenditure of four dollars. The core set of interventions at the outreach, health center, and first referral hospital level, should be the priority for the government of Guinea as it plans further health system reforms. vii I ACKNOWVLEDGMENTS The authors thank the many individuals whose input helped complete this study. Sergiu Luculescu was chiefly responsible for the initiation of the work, and supervised its early stages. Numerous officials in the Ministry of Health, Republic of Guinea, provided data for the study, notably Dr. Ousmame Bangoura, Secretary-General, and Dr. Robert Tolno. The Ministry of Health held two review meetings for the study in Guinea. Dr. Alan Lefevre was the consultant for the study. Marisol Ravicz provided the templates for the cost-effectiveness analysis. Pascale Kervyn provided useful information from the public expenditure review. Mariari Claeson, Peter Cowley, Ed Elmendorf, Yves Genevier, Salim Habayeb, Dean Jamison, Varsha Malhotra, David Peters, Alexander Preker and Helen Saxenian gave constructive comments. Alberto Harth chaired a review meeting for the study within the World Bank. Mendy Woodly and Lauren Ptito helped edit the report. The authors extend their special thanks to Ok Pannenborg, Technical Manager, Africa Technical Human Development 2, Western Africa Population and Human Resources, and to Richard Feachem, Senior Adviser, Human Development Departrnent, for their enthusiastic and continuous support for the study. ix PURPOSE OF STUDY Like most other developing countries, the Republic of Guinea faces difficult choices in allocating limited financial resources to its people's health needs. Making these choices depends on accurate information about the country's health status and on the effectiveness of various interventions the country might undertake to improve health status. The 1993 World Development Report of the World Bank, which focused upon investments in health (World Bank, 1993b), outlined two analytic steps to guide allocation of scarce health care resources in low and middle income countries. The first was the application of burden of disease measures to obtain estimates of deaths and disability from various diseases. The second was cost-effectiveness analysis to identify health interventions that make best use of money in reducing mortality and morbidity. The report suggested that addressing diseases with a high burden with the most cost-effective interventions could do much to reduce disease in the population. To yield meaningful results, burden of disease and cost-effectiveness studies should be made at a country or sub-country level. This is because large variation exists in epidemiological conditions, health services, and unit costs across countries. This report presents the results of studies in Guinea based on the World Development Report's methodology, the outcome of discussion between Guinea's Ministry of Health and staff in the World Bank. While these analyses offer technical information to guide policy, actual choices must also take into account other factors: the feasibility of the interventions; institutional capacity; political support; and the results of macroeconomic, financial, and operational research. As well, consumer preferences for various interventions are important. Moreover, policy choices are influenced by the way the existing system of health care delivery works, including the allocation of investment and recurrent spending. The analyses had five objectives: (a) to provide a common framework for the government of Guinea in prioritizing health care services. Prioritization includes both current health expenditures and planned external aid from the World Bank and other international donors; (b) to provide a framework for similar analyses at regional levels within Guinea; (c) to permit a review of current health programs in the context of disease burdens and cost-effectiveness; (d) to create a set of baseline data and methods that could become a template for similar analyses in other low-income African countries; and (e) to create a data base that would permit selected inputs into economic analyses of projects in low-income African countries. I To achieve these objectives, the Population and Human Resources Division of the Western Africa Department, and the Human Development Department of the World Bank in conjunction with the Ministry of Health of Guinea undertook a study of the burden of disease and cost-effectiveness of various interventions. The study began with preliminary discussions in March 1994, followed by a mission by Bank staff in November 1994 to collect detailed data. Preliminary estimates were completed in March 1995 and shared with counterparts in Guinea. A second mission in July 1995 validated detailed assumptions. It also included a consultant who verified data sources and numbers for the cost-effectiveness analysis.' ' The results we present here for burden of disease, and cost-effectiveness analyses are the major results only. To obtain an annex containing more detailed data, or computer files of the analyses, please contact Prabhat lha in the Human Development Department of the World Bank. 2 DEMOGRAPHIC AND HEALTH SYSTEM PROFILE OF GUTINEA Guinea is located in West Africa, north of the Ivory Coast. It is one of the poorest nations in Africa with a gross domestic product of $460 (US dollars2) per person per year (World Bank, 1993b). Moreover, Guineans suffer among the worst health indicators in the world. Life expectancy at birth is only forty-four years and newborns face a one in four chance of dying before age five. In 1993, according to government surveys, infant mortality was about one- hundred thirty-three deaths per thousand live births and maternal mortality was eight-hundred per hundred thousand live births. Both these measures are among the worst in Africa. The population of Guinea was 6.6 million in 1995 and was 6.1 million in 1992, the year upon which we base the burden of disease analysis (World Bank, 1993a). Guinea's population is young: forty-seven percent of its people are under fifteen and only four percent are sixty or older. World Bank projections suggest children under fifteen will constitute over forty percent of the population until the year 2025. Life expectancy will increase only slightly, to forty-six years by 2005. Women constitute fifty-one percent of the population, because of a slightly higher mortality among male children. The sex ratio is generally constant over different age groups. Approximately seventy-two percent of Guinea's population live in rural areas (DNSI, 1994). Apart from having less access to health services than urban inhabitants, only half of those in rural areas have access to drinkable water compared with almost three-quarters of urban residents; and only one percent have access to basic sanitation facilities compared with nearly a fifth of urban residents. About twenty-eight percent of the population live in Maritime Guinea and one million in Conakry, the capital city. Middle Guinea, Upper Guinea, and Forest Guinea constitute thirty-two, twenty-one, and twenty percent, respectively of the total population. International migration is responsible for 0.3 percent to 0.6 percent net migration out of the country. The Ministry of Health is responsible for organization and delivery of health services in Guinea. An informal system of private and traditional practitioners receives about half of per capita health spending in Guinea. However, the major infrastructure of health facilities is in the public system. The public system relies upon thirty-three prefectures containing three-hundred and seventy-six centre de sante or health centers and three-hundred and forty-seven poste-de sante or health posts. There are thirty-three prefectoral and four regional hospitals and two teaching hospitals in Conakry. A 1993 inventory study by a consulting firm found two-hundred and sixty- nine health centers were part of the Extended Program of Immunization, Basic Care and Essential Medicines, (known as Programme Extendee Vaccination, here referred to as the extended program). Twenty-one health centers were not active due to construction or renovation, and eighty- six were not integrated into the extended program. This study also found one-hundred and fourteen of the health posts were not active and only thirty-five were part of extended program. The activities of the health centers are primary care, infant health (immunization, growth monitoring and nutrition), normal maternal labor (in rural areas; in urban areas, hospitals largely perform 2 All dollar figures in this report are in 1994 US dollars. The conversion rate used in the report is 976 Guinean francs for every US dollar. 3 maternal care), family planning, minor surgery, simple diagnostic tests, sale of essential medications, dental care (in certain urban centers), and provision of information. Health posts promote health education, monitor the growth of infants and progress of pregnant women, offer limited primary care, sell essential medications, and make referrals to the health centers. Both health centers and health posts follow written protocols for presentations of common causes of illness (SEATS/JSI, 1994). According to the government and the World Bank, most of Guinea's health centers and hospitals lack drugs and operating facilities, good management and trained staff, so that even where health services are available, they are inadequate and poorly utilized (World Bank, 1986 and 1993a; SEATS/JSI, 1994). Two other key problems include weak sector organization and institutional capacity and financial constraints and inefficient use of resources (World Bank, 1993a). In keeping with the international declaration of Alma-Ata in 1978, the government aims to make health services accessible to the largest number of people consistent with reasonable costs to the country. The focus is on the implementation of primary health care, particularly through the expanded delivery of immunization and essential drugs, and rehabilitation of physical structures and equipment at health centers and hospitals. The government has set a goal of activating all health centers by the end of 1996. The average public spending on health in Guinea from 1991 to 1994 was slightly under four dollars per capita in real terms in 1994 US dollars (table 1). A 1991 household survey conducted by an international aid agency suggested private expenditures were about double this amount per capita. Of the private spending, the public system realized a cost-recovery of only ten to twenty percent. Thus, the actual total per capita spending of about eight dollars falls below the level of twelve dollars suggested by the 1993 World Development Report for a minimum package of health services (World Bank, 1993b). A recent review of public expenditures in Guinea showed health expenditures declined in real terms from 1988 to 1994 (World Bank, 1996). On a per capita basis, the drop was significant: from seven dollars in 1988 to under two dollars in 1994. Similarly, expenditures dropped from 1.3 percent of gross domestic product in 1988 to 1.0 percent in 1994 -- a performance that compares poorly with similar low-income countries. However, as a share of total public expenditures, health expenditures remained fairly constant at about 4.5 percent. In 1995, the government increased the health recurrent budget from 3.9 percent of the total public budget in 1994 to 5.4 percent in 1995. Despite this increase, the recurrent budget for the sector remains inadequate, exasperated by the large proportion of funds spent on salaries (which reached over ninety percent of the recurrent budget in 1994), and low budget implementation, especially for non-salary expenditures. During the years 1991, 1992 and 19943, expenditures for primary care represented an average of twenty-eight percent of total expenditures on health. Secondary-care and tertiary-care facilities captured twenty-four and twenty-seven percent, respectively, and administration captured twenty-one percent. Three factors are noteworthy in this expenditure pattern. First, 3Data for 1993 budget expencitures were incomplete at the time for this analyses, and thus are excluded. 4 about three-quarters of expenditures at the secondary-care level went to prefecture hospitals that also deliver primary health services. Second, the increase in expenditures for tertiary care in 1994 was only temporary, largely for rehabilitation works performed at the two large hospitals in the capital. Lastly, hospitals included in the tertiary care category performed largely secondary- care tasks. Recurrent non-salary expenditures for primary health care facilities are inadequate. Public expenditures for medicines increased in 1994 to about sixty-four percent of recurrent non- salary expenditures. However, this amounted to less than twenty-five cents per capita, and compares poorly with the average expenditure of two dollars in other sub-Saharan countries (World Bank, 1994). The share remaining for other recurrent expenditures such as the maintenance of infrastructure and vehicles is insufficient. This creates deficiencies in supervision, patient referral, and vehicle operation. Finally, the expenditure on administration in Guinea is higher than the average for sub-Saharan Africa (World Bank, 1994). 