r 2 6 26893 September 2003 - ORIENTATIONS IN DEVELOPMENT SERIES HIV/AIDS in the Middle East and North Africa The Costs of incarC lCion Carol Jenkins and David A. Robalino V :7~~~V- . . .~r iaS~; HIV/AIDS in the Middle East and North Africa ORIENTATIONS IN DEVELOPMENT HIV/AIDS in the Middle East and North Africa The Costs of nacu3on Carol Jenkins David A. Robalino THE WORLD BANK Washington, D.C. © 2003 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved. 1 2 3 40605 0403 The findings, interpretations, and conclusions expressed here are those of the author(s) and do not necessarily reflect the views of the Board of Executive Directors of the World Bank or the gov- ernments they represent. The World Bank cannot guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply on the part of the World Bank any judgment of the legal status of any territory or the en- dorsement or acceptance of such boundaries. Rights and Perniissions The material in this work is copyrighted. No part of this work may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or in- clusion in any information storage and retrieval system, without the prior written permission of the World Bank. The World Bank encourages dissemination of its work and will normally grant permission promptly. For permission to photocopy or reprint, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, World Bank, 1818 H Street, NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. Cover photo: ©Dave Bartruff/CORBIS ISBN 0-8213-5578-3 Library of Congress Cataloging-in-Publication Data has been appliedfon. Preface ix Acronyms and Abbreviations xi Executive Summary xiii Introduction I 1. Global Experience with HPDV/AIDS 3 Note 11 2. Review of the HOWAIDS Situation in the MENA/EM Region 13 Need for Second-Generation Surveillance 17 Notes 20 3. A Typology of Risk Factors 25 At-Risk Groups 25 Injecting Drug Users 26 Sex Workers 29 Men Who Have Sex with Men 31 Vulnerable Groups 32 Migrants 33 Youth 36 STI/STD Patients 40 Structural Vulnerability 42 Notes 45 v vi Contents 4. Assessing the Potential Economic Impact of HOV/AIDS in the MENA/EM Region 47 Exploratory Analysis of the Socioeconomic Determinants of HIV/AIDS Prevalence Levels 48 Potential Socioeconomic Consequences of the HIV/AIDS Epidemic 52 Impacts on Health Fxpenditures 61 Implications for Poverty Reduction 62 Policy Implications 64 Notes 67 5. Responses to 111V/AIDS 69 National Health Systems 70 Other Sectors 73 International and Local Nongovernmental Organizations 74 Private Sector 77 United Nations Agencies 77 Bilateral Donors 79 6. A WayForward 81 Raising the Priority of HIV/AIDS 81 Reaching Hidden Populations 84 Reducing the Vulnerability of Migrants 86 Providing Practical Information and Means of Prevention 87 Reducing the Vulnerability of Youth 88 Care and Support 88 Structural Factors 89 Working with Communities 89 Monitoring and Evaluation 90 The Time Required 90 Financing the Effort 90 Immediate Next Steps 91 Technical Appendix 93 References 131 Index 169 Boxes 1.1 Nepal: Too Little, Too Late 6 1.2 Overlooking the Epidemic: The Case of Indonesia 7 Contents vii 1.3 HIV Behavioral Research in a Conservative Society: Experience in Bangladesh 9 2.1 HIV Epidemic Profiles in the MENA/EM Region 19 3.1 Injecting Drug Users in Egypt 28 3.2 Men Who Have Sex with Men, in Egypt 32 3.3 Young People in Egypt 39 5.1 Situation and Needs Assessment in the Republic of Yemen 72 5.2 VCT and Harm-Reduction Programs in the Islamic Republic of Iran 75 5.3 Tunisia: Youth and Condoms 76 6.1 Strategic Planning for HWV/AIDS in Morocco 82 6.2 Reaching MSM in Morocco and Bangladesh 85 Figures 1.1 The Hidden Epidemic in Nepal: HIV Prevalence among Injecting Drug Users and Sex Workers in Kathmandu 5 1.2 HIV Prevalence in Blood Donors in Indonesia 8 2.1 Total Reported AIDS Cases per 100,000 People in Selected Countries in the MENA/EM Region, with East African Comparisons 14 3.1 Evolution of the Seroprevalence of HTV among Bar Hostesses and Sex Workers between 1987 and 1996 in Djibouti 31 3.2 Estimated Annual Incidence of STIs: Africa, Middle East, and Djibouti 42 4.1 Poverty, Inequality, and HITV/AIDS Prevalence 49 4.2 Gender Inequality and HIV/AIDS Prevalence 51 4.3 Potential Underestimation of Current HITV/AIDS Prevalence in Selected MENA/EM Countries 53 4.4 HIV/AIDS Prevalence and Worker Turnover 54 4.5 Alnual Cost of Treating an AIDS Patient versus Annual Cost of Primary Education 55 4.6 A Model of Growth to Evaluate the Macroeconomic Impacts of HIV/AIDS 57 4.7 Illustration of Diffusion Profiles for the HrV/AIDS Epidemic in MENA/EM Countries 58 4.8 Jordan: Plausible Predictions for HTV/AIDS Prevalence 59 4.9 HFV/AIDS-Related Average Health Expenditures in 2015 62 4.10 Number of People in the MENA/EM Region Who Would Remain below the Poverty Line as a Consequence of NIV 64 viii Contents 4.11 Benefits from Expanding Access to Condoms and Safe Needles for IDUs and Costs of Delaying Action 66 A. 1 Distribution of GDP Losses (2000-25) Resulting from the HIV/AIDS Epidemic 110 A.2 Distribution of Reductions in the GDP Growth Rate (2000-25) Resulting from the HIV/AIDS Epidemic 113 A.3 Distribution of Reductions in the Labor Force in 2025 116 A.4 Distribution of the HIV/AIDS Prevalence Level in 2015 119 A.5 Distribution of HIV/AIDS-Related Health Expenditures in 2015 122 1.1 Global Regional Variation in HIV Prevalence, December 2002 4 2.1 Reported AIDS Cases in the MENA/EM Region, 1987-2001 15 2.2 Profiles of the HIV/AIDS Epidemics in Countries of the MENA/EM Region 21 3.1 Selected Indicators of Development by Country and Human Development Index 43 4.1 Population Living on Less than US$2 a Day, by Region 50 4.2 Economic Impacts of the HIV/AIDS Epidemic 60 A.1 Estimates of Model 1 94 A.2 Estimates of Model 2, Excluding Female Participation in the Labor Force 94 A.3 Estimates of Model 2, Including Female Participation in the Labor Force 95 A.4 Model Parameters 104 A.5 Average Number of Partners and of Occurrences of Intercourse per Year 108 A.6 HIW/AIDS Transmission Probabilities 108 A.7 Descriptive Statistics for Output Variables: Status Quo 109 A.8 Two Classical Interventions: Distributing Condoms and Needles 125 A.9 Unit Costs for Needle Distribution Intervention 126 A.10 Output Variables under Policy Interventions 127 A.11 Output Variables if Policy Intervention Is Postponed for Five Years 128 Preface This volume is the product of a joint exercise among the World Bank, the Joint United Nations Programme on HI`V/AIDS (UNAJDS), and the Eastern Mediterranean Office of the World Health Organization (WHO/EMRO). It was produced under the direction of Maryse Pierre- Louis, Lead Public Health Specialist, and Francisca Ayodeji Akala, Pub- lic Health Specialist, Middle East and North Africa Region, Human De- velopment Sector, the World Bank (MNSHD); by Carol Jenkins, World Bank consultant and United States Agency for International Develop- ment (USAID) Asia/Near East Regional HIV Adviser; and David A. Robalino, Senior Economist, MNSHD, World Bank. Jacques Baudouy, Sector Director, MNSHD, and George Schieber, Principal Economist, Health and Social Protection Manager, MNSHD, provided valuable comments to improve the document. Peer reviewers were Debrework Zewdie, World Bank Global AIDS Adviser; Mead Over, Senior Econo- mist, World Bank Development Economics; Oussama Tawil, Team Leader, UNAIDS Inter-Country Team for MENA, Cairo; and Jihane Tawilah, Regional Adviser, AIDS & Sexually Transmitted Diseases, WHO/EMRO. Henriette Folquet and Darcy Gallucio, Team Assistants, MNSHD, provided support for the presentation of the document. Because fieldwork was limited, gathering information from credible sources was of utmost importance. After presentation of the main find- ings of this study at the MENA Regional Public Health Conference in Beirut in June 2002, as well as at the XrV International AIDS Confer- ence in Barcelona, July 2002, a substantial effort was made to update its contents for publication. Between February and May 2003, numerous people helped with this effort. We are particularly grateful to Jihane Tawilah for providing the most recent epidemiological data for the EM Region and to Oussama Tawil for providing up-to-date information on the changing responses in the region and reviewing contents in great de- tail. In Geneva, Neff Walker (UNAIDS) and Ties Boerma (WHO) re- sponded rapidly to our inquiries. ix x Preface Country visits, though limited to Djibouti, the Arab Republic of Egypt, Tunisia, and the Republic of Yemen, were invaluable. In those countries several people were particularly helpful. We would like to mention the following: In Djibouti: Pierre Luigi, World Food Programme; Samira Ali EHgo and Michel Etchepare, World Bank consultants; and Dr. A. Sow, WHO. In Egypt: Dr. Saleh, WHO/EMRO; Dr. Cherif Soliman, Family Health International; Dr. Sussan Bassiri, WHO/EMRO; Wolfgang Schiefer, United Nations Office of Drug Control and Crime Prevention (UNODCCP); Dr. Maha Aladowy, Ford Foundation; Karima Haluf, Population Council; Dr. Alaa Hamed, World Bank; Dr. Nasr el-Sayed, National AIDS Control Programme; and Mark White, USAID. In Tunisia: Dr. Mourad Ghachem, United Nations Population Fund (UNFPA); Helene Zoughlami, UNFPA; Dr. Amel Ben Said, Ministry of Public Health; Dr. Zied Latiri, Director of Communication, National Office of the Family and Population, Ministry of Public Health; Eliza- beth Bennou, International Planned Parenthood Foundation; Ariel Francais, United Nations Development Programme (UNDP); and Fatma Felah, World Bank. In Yemen: Amer Majeed Abdul, UNAIDS/UNDP; Dr. Rogers Busulwa, International Cooperation for Development; Hashem Awnal- lah, World Bank; and Abdul Wahab Anisi, National Centre for Health Education. Acronyms and Abbreviations AIDS Acquired iinmunodeficiency syndrome ART Antiretroviral therapy AZT Zidovudine, an antiretroviral drug BSS Behavioral surveillance surveys CBO Community-based organization CIS Commonwealth of Independent States ELISA Enzyme-linked immunosorbent assay EMRO Eastern Mediterranean Regional Office, WI-HO FSW Female sex worker GDP Gross domestic product GFATM Global Fund to Fight AIDS, Tuberculosis and Malaria GNP Gross national product HIV Human immunodeficiency virus ICDDR, B International Centre for Diarrhoeal Research, Bangladash IDP Internally displaced persons IDU Injecting drug user INGO Tnternational nongovernmental organization IOM International Office of Migration IPPF International Planmed Parenthood Foundation KAP Knowledge, attitude, practice MENA Middle East and North Africa Region MNSHD Middle East and North Africa Region, I-luman Development Sector, the World Bank MoH Ministry of Health MSM Men who have sex with men (or males who have sex with males) MTCT Mother-to-child transmission NAP National AIDS Program NGO Nongovernmental organization PLWHA People living with HIV and AIDS pvGDP Present value of gross domestic product xii AcroRyms and Abbreviations RTI Reproductive tract infections STD Sexually transmitted disease STI Sexually transmitted infection TB Tuberculosis UAE United Arab Emirates UNAIDS Joint United Nations Programme on HIV/AIDS U,NDP United Nations Development Programme UNFPA United Nations Population Fund UNHICR United Nations High Commissioner for Refugees UNICEF United Nations Children's Fund UNODC United Nations Office on Drugs and Crime UNODCCP United Nations Office on Drug Control and Crime Prevention (now UNODC) UNGASS United Nations General Assembly Special Session on HFIV/AIDS USAID United States Agency for International Development VCT Voluntary counseling and testing WB The World Bank WHO World Health Organization This book reviews the human immunodeficiency virus/acquired immun- odeficiency syndrome (HIIV/ATDS) situation in the Middle East, North Africa, and Eastern Mediterranean (MENA/EM) region, and is intended to stimulate discussion and promote dialogue among the region's policy- and decisionmakers. It seeks to provide a framework for multisectoral strategic action to reduce behaviors that risk spreading HIV, to care for and support those who become infected, and to diminish vulnerability among specific segments of society. Although most evidence suggests that overall HIV prevalence is low in the region, greater investmnents in improved surveillance, prevention, and care are needed now-to main- tain low prevalence levels and preserve the focus on national and regional development goals. The production of this volume has been a joint effort of the World Bank, the Eastern Mediterranean Regional Office of the World Health Organization (WHO/EMRO), and the Joint United Nations Pro- gramme on HIV/AIDS (UNAIDS). Interviews were conducted with 51 people-including bilateral donors and those in government, U.N. agen- cies, and international nongovernmental organizations (INGOs)-in four countries of the region (the Arab Republic of Egypt, Djibouti, Re- public of Yemen, and Tunisia) and in Geneva between August 18 and September 20, 2001. Newly updated information, as of the beginning of 2003, has been incorporated. Documents were gathered from all perti- nent sources, especially WHO, UNAIDS, U.S. Census Bureau NTV/ATDS Database, and the United Nations Office on Drugs and Crime (UNODC). Because full information for every country in the re- gion was not available from any of these sources, there are inevitable gaps and possible inaccuracies. Global Experience Responding to HIV/AIDS In 1991, with the evidence available at the time, experts estimated that 15 million to 20 million adults and 5 million to 10 million children cumula- xiii xiv Executive Summary tively would become infected with HIV by 2000. By the end of 2002, UNAIDS/WHVO reported that 42 million people were living with IATV/AIDS and 5 million new infections had occurred during the year. The substantial gap between earlier projections and current estimates re- flects both the unexpected spread of the virus and the inadequacy of sta- tistics used to track the epidemic. FI-V does not always spread rapidly, though it certainly can if conditions permit. Rapid social and economic changes underlie the virus's spread in most of Africa, Russia, Central Asia, Eastern Europe, China, and elsewhere. HIV epidemics have been particularly sensitive to large migrations of people, wars, economic downturns, and other alterations in social stability. UNAIDSM71HO es- timates that 1.2 percent of all adults worldwide are currendy infected with I-l and all experts agree that the pandemic's worst effects are yet to come. Societies cope with HIV and prevent its spread best where govern- ments are open about the issues, provide information and services, and partner with organizations representing affected communities. HTV/AIDS is not simply a pathogen that can be managed with a typical public health approach, and health services alone cannot tackle the breadth of issues that produce vulnerability. A coordinated, multisectoral response is needed that includes appropriate government departments, nongovernmental organizations and community-based groups, bilateral donors, and U.N. agencies. The collective experience of the past 20 years has shown that the spread of HIV and its effects can be slowed through an approach that re- duces risk, vulnerability, and impact as reflected in the U.N. Global Strategy Framework. The highest levels of political comrmitment are needed to ensure success. The best possible implementation of programs will be achieved with the right mix of policy processes, donor collabora- tion, and governments working together with civil society actors. An Opportunity for Prevention: HIVIAIDS Situation in the MENA/EM Region UNAIDS/WHO estimated that approximately 83,000 people were newly infected with HJV in the MENA/EM region1 in 2002, and about 0.3 percent of the region's adu]ts are currently infected (UNAIDS/WI7O 2002a). Recent evidence suggests that the incidence of sexually transmitted infections, including I-l/AIDS, is increasing, and the total number of AIDS deaths has increased almost sixfold since the early 1990s. In the low- and middle-income countries of the region, HIV/AIDS was the third leading cause of morbidity among people 14 to Executive Summary xv 44 years old in 1998. Levels of HITV infection among tuberculosis (TB) patients are also rising and, by mid-2001, had reached 26 percent in Dji- bouti, 4.2 percent in the Islamic Republic of Iran, and 4.8 percent in Oman (WHO/EMRO 2001a). And an alarmiing rise in HIV continued through 2001-02 among drug injectors in both the Islamic Republic of Iran and Libya. W'hile these regional figures are relatively low compared with Africa, South and Southeast Asia, and the Caribbean regions, low prevalence does not equate to low risks. Inadequate surveillance methods, which is a universal weakness in the region, can overlook outbreaks in marginalized social groups. Furthermore, even in low prevalence nations, the situation can change rapidly, as has occurred in Indonesia and Nepal. Given that HIV/AIDS epidemics tend to exhibit exponential growth, experts expect that the HIV pandemic's worst effects are still to come. Many couniitries in the MENA/EM region have enough evidence of risk factors to warrant imime- diate investments in improved prevention programs. HIV epidemics are also sensitive to changing economic and social factors, and it is noteworthy that current methods of surveillance in the MENA/EM region will fail to detect meaningful changes where they are most likely to occur. Be- cause the cost of the epidemics to society and economic development can be tremendous, good surveillance and effective prevention programs are relative bargains compared with the cost of epidemics. Unfortunately, mostly mandatory screening has produced continued low levels of case detection, appropriate behavioral data are lacking, and the region's governments are overconfident in the protective effects of social and cultural conservatism. These factors combined have dictated a low priority for HIIV/AIDS. Globally, solid evidence has been amassed that justifies a multisectoral investment in intensified prevention while prevalence is low. Waiting until there is an appreciable rise in prevalence is a costly delay that leads to tremendous human, development, and fi- nancial costs. By then, an epidemic may be under way, and it will be too late to prevent the inevitable reductions in human well-being and socie- tal stability, as well as losses in labor productivity, capital investment, and work force availability. A thorough review of available data exhibits numerous gaps and epi- demiological inadequacies. No country, for example, systematically sam- ples and surveys the high-risk groups; instead, the general population, represented by low-risk groups such as antenatal mothers and blood donors, is extensively screened. Although high-prevalence countries can benefit from this type of surveillance, it may fail to record rising rates of HTV among hidden or marginalized groups in a nation thought to have low prevalence. UNAIDS/WIHO recommends that second-generation sur- veillance, consisting of targeted HIV serosurveillance, behavioral surveil- xvi Executive Summary lance, and -sexually transmitted infection (STI) surveillance, can reveal the epidemic as it emerges in the most at-risk groups and identify those who are potentially at risk in the immediate future. Instituting this type of surveillarce requires considerable collaboration with nongovernmen- tal organizAtions (NGOs), community groups, and social scientists to gain access to otherwise marginalized groups. But to gain an essential and realistit vision of their national situations, countries-must overcome obstacles to collaboration with such groups, and they may require ap- propriate technical assistance. Workin-g with the available information, we attempted a rough classifi- cation of epidemic type by nation on the basis of the most recent and least ambiguous-statistics from the end of 1999 to mid-2001. Overall, despite early cases detected among foreigners and returning migrants, IIV has now begun to spread among citizens in all the region's countries. Patterns are shifting, and a rising proportion of cases are resulting from sexual transmission. These cases may well have been sparked, as demonstrated in Asia (Saidel and others 2003), by earlier clusters of cases among injecting drug users (IDUs), and from them to their sexual partners. Increasingly, in- fections are occurring in equal proportions of females to males. Djibouti has one of the highest prevalence levels of IHIV and has the highest level of sexual transmission. In at least one-half of the countries, significant foci have occurred among IDUs in the past and, in some countries have con- tinued. Other infected groups include males who have sex with males (MSM), sex workers and their clients, prisoners (who are frequently drug users), and patients with sexually transmitted diseases (STDs). Social and Structural Vulnerability in the Region Without adequate social and behavioral research, effective HIV preven- tion programs cannot be planned and carried out. Most published HIV- related research on the MENA/EM region concerns clinical and bio- medical issues such as transmission through dialysis. Little substantial HIV-related social or behavioral research has taken place in the region. Designing, implementing, and mronitoring prevention programs without information on the sexual and drug-taking behavior of a population and its subgroups is an exercise in failure. Although the expertise exists in most couiitries of the region to conduct such research, institutional and political slipport for such research is desperately needed. Based on the little research available and a variety of other sources of information, we constructed a typology of risk factors. It should be noted that much of this information is unpublished, unscientific, or an- ecdotal, u-nderscoring the great need for well-conducted research. Sim- Executive Summary xvii ply stated, people are at risk of acquiring an HIrV infection because of what they are doing or what they might do if placed in a facilitating sit- uation. We identified two primary groups: an at-risk group and a vul- nerable group. In the MENA/EM region, as elsewhere, people with known risky behaviors, such as sex workers, their clients, IDUs, MSM, and those who acquire STDs, are immediately at risk. Wbile it is likely that they represent a minority, any such group can form the core of spread into the rest of the population, depending on the extent and na- ture of social linkages and networks. Prevention strategies differ consid- erably for these different groups. The risk factors associated with these subpopulations must be researched and brought to light, if they are to be addressed effectively. The next group comprises those who may be considered vulnerable, that is, they xnay be at risk if and when their life situation changes. This group includes, for example, migrants going to work abroad, refugees, mobile workers such as truckers, tourists traveling for fun and recreation, noninjecting drug users who may switch to injecting when the availabil- ity or price of the drug changes, and young people in general, in that some proportion will engage in nonmarital sex under certain conditions. An integral part of HIV prevention consists of reducing the vulnerabil- ity of people in these social categories. In the MENA/EM region, the HIV-related issues concerning migrants, internally displaced persons (IDP), and