China - Issues and options in health financing

Report No. 15278-CHA China: Issues and Options in Health Financing August 12, 1996 Chlin, in Mn( N'loli(i Dep,irtmenelt I Ii111,1n I)evelOpmrnnt Dep,ilrillent Document of the World Bank  Exchange Rate CuLrrency Unit = Yuan (Y) 1993 US$1.00 = Y 8.0 (commercial rate) January, 1996 US$ 1 .00= Y 8.3 (official rate) Note: Ulnless otherwise specified, all yuan values have been adjusted to 1 993 constant yuan. Acronyms CEM Country Economic Memorandumii (World Bank) CMS Rural Cooperative Medical System DALY Disability-adjusted Life Year DRG Diagnosis Related Grouips EPI Expanded Program on Immunization EPS Epidemic Preventionl Service FP Family Plannilig and Reproductive Healthi Program GBD Global Burden of Disease GIS Government Employees' Health Insurance System [also retirees, military, university studenits] HEI Health Economics Institute; sometimes called National Healthi Economics Researcih Institute HFA Health-for-All in the Year 2000 (Goal of the WHIO) IMR Infanit Mortality Rate LIS Labor Health Insurance System (covers employees of state-owned enterprises) MCH Maternal and Child Health Program MOA Ministry of Agriculture MOF Ministry of Finance MOH Ministry of Health NCD Nonconimunicable disease OECD Organization for Economic Cooperation and Development PRC People's Republic of Chinia SOE State-owned enterprise  SPC State Planning Commission SSB State Statistical Bureau TCM Traditional Chinese Medicine TFR Total fertility rate THC Township health centers (middle level of the rural three-tier health system, between village clinic and county hospital) U5MR Under-five mortality rate UNICEF United Nations Children's Fund WHO World Health Organization Glossary This glossary provides definitions of some of the key terms used in this report. Adverse Selection - The tendency of individuals expecting high health expenditures to purchase insurance or purchase a more generous insurance package than people expecting low levels of expenditures. Basic Benefits Package - A minimum set, or core, of health services. Capitation Payment - A fixed payment to a provider for each listed or enrolled person served per period of time. Payments will vary according to the number, age, and sex of patients enrolled but not witlh the number of services rendered per patient. Community Financing - A community-managed scheme whereby a local community group collects and manages funds from households, government, and local industries and organizes the delivery of a package of health benefits for community members. The health risks of community members are pooled, and the benefits include some catastrophic as well as basic care. Cost Containment - A set of steps to control or reduce inefficiencies in the consumption, allocation, or production of health care services whicih contribute to higher than necessary costs. Inefficiencies in consumption can occur when health services are inappropriately utilized; inefficiencies in allocation exist when a different mix of services could produce greater health benefits; and inefficiencies in production exist wlheni the cost of producing health services could be reduced by using a different combination of resources. Cost-sharing - A provision of health insurance or third-party payment that requires the individual who is covered to pay part of the cost of medical care received. This is distinct from the payment of a health insurance premium, contribution, or tax, which is paid whether medical care is received or not. Ccst-sharing may be in the form of deductibles, co-insurance, or co-payments. Disability-adjusted Life Year (DALY) - A unit used for measuring both the global burden of disease and the effectiveness of health interventions, as indicated by reductions in the disease burden. It is calculated as the present value of the future years of disability-free life that are lost as the result of the premature deaths or cases of disability occurring in a particular year. Diagnosis Related Groups (DRGs) - Groupings of diagnostic categories drawn from the Internrational Classification of Diseases and modified by the presence of a surgical procedure, patient age, presence or absence of significant comorbidities or complications, and other relevant criteria. . .  DRGs are the case-mix measure used in Medicare's prospective payment system in the United States. Modifications of DRGs have been adopted in several other countries as a tool for hospital management and/or reimbursement. Externalities - Costs or benefits arising from production or consumption that fall on individuals and groups not directly involved in the production or consumption concerned; and which are not compensated for by exchange. For example, immunization of an individual against an infectious disease can block the transmission of the disease to other individuals who are not directly involved and who pay nothing for the protection they receive. This creates an external benefit. And one person's inappropriate use of an antibiotic will frequently hasten spread of resistance to it, creating external costs borne by others. Fee-for-Service - Payments to a provider for each item or service rendered. Global Budget - An aggregate cash sum, fixed in advance, intended to cover the total cost of a service, usually for one year ahead. Health Insurance - Financial protection against the medical care costs arising from disease or accidental bodily injury. Such insurance usually covers all or part of the medical costs of treating the disease or injury. Insurance may be obtained on either an individual or group basis. Health Maintenance Organization (HMO) - An organization that accepts responsibility for organizing and providing a defined set of health services for its enrolled population, in exchange for a predetermined fixed, periodic payment for each person or family unit enrolled. The payment is fixed without regard to the amount of actual services provided to an individual enrollee. Health Planning - Planning concerned with improving health, whether undertaken comprehensively for a whole community or for a particular population, type of health service, institution, or health program. The components of health planning include: data assembly and analysis, goal determination, action recommendation, and implementation strategy. Managed Competition - Government regulation of health insurance and health care markets using competition as the means to achieve efficiency objectives within a framework of government intervention. Moral Hazard - In health insurance, this refers to the fact that people who are insured tend to act in ways that increase health expenditures. They may take fewer precautions against avoidable events, or, more importantly, patients and providers increase the use of services because insurance subsidizes prices at the point of service. Out-of-Pocket Payments - Payments borne directly by a patient without reimbursement by any insurance. They include cost-sharing. Prepayment - Usually refers to any payment to an organization for anticipated services (such as an expectant mother paying in advance for maternity care). Prepayment is distinguished from insurance because it involves payment to organizations which, unlike an insurance company, take responsibility for arranging and providing needed services as well as paying for them (such as health maintenance organizations, prepaid group practices, and medical foundations). Public Good - A good or service whose benefits may be provided to a group at no more cost than that required to provide it for one person. The benefits of the good are indivisible and individuals cannot be excluded. For example, a public health measure that eradicates smallpox protects all, not just those paying for the vaccination. iv  Public Health - The science dealing with the protection and improvement of the population's health by organized community effort. Public health activities are generally those that are less amenable to being undertaken by individuals or that are less effective when undertaken on an individual basis. They do not typically include direct personal health services. Public health activities include: immunizations; sanitation; preventive medicine, quarantine, and other disease control activities; occupational health and safety programs; assurance of the healthfulness of air, water, and food; health education; epidemiology, and others. Risk Pooling - The process of distributing the probability of financial loss across multiple parties. Health insurance, for example, is a way of distributing the uneven burden of high medical expenditures across a subgroup of the population. Salary Payment - Remuneration that is fixed per period of time and does not vary either with the number of individuals served or with the number of services rendered, although rate of salary change can depend on performance. Selection Bias - The tendency for multiple health plans or providers to attract, intentionally or unintentionally, an uneven (biased) distribution of health risks. Also known as risk selection. When done intentionally by insurance companies to discourage participants likely to need substantial care, the practice is sometimes known as "cherry-picking". Social Health Insurance - A term mainly used to denote compulsory, or public, health insurance, usually part of a social security system, which is funded from specific (mainly payroll) contributions and is managed by a government agency or autonomous organization such as a sickness fund, mutual aid society, or private insurer. Supplier Induced Demand - The ability of providers to use their authority to boost demand for their services, for the services of colleagues, or for the sale of drugs or tests in which they have a financial stake, above the level that would be demanded if patients had adequate and accurate information on medical benefits and costs. Third Party Payer - Any organization, public or private, that pays or insures health care expenses for beneficiaries at the time they are patients. The first party is the patient and the second party is the provider. Third parties may be private insurers; quasi-public bodies such as sickness funds; and government bodies themselves. Voluntary Health Insurance - Health insurance that is taken up and paid for at the discretion of individuals, or employers on behalf of individuals. Voluntary insurance can be offered by a private, public, or quasi-public body. Yellow Book - Generic term in China for hospital and health service price lists for the thousands of products and services sold. References OECD: The Reform of Health Care, A Comparative Analysis of Seven OECD Countries, pp. 9-11, Health Policy Studies No. 2, Paris, 1994. Glossary of Terms as Commonly Used in Health Care, Alpha Center, Washington DC. v   Contents Page Executive Summary xiii 1. Introduction: Health and Health Policy in China 1 2. Health Services And Their Financing 13 3. Strengthening Public Health Programs 25 4. Meeting the Needs of the Poor 33 5. Implementing Reforms in Pricing and Planning 41 6. Options for Efficient Risk Pooling in Rural Areas 49 7. Options for Efficient Risk Pooling in Urban Areas 59 8. Recommendations and Implications for Public Finance 67 Annexes 1. China's Health Expenditure and Health System Data 75 2. Outside Influences on Health Status 83 3. Deaths and Disease Burden in China 89 4. Child Mortality Trends in China 115 5. Health Indicators from Disease Surveillance Points System 121 Background Papers and Consultant Reports Prepared for this Study 127 Bibliography 129 vii  Tables in Text Table 1 Summary of Recommended Health Policy Actions, 1996-2001 Table 1.1 Trends in Under-Five Mortality Rates--China and Other Asian Countries, 1960-90 Table 1.2 Recommended Health Finance Policies: Impact on the Health Sector Table 2.1 Sources and Uses of Health Financing, 1993 Table 3.1 Government Finance of EPS and as a Share of GDP, Selected Years Table 3.2 Immunization Coverage in China, 1993 Table 3.3 Characteristics of Smokers in China, 1984 Table 3.4 Estimates of Tobacco Price Elasticity of Demand, Various Countries Table 4.1 Incidence of Absolute Poverty, 1978-90 Table 4.2 Health Status by Income, Rural China, 1993 Table 4.3 China's 592 Poor Counties: Comparison of Average Income, 1992-93 Table 4.4 Poverty-Related Disease Conditions, China, 1990 Table 5.1 Comparison of Production Costs and Fees for Selected Procedures in Shanghai Hospitals, 1989 Table 5.2 Comparison of Production Costs and Fees for Body CT Scan, 1988 Table 6.1 Prevalence and Benefits of Community Financing in Thirty Poor Counties, 1993 Table 6.2 Management of Community Financing, 1993 Table 6.3 Percentage of Community Financing by Source Table 6.4 Prevalence and Benefits of Community Financing in Five Provinces, 1991 Table 8.1 Summary of Major Health Care Issues in China Table 8.2 Public Finance Implications of Proposed Health Initiatives: Illustrative Costs Table 8.3 China's Government Expenditures in International Perspective Table 8.4 Revenue Implications of Recommended Tax Measures in the Year 2000 Table Al.l China: GDP, Nominal Exchange Rates and Price Deflators, 1978-1994 Table Al.2 GDP, Health Spending and Population Growth, 1978-93 Table A1.3 China: National Health Expenditures, 1978, 1986, and 1993 Table A1.4 China: Government Spending on Health, Selected Years, 1978-93 Table A1.5 China: Components of Recurrent Health Budget, Selected Years, 1978-93 Table Al.6 China: Health Insurance Coverage and Spending, Rural Population, 1993 Table Al.7 China: Revenues and Government Subsidies of Health Institutions Providing Hospital Services, 1993 Table Al.8 China: Average Number of Hospital Beds and Health Care Personnel, 1993 and 1994 Table A 1.9 China: Utilization of Hospital Beds at County Level and Above, Selected Years, 1985-94 Table A2.1 Income and Poverty in China, 1978-94 Table A2.2 Illiterates and Semi-literates, Selected Provinces, 1982 and 1990 Table A2.3 Access to Safe Drinking Water and Sanitation, China and Comparators, 1990 Table A2.4 Per Capita Nutrient Availability in China, 1950-95 Table A2.5 Key Demographic Indicators for China: 1970, 1995, and Projections to 2020 Table A3.1 Number of Deaths in China by Age, Sex, and Cause, 1990 Table A3.2 Number of Deaths in China by Age, Sex, and Cause, 1990, by Percent Table A3.3 Burden of Disease in China by Age, Sex, and Cause, 1990 (in DALYs) viii  Table A3.4 Burden of Disease in China by Age, Sex, and Cause, 1990 (in DALYs) by Percent Table A3.5 DALYs by Cause, as a Percentage of Totals, China and other Regions, 1990 Table A3.6 DALYs by Cause, as a Percentage of Totals, China and other Regions, 2020 Table A3.7 Deaths and Burden of Disease Attributable to Tobacco Use, China, 1990 and Projections to 2020 Table A4.1 Estimated Rates of Change in Under-Five Mortality, China, 1960-91 Table A4.2 Estimates of China's Under-Five Mortality, 1960-90 Table A4.3 Income Growth Rates, China and Other Countries, 1960-90 Table A5.1 Age-Specific Mortality Statistics, Urban and Rural Areas, 1990-94 Table A5.2 Male Age-Specific Mortality Statistics, Urban and Rural Areas, 1990-94 Table A5.3 Female