5 Table 1. Health expenditures in Guinea from 1988 to 1994 (in millions of Guinean francs), and by level of care (as a percentage of total health expenditures) Year Averagefor 1988 to 1994, excluding 1993 In nominal terms Recurrent 7,600 Investment 8.300 Total health expenditure 15,900 In real terms (1994 values) Total health expenditure 23,300 Percent of total govermment expenditures 4.5 Percent of gross domestic product 0.78 Per capita in constant US dollars 3.9 By level of care, as percentage of total health expenditures Average for 1991-1994, excluding 1993 Primary 28 Recurrent 13 Non-salary 1 Drugs 1 Investment 15 Secondary 24 Recurrent 17 Non-salary 8 Drugs 1 Investment 8 Tertiary 27 Recurrent 9 Non-salary 2 Drugs 1 Investment 18 Administration 21 Recurrent 7 Non-salary 1 Investment 13 Total 100 Recurrent 47 Non-salary 13 Drugs 3 Investment 53 6 MEASURING MORTALITY AND DISABILITY IN GUINEA This section presents detailed estimates of mortality in Guinea, by cause of death in 1992. The methodology relies upon data specific to Guinea and on an external database for sub- Saharan Africa and uses a rapid methodology. Figure 1 shows the various steps in the rapid methodology. The sources of data are given below. DATA SOURCES A) Data on population and mortality for Guinea We obtained numbers of age and sex-specific data on population and mortality in Guinea for 1992 from the World Bank demography database. These population projections are derived from data collected in Guinea in a full census in 1955 and a partial census in 1983. A demographic household survey conducted in 1992 showed a population and age distribution similar to the World Bank estimates (Chi square p value =0.39 and 0.92 for men and women, respectively; DNSI, 1994). B) Data on all possible causes of death We used estimates of the specific causes of death for sub-Saharan Africa from the Global Burden of Disease Project organized by the World Health Organization and the World Bank (known as Version 4 of the GBD; Murray and Lopez, 19944). This consisted of age and sex- specific estimates for disease groupings based on the International Classification of Disease-9th version (ICD-9). The ICD-9 system covers nearly 100 percent of all causes of deaths, classified into three major categories: (1) communicable, matemal and perinatal; (2) injuries; and (3) non- communicable. We maintained the ICD-9 classification for specific diseases as utilized by the Global Burden of Disease Project. We combined all diarrhea diseases into one category, all maternal causes of death into one category, and all trypanosomiasis, schistosomiasis, and onchocerciasis into one category. We separated injuries into unintentional and intentional injuries. We also established four categories consisting of the following: (1) all neoplasms; (2) all nutritional and endocrine diseases; (3) epilepsy, psychosis, alcoholism, dementia, Parkinson's disease, and multiple sclerosis; and (4) rheumatic, ischemic and inflammatory heart disease, and cerebrovascular disease. Finally, we placed components of the respiratory, digestive, and genito-urinary diseases into three broader categories. The details of the methodology for the GBD are described elsewhere (Murray, 1994). C) Local studies of childhood and maternal deaths, HiIV infection, injuries and malnutrition Given the high mortality and high fertility rates in Guinea, we adjusted estimates so as to optimize the quality of data for early childhood (age zero to four) and for maternal deaths. These relied upon: (1) local unpublished surveys of childhood mortality done in 1986 by the U. S. Agency for International Development in rural parts of Kindia province in Guinea; (2) studies of childhood 4These data are consistent with newer Version 6 estimates for mortality in sub-Saharan Africa, although injuries are a higher percentage of deaths in the latter. The rank order and percentage of deaths from the top five causes of death are similar. 7 mortality in Conakry conducted by Cornell University; and (3) maternal mortality estimates in Guinea (data on file, Ministry of Health, Guinea 1994). These data permitted adjustments for deaths from malaria, respiratory illness, malnutrition and diarrhea. Attributing deaths to malnutrition among children is controversial (Pelletier et al., 1993) because the direction of causation may be unclear: malnourished children are at increased risk of acquiring and dying from an infection; conversely, certain infections may lead to malnutrition. Using results from a similar study in Eastern Africa, we assumed that malnutrition was the direct cause in only eighteen percent of all deaths attributed to malnutrition in the AID and Cornell studies (World Bank, 1995). This avoided double-counting of causes of deaths. We used two additional pieces of data to adjust the estimates of deaths. First, we reduced deaths from human immunodeficiency virus (HIV) infection by half to reflect a 1993 finding that, among patients at a sexually-transmitted disease clinic in Conakry, only 2.7 percent were seropositive for the HIV (data on File, Ministry of Health, 1994). This prevalence is significantly lower than that among other high-risk populations in sub-Saharan Africa (US Bureau of the Census, 1994). Second, we reduced estimates of deaths from war by an arbitrary one-quarter because Guinea has been at peace in recent history. METHOD FOR DERIVING MORTALITY AND MORBIDITY ESTIMATES After deriving final estimates for children aged zero to four, we used these to provide estimates for deaths at older ages, with a separate analysis for males and females. Figure 1 shows the details of this approach. Our method ensured that each death had only one cause and those total deaths in any age group did not exceed the data based upon the census information for Guinea. We explicitly assumed that the pattern of deaths across age groups was similar for the populations covered by the estimates. To validate our approach, we also calculated deaths at older ages based upon the cause-of-death structure taken as that within age groups for the Global Burden Disease Project for sub-Saharan Africa. These two approaches yielded virtually identical results (Chi square P value =0.99) in the final distribution of major causes of deaths. A) Discounting of life years To calculate the discounted years of life lost, we derived the average age of death for each age and sex group for the various diseases using a study in Ghana (Ghana Health Assessment Project Team, 1981). We subtracted the mean age of death from the maximum time lived in developed countries: eighty for women and seventy-eight for men (Ansley et al., 1982). We discounted years of life lost from death at a rate of three percent to take into account that society has a preference for consumption now rather than later (World Bank, 1993b). The higher the discount rate, the more years of life in the present are valued relative to those in the future. 8 Figure 1. Rapid method to calculate mortality data for Guinea DATA SOURCES METHOD, BY STEPS Global Burden of Step 1: Abstract the age-, sex-specific population values and age-, sex-, cause- Disease estimates > specific numbers of deaths for sub-Saharan Africa. These data are presented in seven age categories; 0 to 4, 5 to 14, 15 to 29, 30 to 44, 45 to 59, 60 to 69 and 70 or more. v Step 2: In place of zero (published) values, assume the number of deaths due to most diseases to be 500 for each age and sex-category in the entire sub-Saharan population. V Step 3: Adjust proportionally the number of age and sex-specific causes of death such that their sum matches the total number of deaths in sub-Saharan Africa. e Step 4: As described in the text, collapse the data into twenty-five disease categories, based upon the Intemational Classification of Diseases system. v Step 5: Calculate mortality rates for each age and sex-stratum. v Step 6: For each stratum, calculate the ratios of the mortality rates among older age groups to the corresponding mortality rate among 0 to 4 year-olds. For "maternal causes" calculate the ratios of mortality rates among older age groups to the mortality rate among 5-14 year-olds. e Guinean studies of Step 7: Adjust the mortality rates for the group aged 0 to 4 years to reflect data childhood mortality > available from Guinea, using local studies with data on deaths in early childhood, and data on HIV studies of HIV prevalence and injuries. infection and injuries e Guinean data, derived Step 8: Multiply the sex-specific mortality rates generated in Step 7 for each disease from censuses, for > by the corresponding population of 0 to 4 year-olds. Sum the resulting values to yield population and the total predicted number of deaths among 0 to 4 year-olds. Adjust proportionally mortality the sex-specific mortality rates such that the predicted number of deaths matches the actual number of deaths among 0 to 4 year-olds. V Step 9: Calculate mortality rates for the older age categories by multiplying the sex- specific mortality rates for each disease for 0 to 4 year-olds by the ratios calculated in Step 6. e Guinean studies of Step 10: Divide the total number of maternal deaths in Guinea into the 5 age maternal mortality > categories 5 to 14, 15 to 29, 30 to 44, 45 to 59, and 60 to 69 years of age such that the ratios of maternal mortality rates among older groups to the rate among 5 to 14 year- olds is the same as the ratios calculated in Step 6. V Guinean data, derived Step 11: Multiply the age and sex-specific disease mortality rates by the populations from censuses, for > of the appropriate age-and sex-specific groups in Guinea. Sum the resulting values population and to yield the total predicted number of deaths for each age category. Adjust mortality proportionally the sex and age-specific disease mortality rates (except maternal mortality rates) such that the predicted number of deaths for each age group matches the actual number of deaths. 9 B) Estimates of burden of disability Unlike for deaths, there was no feasible method to estimate years of disability by age and sex group. Thus, we used a simplified methodology adjusting the Global Burden of Disease Project estimates of years lived with a disability in sub-Sahara Africa among age groups' 0-4, 5-14, 15-44, 45-59 and 60+ for each sex to the relative population size for Guinea. As with mortality estimates, we combined certain categories of diseases. These estimates are discounted at three percent per year and have been age-weighted with a maximum social age preference of twenty-four years (Murray, 1994). In contrast, the estimates for mortality are not discounted and are not age- weighted; the estimates for years of life lost are discounted but not age-weighted. Thus the three measures are not directly comparable, but as reported by Murray et al. (1994a), these different weighting schemes do not greatly alter the relations among the major causes of death -- communicable, maternal, and perinatal disease, injuries, and non-communicable diseases. RESULTS FOR BURDEN OF DISEASE Among a Guinean population estimated at 6.09 million, 124,078 people died in 1992 or 20.4 deaths per thousand population. As table 2 shows, the mortality rates among females and males were very similar--20.2 deaths and 20.5 deaths per thousand female and male population, respectively. A child born in Guinea faces a nearly one in four chance of dying before age five and has a life expectancy at birth of only forty-four years. A child reaching age five has about a one in fourteen chance of death before age fifteen. A child at age fifteen has a twenty-four percent chance of dying before age forty-five and adults who reach age forty-five have about a twenty-three percent chance of dying before age sixty. The table also reveals that mortality rates are higher for males than for females at younger ages but lower at older ages. Background annex 1, tables 1 to 4 provide detailed data for the results of the burden of disease analysis. Causes of death, Overall, communicable, maternal, and perinatal diseases caused 86,400 deaths (seventy percent), injuries caused 9,000 deaths (seven percent) and non-communicable diseases caused 28,700 deaths (twenty-three percent). The proportions in the first category are similar for females and males. Deaths from injuries are more common among males, and deaths from non-comrnmunicable diseases are more common among females. As table 3 shows, the top seven causes of death are the same for both sexes, ranging down from respiratory infections, malaria, diarrhea, childhood infections, cardiovascular diseases, tuberculosis, and perinatal causes. Of those diseases shared by males and females, the gap between the percentages for the two sexes is widest --being more common