refugees are especially significant. Given the large numbers involved in foreign labor migration, AIDS prevention takes on a truly in- ternational perspective and must be approached accordingly. Although a significant number of countries in the region have recorded foci of HIV among IDUs, for example, few countries have made serious attempts to find out how many people are at risk, where they are located, and how to reach them with information and services. In the short run, the perceived negative consequences of public knowl- edge of such activities may be a valid concem, but in the long run, inac- tion can adversely affect HIV prevention programs. Learning how to reach at-risk and vulnerable subgroups in discreet, unpublicized ways would contribute significantly to the national AIDS programs through- out the MENA/EM region. This effort will require a special policy process, political commitment, and the creative collaboration of NGOs, social scientists, social workers, and AIDS program managers. Structural factors, such as poor and dysfunctional health care systems, high rates of unemployment, a lack of access to information and con- doms, and inadequate STD and drug-dependency treatment, also in- crease overall vulnerability. Other structural factors exist as well, includ- ing legal restraints on NGOs in some countries or health policies that disadvantage the young and unmarried, the poor, and noncitizens. There xviii Executive Summary is a general lack of policies and legal frameworks for the protection of people living with HIV, particularly in the workplace. In each situation, relevant policies may require review and modification to alleviate barri- ers to improved prevention and care. The special needs of refugees and others displaced by conflicts must be highlighted in the region. Health authorities cannot ignore the likelihood of spread from migrants, whether legal or illegal, to their local populations. Macroeconomic Impact The human cost of HIV/AIDS is incalculable, from the pain and guilt surrounding personal and intimate relationships, to threatened social and political security at the state level. It is nonetheless revealing to examine the costs and losses that can be calculated, for these too are great and can have a major impact on a nation7s future. Using the most recently published estimates of HIV prevalence, losses in gross domestic product (GDP) and consumption resulting from the diffusion of HIV/AIDS could be significant in many MENA/EM countries. On the basis of data from nine MENA/EM nations (Algeria, Arab Republic of Egypt, Djibouti, Islamic Republic of Iran, Jordan, Lebanon,- Morocco, Republic of Yemen, and Tunisia), calculations for a broad range of diffusion scenarios indicate that the average growth rate of potential GDP could be reduced by 0.2 to 1.5 percent per year for the period 2002-25. The future losses of potential output and consumption during that period could be equivalent to 35 percent of today's GDP, even under conservative assumptions. These losses occur as rising mor- tality and morbidity reduce labor productivity, capital investments are re- duced, and the labor force shrinks. The analysis also reveals that there could be a considerable impact on health expenditures. By 2015, annual expenditures to treat all AIDS pa- tients may have increased by 1.2 percent of GDP, on average-even with only limited use of antiretroviral drugs. It is possible to find cases in which HF//AIDS-related expenditures surpass 5 percent of GDP. Govermnents have a key role in developing and financing the imple- mentation of policies to confront HfV/AIDS. Indeed, individuals alone cannot devise appropriate mechanisms to contain the epidemic. To achieve significant results, governments can only intervene if cost-effective in- terventions are available. Fortunately, international experience shows that low- cost prevention strategies are effective in slowing the spread of HIV/AIDS. In the case of MENA/EM countries, our analysis shows that increasing condom use and expanding access to safe needles for IDUs can generate savings equivalent to 20 percent of today's GDP. WVe also show that de- Executive Summary xix laying the implementation of these policies could give rise to accumu- lated costs for the period 2000-15 that are equivalent to 1.5 percent of today's GDP for each year of delay. The main messages from these analyses are as follows: * The risk of an increase in the HlV/AIDS prevalence level in MENA/EM countries is real. * Expected costs over the next 25 years could be considerable-on the order of 3 5 percent of current GDP even under conservative assump- tions. * Effective actions can be implemented to prevent the spread of the epi- demiic, and the costs of these actions would be more than compen- sated by the savings they generate. and * The time to act is today, when prevalence levels are still low. Responses Timing is critical. While national HIV surveillance concentrates on rel- atively low-risk groups, the virus can reach others. It takes a number of years, particularly in low-prevalence settings, to convince those at risk to alter their behaviors. VVhere skills are scant and NGOs or community- based agencies have little experience in HIV/AIDS prevention, it takes several years to develop these skills. Finances must be mobilized and ef- ficiently directed, an effort that often requires new administrative struc- tures and mechanisms. Popular, political, business, and religious leaders must be educated to help create an enabling enviromnuent in which ef- fective prevention activities can be carried out. Usually, legislative change is required, and legal reform takes time. The MENA/EM region is lag- ging on most fronts in its defense against HIV To date, most decisionimakers in the MENA/EM region have not con- sidered investment in HIV prevention a high priority. Although all cowI1- tries in the region have National AIDS Committees and 14 of 18 coun- tries have U.N. Theme Groups, their functioning varies considerably. Most countries have instituted actions to ensure safe blood supplies, al- though this effort is incomplete in some. Efforts to install universal pre- cautions and safe waste management in health services are also inade- quate in some countries. Medical management and counseling for people with HlV/AIDS have been set up in many countries, and antiretroviral therapy (ART) is available in several nations. Use of ART to reduce mother-to-child transmission (MTCT) is found less frequently. These Executive Summary measures, however, are only the most basic foundation for a successful HIV/AIDs prevention plan. Little has been done on the essential work of reducing stigma and dis- crimination, except among health workers in some countries. Educating the populace about FIIV has been spotty and seldom emphasizes the use of condoms for prevention. Condom promotion is nearly absent in the region. High-risk groups are rarely reached with targeted interventions or with systematic surveillance. Harm reduction for IDUs has not been discussed, except in the Islamic Republic of Iran. In general, the region has been slow to respond to the challenges posed by HIV. IIV epidemics are sensitive to changing economic and social factors and, in the lMENA/EM region, current methods of surveillance are un- able to detect changes where they are most likely to take place. The response from U.N. agencies has, until recently, been mainly from WH1O/EMRO, which has supported numerous meetings, small studies, and basic program reviews since the 1980s. One or two other agencies have been active in supporting I1WV/AI1DS prevention in either selected countries or within specific programinatic areas, such as school education. Nearly all other U.N. agencies concur in recognizing that their involvement has been limited to date, but is now changing. UN- AIDS has begun funding a series of small activities, mainly through WHO/EMRO, most of which have not vet been evaluated. In general, U.N. agencies have only recently become active in the field of HIV/AIDS in the region. Few countries have developed an IITV/AIDS policy or strategic na- tional plan that involves all stakeholders, including civil society and groups affected by HAV Although bilateral donors are present, along with several international and local NGOs, these organizations are inad- equately utilized. Only a few nations are demonstrating a conmnitment to prevention on a scale large enough to make an impact. For example, with funding from the United Nations Population Fund (UNFPA) through a consortium of NGOs, Tunisia has piloted a project for young people that offers appropriate education, counseling and testing, con- doms, and STD services. The project has reached about 10 percent of youth and is being scaled up to reach more. In the Islamic Republic of Iran, a large-scale response to rising rates among IDUs has recently been set in motion, including voluntary counseling and testing (VCT), methadone treatment, needle exchanges, and associated activities such as drug demand-reduction programs. Morocco has instituted a large-scale attempt to upgrade STD services and has developed and approved a comprehensive strategic plan. Djibouti is undertaking a similar process, but with its epidemic at a generalized stage, considerable resources will be required to diminish the disease's impact on many citizens and resi- Executive Summary xxi dents. Both the Republic of Yemen and Sudan recently demonstrated public and political concern about HIV/AIDS, a welcome development. During the first two rounds of grants from the Global Fund to Fight AIDS, Tuberculosis and Malaria (GFATM), the Islamic Republic of Iran, Jordan, and Morocco succeeded in acquiring funding. A Way Forwalrd The actions urgendy needed in each country to respond effectively to HIV/AIDS depend on the stage of the epidemic and other essentially local issues. Recommendations for improving the current situation include: * Raising the priority given to HIV/AIDS through research, media, and advocacy * Evaluating what national HIV/AIDS/STD programs have achieved, including proposing ways to strengthen them * Developing national policy and strategic plans, with associated budg- ets and the identification of potential resources * Instituting second-generation surveillance, including STD and be- havioral surveys * Learning to conduct targeted interventions with at-risk groups in a discreet, culturally appropriate manner, in collaboration with NGOs * Reducing the vulnerability of migrants, IDP, mobile populations, and refugees. This should involve U.N. agencies and appropriate INGOs, and should begin with research to assess situations of risk followed by a process involving all stakeholders to design appropriate and coordi- nated interventions * Improving the provision of clear information and means of protec- tion, and taking small steps toward the development of adequate pro- motion of condoms * Developing life skills and drug demand-reduction education for youth that is culturally appropriate and effective and recognizes the structural factors associated with drug use * Reducing vulnerability among youth through multisectoral planning to affect sexual and reproductive health, unemployment rates, educa- tional costs, and information access, among other critical issues * Ensuring that ART treatment programs are provided and include ad- equate prevention services as well xxii Executive Summary o Empowering affected communities by encouraging local NGO devrel- opment, including increased participation of people living with HIV and AIDS (PLWHA) o Developing regional networks of experts to fulfill technical needs, while developing local capacity o Developing programs that are based on sound knowledge of situa- tional context, include increased participation of a wide range of sec- tors and partners, address vulnerability factors, and have monitoring and evaluation plans and budgets o Promoting future sustainability of low prevalence through insurance and other health-financing schemes that can help ensure equitable ac- cess to health care for people of all socioeconomic strata o Strengthening country-level information systems in order to monitor and evaluate the situation and response to HJV/AIDS o Mobilizing additional human and financial resources in support of na- tional responses through collaborative efforts involving a wider range of international, regional, and national partnerships. These actions will require the collaborative efforts of many partners, including multiple sectors within government, community-based groups, religious and other local leaders, INGOs, bilateral donors, and various U.N. agencies. Most important, the coordination of such a pro- grain requires political commitment at the highest levels. If that com- mitment is attained, the threat that HIV represents to the development goals of the MENA/EM region can be averted. The time to act is today-when the prevalence levels are still low. 1. For the purposes of this book, the MENA/EM region includes Al- geria, Arab Republic of Egypt, Bahrain, Djibouti, Islamic Republic of Iran, Iraq, Jordan, Kuwait, Lebanon, Libya, Morocco, Oman, Qatar, Re- public of Yemen, Saudi Arabia, Syrian Arab Republic, Tunisia, and United Arab Emirates. Information is included on important neighbor- ing countries that are in the operational region of collaborating U.N. agencies, for example EMRO (Sudan and Somalia) and the United Na- tions Children's Fund (UNTCEF) (Sudan only). West Bank-Gaza (Pales- tine) has been omitted because of inadequate comparative information. This book reviews the human immunodeficiency virus/acquired immun- odeficiency syndrome (HTV/AIDS) situation in the Middle East, North Africa, and Eastern Mediterranean (MENA/EM) region. It is intended to stimulate discussion and promote dialogue among policy- and deci- sionmakers in the region. It will also provide a framework for multisec- toral strategic action to reduce risk behaviors that spread HrV, to care for and support those who become infected, and to diminish vulnerability among specific segments of society. Although most evidence suggests that overall HIV prevalence is low in the region, intensifying investment in improved surveillance, prevention, and care 7now can ensure the con- tinuation of low prevalence rates and can have a major impact on na- tional and regional development goals. The production of this volume has been a joint effort of the World Bank, the Eastern Mediterranean Regional Office of the World Health Organization (WHO/EMRO), and the Joint United Nations Pro- gramme on HIV/AIDS (UNAIDS). Interviews were conducted with 51 people-including bilateral donors and those in government, U.N. agen- cies, bilateral donors, and international nongovernmental organizations (INGOs)-in four countries of the region (the Arab Republic of Egypt, Djibouti, Republic of Yemen, and Tunisia) and in Geneva between August 18 and September 20, 2001. Documents were gathered from all pertinent sources, especially WHO, UNAIDS, the U.S. Census Bureau HIV/AIDS Database, and the United Nations Office on Drugs and Crime (UNODC). Because full information for every country in the region was not available from any of these sources, there are inevitable gaps. An up- date with newly available information, as of early 2003, is included. Chapter 1 briefly reviews world patterns and focuses on experience in low-prevalence countries. Chapter 2 examines available epidemiological data in the region, groups countries in probable epidemic stages, and re- ports on the little information known about the genetic variation of the virus in the region. Chapter 3 develops a typology of risk factors using 2 HIVthIDS in the Middle East and North Africa detailed information gathered in four country visits (Arab Republic of Egypt, Djibouti, Republic of Yemen, and Tunisia) as well as from an ex- tensive bibliography covering all countries in the region.1 This chapter examines known groups that are practicing risky behavior as well as vul- nerable populations who are potentially at risk and explores structural factors contributing to vulnerability. Chapter 4 models the macroeco- nonmic impact of a variety of diffusion scenarios based on the spread of HIV to and from unprotected injecting drug users (IDUs) and sex work- ers, which are the best understood examples of high-risk groups and are pertinent to specific countries in the region. Chapter 5 reviews the re- sponses to date of health systems, other government sectors, the U.N. and bilateral donors, and other actors in the region. Chapter 6 recom- mends next steps to improve the current status of programs in the re- gion. The volume concludes with a wide-ranging reference list that was used as background material. Although we extensively reviewed material to prepare this book, it is likely that we were not able to access all exist- ing information at the national level because of time constraints, a lim- ited number of country visits, and, in some cases, difficulty in accessing certain reports. 1. For the purposes of this book, the MENA/EM region includes Al- geria, Bahrain, Djibouti, Egypt, Islamic Republic of Iran, Iraq, Jordan, Kuwait, Lebanon, Libya, Morocco, Oman, Qatar, Republic of Yemen, Saudi Arabia, Syrian Arab Republic, Tunisia, and United Arab Emirates. Information is included on important neighboring countries that are in the operational region of collaborating U.N. agencies, for example EMRO (Sudan and Somalia) and the United Nations Children's Fund (UNICEF) (Sudan only). West Bank-Gaza (Palestine) has been omitted because of inadequate comparative information. CHAPTER 1 Moth NNVAMSU In 1991, with the evidence available at the time, experts estimated that by the year 2000, 15 million to 20 million adults and 5 million to 10 mil- lion children cumulatively would become infected with HIV (UN- AIDS/WVHO 2001). By the end of 2002, UNAIDS/VHO reported that 42 million people were living with HIV/AIDS and 5 million new infec- tions had occurred during the year; nearly half of those infections were in females. The worldwide adult prevalence rate had reached 1.2 per- cent. The great difference between earlier projections and current esti- mates reflects both the unexpected spread of the virus and the inade- quacy of statistics used to track the epidemic. HIV does not always spread rapidly, though it certainly can if conditions permit. Rapid social and economic changes underlie the virus's spread in most of Africa, Rus- sia, Central Asia, Eastern Europe, China, and elsewhere. HIV epidemics have been particularly sensitive to large migrations of people, wars, eco- nomic downturns, and other alterations in social stability. Table 1.1 re- flects the global situation at the end of 2002. In 1997, the World Bank developed a classification system for differ- ent types of HTV epidemic situations that was widely adopted (World Bank 1999a). The term nascent was used as a label for epidemics in which less than 5 percent of people in presumed high-risk groups were in- fected.1 This classification system was altered in 1999 by UNAIDS/WHO, which switched to using the term low prevalence instead of nascent to avoid the connotation that the low levels of HTV prevalence always represented the beginning of a greater epidemic (UNAIDS/WHO 2000a). However, the term low prevalence has too often been equated with low priority for national AIDS programs, lending credibility to unsup- ported notions that little risk or vulnerability exists. UNAIDS/WHO has subsequently promoted second-generation surveillance, a combination of I-UV and sexually transmitted disease (STD) serosurveillance with behav- ioral surveillance of the most-at-risk groups, to provide nations with ac- curate, comprehensive, and forward-looking information about levels of infection as well as levels of risk and vulnerability. 4 HIWAIDS in the Middle East and North Africa Th A3 r1. 5 GDobni Regional Varaton on igV Plrevalence, [ecemnbtr 2002 HOl/AIDS prevalence Adults and chi0rern HIV-poslTve adults among adults nrwly infecte, - who are women Reg on (percent) with liIV (pefcent) Sub-Saharan Africa 8.8 3.5 million 58 North Africa and Middle East 0.3 83,000 55 South and Southeast Asia 0.6 7C0,300 36 East Asia and Pacific 0.1 270,000 24 Latin America 0.6 150,000 30 Caribbean 2.4 60,030 50 Eastern Europe and Central Asia 0.6 250,000 27 Western Europe 0.3 30,000 25 North America 0.6 45,020 20 Australia and New Zealand 0.1 500 7 Total 1.2 (average) 5 million 50 (average) Source: UNAIDS/WHO 2002a. Currently, most of the world's countries with generalized epidemics have functional and appropriate surveillance systems of antenatal moth- ers or other sexually active members of the general population. But countries with low and concentrated epidemics often do not have the surveillance systems needed to track their HiV epidemics properly. This problem clearly exists in the MENAJEM region. No country in the re- gion has set up the type of systematic surveillance of high-risk groups, including their STD and behavioral indicators, required to yield dear in- formation on major risk groups or the effectiveness of AIDS prevention programs (Walker and others 2001). This lack of clear information has led to a lack of visibility, the continued perception of low risk and low priority accompanied by low levels of investment, and, therefore, inade- quate levels of protection. Hidden epidemics in marginalized social groups pose a threat of an expanded epidemic even when repeated testing indicates low prevalence in the general population (Family Health International 200 1b). This has happened elsewhere and, based on the scant information available on risk factors, is likely to occur at some time in the AiMENA/EM region. Figure 1.1 illustrates clearly the progression of the epidemic in Nepal, a pro- gression that took place even as the National AIDS Program continued to monitor the general population. Sentinel surveillance, undertaken semiannually in STD and antenatal clinics throughout the country, showed levels of less than 3 percent and less than 0.1 percent, respec- tively, as of 1999, but the epidemic continued to grow among groups with whom the surveillance system had Little contact. These data illus- Global Experience with HIV/AIDS 5 FIGURE 1.1 The Hidden Epidemic in Nepal: HIV Prevalence among Injecting Drug Users and Sex Workers in Kathmandu HIV+ (percent) 80 70 rIDU 60 - SW 50 SW/IDU 40- 30- 20- 10 - 1992 1993 1996 1997 1998 1999 Year Note: IDU, injecting drug users; SW, sex workers; HIV+ = HIV positive. Source: Courtesy of J. Ross, Family Health International, Nepal, and United States Agency for International Development (USAID); presented at ICAAP, Melbourne, Australia, Oct. 13, 2001; U.S. Census Bureau 2001. trate further the intensity of HIIV prevalence among unrecognized sub- groups that cross risk zones, for example, the prevalence levels among people who are both IDUs and sex workers is, in this instance, as high as 75 percent. In Asia, both Nepal and Indonesia have recently experienced expand- ing epidemics, especially among IDUs (Family Health International 2001a). Advice from experts has for years led these and other similar na- tions into believing that their epidemics would be self-contained (The Econonist 2001). Behavioral surveillance, however, has indicated other- wise, showing that IDUs frequently have sexual partners among non- IDUs, including their own wives as well as others. During the past few