Age-Specific Mortality Statistics, Urban and Rural Areas, 1990-94 Table A5.4 Male Age-Specific Mortality Statistics, Rural Areas, by Income Quartile, 1990-94 Table A5.5 Female Age-Specific Mortality Statistics, Rural Areas, by Income Quartile, 1990-94 Boxes in Text Box 2.1 Health Research and Development: A Neglected Component of the Chinese Health System? Box 3.1 The Challenge of STDs and HIV in China Box 5.1 The Leveraging Effect Box 5.2 The High Technology Equipment Race Box 5.3 Problems of Rural Health Delivery: A Case Study of Maternal and Child Health Programs Box 5.4 Regional Planning for Health Box 6.1 Estimating the Costs and Content of a Basic Benefit Package Box 7.1 Wage Tax Flows from Contributors to Individual and Common Accounts Box 7.2 The Three Tiers of Financing Health Services in the Jiujiang and Zhenjiang Experiments Box 8.1 Lessons from Reform of OECD Health Systems and Singapore Figures in Text Figure I Estimates of Under-Five Mortality Rates in China, 1960-91 and Projections to 1995 Figure 2 Insurance Status and Sources of Health Finance, 1993 Figure 1. I Deviation of China's Annual Rate of Decline in Under-Five Mortality from Rates Predicted by the Growth Rates in Income Figure 1.2 Distribution of Disease Burden by Broad Groups of Cause for China and Comparators, Estimates for 1990 and Projections to 2020 Figure 1.3 Insurance Status of China's Population, 1981 and 1993 Figure 1.4 Growth in Health Spending as a Share of GDP, Selected Countries, 1960-93 Figure 2.1 Growth in Health Expenditures in China by Source, 1978-93 Figure 2.2 Per Capita Annual Health Spending on Population Groups, 1993 Figure A2.1 Evolving Patterns of Age Distribution and Mortality in China, 1970, 1995, and 2020 Figure A4.1 Trends in China's Under-Five Mortality Rate, 1960-91 and Projections to 1995 ix  Figure A4.2 Relation Between Change in Child Mortality and Change in Per Capita Income, China and Other Countries, 1960-89 Figure A4.3 Deviation of China's Annual Rate of Decline in Under-Five Mortality Rates (from Rates Predicted by Growth Rates in Income) x  Acknowledgments This study was prepared by the World Bank at the request of the government of the People's Republic of China, which is considering policies to improve the equity and efficiency of health services. It is the third World Bank policy study of China's health sector. The series began with a general study entitled China. The Health Sector (1984), followed by China: Long Term Issues and Options in the Health Transition (1992). This report, China. Issues and Options in Health Financing (1996) looks in greater detail at the impact of China's move toward a market economy on the financing and organization of health care, and recommends a number of steps China can take to reform the financing of health care. Research for this report was carried out jointly by the Chinese government and World Bank staff. Work in China proceeded under the direction of Mr. Liu Peilong, Director of the Foreign Loan Office of the Ministry of Health. Two World Bank task managers in the Human Development Department managed the project: William P. McGreevey from June 1994 to February 1996 and Helen Saxenian from March to August 1996. The report was authored by Helen Saxenian, together with William Hsiao, Dean T. Jamison, William P. McGreevey, and Winnie Yip. Initial work at the World Bank was carried out under the direction of Mr. Vinay Bhargava, Chief, Human Resource Division, China and Mongolia Department, followed by Mr. Joseph Goldberg, Chief, Rural and Social Development Division. Mr. Jagadish Upadhyay, health group manager, and Dr. Janet Hohnen, public health specialist, managed the work on behalf of the Division. Mr. Nicholas Hope, Director, China and Mongolia Department, guided the preparation work and Mr. Richard Newfarmer, Lead Economist in the Department, and Mr. Michael Walton, Chief Economist, East Asia and Pacific Region, helped set the overall context for the report. William P. McGreevey and Helen Saxenian worked under the general direction of Dr. Richard Feachem, Senior Adviser, Human Development Department. World Bank staff in Beijing, including Mr. Pieter Bottelier, Resident Mission Director, Mr. Ramgopal Agarwala, Ms. Kathy Ogawa, and Mr. Zhao Hongwen, also provided assistance. China's Minister of Health, Professor Chen Minzhang, guided the early report preparation in October 1994. Ministry of Health staff, including Mr. Liu Xinming, Deputy Director, Planning and Finance Department, Professor Cai Renhua, Director, Legal Affairs Department, Ms. Liu Yingli and Mr. Liu Junguo of the Foreign Loan Office, and Mr. Fei Zhao Hui, now of the Ministry of Finance, provided considerable help during field visits. Mme. Sun, Ministry of Finance, and Mr. Ying Li, State Council, reviewed many phases of the study's work and provided advice throughout. The study draws extensively on 15 background papers prepared between November 1994 and September 1995 by leading specialists in China's health economics and finance. A list of the papers and their authors appears before the bibliography attached to this report. A related study of China's national health accounts was partially funded by the World Bank and a special grant from the Canadian government. Contributors to that study include Professor Peter Berman of Harvard University, Mr. Gilles Fortin of the Canadian Institute for Health Information, Mr. Vernon Hicks of Health Economics Consulting Services in Halifax, Nova Scotia, and Mr. J. Brad Schwartz, consultant, of Chapel Hill, North Carolina. An advisory group has been helpful throughout in defining the principal issues and approaches in this study. This group consists of Professor William Hsiao, the K. T. Li Professor of Health Economics at Harvard University; Professor Hu Shanlian of Shanghai Medical University and xi  Deputy Director, Health Economics Institute, Beijing; Professor Dean T. Jamnison, Director, Center for Pacific Rim Studies, UCLA; and Professor Wei Ying, Beijing Medical University and Director, Health Economics Institute. Peer reviewers in the World Bank include Mr. Willy de Geyndt, Mr. Charles Griffin, Mr. Jeffrey Hammer, Mr. Emmanuel Jimenez, and Mr. Nicholas Prescott. Members of the World Bank's Health Group in the Human Development Department provided valuable comments, including Mr. Howard Barnum, Dr. Denis Broun, Mr. Philip Musgrove, Dr. Mary Young, and Mr. George Schieber. Mr. Richard Bumgarner, Mr. Yuanli Liu, and Mr. Richard Peto also provided valuable input. Bank staff visited the provinces of Shanxi and Jiangsu in October 1994, and Sichuan, Hebei, Jiangxi, and Guizhou, in April 1995. Bank staff members and consultants who joined these missions and contributed to the report include Professor John S. Akin, Dr. Harry E. Cross, Mr. Jeffrey Hammer, Professor Winnie Yip, and Mr. Zhou Ji An. Ms. Helene Genest and Mr. Paul Hutchinson provided valuable assistance to the mission work. An initial draft of this report was discussed with the Chinese government in October, 1995. Follow-up work on national health accounts was done during a December, 1995 mission. The full report was discussed with Chinese government officials in a workshop from April 22 - 24, 1996, and this version incorporates their comments. This report was edited by Madelyn Ross. Mylene Domingo prepared the manuscript with the help of Susan Sebastian, Akosua Hudgens, Yvette Atkins, and Euna Osbourne. Jillian Cohen, Ellen Lukens, and other Health Group staff helped in checking sources and manuscript review. xii  Executive Summary BACKGROUND 1. Before 1949 China's population was among the least healthy in the world. This burden of disease was both a consequence and a cause of the nation's poor economic performance. Subsequent investments in improving health have contributed to China's high rate of economic growth, while at the same time directly improving human well-being, particularly among the poor. China's dramatic success in improving health conditions--life expectancy rose from less than 40 years in 1950 to 69 years by 1982--was accompanied by two related but less frequently noted achievements: * By 1975, insurance coverage and the rural Cooperative Medical System (CMS) reached close to 90 percent of the population--almost all the urban population and 85 percent of the rural. China's citizens thus had reasonable access to a broad range of cost-effective preventive and curative health services while sharing the risks of medically-caused financial misfortune. * The system for finance and delivery of health services contained costs. Health care costs were held by 1981 to just over 3 percent of GDP, despite the remarkable gains in health status and insurance coverage. 2. Beginning in 1978, the Chinese government introduced radically new economic policies that moved China away from a centrally planned economy and toward a competitive market system. This change in economic policies was accompanied by devolution of power from the central government to provincial governments. Many of these changes had profound repercussions for the health system. For example, the transition from agricultural collectives to the agricultural household responsibility system weakened the financial base of the Cooperative Medical System in rural areas. Some changes were also introduced to the health sector: the government gave health programs and facilities a great deal of financial independence and encouraged them to rely on user fees to support their operations. However, the government continues to administer many input and output prices in the health sector, the majority set below cost, and to control staffing in public facilities. 3. China's health sector is facing deep problems today as measured by four dimensions of health sector performance: health status, access to health care, efficiency, and total costs. Some of these problems are common to many countries. Others stem from, or are exacerbated by, the Chinese government's failure to reformulate health finance and to redefine its roles in health. China needs to act now to correct these problems, before they deepen and become more entrenched. Corrective action will require high-level and joint inter-ministerial action. Health is a sector that cannot simply be left to market forces. Problems in Recent Health Sector Performance 4. Health status. While China's overall health status is excellent compared to other countries of similar income level, recent trends in child mortality are a cause for concern. After falling steadily for 40 years, China's infant mortality and under-five mortality rates appear to have leveled off in the mid- 1980s. Figure I shows trends in under-five mortality in China for the period 1960-91, with projections to 1995. The graph indicates that the steady decline in the mortality rate from 1960 ended by 1985. From this point on, the under-five mortality rate apparently remained stagnant, or xiii  even rose slightly, to a level in 1991 of about 45 per 1000. (For comparison, the under-five mortality rate is about 22 in Sri Lanka, 36 in Thailand, 111 in Indonesia, and 130 in India.) Figure 1. Estimates of Under-Five Mortality Rates in China, 1960-1991 and Projections to 1995 200 Under-5 150 Mortality Rate 100 per 1,000 50 1960 1965 1970 1975 1980 1985 1990 1995 Year Source: See Annex 4. Data for 1992-95 are projections. 5. Mortality rates are, of course, determined by many factors, including income and education. China performed as well or better than other developing countries in terms of the rate of decline of under-five mortality relative to the rate of income growth before 1985. This positive performance relative to other countries sharply reversed during the period 1985-90, despite continued strong increases in China's real income per capita of 6 percent per year. However, even as per capita incomes rose overall, the incidence of rural poverty--which fell from 33 percent to 12 percent between 1978 and 1985--stopped declining in the second half of the 1980s and subsequent declines have been slower than during the early years of reform. It is unclear how much of the stagnation in child health from 1985 onwards is due to the decline in the availability of basic health services in rural areas following the dismantling of the CMS, to the persistence of rural poverty, or to other factors. What is clear is that China's performance in child health improvement since 1985 has been disappointing. 6. Access. Access to health care in China is inequitable, with especially deep divisions between the urban and rural population. For China as a whole, health spending per capita (public and private) was estimated at 110 yuan per year, or US$13.50 in 1993. (Note that health spending in purchasing power parity terms would be 4.8 times higher because of international price level differentials.) But the average health spending of 235 yuan per capita in urban areas was almost four times the average of 60 yuan per capita in rural areas. And the poorest quartile of the rural population accounted for only about 4 percent of all health spending in 1993. Only 10 percent of the rural population is insured compared to 50 percent of the urban population. While the two urban insurance systems--the Government Employees Health Insurance System (GIS) and the Labor Health Insurance System (LIS)--cover only 15 percent of China's total population, they absorb two-thirds of all public spending on health, and 36 percent of total health spending (Figure 2). 7. Health insurance coverage has declined rapidly in rural areas since the late 1970s due in part to new agricultural policies making farmers more autonomous economic agents and in part to government policies that discouraged reliance on the rural Cooperative Medical System (CMS). As xiv  communes gradually disappeared, so did the CMS that the communes had supported. Only about 10 percent of the rural population is now covered by some form of community-financed health care, down from a peak of about 85 percent in 1975. There is also much variation in coverage across provinces depending on interpretation of national policy at the provincial level. The decline of the CMS has led to the loss of prepayment or insurance coverage by perhaps 700 million rural Chinese, requiring them to pay out-of-pocket for virtually all health services. Without insurance, medical expenses can lead to deferral of care, untreated illness, financial catastrophe, and poverty. 8. Efficiency. Facility duplication and excess capacity from some vertical national programs result in inefficiency and waste. In urban areas, the overlap involves Ministry of Health facilities, state-owned enterprise facilities, and traditional Chinese medicine facilities. In rural areas, there is growing duplication and overlap of services between maternal-child health centers, family planning services, township health centers, and epidemic prevention stations. 