in women -- for cardiovascular diseases, and -- being more common in men -- for unintentional injuries. Results for Age Groups. Figure 2 illustrates the relative importance of the major categories of cause of death in Guinea by age group. Table 4 lists, in decreasing order, the top ten causes of death by age group in Guinea. The mortality rate among children aged zero to four is 54.1 deaths per thousand population. Among this age group, ninety-one percent of deaths are due to communicable, maternal, and perinatal diseases, led by diarrhea, respiratory infections, malaria, 10 Table 2. Distribution of population and number of deaths in Guinea, by sex and age, 1992 Sex and age Percent of total Number of Deaths per Probability of group population deaths thousand dying in this population age group Females 0-4 9.7 29,624 50.4 23.7 5-14 13.9 7,103 8.4 7.6 15-44 20.8 11,716 9.3 24.4 45-59 4.1 4,788 18.9 22.8 60+ 2.3 9,328 66.7 100.0 Total 50.8 62,559 20.2 ... Males 0-4 9.5 33,639 58.0 23.0 5-14 13.7 6,885 8.3 6.7 15-44 20.3 9,689 7.8 24.8 45-59 3.8 4,147 18.1 25.9 60+ 1.9 7,159 60.3 100.0 Total 49.2 61,519 20.5 ... Total males and females 100.0 124,078 20.4 ... Table 3. Distribution of deaths by cause in Guinea, by sex 1992 Botb sexes Females Males Ranked cause Percentage Ranked cause Percentage Ranked cause Percentage of death of death of death Respiratory infections 13.3 Respiratory infections 13.4 Respiratory infections 13.2 Malaria 12.3 Cardiovascular 12.2 Malaria 12.6 Diarrhea diseases 11.8 Malaria 12.1 Diarrhea diseases 12.1 Childhood infections 10.5 Diarrhea diseases 11.6 Childhood infections 10.9 Cardiovascular 10.2 Childhood infections 10.2 Cardiovascular 8.2 Tuberculosis 7.1 Tuberculosis 6.9 Perinatal causes 7.6 Perinatal causes 6.3 Perinatal causes 5.2 Tuberculosis 7.3 Unintentional injuries 4.4 Maternal causes 4.6 Unintentional injuries 6.2 Malignant neoplasms 3.5 Malignant neoplasms 3.7 Intentional injuries 3.7 Intentional injuries 2.9 Nutritional/endocrine 2.7 Malignant neoplasms 3.4 Nutritional/endocrine 2.4 Unintentional injuries 2.6 Nutritional/endocrine 2.1 Maternal causes 2.3 Syphilis 2.2 Digestive 2.0 Syphilis 2.1 Intentional injuries 2.0 Syphilis 2.0 Digestive 2.0 Digestive 1.9 Tropical diseases 1.5 Tropical diseases 1.6 HIV 1.8 Congenital 1.4 HIV 1.5 Tropical diseases 1.6 HIV 1.3 Note: total deaths 124,078 62,559 61,519 11 childhood infections, and perinatal diseases. Among non-communicable diseases, congenital diseases are the largest causes of death. The mortality rate among children aged five to fourteen is much lower, 8.3 deaths per thousand population. Among this age group, communicable, maternal, and perinatal diseases continue to predominate, constituting seventy-three percent of deaths; these are led by malaria, childhood infections, diarrhea, respiratory infections, and tuberculosis. Unintentional injuries constitute nine percent of deaths in this age group. Neuro-psychiatric disease is the leading non- communicable cause of death. The mortality rate during young adulthood and early middle age (ages' fifteen to forty-four) is 8.6 deaths per thousand population. Three in five deaths in this group are caused by communicable, matemal, and perinatal diseases, led by tuberculosis and maternal causes (one in four of all deaths among women in this age group). Intentional and unintentional injuries together account for more than twenty-two percent of deaths. Non-communicable diseases are not yet prominent. The mortality rate among people aged forty-five to fifty-nine years of age is 18.5 deaths per thousand population. The leading causes of death in this age group are non-communicable diseases (sixty-four percent), led by cardiovascular diseases and cancer. However, even in this age group, one in five deaths are due to tuberculosis. Finally, as expected, people over sixty years have the highest mortality rate of 63.7 deaths per thousand population; three in four succumb to non-communicable diseases, mainly cardiovascular disease and malignant neoplasms. Even in this older age group, respiratory infections and tuberculosis cause substantial numbers of deaths. YearL of life lost Table 5 presents years of life lost discounted at a rate of three percent. These show a pattern similar to that for deaths, with communicable, maternal, and perinatal diseases accounting for more than three-quarters of years lost. Non-communicable diseases account for sixteen percent, and injuries cause seven percent. Diseases of childhood obviously are responsible for the most lost years: diarrhea, malaria, respiratory infections, injuries, childhood cluster, and tuberculosis (see table 6). Simple estimates of years lived with a disability. Table 7 provides estimates for years lived with a disability, discounted at three percent and with a maximum time preference of twenty-four years. This analysis shows that the group of communicable, maternal, and perinatal diseases account for about forty-three percent of these years lived with a disability. Non-communicable diseases account for a similar percentage. Injuries cause the most years lived with a disability, followed by neuro- psychiatric illness. Approximately thirty-five percent of all years lived with a disability begin at ages' fifteen to forty-four and another thirty-five percent in children under age five. Years lived with a disability from ages' five to fourteen contribute sixteen percent to the total. There are important sex differences: sexually transmitted diseases account for nearly six times more such years in females 12 Figure 2. Distribution of deaths, by major cause and age in Guinea, 1992 100% -1 90% 6%1 80% 70% 2 L 4;~~~~~~~~~~~~~~~~4 60% e~~~~~~~~~~~~g Gru\yas 0 o 4 \o1 St 44 o5 0 50%\ 40%\ \ 30%~ 9 20%~~~~~~~~~~~~~~~~2 20%~~~~~~~~~~~~~~0 10% 21% 0% OA4 (63,264 deaths) 5-14 (13.988 deaths) 15-44(21.405 deaths) 45-59 (8,935 deaths) 60+ (16,486 deaths) Age Group (years) *g Communicable, Maternal & Perinatal a Injuries [jNoncommunicable Table 4. Top ten causes of death in Guinea, by age, 1992 (ranked in order and percentage except as noted) 0 to 4 5 to 14 15 to 44 45 to 59 60+ Diarrhea diseases 20.3 Malaria 19.0 Tuberculosis 19.4 Cardiovascular 29.4 Cardiovascular 50.0 Respiratory 19.3 Childhood 14.8 Matemal causes 12.8 Tuberculosis 19.8 Malignant 12.6 infections infections neoplasms Malaria 17.6 Diarrhea diseases 10.5 Intentional injuries 11.4 Malignant 15.9 Respiratory 12.1 neoplasms infections Childhood infections 17.3 Respiratory 10.5 Unintentional 11.0 Digestive 6.5 Tuberculosis 7.4 infections injuries Perinatal causes 12.4 Tuberculosis 8.9 Syphilis 6.5 Unintentional 4.2 Respiratory 3.2 injuries Congenital 2.2 Unintentional 8.8 Malaria 6.1 Nutritional/endoc 3.4 Digestive 3.1 injuries Nutritional/endocrine 2.1 Tropical diseases 4.8 HIV 5.9 Tropical diseases 3.2 Genito-urinary 2.3 Unintentional injuries 2.0 Neuro-psychiatric 4.2 Cardiovascular 5.7 Respiratory 2.9 Nutritional/ 2.2 endocrine Syphilis 1.7 Meningitis 2.7 Tropical diseases 3.4 Intentional 2.6 Unintentional 1.5 injuries injuries Digestive 0.8 Nutritional/ 2.4 Nutritional/ 3.3 Genito-urinary 2.6 Diabetes mellitus endocrine endocrine Total Deaths 63,264 13,988 21,405 8,935 16,486 Probability of dying in this age group 23.4 7.1 24.6 24.3 100.0 13 Table 5. Distribution of years of life lost in Guinea, by sex and cause, 1992 Both sexes Females Males Ranked cause Percentage Ranked cause Percentage Ranked cause Percentage of years of life lost of years of life lost of years of life lost Malaria 14.2 Respiratory infections 14.2 Malaria 14.4 Respiratory infections 14.2 Malaria 14.1 Respiratory infections 14.1 Diarrhea diseases 13.9 Diarrhea diseases 13.7 Diarrhea diseases 14.0 Childhood Infections 12.3 Childhood Infections 12.1 Childhood Infections 12.6 Perinatal causes 7.5 Tuberculosis 6.6 Perinatal causes 8.8 Tuberculosis 6.3 Cardiovascular 6.3 Unintentional injuries 6.2 Cardiovascular 5.3 Perinatal causes 6.2 Tuberculosis 6.0 Unintentional injuries 4.5 Maternal causes 4.7 Cardiovascular 4.2 Intentional injuries 2.9 Unintentional injuries 2.7 Intentional injuries 3.6 Nutritional/endocrine 2.4 Nutritional/endocrine 2.6 Nutritional/endocrine 2.2 Maternal causes 2.4 Malignant neoplasms 2.4 Syphilis 2.1 Syphilis 2.2 Syphilis 2.3 Malignant neoplasms 1.9 Malignant neoplasms 2.2 Intentional injuries 2.1 Digestive 1.6 HIV 1.7 HIV 2.0 Congenital 1.6 Digestive 1.7 Digestive 1.7 Tropical diseases 1.5 Tropical diseases 1.5 Tropical diseases 1.6 HIV 1.4 Total years of life lost 3,157,636 1,579,520 1,578,117 Table 6. Top ten causes of years of life lost in Guinea, by age, 1992 (ranked in order and percentage except as noted) O to 4 5 to 14 15 to 44 45 to 59 60+ Diarrhea diseases 20.3 Malaria 19.2 Tuberculosis 19.8 Cardiovascular 29.5 Cardiovascular 47.9 Respiratory 19.3 Childhood Infections 14.9 Maternal causes 13.2 Tuberculosis 19.6 Malignant neoplasms 13.6 infections Malaria 17.6 Diarrhea diseases 10.6 Intentional injuries 11.5 Malignant neoplasms 15.9 Respiratory 12.3 infections Childhood Infections 17.3 Respiratory infections 10.6 Unintentional injuries 11.1 Digestive 6.5 Tuberculosis 8.6 Perinatal causes 12.4 Tuberculosis 9.0 HIV 6.7 Unintentional injuries 4.1 Digestive 3.3 Congenital 2.2 Unintentional injuries 8.8 Syphilis 6.6 Nutritional/endocrine 3.5 Respiratory 3.1 Nutritional/endocrine 2.1 Tropical diseases 4.8 Malaria 6.2 Tropical diseases 3.4 Genito-urinary 2.4 Unintentional 2.0 Neuro-psychiatric 4.2 Cardiovascular 5.4 Respiratory 2.9 Nutritional/endocrine 2.3 injuries Syphilis 1.7 Meningitis 2.7 Tropical diseases 3.6 Genito-urinary 2.6 Diabetes mellitus 1.7 Digestive 0.8 Nutritional/endocrine 2.4 Nutritional/endocrine 3.2 Intentional injuries 2.6 Tropical diseases 1.4 Note: total years of life lost, in thousands 1,902 406 539 161 148 14 Table 7. Distribution of years of life lived with a disability in Guinea, 1992, by cause and sex Both sexes Females Males Ranked cause Percentage Ranked Cause Percentage Ranked Cause Percentage Injuries 14.3 Maternal 11.6 Injuries 19.7 Neuro-psychiatric 11.4 Neuro-Psychiatric 9.8 Neuro-Psychiatric 13.1 Nutritional/endocrine 8.9 Injuries 9.0 Nutritional/endocrine 9.1 Malaria 7.3 Nutritional/endocrine 8.6 Malaria 7.4 Perinatal 7.1 Malaria 7.2 Perinatal 7.3 Maternal 5.9 STD's except HIV 7.0 Tropical cluster 7.2 Tropical cluster 5.5 Perinatal 7.0 Congenital 4.8 Congenital 4.8 Cardiovascular disease 4.7 Cardiovascular disease 4.3 Cardiovascular disease 4.5 Congenital 4.7 Respiratory-chronic 3.8 STD's except HIV 4.2 Respiratory-chronic 3.9 Digestive 3.6 Respiratory-chronic 3.8 Tropical cluster 3.8 Childhood cluster 2.6 Digestive 3.5 Digestive 3.4 Respiratory-acute 2.6 Respiratory-acute 2.6 Respiratory-acute 2.6 Tuberculosis 2.2 Childhood cluster 2.3 Sense organs 2.2 Genito-urinary 1.8 Tuberculosis 2.0 Trachoma 2.0 Sense organ 1.7 Sense organ 1.9 Childhood cluster 2.0 HIV 1.6 HIV 1.6 Tuberculosis 1.7 STD's except HIV 1.3 Genito-urinary 1.6 HIV 1.6 Intestinal worms 1.2 Note: total years lived with a disability 789,370 401,802 387,569 than in males, and men have more years than women from the tropical cluster of disease and from neuro-psychiatric illness. Comparison with other regions, The distribution of causes of death in Guinea is similar to that in sub-Saharan Africa, which is not surprising given the methodology (figure 3). In comparison with developing and developed countries, the overall burden of mortality in Guinea is much higher. The explanation exists partly in Guinea's younger age structure, and thus a higher probability of deaths from diseases associated with childhood and childbearing. However, even among children under age five, those in Guinea have a more than two-fold higher probability of death than do children in other developing countries. Among the middle aged (fifteen to fifty-nine), Guineans have a higher probability of death than people in other developing countries (forty percent versus twelve percent, respectively). These higher risks are largely due to a higher burden of communicable, maternal, and nutritional disease, that carry a high chance of death in the absence of available medical care. Thus, these higher risks also represent the relatively low access to medical care in Guinea. 