years, HIV levels have begun to reflect what behavioral studies had re- vealed earlier. A similar scenario of spread can plausibly be predicted for several countries in the MENA/EM region as well. Significant lessons from elsewhere cannot be ignored (see boxes 1.l and 1.2). Where HIV has been allowed to spread, the impact on affected coun- tries has been devastating. In the worst scenarios, teachers die faster than they can be trained, health workers are so often infected that services are crippled or canceled, and farmers can no longer maintain normal agri- cultural production. The military is almost always affected-often, as in many African countries, to the point that defense functions are seriously hampered. More commonly, military actions taken because of political unrest contribute to the spread of the epidemic. Health and development 6 HIt//AIDS in the Middle East and North Africa BOX 1.1 Nepal: Too Little, Too Late In Nepal, as in almost all other countries in Asia, the first cases of HrV/AIDS were de- tected during the late 1980s or early l990s in either a foreign visitor or a returning Nepalese migrant. During the early 1990s, I-UV seroprevalence surveys detected EIIV infections among STD patients and female sex workers throughout most regions in Nepal. Because of the well-known traffic of both young girls as sex workers and young men as migrant workers between Nepal and India, public health concern focused on the spread of HIV to and from these groups. IDUs in Nepal were believed to share injection equipment in relatively small and iso- lated networks and therefore were thought not to pose a threat to the larger society. Sur- veys undertaken among IDUs in Kathmandu between 1991 and 1994 showed nearly zero prevalence and declining needle-sharing behavior among the men who were in con- tact with a small pilot needle exchange and education program. But an explosive increase of HIV infections in about one-half of all IDUs occurred during the late 1990s nation- ally. No other surveillance or study among IDUs took place until 1999, when a national study showed that 40 percent of IDUs nationally were infected with HfV, and more than 50 percent of IDUs in Kathmandu were infected. A smaller study in Kathmandu seemed to show that those who had been associated with the needle exchange program were less likely to be infected. However, it was clear that the project, while starting early, had not reached a large enough proportion of all IDUs with enough needles to diminish the emerging epidemic. Program expansion was difficult because no policy was in place to support it, and many authorities in the drug control arm of government did not support the harm-reduction approach. It is now too late to avert an IDU epidemic in Nepal. The estimated HIV prevalence in Nepal in 2000 was greater than 50,000 people or close to 0.5 percent of the total 15- to 49-year-old population. A growing proportion of this increased HIV prevalence is the result of in- jecting drug use, sex work, or a combination of the two. Source: Famnily Health International 2001a, 2001b; Karki 2000; Oelricbs and others 2000; Peak and oth- ers 1995. gains reflected in increased life expectancy in developing countries are lost, millions of children are orphaned, families' savings and property are lost, and, because of uncontrolled discrimination and stigma, the usual safety nets based on kinship and community support are wvithdrawvn. While both the rich and the poor can acquire HIV, the rich cope more easily. Likewise, both the rich and poor can be vulnerable to HIY, but vulnerability is far greater where people are illiterate, unable to access in- formation and basic services, trapped in unstable survival circumstances, Global Experience with HIV/AIDS 7 BOX 1.2 Overlooking the Epidemic: The Case of Indonesia Indonesia is an island republic off the coast of mainland Southeast Asia. As the fourth most populous country in the world with an estimated total population of about 209 mil- lion, of which more than 80 percent are Muslim, it is the largest Islamic population in the world. Over the years, Indonesia appeared to have low prevalence, despite projec- tions made in 1994 that it would have 400,000 HTV infections by 1996. "Registered" brothels have existed for more than a century in Indonesia, and sex workers were tested. Even these women showed infection levels of less than 5 percent through 1998. In 1999 and again in 2000, several sentinel sites for female sex workers began to record increas- ing numbers of HIV infection, with rates fi-om 1.5 percent to 8 percent, and a sample of prisoners reached as high as 17.5 percent. Despite outreach programs, condom social marketing, and mass media campaigns, behavioral surveillance conducted annually be- tween 1996 and 1999 showed that condom use between sex workers and clients did not rise until HIV levels were also rising in 1999. IDUs were never included in routine serosurveillance in Indonesia. In fact, informally, many people thought there were too few IDUs to bear any significance for an HIV epi- demic, and most attention was focused on sex workers. In late 2000, several ad hoc sur- veys of IDUs throughout Indonesia detected sharply increasing HTV prevalence, up to more than 50 percent in Jakarta. This increasing trend of HlV prevalence can be seen in the blood donor data from the Indonesian Red Cross from 1992 through 2000 in figure 1.2. In recent years, approximately 750,000 to nearly 1 million blood donors have been screened annually for HTV A marked increase was seen in 1999-2000 and may well re- flect the increase among IDUs noted during this same period. Based on the most recent HIV prevalence findings among different risk groups and on estimates of the size of these groups, a national consensus workshop in Jakarta in March 2001 estimated that there may have been from 80,000 to 120,000 HTV infections in Indonesia in 2000. Indonesia is now classified as a country with a concentrated HIV epidemic primarily among its IDUs, but experience elsewhere indicates that IDUs are not isolated and that the epidemic can easily spread out:ward to their sexual partners, their children, and others. Source: Center for Health Research, University of Indonesia 2000; Famiiily Health International 2001a; U.S. Census Bureau 2001. or powerless to alter their situations. Where women have little economic power, they are usually unable to negotiate safety in their sexual rela- tions. Societies cope with HNV and prevent its spread best where gov- ernments are open about the issues, provide information and services, and partner with organizations representing affected communities. 8 HIV/AIDS in the Middle East and North Africa I jJ - 1 .2 GWV Prevalence in BMood Dlonors in Indonesla Percent HIV positive 0.01 0.008 ' 0.006 0.004 0.002 - 0 92/93 93/94 94/95 95/96 96/97 97/98 98/99 99/2000 Year of blood donation Source: Center for Health Research, University of Indonesia 2000. HIV/AIDS is not simply a pathogen that can be managed with a typical public health approach, and health services alone cannot tackle the breadth of issues producing vulnerability. The collective experience of the past 20 years has shown that it is fea- sible to reduce the spread of HIV and its effects through an approach that reduces risk, vulnerability, and impact as reflected in the U.N. Global Strategy Framework (UNAIDS 2001). The highest political commitment is needed to achieve success, with the right mix of policy processes, donor collaboration, and governments working together with civil society to ensure the best possible implementation of programs. Care and support to those already infected and affected are best inte- grated with prevention programs from the beginning. HIV-infected peo- ple and affected communities are powerful advocates and actors when supported positively. No one agency, government or nongovernment, U.N. or bilateral donor agency, can handle the breadth of response nec- essary. Effective national responses to HrV are consistently hampered by attitudinal barriers to interagency collaboration; bureaucratic inefficien- cies; a paucity of skills; fear of social or political repercussions; poor man- agement, including the mishandling of funds; and discriminatory hiring practices. Highly motivated leadership can reduce these factors and dis- cover creative ways to handle the sensitivities involved. Youth, migrant workers, refugees, traveling businessmen, drug users, and sex workers each have different life conditions. To design HIIIV pre- vention programs that are meaningful to people in all these situations, sound social, behavioral, and epidemiological research is essential. In the Global Experience with HIV/AIDS 9 MENA/EM region, such research is not only unavailable and frequently considered to be impossible to carry out but also, even when it has been conducted, is often not shared because of intense sensitivities about re- vealing unacceptable behaviors. The sharing of results of sociobehavioral studies can have a strong and salutary effect on the evolution of I-HV/AIDS prevention programs and should be encouraged, at least to a small group of legitimate prevention partners. The Bangladesh experience is illustra- tive of the effective use of such studies for advocacy (see box 1.3). BOX 1.3 HIV Behavioral Research in a Conservative Society: Experience in Bangladesh When, in 1995, the Ford Foundation gave a grant to the International Centre for Diar- rhoeal Disease Research, Bangladesh (ICDDR, B) to develop capacity in reproductive health social science research, sexuality research was not even mentioned. Most people at the time did not think such research was possible in Bangladesh, and the research re- view committee refused to allow any interviews with unmarried people. The program began slowly, and invested a long period in training selected young researchers. Small studies were conducted in a rural area in which ICDDR, B personnel were well-known and appreciated for their programs in diarrhoeal diseases and family planning. These studies examined the terminology and concepts associated with reproductive tract symp- toms among men and women. The researchers accumulated terminology used by ordi- nary people for sexual organs, their functions, and symptoms of disease. A small study was attempted of clients of brothel sex workers, but it was found that men in rural villages who knew the research team well were more willing to discuss their sex lives. Another small, ethnographic study was conducted of IDUs in a Dhaka slum. In 1997, in cooperation with the Red Crescent Society, an attempt was made to study the sexual behavior of urban young people but, while young men appeared to speak freely, young women were extremely frightened of such discussions. WArith funding from Fam- ily Health International, a large-scale study was conducted of at-risk groups in Chit- tagong, the nation's largest port. Sailors, fishermen, dockworkers, long-distance truck drivers, as well as both male and female sex workers were interviewed on audiotape, pri- vately and anonymously. This time, more than 600 interviews were acquired, and a deeper understanding was attained of the range of sexual behaviors among these people, the contexts in which these encounters took place, as well as a glimpse into how a hid- den sex trade operated in a conservative city. This study eventually led to a large, well- funded intervention on a national scale. Throughout this period, a quietly conducted, in- depth study took place among men who have sex with men (MSM) in Dhaka, using (Box continues on the following page.) 10 HIV/AIDS in the Middle East and North Africa BOX 1.3 (CONTINUED) trained men from, this community as initerviewers. Fventually, even the clients of street sex workers were interviewed in unmarked cars parked on the streets at night. With this experience, the Social and Behavioral Sciences Program of ICDDR, B agreed to conduct the first national HI-V behavioral surveillance surveys (BSS) for the government, in collaboration with laboratory scientists who conducted the serosurveil- lance. From 1998 to 1999, 3,363 people (IDUs, street and brotlhel female sex workers (FSWs), MSM, and truckers) were interviewed with a structured questionnaire, but the sampling was not random (except at the brothels) because it was believed to be too dif- ficult to map such groups first in order to acquire a real probability sampling frame. In the following year, with the help of guides from each of the at-risk communities, map- ping was accomplished in three cities and true probability samples were attained of 4,449 people (street FSWs, male sex workers, transgender sex workers, brothel sex workers, IDUs, and rickshaw pullers). During these studies, a senior expatriate social scientist served as the Principal Investigator and intensively trained young Bangladeshi re- searchers. By the third round of surveillance, the Bangladeshi researchers handled an even larger set of probability samples, and full second-generation surveillance was estab- lished in the countrv. The findings of these studies were disseminated to all agencies in- terested in HTV/AIDS and were instrumental in helping the government decide to make proactive investments in prevention while the nation was still experiencing low preva- lence. source:Jenkins, Ahmed, and others 2001;Jenkins, Rahman, and others 2001; National AIDS/STD Pro- gram, Bangladesh 2001. Timing is critical. While national I-UV surveillance concentrates on relatively low-risk groups, the virus can reach others. It takes a number of years, particularly in low-prevalence settings, to convince those at risk to alter their behaviors. Where skills are scant and nongovernmental or- ganizations (NGOs) or community-based agencies have little experience in HIV/AIDS prevention, it takes several years to develop these skills. Finances must be mobilized and efficiently directed, an effort that often requires new administrative structures and mechanisms. Popular, politi- cal, business, and religious leaders must be educated to help create an en- abling environment in which effective prevention activities can be car- ried out. Usually, legislative change is required, and legal reform takes time. The MENA/EM region cannot afford to delay any longer in its de- fense against HIV Global Experience with HIV/AIDS 11 Noe 1. Concentrated = 5 percent or more among high-risk groups; gener- alized = 5 percent or more within general population, represented by women visiting antenatal clinics, with much higher levels in high-risk groups (World Bank 1999a). UNAIDS/WHO changed the cutoff for generalized epidemics to 1 percent or more in the general population in 2000 (UNAIDS/WIHO 2000a). CHAPTER 2 Revoew of the NNSAWS$ $0Ruatfi' Mn the MENA/M Regofl0 UNAIDS/WHO estimated that approximately 83,000 people were newly infected with HIV in the MENA/EM region in 2002 and that about 0.3 percent of adults in the region are currently infected. Recent evidence suggests that the incidence of sexually transmitted infections (STIs), including HIV/AIDS, is increasing, and that the total number of AIDS deaths has increased almost sixfold since the early 1990s (UN- AIDS/MENA 2000; UNAIDS/WAfHO 2001, 2002a). In the low- and middle-income countries of the region, HIV/AIDS was the third leading cause of morbidity among those 14 to 44 years old in 1998. Levels of HIV infection among tuberculosis (TB) patients are also rising and, by mid-2001, had reached 26 percent in Djibouti (UNAIDS/WHO 2002d), 4.2 percent in the Islamic Republic of Iran (UNAIDS/WIHO 2002g), and 4.8 percent in Oman (UNAIDS/WHO 2002m). While there may be a reluctance to study the often illegal and socially unacceptable behaviors associated with a high risk of acquiring an HIIV infection, some national AIDS programs have made efforts to test high- risk groups for HIV In most cases, these were institutionalized popula- tions (in prisons, detention, or drug treatment centers) and are not likely to be representative of these groups. Because such surveillance has not been conducted systematically, trends cannot be examined, and a hidden epidemic among these groups could be spreading without being detected. Figure 2.1 compares the rates of AIDS cases per 100,000 people for many countries in the region with a few countries in East Africa. Table 2.1 provides the most recently available official statistics on AIDS as of the end of 2001 in the region, gathered from sources in WHO (EMRO and Geneva), as well as at the country program level. Based on reported AIDS cases to the end of 2001, these figures represent registered cases throughout the region and suggest the levels of transmission taking place 5 to 10 years ago. Such case detection is usually more accurate in coun- tries that have well-functioning health care systems with high coverage of their populations. Reporting of cases is subject to numerous con- straints, including the effects of stigma, fear, and ignorance on the part 13 14 HIV/AIDS in the Middle East and North Africa FIGURE 2.1 Total Reported AIDS Cases per 100,000 People in Selected Countries in the MENA/EM Region, with East African Comparisons AIDS cases/100,000 '350 339 300 - 275 250 200 - 156 150 - 100 - 50- 0.5 0.5 0.5 1.2 1.9 3.2 3.3 12l6 12. 15. 1 0 - I I - Countries Source: UNAIDS/WHO 2000b, 2002c, 2002d, 2002e, 2002f, 2002g, 2002j, 20021, 2002m, 2002n, 2002o, 2002p, 2002q, 2002r. of health professionals regarding diagnosis and the lack of reporting from the private sector. Sometimes expatriates are specifically included (as in Kuwait), and sometimes they are not; in most cases, available data do not clarify this issue. Nonetheless, where other data are scarce, these case reports present a relatively cohesive overview of the region. Djibouti stands out clearly as having a generalized advanced epidemic among other countries in which prevalence remains low or emerging epidemics are concentrated in specific groups. The next most affected countries in- clude Algeria, the Islamic Republic of Iran, Libya, Sudan, and several Gulf states, including Kuwait, Oman, and Qatar. The relatively low lev- els in all nations but Djibouti must not be considered stable and pre- dictable. History has repeatedly shown that the HIV epidemic shifts in accordance with numerous factors, and only proper serological and be- havioral surveillance can detect these changes. AIDS case reporting is ca- pable of detecting past patterns but cannot predict current or future pat- terns of transmission. TABLE 2.1 Reported AIDS Cases in the MENA/EM Region, 1987-2001 1987 Rate and Population, per Country before 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 Total '000s 100,000 Algeriaa 5 16 17 27 35 34 31 53 32 48 39 49 24 58 17 501 30,841 1.6 Bahrain 1 - 1 2 - 6 3 6 10 11 14 11 9 8 7 89 652 13.6 Djibouti - 1 6 51 107 144 144 196 231 358 434 111 267 131 - 2,181 644 339.0 Egypt, Arab Rep. of 5 6 9 7 12 23 29 22 16 14 25 33 34 44 42 321 69,080 0.5 Iran, Islamic Rep. of 1 3 5 10 25 16 32 19 16 27 40 21 27 68 76 386 71,369 0.5 Iraq _ - - - 7 6 21 37 16 15 2 4 6 6 4 124 23,584 0.5 Jordan 5 1 6 1 8 7 8 6 2 4 12 1 1 3 14 10 98 5,051 1.9 Kuwait 1 - 1 1 3 2 2 5 4 5 2 19 4 12 10 71 1,971 3.6 Lebanonb 4 3 5 10 13 7 22 12 18 5 8 37 32 21 21 218 3,556 6.1 Libya - 24 5 11 6 9 2 11 16 21 38 396 72 - - 611 5,471 11.2 Morocco 10 14 20 26 28 30 44 77 57 66 92 93 165 118 129 969 30,430 3.2 Oman 17 26 26 22 25 32 37 51 28 24 36 24 28 30 35 441 2,622 16.8 Qatarc 31 12 9 10 13 5 8 8 6 2 4 3 9 3 2 125 575 21.7 Saudi Arabia 16 6 7 5 10 6 12 38 37 100 112 39 24 24 29 465 21,028 2.2 Somalia 1 4 3 5 - - - - - - - - - 73 - 86 5,892 1.5 Sudan 4 64 122 130 188 184 198 201 250 221 270 511 517 652 354 3,866 31,809 12.2 Syrian Arab Rep. 2 2 8 1 7 3 3 4 6 9 8 8 7 7 14 89 16,610 0.5 Tunisia 20 23 19 36 36 38 52 50 65 54 62 44 42 44 48 633 9,562 6.6 United Arab Emirates 0 0 0 8 1 3 1 2 1 2 1 1 2 - - 22 2,398 0.9 Yemen,Rep.of 0 0 0 1 0 3 4 3 11 60 40 34 34 18 31 239 19,114 1.3 -No data. a. Algeria is included in the World Bank MENA region but not in the WHO/EMRO region. b. Lebanon's total includes four cases with unknown date of reporting. c. Qatar's total includes two cases with unknown date of reporting. Source: Al-Jowder 2001; As'ad and Al-Azzeh 2001; Ba-Omer 2001; Ben Said 2001 b; Egyptian National AIDS Program 2001; El Nakib 2001; WHO/EMRO, personal communication 2003; WHO/EMRO 2001a; Islamic Republic of Iran 2001; Kuwait National AIDS Program 2001; Ministry of Health and Population, Egypt 2001; Ministry of Public Health,Tunisia 2000, 2001; Morocco National AIDS Program 2001; National AIDS Pro- gram, Iran 2001; National AIDS Program, Sudan 2001; Sow 2001; Syrian Arab Republic 2001 b; United Arab Emirates 2001; Ba-Omer 2002. 