9. Public spending on health is skewed towards hospitals, while priority public health programs are increasingly underfunded. And with fiscal decentralization, the poorest counties are least able to finance public health programs. As a result of funding difficulties, public health workers have been diverted from important public health work, such as immunizations and disease surveillance, to activities such as routine testing of water and food in urban areas, for which they can more easily charge fees. In many parts of the country, the Epidemic Prevention Service (EPS) is now even charging for immunizations and tuberculosis treatment. This has reduced coverage and--in the case of tuberculosis treatment--led to medically inappropriate but profitable patterns of care. 10. Prices of most health services and many inputs to the health sector are fixed well below costs under guidelines issued by the Price Commission. Health care providers inappropriately promote certain profitable items--especially pharmaceuticals and high-technology diagnostic tests--in order to generate profits and cross-subsidize under-priced products and services. This leads to misallocation of spending, medically inappropriate services, and upward pressure on overall health expenditures in both rural and urban areas. Given the incentive structure, it is not surprising to find that pharmaceuticals account for a remarkably high share of aggregate health spending--52 percent in 1993. 11. Total Costs. Total health expenditures per capita grew by 8 percent per year in real terms from 1978 to 1986, accelerating to 11 percent per year from 1986 to 1993. Health expenditures now account for about 3.8 percent of GDP. Over the same time period GDP per capita grew by 7.7 percent per year in real terms. Health spending will continue to grow in real terms as China's income grows, but this growth is likely to be accelerated by China's price distortions in the health sector and heavy reliance on fee-for-service provider payment methods--particularly fee-for-service under third party insurance systems like the GIS and LIS. Spending growth in the GIS and LIS systems is simply not sustainable. 12. The aging of the population will increase health care costs, because the elderly have higher health costs than the young. China's population aged 55 and over now makes up 13 percent of the population, and will reach 18 percent by the year 2010. While the overall aging of the population is inevitable, government policies can influence how efficiently the health care system is prepared to address the elderly's health care needs. Further, effective health promotion and disease prevention programs implemented now--particularly to control tobacco use--will improve the health outlook for China's elderly. xv  Figure 2. Insurance Status and Sources of Health Finance, 1993 Insurance Status by Population Subgroups, 1993 (as percent of total population) URBAN UNINSURED 15% GIS COVERED TOTAL (Government Insurance INSURED System) 2% 21% LIS COVERED RURAL UNINSURED (Labor Insurance 64% ~~~~~~~~~~~~System) 64% ~~~~~~~~~~~12% TOTAL RURAL INSURED UNINSURED 7% 79% Health Expenditures by Sources of Finance, 1993 (as percent of total spending) OTHER TTLPBI (public and private) TOTAL PUBLIC 6% GOVERNMENT 55% RURAL CMS BUDGET (largely public) % 2%_ GOVERNMENT INSURANCE SYSTEM RURAL OUT-OF- (GIS) POCKET 9% (private) 26% _ V TOTAL PRIVATE 45% LABOR INSURANCE URBAN OUT-OF- SYSTEM (LIS) POCKET 27% (private) 16% Total 1993 Health Expenditures = 132 billion yuan Per Capita 1993 Health Expenditure = 110 yuan (60 yuan in rural areas and 235 yuan in urban areas) Source: Wei 1995. xvi  13. Another cost to the economy of the present health financing and delivery system is the economic distortions that result from the current urban insurance systems. Because health coverage is tied to the employer--the government or state-owned enterprise--workers cannot retain their social benefits if they move from one job to another. Reforms, therefore, are needed so that workers can transfer jobs without jeopardizing their health (and pension) benefits. 14. In short, despite the remarkable early successes of China's health policy, achievements in child mortality, access, efficiency, and cost containment appear to have been either stagnating or eroding for over a decade. The Chinese government has reached consensus that these important concerns must be addressed by strong policy initiatives. The rest of this summary presents recommendations for dealing with these issues, following the same order as the chapters in the report. Finally, the implications of the recommendations in this report for public sector expenditures are discussed. REPORT RECOMMENDATIONS 1. Strengthen Public Health Programs (Chapter 3) 15. Beginning with the founding of the People's Republic in 1949, China has complemented the development of local health services with a series of strong national programs for high priority public health activities, including disease surveillance, environmental monitoring and improvement, mass immunization, and health education. The government also supported treatment of infectious diseases, implicitly recognizing the large social benefits from infectious disease control. This was an appropriate role for government, since most public health programs provide services for which there are large social benefits, but for which individuals are less willing to pay on their own. 16. Three interrelated problems increasingly limit the effectiveness, scope, and population coverage of China's national public health programs, however. The first is simply that budgetary pressures constrain the operation and efficiency of existing programs, even though resource requirements are modest in the context of overall health expenditures. Part of this is due to the fact that almost all spending on public health is from provincial and local governments, and the poorest areas--which experience the most severe public health problems--have the least capacity to finance these programs. The second problem, related to the first, is that cost recovery policies constrain the demand for certain public health services (e.g. immunization), particularly among the poor. Third, the general movement toward fee-for-service payment has diverted the energies of public health providers to those activities for which fees can most easily be charged, instead of the highest priority activities for public health. 17. China needs to return to policies that vigorously finance and support public health, recognizing that these services must be financed by the govemment if they are to be provided at socially optimal levels. Particular attention needs to be given to reaching the unregistered urban population with public health programs. This report recommends that the Epidemic Prevention Service's budget of 1.3 billion yuan in 1993 be increased to at least 6.5 billion by the year 2001 and that the EPS be prohibited from charging user fees for most of its services. The budgets of other agencies carrying out priority public health activities also need additional support. At the same time, the government must ensure that public health programs are implemented efficiently and that China's highly effective disease surveillance system is maintained and adapted to the changing pattern of disease burden. xvii  18. Nearly one million Chinese die each year of smoking-related diseases. Given the looming problem of a massive burden of costly illness and premature death from tobacco-related diseases (tobacco-related deaths are projected to increase to over two million per year by the year 2020), another recommended public health action would be to increase tobacco taxes substantially, accompanied by other measures to reduce smoking. This would help reduce eventual illness and death from smoking-related diseases and, if the incremental revenue were applied to public health (as in Australia), would help to finance public health programs adequately. 2. Ensure Essential Health Services for the Poor (Chapter 4) 19. The second priority for government health spending should be to ensure that the country's neediest citizens have access to priority health services. China's absolute poor reside almost entirely in rural areas. They are more likely to suffer from ill health, and their health problems can contribute to keeping them in poverty. Among the poorest quarter of the rural population, the infant mortality rate is 3.5 times greater than among city dwellers. The urban poor, especially unregistered migrants, also face relatively high health risks and need to be reached more effectively. There is strong justification on poverty assistance (or equity) grounds for government subsidies aimed at improving the poor's access to important health services. The Chinese government's role at present is minimal in this area. Resources need to be redirected or expanded to assure key health services for the poor. At the same time, given scarce public resources, subsidies need to be carefully targeted. 20. There are several ways to target the poor: - Geographical targeting to areas where the poor are concentrated. For example, poor administrative villages in China's 592 officially designated poor counties could be targeted for subsidized services. This population totals about 75 million. e Individual or household targeting, by identifying and certifying the poor for subsidized services. (China may be one of the only developing countries in the world where this targeting method is actually feasible, given good government records, but it entails heavy administrative costs); * Program targeting to health services that particularly benefit the poor in both rural and urban areas, such as deworming and management of acute infections in children. (This would require subsidies in addition to those for the public health programs discussed in the previous section). This report recommends phasing in a blend of geographic targeting (probably most practical at the village level) and universal finance of a limited number of services that particularly benefit the poor. It also recommends monitoring these approaches to guide policy improvements over time. 3. Reform Prices and Provider Payment Mechanisms (Chapters 5, 6 and 7) 21. Price distortions and irrational allocation of existing health resources have diminished the quality and effectiveness of China's existing health services. Prices for most medical services are set by the government at levels well below cost. To offset losses on basic services, the government has allowed high prices to be charged for drugs and high technology equipment. This results in a distorted pattern of services with overprovision of some services, such as CAT scans and ultrasounds, and underprovision of other services, especially those with a high labor content. xviii  Providers have strong incentives to overprescribe drugs of ail kinds, especially expensive drugs, in order to bring in additional income. Fee-for-service provider payment methods also encourage overprovision of services. Outpatients, for example, are frequently treated with intravenous drip solutions of glucose, vitamins, antibiotics, and other drugs. In almost all cases, this does not constitute justified medical practice. Given the incentive structure, it is not surprising to find that spending on drugs accounts for over half of all health spending in China, compared with 5 to 20 percent in OECD countries and 15 to 40 percent in most developing countries. Government action is needed for a variety of efforts to promote price and provider payment reform. These efforts will help to contain costs, reduce waste, and improve overall efficiency. 22. Price reform is urgently needed to bring administered prices in the health sector in line with marginal costs. A major study of pricing is needed to lay the basis for reform. Price and related provider payment reform would likely need to be implemented gradually over several years to be politically acceptable. Price reforrn in the health sector is an integral part of addressing an economy- wide problem of price distortions. 23. China also needs to move away from its dependence on fee-for-service provider reimbursement. Unconstrained fee-for-service reimbursement promotes the excessive use of such services because consumers rely on providers to recommend needed services, while providers, in turn, have a financial incentive to increase the volume of services. Experience from other countries shows that case-based reimbursements and various partial and full capitation payments can help contain costs and improve quality. Alternatively, fee-for-service payment methods can be used under a global budget constraint, or a mix of provider payment approaches can be used to improve incentive structures. 24. Options for provider payment reform, of course, hinge fundamentally on what types of risk- pooling arrangements are developed in rural and urban areas. One approach to provider payment would be to move, initially under urban insurance and community-financing schemes, from fee-for- service toward more aggregated products, as has already been tested in Zhenjiang, and finally to prepayment for a complete package of services. Under such a system, the provider would assume more risk and would have a decreasing incentive to over-provide services. 4. Control Investments and Improve Regional Planning (Chapter 5) 25. Governments can play an important role in cost containment through exerting oversight and controls on major human and capital investments in the health system. China is moving away from a centrally planned economy, but some oversight needs to be retained over new investments (across the many delivery systems in China) in hospital beds (especially at the tertiary level), expensive medical equipment, and the mix and numbers of health personnel. Experience from other countries shows that, once excess supply is created in any of these areas, the problem is politically difficult to correct. This excess capacity also drives up spending due to supplier-induced demand. 26. Government supply-side controls could be complemented by efforts to improve regional planning in health. China's health system is plagued by poorly coordinated vertical delivery systems in both rural and urban areas. Regional planning efforts should involve all relevant actors, including GIS, LIS, traditional Chinese medicine facilities, medical schools, and the MOH. As urban insurance centers are expanded, they would also become major stakeholders in regional planning efforts. China can build on experiences from planning exercises already underway in Baoji in Shaanxi, Jiujiang in Jiangxi, and Jinhua in Zhejiang provinces. xix  5. Promote Efficient Risk Pooling in Rural and Urban Areas (Chapters 6 and 7) 27. In the poorest countries of the world, more than half of all health spending comes from private sources, mainly in the form of out-of-pocket expenditures. Within these countries, the poor purchase most of their services themselves. As country incomes rise, there is a remarkably uniform trend around the world: the share of out-of-pocket health spending gradually falls as prepaid risk- pooling mechanisms develop. Without risk-pooling mechanisms in place, catastrophic health expenses are a major hardship for both the poor and nonpoor. Government support to encourage the development of risk-pooling mechanisms can make health services more accessible and efficient. Risk-pooling mechanisms can be financed by general tax revenue, social health insurance (mandated payroll taxes), private voluntary insurance, or community financing (a package of services funded from a combination of households, local communities, and the government, which are organized and managed by the local community). 