15 Figure 3. Relative importance of causes of death in Guinea, sub-Saharan Africa, developing and developed countries 1000/; - 900/0 23 [13 O/o 26% 800/o 7% ~~~~~~~~~~50%/ 70D/o 8/- 600/o 88- a~500/o-- 400/o - 300% I100/0 7% 5% 0% Guinea SSA Developing Developed Countries Countries | Communicable, Maternal & Perinatal * Injuries j Non-communicable Item Guinea Sub-Saharan Developing Developed Africa countries countries Probability of dying before age 5, in percent 23.5 15.8 9.7 1.7 Probability of dying between ages 15 and 60, in percent 39.9 35.1 21.8 11.7 16 Impact of population growth on burden of disease. World Bank estimates suggest that population growth in Guinea was 2.8 percent per annum in 1992 and it is projected to increase to 2.9 percent by 2000. The growth rate in Conakry is about five percent per annum. The population of Guinea is expected to reach over 7.6 million by 2000 and 11.6 million by 2015. No major declines in the fertility rate are expected before 2010, and therefore the number of births is expected to increase annually by 2.5 percent (from 298,000 in 1992 to 446,000 in 2012). Figure 4 depicts future numbers of deaths based upon population growth and changes in death rates. Detailed estimates are in Background annex 1, table 4. These estimates assume that HIV-related deaths will increase at a net rate of one percent per year, that current levels of health service provision will be maintained, and that no major epidemics will occur. They suggest that the total number of deaths will increase by about 1.5 percent a year, despite a falling death rate. Given shifts in the age of the population (Figure 5), more deaths are likely to occur from tuberculosis and injuries, diseases that have a higher death rate in adolescents and young adults. These simple linear projections are not reliable for each specific disease, but they do suggest that the overall pattern of high mortality from communicable, maternal, perinatal, and nutritional diseases will continue among an ever-increasing absolute number of children born over the next decade. Figure 4. Projections of number of deaths in Guinea, by cause 180 -- 160 140 43 _________ ; 140 ,- 35 CNonconmnunicable 120 .0 r- 100 ~~~~~~~~~~~~~~~~~*Injuries 80- 0 _ * ] Comrnunicable, c) 60 106 Matemal & Perinatal z 40 86 20 0 1992 2000 2010 Year 17 Figure 5. Distribution of deaths by age group in Guinea, 1992, 2000, and 2010 90% 80% 70% 0o6+ < 60% -J45-59 ° 50 . _ 15 44 0 °/ 50% 40% ~- a. 30% 20% 10% 1992 2,000 2,010 Year CONCLUSIONS AND POLICY IMPLICATIONS OF DISEASE BURDEN This analysis underscores the need for national health care planners and international donors to focus on diseases that cause the most premature deaths. In Guinea, these are largely avoidable or treatable communicable disease, maternal, and malnutrition diseases that together account for nearly three in four of all deaths and, with perinatal disease, account for nine in ten deaths before age five. Moreover, thirty-one percent of deaths during the economically-productive ages of fifteen to fifty-nine are caused by tuberculosis, injuries and maternal deaths, which also may be treated or avoided. Evaluating Health Programs and Performance Using Deaths and Disability The Global Burden of Disease Project suggested that premature deaths cause over three- quarters of all the burden of disease, death and disability in sub-Saharan Africa (Murray et al., 1994a). The same is likely to be true for Guinea. Thus, reduction in deaths is a plausible goal for health care provision and is also a reasonable measure for evaluating the performance of the health care system. Nonetheless, disability is important, especially when its burden is high and cost- effective and feasible programs for addressing it are available. One example is intestinal helminthiasis among school-aged children (five to fourteen), which is responsive to school deworming programs (World Bank, 1993b). While the reduction in deaths is a valid objective for health systems, people in Guinea may utilize health services more for relief of disability and even anxiety. Moreover, political and social 18 factors may influence the choice and location of services offered. Historically, the evaluation of delivery systems in Guinea and most other parts of the world has not relied upon reduction of premature deaths as an outcome measure. It would not be reasonable to assume that, in the short term, health systems may be modified to measure program performance by premature mortality reduction. In the longer term, these estimates of the burdens of mortality and morbidity are useful additions to health system management and organizational structures. The estimates are not as reliable as those gathered from comprehensive and time-consuming prospective epidemiological studies. However, they do provide a common framework for comparing reduction in mortality for one disease or reductions attributable to one program. Regional and Equity Implications While we do not describe them in detail here, regional differences in mortality within Guinea are likely important. For both demand and supply reasons, access to health services is lower in Upper and Forest Guinea, and malaria, tropical illness, maternal deaths and nutritional diseases are more prevalent in these regions than in Middle and Maritime Guinea. Data from a household survey (DNSI, 1994) suggests that burdens from communicable, maternal, and nutritional diseases are twenty to thirty percent higher among rural people and those with low educational status, conditions associated strongly with poverty. Thus, much of the burden of disease likely falls upon the poor, and effective health care provision yields improvements in equity. Even if it reduces disease by the same relative amount across income groups, the absolute reductions would be larger among the poor. Addressing Population Growth and Future Burdens of Disease This analysis reinforces the suggestion that the growth of population will result in substantial increases in absolute health burdens (Ainsworth, 1994). Health care planners must pay attention to allocating resources to cover not only today's population but also the larger numbers of children born in the future. The projections of future disease burden cannot be taken as precise, as more intense application of appropriate health services to the most prevalent diseases may occur, such as wider immunization campaigns against childhood infections and oral rehydration treatment for diarrhea. Past efforts at family planning services in Guinea have had limited success (World Bank, 1986; SEATS/JSI, 1994). Future efforts will need to focus on childhood mortality and the lowering of female fertility by improving its determinants--such as education, birth spacing, and contraceptive use. Only these efforts, along with economic growth may help avoid increases in absolute mortality. A reasonable goal is to narrow the gap between mortality and fertility decline (Bobadilla et al., 1993). Finally, growth in some diseases may be expected. Deaths from tuberculosis and injuries may grow, largely because the numbers in the age group affected by these conditions will increase. Thus more resources should be allocated in the future to controlling these causes. The Role of Vital Registration and Epidemiological Studies We analyzed all the major causes of death in Guinea and thus avoid the common scenario whereby some experts overestimate the mortality attributable to disease of interest. Our estimates are consistent with the pattern of utilization at Guinea's health centers and hospitals (World Bank, 1993a; SEATS/JSI, 1994). They are consistent also with estimates of mortality from a more formal epidemiological study done in Ghana in the late 1970s (Ghana Health Assessment Team, 1981; 19 table 8) and a recent study in Tanzania (Kitange et al., 1996). While the Guinea estimates likely suffer from imprecision, the magnitude of error is not likely to compromise the decision, based upon cost-effectiveness or other analyses, about allocation of resources among diseases. In addition, we have minimized error by focusing upon the maximum quality of information for early childhood and by ensuring that the overall total of each cause matched independent estimates of numbers of deaths. The estimates for disability are less robust than those for mortality as there was no clear way to calculate an upper limit for disability within each age and sex. Table 8. Relative importance of causes of death in Guinea and Ghana Guinea, 1992 Ghana, 1979 Ranked cause Percent Ranked cause Percent of death deaths of death deaths Respiratory infections 13.3 Perinatal causes 13.6 Malaria 12.3 Childhood infections 12.2 Diarrhea diseases 11.8 Cardiovascular 11.1 Childhood infections 10.5 Respiratory infections 9.9 Cardiovascular 10.2 Other non-communicable (1) 9.0 Tuberculosis 7.1 Diarrhea diseases 5.9 Perinatal causes 6.3 Malaria 5.5 Unintentional injuries 4.4 Nutritional/endocrine 5.4 Malignant neoplasms 3.5 Injuries (2) 4.6 Intentional injuries 2.9 Digestive 4.5 Nutritional/endocrine 2.4 Tuberculosis 4.2 Maternal causes 2.3 Malignant neoplasms 3.2 Syphilis 2.1 Other infections (3) 2.6 Digestive 2.0 Maternal causes 1.9 Tropical diseases 1.6 Tropical diseases(4) 1.8 Notes: t ) Other non-communicable: sickle-cell disease and cirrhosis. (2) injuries: Unintentional and intentional injuries. (3) Other infections: leprosy, chicken pox, hepatitis and influenza. (4) Tropical diseases: hookworm anemia, trypanosomiasis, schistosomiasis and onchocerciasis. Other methods exist for deriving mortality estimates, including more sophisticated regression models that make various assumptions about fertility and life expectancy (Murray, 1994). Although these may yield more precise numbers, the marginal improvement in their validity over those of cruder models remains unknown. The major causes of death may be determined definitively only by epidemiological studies and vital registration. Limited efforts with epidemiological monitoring in sub-Saharan Africa have been done successfully in a variety of local settings (Feachem et al., 1991). National efforts on representative samples are lacking. Some newer epidemiological methods are applicable to low-income countries. Among them are verbal autopsy (Gray et al., 1990) for childhood deaths, the sisterhood method for maternal mortality (David et al., 1991), capture-recapture methodology for completeness of data (Laporte, 1992), and the spouse case-control methods (R. Peto, unpublished data) exist. These may significantly enhance the quality of baseline data and thus national mortality and disability estimates, cost-effectiveness analysis, and examination of risk factors. Such efforts lie within the comparative advantage of govermnents (World Bank, 1994) and thus should be priorities as Guinea reforms its health system. 20 COST-EFFECTIVENESS OF HEALTH INTERVENTIONS IN GUINEA Cost-effectiveness analysis is a method of comparing the costs of a health care intervention with its health impact. It provides a comparison of discounted years of life saved to a unit cost, for various interventions. This analysis, like that in the last section, utilizes mortality, as defined by years of life saved. In the context of the burden of disease described in the previous section, the prime purpose of cost-effectiveness analysis is to guide the allocation of the scarce resources in Guinea toward maximum benefit in health gains. It also offers governments, communities and foreign donors a common framework for evaluating health interventions. Cost-effectiveness analyses do not offer fully valid estimates for each specific intervention, but do permit comparison and ranking across interventions. METHODOLOGY FOR COST-EFFECTIVENESS ANALYSIS We analyzed a total of forty interventions selected from treatment protocols at health centers (or centre de sante), health posts (or poste de sante), and prefectoral (or first referral) hospitals in Guinea. We chose these in consultation with health care providers as the most commonly performed interventions. We also included several interventions that are planned but not yet implemented, such as school-based health programs, or that have been cost-effective in similar countries, such as anti-tobacco programs. We chose four levels of interventions: (1) those at outreach level which rely upon reaching members of the community in their own surroundings; (2) at the health post level; (3) at the health center level; and (4) at the prefectoral hospital level. As much as possible, we structured the analysis to enable comparisons across current common interventions by using estimates of actual, rather than ideal practice. The interventions are defined as preventive or curative, and are defined by their referral level. All outreach, or community-based programs were preventive interventions. At the health post, health center and hospital levels, interventions were curative except for the immunization program. Cost-effectiveness analyses need to consider how many people without the condition need to be screened in order to treat an affected person. As the referral rate in Guinea from health center to hospital is less than one percent, we adjusted the analysis to account for visits for those interventions available at both health centers and hospitals. Overall, twenty-two of the forty interventions are directed towards the health of children under age five or maternal health. For each intervention, we assigned a value for efficacy, or the reduction in disease incidence or the reduction in case fatality rate among untreated populations under ideal conditions. Background annex 2, table I contains the detailed calculation for all interventions used. The major sources of these data were a recent World Bank book (Jamison et al., 1993), a study in Ghana (Ghana Health Assessment Team 1981), and the