16 HIV/AIDS in the Middle East and North Africa In almost all countries, the first case of AIDS was discovered in the middle to late 1980s. Although almost every nation formed a National AIDS Committee and a National AIDS Control Program within a year of having detected the First case, the surveillance systems were largely de- signed to protect blood supplies and medical services (Bernvil and others 1991; Constantine and others 1990; WHO/EMPRO 2001c). In several countries (Arab Republic of Egypt, the Islamic Republic of Iran, Libya, Oman, and Saudi Arabia), outbreaks were detected in hospitals among dialysis patients, cancer patients, and children receiving multiple transfu- sions for hemophilia or other inherited blood disorders (Aghanashinikar and others 1992; Al-Mahroos and Ebrahim 1995; Al-Owaish and others 2000; Bakir and others 1995; El-Hazmi and Ramia 1989; El-Sayed and others 2000; Hachicha and others 1995; Hassan and others 1994; Hnmida and others 1995; Leonard and others 1990; Massenet and Bouh 1997; Watts and others 1993; Yerly and others 2001; Zawawi and others 1997). In most countries, but not all, universal precautions and screened blood supplies have greatly diminished these iatrogenic risks (Daar 1991; Kennedy, O'Reilly, and Mah 1998; Shanks and al-Kalai 1995). Moderate to high levels of hepatitis C in several nations, however, suggests that ad- equate prevention against iatrogenic transmission of blood-borne viruses does not exist in all medical facilities (Aghanshinikar and others 1992; Al- Mahroos and Ebrahim 1995; Amini and others 1993; Attia 1998; Bakir and others 1995; Benjelloun and others 1996; Cacoub and others 2000; El-Hamzi and Rarnia 1989; El-Sayed and others 1996, 2000; Hachicha and others 1995; Kassem and others 2000; Mohamed, al Karawi, and Mesa 1997; Nafeh and others 2001; Groterath and Bless 2002). Surveillance for HIV infection, as opposed to AIDS case detection, is seriously deficient in the region, as it is only systematic in low-risk groups representing the general population (blood donors, antenatal mothers, and TB patients1) and only in some countries. This type of sur- veillance is not recommended for low-prevalence countries where it is essential to monitor prevalence among those most likely to become in- fected. Where STD clinics are sentinel sites, at-risk people may be sam- pled, but there is no information on who they are in a socioeconomic sense, what proportion of those in the community with STDs actually come to those clinics, or if these are the most high-risk people. Stan- dardization of reporting across the region has yet to be achieved. Some nations simply do not report at all some years, while overall poor man- agement of data is evident. Nonetheless, examining the available data on HIV prevalence reveals a set of different scenarios. Table 2.2 (at the end of the chapter) summa- rizes the most recent and least ambiguous statistics on HrV available from the end of 1999 to the end of 2002. It lists actual detected H1V/AIDS Review of the HIV/AIDS Situation in the MENA/EM Region 17 cases among adult citizens as well as the estimates made by WHO at the end of 1999. Modes of transmission are nearly always calculated on cu- mulative AIDS cases; if otherwise, this is noted in the table, particularly for recent updates using the last few years only. Where sentinel surveil- lance has been instituted, it takes place at antenatal, TB, and STD clinics. In Tunisia, a small number ( approximately 300) of sex workers are "reg- istered" and, in other nations, the morals police may have access to sex workers, usually by arresting them. No country in the region has access to IDUs outside of treatment centers and the legal and security systems. These systems are often linked. MSM are more likely to be tested through confidential and voluntary services, but this has taken place in only a few countries as small cross-sectional studies (for example, in Egypt and Morocco). Other groups assumed to be at risk who have been tested include soldiers, truck drivers, prisoners, refugees, tea sellers, tourist in- dustry and hotel workers (El-Sayed and others 1996), travelers, nonin- jecting drug users under treatment or arrest, and taxi drivers. Migration figures prominently in the economic life of the region, and legal out-mi- grants are nearly always tested before being permitted to work abroad or they are tested in the host country as in-migrants. Many people are tested when they return to their home countries as well. Almost all of this test- ing is mandatory and without adequate pre- and posttest counseling. In almost all countries, the first years of testing found most positive results among foreigners, returning citizens, or those infected through blood or blood products. In Lebanon, affected migrants have often been businessmen in trade with or residing in West or Central Africa. In Tunisia, a high proportion of infections were found in IDUs returning from France, where, if caught for drug-related offenses, they are sent home. In Egypt, many were workers returning from the Gulf states. Dji- bouti has the distinction of being the main port for Ethiopia and also has a large French and, more recently, a U.S. military presence, making it a magnet for sex workers from neighboring countries. Both Djibouti and the Republic of Yemen are hosts to many thousands of Somalian and Ethiopian refugees as well. These conditions may have helped the re- spective epidemics begin, but cannot be assumed to be the forces that maintain them or that will influence outbreaks in the future. N!eed for Second-GeneratV.ion Surveilance Surveillance in the region has not yet evolved into second-generation surveillance,2 including STI and behavioral surveillance, as recom- mended by WHO. Sampling is not systematic of most high-risk groups, which makes it impossible to examine claims of low prevalence. As seen 18 HIV/AIDS in the Middle East and North Africa clearly in the Djibouti material, and corroborated in many other nations, low or absent levels of infection among antenatal women or blood donors can continue for several years while prevalence rises in high-risk groups such as sex workers or IDUs. In several nations, the movement of people at high risk in and out of the country certainly affects the meas- ureinent of prevalence. For example, in Djibouti, many thousands of Ethiopians returned to their homes after the conflict with Eritrea sub- sided. Refugee camps documented in 1996 to have brothel-like condi- tions for very poor sex workers were nearly empty by 2000 (Tchupo 1998). Under the pressure of law enforcement authorities, shifting pat- terns of sex work have played a role in the HIV epidemics of several countries. For example, in Thailand and Bangladesh, such pressures have caused sex workers to change locations or venues. Other rapidly chang- ing phenomena include return-migration and refugee flows, as in the Republic of Yemen as a result of political and historical events in the re- gion, and patterns of drug use. Evidence from Asia warns of the rapidity with which noninjecting drug use can shift to injecting drug use (Task Force on HIV and Vulnerability 2000). Repeated monitoring of drug- use patterns has become a necessity in much of the world and is sup- ported by TNODC and WVHO using rapid assessment methodology. Behavioral surveiLlance of high-risk groups has not begun in the re- gion, with the exception of Djibouti, where such surveys are now under development. Family Health International will be overseeing these sur- veys as part of an intervention to be carried out by Save the Children (USA) along the road corridor from Mekele in Ethiopia to Djibouti Ville. UNICEF supported an attempt to conduct such studies in Soma- lia, but it failed to obtain useful information on sexual behavior because of inadequate training of research personnel, among other technical problems. Several nations, such as Djibouti, Egypt, Jordan, Lebanon, Oman, Somalia, Syria, and Tunisia, have determined that their youth are at risk and have carried out surveys of young men and sometimes young women. These surveys cover attitudes and knowledge, but sexual behav- ior was queried in only a few cases. Among those surveys that did inquire, levels of premarital sex were considerable among males in some coun- tries and very low among females, suggesting that anecdotal accounts of male youth going to commercial sex workers may be accurate. Without information on the sexual behavior of a population as a whole and its subgroups, including nonnarital sexual activities, it is not possible to de- sign, implement, or monitor effective prevention programs. Conducting both behavioral surveillance and serosurveillance of high-risk groups is dependent on having services to offer, for example, treatment after testing for syphilis. Among hidden populations, develop- ing surveillance means reaching people with services, and, in order to do Review of the HIV/AIDS Situation in the MENA/EM Region 19 this, formative documentation of the nature of the drug and sex trade is essential. In sum, the most important information needed to plan anid design HIV/STD prevention programs in the region is lacking. Nonetheless, given the information available, one can attempt to clas- sify countries having similar epidemiological profiles into groups (see box 2.1). This classification is based on a combination of the quality and quan- tity of information accessible and the prevalence levels revealed. The first level is composed of countries that have gathered some information through extensive or repeated testing, albeit not always of the most im- portant groups, and that consistently reveal low levels of transmission. Egypt exemplifies this group; Jordan, Syria, and possibly Saudi Arabia and Traq also fall into this category. Egypt and Syria do report some informa- tion on high-risk groups, but sampling methods and sizes are not always reported; information on the other countries was limited. The second level consists of those countries with a gradually growing accumulation of in- fections and at least some high-risk groups identified. This second group includes Algeria, Bahrain, the Islamic Republic of Iran, Kuwait, Lebanon, Libya, Morocco, Oman, the Republic of Yemen, Tunisia, and possibly Qatar and the United Arab Emirates (UAE). The third profile is one in which the level of infections appear to be high in the general population. In both Djibouti and Sudan, available data indicate widespread generalized epidemics. A general population survey conducted in Djibouti durinig 2002 confirmed the presence of a generalized epidemic, but at a lower level of prevalence than previously estimated by UNAIDS/WHO. Somalia is likely to fall into this third category also, though adequate data are lacking. BOX 2.1 HIV Epidemic Profiles in the MENA/EM Region Type 1. Repeated testing, consistently low rates, but no consistent, systematic testing (or reporting) of high-risk groups: o Egypt, Jordan, Syria, and possibly Saudi Arabia and Traq. Type 2. Accumulating levels of infection; gradual and slow; some rapid increases in iden- tified high-risk groups: o Algeria, Bahrain, the Islamic Republic of Iran, Kuwait, Lebanon, Libya, Morocco, Oman, the Republic of Yemen, Tunisia, and possibly Qatar and the UAE. Type 3. Generalized epidemic levels of HTV o Djibouti, Sudan, and probably Somalia. 20 hIV/AIDS in the Middle East and North Africa Viral subtypes have not been identified in all countries of the region, or, if they have, the information has not been published in standard med- ical journals. However, in Djibouti, subtypes A2, E, and C have been found, indicating sources from West/Central Africa as well as Asia. For viral genetics, see Abid and others 1998; Carr and others 1998; Con- stantine and others 1990a; Etharti and others 1997; Fox and others 1992; Giuman and others 1990; Laskv and others 1997; Mboudjeka and others 1999; Montavon and others 2000; Voevodin and others 1996. 1. TB patients are often surveyed in low-prevatence countries because they are accessible and because those in charge of TB programs wish to know what proportion of patients is simultaneously infected with HIV. Such patients are to be considered at low risk for HIlV however, as the behaviors associated with acquiring IEIV are not factors in becoming in- fected with TB. 2. Earlier surveillance recommendations were for serology only and did not include behavioral surveillance or STI surveillance. The emerg- ing epidemic scenarios have shown that resources are wasted if they are not targeted at the groups with the most risky behaviors. Second-gener- ation surveillance is the term used to designate an upgraded approach, in which serosurveillance is adjusted for the stage of the epidemic and linked to repeated behavioral surveys. TABLE 2.2 Profiles of the HIV/AIDS Epidemics in Countries of the MENA/EM Region Adults and children Female Estimated adult living with HIVWAIDS reported prevalence (reported [r] and AIDS cases Features of epidemic transmission modes Date of first level estimated [el or reported among reported AIDS cases, 997-2001; recorded (percent) by UNAIDS/WHO, HIV infections HIV in high-risk HIV prevalence in general population; Country AIDS case (UNAIDS/WHO) end 2001) (percent) groups (percent) mnajor risks; indicators of change Algeria 1985 0.1 1,067 (1/01) [r]; 27 (AIDS) 1.2 in FSWs (1988).10 in 20 FSWs (2000); 41% heterosexual,5% homosexual, 18% IDU, 10% blood,2% 13,000 (adults) [e] 3 in 139 FSWs (2000);STD clinics at 6 sites, MTCT,<1 other,24% unknown;0.9% HIV prevalence at 4ante- 1.3 at Tamanrasset, 4 at Oran, rest zero (2000) natal sites (2000); 0.3% in 345 TB patients (1998); 0.4 in 1,984 antenatal mothers ('00); high levels of migrants from West and Central Africa transiting through southern area; FSWs from Algetia and elsewhere. Bahrain 1985 0.3 216 (December 7 (AIDS); 11 (HIV, 1.6 in 242 multitransfused children with 11% heterosexual,4% homosexual/bisexual; 72% IDU,2.4% 2000 [r]); July 2001) hemolytic anemias (1995); 0.3 in 291 IDUs MTCT, 10% blood; 67% all cases in IDUs (2001); 0 in 2,079 <1,000 [e] (2000);0.9-2.3 in IDUs (1998). blood donors (1999);0.2 in 627 antenatal women (1998); migrant sex trade, opiates. Djibouti 1986 2.8 3,500-14,500 [e] 21 (AIDS, 1997-98); 22 in STD patients (1996); 28 in FSWs (1998); 99% heterosexual, 1% MTCT (1997-98); 1.9% in private ante- (1999) 54among 15- to >50 in street sex workers and 26 in bar sex natal clinic (1999);26% in TB patients (2001);1.8%-3.1% in 29-year-old cohort. workers (1996). blood donors (2000); STDs: 3.2% syphilis in antenatal mothers (1997); in 2002, general population 2.8% (950h Cl 1.2% to 4.5%, 1999); high levels of commercial sex, refugees. Egypt, 1986 <0.1 1,291 (June 2002) 11; 17.3 (NAP) 1 in 102 MSM (1999); 0.79 in 382 MSM (2000); 45% heterosexual, 21% homosexual, 6% IDU, 16% blood, <1% Arab Rep.of cumulative [r]; 0.86 in 815 MSM (2001);0 in 129 FSW (2001); MTCT, 11% unknown;0.006% in blood donors (2000);0 ante- 8,000 [el 0 in 920 STD patients (2001). natal; 0.6% in TB patients; 75% infections acquired in Egypt. Iran, Islamic 1986 <0.1 20,000 [el 8 0.72 in 140,277 drug users tested over time; 10% sexual; 64% IDU, 6% blood, 1% MTCT, 19% unknown; 4.2% in Rep.of 0.5 in 8,202 IDUs (1998); TB patients (2000);rapid threefold increase in HIV/AIDS in 2001; 2.3% among prisoners, mostly drug users, 4% among VCT center users (2001); sex work, polygamy also risk in 2000;0 in 5,700 STD patients (1998); 0 in factors. 1,605 FSWs (1998). (Table continues on the following page.) Adults and children Female Estimated adult liv;ng with I-V/AIDS eoorted prevalence (reported [r] and AIDS cases Features of epicemic transrniss on modes Date of fir5t level estimated (el or reported among reported AIDS cases,1997-201; re.orded (oe'cent) by UNAIDSNWHO, HIV infections HIV in high-risk hlV ooevannece in g'neral populatior; Country AIDS case (UNA DS/WHO) end 2001) (percent) groups (percent) malor risks; ind:cato,s of crange Iraq 1991 < 0.1 <1,000 (e] 9 Mandatory screening for prisoners, STD patients, 9.39S heterosexual, 86.1S% blood, 4.6% MTCT (1 999); STDs: health and hotel workers; premarital tests, etc. reported cases increased between 1999 and 2000; 0.1 9o syphilis and 0 HIV in pregnariL vvomen (2000). Jordan 1986 <0.1 I18 (June 2001i) 13 (AIDS); No surveilance except prisoners; none infected 40% neterosexual, 3.2',r homosexua, 3.2- ,. DC, 38.9-,i bhood, [r]; <1,000 [e] 26 (NAP) of 945 tested (2000). (1997-2000),1.1°SMTCT,13.7% unknown;only 1/281 IB patients infected;0.0396 among blood donors; none among antenatal mothers tested 1992,1994,1999; KAP study (1999) of 3,2C0 people revealed 495-1695 sex outside of marriage in last year, of which 1G', MShi. Kuwait 1984 0.12 835 (December 18 0 in 2,600 STD patients; mandatory testing /3SS heterosexual, 655 homosexual, 6°S IDUs; 2%1 IVTCT; 6% (includes 20C0) [r); 1,300 [e] for sexual offense prisoners, IDUs in treatmert/ b ood, 8SS unknown; 275,307 people screeried ir 2000. 1.7%S I-V expatriates) custody; 0 in 193 IDUs (2W00). positive; 0 in pregnant women; STDs: increased from 1,002 in 1991 to 6,043 in 1997; 30°S gonorrhea, 1.6% syphilis; HIV types B & C, via India; migrants/sex/heroir. Lebanon 984 0.09 613 (Cecember 6 (AIDS);21 In 1999,Cinfected of 205 select FSWs; 47SS heterosoxiai,28%S nomosexual,3%IDUs, 15.6%i blood, 2000) [r]; 1,500 [el (December 1999 0.2 in prisoners; 6.3 of all reported HIV 6,71 MTCT (cumulative); >50% recent cases local origin; rising HIV/AIDS cases) cases are ICUs, all males. percentage of women; general populat:on behavioral surveys in 1991 and 1996 show drop in ever use of condom from 40% to 33°i. Libya 1990 0.2 611 (1999) [r]; 571 new infections in 2C00,98 among IDUs. 56% heterosexual,22% blood,22SSMTCT (cumulative),but not 1,182 (end 2000) currently accurate; outbreaks in hospitals from lack of infection [r]; 7,000 [el control, 370 children in 1998; 22S in b'ood donors (1 998); 0.3% in 296 TB patients (1998). Morocco 1986 0.1 809 (September 36:50 (among 0.16 in STD patients (2000); no surveillance 70%heterosexual,9% homosexual,6% IDUs,3% blood,2% (includes 2000) [r]; new cases) among FSWs or IDUs or MSM. MTCT, 6% other,4% unknown; <1% in antenatal women (2001); expatriates) 13,000 [el rising rates in some areas,Tangiers higher in IDUs, Marrakech higher in MSM. 94% of all cases among Moroccans. Oman 1987 0.1 600 (February 32 In 1999,5 in 135 arrested IDUs; 8.3 among 60 41% heterosexual, 11% homosexual,2% IDUs,22% blood,6% 2001) (r]; IDUs (2000); 0 HIV in 337 STD patients (2001). MTCT, 2% other, 17% unknown; 2% in TB patients (2000); STDs: 1,300 [el incidence rate of reported cases dropped from 92 to 48.6 per 100,000 between 1996 and 2000; among 245 men in social clubs, 13% nonmarital sex in past year (1995). Qatar _ 0.09 300 [e] 29 (cumulative In 1999,5 in 2,249 STD patients. 20% heterosexual,4.8% homosexual, 58% blood, 8% MTCT, AIDS) 9.6% unknown;0 in 2,464 blood donors (1999). Saudi _- - 34 (AIDS 1997-98); 2.3 in multitransfused children (1989); 72%heterosexual,6% homosexual, 2% IDUs,15% blood,4% Arabia Two-thirds of AIDS 0.14 in 2,102 IDUs (1997). MTCT (1997-98). cases among male expatriates. Somalia 1987 1.0 43,000adults (e] - 2to4 in FSW(1990) Little information,MTCT3%;3%-9%in antenatal mothers (2000); blood donors range from 0.8% to 4.4%; 6.9% in 649 TB patients (1999). Sudan 1985 1.6 450,000 [e) 29.8 (AIDS) 1 in out-migrants, 4.3 among 470 refugees, 97% heterosexual, 3% MTCT (1997-2001),1.4% HIV in blood 4.4 among sex workers, 2.5 among tea sellers, donors (2000); 1%-5% antenatal clinics, 7.7%-20% TB patients; 1.6 among TB patients (2002). clear rise starting late 1980s, generalized by mid-I 990s; STDs: 2% syphilis in antenatal mothers, 9% in STD patients (2001); national survey n = 7,385,1.6% (2002). Syrian 1987 0.01 145 (July 1999) [r]; 21 (July 1999, NAP) As of 1998, STD patients 0.12; FSWs 0.12; 73% heterosexual, 8% homosexual, 8% IDUs, 8% blood, 4% MTCT Arab 800 [el bar girls 0.04; MSM 0.59; 0 in IDUs. (1997-2000); 250,000 people tested yearly, almost all mandatory; Rep. 0.0015% in blood donors, 0.005% in Syrian travelers before departure; few cases detected; STDs: April to June 1999,2,342 cases reported at four centers according to syndromic method. (Table continues on the following page.) TABLE 2.2 (continued) Adults and children Female Estimated adult living with HIV/AIDS reported prevalence (reported [r) and AIDS cases Features of epidemic transmission modes Date of first level estimated [el or reported among reported AIDS cases.1997-2001; recorded (percent) by UNAIDS/WHO, HIV infections HIV in high-risk iIV prevalence in general population; Country AIDS case (UNAIDS/WHO) end 2001) (percent) groups (percent) major risks; indicators of change Tunisia 1985 0.06 (March 985 (December 40 (AIDS, 1998-99) 0 to < in registered FSWs throughout 1990s; 51% heterosexual, 10% homosexual, 27% IDUs,8% blood,4% (includes 2001) 2000) [r]; 0 in 570 FSWs (1999),0.22 in 458 FSWs (2001). MTCT (1998-99); 0 in 108 antenatal mothers (1999); >50% expatriates) 2,200 [el detected as AIDS; all infected females and 30% males acquired infection in Tunisia; high proportion among expatriates; 0.2596 inTB patients (1996);0.003% in blood donors, United 0.18 2,300 (January - - Arab 2000) [el Emirates Yemen, 1990 0.01 960 (December 33 (2000); 5 in 88 FSWs (1998);3 in 585 STD patients 77.3% heterosexual, 15.9% homosexual,6.8% blood (1998);rmore Rep.of 2000) [rI; 9,900 [e] -50 HIV (2000) (1998); 27 in 147 prisoners (1 998);. 2.7 in FSWs than 50% infections acquired in Republic of Yemen; gender (1999); 7 in FSWs (2001); 1.8 in 284 STD ratio changed from 4:1 (male to female) in 1995 to 2:1 in 1999 patients (2000). and 1:1 in 2000; 0.7% in 11,070'1ow- risk'people (1998);0.04% in 19,813 blood donors (1998) rose to 0.28% in 2000; 6.9% in T8 patients (1999);45% HIV infections among Yemenis. -Not available. Note: Proportions represent AIDS cases and assumed modes of transmission according toWHO 2001. Cl,confidence interval; FSW,female sex worker;IDU,injecting drug user KAP, knowledge,attitude, practice; MTCT,mother- to-child transmission; NAP, National AIDS Program; STD, sexually transmitted disease;TB, tuberculosis;VCT, voluntary counseling and testing. Source: Abdelmajid 1999; Al-Jowder 2001; Al-Owaish and others 2000; As'ad and Al-Azzeh 2001; Ba-Omer 2001; Ben Said 2001 a, 2001 b; Bouakaz 1998; Bouakaz 2000; Egyptian National AIDS Program 2001; El Nakib 2001; WHO/EMRO 1995,2000a; 2001a; EI-Sayed 2002; Etchepare 2001, 2002; Family Health International 2001 c; Farza 2001; Hermez 2002; Islamic Republic of Iran 2001; Kuwait National AIDS Program 2001; Ministry of Health, Dji- bouti 2001; Ministry of Health in collaboration with UNAIDS and WHO, Jordan 1999; Ministry of Health and Population, Egypt 2001; Ministry of Public Health,Tunisia 2000, 2001; Morocco National AIDS Program 2001; Mos- bah and Ben Yahi 1998; National AIDS Program, Iran 2001; National AIDS Program, Sudan 2001; Njoh and Zimmo 1997; Preble 1996; Republique Algerienne Democratique et Populaire Ministere de la Sante et de la Popula- tion/Direction de la Prevention/Institut National de Sante Publique 2000; Rosa 1999; Sow 2001; Sudan National AIDS Control Program 2000,2001,2002; Syrian Arab Republic 2001a ;Tchupo 1998; UNAIDS/WHO 2002a, 2002b, 2002c, 2002d, 2002e, 2002g, 2002h, 2002i, 2002j, 2002k, 20021, 2002m, 2002n, 2002o, 2002p, 2002q, 2002r, 2002s, 2002t. CHAPTER 3 The lack of solid epidemiological and sociological understanding of the contexts and determinants of risk behaviors throughout the region seri- ously limits intervention strategies. In the interest of facilitating a dis- cussion concerned with devising new and creative means to implement properly targeted prevention programs, the following section examines the presumed main risk factors in the region and attempts to examine se- lected social, economic, and structural factors that can influence vulner- ability. At-Risk Groups Simply stated, people are at risk of acquiring an HI-IV infection because of what they are doing or what they might do if placed in a facilitating situation. The first group constitutes those who currently have multiple, concurrent (within the past year) sex partners or inject drugs; the second group consists of all those whose lives may place them in situations in which these behaviors are more likely, for example, migration away from home, particularly when under insecure conditions.1 A common situa- tion of vulnerability is when noninjecting drug use shifts to injecting drug use because of changes in availability and price of drugs. Prevention strategies differ considerably for these different groups. In the MENA/EM region, there is evidence of immediate high risk for the fol- lowing groups: IDUs, sex workers, their clients, and MSM. Few countries in the region have made serious attempts to find out how many of these people are at risk, where they are located, and how they may be accessed for education and services. In the short run, po- tential negative consequences of public knowledge of such activities may be a valid concern in many circumstances, but in the long run, lack of knowledge can seriously impair planning and design of HIV prevention programs. Learning how to reach these people in quiet and unpublicized ways would make a great contribution to the National AIDS Programs 25 26 HIV/AIDS in the Middle East and North Africa throughout the MENA/EM region. This effort will require a special policy process, political commitment, and the creative collaboration of NGOs, social scientists and social workers, and H1V/AIDS program managers. gnDecVng Dirug Users The UNODC reports that drug-trafficking routes have been shifting in the eastern Mediterranean and parts of northern Africa in recent years. In addition to the well-known opium and heroin traffic from Afghanistan to the Islamic Republic of Iran, a major transit route through Turkey and neighboring countries provides heroin and stimulants to the markets of the Gulf states. In addition, illicit drugs are smuggled into the Arabian peninsula from southwest Asia, and synthetic drugs are produced in some countries of the region. As elsewhere, changing drug-trafficking routes produce localized changes