28. By the 1970s, China had risk-pooling mechanisms in place for a remarkably high percentage of its population relative to its still very low income level. With the widespread dismantling of the CMS in the 1980s, however, China has become much more like the rest of the low-income world in this area: perhaps 700 million rural people in China have lost their access to prepaid care and are now paying out of pocket for almost all of their health care. As incomes rise, they are seeking to regain such coverage. 29. Urban populations still have relatively high health insurance coverage but the two urban insurance systems, the GIS and the LIS, are in urgent need of reform. The GIS now covers about 30 million people, including current government workers, government retirees, the military, and university students. It spends almost four times the average health spending in China--389 yuan per covered member, compared to the national average of 110 yuan. The LIS covers an estimated 140 million people who are employees and retirees of state-owned enterprises (SOEs). The LIS currently spends about 259 yuan per covered member. 30. Rapid cost escalation in the GIS and LIS has led to a fiscal crisis in both systems. Government spending on the GIS system grew by 15 percent a year in real terms from 1978 to 1993. In recent years, the government has had to allocate additional funds to cover deficits. In the LIS system, some enterprises have been unable to cover the health care bills of their employees and retirees. Since the systems are pay-as-you-go financed, state-owned enterprises (SOEs) and government units with large numbers of retirees have particularly high costs. 31. Both the GIS and LIS systems also have major inefficiencies. Except for dependents in the LIS system, enrollees do not make significant copayments and therefore have few financial constraints on their consumption of medical services. In addition, the care that is not provided directly is reimbursed on a fee-for-service basis, resulting in incentives for overprovision of services. Finally, and most importantly, health insurance coverage is tied to place of work and therefore impedes the labor mobility essential for a modem economy. For a Chinese worker the loss of a position in a state-owned enterprise means loss of health benefits, housing, and pension rights. 32. In both rural and urban areas, the govemment can play an important role by promoting an appropriate policy framework for the development of efficient and equitable risk-pooling, or health insurance, mechanisms. The government also needs to monitor and systematically evaluate new approaches to rural and urban insurance in order to adapt and improve the policy framework. xx  33. Rural reform options. Because of public finance constraints, a health system for the rural population financed largely from general revenue does not appear feasible over the medium term in China. (Government subsidies at present cover only a small fraction of publicly provided services: about 85 percent of costs are recouped from fees.) Mandatory wage taxes are also not feasible given that the majority of the rural population are self-employed. Community financing appears to be the most promising way to promote universal, or near-universal, health coverage in local communities and efficient service delivery, without being a major drain on government funds. China has considerable experience with community approaches to rural health insurance, beginning with the commune- based CMS system, and including a number of ongoing community-financing schemes that cover about 10 percent of the rural population today. A recent study of thirty poor Chinese counties showed that villages with community financing (compared to those without) are associated with higher use of health services at lower level facilities, lower rates of morbidity, lower fees for primary care services, and a lower proportion of income from drug sales in township health centers and county hospitals (see China Network 1996, Jin 1995a). To the extent possible, coverage should be universal at the local level. 34. China's experience suggests that, with appropriate government commitment, community financing is likely to be both administratively and financially feasible in many rural areas. Community financing has many advantages over the private, voluntary insurance that conceivably might develop in the richest rural areas. Private, voluntary insurance would exclude both the poor and individuals with high health risks and, if based primarily on fee-for-service payments, would also lead to cost escalation. And community financing has many advantages over the present system, in which rural residents must pay out-of-pocket for services on a fee-for-service basis. The government would need to monitor community-financing approaches in different areas, however, in order to evaluate their impact on health spending, efficiency, equity, and consumer satisfaction to inform medium-to long-term policy choices for rural health. Some of the key design features to be evaluated include the content of the basic benefits package, the size of the risk pool, different management models, reimbursement methods for individual doctors, township health centers, and county hospitals, the level of copayments by type of service and facility, and alternative measures to rationalize pharmaceutical use. 35. Urban reform options. Short- and-medium-term measures are urgently needed to reform the GIS and LIS insurance systems. * Management of insurance needs to be taken out of the SOEs and government units in order to achieve larger risk pools and economies of scale in administration. * Health insurance needs to be portable so that workers can move from one job to another without losing social benefits. * The system needs to move from a pay-as-you-go system to one in which contributions allow for expected expenditures in old age. (Or, alternatively, other mechanisms need to be developed to cover workers in old age). * Benefit packages need to be redesigned to be financially sustainable. * Provider payment needs to move from fee-for-service to mechanisms that help contain costs, such as capitated payment methods. xxi  36. The State Council has sponsored experimental health insurance centers that are testing key reforms of the GIS and LIS. Started in December, 1994, experiments in the cities of Jiujiang and Zhenjiang provide for wage-based enterprise/employee contributions into individual and pooled accounts managed by the insurance centers. The State Council decided to expand these experiments to fifty more cities throughout China in 1996. All of these experimental health insurance programs need to be carefully monitored and evaluated in order to inform policy-making. Government financing is needed for technical assistance, monitoring and evaluation, and strengthening regulatory capacity. Many issues still remain to be addressed, however, including how to pay for the unfunded health obligations of GIS and LIS workers (for example, for retirees of SOEs that will close) and what to do with SOE health facilities. 37. Over the medium- to long-term, urban health insurance coverage will also need to be broadened to include the increasing proportion of workers outside the state-owned sector and their dependents, as well as the rapidly growing unregistered or "floating" urban population. Urban insurance centers would, over time, need to harmonize benefits and contributions and portability across different areas of China. Implications for Public Finance (Chapter 8) 38. Because of the erosion in public health programs and the unmet demand for health services outside the insured population, and especially among the rural poor, the central government should consider a substantial increase in its health spending for the purposes discussed above. In terms of broad direction, this report argues that spending on strengthened national public health activities should increase as a share of GDP by 2001. Public spending is also needed for new programs to provide essential services for the poor and to promote risk pooling in rural and urban areas. 39. China can afford these priority programs. Moreover, anticipated improvements in China's tax revenue performance will make them even more affordable. Today, China's budgetary expenditures as a share of GDP are far below those of other countries. Government budgetary expenditures were 14.1 percent of GDP in 1994, and extra-budgetary government expenditures were an additional 3.8 percent. Central government expenditures as a share of all government expenditures are also unusually low--only 40 percent compared to an average of 78 percent in other developing countries. According to World Bank projections, China's GDP will double between 1993 and 2001, from about 3,450 billion yuan, to 7,500 billion yuan, in 1993 prices. The World Bank also projects that government revenues, as a share of GDP, will rise significantly during the Ninth Five-Year Plan (1996-2000) if China implements suggested changes in tax administration and tax structure. With these changes, tax revenues should increase by 6 percent of GDP. The priority health programs recommended in this report might cost on the order of 13 billion yuan by the year 2001, which would be less than one percent of projected government revenue. Most of this spending--90 percent--would be for public health activities and subsidized services for the poor. 40. One related recommendation involves increasing the tax on tobacco by 20 percent or more. A 20 percent increase in the tobacco tax is projected to generate 10 billion yuan in incremental revenue annually. By decreasing smoking, this tax would produce important health benefits as well as revenues. These incremental revenues might also be used to help finance the public spending increases recommended in this report. 41. Experience from the high-income countries suggests that China may now be at a critical juncture for redirecting its health policies. If China adopts the package of policy reforms xxii  recommended in this report, it could expect, within ten to fifteen years, to achieve much higher levels of prepaid health coverage, to eliminate most of the excessive disease burden among the poor, to have maintained general improvements in life expectancy, and to have stabilized health expenditures at 5 to 7 percent of GDP (just below the range for OECD countries). Failure to adopt these policies would risk leaving a large proportion of the population without health insurance, jeopardize health indicator improvements, and encourage growth in health expenditures to 10 percent of GDP or more (as in Argentina, France, and the U.S.A.). 42. The choices that China makes in its health financing policy in the coming years will not rest only on financial and economic analyses. These choices hinge fundamentally on judgments about what kind of society China wishes to be and what value it places on social cohesion, poverty alleviation, equity, consumer choice, and quality of care. This report argues that achieving these broader social goals can be combined, in the right policy environment, with measures that improve economic efficiency in the health sector. xxiii  Table 8.5. Summary of Recommended Health Policy Actions, 1996-2001 Objectives and Short-Term (1 to 2 Years) Medium-Term (3 to 5 Years) Programs 1. Strengthen public Provide full funding for the current EPS. Increase funding for public health to health programs Prohibit EPS from collecting user fees for about 6.5 billion yuan by 2001. (Chapter 3). public health services. Increase the tobacco tax by 20 percent or Strengthen anti-tobacco programs, such more to reduce smoking. Earmark the as health education, prohibition of tax revenue to fund public health smoking in public places, and regulations activities and health services for the poor. on levels of tar and nicotine in tobacco products. Upgrade the skills of staff in EPS and other agencies involved in public health Develop a strategic plan for public health, to carry out their new mandate. given China's changing pattern of disease burden and risk patterns, to guide the Ensure that priority public health work of EPS and other agencies carrying programs reach the poor in urban areas, out public health functions. particularly the unregistered urban poor. 2. Ensure essential Phase in a program of geographically- Continue to subsidize services for the health services for the targeted subsidies for health services for poor and expand coverage. poor (Chapter 4). the population in poor administrative villages in China's 592 officially- Monitor impact on the poor to guide designated poor counties. If the policy improvements over time. administrative village has a viable community-financing scheme, channel Over the long term, when more the subsidies through the scheme; comprehensive urban insurance systems otherwise direct them to health providers. are set up, the government might consider subsidizing the poor's contribution to Phase in program-targeted subsidies for a such schemes. limited number of health programs that particularly benefit the poor. 3. Reform prices and Carry out a major study on price reform, Gradually bring prices more in line with provider payment focusing both on prices of individual marginal costs, with periodic updates to mechanisms (Chapters health services and ways to bundle account for inflation. 