literature. Ideal efficacy is that found in clinical trials under well-controlled situations, but is seldom seen in actual practice. After discussions with clinicians, and reviewing results from other countries, we adjusted the ideal efficacy to reflect 21 diagnostic accuracy, the quality of service delivery, the availability of drugs and equipment, patient compliance, and other factors (Drummond et al. 1987). We used maximum downward adjustments to derive conservative estimates for ideal efficacy: for outreach programs the adjustment ranged from thirty to seventy percent, depending upon the expected compliance by the user. At a health post, adjustment ranged from twenty to fifty percent. At a health center, adjustment was a generic seventy percent. A hospital had a higher value of eighty percent to reflect the better quality of care at hospitals. The overall effectiveness of an intervention is the efficacy multiplied by the adjustment factor. For example, immunization has an efficacy of ninety-five percent, and an adjustment of seventy percent because of poor drug supply, inadequate refrigeration of the vaccine, and other factors. Thus the overall effectiveness is 66.5 percent (0.95 x 0.70). For each disease category, we derived a median age of death from the literature and the Ghana study (Ghana Health Assessment Team, 1981) and we assigned discounted life years to each disease using a Coale and Demeny, Model West 26 life table (Ansley et al., 1982). The latter used a life expectancy of eighty for women and seventy-eight for men, with future years of life discounted at three percent. The years of life saved represents the difference between the life years lost among those without treatment and those with treatment, after taking into account the proportion of the population already receiving the treatment (or treatment coverage). For each intervention, we assigned a target population of people who could potentially benefit from the intervention, depending upon the level of the intervention, population size and expected number of cases. Costs of Labor. Drugs. Equipment, and Overhead. To determine labor costs we calculated the total monthly official salary for four levels of government staff: a senior doctor; a new graduate doctor; a nurse, midwife, health aide, or laboratory technician; and an outreach worker or assistant agent. We made a forty percent adjustment for pensions and a fifty percent multiplier to reflect poor manpower planning, an inefficient referral system, and other inefficiencies in the system. We divided these costs by one-hundred and seventy-one working hours per month to derive hourly salaries. In Guinea, very few health care workers receive housing or other benefits, so we excluded this category of benefit. We obtained data on the types of drugs, their dosages, and the length of treatment from protocols of the health centers and from hospital guidelines, as well as from expert opinion. For each drug, we used the highest price obtained from the Central Pharmacy, direct health center purchases, and independent private distributor, and we used a two-fold multiplier to account for transportation costs, spoilage, theft, and loss. The cost per use of equipment was the replacement cost of the equipment divided by the number of uses in its lifetime and multiplied by the number of uses for that intervention. We derived the overhead for an intervention by taking the national non recurrent salary budget, the amnortized investment budget, and the administrative budget weighted by number of visits across the levels of care (health center, health post and hospital). This figure includes vehicle and other transport costs for the facility. We derived overhead costs for hospitalizations similarly. Overhead costs for hospital outpatient interventions were one-seventh of those for in-patients, as noted in a recent study (World Bank, 1996) 22 RESULTS OF COST-EFFECTIVENESS ANALYSES OF INTERVENTIONS Table 9 lists the main results of the total cost of the interventions, target population, present coverage, and overall cost-effectiveness expressed as 1994 US dollars per life year saved. Of the interventions evaluated, eight were at the outreach level, four at the health post level, twelve at the health center level and sixteen at the hospital level. Table 10 ranks the interventions from most to least cost-effective along with the percentage of years of life lost in their target population. According to these analyses, several interventions cost about one-hundred dollars per year of life saved or less, and address more than five percent of their target population. These include treatment of childhood pneumonia at three dollars per years of life saved, rehydration therapy for diarrhea at health centers (seven dollars) and at health posts (eight dollars), treatment of childhood malaria at health centers (eight dollars), childhood vaccination at health centers (twenty-five dollars), short-course tuberculosis treatment at health centers (eleven dollars), treatment of sexually-transmitted diseases at health centers (sixty-seven dollars) and impregnated bed nets against malaria (forty-three dollars). Maternal and perinatal diseases, which together account for nine percent of years of life lost in the target group, have slightly less cost-effective interventions: prenatal and delivery care at health centers (one-hundred and thirty-three dollars) or programs using outreach trained birth attendants or TBA (two-hundred and sixty-one dollars). For non-communicable diseases, the most cost-effective interventions appear to be preventive outreach programs. These include tobacco legislation and warnings and an injury prevention program (both at sixty-five). Among curative programs for non-communicable disease, appendectomy (thirty-six dollars) and hernia repair (seventy-four dollars) are the most cost- effective. Finally, although it is not examined in this analysis, many of these interventions also reduce morbidity. For example, school-based health programs should reduce the relatively large burden of intestinal worms and micronutrient deficiency in children and also should reduce the large percentage of visits to health center arising from intestinal worm infection (World Bank, 1993a). These analyses reveal tremendous variation in the costs involved in gaining one life year. It is not obvious that preventive interventions are always the most cost-effective. For example, the cost of gaining one year of life from water and sanitation improvement is three-hundred and forty- two dollars versus seven dollars for treatment of diarrhea in children at health center. The ratios reflecting the most cost-effective interventions are driven largely by low cost rather than by high efficacy. Among outreach interventions with cost-effectiveness ratios below one hundred dollars per years of life saved, all but one (condom use for bar ladies) have absolute costs of less than four dollars per person. Similarly, for health center and health post interventions with cost-effectiveness ratios below one-hundred dollars, all except two (short-course tuberculosis treatmnent and childhood vaccination) have absolute costs of less than four dollars per person. At health center and health post, four of the top most cost-effective interventions have costs less than three dollars per person. In contrast, the average cost per treatment is at least ten times higher among hospital-based interventions with cost-effectiveness ratios below one-hundred 23 Table 9. Summary of population, coverage, and cost-effectiveness for forty health interventions in Guinea, 1994 Years of Total cost in Cost per year of life Adjusted Size of Current life saved thousands saved in Level Disease and intervention Name of target group efficacy target coverage per person of Guinea thousands US dollars (percent) group (percent) with francs of Guinea treatment francs Malaria Use of impregnated bednets Age <5in malaria areas 16.5 818,300 1 0.070 2,920 42 43 Treatrnentofchildren athealth post Age<5 inmalariaareas 19.5 818,300 22 0.134 1,053 8 8 Treatment ofchildren at health center Age <5 in malaria area 32.5 818,300 22 0.223 2,844 13 13 Hospital treatment of severe childhood malaria Severe malaria age <5 52.0 40,915 8 0.512 43,555 85 87 Acute respiratory infection Educate mothers on childhood pneumonia at Age <5 0.6 1,169,000 8 0.002 229 105 108 health center Treatment of children at health center Age <5 40.0 701,400 45 0.753 2,204 3 3 Treatment ofadults at health center Age >5 6.0 436,020 30 0.023 2,993 130 134 Hospital treatment of children with severe Age <5 64.0 35,070 8 2.280 68,441 30 31 pneumonia Hospital treatment of adults with severe Age >5 24.0 436,020 13 0.306 63,550 208 213 pneumonia Maternal and family planning Outreach family planning services Married women age 15 to 49 27.0 1,334,000 3 0.267 73,857 276 283 Trained birthattendants(TBA)prenatal and Births plusmothers 15.0 166,800 15 0.099 25,172 254 261 delivery care at health post Normal prenatal and delivery care at health center Newborns and delivering 28.0 222,400 28 0.185 23,936 130 133 mothers Normal prenatal and delivery care at hospital Newborns and mothers 32.0 111,200 37 0.211 26,090 124 127 Obstructed birth/Cesarean section Referrals from lower levels 56.0 5,560 10 2.309 41,481 18 IS Nutritional deficiency School-based vitaunin A/iodine/deworming Age 5 to 14 in school 28.0 673,600 5 0.028 1,849 66 67 Surveillance and treatment of mild malnutritioi at Age <5 17.5 1,169,000 25 0.021 15,855 760 779 health center Hospital treatment of severe malnutrition Severely malnourished <5 24.0 58,450 5 1.687 68,848 41 42 Diarrhea Construction of pit latrines and safe water supply Age <5 21.0 1,169,000 15 0.080 26,775 335 343 Rehydration therapy at health post Age <5 27.0 233,800 10 0.175 1,300 7 8 Rehydration therapy at health center Age <5 45.0 233,800 20 0.291 1,876 6 7 Hospital treatment for severe diarrhea in children Age <5 48.0 4,676 29 0.569 41,305 73 74 Cardiovascular Tobacco legislation and warnings Age >15 0.2 1,622,500 5 0.000 10 63 65 Hypertensive treatment at health center Hypertensive adults 15.0 37,100 16 0.029 67,746 2,323 2,381 Aspirin in pre-existing cardiovascular disease Pre-existing CVD in adults 12.0 31,100 5 0.070 17,534 251 257 (CVD) at health center Hospital treatment ofrheumatic fever Age 5-19 with RHD 10.0 1,665 5 0.457 30,851 67 69 Hospital treatment for stroke or heart attack Heart attack/stroke patients 17.5 16,540 20 0.549 153,280 279 286 AIDS and sexually-transmitted disease programs AIDS education via media Pop >15 0.1 3,245,000 10 0.000 4 12 12 Condom use/provision for bar ladies Bar Ladies 16.0 16,458 10 0.531 41,903 79 81 Condom distribution for public at health post Age >14 and < 60 18.0 2,986,000 3 0.055 12,379 225 230 Treatment of sexually-transmitted diseases Age >14 and < 60 with STDs 30.6 318,540 20 0.179 11,730 66 67 (STDs) at health center Screening blood for transfusion at hospital Annual transfusions 69.3 10,350 90 0.214 22,421 105 107 Hospital treatment of AIDS patients People with AIDS 14.4 1,681 14 0.108 138,476 1,282 1,314 Childhood infections Childhood vaccination at health center Age <5 66.5 1,169,000 40 0.690 16,658 24 25 Hospital treatment for complicated measles Age <5 24.0 5,845 20 0.867 40,413 47 48 Tuberculosis (TB) Short-coursetreatmentforTBathealthcenter NewcaseswithsputumTB 49.0 9,200 35 4.715 51,716 11 11 Hospital short-course treatment for TB New cases with sputum TB 56.0 9,200 10 5.388 228,746 42 43 Injury or trauma Legislation/fines/seat belts for injury prevention Age >5 0.2 1,873,000 10 0.000 10 65 67 Hospital treatment for severe injury Chest/orthopedic trauma 21.0 12,400 30 0.837 227,229 271 278 Surgical Appendectomy Appendicitis 35.0 12,1% 30 1.860 65,819 35 36 Hernia repair Hernia 28.0 24,392 20 0.714 51,768 72 74 24 twenty-five dollars. This difference stems partly from the higher effectiveness of these interventions at hospital levels, versus at the health center and health post level. Grouping interventions by level of activity can tell us something about the willingness of users to pay for various services (table 11). For example, only government is likely to have much incentive to pay for outreach programs with a high public goods content. Willingness to pay for curative treatment with a high private content is likely higher. In Guinea, previous studies suggest that up to forty percent of people presenting at health centers cannot pay basic fees for treatment (World Bank, 1996). Thus, prioritizing key interventions within a level of care, such as a health center, may help achieve efficient use of public funds to provide free or subsidized care to those unable to pay. Table 11 also presents the costs of achieving the government's objectives for the year 2000. The numbers assume the following: (1) coverage for childhood respiratory, diarrhea diseases and immunization will be stated government goals; (2) a population growth rate of 2.9 percent between 1994 and 2000; and (3) no increase in marginal unit costs of incremental