in patterns of drug consumption along those routes. Social change and the long-term consequences of the drawnm-out conflicts that are characteristic of the region may also have an impact on drug use (Toufic 1996-97). Most countries in the MEiNA re- gion report dara that point to increasing drug use, especially amolng young people, mainly of cannabis, h-eroin, and stimulants, including large-scale use of fenetylline. Strong cultural and social stigma is likely to deter users from admitting to their addictions and seeking help (UNODC 2003). Coupled with inadequate HIV surveillance as well as a real paucity of information on the drug trade and drug-use patterns, it is clearly possible that hidden HIV epidemics among drug users are to be found throughout the region. A significant number of countries in the region has recorded foci of HTV among IDUs. Recent outbreaks have been documented in Bahrain, the Islamic Republic of Iran, and Libya and concentrations of infections have been reported in Algeria, Egypt, Kuwait, Morocco, Oman, and Tunisia. To a lesser extent, Jordan, Lebanon, Qatar, Saudi Arabia, and Syria have also reported such clusters of cases. No information was avail- able on the UAE or the Republic of Yemen, although in the Republic of Yemen there is some acknowledgment of the potential for IDU trans- mission. In 1998, IDUs represented two-thirds of all cases of AIDS re- ported in Bahrain and half of all cases reported in the Islamic Republic of Iran. In the same year, IDUs represented 30 percent of all recorded HIV (not yet AIDS) cases in the Islamic Republic of Iran, rising to about 65 percent by 2001. In one study in 1991, 242 IDUs under treatrnent in Manama, Bahrain, had an EIIV prevalence of 21.1 percent (Al-Haddad and others 1994; WHO/EMRO and UNAIDS 2001), though subse- quent studies revealed much lower levels. Oman reported a level of 5 A Typology of Risk Factors 27 percent HTIV among ID Us tested in custody in 1999 (WHO/EMRO and UNAIDS 2001). Morocco reports 7 percent of all cumulative AIDS cases among IDUs (Morocco National AIDS Program 2001). In these countries, IDUs were tested when in contact with authorities, either after arrest or in treatment facilities. Although about one-third of reported AIDS cases were among IDUs in Ttnisia, it is thought that they all acquired their infections when outside of the country. Because of the strict prosecution of illegal drug users in most countries of the region, access to out-of-treatment IDUs is difficult, and sentinel surveillance of street-level IDUs has not been possible. Hence, the bias in prevalence levels cited is unknown. Most estimates are based on analysis of cumula- tive reported AIDS cases. With the exception of Egypt (United Nations Office of Drug Control and Crime Prevention [UNODCCP] 2001b, 2001c), the Islamic Re- public of Iran (Razzaghi and others 1999), Lebanon (Ingold and the Lebanese Research Cooperative 1994), and Morocco, to our knowledge, there have been no rapid assessments or other studies to establish the profile of IDUs in each country. Even where such assessments have taken place, these can become outdated quickly as drug-use patterns can shift rapidly and remain hidden. The 1994 assessment in Beirut estimated that 10 percent of all users were IDUs and 40 percent of these IDUs shared needles. In Egypt, several studies suggest that injecting poses a possible serious problem, although better sampling is required in order to be cer- tain (see box 3.1). In Jordan, it is said that there has been a recent shift from use of hashish to the use of injectable heroin, cocaine, and opium (UNODCCP 2000c). In Kuwait, according to local press reports, there are more than 29,000 drug addicts in a population of 1.7 million (Reuters 2001a; Whitaker 2000). Although heroin and opium are reportedly in use, no estimates were found of the proportion of users who are inject- ing. Libya is estimated to have about 7,000 drug users, most of whom in- ject heroin. Almost all (564) of the 571 new HIV infections reported in Libya during 2000 were among drug users (UNAIDS/WHO 2002k). In the Islamic Republic of Iran, the situation among IDUs is better un- derstood than elsewhere (Aqaie 2001; Arbesser, Bashiribod, and Sixl 1987; Iran Daily 2001b; Islamic Republic of Iran 2001; Moore 2001; Muir 2002; National AIDS Program, Iran 2001; Razzaghi and others 1999). The first real outbreaks were seen in 1991 in prisoners, most of whom were ar- rested for illegal drug use. Little evidence of HTV was found again until 1996, when 29 percent of injectors in two prisons were found to be in- fected. By 2001, 10 prisons had reported IIIV infection among injectors, with one site as high as 63 percent. IDUs under treatment have lower prevalence levels, which were recorded recently at 12 percent, indicating a high probability of transmission taking place in prisons. 28 HIV/AIDS in the Middle East and North Africa BOX 3.1 Injecting Drug Users in Egypt In Egypt, more information is available on drug-related risk for HIV infection than on sexual behavior-related risk. Since the mid-1990s at least three key studies have sup- ported the need for considerable HIV prevention efforts among drug users. The first study revealed that 6 percent of 16,645 young people in five govemorates had used illicit drugs at some time in their lives. The second study, a situation assessment conducted by the Ministry of Public Health and UNODC in five governorates (n = 697) found that 16.4 percent of a sample of 175 drug users reported injecting. Injecting was most com- mon in Cairo, followed by Sinai. The drugs injected were heroin and psychotropic med- icines, often purchased as tablets and crushed. These IDUs were more often older, sinl- gle, skilled laborers and had higher incomes than other drug users. Sexual orientation was queried among all drug users; 1.1 percent were homosexual and 2.2 percent were bi- sexual. The third, most recent study, of those in drug-treatment centers (n = 152) found that 41 percent were IDUs. The sharing of equipment and unsafe sexual practices were both reported as risks for HIV: These studies point to the great need for reaching drug users outside of treatment through outreach in order to provide suitable programs for treat- ment, education, and HLV testing. Source: El-Sayed 2002; UINODCCP 2001b, 2001c. A rapid assessment by Razzaghi and others (1999) was carried out in the Islamic Republic of Iran in 1998 under the guidance of the UNODCCP (now UNODC). This well-executed study showed that heroin and opium use was widespread, with about 2 percent of the population using these drugs, a level comparable to that in Russia and many times greater than in the United States. An estimated 10 percent to 18 percent were currently injecting, for an estimate of about 200,000 IDUs. More than 90 percent were men, but women were probably underrepresented. Among men, the mean age at onset of illicit drug use was 22 years old + 7 years. About one-half the men were married, and one-third of these reported having had extramarital sex, mostly with sex workers. About 70 percent of the unmarried men were also sexually active, 74 percent with sex workers and 30 percent with other men. The assessment also showed the serious lack of H1V prevention among IDUs and the inadequacies in treatment programs. Since that time, the Islamic Republic of Iran has attempted to expand and improve treatment options, including trials with methadone, and has begun harm-reduction programs that include nee- A Typology of Risk Factors 29 dle exchanges (U.S. Census Bureau 2001; Yeghaneh 2001). Recent stud- ies confirm widespread injecting drug use by prisoners. The Islamic Re- public of Iran's national organizations have actively responded to the well-documented serious risk of HIV transmission in their nation and have begun to examine issues surrounding sex work as well (Dareini 2001; Reuters 2001b). Similarly well-conducted studies have not yet taken place in most other countries. Coupled with the fact that at least nine other countries have recorded a significant number of infections among IDUs, the lack of data on this issue reveals widespread ambivalence about drug use in the region. In 1999, WHO/EMRO organized an intercountry consul- tation on demand reduction in which most countries reported having a mixed policy, treating the drug addict as both a criminal and a patient (WHIO/ EMRO 1999). There are few treatment facilities; they are generally based on detoxification and often do not have programs to prevent relapse upon release. Methadone or other substitution treat- ment is not commonly available. Harm reduction to diminish the spread of HIV or other blood-borne infections in prisons has not been considered except in the Islamic Republic of Iran, but may be needed in several cotntries. Sex Workers Women and sometimes men engage in commercial sex throughout the world, and the MENA/EM region is no exception. Religious principles, social pressures, and, in some instances, punishments are strong and keep the various types of sex trade hidden. Nonetheless, numerous reports point to the high-risk situation surrounding sex workers in many coun- tries in the region. The best documentation regarding sex work in the region comes from Djibouti, where the trade is far less hidden than else- where (Constantine and others 1989, 1992; Couzineau and others 1991; Etchepare 2001; Fox and others 1988, 1989a, 1989b; Philippon and oth- ers 1997; Rodier and others 1993a, 1993b; Tchupo 1998). Previously, women from neighboring countries dominated the trade, but, of late, Djibouti women are increasingly involved. In 1998, a brief assessment of the STI situation in Djibouti described several levels of sex work. Start- ing with relatively highly paid women accessible at five-star hotels serv- ing the wealthy, foreigners and Djibouti nationals alike, the list works its way down to street children sex workers and brothel-like scenes for women in refugee quarters that are reportedly now mostly abandoned. In between, there are bar and restaurant workers serving the military; street workers providing "quick" sex, mainly to Djibouti men at various sites around town; as well as geeza women (Somalian) who serve where 30 HIV/AIDS in the Middle East and North Africa men chew qat (Tchupo 1998). Reaching these many different women would require customized strategies. Collected epidemiological material shows distiietly different rates of HTV among bar girls or hostesses com- pared with street sex workers (see figure 3.1). In Djibouti, bar girls have been checked and offered STI services by French army authorities as a matter of prophylaxis for manv years. Algeria, Lebanon, Morocco, Tunisia, and Syria also have data on HIV among sex workers. Tunisia has a tradition of registering sex workers, but these number only about 300, and many more are thought to be in the clandestine trade. In Sana'a, Republic of Yemen, sex workers frequent known streets, working in a system in which rooms can be rented in apartments of women who participate in the network. It is thought that- many of these women are from neighboring African countries. One re- cent report states that HIV prevalence among arrested sex workers in the Republic of Yemen reached 7 percent (Al-Qadhi 2001). Other countries have locales, some more hidden than others, to which it is reported that women from a variety of countries, particularly South and Southeast Asia, are recruited (Blanchet 2002). Egypt has its "red flats," rental spaces kept especially for this activity (El-Gawhary 1995). Street-based and other forms of sex work have been documented in writings about Cairo over recent years. Higher-paid "call girls" operate in better hotels in many countries. Because of the great need for keeping the business hid- den, cell-phone-carrying brokers play a major role, and most appear to be female. Sex work is increasingly being openly recognized as a social issue in the Islamic Republic of Iran and Lebanon. Concern has been ex- pressed in the Islamic Republic of Iran about the potential for temporary marriage, known as sigheh or mnuta,, to contribute to the spread of HIV (Haeri n.d.; Heise, Moore, and Toubia 1996). Bahrain and the Islamic Republic of Iran have also attempted to tackle the trafficking of women as sex workers (Reuters 2001a, 2001b). Hardly any in-depth studies have been made on commercial sex in the MENA/EM region, and some of the information reported in this book is based on unconfirmed discussions with local people. The main clients of sex workers have not been identified. Clearly, however, adequate data, such as increasing rates among STD patients, exist to indicate the po- tential for spread of HYV, first to sex workers and, if the virus accumu- lates among them, to their clients and on to clients' wives and children. The fact that sporadic testing of a few hundred sex workers who are in custody has not revealed high levels of HWr in most countries should be no excuse for complacency and neglect. As long as these people are not reached with prevention education and specially managed STD services, the national risk will grow. It is important that these aspects be con- fronted by all concerned parties in a special policy process that seeks a A Typology of Risk Factors 31 FIGURE 3.1 Evolution of the Seroprevalence of HIV among Bar Hostesses and Sex Workers between 1987 and 1996 in Djibouti HIV seroprevalence (percent) 60 55.8 51.3 >50 50 41.4 395 40 30 25.6 21.7 20 14.2 109 0 3.49 2 5 < 2.7 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 -_-Street sex workers --Bar hostesses Source: Etchepare 2001. culturally tolerable level of outreach and service provision for the sake of the public's health. Men Who Have Sex with Men As elsewhere, the least acceptable sexual behavior in the region is homo- sexuality. Strictly forbidden by religious teachings, MSM hide their pref- erences from family and friends, but homosexual transmission of HIV has been recorded in Algeria, Egypt, Jordan, Kuwait, Lebanon, Mo- rocco, Oman, Qatar, the Republic of Yemen, Syria, and Trunisia, and heightened risk practices are thought to be found in this group (Khan 1995, 1996; Schmitt and Sofer 1992). In Egypt, one of the few countries in which samples of MSM have been repeatedly tested for HIV, preva- lence has hovered around 1 percent for the past few years, while MSM represent 32 percent of recently (1997-2001) recorded AIDS cases (USAID/Egypt 2001b) (see box 3.2). While the National AIDS Program recognizes the need for active prevention, the recent publicized arrest of young men who were accused of being involved in homosexual relations has made its efforts more difficult (El Deeb 2001; Fam 2001). 32 HIV/AIDS in the Middle East and North Africa BOX 3.2 Men Who Have Sex with Men, in Egypt The National AIDS Program (NAP) conducted one small study of 58 MSM in early 1994. This Cairo study was achieved through snowball sampling, producing a strong participation bias toward the well-educated middle class, with 96.5 percent of the men literate, 26 percent university graduates, and 37 percent of professional occupational groups. In this sample, more than 30 percent were married anad 44 percent were bisex-u- ally active; 20 percent were male sex workers and 67 percent admitted to having more than five male partners concurrently, though only 21 percent ever used condoms, with fewer than 2 percent using them consistently. No organized programs for MSM exist, al- though socially interconnected groups have been in touch with the NAP. The study's small size and method of sampling indicates a lack of representativeness and points up the importance of further well-conducted social and behavioral research. Source: El-Sayed and others 1994. It is maintained, however, that discreet activities can be carried out with MSM for the purposes of HIV prevention. Projects that reach MSM in culturally similar nations do exist, for example, in Morocco (Tawil and others 1999), Pakistan, and Bangladesh (enkins, Abmed, and others 2001), and can provide templates for such projects elsewhere. Fol- lowing a situational assessment, HIIV/AIDS prevention outreach activi- ties have recently been initiated among MSM in Beirut, Lebanon. Morocco appears to be the first country in the region to have devel- oped H[V prevention programs for male sex workers (Boushaba and Hammich 2000; Boushaba and others 1999); more recently, such a pro- gram has been started in Lebanon as well. Not enough is known about the level of commercial male-to-male sex elsewhere in the region. WXuneirable Giroiups For the most part, the behaviors mentioned above are believed to be prac- ticed by a small minority of people who, nevertheless, could spark a more widespread epidemic. The majority of people in all countries are not prac- ticing high-risk behaviors, nor do we expect the majority to become in- fected with HV There are, however, a large number of people in the MENA/EM region, as elsewhere, in a state of vulnerability, which simply means that if certain factors shift in their lives, they will be at risk. Pre- venting infections in this group is a longer-term investment in overall A Typology of Risk Factors 33 health, particularly sexual and reproductive health, and contributes to re- ducing both adult and infant/early childhood mortality rates as well as the extraordinary cost of !lTV/AIDS to a nation as a whole. In the MENA/EM region, these groups encompass, among others, migrants (including work- ers, refugees, and tourists), noninjecting drug users, and youth. Migrants It is important to state at the outset that studies indicate that being a mi- grant per se is not a risk factor for HIV. But the reasons people migrate and the conditions under which they live during and after migration often increases vulnerability and can trigger high-risk behavior. Migra- tion for work is largely driven by poverty, unemployment, and a search for a better livelihood. In the MENA/EM region, the nation with the largest number of out-migrant workers is Egypt, with an estimated 3 million people, the majority of whom are men, working mostly in the Gulf countries. Reportedly, between 150,000 and 350,000 people are screened for HIV yearly before they work abroad. Algeria, the Islamic Republic of Iran, Jordan, Lebanon, Libya, Morocco, Syria, and Tunisia also report high levels of out-migration. Exact figures are difficult to ob- tain, and most experts believe the figures are misleading because the level of clandestine migration is high everywhere. From the host country per- spective, Oman's total population includes 25 percent who are migrants from South and Southeast Asia. Saudi Arabia hosts 850,000 Filipinos. AIDS prevention takes on a truly international perspective in such cir- cumstances and requires multisectoral and multinational planning and cooperation. Returning migrants have often been among the first men registered with HIV or AIDS in numerous countries and, after transmitting the in- fection to their wives, are responsible for many of the first cases of pedi- atric EIIV infection. The associated general pattern appears to consist of several components. First, where migration is largely to Europe, for ex- ample, from the Maghreb, most infections have been associated with in- jecting drug use. Second, where migration has largely been to the Gulf countries (or for Lebanon, to West and Central Africa), most infections are associated with sexual activities, either commercial or otherwise. Third, it is difficult to prove that returning infected migrants become a core trans- mitter group. Although some of their wives and newborns may acquire the virus, the spread from them generally stops there. Changed life circum- stances, including learning that one has FElV, appear to effectively diimin- ish risk-taking for many people, and the force of infection does not sustain an epidemic from this group. There are, however, other concerns because 34 HIViAIDS in the Middle East and North Africa continuing migration means continuing addition of new infections in home countries, some inevitable transmission to local people, and, more critically, large numbers of uninfected men who return with newly ac- quired habits, either of drug use or commercial or casual sex. Well over a million Bangladeshi, Ethiopian, Indian, Philippine, Thai, and Sri Lankan women are contracted as domestic servants, cleaners with various companies and schools, and for other services in the region; a large proportion of these women may be at risk for commercial or coer- cive sex (Blanchet 2002; Heise, Moore, and Toubia 1996; Khaled 1995). As the women are rarely present for long periods, a quick turnover moves any HITV-infected woman out of these countries rapidly. In Sri Lanka, health authorities estimated in the mid-1990s that up to 40 percent of H-IV infections among women involved those who had been working in Gulf states, in particular as domestic workers (WHO/EAMRO 2000a). Bangladesh and Pakistan have also documented high rates of I IIV among returning migrants from the Gulf regions. WAomen from the Common- wealth of Independent States (CIS) are also brought in by agencies, and, recently, Bahrain has attempted to crack down on this aspect of the trade (Reuters 2001a). Migrant women are subject to sexual and other abuses around the globe because of their precarious financial and legal circum- stances. These issues have been reported repeatedly by the International Office of Migration (JOM), and their rmitigation will require excellent in- ternational and regional consultation and cooperation. Short-term or transit migration plays a part in the overall story in some nations. The southern portion of Algeria, having borders with Maurita- nia, Mali, Niger, and Libya, appears particularly affected (Bouakaz 1998; Bouakaz 2000). At Algeria's Malian and Niger borders, registered mi- grants rose from 37,054 in 1978 to 61,444 in 1994; in addition, 263,322 people came from Morocco and 851,601 from Libya during these years. These large numbers of people, in addition to the even larger number who are unregistered, often include military men seeking recreation, women seeking work and migrating toward Europe, traders going back and forth, and West and Central Africans in search of work. Because of the commercial market they create, sex workers from elsewhere in Alge- ria also migrate to the area, which has seen marked rises in reported STD rates and AIDS cases in recent years. Surveillance in the southern border region in 2000 showed that about 1 percent of antenatal women were in- fected with l IlI The rising local epidemic is understood to create a seri- ous strain on health services in the southern region of Algeria and con- tribute directly to the risk of HIV spread in the country. Refugees are an especially vulnerable group of migrants in the MENA/EM region. In the Republic of Yemen, for example, more than 70,000 refugees, mostly from Somalia and Ethiopia, have entered over A Typology of Risk Factors 35 the past decade (United Nations 2001). Marginalized, poor, and up- rooted, these men and women are vulnerable both to acquiring HTV as well as being deported rapidly if their infections are detected. Despite in- ternational conventions, the United Nations High Commissioner for Refugees (VNI-ICR) must convince any HIV-infected person to be repa- triated voluntarily or face incarceration in the Republic of Yemen. A re- cent assessment (see box 3.3) has shown that the Republic of Yemen has developed little capacity to manage HIV-infected people, whether they are Yemeni or foreign-born. The reality of refugee vulnerability to HITV is highlighted by the find- ings of a recent survey in Sudan, which showed that 4.3 percent of a sam- ple of 470 refugees were infected with I-IIV (Sudan National AIDS Con- trol Program 2001). Little is known about the risks of acquiring HIV among the many Palestinian refugees in the Middle East or the millions of Afghani refugees in the Islamic Republic of Iran and Pakistan. The special problems of HIV prevention and care in conflict situations, as well as postconflict periods, have become increasingly salient and recog- nized (Hankins and others 2002). Sudan represents an emerging post- conflict situation in which greater freedom and mobility could increase the spread of HITV throughout the country and its neighbors. Recent po- litical circumstances affecting the Middle East and North Africa are likely to worsen this scenario and require considerable attention. Typically, migrants have less access to health and educational services than residents. AIDS educational campaigns, such as those on World AIDS Day, do not address them, yet they and tourists are easy targets in the common discourse of AIDS as a foreign disease. In Jordan and Egypt, for example, contact between tourists and nationals working in the tourist industry is seen as a situation creating risk of HIV infection. Wherever testing or other studies of local people in the tourist industry have taken place, as in Egypt, there was no evidence of HI\V transmission; however, sampling was by convenience and cainot be considered representative. Tunisia has 5 million tourists per year, but local experts recognize that local citizens may be at risk even without contact with tourists. Tourism among well-off young men in the MENA/EM region may also contribute to the overall risk of HIV spread. It is estimated that Egypt alone receives 1 million Gulf tourists yearly. In a survey of male members of social clubs in Muscat, Oman, in 1995 (n = 245), 13 percent reported extramarital relations over the past 12 months (\07HO/EMRO and UNAIDS 2001). One report states that tests of blood donors in the Republic of Yemen showed that HIV was more prevalent among those who had traveled out (1.5 percent) than those who had not (0.05 percent) (U.S. State Department 2001). 