5, 6 and 7). services into packaged fees (this relates to reforms of the GIS and LIS, below). Move away from the reliance on fee-for- Assist rural community financing plans in service provider payment methods. establishing provider payment These payment methods encourage mechanisms that encourage efficiency overprovision and drive up costs. (such as salaries with performance bonuses for village doctors, and Experiment with and evaluate alternative capitation payments to county hospitals). provider payment methods that encourage efficiency in urban insurance experiments. xxiv  Objectives and Short-Term (I to 2 Years) Medium-Term (3 to 5 Years) Programs 4. Control investments Improve and expand regional planning Institute nationwide regional planning and improve regional techniques already developed under guidelines. planning (Chapter 5). World Bank Health Loan III to better integrate various hospital systems. Develop government oversight mechanisms to avoid creating Change government's role in manpower oversupplies of physicians, particularly from central management of staffing of specialist physicians, tertiary hospital individual health facilities to more macro beds, and expensive diagnostic manpower planning nation-wide and equipment. province-wide to support the broad health reforms proposed here. 5. Promote efficient Rural areas: risk pooling in rural and urban areas Develop national guidelines for the Provide training, technical assistance, and (Chapters 6 and 7). operation of community financing that modest central government subsidies to address the size of the risk-pooling base, rural communities that establish management structure, and community community health financing. control mechanisms. Monitor impact of alternative forms of Organize provincial technical assistance community financing on health status, teams to assist community programs. access, organization and delivery of services, pharmaceutical use, health Implement pilot programs. spending, etc. Urban areas: Extend urban insurance experiments to Gradually extend social health insurance other provinces. Systematically evaluate to cover workers employed by joint these additional experiments. Some of ventures, smaller collective industries, the start-up costs of new schemes (such and private enterprises. Mandate open as information systems, capital access on a voluntary basis to all urban equipment, training) might be financed residents not covered by their employers. with government assistance. The government could eventually finance participation by the indigent in insurance. Experiment with risk pooling in larger areas such as the provincial level. Reform provider payment methods. Insurance Fund Centers could move toward negotiating capitation contracts with providers covering all levels of services, under risk-adjusted capitation rates. Alternatively, they could move towards DRG-type payments for hospital services, based on reasonable cost, with periodic adjustments as needed. xxv   1. INTRODUCTION: HEALTH AND HEALTH POLICY IN CHINA 1.1 A nation's health policies affect directly both the health of the population and the operation of the health care system. A growing body of evidence and analysis from around the world allows individual countries to learn from the policies of others. This chapter sets the analytical framework for evaluating health finance policies in China by assessing the performance of China's health sector and examining health policy issues raised by this performance. 1.2 To assess China's health sector, four broad measures of performance are used in this chapter. The first measure is the health status of China's population. Overall China's health status has improved enormously since 1949. A good measure of this is life expectancy: A typical country of China's income level in 1990 achieved a life expectancy of about 64 years, whereas China achieved 69 years. As will be discussed in detail below, however, other important health status indicators in China appear to have stagnated since the early 1980s. The next three performance measures relate to the financing and provision of health services: access to health services, efficiency, and total costs. Ensuring broad access to health services helps ensure that health care is delivered equitably. This is a desirable goal best achieved through risk-sharing (insurance) mechanisms that provide prepaid coverage for a reasonable range of services. The percentage of a population covered by such risk-sharing mechanisms is therefore an important indicator of a health system's performance. Health policies also affect the efficiency of care--in other words, whether services are produced at the least possible cost and whether spending is efficiently allocated (producing value for money). Finally the total cost of the health resources used by a nation is another important policy outcome. Experience around the world suggests that spending more on health is not required for improving health outcomes, since some high spending countries (such as the United States) get relatively little return for their resource commitments. Another dimension of costs is indirect economic losses that result from the health care system. For example, labor mobility may be impeded by insurance systems that tie the possibility of health insurance coverage just to certain employers. This report deals with policy issues and alternatives in all four of these areas. 1.3 It should first be noted, however, that health policy choices are not the only influences on a nation's collective health. Outside factors also have a major impact on a nation's health status and the operation of its health system. Income and education levels, for example, form the foundation upon which a nation's health policies are built. As income increases, so does the ability to acquire the necessities for good health--adequate food intake, clean water and sanitation, satisfactory shelter, and access to health services. Similarly, as education levels rise, so does the capacity of the population to make informed choices concerning health, disposition of income, and personal behavior. 1.4 Demographic changes are another important outside influence on the health sector. China's demographic patterns have changed dramatically over the past four decades. Rapid declines in fertility and mortality, the aging of the population, and the potential for explosive urban growth affect both health conditions and planning for the evolution of the health I  system and its finance. Annex 2 examines these and other outside influences on health status in greater detail. Health Status: Progress and Problems 1.5 In 1984 the World Bank's first health sector report on China (World Bank 1984) called for completing the first Chinese health care revolution: extending successful programs for improving child health and control of endemic infections into poor rural areas; consolidating and deepening the health gains achieved in most of rural China by reversing the ongoing breakdown in cooperative medical services; and seeking new means to finance public health programs that were being neglected by local providers embracing fee-for- service practices. The report's other concern was to encourage a second revolution: preventing and managing the emerging epidemic of noncommunicable disease (NCD) at much lower cost than in the high-income countries, which preceded China in emergence of NCDs as dominant problems. The World Bank's second health sector report (World Bank 1992a) dealt much more extensively with problems of NCDs and their risk factors. This section reviews the status of these two health issues in China today. Child Health and Infectious Disease 1.6 Despite rapid income growth in the past decade, China's earlier progress in improving child health appears, in the aggregate, to have come to a stop. Analyses from inside and outside China suggest that the infant mortality rate (IMR) stopped declining in about 1982 (World Bank 1992a, pp 6-7). A more recent overview (Parker n.d.) reaches a similar conclusion, pointing to a decline in immunization coverage beginning about 1988 and to recent unexpected outbreaks of immunizable diseases. Parker also presents evidence from surveys in nine provinces that a key indicator of child malnutrition--the percentage of children with very low height for their age--increased in rural areas between 1987 and 1992. Urban malnutrition, by contrast, declined sharply in those provinces during this period. These overall findings are even stronger in poor areas. The Study of Thirty Poor Counties found that the median infant mortality rate increased in those counties from about 50 to 72 deaths per 1,000 live births from the late 1970s to the late 1980s (China Network 1996 and Liu Yuanli and others, 1996). 1.7 Recently available census and survey data now permit a careful reassessment of trends in child mortality. This report commissioned a complete analysis of available data on national trends in the under-five mortality rate (U5MR). While this assessment could also have been undertaken for the infant mortality rate, demographers have concluded that under- five mortality estimates are consistently more robust and reliable than those for the IMR. UNICEF regards U5MR as the best indicator of social development because it accounts for the mother's health and level of knowledge, immunization levels, use of appropriate health services, access to water supplies, sanitation conditions, and the overall safety of the child's environment (UNICEF 1989, p. 82). Table 1. I compares China's results to those of other Asian countries; Annex 4 provides more detail on methods and results. The analysis concludes that U5MR in China declined steadily until the early 1980s and then began a slight upward drift. 1.8 Experience from other countries suggests that the under-five mortality rate need not plateau as China's has. Sri Lanka's per capita income is slightly higher than China's and its 2  Table 1.1. Trends in Under-Five Mortality Rates-China and Other Asian Countries, 1960-90 Year Under-Five Mortality Rate China Hongkong India Indonesia Japan Sri Lanka Vietnam 1960 173 53 235 214 37 140 105 1965 144 1970 115 1975 85 17 195 151 11 69 68 1980 60 1985 44 1990 44.5 7 127 111 6 22 46 Sources: China -- Annex 4; Other Countries -- World Bank, World Development Report 1993. Note: The under-S mortality rate is defined as the number of children who die between birth and their fifth birthday, expressed here as the number per thousand live births. 1975 under-five mortality rate was moderately lower, but by 1990 its under-five mortality rate had dropped to half that of China's. Japan's IMR in 1951 was about the same as China's was in 1976, but it then dropped by a third in six years. Indeed, until 1951, the decline in Japan's IMR was remarkably similar to that which occurred in China 25 years later; Japan's IMR, however, continued to decline while China's decline soon stopped (Parker, n.d.). 1.9 This suggests that China's performance has deteriorated not only in absolute terms but also relative to other countries. To address this question, data from China and more than 80 other countries were analyzed to assess the relationship between changes in under-five mortality and changes in income. The actual rate of decline in China's U5MR was then compared to that which would be predicted based on the experience of the sample of countries. Figure 1.1 plots the results for each of the time periods into which the data were divided. China's best relative performance was in the period 1975-79; the under-five mortality rate declined a full 3 percent per year more quickly in China than would have been predicted. Within a decade, in 1985-89, the position reversed and China's U5MR was changing 3 percent per year less favorably than predicted. (Indeed China's U5MR was by then actually increasing slowly.) 1.10 What might explain this major and disturbing change? The large decline in China's performance in child health relative to other countries suggests that multiple factors are at work, including at least the following three possibilities: * First, China's best performing period corresponds reasonably closely with China's period of rapid fertility decline, which stopped by the early 1980s. To the extent that the total fertility rate (TFR) decline reduced the number of closely spaced births (with their known risks for child health and survival), it would have contributed to China's good performance in the 1970s and early 1980s. The birth spacing effect, however, may be partially counterbalanced by the greater risk associated with first births. 3  Figure 1.1: Deviation of China's Annual Rate of Decline in Under-Five Mortality Rates from Rates Predicted by Growth Rates in Income 0.060 0.050 0.040 t 0.030 Deviation 0.020 -1 from 0.010 predicted 1 I 0.0 1960-64 16-69 1980-84 1985-89 -0.020 197 7 -0.030-1907 _ 0.040 ~~~~~~1975-79 -0.040 Source: Adapted from Annex 4. Note: Negative deviations indicate that China's under-five mortality rates were declining more rapidly than would be predicted--in other words, China was doing well. Positive deviations (e.g. 1985-89) point to poor performance--the decline was slower than would be expected. * Second, China's success in the period 1978-85 in reducing the percentage of the population in absolute poverty (from 33 percent to about 12 percent) would have helped improve health conditions overall. The percentage in absolute poverty, however, decreased no further from 1985 to the early 1 990s,. Although changes in the rate of poverty reduction almost certainly contributed, China's improvements in relative U5MR performance began earlier and started to decline earlier than the changes in poverty. Changes in poverty rates probably contributed to, rather than drove, China's relative performance in U5MR. * Finally, China's child health performance over this entire period tracks closely with the rise and fall in the percentage of the rural population covered by the rural Cooperative Medical System (CMS). This system reached its peak coverage of about 85 percent in 1975, and declined rapidly after 1979 (see Chapter 6). The CMS did a good job of providing cost-effective interventions to address major causes of child mortality. Both statistical information (e.g. on immunization coverage) and anecdotal accounts suggest that China's public health function has suffered widely with the decline of the CMS, and its rise and fall probably account for some of the change in China's relative performance in reducing the U5MR. 4  1.11 The relative importance of these various causes will be clearer when analyses can be undertaken on a province-by-province level. 1 The deterioration of China's under-five mortality rate relative to the rest of the world suggests the need for serious reconsideration of how health policies can be appropriately adapted to market-oriented reforms in the economy. This report explores options for beginning that process. Noncommunicable Disease and Injury 1.12 The young and the old are afflicted by very different health problems. Stroke, cancer, ischemic heart disease, and chronic lung disease account for most of mortality in people of late-middle and older ages. Children, on the other hand, die from a relatively short list of infections, most of which are inexpensive to treat or prevent. Figure 1.2 illustrates the increasing importance of noncommunicable diseases in China and other countries, with projections to 2020. As noncommunicable diseases increase their share of the disease burden in China, the contribution of injury is expected to change very little, while infectious diseases continue their steady decline in relative importance. Annex 3 provides more detail. 1.13 Much of the projected increase in the relative importance of noncommunicable disease results from unalterable demographic changes. Part, however, comes from projected patterns of disease incidence that are substantially higher than they would be if patterns of risk were better controlled. Tobacco use provides the most important example. By 1990 tobacco use already accounted for about 800,000 of China's 8.9 million annual deaths; projections of the effects of past and future tobacco use indicate that over 2 million deaths related to tobacco use are likely to occur in 2020. Tobacco-attributable deaths will have almost tripled between 1990 and 2020 and their percentage of all deaths almost doubled unless tobacco use can be curtailed. (See Annex 3, Table A3.7). 