additions of coverage. This analysis reveals that in the year 2000, provision of the current forty interventions will cost over eighty-seven million dollars, at 1994 dollar values. The total represents about eleven dollars per capita in the year 2000, and is nearly quadruple that of current per capita public spending on health. Of the total amount, over eighty-five percent will be needed to expand activities at the outreach, health center, and health post levels. Goals for expanding hospital services are more modest, thus they will require the remaining fourteen percent. The incremental cost for meeting coverage goals for outreach, health center and health post levels is about forty-three million dollars, and that for hospitals about two and half million dollars. A MINIMUM PACKAGE OF HEALTH SERVICES Based upon cost-effectiveness that is around one-hundred dollars per years of life saved and that addresses a substantial percentage of target years of life lost, we suggest one possible minimum package of health services in table 12. This includes outreach programs for AIDS prevention, tobacco control, injury prevention, condom use for bar ladies, school-based health programs, impregnated bednets for malaria, and outreach family planning. These outreach programs have a large public goods' component. The per capita cost for these outreach programs to achieve the government goals for the year 2000 is slightly over one dollar, and the programs address about twenty-one percent of the total years of life lost. For interventions at the health center or health post level, the minimum package has a combination of single and integrated treatments. The single treatments include short-course treatment of tuberculosis, childhood vaccination, treatment of adult pneumonia, and treatment of sexually-transmitted diseases. The two integrated packages are integrated management of the sick child (malaria, pneumonia and diarrhea), and integrated antenatal care, delivery and family planning services. The cost per life year saved for integrated care of the sick child is fifteen dollars. The cost per life year saved for the integrated antenatal care and family planning is above one- hundred dollars, but is included because it address major causes of deaths in Guinea, and because 25 Table 10. Health interventions in Guinea, ranked by cost-effectiveness Intervention US dollars Percentage of years of per life year life lost for target saved population Under fifty dollars per life year saved Treatment of children with pneumonia at health center 3 13.5 Rehydration therapy at health center 7 14.2 Rehydration therapy at health post 8 5.1 Treatment of children with malaria at health post 8 4.4 Short-course treatment for tuberculosis at health center 11 5.7 AIDS education via media 12 1.7 Treatment of children with malaria at health center 13 12.3 Obstructed birth/Cesarean section 18 0.5 Childhood vaccination at health center 25 15.5 Hospital treatment of children with severe pneumonia 31 1.0 Appendectomy 36 0.2 Hospital treatment of severe malnutrition 42 0.2 Use of impregnated bednets for malaria 43 17.6 Hospital short-course treatment for tuberculosis 44 0.6 Hospital treatment for complicated measles 48 1.7 Fifty to one-hundred dollars per life year saved Tobacco legislation/warnings 65 2.4 Legislation/fines/seat belts for injury prevention 65 2.2 School-based vitamin A/iodine/deworming 67 2.1 Treatment of STDs at health center 67 1.7 HIospital treatment of rheumatic fever 69 0.3 Hernia repair 74 0.2 Hospital treatment for severe diarrhea in children 74 1.0 Condom use/provision for bar ladies 81 1.7 Hospital treatment of severe childhood malaria 87 0.9 One-hundred to three-hundred dollars per life year saved Screening blood for transfusion at hospital 107 0.1 Educate mothers on childhood pneumonia at health center 108 0.7 Normal prenatal and delivery care at hospital 127 1.0 Normal prenatal and delivery care at health center 133 4.9 Treatment of adults with pneumonia at health center 134 8.0 Hospital treatment of adults with severe pneumonia 213 0.4 Condom distribution for public at health post 230 1.7 Aspirin in pre-existing CVD at health center 257 0.5 TBA pre-natal and delivery care at health post 261 3.9 Hospital treatment for severe chest/orthopedic injury 278 0.6 Outreach family planning services 283 2.4 Hospital treatment for stroke or heart attack 286 2.6 Over three-hundred dollars per life year saved Construction of pit latrines and safe water supply 343 0.2 Surveillance and treatment of mild malnutrition at health center 779 0.6 Hospital treatment of AIDS patients 1,314 0.0 Hypertensive treatment at health center 2,381 0.3 26 Table 11. Cost-effectiveness of health services in Guinea, by level of care and future costs from expanded coverage of health services (assuming no improvement in quality of care) Coverage (percent) Costs in thousands of US dollars Level and intervention US dollars Present Planned for Total for Additional per year of 2000 2000 to expand life saved coverage OUTREACH LEVEL AIDS education via media 6 10 50 9 7 Use of impregnated bednets for malaria 43 1 50 1,453 1,424 Tobacco legislation/warnings 65 5 50 10 9 Legislation/fines/seat belts for injury prevention 65 10 50 12 10 School-based vitamin A/iodine/deworming 67 5 25 379 303 Condom use/provision for bar ladies 81 10 50 419 336 Outreach famnily planning services 283 3 5 5,992 2,397 Construction of pit latrines and safe water supply 343 15 60 22,842 17,131 Total ... ... ... 31,115 21,616 HEALTH POST LEVEL Rehydration therapy 8 10 30 III 72 Treatment of children with malaria 8 22 40 419 184 Condom distribution for public 230 3 5 2,248 899 TBA prenatal and delivery care 261 15 25 1,277 511 Total ... ... ... 4,055 1,666 HEALTH CENTER LEVEL Treatment of children with pneumonia 3 45 80 1,504 658 Rehydration therapy for diarrhea 7 20 50 267 160 Short-course treatment for tuberculosis 11 35 50 289 88 Treatment of children with malaria 13 22 40 1,132 497 Childhood vaccination 25 40 80 18,948 9,474 Treatment of sexually-transmitted diseases 67 20 50 2,272 1,363 Educate mothers on childhood pneumonia 108 8 25 81 57 Normal prenatal and delivery care 133 28 50 3,237 1,426 Treatment of adults with pneumonia 134 30 30 476 0 Aspirin in pre-existing CVD 257 5 5 33 0 Surveillance and treatment of Mild malnutrition 779 25 50 11,272 5,636 Hypertensive treatment 2381 16 16 489 0 Total ... ... ... 40,002 19,360 HOSPITAL LEVEL Obstructed birth/Cesarean section 18 10 50 140 112 Treatment of children with severe pneumonia 31 8 25 730 486 Appendectomy 36 30 50 488 195 Treatment of severe malnutrition 42 5 25 1,224 979 Short-course treatment for tuberculosis 44 10 10 256 0 Treatment for complicated measles 48 20 40 115 57 Treatment of rheumatic fever 69 5 5 3 0 Hernia repair 74 20 40 614 307 Treatment for severe diarrhea in children 74 29 50 117 49 Treatment of severe childhood malaria 87 8 25 542 375 Screening blood for transfusion 107 90 99 279 25 Normal prenatal and delivery care 127 37 37 1,297 0 Treatment of adults with severe pneumonia 213 13 13 4,230 0 Treatment for severe chest/orthopedic injury 278 30 30 1,028 0 Treatment for stroke or heart attack 286 20 20 617 0 Treatment of AIDS patients 1314 14 14 40 0 Total ... ... ... 11,720 2,586 Total for all programs 86,892 45,228 Excluding water and sanitation 64,050 28,096 Excluding water and sanitation and hospitals 52,330 25,510 27 Table 12. Total and per capita costs of a minimum package of health services in Guinea (assuming no improvement in quality of care) Coverage (percent) Costs in thousands of Per capita costs in Percent of US dollars US dollars 1994 US dollars total years Level and intervention per year of Present Planned Total for Additional to Total for Additional of life lost life saved for 2000 2000 expand 2000 to expand addressed coverage coverage OUTREACH AIDS education via media 6 10 50 9 7 0.00 0.00 Use of impregnated bednets 43 1 50 1,453 1,424 0.20 0.20 Tobacco legislation/warnings 65 5 50 10 9 0.00 0.00 Legislation/fines/seat belts for injury prevention 65 10 50 12 10 0.00 0.00 School based vitamin A/iodine/deworming 67 5 25 379 303 0.05 0.04 Condom Use/Provision for bar ladies 81 10 50 419 336 0.06 0.05 Outreach family planning services 283 3 5 5,992 2,397 0.83 0.33 Total 8,273 4,484 1.14 0.62 20.3 INTEGRATED HEALTH CENTER OR POST Integrated management of sick child 3 25 80 3,774 2,595 0.52 0.36 Short-course treatment for tuberculosis I1 35 50 289 88 0.04 0.01 Childhood vaccination 25 40 80 18,948 9,474 2.62 1.31 Treatment of sexually-transmitted diseases 67 20 50 2,272 1,363 0.31 0.19 Integrated antenatal care/childbirth/family planing 123 25 50 58,518 29,259 8.08 4.04 Total 83,801 42,779 11.57 5.91 39.8 Total for outreach and health center or post 92,074 47,263 12.72 6.53 60.4 HOSPITAL Obstructed birth/Cesarean section 18 10 50 140 112 0.02 0.02 Treatmentofchildren with severe pneumonia 31 8 25 730 486 0.10 0.07 Appendectomy 36 30 50 488 195 0.07 0.03 Treatment of severe malnutrition 42 5 25 1,224 979 0.17 0.14 Short-course treatment fortuberculosis 44 10 10 256 0 0.04 0.00 Treatment for complicated measles 48 20 40 115 57 0.02 0.01 Treatment of rheumatic fever 69 5 5 3 0 0.00 0.00 Hemia repair 74 20 40 614 307 0.08 0.04 Treatment for severe diarrhea in children 74 29 50 117 49 0.02 0.01 Treatment of severe childhood malaria 87 8 25 542 375 0.07 0.05 Screening blood for transfusion 107 90 99 279 25 0.04 0.00 Total 4,509 2,586 0.62 0.36 2.0 Totalfor allprograms 96 583 49,849 13.34 6.88 62.4 antenatal care is a common service provided at health centers (SEATS/JSI, 1994). The total package of services at the health center or health post level would cost about twelve dollars per capita and would address about forty percent of total years of life lost. Components of a minimum package at the hospital level include: emergency obstetric procedures; treatment of severe cases of childhood pneumonia, malnutrition, diarrhea, and malaria; appendectomy and hernia repair; treatment of tuberculosis; and screening of blood transfusions. As the coverage goals for these are modest, the hospital services would cost only about sixty cents per capita but would address only two percent of total years of life lost. This analysis suggests that for approximately thirteen dollars per capita, a comprehensive set of basic public health and clinical service goals may be offered to Guineans, which would address a large proportion of major causes of premature mortality, and by extrapolation, much of 28 morbidity. Reaching such goals is not possible from the current total public spending of about four dollars per capita. The minimum package presented here offers only a broad guideline to the types of allocation decisions needed to maximize health for Guineans at the lowest possible cost. The package does not include outreach programs such as improved water and sanitation, which cost about three dollars more per capita. While water and sanitation services have externalities far beyond reducing diarrhea and waterborne diseases, and are often demanded as a priority item by people, it is probable that they would be financed from sectors outside health. Furthermore, the package does not include several interventions, such as simple treatment of skin infections, that might encourage more people to use the services as they perceive its benefits. An approximate guideline for such interventions would be to allocate an additional twenty percent (or two dollars per capita) to the cost of the package. Finally, the outreach programs for AIDS education, anti- tobacco campaigns and injury prevention may not be implemented in the short-term as they rely upon widespread literacy and enforcement of legislation. Improvements in community effectiveness of services offered such as better compliance with family planning or with diarrhea treatment, would probably improve the overall cost- effectiveness of health care service delivery. Cost savings may also result from improved quality and coverage from individual interventions, although these are much harder to quantify. As noted, we have made generous adjustments to ideal efficacy. Better quality of services would improve cost-effectiveness. For example, the integrated management of a sick child is not yet commonly practiced in Guinea, but shows a higher cost-effectiveness than do individual programs. This improvement comes from limiting the percentage of missed diagnoses among children presenting with symptoms attributable to diarrhea, malaria, or pneumonia. The minimum package presented includes costs for hospital care. Better Health in Africa (World Bank, 1994) recognized that a well-functioning first-referral hospital is crucial to the entire package of health services delivery. The analysis here suggests that there are several cost-effective health interventions at hospitals. While the per capita figure for curative hospital services is low, it represents a considerable expense per patient. In addition, we assume only modest goals of coverage for in-hospital