'While adequate in-depth studies have not been undertaken to verify actual patterns, the assumption made most 36 HIV/AIDS in the Middle East and North Africa often is that young men (and to a smaller extent, young women) engage in risky sexual behavior when away from home or with foreigners be- cause of the lower risks of discovery. This assumption seems reasonable and, for the past and present time, would help explain the continued low prevalence of HIV in the region as a whole; that is, wvith the exception of Ethiopia, Thailand, and certain states of India, countries sending work- ers into the region or countries heavily visited by tourists from the re- gion are themselves low-prevalence countries. This fortunate circum- stance cannot be depended on in the future. Programs to prevent HIIV spread among mobile people of all sorts must begin with acknowledging the numerous structural features of vul- nerability in each case. These features may include severe poverty in the home country, recruitment by illegitimate brokers, debt bondage, con- tracts with few safeguards, inadequate legal and other support services, language difficulties, crowded or restrictive living conditions, and, with illegal migration, great reluctance to utilize any services at all. Often mi- grant workers consider saving money to send home the highest priority, forgoing even the simplest self-care and placing themselves at great risk of abuse from others. Loneliness and the lack of a support network drive many to activities they would not engage in at home, including sex work, pimping, and so forth. Where migrant labor is well organized, these neg- ative features can be minimized, with both sending and receiving coun- tries providing coordinated HIV prevention education and services. Where migration is of the short-term, transit variety, cross-border pro- grams implemented by INGOs and the IOM in the Mekong region of Asia, as well as those supported by the World Bank, UNAIDS, and other partners in West Africa, could serve as models to be examined and adapted for the unique conditions in the MENA/EM region. The single greatest factor influencing sexual behavior in all societies is age. Even where social, religious, and cultural values curtail the levels of experimenting with sex before marriage, vulnerability among youth is great. A substantial portion of youth in the region appears to be at risk. In Tunisia, 21 percent of all people with I-V are between 15 and 24 years of age, and 93 percent are single. In Morocco, 63 percent of AIDS cases are among single people, and 40 percent of STDs are recorded among young adults between 15 and 29 years old. The HIV epidemic in Morocco is increasingly reaching young women as often as young men. In Djibouti, among recorded AIDS cases, 3.8 percent are found among those 15 to 19 years old and 43.6 percent anmong those 20 to 29 years old A Typology of Risk Factors 37 (Etchepare 2001; Rodier and others 1990). The sex ratio shows that younger women are more often infected, with 54.3 percent versus 42.7 percent among young men. As these figures from Djibouti reflect infec- tions that occurred 5 to 10 years ago, the early age at which a large pro- portion of people became infected is striking and disturbing. The tension between maintaining traditional values and modernizing societies through improved educational, health, and communications opportuni- ties is a serious source of social and political conflict. In the MENA/EM region, as elsewhere, religious and other authorities use their influence to counter activities viewed as threatening to the maintenance of tradi- tional forms of social control, particularly those surrounding sexuality, and, even more specifically, female sexuality (Althaus 1997; Anderson 2001; Bourqia 1992; Davis 1992; Ghannam 1997; Ghurayyib 1992; Goodwin 1995; Government of Yemen/UNICEF/World Bank/Radda Barnen 1998; Govindasamy and Malhotra 1994; Guenena and Wassef 1999; Hal 1997; Heikel and others 1999; Heise, Moore, and Toubia 1996; Helmy 1999; Ilkkaracen 2000; International Planned Parenthood Foundation [IPPF] 2001a; Khattab 1996; Khattab, Younis, and Zurayk 1999; Lane 1992; Masrerson and Swanson 2000; Mehra and Feldstein 1998; Mernissi 1985, 2000; Obermeyer 2000). Such phenomena as tra- ditional female genital surgery, minimum age at marriage, and related is- sues have periodically become contentious symbols of this struggle. Meanwhile, forces of economic globalization and communications proceed apace, with many nations undergoing extremely rapid social and economic changes. The rise in family planning acceptance throughout the region over the past decade or two has significantly reduced fertility almost everywhere (except the Republic of Yemen), and given the demo- graphic momentum, has resulted in a large cohort of youth. In Egypt, Jordan, and Morocco, about one-third of the entire population consists of those between 15 and 29 years of age. Similar proportions are found elsewhere. Coupled with economic forces such as high levels of unem- ployment among youth nearly everywhere, improved levels of female education, and increasing bride prices in some areas, the average age at marriage has been increasing rapidly in several countries, giving rise to widely acknowledged concerns about I1TV vulnerability among the re- gion's youth. Efforts have been made to conduct knowledge, attitude, and practice (KAP) surveys among youth in Algeria, Djibouti, Egypt, Jordan, Kuwait, Lebanon, Morocco, Oman, Saudi Arabia, Somalia, Syria, Tunisia, the UAE, and elsewhere, but sexual practices were queried in only a few coun- tries (Al Mulla and others 1996; Al-Owaish and others 1999; Essghairi 2000; Family Health International 2001c; Farghaly and Kamal 1991; Faris and Shouman 1994; Gueddaina and others 1996; Ministere de la 38 HINAIDS in the Middle East and North Afica Sant6/PNILS 2001; Ministry of Health in collaboration with UNAIDS and 'WVHO 1999; Petro-Nustas 1999; Reysoo 1999; Saleh and others 2000; Seif El Dawla, I ladi, and Wahab 1998). A recent survey in Sudan has gath- ered considerable useful behavioral information (Sudan National AIDS Control Program 2002), wvhereas in Somalia, one survey appears to have not trained interviewers adequately and little useful information on sexual behavior was obtained (WHO/EMRO 2000c). In Djibouti, another study conducted as a stratified sample among 200 young people (100 males, 1 00 females) at six schools and two out-of-school venues clearly demonstrated the sexual risk-tak-ing of youth, with 71 percent of those between 15 and 19 years old admitting to having had sexual relations (undefined); only 39 percent claimed to have ever used a condom (Ministere de la Sant6'PNLS 2001). The data, which are unanalyzed by sex, seem to demonstrate that an increasing proportion of very young people are initiating sexual rela- tions at an early age. Among those currently between 15 and 19 years old, 11 percent admitted to having had their first sexual relations at the age of 13; but among those between 10 and 14 years old, 18 percent stated that they had sex at the age of 13. Discussions about sex and HIlV were still con- sidered taboo although, with increasing age, larger numbers of young peo- ple had learned about the disease and about condoms. Less detailed information is available from elsewhere, but a 1994 study in Jordan showed that 7 percent of college students admitted to nonmarital sex, while another national study in 1999 among the general population between ages 15 and 30 showed that 4 percent admitted the same. Among the sexually active, 90 percent mentioned opposite-sex partners only, while 10 percent had partners of either the same sex or of both sexes. Less than half reported ever using condoms. Levels of knowl- edge about HIIV had dropped between surveys (Family Health Interna- tional 2001c). In Morocco, qualitative studies seem to show an increasing accept- ance of premarital sex among youth, including sex for material exchange, and a certain tolerance of sex work as a means of improving economic status. Simultaneously, young people seem to believe that condoms de- stroy pleasure and are associated with flagrant misbehavior such as adul- tery; that admitting to having an STD will lead to punishment, so they prefer to go to pharmacies for treatment because they distrust public fa- cilities; and that AIDS is not yet perceived as a problem for Moroccans (Davis 1992; Farza 2001). An excellent national study of youth in Egypt did not examine sexual risk-taking but, nonetheless, yielded very useful information (see box 3.3). Like parents everywhere, MENA/EM parents seem convinced that sexual activity among youth is increasing relative to their own youth. Elsewhere, improved nutrition and health care have been shown to influ- A Typology of Risk Factors 39 BOX 3.3 Young People in Egypt Egypt has tlhe largest population in the Arab world with about 68 million people, and, with a current population growth rate of only 1.72 percent, the demographic momen- tum has created the largest cohort of youth in Egypt's history. There are at least 13 mil- lion young people between the ages of 10 and 19, representing one-fifth of the total pop- ulation. Educational enrollment has increased at all levels, although marked disparities occur by region. Yet one-tliird of all 10-year-old girls have never been to school, and 12 percent are married before 19 years old, with an average age at first pregnancy of 17.6 and associated high levels of stillbirths and infant deaths. In some regions, as high as 20 percent of girls are married by age 19. These married adolescents are also often the least educated. Regardless of education, few youths understand the nature of their changing bodies, reproductive functioning, and sexuality. In 1997, a national survey was conducted of 9,128 adolescents and parents. Results showed that 86 percent of unmarried girls continue to be circumcised, despite a grow- ing conviction that it is not necessary, especially among the more educated. In addi- tion to the gender-biased problems in education and employment that girls face, the study revealed that boys also must handle serious obstacles, including unexpectedly high rates of undernourishment, stunted growth, exposure to violence, and excessive work force participation. On a positive note, many Egyptian youngsters expressed con- fidence in their future and felt able to discuss a wide variety of topics with their par- ents, except for sex. Sex education in school is confined to biology, but does not seem to be well executed, given the level of understanding revealed in this survey. Most young people have not learned even basic information from schools, and while they would like to discuss these issues with their families, parents do not feel comfortable giving them this information. The hotline established in Cairo for AI1DS prevention yields evidence that many mis- conceptions and dangerous gaps in understanding exist. Sexual mores are well under- stood, and young women know they are expected to preserve their virginity to maintain family honor. Some will be subject to virginity examination at the time of marriage, and therefore practice various ways to avoid discovery. Young men may avoid social appro- bation through use of sex workers and sex with other males. These realities are known to exist but little progress has been made in addressing them, despite the fact that all concerned recognize they place the coming generation of adult Egyptians at risk of H:V infection. Souirce: El-Gawhary 1998; Population Council 1998. 40 HIV/AIDS in the Middle East and North Africa ence the age of sexual maturity, lowering it and creating considerable dis- junctive problems with a rising age at socioeconomic maturity. It is likely that this has been occurring in the MENA/EM region as well, although adequate studies have not been conducted to demonstrate this or other factors directly contributing to vulnerability among youth. However, studies in Europe and, more recently, in Uganda and the United States, have shown how improved adolescent education and services can reverse these trends, or, at a minimum, significandy reduce the harm caused by earlier and more frequent premarital sexual activity. Tunisia has taken steps to bring these services to its young; Egypt plans to develop an im- proved reproductive health program that will incorporate the "life skills"2 approach. There is clearly a great deal of scope in the region to develop culturally sensitive reproductive health education and services for young people that could affect decisionmaking regarding risky behavior. Children under specially disadvantaged circumstances must be noted. Refugee children are a case in point, as well as specially marginalized groups such as the caste in the Republic of Yemen known as Akhdam (lit- erally meaning "servants"), a historical minority kept at the bottom of so- ciety and considered by observers to be at greater risk of acquiring HIV (Republic of Yemen and UNDP 1998; UNICEF 2001b; United Nations 2001). In addition, there are an estimated 200,000 street children in Egypt, and increasingly large numbers working or begging on the streets in the Republic of Yemen (Government of Yemen/UNICEF/World Bank/Radda Barnen 1998), the Islamic Republic of Iran, and elsewhere. Studies suggest that such children are vulnerable to sexual and drug abuse in most cities of the world. ST[/1TD Patients Levels of STIs are generally unknown in the region. Because most in- fections are asymptomatic in women and many, especially viral infec- tions, are asymptomatic in men, few data exist that would reflect actual prevalence of infections. More commonly, registered patients with dis- ease symptoms (STDs) are reported. Few proper studies of STIs among high-risk groups have been conducted, and where they have, these re- sults usually have not been officially published. Fear and shame continue to characterize STD issues in much of the region. By law in most countries, STDs are notifiable diseases, but reporting is generally limited, data are confusing, and private practitioners seldom comply. The high level of stigma associated with STDs drives the ma- jority of patients to private practitioners. Furthermore, WHO has en- couraged most countries to report using the syndromic method,3 but the A Typology of Risk Factors 41 accuracy of this approach without validation both for treatment and re- porting, particularly for women, has been repeatedly questioned (Garg and others 2001; Hawkes and others 1999; Heikel and others 1999; Ryan and others 1998). Where assessments have been made of service providers, preventive counseling and condom promotion and distribu- tion have not been observed. Proper STI surveys are needed in every country in the region both to document community levels of pathogens and the main service indicators. Improvement of STD treatmnent pro- grams is a major investment in HIV prevention. Anecdotal evidence suggests high levels of STDs in the Republic of Yemen, but the source of these data has not been found. The Republic of Yemen has no STD clinics, and women only have access to gynecol- ogy clinics. The prevalence of syphilis by serology among blood donors is reported as 2 percent (Global Fund to Fight AIDS, TB, and Malaria [GFATM] 2002). One study in Jordan found high levels of chlamydia while another found none. Laboratory issues may be a concern in Jordan (Family Health International 2001c). In Oman, where health facilities, including laboratories, are considered to be good, antenatal women are screened for syphilis, but rates are very low; 2 percent of blood donors test positive. STIs are reportedly rare in Lebanon, which one study in the Bekaa Valley seems to confirm (El Nakib 2001; WHO/EMRO 2001b, 2001c). STI/STDs among known high-risk groups, however, have not been reported. In Syria, syndromic reporting has been intro- duced in four governorates. These reported 2,342 STD cases for the pe- riod of April-June 1999. Other studies have been done but remain unre- ported (Syrian Arab Republic 2001a, 2001b). In Egypt, some studies show significant levels of reproductive tract infections (RTIs) among women, but not necessarily STDs (Khattab and others 1996). More re- cently, an important study conducted by Family Health International has shown that 36.5 percent of sex workers (n = 52), 23.8 percent of MSM (n = 80), 5.3 percent of drug users (n = 152), 8.3 percent of family plan- ning clinic attendees (n = 108), and 4 percent of antenatal clinic atten- dees (n = 607) had detectable STIs (Family Health International 1998-2000). In Algeria, STD notifications rose in the south between 1990 and 1995 by 22 percent and again in 1996 by 26.7 percent (Bouakaz 2000). In Morocco, increasing numbers of STI cases are reported, but that may well be the result of increasing coverage of improved STD clin- ics. It is recognized that at least 50 percent of all patients go to private practitioners or use self-treatment, and much more work is necessary to educate the populace about STDs (Heikel and others 1999). Among an- tenatal women, 1 percent are reported to test positive for syphilis. A study of 294 secondary students in Djibouti found no syphilis despite the presence of 46 percent in sex workers (Rodier and others 1993c). In 42 HIV/AIDS in the Middle East and North Africa other words, data are sketchy, do not reflect community levels, and indi- cate that a considerable proportion of infected people do not seek proper treatment. A summary estimate of STI annual incidence levels for 1999 by region compared with Djibouti shows that reported levels in the MENA region are considerable, even if lower than Djibouti itself and significandy lower than Sub-Saharan Africa (see figure 3.2). Structural Vulnerability Structural factors that contribute to risk can be considered to be those that usually affect more than one person and result from the way socie- tal institutions are structured. For example, these factors may include health services and health service policies, the labor market, and educa- tion policies, or the structure and functioning of a refugee camp or prison. As in any complex etiology, no one factor is fully determinative, and the conditions in communities with increased vulnerabilty to HTV vary considerably. However, patterns have emerged and regularities are discernable. Where overall social and economic conditions are poor, not only is there a greater chance of HTV spreading, but there also is less ca- pacity to handle both prevention and care. Measures of larger social and economic factors, such as the Human Development Index, literacy rates by gender, unemployment, expendi- tures on health and, as a measure of women's health, maternal mortality, FIGURE 3.2 Estimated Annual Incidence of STIs: Africa, Middle East, and Djibouti Incidence (per 1,000) 300 254 250 200 150- 85 100 60 50- Sub-Saharan Africa Middle East and Djibouti North Africa Source: Etchepare 2001. A Typology of Risk Factors 43 give an indication of the relative state of development in each country in the region (see table 3.1). The proportion of the population that is urban also is an important indicator of potential vulnerability. Although, ordi- narily, all services are better in urban areas, drug use and sex work are both more common in urban than in rural areas. These indicators, like all such measures, obscure real differences within population segments in a country. For example, although the total proportion of Moroccans living in poverty is estimated variously at be- tween 13 and 16 percent, rural Moroccans are worse off, representing 60 percent of all the poor. This situation drives high levels of internal mi- gration, with increasing numbers of young men and women searching for jobs in urban areas. While the national unemployment rate is 22 per- cent, the rate is 35.5 percent among youth 15 to 24 years old. Similar dis- parities are found elsewhere, for example, in Tunisia, with double the rate of unemployment among youth (36 percent for 18- to 19-year-olds, 31.1 percent for 20- to 24-year-olds) compared with overall national fig- ures (16 percent). In Algeria, a general unemployment rate of 30 percent particularly affects young graduates, who may account for 75 percent of total unemployment (World Bank 2002a). With increasing investmlent in education and stagnating overall economic growth, the levels of unem- TABLE 3.1 Selected Indicators of Development by Country and Human Development Index Human Male Female GDP spent Maternal Development literacy literacy on health mortality Urban Country Index (percent) (percent) Unemployment (percent) (percent) (percent) Kuwait 35 82 75 - 2.9 5 97 Bahrain 37 89 79 - 5.7 60 92 Qatar 41 79 80 - 2.8 - 90 UAE 43 79 80 - 2.5 26 83 Libya 65 88 63 - - 220 84 Lebanon 69 95 90 - 5.3 100 86 Saudi Arabia 78 72 50 - - 130 82 Oman 89 95.5 79 - - 190 77 Jordan 94 93 79 14 7.9 41 70 Iran, Islamic Rep. of 95 77.5 56.4 12.7 6.0 37 61 Tunisia 102 82 64 16 5.9 70 63 Algeria 109 74 49 30 4.6 160 56 Syrian Arab Rep. 111 87 56 - _ 180 52 Egypt, Arab Rep. of 120 64 39 11.3 1.8 170 45 Morocco 