1.14 It is possible to devote enormous resources to interventions for management of NCDs that are costly and of limited efficacy. More cost-effective approaches to the prevention and treatment of NCDs may be neglected unless the incentive environment encourages experimentation with and adoption of these approaches. This report points to options for improving China's approach to decisions about NCD prevention and management. Performance of the Health Care System 1.15 A nation's health policies have a direct impact on the operation of the health care system as well as on the health of the population. This section highlights trends and policy issues associated with these non-health outcomes of health policy--access, efficiency, and costs of China's health care. I China's Disease Surveillance Points data also provide information on age-specific mortality rates. These data are presented in Annex 5. The data, unfortunately, are only truly comparable for the period 1990-94 as the sampling base was changing in earlier years. Over this short period, no clear trend emerges in life expectancy, the probability of dying before age one, the child mortality rate, or the adult mortality rate. 5  Figure 1.2. Distribution of Disease Burden by Broad Groups of Cause for China and Comparators, Estimates for 1990 and Projections to 2020 1990 2020 1 24g /ii CHINA II 68%79 18% 2 9 k 29% 24 INDIA 56% , / / 15% s % 7% lil 5% 1:2% /~ R10% HIGH-INCOME COUNTRIES -i (Europe, US, 81% 85% Japan) 1: Communicable, Maternal, Perinatal, and Nutritional Conditions a II: Noncommunicable Diseases 0111: Injuries Source: Murray and Lopez (1996); See Annex 3. 6  Issues of Access and Equity 1.16 The Chinese public has relatively good physical access to basic health care services. High population densities and a well developed health infrastructure make geographical barriers relatively modest in China, except for a significant minority living in mountainous or remote rural areas. Until recently, the government also made cost-effective public health services widely available, thus minimizing financial barriers to public health services. Finally, the cost of routine, basic outpatient health services is low enough that most nonpoor Chinese households are able to pay for such services out of their own current income or savings. 1.17 Catastrophic care poses more of a problem, however, because it involves services that are so expensive, relative to household income, that many households must either forego treatment or go deeply into debt to pay for them. As much as 70 percent of total health expenditures in many countries are for catastrophic care. An important performance indicator for a nation's health system, then, is how well it solves this problem of efficiently arranging for these services to be financed and provided. Given limited resources, what can be included under "catastrophic care" coverage will vary with a country's income level and health infrastructure. 1.18 Since only a fraction of the population needs catastrophic health care in any given year, pooling of risks provides the best mechanism for financing these services--either under a system financed by general tax revenue, social insurance, or private voluntary insurance. This is an issue both of equity--the poor will require subsidies--and of efficiency, in that all but the very wealthy (who can be self-insured) will generally benefit from risk pooling. Low-income countries typically lack the institutional and financial capacity to offer most citizens the advantages of risk-pooling arrangements. As development proceeds, however, risk pooling typically benefits an increasing percentage of the population--often, at least initially, through straightforward government or collective finance of clinics and hospitals open to everyone. 1.19 From the late 1960s through the early 1980s, China provided an exceptionally high percentage of its population relative to its income level with some form of risk pooling that guaranteed access to care. Figure 1.3 illustrates that only 29 percent of China's population in 1981 lacked access (i.e. were uninsured) but that by 1993 the uninsured proportion had grown to almost 80 percent. This change occurred largely as a result of fundamental economic changes in China's economy. The challenge is to reintroduce broad access to health care in the context of China's new economic environment. Efficiency 1.20 The fact that total health expenditures have been growing rapidly in China, even as some key indicators of health status have stagnated or even begun to decline, suggests a mounting problem of inefficiency. The allocation of public spending favors less cost- effective hospital services over highly cost-effective public health activities. Distorted prices encourage the overuse of drugs and high technology tests. Fee-for-service payment encourages overprovision. Multiple vertical health delivery systems have led to excess capacity and waste. Reversal of these trends will require reallocation of resources, both 7  Figure 1.3. Insurance Status of China's Population, 1981 and 1993 1981 Government Employees' State-Owned Enterprise Insurance (GIS) Employees' Insurance TOTAL 2% (LIS) TOTAL ~~~~~~~~~12% UNINSURED 29% Collective Industry Insurance Uninsured 5 29% Commune Industry Insurance 4% TOTAL Rural Insured INSURED 48% 71% 1993 Government Employees' Insurance (GIS) State-Owned Enterprise TOTAL 2% Employees'lnsurance INSURED Urban Uninsured (LIS) 21% 15% 12% Rural Insured 7% Rural Uninsured t ;;1. 64% , 5 ! i,. if ' X ' 2 )~~M i. ~i X TOTAL UNINSURED 79% Source: World Bank 1984 and Wei 1996. 8  public and private. Risk-pooling and provider compensation arrangements need to be designed to contain costs, extend access, and promote attention to value-for-money and quality of care. Cost of Health Services 1.21 The cost to the economy of operating the health system is another important performance indicator. Demographic change and economic growth virtually ensure that health expenditures in China will grow as a percentage of GDP. Policy choices can, however, influence whether that expenditure growth is excessive, and whether it efficiently expands access and improves health outcomes. 1.22 Figure 1.4 shows that some countries have controlled health care costs much better than others. In 1960, Canada, Japan, the United Kingdom, and the United States were all spending 3-5 percent of national income on health, a share similar to what China spends today. But spending rates diverged sharply over the next 30 years, with health spending reaching 14 percent of GDP in the United States by 1993--even though 15 percent of the population still is not insured. By contrast, Japan managed much more successfully to contain spending. In 1993 Japan spent only 7.3 percent of GDP on health, with nearly universal coverage and the world's highest life expectancy. Figure 1.4 Growth in Health Spending as a Share of GDP (in percent), Selected Countries, 1960-93 16 14- g' 12 - 2L910 v 0 g! 8 s 6 4 2 0 1960 1970 1980 1990 Year Source: OECD Health Data 1995. 9  1.23 Perhaps the most important factor outside the health sector that leads to rising costs is the aging of the population and accompanying epidemiologic changes. The World Bank has concluded that the agingfactor alone would increase health expenditures from about 3.2 percent of GDP in 1992 to 5 percent in 2010 and 7 percent by 2030 (World Bank 1992a). 1.24 Factors within the health sector influence costs even more, however. China's health expenditures grew rapidly between 1978 and 1993--averaging 10.9 percent per year over the full 15-year period. Factors placing upward pressure on expenditures include a worldwide tendency for the utilization of health services to rise faster than income, and the transition from salaried compensation of providers to fee-for-service compensation. Distorted prices and high profitability from drugs and high technology diagnostic tests (discussed in detail in Chapter 5) also put upward pressure on health care costs in China. 1.25 During the same 15-year period, risk-pooling arrangements in China underwent significant changes as shown in Figure 1.3. Risk pooling in rural areas declined sharply after 1978, from 48 percent in 1981 to 7 percent in 1993. Under these conditions, the willingness and ability of patients to pay personally for services limits providers' capacity to overprescribe or even to provide genuinely necessary services. No such constraint operates under existing urban insurance programs such as the Government Health Insurance System, or GIS. Under the GIS, a "third party" (the government) reimburses providers for essentially all the patient procedures that GIS agrees with. Consequences of these different rural and urban incentive regimes were clear: Between 1981 and 1993 per capita expenditures in rural areas increased from 21 yuan per capita to 60 while GIS expenditures increased much more rapidly, from 96 yuan per capita to 389. As this report will discuss, there are many ways to achieve the efficiency gains of wide risk pooling without creating incentives that lead to excessive cost escalation. The GIS, however, embodies incentive arrangements of the type that have led to excessive growth in costs in other countries--such as Korea, Singapore, and the United States. And, as incomes grow in rural China, the demand for risk pooling and prepaid care will grow apace. Reliance on out-of-pocket financing to keep costs down is not only undesirable because of the efficiency losses but also impractical in the face of probable demand for prepaid arrangements. 1.26 Over the medium term, health-financing policies also need to minimize indirect costs to the economy (principally distortions in the labor market) by, for example, separating the provision of health services from employment. Conclusion 1.27 China faces some disturbing trends and challenging policy issues in all four of the health sector performance areas reviewed in this chapter. There is stagnation (and perhaps even deterioration) in the child mortality rate. Insurance coverage and access to care have markedly declined. At the same time, real expenditures per capita have increased by a factor of more than 2.5 in rural areas and by as much as 4 times in urban areas since 1978. Incentives for preventive care have eroded while incentives (and opportunities) for providing excessive or inappropriate care have multiplied. And initiatives to undertake the collective action required for efficient risk pooling have not been widely or consistently implemented. 1.28 Clearly, major economic reforms begun in China in the late 1970s have brought rapid economic growth, but they have also had unintended and sometimes detrimental 10  consequences for the health sector. This report lays out options for adapting health policy to China's new macroeconomic environment. The following chapters recommend specific policy measures to deal with each of these challenges. Table 1.2 provides a preview of these measures, and shows the impact that each would have on the health sector issues outlined above. 11  Table 1.2 Recommended Health Finance Policies: Impact on the Health Sector POTENTIAL IMPACT ON HEALTH SECTOR PERFORMANCE REPORT RECOMMENDATIONS Promoting Improving Access and Improving Containing Health Equity Efficiency Costs Status Fully Fund Key Public Health ff f f t Programs Subsidize Essential Health Services ft ft t for the Poor Reform Prices and Provider Payment ft t Mechanisms Control Investments in Capital and ft ft Manpower and Improve Regional Planning Promote Universal Risk Pooling ff W2 = no significant impact E = moderate impact ft =strong impact 12  2. HEALTH SERVICES AND THEIR FINANCING 2.1 This chapter describes China's current health delivery system: infrastructure, personnel, and the mechanisms used to finance services. It updates earlier World Bank health sector reports on China done in 1984 and 1992 (World Bank 1984 and 1992a) and sets the stage for the analysis and recommendations that appear in subsequent chapters of this report. Infrastructure and Personnel 2.2 To provide health services to its population of 1.2 billion, China has some 200,000 health establishments, and a wide array of supporting research organizations. The country has 5.3 million health professionals, who make up about 0.8 percent of the total labor force. This includes 1.9 million doctors, or about 1.6 doctors per thousand people, and I million nurses. Doctors are trained in one of three categories: junior doctors (19 percent of the total); senior doctors (62 percent); and doctors oftraditional Chinese medicine (19 percent). Senior doctors are concentrated in medium- and large-sized cities. Village doctors (also known as Barefoot Doctors) with much less training are excluded from these estimates and categories. 2.3 China has some 3 million hospital beds in operation, averaging 2.4 beds per thousand people. This is a higher per capita ratio than is found in other low-income regions of the world, including Africa and the rest of Asia, and nearly as high as in Latin America and the Caribbean region. Only the OECD and Eastern European countries have higher bed-per-capita ratios. In 1994, China's bed-occupancy rate was 69 percent, and the average length of hospital stay was 15 days (State Statistical Bureau 1994; World Bank 1993a; Ministry of Health 1993a.). 2.4 Three-quarters of China's health spending goes to pay for inpatient or outpatient hospital care (Table 2.1). About 60 percent of hospital spending is on pharmaceuticals. Most hospital facilities are part of the Ministry of Health (MOH) system, including its provincial and county affiliates, or are operated by state-owned enterprises (SOEs). Others are run by village and township collectives. China also has an estimated 161,000 clinics operated by private practitioners in urban areas, although there is no official estimate of the numbers of patients they serve. 2.5 More than half of China's health workers are employees of the Ministry of Health or its provincial Health Bureaus. At the end of 1993, the Ministry of Health and its provincial affiliates employed 1.7 million hospital workers, 1. I million health workers in township health centers (THCs), 250,000 workers in the Epidemic Prevention Service (EPS), and a little under 100,000 in maternal and child health (MCH) programs (MOH 1995). MOH also finances the education of 220,000 medical students enrolled at 120 medical schools. Thirty of these schools belong to the MOH; others belong to local or provincial governments. 