services. Higher coverage would increase costs considerably. The analysis does consider the current lack of referral system in Guinea. Should primary level care erode further, then costs of coverage for the population at the more expensive hospital levels would rise. POLICY IMPLICATIONS OF COST-EFFECTIVENESS ANALYSIS The cost-effectiveness analysis holds several implications for health planners. First, it suggests that cost-effective interventions are available for the major causes of death in Guinea, and are accessible to its people. Communicable, maternal, and perinatal diseases, the largest causes of death in Guinea, are subject to several highly cost-effective curative interventions. The same is true for preventive programs for injuries and anti-tobacco programs to reduce cardiovascular disease and cancer. Second, the analysis confirms that provision of a basic package of health services will require substantial increases in funding from the government of Guinea and donors. A minimum 29 package costs about thirteen dollars per capita, which is about three times the current public per capita expenditure of four dollars. Third, a focus on the core set of interventions at the outreach, health center and health post levels shown in table 12 should be the top priority for government as it plans further health system reforms. The analysis also has implications for investment patterns. First, it assumes the presence of a large base of investment in infrastructure, drugs, equipment, work force, training, and other delivery components needed to offer a minimum package of services to the population at a particular unit cost. To a large extent, the Guinean government currently focus on building and refurbishing physical structures (World Bank, 1993a; SEATS/JSI, 1994). As well, its goal is to activate all health centers within a short time frame. The proposed minimum package allocates only small amounts to investment items, largely selected equipment. For recurrent budget items, it allocates approximately ninety percent for non salary items, and ten percent for salary items. The requirement for large non-salary expenditure exists even considering that the analyses adjust for the large current inefficiencies in drug and personnel allocation (as reflected in the generous multiplier for these in the analyses) and the current inefficient referral system. In marked contrast, actual recurrent spending in Guinea is about eighty-five percent for salary items and fifteen percent for non-salary items (World Bank, 1996). While a complete change in this spending pattern is impossible, it does suggest that the implementation of minimum package is consistent with previous efforts to inon salaryn-salary recurrent expenditures (World Bank, 1993a; SEATS/JSI, 1994). Cost-effectiveness and burden of disease analyses are only two methods that measure health system performance, and have complimentary roles to other analyses (table 13). A plausible set of variables to measure health system performance include health status, equity or poverty reduction, access to services by the population, macroeconomic and microeconomic efficiency, clinical effectiveness, consumer satisfaction and long-run sustainability (Schieber, 1995). Cost- effectiveness analyses provide information on microeconomic efficiency and clinical effectiveness, by estimating how improved effectiveness may reduce costs. An example is the case for integrated management of the sick child. Although not done here, cost-effectiveness analyses may also provide estimates of the costs for improved access, given that unit costs for incremental expansion of coverage have increasing costs. Finally, they may provide estimates for future costs for coverage, and thus help establish long-run sustainability. Burden of disease analyses provides a measure of health status by providing a static measure of disease burden. It provides limited information on equity, as disease burden is strongly associated with poverty, especially for communicable and maternal diseases. Burden of disease provides limited information on access and microeconomic efficiency, in that lower existing access and poor efficiency should result in high burdens of disease. Finally, projections of future burden of disease may help guide estimating costs for meeting disease coverage, and thus help inform policy makers of the long-run sustainability of current health spending in relation to needs. Table 13 lists the key conclusions from the analyses in Guinea, namely that the health status reflects a high burden of preventable or curable communicable and maternal diseases, and that cost- effective interventions exist that address these diseases. Effective responses to the high disease burden necessitate increases in spending on health, increased clinical effectiveness and quality, all 30 Table 13. Contributions of burden of disease and cost-effectiveness analyses to evaluation of health sector performance Variable Burden of ilisease Cost-effectiveness analyses Some otlher available tools Resultsfor Guinela Health status Covers all diseases and provides N/A Vital registration, cause of High burden from communicable static measure of burden death, morbidity surveys, disease, mostly in under age 5 Equity or poverty Higher burdens are in poor, thus N/A Expenditure Incidence Rural areas likely have higher disease reduction spending prioritized by burden Analyses (or subsidy burden. helps equity impact analyses) Access to services Low access equals higher burden May provide information on Utilization surveys and Poor access means high communicable how incremental increases in physical inventories of disease burden coverage are more expensive facilities Macroeconomic N/A N/A Public expenditure analyses N/A efficiency (overall spending in relation to outcomes) Microeconomic Provides indirect information only: Allows comparison of actual Several: reviews of payment I) Poor efficiency of public facilities is efficiency (including high burdens should reflect poor interventions to ideal practice schemes, referral systems, seen in high burden and low W spending by level of efficiency etc. effectiveness of care care or by disease) 2) Improved quality of care requires more spending on non-salary, recurrent items Clinical N/A Better effectiveness means Performance versus other Better quality is possible with specific effectiveness lower cost-effectiveness facilities inputs, e.g., integrated management of sick children Consumer N/A N/A Consumer surveys satisfaction Long-run Provides indirect information only: Helps estimate future costs for Public expenditure analyses I) Future burden of disease will be sustainability estimates of burdens from coverage of a minimal package driven largely by fertility and not population growth and the epidemiological transition epidemiological transition 2) Costs for the minimum package alone are three times current public .__ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ sp e n d in g of which may be done by focusing on a core package of clinical and preventive interventions. Fortunately, methods complementary to burden of disease and cost-effectiveness have examined health system performance. These include utilization surveys, public expenditure reviews, expenditure-incidence analyses (SEATS/JSI, 1994; World Bank, 1996). However, other aspects of system performance remain unmeasured, including vital registration, cause-of-death surveys, reviews of payment and finance schemes, and consumer surveys of satisfaction. The cost-effectiveness analyses reported here have several limitations. First, they do not provide a direct analytic basis for managing issues such as decentralization, drug delivery, personnel planning, infrastructure development, quality of care, and other aspects of health care delivery. Second, as the analyses depend upon data for both the numerator and denominator, even minor misclassification of either could greatly alter the estimated ratios. However, such error likely applies across all interventions and is below the level of error that would distort the decision to allocate resources across levels. Third, while these interventions address major causes of death in Guinea, their full implementation will not necessarily reduce deaths dramatically for the population as a whole. Allocation of resources toward an intervention is generally on a marginal basis and results in slow, modest reductions in overall mortality. In addition, the estimates of disease-specific mortality do not rely upon rigorous epidemiological data and thus are likely to contain measurement error. For this reason, we have not calculated the reduction in mortality expected from these interventions. Fourth, some interventions that do not appear cost-effective in this analysis, for example, provision of adequate water and sanitation, are nonetheless important in improving the general quality of life in a community. Fifth, these cost-effective analyses do not take into account private costs incurred by a patient, such as the costs of travel or of time lost from work, and the like. The analysis reflects the actual practice in Guinea, rather than an ideal effectiveness of interventions. Significant reductions in total costs (and thus in per capita costs) would result from enhanced community effectiveness of the interventions. Such effectiveness involves improved access, compliance, drug supply, and training. Quantifying the magnitude of such internal efficiency gains is difficult. Improving quality requires additional expenditure on training, planning, and management. These, in turn, tend to require non-salary recurrent expenditures. To guide policy choices, the minimum package requires testing in focal areas, and such testing must include careful observation of costs and consequences, and must be accompanied by collection of better epidemiological data, health service utilization, and consumer satisfaction surveys. Guinea has a strong record of innovation. Such piloting, drawing upon similar experiments in other countries, would be a useful guide to widespread implementation of a program of cost-effective interventions addressing the major causes of death and disability in Guinea. 32 BIBLIOGRAPHY Ainsworth, M. 1994. Socioeconomic determinants offertility in sub-Saharan Africa. Washington, D.C.: World Bank. Ansley, J., Demeny, P. and Vaughan, B. 1982. Regional Model Life Tables and Stable Populations: Second Edition. New York: Academic Press. Bobadilla, J.L. et al. 1993. "The Epidemiological Transition and Health Priorities." In Jamison D.T. et al., eds., Disease Control Priorities in Developing Countries. Oxford: Oxford University Press. David, P., Kawar, S. and Graham, W. 1991. "Estimating maternal mortality in Djibouti: an application of the sisterhood method." International Journal of Epidemiology 20:551-57. DNSI (Direction Nationale de la Statistique et de l'Information). 1994. Enquete Demographique et de Sante, Guinee, 1992. Conakry: DNSI. Drummond, M.F., Stoddard, G.L. and Torrance, G.W. 1987. Methodsfor the economic evaluation of health care programmes. Oxford: Oxford Medical Publications. Feachem, R-G., Jamison, D.T. and Bos, E.R. 1991. "Changing Patterns of Disease and Mortality in Sub-Saharan Africa." In Feachem, R.G. and Jamison. D.T., eds., Disease and Mortality in Sub- Saharan Africa. Oxford: Oxford University Press. Ghana Health Assessment Project Team. 1981. "A quantitative method of assessing the health impact of different diseases in less developed countries." International Journal of Epidemiology 10:73-80. Gray, R.H., Smith, G. and Barss, P. 1990. The use of verbal autopsy methods to determine selected causes of deaths in children. Johns Hopkins Occasional Paper No 10. Baltimore: Johns Hopkins University. Jamison, D.T., Mosley, W.H., Measham, A.R. and Bobadilla, J.L., eds. 1993. Disease Control Priorities in Developing Countries. Oxford: Oxford Medical Publications. Kitange, H.M., Machibya, H., Black, J. et al. 1996. Outlook for survivors of childhood in sub- Saharan Africa: adult mortality in Tanzania. British Medical Journal 312:216-20. Laporte, R.M., Tull E.S. and McCarty, D. 1992. "Monitoring the incidence of myocardial infarctions: Applications of the capture-mark-recapture technology." International Journal of Epidemiology 21:258-63. 33 Murray, C.J.L. 1994. "Quantifying the burden of disease: the technical basis for disability- adjusted life years." Bulletin of the World Health Organization 72:3: 429-45. Murray, C.J.L. and Lopez, A.D. 1994. "Global and regional cause-of-death patterns in 1990." Bulletin of the World Health Organization 72:3:447-80. Murray, C.J.L., Lopez, A.D. and Jamison, D.T. 1994a. "The global burden of disease in 1990: summary results, sensitivity analysis and future directions." Bulletin of the World Health Organization 72:3:495-509. Murray, C.J.L., Govindaraj, R. and Chellaraj, G. 1994b. Global Domestic Expenditures in Heath. World Development Report 1993 Background Paper Number 13. Washington D.C.: World Bank. Pelletier, D.L, Frongillo, E.A., Schroeder, D.G. and Habicht, J.P. 1994. "A methodology for estimating the contribution of malnutrition to child mortality in developing countries." Journal of Nutrition 124:2106S-22S. Scheiber, G.J 1995. "Preconditions for health reform: experiences from the OECD countries." Health Policy 32:279-93. SEATS/JSI ("Family Planning Service Expansion and Technical Support"/ John Snow Inc.). 