126 62 34 22 3.6 230 52 Yemen, Rep. of 148 63 24 18 1.1 1,400 25 Djibouti 157 73 43 45 7.0 740 82 -Not available. Source: UNDP 2000; World Bank 2001 a,World Bank 2002b. 44 HIV/AIDS in the Middle East and North Africa ployment among college graduates could negatively affect political sta- bility, drug use, and I-HV prevalence levels alike. Rates of literacy are also somewhat misleading with reference to HIV prevention because few of the messages and materials provided by AIDS programs are clear, give explicit information about risk, or even mention the use of condoms for prevention. Most educational efforts are inade- quate, and surveys of knowledge among people in the most literate coun- tries reveal large gaps in understanding. For example, in one country, while 71 percent of adolescents aged 16 to 19 reported knowing about HIV/AIDS, only 19 percent knew that condoms are protective. In Jordan, a large survey showed high levels of knowledge about modes of transmis- sion, yet one-third of respondents thought IIIV could be acquired from mosquitoes and toilets, and only 29 percent had ever seen a condom. Half of the respondents thought they could see if a person had lIIV, an ex- tremely dangerous belief (Family Health International 2001c). A large study in 1999 showed that, while most of the people in Kuwait were aware of the main modes of liP! transmission, a gap existed about modes that did not transmit the disease. This lack of understanding was reflected in their attitudes and practices toward FIIV/AIDS patients and a lack of pro- tective behavior change (Al-Owaish and others 1999). In most of the re- gion, media constraints limit discussions. Access to the Internet remains relatively low in much of the MENA/EM region. The lack of access to information is a serious factor increasing vulnerabiiity. Inequity in resource and income distribution appears to play a major role in fueling HIiV epidemics. The exact mechanisms through which this occurs are not clear, but are likely to include differential access to health care, pertinent information, and the power to protect oneself from HIV. Sociological studies have long shown that rising expectations not accompanied by rising opportunities produces fertile ground for so- cially disruptive forces and movements of many kinds. Stagnant economies and high unemployment among youth lead to increased drug-, violence-, and sex-related risk scenarios. Strong family ties can counter this trend, but where families are weakened by long-term mi- gration, low incomes, or family disruption, youth are unprotected. Gender bias in educational and employment opportunities specifically places women at an even greater disadvantage. In Uttar Pradesh, India, studies have revealed the close link between domestic violence and the risk a wife has of acquiring an STD from her husband (Martin and oth- ers 1999). In Bangladesh, studies suggest that accessing Grameen Bank-type loans plays a role in reducing domestic violence by making women's lives more public (Schuler and others 1996). Women without economic power are less able to protect themselves from the risk of HIV; even more threatening is the fact that most women never perceive that A Typology of Risk Factors 45 they are at risk of acquiring HIV at all, until it is too late. The highest risk group in any society consists of faithful, uninfected wives married to (unfaithful) lIIV-infected husbands. As only a small proportion of men ever reveal their infections (if known) to their wives, these women are thoroughly vulnerable, as are their unborn children. Thus, improving women's economic power can have direct and indirect benefits for HIV prevention. Inequity in resources takes on new and ominous meanings in the era of AIDS, and governments have an obligation to reduce poverty and provide the means of protection to all their citizens. Long-term invest- ment in reducing the structural factors contributing to vulnerability for HI-V infection in the region can only be approached through multisec- toral responses. This objective will require a significant effort in plan- ning, through consultative and collaborative processes, appropriate re- sponses that can shore up the gaps that allow HIV to enter social networks and spread. 1. Others include health workers and multiply exposed patients but, under ordinary conditions, the risk to these groups is easily reduced. 2. Life skills programs have been developed by UNICEF, the United Nations Population Fund (UNFPA), and many other agencies, for both in-school and out-of-school youth. They are locally adapted to the needs of young people, incorporate local values, teach skills, are often peer-led, and view sexual and drug-taking behavior within the larger context of re- sponsible decisionmaking for healthier life styles. 3. Syndromic management and reporting of STDs focuses on symp- toms only in an effort to avoid the use of laboratory tests in resource- poor settings. CHAPTER 4 $$sess'Mg the potentDaW Fconomk ftad@ Of MYNDS0 The human cost of HIV/AIDS is incalculable, from the pain and guilt surrounding personal and intimate relationsllips to threatened social and political security at the state level. It is nonetheless instructive to exam- ine those costs that can be calculated, for these too are great and can have a major impact on the futLre of a nationi. This chapter summarizes the results of an assessment of the potential macroeconomic impacts of HTV/AIDS in selected MENA/EM coun- tries. Where prevalence levels are low, studies of eventuLal economic im- plications are useful to demonstrate the costs and losses that will accrue if investments are not made rapidly to avoid the outbreak of an epidemic. As discussed previously, HIV epidemics are sensitive to changing eco- nomic and social factors and, in the MENA/EM region, current meth- ods of surveillance are unable to detect changes wlhere they are most likely to take place. Therefore, it is essential to raise awareness for all concerned of the importance of investing now to avoid serious conse- quences later. This chapter first summarizes the results of recent research on the socioeconomic determinants of the HIV/AIDS epidemic and extrapo- lates these results to the case of MENTA/EM countries. One of the goals is to assess how current official estimates of prevalence levels compare with those predicted by models on the basis of cross-country data. The chapter then develops and exploits a model of growth to evaluate the potential socioeconomic impacts of the HIV/AIDS epidemic in selected MENA/EM countries. The focus is on output, health expenditures, and poverty. Finally, the chapter assesses the gains from preventive invest- ments made promptly and widely. 47 48 HIVWAIDS in the Middle East and North Africa Exploratory Analysis of the Socioeconomic Determinants of HIV/AIDS Prevalence Levels The HIIV/AIDS epidemnic and economic and social environments are in- terrelated. On the one hand, social norms and economic incentives and constraints shape individuals' behaviors, which are among the main de- terminants of the evolution of the epidemic. On the other hand, by af- fecting the health of individuals, reducing life expectancy, and increasing mortality, -HIV/AIDS influences the economy and individuals' behav- iors. This discussion of the socioeconomic factors that facilitate the spread of the epidemic is based on the results of two exploratory statisti- cal analyses of cross-country data on HIV/AIDS prevalence levels and indicators such as income per capita, inequality, male and female literacy rates, female labor force participation, the share of tourism in total value added, and migration. Not surprisingly, these macrolevel indicators can only explain part of the international variation in NW1V/AIDS prevalence levels. Much of this variation remains unexplained and can be attributed to the heterogeneity of the economic, social, and cultural environments. Nonetheless, the goal of the analysis is not to provide rigorous estimates of the impact that changes in economic growth or income distribution could have on NTV prevalence, which would be a hopeless exercise given our rudimentary knowledge about how the epidemic and the economy interact. Instead, the purpose is simply to identify broad categories of so- cial and econornic factors that are likely to influence the diffusion of the epidemic in MENA/EM countries. The first analysis explores correlations between urban 1HIV/AIDS prevalence levels and eight socioeconomic indicators (Over 1997). These indicators include a crude measure for the age of the epidemic computed as the time elapsed since the first case was identified, the gross national product per capita, the foreign-born percentage of the population, the percentage of Muslims in the population, the Gini index of inequality,l the male-female literacy gap, the 'urban male-female gender ratio for ages 20 to 39, and the share of militarv forces in the total urban popula- tion. The study uses 1997 data for 17 countries from Sub-Saharan Africa, 15 from Latin America and the Caribbean, 14 from Asia (including India and China), and 4 from the Middle East. The second analysis, developed for the current study, predicts inter- national variations in 1I1V/AIDS prevalence levels as a function of in- come per capita, female participation in the labor force, female literacy, the Gini index of inequality, the share of tourism-related activities in gross domestic product (GDP), and migration. The study uses 1997-99 data for 92 countries, including 27 from Africa, 17 from Asia, 13 from Eastern Europe and Central Asia, 21 from Latin America, 9 from the Assessing the Potential Economic Impact of HIV/AIDS in the MENA/EM Region 49 Middle East, and 5 from North America and Western and Central Eu- rope. The quantitative results of these studies are summarized in tables A.1 to A.3 in the technical appendix. Below we discuss the main mes- sages. We wish to emphasize that, while the two studies use different data and indicators, both arrive at similar conclusions. One of the results of the studies is that poverty and income inequal- ity facilitate the diffusion of the HIV/AIDS epidemic, although the exact mechanisms remain unknown. Over's study (1997) shows that HIV prevalence levels increase when income per capita declines and inequal- ity increases. On average, a US$2,000 increase in income per capita is as- sociated with a 4 percentage points' lower level. Similarly, if inequality (as measured by the Gini coefficient) is reduced from 50 to 40 (the dif- ference between Honduras and Malawi), the prevalence level would de- crease by 3 percentage points (see figure 4.1). Our analysis also finds a strong correlation between inequality and HIV/AIDS prevalence. The impact of income per capita, however, varies by region. There are various plausible explanations for the finding that poor countries with an unequal distribution of income are likely to be more vulnerable to the epidemic. The poor have less knowledge about the risks of HIAV/IDS and prevention methods; poor populations may have less freedom to define their choices or behaviors (particularly women); poverty may induce low-income individuals to engage in commercial sex (fueled by migration from rural to urban areas prompted by differentials in expected income); poor men may need to delay marriage and there- FIGURE 4.1 Poverty, Inequality, and HIV/AIDS Prevalence Urban adult HIV prevalence 35 35 Botswana ZambiaGuyana Zirnbabwe 20 . i , ZimbabweBotswana 20 Guyana HaitiZambii., Honducris.. Rep. of Korea Rep. of Korea i Ma wi M wi .Pa.N.- ; .'rhail4t Thailand * . . .. . . ~~~~~pain Plsai !:~ . ' . . ARE China C China .01 .. .01 ._ ._ I _ I 1,100 3,000 8,100 0,3 0.4 0.5 0.6 GNP per capita (1994 US$) Inequality of income Note: ARE, Arab Republic of Egypt. Source: Over 1997. 50 HIV/AIDS in the Middle East and North Africa fore delay having a stable paruter; and among the poor, spouses may need to leave the household frequently to find jobs. Regardless of the mechanisms, however, international evidence suggests that sustainable growth and a more equal distribution of income are factors that con- tribute to diminishing the spread of HITV In MENA/EM, 30 percent of the populationi is living on less than US$2 per day (see table 4. 1). Growth remains elusive, and therefore it is unclear that poverty will retreat over the medium term. Hence, unless actions are taken to tackle HIV/AIDS, the conditions to facilitate the diffusion of HIV epidemics in the region are falling into place. The studies use various indicators to proxy women's rights in society and provide evidence that gender inequalities are associated with higher HIV/AIDS prevalence. Over (1997) finds that indicators such as the share of female population in urban centers and the gap benveen male and female literacy rates are correlated with HIV/AIDS prevalence lev- els. Urban centers where men aged 20 to 39 outnum3ber women have considerably higher levels of HIV/AIDS prevalence. Countries where women are less educated than men also tend to have higher prevalence (see figure 4.2). Our study finds that, other things being equal, countries with a high female illiteracy rate are also likely to have higher liY/AIDS prevalence levels. The results also suggest that low female participation rates in the labor force (less than 30 percent) are correlated with high ErV/AIDS prevalence. High participation rates (above 40 percent), however, are also associated wvith higher EITV prevalence. One possible explanation is that, other things being equal, societies with high female participation in the labor force might also have more liberal sexual be- havior. The key message from the analysis is that countries where women are not empowered, and therefore have less or no control over TABLE 4.1 Population Living on Less than US$2 a Day, by Region Number in 1987 Number in 1988 Region Percentage in 1997 Percentage in 1998 (miilion) (miilion) EAP 67.0 48.7 1,052.3 884.9 ECA 3.5 20.7 16.3 98.2 LAC 35.5 31.7 147.6 159.0 MENA 30.0 29.9 65.1 85.4 SA 86.3 83.9 911.0 1,094.6 SSA 76.5 78.0 356.6 489.3 Total 61.0 56.1 2,5490 2,811 5 Note: EAP,EastAsian and Pacific region; ECA,Europe and Central Asia region;LAC,Latin America and Caribbean region;MENA,Middle East and North Africa region; SA, South Asia; SSA, Sub-Saharan Africa Source: World Bank 2001 f Assessing the Potential Economic Impact of HIV/AIDS in the MENA/EM Region 51 FIGURE 4.2 Gender Inequality and HIV/AIDS Prevalence Urban adult HIV prevalence 35 35 - Botswana Botswana Zimbabwe 20 - -eh . 2Rep. of Koreazh. of.Korea 20 - ~~GuyaKllbabweZabita Zambia Zrbi R-aiti. 5 H ao V 5 Spain Hond$hajTbailaM jaR Hon uras.'. aMzalvtjr stn ,irocco ARE~ ~ ~ ~ ~~~~~~~~~~~~R 1 . 1. - 7.;* : AREChina China. .01 1.01 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 0 10 20 30 Ratio of urban males to females, age 20-39 Male-female literacy gap Note: ARE, Arab Republic of Egypt. Source: Over 1997. the frequency and type of sexual encounters (in or out of the workplace), provide an ideal environment for the spread of the HIV/AIDS epi- demics. These results are backed up by other studies showing that the risk of acquiring an STD for married women is associated with domes- tic violence (Martin and others 1999). In MENA/EM countries, gender inequalities in terms of access to labor markets and education have been reduced over time. Nonetheless, the gaps are still considerable and this may also contribute to the spread of HITV/AIDS. Migration and tourism could contribute to the diffusion of HIV, but the evidence from international statistics is not robust. Over's study (1997) on urban HIV prevalence levels finds that countries where 5 per- cent of the population are immigrants (implying a more important flow of immigrants) can have an HIV/AIDS prevalence level that is 2 per- centage points higher than the prevalence of a country with no immi- gration. However, migration per se is not the risk factor. Risks are de- termined by the recruitment, labor, and living conditions for migrants in host cotntries and, importantly, the differential in I-IIV prevalence lev- els between source and destination communlities. Nonetheless, as the density of flow increases, the probability increases of finding migrants who become part of high-risk groups. Migration and tourism can also contribute to fueling the epidemic by creating interaction channels with high-risk population groups in other 52 HIV/AIDS in the Middle East and North Africa countries affected by 4-V/AIDS. However, despite the intuitive sound- ness of these arguments, when looking at the role of tourism and migra- tion, our study does not find statistically significant correlations with HIV/AIDS prevalence levels. One explanation is that in countries that are more exposed to the risk of increased tourism (Thailand), public in- terventions not accounted for in the model have been put into place to reduce HIV prevalence levels. Furthermore, differences in the HIV prevalence in source and destination countries will affect the outcome. Another explanation is that the potential negative effects of increased tourism on F-IT\ prevalence are neutralized by the positive effects on higher economic growth. In NMENA/EM countries, both migration and tourism are essential parts of the economy. As liberalization takes place and restrictions on trade and capital flows are reduced, migration and tourism are expected to increase, thus bringing economic benefits. Col- lective risk will be enhanced, though, where part of the tourism is related to commercial sex or drug use, or when migrants access or act as sex workers. When the model developed for this study is applied to MENA/EM countries, it predicts prevalence rates that are higher than current offi- cial estimates.2 Given current levels of output per capita, participation of women in the labor force, the level of income inequality, and rates of fe- male illiteracy, I-IIV prevalence levels in MENAIEM countries could be between 0.2 and I percentage point higher than current estimates (see figure 4.3). The predicted prevalence levels are sensitive to model spec- ification but, in all cases, they are higher than those currently reported. Differences could be explained by the absence of appropriate surveil- lance systems. Potental Sod econolrnki Consequences of the H0llWAlDS Ep5dernk The more visible impact of HIV/AIDS is on health and life expectancy. It is estimated that in countries with relatively high prevalence levels (above 5 percent), life expectancy has been reduced to the levels observed 10 years ago (World Bank 2000e). Worldwide, while the share of deaths from infectious diseases is expected to decrease from 30 percent today to 14 percent in 2020, the share of these deaths from AIDS could increase during the same period from 2 percent today to 14 percent (Murray and Lopez 1996). AIDS is, of course, not the only health problemn. Malnu- trition and childhood-related diseases are currently responsible for more than 1.8 million deaths per year. TB kills 2 million people per year and malaria close to 800,000. It is expected that the number of deaths caused Assessing the Potential Economic Impact of HIVIAIDS in the MENA/EM Region 53 FIGURE 4.3 Potential Underestimation of Current HIV/AIDS Prevalence in Selected MENA/EM Countries Percentage 1.0 0.8 - 0.9 0.7 0.6 0.5 0.4 0.3 0. Source: Authors' calculations. by tobacco will increase from 3 million today to 8.4 million in 2020 (Murray and Lopez 1996). The particularity of HITV/AIDS is that the health impacts are likely to concentrate on young adults in their most productive ages. Thus, AIDS could become the second major cause of death among adults of working age in the world, posing a serious eco- nomic threat. Below, we discuss potential implications as they relate to MENA/EM countries. Ec070nZiC impacts through t.he loss of humnian capital. One channel through which HIV/AIDS affects the economy is by disrupting the human capi- tal accumulation process. Indeed, the human capital of a country is given by the size and quality of its labor force. FITV/AIDS is likely to affect both. First, there are effects in the size and productivity of the current labor force because of higher mortality and morbidity. Premature deaths represent not only losses in a productive factor, but also losses of the knowledge and experience embedded in it. Higher morbidity can also re- duce labor productivity, for instance, through high labor turnover (see figure 4.4), higher health insurance expenditures, and the need to put in place preventive measures.3 Second, HIV/AIDS can affect the accumu- lation of future human capital. Indeed, premature deaths tend to increase the number of orphans who are less likely to fully clevelop their physical and intellectual capacities.4 Morc important, in the absence of appropri- ate social protection systems, the income shock associated with the death of a household member, particularly if he or she is the main provider of income, can reduce investment in health, education, and nutrition for children (see Bell, Devarajan, and Gersbach 2003). Shorter life ex- 54 HIVAIDS in the Middle East and North Africa FlGi;RE 4.4 HIM/1O$S prevalence and Work@tr Turn¶ver Workers leaving the firm (percent) 3.0 2.5 2.0 1.5 1.0 0.5 0 15.0 17.0 19.0 21.0 23.0 25.0 27.0 HIV/AIDS prevalence rate in cities (percent) Source: Biggs and Shah 1996. pectancies resulting from higher mortality also reduce incentives to in- vest in education and health. Economnic impacts through lower efficiency. The need to finance additional health expenditures and reallocate resources toward curative anid pre- ventive measures may also reduce economic efficiencv. The impact of the epidemics on the health system is likely to vary widely from one countrv to another, depending on the type of technologies and services provided to IEITV/AIDS patients. In general, howevcr, the epidemic will bring higher demand for health services, increased health care costs, and there- fore higher expenditures. The macroeconomic consequences will de- pend, in part, on how additional expenditures are financed. In the case of public expenditures, governments will have to choose between increas- ing taxes, issuing debt, or simply reducing other types of expenditures (consumption or investment). For instance, higher expenditures in F-IV/AIDS curative interventions could crowd out other health or edu- cation expenditures (see figure 4.5). Higher private expenditures could also be financed either by reducing savings5 or substituting consumption. If lower savings dominate, then new investments and growth would be compromised. In all cases, necessary adjustments to finance additional health expenditures are likely to decrease welfare. Efficiency losses may also result as private and public resources used to supply HITV/AIDS- related services are reallocated away fir-om other more productive uses. Assessing the Potential Economic Impact of HIV/AIDS in the MENA/EM Region 55 FIGURE 4.5 Annual Cost of Treating an AIDS Patient versus Annual Cost of Primary Education Cost per person per year (US$) 7,000 - + Primary education 6,000 - - Treatment HIWAIDS 5,000 - 4,000 - 3,000 - 2,000 - 1,000 - 0- 400 1,400 2,400 3,400 4,400 Income per capita (USS) Note: HIV/AIDS expenditures as a function of income per capita in USS (Y) are given by: In E.,d, = 0.63931 + 0.95 In Y Primary education expenditures as a function of income per capita in US$ (Y) are given by: In Eed. = -3.3554 + 