2.6 China's SOEs employ another 1.4 million health workers and operate 700,000 hospital beds, both accounting for roughly one-quarter of the national total. Since the 1950s, SOEs have provided health services directly to their employees, retirees, and their families. Most SOE health facilities are small clinics and health posts, but they include some hospitals and large service centers as well. Some large SOE hospitals and clinics serve third parties on a fee-for-service basis. The Taiyuan Machinery Works in Shanxi province, for example, owns and operates a 13  Table 2.1. Sources and Uses of Health Financing, 1993 (millions of 1993 yuan) Sources of Finance Government Eudget (excluding GIS) Insurance Society Finand ng 2/ Out of ocket Uses of Health Traditional Finance Recurrent Chinese Community Govt. Other Total Expenditure 1/ Medicine Others GIS LIS Financing and NGOs Enterprises Village Private Other Urban Rural (Percerit) Hospitals Outpatient 1,912 349 1,266 3,737 737 1,276 5,828 9,459 24,564 Inpatient 2,571 445 2,569 7,583 139 0 1,096 1,170 15,573 Total 4,483 794 3,835 11,320 876 1,276 6,924 10,629 40,137 ________ 30.4% Pharmaceuticals Hospitals 7,639 22,552 1,367 10,812 17,844 60,214 Individual providers, 45.6% retailers 3,300 5,417 8,717 6.6% Testing 112 103 215 ________ ~~~~~~~~~~~~~~~~~~~~~~~~0.2% . Epidemic 1,305 137 195 1,604 3,241 Prevention Service 2.5% Maternal/Child 324 193 569 137 195 534 381 799 3,132 Health Program _ 2.4% Other Primary 1,344 78 610 2,032 Health Care 1.5% Family Planring 2.292 677 2,969 2.2% Medical Research 408 42 1,160 816 2,426 and Education 1.8% Construction 2,810 3 1,144 342 877 143 1,067 6,386 4.8% Other 2,579 2.579 Total 10,786 917 7,175 11,667 34,441 2,243 1,618 1,151 1,067 143 4,734 21,417 34,689 132,048 Percert 8.2% 0.7% 5.4% 8.8% 26.1% 1.7% 1.2% 0.9% 0.8% 0.1% 3.6% 16.2% 26.3% 100.0% 1/ Health recurrent budget includes Ministry of Health and local government public health departments. Most spending is at the local, not central, level. 2/ In the Chinese health accounts data, society financing can be either public or private finance. For example, under EPS and MCH, society financing from others' refers to user fees. Source: Wei Ying, 1996.  300-bed hospital that receives a quarter of its revenue from patients not affiliated with the enterprise. In 1993, China's SOE health facilities delivered 18 percent of all outpatient and emergency care in China, and 13 percent of all inpatient treatment. The proposed reforms of SOEs would separate such services from the regular business of the enterprises. In some cases, however, the SOE may prefer to expand its medical care business rather than to give it up. The Taiyuan Machinery Works plans to expand its hospital to 500 beds, for example. 2.7 In addition to the health care infrastructure of the MOH and the SOE systems, thousands of health workers are employed by other government institutions such as the military and prison systems. In addition, there were some 150,000 health workers in private practice in 1990, and an estimated 190,000 in 1993 (excluding village doctors). The Three-Tier Rural Delivery System 2.8 The rural three-quarters of the Chinese population is served by a three-tier system of health services and referral. Normally, farmers and their families enter the system through a visit to one of China's approximately 1.44 million village doctors. These health workers engage in both medical service and farming, and often earn as much from farming as medicine. Many of them received rudimentary training as Barefoot Doctors in the 1960s and 1970s and continue working in the villages of their birth. They work independently and are not counted among the 5.2 million health personnel who work in China's medical institutions. 2.9 In this first tier of the system, village doctors diagnose and treat patients, prescribe pharmaceuticals, and refer patients to higher levels of service as warranted. Village doctors generally operate on a fee-for-service basis, but they also depend on the mark-up on pharmaceutical prescriptions (typically about 15 percent) as a source of medical income. 2.10 The village doctor may also refer patients to the nearby township health center/hospital (THC). This second tier of rural health care consists of some 52,000 rural township health centers operating 730,000 beds, or about one-quarter of all hospital beds in China. Township health centers are staffed by junior doctors and other medical personnel, and can deliver babies, treat infections and wounds, and provide some basic surgery such as appendectomies. The THCs depend substantially on patient fees to finance their services, but local governments also provide subsidies that cover part of their costs. These facilities tend to have lower bed-occupancy rates than the higher-level hospitals. Some farmers referred for hospital attention bypass the township facility because they question its quality, and go directly to the county hospitals. 2.11 There are about 4,000 county hospitals in China, which make up the third tier of the rural health care system. These hospitals are usually the last point of referral for inpatient treatment of rural residents, since few farmers can afford to be treated at specialized big-city hospitals. County hospitals have on average about 300 beds. They typically have five departments: obstetrics and gynecology, pediatrics, general surgery, internal medicine, and laboratories/X- rays, as well as emergency room facilities. 2.12 There are also three important vertical public health services that function independently of this three-tier system in rural China. These are the Epidemic Prevention Service (EPS) and the Maternal and Child Health Program (MCH) under the MOH, and the Family Planning and Reproductive Health Program (FP) under the Family Planning Commission. All of these programs receive budgets from provincial and county governments and also collect fees from 15  their clientele. User fees supplement the public funding to a greater or lesser degree depending on economic conditions in the area where the services operate. In wealthier provinces, such as Jiangsu, these services are largely self-supporting and receive financing, if needed, through rural collective enterprises. In poorer provinces, such as Shanxi and Guizhou, these services are less able to obtain reimbursement from clients and enterprises and are more dependent on government finance. Sources of Health Spending 2.13 China allocated about 3.8 percent of GDP to health in 1993, the most recent year for which comprehensive national health statistics are available.' Chinese society spent somewhat less on health care in earlier years--an estimated 2.9 percent in 1978 and 3.0 percent in 1986. Figure 2.1 shows the growth in China's health expenditures by sources and uses. The most dramatic change in health financing between 1978 and 1993 was the demise of the rural Cooperative Medical System (CMS) and its replacement mainly with private out-of-pocket spending for health care services in rural areas. The change from cooperative to personal financing may have hit the poor hardest, since they are among the least able to pay for health care out of savings. Other significant changes were as follows: * The share of health spending provided by the government budget, exclusive of subsidized care for government workers provided through GIS, declined from 36 percent to 16 percent of national health expenditures between 1986 and 1993; - The share of spending contributed by the CMS fell from 20 percent of health spending in 1978 to 2 percent by 1993; - Out-of-pocket payments rose from 20 percent of revenue in 1978 to 26 percent in 1986 and to 42 percent in 1993, transforming the financing base of the health sector. * In constant-price terms, government health spending (excluding GIS) increased threefold between 1978 and 1993. However, the rapid growth of the economy induced even faster growth in private health spending, which went up by a factor of 10 in constant-price terms. Spending by the Government Employees Health Insurance System (GIS) and the Labor Health Insurance System (LIS) rose only slightly as a share of national health expenditures, from 30 percent of all health spending in 1978 to 33 percent in 1986 and 36 percent in 1993. Uses of Health Spending 2.14 The main uses of China's health sector funds--public, private, and insurance-based--are for hospital services and the purchase of drugs. Public-sector staff and institutions deliver most health services, collecting fees from patients or their employers. 2.15 Most hospitals manage to balance revenues and expenditures. Hospital managers are expected to cover 85 percent or more of their costs from patient revenues. To attract patients, hospitals sometimes borrow from suppliers or their own staff to purchase highly profitable high ' The Health Economics Institute in Beijing (HEI), produced a revised estimate of national health expenditures in January 1996 for spending in 1993. The revision raised the 1993 estimate from 3.6 percent to 3.8 percent of GDP. HEI has not revised estimates for earlier years. 16  Figure 2.1. Growth in Health Expenditures in China by Source 1978-93 140 120 ,f, 100 - L CD Out of Pocket A 80/... Govemment 60 ~~~~~~~~~~~~~~~~~~~Insurance System (GIS) + Labor Insurance X - System (LIS) c 40 : Govemment 20 Budget (excluding GIS) oI Rural Cooperative Medical System 1978 1986 1993 technology diagnostic equipment. In 1993, 76 percent of all spending on health was used to finance hospital inpatient and outpatient services, including pharmaceuticals. 2.16 Chinese public hospitals are fairly autonomous compared to public hospitals in many other countries. Basic salaries are set by the government, but the hospital establishes bonuses, which can be two to three times the basic salary. Hospitals can move funds across budget categories, and can make their own capital investment decisions. They have little autonomy over personnel, however. Personnel are assigned to hospitals by the Ministry of Health and provincial health bureaus. 2.17 A 52 percent share of China's total health spending went to purchase pharmaceuticals in 1993. About 85 percent of pharmaceutical sales occurred in hospital inpatient or outpatient settings (Table 2.1). This 52 percent share of national health expenditures allocated to pharmaceuticals, whether in or out of hospitals, is higher than in most low-income countries. For OECD countries, spending on drugs averages 14 percent of health spending. 2.18 Overprescription and misallocation of resources towards drugs is one major efficiency issue in China's health system. Other problems are overuse of high technology diagnostics, and the long average length of hospital stays (15 days). These efficiency issues are discussed in more detail in Chapter 5. Government Recurrent Budget 2.19 In China's national health accounts, spending under the government's "recurrent health budget" (spending by the Ministry of Health and related departments at the provincial and 17  county levels) amounted to about 8 percent of all health spending in 1993 (Table 2.1). It includes spending on public hospitals (40 percent of the total), EPS (12 percent), and maternal child health services (3 percent). Public spending on family planning is outside of MOH's recurrent health budget. It amounts to about 1.7 percent of all health spending. Public spending on traditional Chinese medicine facilities (also separate from the recurrent health budget) is smaller yet, accounting for 0.7 percent of all health spending. 2.20 Public spending for the health care of government employees and related groups (8.8 percent of total health spending) is included as part of the government health spending, but is not part of the recurrent health budget. OECD countries generally keep this account separate as part of government non-wage employee compensation. GIS payments rose from 13 percent of government spending on health in 1978 to 34 percent in 1993. The bulk of increased government spending in the government health budget therefore went to pay the health care costs of government employees. There was also a big increase for the high priority family planning program. These changes have taken funds away from basic public health services. 2.21 China's EPS programs, which were entirely subsidized before 1985, are among those being pushed hardest to finance their work through user fees. Government regulations now require restaurant workers, for example, to be tested by the EPS, while their employers must pay for these tests. Shanxi's provincial EPS was entirely financed by the government until 1991. By late 1994, it received only 65 percent of its income from the provincial treasury, with the remainder coming from fees for testing services in cities and towns. There has been a significant decline in the frequency of field visits to villages, however, where costs cannot be recovered as easily. The EPS in Jiangsu province now receives just 60 percent of its income from the government treasury. In the relatively poor province of Guizhou, however, most EPS income still comes from the government because there is little capacity to sell testing services even in cities and towns. The EPS of Guizhou has cut its village field work even more than wealthier provinces like Jiangsu (Jin 1995a, pp. 25-26). Capital Expenditures 2.22 Duplication of facilities is a serious issue in China today. In urban areas, the overlap involves the Ministry of Health, state-owned enterprises, and traditional Chinese medicine facilities. In rural areas, there is some duplication and overlap of services between maternal and child health centers, family planning services, township health centers, and epidemic prevention stations (see also Chapter 5, Box 5.3). 2.23 From 1985 to 1989, as much as 80 percent of health investments went into hospital construction and equipment and less than 10 percent supported public health and high priority basic clinical services. In the Eighth Five-Year Plan period (1991-1995) the central government established a special fund of 1.1 billion yuan to strengthen public health and basic health care, known as the Three Items Construction Program. Local governments had to provide complementary funds for EPS and MCH programs at the provincial and county levels (Hou and Zhou 1995). The idea was sound, but the poorest counties and townships have been unable to generate counterpart funds, and have thus been the least able to benefit from this program. 2.24 Public hospitals have the autonomy to make their own capital investment decisions. Hospitals' reputations and their ability to attract clients now depend on having high technology equipment because CT scanners, ultrasound, and other diagnostics have come to symbolize 18  satisfactory health care to Chinese consumers. At the end of 1993, China had 1,300 CT scanners, 200 MRIs, and 1,200 color Dopplers (Hu Haobo 1995). Many specialists believe these instruments are not the most cost-effective investment for a country at China's stage of development, or indeed even for much wealthier countries. 