1994 Health and Population Assessment: Republic of Guinea, March-April, 1994. Report to U. S. AID Shaw, R.P. and Griffen, C.C. 1995. Financing health care in sub-Saharan Africa through user fees and insurance. Washington, D.C.: World Bank. Shepard, D.S. and Thompson, M.S. 1979. "First principles of cost-effectiveness analysis in health." Public Health Reports 94:535-43. Sow, O., Diallo, M.P., Camara, B. et al. 1990. "Evaluation du risque annuale d'infection tuberculeuse au moyen d'une enquete tuberculinique dans la ville de Conakry." Revue Maladie Respiratoire 7:355-60. US Bureau of the Census. 1994. Recent HIV Seroprevalence levels by Country: June 1994. Research Note 13. Washington, D.C.: US Bureau of the Census. Weinstein, M.C. and Stason, W.B. 1977. "Foundations of cost-effectiveness analysis for health and medical practice." New England Journal of Medicine 296:716-21. World Bank. 1986. Guinea, Population, Health and Nutrition. Sector Report No 6229-GUI. Washington, D.C.: World Bank. 1987. Financing Health Services in Developing Countries: An Agenda for Reform. World Bank Policy Study. Washington, D.C.: World Bank. 34 * 1993a. Staff Appraisal Report, Republic of Guinea, Health and Nutrition Sector Project. Report No 11684-GUI. Washington, D.C.: World Bank. . 1993b. The World Development Report 1993: Investing in Health. New York: Oxford University Press. -. 1994. Better Health in Africa. Report No. 12577-AFR. Washington, D.C.: World Bank. 1995. Eastern Africa Burden of Disease Cost-effectiveness, and Health Policy Study. Report No EA 14040. Washington, D.C.: World Bank. 1996. Republic of Guinea: Public Expenditure Review. Report No 15147-GUI. Washington, D.C.: World Bank. World Health Organization. 1996. Investing in Health and Research Development: Report of the Ad Hoc Committee on Health Research Relating to Future Intervention Options. TDR/Gen/96. 1. Geneva: World Health Organization. 35 Distributors of ^ Pq_IeSA naauuml 04 R bue Wlmeien ,Cu iiapaUWA, V,, Gleam E5eedu* Distributors oF l fR- & E# n 36, Samevr Sb. ye DM Di C.w 1/1 Do*k Dn W I I Y AIR, MM CiLdl-411 World Bank Pi Hm 1ees12Aats ce-52P- p. BS612 A PA C1,4117 8. . 4 Da Fo Si DaV Jo TL (1) 34-16 50125 Ftu N606 OD 5 PEt Pie. LW. T: (01)) 943273 Publications owe F (1) s4a254 Tl u5s5eo4s 4W: s5743 41KalsgPUdft R 04-3 Fax (021)9433805 Prkcs and crdit tems vary t 0se) 10ss Fa CM) 641 257 Fa (22) 68-1901 G Whed ed pm co_ntry to anntry Far 6) 10401-7355 HON KONG, MACAO Sgne 34318 TANAIA Cionsut ur tocad disinlntor AA 200D LWL JAMAICA PAKISTAl :T (65)7415160 OxkId L*Ud PU bejbrt youri lo. it-'btr CL Sin & ceudob Dapitee hi Rude Pubaiue LteaW BookAAgnc Fax (65) 742-936 MWMb Sues before pElacg an orderr . . asL Stud Hmm, iun 1101-4 2a6 Old Hwe Road 65, SIulvt-e uvd-Anrm eat ghpbOahumoee m P0 52t9 ApboAmn 34270 22 W Ses, Ca"l Kt 6 P.O. B No. 729 Der e Seni iT (1) 2-27mS Figr KTi T* C-3M LaIn1,1 040C SLOVAK REPIBJE Tit (51) 2M2 AUAM Far (1) S26 Tat 8522530-l401 Far 80977.843 TW: (42)735301 m 00.T.G. LW. Fax (51)46120 Am ULt Far 522526-1107 Far (42)775283 Kdnu4 PaaW 3 Ut COTE DWO URt hNp9Jin6as9 .rnm mJ* JAPAN PO Box 152 THALAg PA 9, SIL 1, Ap. 4 Cut dtEdm d de Diusin Esar Bak Sta Oadm Undrity Pins 852 99 6indv 5 Cnl Beta Db net Tlb Aliht CEDA) HUNGARY Hap 3ahne. 5 Bngpl Town Td: (7) 83472 308 Slam Rood Tk (42)27419i22172 04 LP. 541 Faatblbfor MBakyo-ku 113 Sham Fid Far (7) 839485 B kk Far (42)27419 Abi 04P Eancey Towo P Box 13033 Tt (2)235-S4t0 i* 225-24-0510 Dmtoed Ut 17-19 Td: (03)3818-0861 KasdhF75350 SOUTH AFRICA, rIWAA Far (2)237-01 ARIUIMA Fac 225,45 67 H1117Bx Far (D3)3919484 Tl: (21)448307 For pb on. dd Lbm hra it 31204 2951 or ULRL hr 9tiJeħ*Muo sjtsr-a Far (21) 454-7640 Odord U riaw Puas TM0DAD TOBAG0, JAIM A. Caie 1977 CYPRUS 36 1 204 2948 E-nail: oupaoup.en1 e.un$comni Sot lAte Sy_awnu Sbn Lt 113109tumesAie Cedeor Apped Rmardh Far 361 204 2953 KBIYA P.Box 1141 4t S rbes iot (1)8114186 CWnuCelu Afria Book Swvta (EA) LWd. PERU) Cwe Tom MM Ccuqt Far (1) 815J4 6. GDug Sbe, Ereigt MM Qur ftm, MXnArto OSba Editaril DeIrr2D SA Te: (21)45-7266 Tnrti, Wed kl P.0. k2D6 A2ied PtUdli Li. P.. Bx 45245 Aptl*do324 Fax: (21)45-7265 Tel: 909-62-5854 A1.1M1RALIA, RLI, iOUA NEW SMEA, N' e 751 Meal Roed Nei Lirr. I Far 0809--5554 9OU4UNLANDS.UUJAJIJAND TdU 1 (2)44-173D Mae-60t aD02 Tl: (2)23E41 Tt: (14)2853E0 Fbrmubwb aord:e WESTRN smOA Fa (2)462051 Te: (44)852-3808 Fac (2) 30272 Fax (14)228682 miesnali Subtdtni Sante U3ANDA DA Wbsnli Swice Far (44)852-09 P.O. Box 410 52ub LW. 648 Wlhuas Red CE REPUBC KOREA, REPUBUC OF PHUNES CM"ld Bl u t tt lhn 3132 NdbD ituul Cerer ONEStA Dem Traeg Co. Ld. ktertalmeal Bedtaouace Cener Inc. Jdwa g 2024 PO Bo 9997 Vkb . ormildssp,lta 5 PLh iat LinS P.O. Be 34 Stle 720, C'ytl 10 Tel: (11)80-1448 PD1t4 JAija Rd. Tit (61)392107777 CS -11357 Pe I JAb Bombdf 20 YaDin Caidwrtt Tower 2 Fae (11)8 O4248 Kla'np Far (61)392107788 T* (2)2422-9433 PO. BO 181 Seaut H.V ddEi Con, mau Ti/Far (41)254763 URL:V hWJAmuw edsAu Far (2) 2422-1484 JeSt. 10320 Te: (2) 785-131/4 Vw St. SPAN URIt hprwSaJcd T: (21)366-4280 Far (2) 784-0315 MaIbbm,oo Muk ILatferma Librs, SA UNrrED KNWOM AUSTA Fax (21) 4214289 Td: (2) 817-9676 Ce1e11 37 Micibt Lid GuiddnCo. DENtCK MIALAYSIA Far (2)817-1741 28801 MAt PO. Oa 3 GIn 31 Su,dwfLajr MM Uret cis Wdey COxp e Td: (1) 431-4 Alabn, Hunpl GU342PG A-101 Wun RessincAblA1 Kw bS PUAders Bmakhp, ULh d POLAND Fax (1)5754986 Engd Ta (1) 512-47-314 DK-1970 Frdlbee C P.O. Box 19575-511 PO. a 1127 TIbrriu PUbiSlg Sanvie itrieS.ri.eafdla Te: (1420) 86848 Far (1)512-47-31-29 Te: (31)4151942 ats dan Pwi Baru Ul. Plai 31M37 Far (142D)89889 Far (31357822 TeL (21) 2a-3723 59700 Kuela Lumwor 00-577 Warzan LrAd-Pina BsnMrib BUtABEI8 Fax 98(21)258-723 TW: (3)h785000 Te: (2)6286089 Card de Ce 391 ZASA 1t. DaGe rt EGYFl AMB RPUSJC OF Far (3) 7554424 Farx (2) 821-7255 08009 Barelona UtwaMdy BDockp A _ Soedy (IIDAS) Al lwn Di_tn Agency Ktb Sue Co. Putos Td: (3) 488-30C9 Grad Ead Food CenipL Ham5, Rd 1 Al Gin SuWe thied F -eS Ave., REXJCO PORTUGAL Fax (3) 487-7659 P.O. Box 32379 Dwxed WAne Carto SO es t NFOTEC Littt PIod Luee Dhba 12i9 T ) 5788003 KuhS Delhdre No.8 AptdoPosPabl 22-80 Rua DoCaeio 70-74 SRI LANKA, TMNMA ES Td: (1)213221 ExL 482 Td O232427 Far (2)57433 Td1es 14060 Tlalen, 120 LieSn Lake Hol Boebhp Far 02)81118 Te: 8717819or8716104 MidcoF.D Ti: (1)347-4982 P.O. B244 Z ABWE The kdds ESCdOswr Far 8862479 Tel: (5)a60E411 Fax (1)347-264 100, 5ir C hbmpalam A. Lalen b(bee (Ple.)L SE 41, Shutr Shes Erd: keab-eatOnede.nd.ir Fax (5) 824-2622 Gusr Mw attet Turet Road, Ardheni Jtm DBe -w Can ROMINIA Cdweto 2 PO. Box ST125 ft du Roi202 Tel: (2)393-732 RELAND NETHERIANDS CarqpniDe Ltati teuremeoSA Tel: (1)32105 Soualteron 136wtIRe Far (2)3939732 GerrteetStppsAeDcy De. LiiutelrvlnOr-Pdtthtes Str. Liscani rio. 26, sedr 3 FarK (1)432104 Hse Tit 2)538-6169 Oii6g antSdA.u P.O. Bx 202 Beutest Tel: (4)6215617 Fac (2)S391-41 F14LAND 4-StAt4-5Huoul Rd 7480AE laaks-gen Tel: (1)6139645 SWEDEN Fax: (4)621670 Alesinen KrpaLps Dubin 2 Tel: (53) 574-240 Far (1) 312 4000 Friies Cudru Servee BRAZL PO. Box128 Te: (1)461-3111 Far (53)5724296 RPapwgele12 P Titaee S T/Ma hwasta FIN-U0101 Hues Far (1)475-2670 RUSSIAN FEDERATIO 5-10647 S5atoin LI _ Tel: (0)12141 NEW ZEALAND datolelbto Vees Mir> Te: (8)6909080 ls Pi G ar, 209 Far (0) 1214441 tRAEL EBSCO NZ Ld. Be, Kolpchniy Pereulk Far (8) 21 47 77 01409 Sho PALtI, SP. URV haptJbeknzc*euldik Yowmot Literamre Ld. Piate Mdl Beg 99914 MD5CDU 101831 Tit (11) 259944 P.O. Box 5055 New MeLm Tel: (95) 917 97 49 Wernergen-W4lus AB Far 411)258499 FRAICE TeldAv6155D Ackatnd Farx (95)9179259 PO.Box1305 URL:h gtJ/uAdOr World Btd t Pttabon TW: (3)5285-397 Tel: (9)524-8119 5-17125 Sobes CW uvtj dl% Far (3) 5285-397 Far. (9) 524-07 SAUDI ARABIA, OATAR Tel: (8) 705-97-50 CANAMI 751161Pafis Jatir Beet Stoem Fox (8( 27-00-71 Rand PL g Co. Ld Td: (1 ) 40-0-4 57 R.O.Y. Inbernatal NIGERIA PO. Box 3196 1284 A mt Pad Far (1) 406-440-8 PO Box 13058 Urleesy Pewen tr/ted Ry.dh 11471 SWITZERLAND 011,aOnsl KiB3W! bA* 61130 ThmreCrownsBuldkgJeudta Tel: (1)477.3140 Lraie Payel it 613-7414333 GERAY Tel: (3)5481423 Pent.Mel Bag 5095 Fax: (1) 477-2940 Srvice lrbtrabonel Fa613-741-5439 UIO-Vs,o Far (3) 5461442 bd CStIe-de-Morrtan 30 IRFtL h /wbklnral-nau PrapdkrAVe 55 Tel: (22)41-1356 1002 Laueanne 53115 Ban Paeetlu eaoteA bdyle Eat Far (22)41-2056 Tel: (521)341-3229 Tt (228)212I40 hwex k*txl_;nt Sereee Fax (021)-341435 Far (4217492 P.OB. 19502 JAna_ Tel: (2 271219 RECENT WORLD BANK TECHNICAL PAPERS (continued) No. 295 Pohl, Jedrzejczak, and Anderson, Creating Capital Market5 in Central and Eastern Europe No. 296 Stassen, Small-Scale Biomass Gasifiersfor Heat and Power: A Global Review No. 297 Bulatao, Key Indicatorsfor Family Planning Projects No. 298 Odaga and Heneveld, Girls and Schools in Sub-Saharan Africa: From Analysis to Action No. 299 Tamale, Jones, and Pswarayi-Riddihough, Technologies Related to Participatory Forestry itn Tropical and Subtropical Countries No. 300 Oram and de Haan, Technologiesfor Rainfed Agriculture in Mediterranean Climates: A Review of World Bank Experiences No. 301 Mohan, editor, Bibliography of Publications: Techinical Department, Africa Region, July 1987 to April 1995 No. 302 Baldry, Calamari, and Yam6ogo, Environmental Impact Assessment of Settlement and Developmtetnt in the Llpper Utraba Basin No. 303 Heneveld and Craig, Schzools Cou nt: World Bank Project Designs and the Qlualiti of Primary Ediucation in Sub-Saharan Africa No. 304 Foley, Photovoltaic Applications in Rural Areas of the Developing World No. 305 John,son, Education and Training of Accoutitants in Sub-Saharan Anglophone Africa No. 306 Muir and Saba, Improving State Enterprise Performance: The Role of Internal and External Incentives No. 307 Narayan, Toward Participatory Researcih No. 308 Adamson and others, Energy Use, Air Pollution, anid Environmental PoliLy in Krakow: Can Economzlic Incentives Really Help? No. 309 The World Bank/FOA/UNIDO/Industry Fertilizer Working Group, Worli and Regional Supply and Demand Balances for Nitrogen, Phosphate, and Potash, 1993/94-1999/2000 No. 310 Elder and Cooley, editors, Suistainable Settlement and Development of thle Oncihocerciasis Control Programme Area: Proceedings of a Ministerial Meeting No. 311 Webster, RiopeLe and Chidzero, World Batik Lendingfor Sm all Enterprises Z989-1993 No. 312 Benoit, Project Finanice at the World Bank: An Overviezo of Policies and Instruments No. 313 Kapur, Airport Infrastriuctu re: The Emerging Role of the Private Sector No. 314 Valdes and Schaeffer in collaboration with Ramos, Surveillance of Agricultural Price and Tradle Policies: A HandbookJfor Ecuador No. 316 Schware and Kimberley, Information Technology and National Trade Facilitation: Making the Most of Global Trade No. 317 Schware and Kimberley, Informiationi Technology and National Trade Facilitation: Gu ide to Best Practice No. 318 Taylor, Boukambou, Dahniya, Ouayogode, Ayling, Abdi Noor, and Toure, Strengthlening National Agricultural Research Systemis in the Humidl atn d Sub-humid Zones of West and Central Afric-n: A Frame-workfor Action No. 320 Srivastava, Lambert and Vietmever, AMedicinal Plants: An Expianding Role in Development No. 321 Srivastava, Smith, and Forno, Biodiversity and Agricu ltu re: Implicationsfor Conservation and Development No. 322 Peters, The Ecologly an7d Management of Norn-Timilber Forest Resolurces No. 323 Pannier, editor, Corporate Governance of Public Enterprises in Transitiorial Economies No. 324 Cabraal, Cosgrove-Davies, and Schaeffer, Best Practices for Photovoltaic Holusehold Electrificatiotn Programs No. 325 Bacon, Besant-Jones, and Heidarian, Estimzating Construction Costs and Schedules: Experience w!ithi Power Generatior Projects in Developing Countries No. 326 Colletta, Balachander, Liang, The Condition of Young Childiren in Sub-Saharan Africa: The Convergence of Health, Nuitrition, and Earlil Eduication No. 327 Valdes and Schaeffer in coLlaboration with Martin, Surveillance of Agricultz ral Price and Trade Policies: A Handbook Jor Paraguiay No. 328 De Geyndt, Social Development and Absolute Poverty in Asia and Lititz America No. 329 Mohan, editor, Bibliography of Publicationis: TeciihnicIl Departmelit, Africa Region, Ju ly 1987 to April 1996 No. 332 Pohl, Djankov, and Anderson, Restru ctu ring Large Indlustrial Firms in Central aid Eastern Europ.e: An Empirical Analysis No. 333 Jha, Ranson, and Bobadilla, Mveasuring thie Burien of Disease and tie Cost-Effectiveness of Health Interventions: A Case Study in Guin1ea II- 0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ I.,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~l in.. ; ii - I - - -