1.1589 In Y Source: Cyrillo, Paulani, and Aguirre 2001; Floyd and Gilks 2001; World Bank 2002b. Empirical studies measuring the economic impact of AIDS find con- tradictory evidence. Some studies find no correlation between 1E]IV/AIDS prevalence levels and economic growth (Bloom andl Mahal 1995). Other researchers (Bonnel 2000) find that growth rates in Africa's most affected countr-ies could have been higher by 1 percentage point in the absence of HTV/AIDS. Most macroeconomic simulation models, however, predict that reductions in economic growth- in affected countries could average 0.5 percentage point per year (Ainsworth and Over 1994). Recently it has been arguecl that by ignoring the impact of the epidemic on the accumulation of future human capital, most studies are underesti- mating the long-term economic costs (see Bell, Devarajan, and Gersbach 2003). Regardless, short-term economic impacts will depend on several factors, including the severity of the epidemic, the degree to which I-IlV/AIDS expenditures are financed from savings, the distribution of HlV/AIDS infections by workers' productivity level, the time lost from work by the person with AIDS, and the efficiency of labor markets. In general, short-termn macroeconomnic impacts would be lower in countries with high unemployment rates. This is the case in MENA/EM countries where today's unemployment rates are high, with official rates attaining 56 HIV/AIDS in the Middle East and North Africa 21 percent in Algeria and more than 30 percent in Djibouti and the Re- public of Yemen. In these cases, the slowdown in the growth rate of the labor force could be translated into a reduction in the unemployment rate and not GDP. Nonetheless, it is important to acknowledge that whether employed or not, a young adult is a potentially productive resource. Thus, the death of an unemployed person is still a loss of human capital. To evaluate the economic cost of HIV/AIDS in MENA/EM coun- tries, we use a model of growth that includes an HIV/AIDS diffusion component that formalizes transmission through two channels: sexual intercourse and the exchange of infected needles among IDUs. The model is necessarily a simplification of the complex mechanisms through which the economy and the HIV/AIDS epidemic interact.6 Nonetheless, it is able to capture the impact of the epidemic on (a) the size of the labor force, (b) labor productivity growth, (c) health expenditures, and (d) the savings rate of the economy. A schematic representation of the model is shown in figure 4.6. A more formal presentation of the model is provided in the technical appendix. The model was calibrated to nine MENA/EM countries (Algeria, Djibouti, Egypt, the Islamic Republic of Iran, Jordan, Lebanon, Mo- rocco, the Republic of Yemen, and Tunisia) on the basis of recent demo- graphic and economnic data and projections used in the World Bank Country Assistance Strategies (World Bank 1997a, 1997b, 1999b, 1999c, 2000c, 2001b, 2001c, 2001d; World Bank Group 2001) and the Algeria Social Expenditures Review (World Bank 2002a).7 The period of analy- sis is 2000-25. The model is not used for prediction purposes, but rather to explore the likelihood of a large number of plausible futures and asso- ciated economic costs. Hence, for each of the countries, we simulate the dynamics of the economy for 100 combinations of the model parameters that determine the diffusion of the epidemic and its economic impacts: the shares of high risk populations (IDUs and sex workers); the HITV/AIDS prevalence rates among these groups; the frequency and het- erogeneity of sexual interactions; the prevalence of STDs, the prevalence of condom use and needle sharing; and the distribution of AIDS-related deaths among unemployed, skilled, and unskilled workers, as well as the reduction in total factor productivity resulting from an increase in the HIV/AIDS prevalence rate (see the technical appendix for a description of the methods used to calibrate the model). As an illustration, figure 4.7 presents two EIV/AIDS diffusion pro- files for a typical MENA/EM country. The profiles depend on the preva- lence of condom use, the level of access to safe needles among IDUs, the presence of STDs, and the intensity and heterogeneity of sexual behav- iors. Clearly, there are high levels of uncertainty in terms of how the epi- demic is likely to diffuse. This level of uncertainty is captured in the case Assessing the Potential Economic Impact of HIV/AIDS in the MENA/EM Region 57 FIGURE 4.6 A Model of Growth to Evaluate the Macroeconomic Impacts of HIV/AIDS < ~~HIWAIDS| [ | ~diffusion model Productivity Population Quality- l Stock of produced |adjusted labor | capitall Savings GDP | *|< ~~~~~~~~~~Health |Consumption | Savings Poverty incidence Source: Authors' design. of Jordan in figure 4.8. Each bar in the figure is associated with a given prevalence rate. The height of the bar gives the number of scenarios that generate that particular prevalence rate. We observe that the prevalence rate could range anywhere from 0 to more than 20 percent. However, more likely, the prevalence rate in 2015 would be below 5 percent. As we will see, the challenge for policyinakers is to reduce this level of uncer- tainty and to insure against it. We need to make two important assumptions to assess the impact on the economy of the various paths that the epidemic can take. First, in terms of the distribution of HITV infections among unemployed, skilled, and unskilled workers, second regarding the link between HIV preva- 58 HIV/AIDS in the Middle East and North Africa FCiURE 6.7 M ustaxtion off DffffuMson :7 fl ffoir the H11VIODS IEpWdemlc Mn MiEWM/IEV1 Countires Profile 1: Low Risk Prevalence rate 1 = 100% 1.0 - Sex workers IDUs,female - IDUs,male 0.9 ------ Low-risk male * - -- Low-risk female 0.8 Total 0.7 0.6 0.5 0.4 0 2000 2004 2008 2012 2016 2020 2024 Year Profile 2: High Risk Prevalence rate 1 = 100% 1.0 Sex workers --- IDUs,female 0.9 ----- IDUs, male - Low-risk male Lo w-risk female 0.8 Totwal 0.7 / 0.6 / 0.4 I 0.3 0.2 I / 0.1 0 ;-. 2000 2004 2008 2012 2016 2020 2024 Year Note: We assume that only high-risk individuals are infected initially.ln Profile 1,we consider a 50 percent prevalence of condom use,only a 10 percent probability of sharing needles among IDUs, 1 percent STD prevalence, low sexual activity and low heterogeneity (as reflected by the parameters in table A.4 in thetechnical appendix). In this case,theepidemic is mostly confined to high-riskpopulation groups,particularly IDUs, where within a few months all the population is infected. In Profile 2, condom use drops to 10 percent,the STD prevalence increases to 5 per- cent, and the parameters defining sexual behaviors are multiplied by 1.5. Now the epidemic spreads to the general population and attains alarming levels among sex workers. Source: HIWAIDS diffusion model. Assessing the Potential Economic Impact of HIV/AIDS in the MENA/EM Region 59 FIGURE 4.8 Jordan: Plausible Predictions for HIV/AIDS Prevalence Frequency (percent) 40 36 32 28 24 20 16 1 2 8 0 L iLIHHnni LO l Iln n 0 5 1 0 1 5 20 25 30 35 HIV/AIDS prevalence in 2015 (percent) Note: Each bar is associated with a given HIV/AIDS prevalence rate.The height of the bar gives the number of scenarios that generate the given prevalence rate. Source: Authors' calculations. lence and total factor productivity growth. There are few data to cali- brate these parameters. Hence, as in the case of the parameters deter- mining diffusion scenarios, we consider a large number of combina- tions.8 For each combination of a diffusion profile and an impact scenario, we coinpute five output variables: the present value9 of total GDP produced between 2002 ancl 2025, the average growth rate of GDP for that period, the size of the population in 2025, and the FllV/AIDS prevalence level and total share of AIDS-related health ex- penditures in GDP in 2015. Given the lack of epidemiological data, there is a wide range of variation for the outcome variables. Descriptive statistics for these variables are presented in table A.7 in the technical ap- pendix. Here we discuss the main messages. In terms of the I-IIV/ATDS prevalence rate, we find several scenarios where it fluctuates around 4 percent by year 2015 (see technical appendix figures A.1-A.5). The fact that prevalence rates are similar across coun- tries reflects little variation in epideiiological initial conditions. The ex- ception is Djibouti, where the prevalence rate is already at 3 percent and could reach 15 percent by 2015 (see table 4.2). In general, once the 60 HIVAIDS in the Middle East and North Africa TABLE 4.2 Economic Impacts of the HIV/AIDS Epidemic (average across agencies) Present value losses Average reduction Population change HIV prevalence 2000-25 in GDP growth in 2025 in 2015 (percent GDP) (percent) (oercent) (percent)' Algeria 41.22 -0.40 4.07 4.46 Djibouti 150.77 -1.34 -16.68 15.88 Egypt, Arab Rep. of 51.33 -0.42 -3.84 4.23 Iran, Islamic Rep, of 38.65 -042 -3.85 4.18 Jordan 33.56 -0.35 -3.16 3.69 Lebanon 30.03 -0.45 -4.44 4.63 Morocco 39.48 -042 -3.97 4.27 Tunisia 54.04 -0.44 -4.21 443 Bahrain 35.58 -0.38 -3.92 4.20 Kuwait 35.70 -0.36 -3.78 4.11 Ornan 35.56 -0.35 -3.63 4.04 Qatar 33.19 -0.38 -4.00 4.27 Saudi Arabia 35.84 -0.31 -3.12 3.72 UAE 25.57 -0.32 -3.40 3.91 a. HIV/AIDS are not projections but simple averages across scenarios. Source: Authors' calculations. FIV/AIDS epidemic reaches 4 percent, its development becomes expo- nential. Hence, across scenarios, the accumulated economic costs for the period 2000-2 5 could approximate 30 to 50 percent of GDP (greater than 150 percent in the case of Djibouti). In all countries it is possible to find scenarios, although rare, where accumulated costs surpass current GDP. GDP losses reflect the slowdown in the GDP growth rate, which in most countries could approximate 0.4 percentage point per year (this is in line with the value found by other studies). Hence, the impacts on econornic growth can be substantial, even if most of the countries face high unem- ployment rates. Costs could be even higher than those reported here, if we had accounted for the impact of the epidemic on the accumulation of future human capital. Because the epidemic is still in its early stages, projected demographic impacts seem relatively small. For instance, in the case ofjordan and the Republic of Yemen, the size of the labor force would be reduced by 3 percent in 2025. In Algeria, the Islamic Republic of Iran, and Morocco, this number would be closer to 4 percent. This is equivalent to a reduc- tion in the average population growth rate of 0.1 to 0.3 percentage point per year.'0 Since the slowdown of the economies is also accompanied by a reduction in population growth rates, the overall impact in the growth rate of GDP per capita is less severe. In some extreme cases, this growth rate could even increase as a result of the epidemic. This would happen Assessing the Potential Economic Impact of HIV/AIDS in the MENA/EM Region 61 if the negative impact of HIV/AIDS on population growth rates was considerably higher than its impact on economic growth. Thus, caution is needed when using GDP per capita as an indicator of the economic impacts of the H-IV/AIDS epidemic.'t lMpacts on @eaM Expend(tnes Given a diffusion profile, the impact of the epidemic on health expendi- tures will depend on the average cost of treating an AIDS patient and the share of the population of patients who have access to treatment. Little information is available about these two factors in MENA/EM coun- tries, and, therefore, the calculation of potential changes in health ex- penditures is based on estimates from the literature. In terms of the av- erage cost of treatment, estimates from cross-country studies suggest a range of two to three times GDP per capita, excluding the cost of anti- retroviral drug (Cyrillo, Paulani, and Aguirre 2001; Floyd and Gilks 2001). In this analysis, we work under the conservative assumptions that the average yearly cost of treating an HIV/AIDS patient is equal to US$1,400 in a country with a GDP per capita of US$1,000, and that it increases by 0.95 percent for each 1 percent increase in GDP per capita. Access to treatment, however, varies widely across countries. The results summarized in this section are based on the assumption that 30 percent of those affected by AIDS would obtain treatment. The simulations confirm that I-IIV/AIDS could have a significant im- pact on health expenditures. Across countries, these expenditures could increase by 1.5 percent of GDP, even if only one-third of the patients have access to health services (see figure 4.9). In the case of Djibouti, health expenditures could attain 6.4 percent of GDP. Notice that these numbers are averages across simulations. However, it is possible that even in countries where current prevalence levels are below 1 percent, health expenditures could increase by up to 5 percent of GDP (see the teclnical appendix, table A.7). These results are consistent with estimates from the literature that suggest that the major economic consequences of the epidemic would be felt within the health system. For instance, it has been found in the case of India (Ellis, Alam, and Gupta 1997) that a severe AIDS epidemic would increase government health care expenditures by US$2 billion per year. Thus, according to the study, by 2010 health expenditures in India would be 60 percent higher than without HIV/AIDS. While the analysis developed in this section is highly simplified, it il- lustrates that the HIV/AIDS epidemic, if uncontrolled, could force im- portant budgetary reallocations. As discussed previously, by taking re- 62 HIV/AIDS in the Middle East and North Africa HDW/AOS-RelaWed Aveirage HeaM IExpendiItures in 20D5 (percentage of GDP) AIDS-related health exp. in 2015 (% GDP) 5.00 - 4.50 4.00 3.50 3.00 2.50 2.00 1.50 1.00 0.50 0 S Algeria Djibouti Egypt Iran Jordan Lebanon Morocco Tunisia Yemen Source: Authors' calculations. sources away from mnore productive economic activities, such as chil- dren's education, these reallocations can reduce aggregate economic ef- ficiency and put additional pressures on long-termn econonmic growth. The epidemic will have direct and indirect implications for poverty re- duction. First, the epidemic, which is likely to affect disproportionately more low-income individuals, will increase vulnerability and poverty. Second, by slowing down economic growth it will reduce the opportu- nities that the poor have to escape poverty. We discuss each of these ef- fects in turn. For thc poor and many households with consumption levels just above the poverty line, the main or only source of revenue is their labor. While the nonpoor can hedge their losses in wage income resulting from AIDS with other assets, coping mechanisms for the poor and vulnerable are more limited and usually involve changes in consumption patterns (reducing ed- ucation, food, and health expenditures) or sending children to work. These mechanisms result in human capital loss as a result of high child malnutri- tion or lower school enrollment rates, among others. Although informal copiig mechanisms to manage risks are diverse in MENAJEM coun- Assessing the Potential Economic Impact of HIV/AIDS in the MENA/EM Region 63 tries-ranging from family support and kinship ties to religious charitable organizations-research has shown that they are usually insufficient to hedge against adverse shocks (AWorld Bank 2001g). Studies show that re- ductions in consumption in low-income households following the death of an adult household member would reduce food expenditures by 32 percent and food consumption by 15 percent (Over and others 2001). This occurs not only as household income is lost and funeral expenditures need to be financed (on average, households spend 50 percent more, or US$800 to US$900, on ftuerals than they do on medical care), but also because households that experience a death cut back on the number of hours they work for wages (Biggs and Shah 1996). Hence, IHlV/A1DS can be the shock that drives many vulnerable households into poverty. In most MENA/EM countries, the poor already face problems of access to health services. As health systems become financially constrained, these problems can be exacerbated. At the same time, the poor are more exposed to infec- tious diseases, and, given complications from un-demutrition, are more vulnerable to the deterioration of their immune system. Even among poor households that are not directly affected by the epi- demic, the associated slowdown in economic growth will reduce their opportunities to escape out of poverty. Indeed, the evolution of the poverty prevalence of a country is determined by the growth rate of av- erage income and the change in income distribution. Research has shown that income distribution remains relatively unchanged over 10- year periods. Therefore, it is mainly economic growth that determines how many people are lifted out of poverty. In general, a 1 percent in- crease in per capita income can be associated with a 1 percent to 2 per- cent reduction in poverty prevalence (Dollar and Kraay 2001). As an il- lustration, if the growth rate of GDP per capita is reduced by 0.5 percent to 1 percent as a result of H.V/AIDS, the number of people who fail to escape poverty could reach 20 million by 2010. In 1998, close to 30 per- cent of people in MENA/EM countries (85 million) were living on less than US$2 per day. With an average growth rate of 3 percent per year, by 2010, poverty prevalence could be reduced to somewhere between 22 percent and 16 percent (depending on how sensitive the poverty preva- lence is to the growth rate of income per capita), and the number of poor could fall to somewhere between 79 million and 58 million. Because of I-fiV/AIDS, however, poverty prevalence would be higher in 2010 as the average growth rate of GDP per capita falls by 0.2 percenitage point to 1 percentage point (depending on the severity of the epidemic and its im- pact on labor productivity). The number of people who would have failed to escape poverty could then range between 8 million and 30 mil- lion, depending on how sensitive poverty prevalence is to economic growth (see figure 4.10).12 64 HIV/AIDS in the Middle East and North Africa Policy Implications The FIIV/AIDS epidemic in MENA/EM presents a typical problem of decisionmaking under conditions of high uncertainty. Prevalence levels could remain at low levels, but there is also a risk that the epidemic could develop through profiles similar to the ones discussed in the previous section, and the human and economic costs could be considerable. We show that waiting to intervene can be cosdy. Societies are better off if they invest today in interventions to reduce and mitigate the risk of an IITV/AIDS epidemic. Governments, whose mandate is to ensure the well-being of the pop- ulation, have the responsibility to develop and finance the implementa- tion of policies to confront HIfV/AIDS. Indeed, individuals alone could not devise appropriate mechanisms to contain the epidemic. First, indi- viduals do not take into account the social costs of the risks they take, or the social benefits of the preventive measures they adopt. In the absence of government intervention, we would observe an excess of risky behav- ior and too little prevention from a social point of view. Second, individ- uals suffer from information problems. Individuals may not have enough information about the risks of HIV and may lack knowledge and skills related to preventive behaviors. Finally, formal and informal institutions (culture or religious values) may constrain individuals' actions in ways that render them, and society, more vulnerable to HIV/AIDS. The role FIGURE 4.10 Number of People in the MENA/EM Region Who Would Remain below the Poverty Line as a Consequence of HIV Percent reduction in poverty prevalence from a 1% growth in GDP per capita 2- 1.5 ~~~~~~1 0 million-20 million i 1 million-10 million \ -0.50 -1.00 -1.50 -2.00 Reduction in GDP per capita growth rate (percent) Source: Authors'calculations. Assessing the Potential Economic Impact of HIV/AIDS in the MENA/EM Region 65 of governments is therefore critical in providing information, subsidiz- ing interventions to reduce risky behaviors, providing care and support for people living with HIV and AIDS (PLWiVHA), reducing stigma and discrimination, as well as creating the enabling environment. Governments can only intervene, however, if there are cost-effective interventions at their disposal (Kremer, 1996a, 1996b). Fortunately, in- ternational experience has demonstrated that there are cost-effective in- terventions to tackle HIV/A1DS. Recent studies and collective experi- ence from countries show that interventions can be highly cost-effective when focused on reducing risks (through information and preventive be- haviors and services) in those population groups most likely to contract and spread HIV Jenkins, Rahman, and others 2001; Kahn 1996). Inter- ventions such as reproductive health and HIV/AIDS education in schools, provision of basic prevention and care packages for highly vul- nerable groups (including STD treatment), and harm reduction for IDUs have also proved to be cost-effective. To implement these inter- ventions, several instruments are available, including direct provision of services, subsidies, taxes, and regulatory power. In general, early inter- ventions bring higher benefits and lower costs. In the case of MENA/EM countries, we simulate the impact of in- creasing condom use by 30 percent and expanding access to safe needles for IDUs by 20 percent (see the technical appendix, tables A.8 and A.9, for a description of the methods used in the simulation and the costs of the interventions). The policy is assumed to be implemented immedi- ately and is applied to each of the 100 scenarios discussed in the previous sections. The results show that these two interventions can considerably reduce GDP losses for the period 2002-25 (see figure 4.11). Across countries, savings (net of the costs of the interventions) could surpass 15 percent of today's GDP (see also tables A.10 and A.1 1 in the technical appendix). On average, this translates into an increase of 0.3 percentage point in the yearly GDP growth rate. We have also simulated imple- menting these two interventions with a delay of five years. Not surpris- ingly, waiting to intervene can cost countries an average of 8 percent of today's GDP (5.6 percent in tlhe case of Djibouti). In the previous simulations, we have applied the same policy to all countries in all scenarios and this is likely to underestimate net bene- fits. Indeed, the total amount of resources that societies ought to invest to fight HTV/AIDS and their allocation across alternative interventions depends on countries' characteristics. For instance, in the case of the Republic of Yemen, the simulated intervention may be too costly. In general, the costs and effectiveness of the different interventions are given by factors such as the level of development of the epidemic, so- cial and economic constraints on safe behavior, underlying patterns of 66 HIV/AIDS in the Middle East and North Africa FIGURE 4.11 Benefits from Expanding Access to Condoms and Safe Needles for IDUs and Costs of Delaying Action Gains from policy intervention (% today's GDP) 40 35 145 _~m 30 _ E * 25 I.E.E. Losses from Syears'delay ;, ^$ ,;> °20° > (%GDP) cr' *>S<>$<< 0§> q