2.25 Investments in health research can contribute substantially to meeting China's remaining health sector problems, and government support of research is an essential element of public health policy (Box 2.1). The Chinese national health accounts data do not disaggregate investments in medical research. These investments are included in the estimate for medical education and research, which totals about 1.8 percent of all health spending (Table 2.1). Government Fiscal Structure 2.26 Two fiscal problems complicate government efforts to finance health services and to promote redistribution between rich and poor areas of the country. First, revenues are decentralized, leaving the central government with few resources to transfer from rich to poor provinces. Only about 4 percent of the total recurrent health budget in 1993 fell under the direct control of the central government (Berman and others, 1995, p. 28). Province, prefecture, county, and township spending accounted for the remainder. This decentralization inhibits special assistance for the poor, since they generally live in provinces with less capacity to tax and redistribute benefits through publicly subsidized health services. Furthermore, regional income disparities are growing. 2.27 A second problem that constrains the government's ability to finance priority programs and promote redistribution has been the steady reduction in government revenue as a share of GDP. China's overall budgetary expenditures declined from 33.8 percent of GDP in 1978 to 13.8 percent in 1994 (World Bank 1996a). 2.28 These problems of decentralization and limited resource mobilization are addressed in a recent World Bank report that outlines the prospects for increasing China's central government revenues through more effective collection of the value-added tax from payroll taxes, and from a tax on pollutants. With these changes, government revenue would rise by 6 percent of GDP by the year 2000 (see Chapter 8, Table 8.4). The central government's share of revenues would increase to 60 percent, from about 40 percent in 1994, thus facilitating fiscal transfers to poorer provinces and counties. A key challenge for intergovernmental fiscal relations is to design and implement a grants scheme that will redistribute the central government's revenue surplus to the poorer provinces (World Bank 1996a). Proposals in later chapters of this report are consistent with those suggestions. Who Gets Health Services? 2.29 Two major concerns of this report are that health insurance coverage in China is declining and that the rural poor have inadequate access to health services. Population sub- groups differ markedly, however, in income levels, health insurance coverage, and health spending (Figure 2.2). In 1993, risk-pooling mechanisms (GIS, LIS and rural community financing schemes) covered only 21 percent of the population, but accounted for 38 percent of health expenditures. This wide difference suggests that those who are covered have far better 19  Box 2.1. Health Research and Development: A Neglected Component of the Chinese Health System? The high-income countries of the OECD invest a much more substantial fraction of their health expenditures on research and development (R&D) than do low- and middle-income countries. In the OECD, 3 percent to 4 percent of health expenditures goes to R&D; in developing countries the amount lies between 1/2 percent and I percent. China's R&D expenditures for health appear to be in the developing country range. In China, as elsewhere, elite universities and specialized institutions (such as the Chinese Academy of Preventive Medicine) undertake most of the health research. These institutions, along with industry, also engage in new product development, including the development of new pharmaceuticals. In the OECD countries, R&D expenditures split approximately evenly between research (mostly public sector) and product development (mostly private sector). There appears at present to be relatively little new product development in the Chinese pharmaceutical industry as indicated by the virtual absence of newly registered drugs from China (or, indeed, from any developing country). The issue for health policy concerns the extent to which China should consider investing a larger proportion of health expenditures in R&D and, if so, how this might best be done. One line of argument is that, since the OECD countries invest so heavily in R&D, the rest of the world need not do so: it can simply use the knowledge generated elsewhere even if, occasionally, such use requires payment of licensing fees or high costs for patent-protected products. A different line of argument is advanced in a recent report of the World Health Organization (Ad Hoc Committee 1996) which suggests four broad areas where China might productively invest R&D resources: (i) Operational research focused on control of the diseases of extreme poverty that still affect perhaps 100 million Chinese; (ii) Biomedical research and new product development to counter still evolving infections (such as AIDS or drug-resistant tuberculosis) with better vaccines, drugs, and diagnostic tools; (iii) Epidemiological, preventive, and clinical research to address rapidly increasing problems of noncommunicable disease and injury with interventions that are inexpensive enough to be widely implemented and sustained; and (iv) Health policy and systems research to address questions of cost containment, access, and quality of service. The case for expanding China's efforts in these four areas appears strong. The 1996 WHO report also stressed the importance of competitive allocation of R&D resources and full engagement with the international R&D communities to help China avoid reinventing what has already been learned elsewhere. Given the importance of new knowledge for improving health outcomes and constraining cost growth, China may wish to undertake an in-depth review of this aspect of its system. One approach might be to forn a committee like the one that prepared the WHO report--on which a Chinese scientist participated--to undertake the task. While most members of the Committee would of course be from China, it might also be useful to include the perspectives of a few eminent outside scholars and industrial scientists. capacity to meet their health care needs than the nearly 80 percent of the population who are not insured. 2.30 In terms of health service access, the 1993 data show that the poorest quartile of the rural population, i.e. the poorest fifth of all Chinese--virtually none of whom have prepayment or insurance arrangements to assure funding for health services--accounted for only about 4 percent 20  of total health service spending. This Figure 2.2. Per Capita Annual Health Spending low share supports the view that the rural on Population Croups, 1993 poor receive an inequitably small share of the available health services, as discussed further in Chapter 4. 250 Urban 2.31 Urban dwellers, in contrast, who 200 - account for approximately 53 percent of c 150 earned income, received at least two- thirds of all health spending. The high >~ 100 Rural income elasticity of demand for health 50 services would explain only part of these o large disparities. Health Coverage in Urban Areas Urban residents account for four times as much on health spending as do rural residents. Spending per 2.32 Official data indicate that 30 person is far lower among poor farmers, than among million persons, or some 2.5 percent of richer farmers, or groups covered by insurance. the Chinese population, are eligible beneficiaries of the Government Employees' Health Insurance System, or GIS. They receive free care in the government's clinics and hospitals or are reimbursed by their employing 400 - government agency. Beneficiaries include employees and retirees of 350 - central, provincial, and local governments, disabled veterans, and 300 - university students. The 1993 health survey revealed an apparent discrepancy 250 between official coverage of GIS and 2 survey responses: 5.8 percent of those > 200 a interviewed said they were covered by 60 _ ffi , = 9 GIS, which would be the equivalent of 70 million persons. The difference 100 3 between the 30 million people officially covered by GIS and these responses may 60i be due to dependents stating they are eligible even though the government may 0 * i _ not count them as eligible. About 10 percent of those who said they were covered by GIS in 1993 lived in rural Health spendingfor the lower-income quarter of areas. More analysis of actual GIS rural Chinese is one-tenth that of the government eligibility and coverage is needed, and SOE workers. especially in light of the very high per capita spending revealed by official Source: See Annex Table Al.6 and Wei 1995. estimates of the costs of this insurance program (Zhao Yuxin 1995). 21  2.33 In terms of insurance coverage and eligibility, the next best-covered group in China is the 11.7 percent of the population employed by SOEs. In 1951, the PRC government adopted a policy that SOEs should provide or finance health services for employees and retirees, with dependents treated on a 50-50 cost-sharing basis. This is the LIS insurance system. The government requires SOEs to contribute 14 percent of wages to cover both health and--to a much smaller extent--welfare benefits such as child care. With the erosion of profitability in recent years, many SOEs now restrict employee eligibility for insurance-paid health care. Surveys in 1992 and 1993 show that many workers who were covered in principle did not receive any insurance-paid care (Hu Teh-Wei 1995; Henderson and others, 1995). 2.34 SOEs spent 34.4 billion yuan for health care in 1993 (Table 2.1)--over a quarter of aggregate health spending in China. One-sixth went to the SOEs' own services, and the rest to other parties, especially public hospitals. SOE spending on their own services is about half as large as that of the government recurrent health budget. 2.35 This comparison underscores the risks for the government in trying to replace the services now delivered directly by SOEs to their associated workers and retirees. The government strategy for reform of SOEs recommends the separation of SOE-owned and managed health services from the principal business of SOEs. The creation of health insurance centers as part of demonstration projects in Jiujiang and Zhenjiang in late 1994 may offer a model for such change. Chapter 7 discusses options for reform of SOE health insurance. Health Coverage in Rural Areas 2.36 In the 1960s and 1970s, the rural Cooperative Medical System (CMS) reached the majority of rural Chinese. Village authorities used funds from agricultural collectives to pay for the salaries and training of Barefoot Doctors. These locally-recruited health workers focused on the villagers' basic health needs. The collectives helped to pay for farmers' health care, but most services required substantial copayments. By 1985, however, less than 10 percent of China's villages maintained CMS arrangements, most of them with village incomes falling within the upper quartile (Zhao Zhuyan and Lusheng Wang 1995). The decline of the CMS program is discussed in more detail in Chapter 6. 2.37 Some vestiges of the CMS system remain even in poor villages, however. Village doctors throughout China sell contract or prepayment insurance for selected EPS and MCH services. There are both one-year and four-year contracts available, depending on how long the series of basic shots and dosages will take to administer. A typical EPS contract provides four years of immunizations (the Expanded Program on Immunization (EPI) standard groups of DPT, measles, and polio) at a prepaid price of a few yuan. The MCH contract covers ante-natal and post-natal care for a child up to age seven and includes nutrition, growth monitoring, and referral, if necessary. These special insurance contracts reach as much as 40 percent or more of the targeted child population. 2.38 Community financing schemes similar to the former CMS system still operate in the rural areas of some wealthy provinces like Jiangsu. Community health funds pay for most medical fees and pharmaceuticals. The funds come from township enterprises, village tax collections, and voluntary contributions. A THC health committee supervises use of the funds by each village. Each village clinic has its own account to pay for drugs and other materials 22  consumed. Salary and subsidies of the village doctors do not depend upon pharmaceuticals sales, but rather on the number of patients served. 2.39 In the village clinics in Shanxi province, a different financing pattern prevails. Community financing still exists and requires a 4 yuan per person-year prepaid fee to cover four services. Pharmaceutical coverage is excluded. The villages provide the clinics with office space, and village doctors make their income from selling drugs: the more they sell, the higher their income. In October 1994, the majority of village clinics in rural Jinzhong prefecture, Shanxi, resembled drug stores. Those visited had stocks of more than 1,000 kinds of drugs (Jin 1995). Trends in Paying for Health Care 2.40 Most Chinese--some 800 million in rural areas, and perhaps 100 million in urban areas-- pay directly for health services when they receive them. Of those surveyed in 1992-93 who had been referred to a hospital for care, 40.6 percent did not seek hospitalization on grounds of excessive cost and their inability to pay (Zhao Zhuyan and Lusheng Wang 1995). Even middle- income farmers would be unlikely to have enough savings to pay for a long hospital stay. The share of health spending that is out-of-pocket has risen steadily since the late 1970s. While many countries are moving towards a curative health system that is financed publicly but provided largely privately, China is moving in the opposite direction. 23  I  3. STRENGTHENING PUBLIC HEALTH PROGRAMS 3.1 Governments have a large role to play in assuring public health services and this role needs to be strengthened in China. Public health programs are designed to address the health problems of entire populations or subgroups in the population. Public health activities may involve specific health services for the community (such as immunizations), the promotion of healthy behavior (reducing tobacco consumption, limiting salt intake, avoiding sexually transmitted diseases), or the promotion of healthy environments (improvements in sanitation). Some clinical services--such as the treatment of tuberculosis and sexually-transmitted diseases-- are considered public health activities because they yield large social benefits by preventing or curing these infectious diseases. Public health programs can also play an important role in providing health education and in ensuring that infectious diseases are detected and adequately treated at the clinical level. Disease surveillance is another public health activity. Most public health programs provide services for which there are large social benefits, but for which individuals are unwilling to pay the full costs. As such, public finance for priority public health programs can correct these "market failures." 3.2 China has a long history of well developed public health programs, but in recent years these programs have faced funding difficulties, as well as coordination and other operational problems. In addition, China's disease patterns and health risks are changing, and public health programs need to adapt to new challenges. This chapter reviews the status of public health programs in China today. It is not intended to be an exhaustive review but instead uses tuberculosis control, immunizations, and tobacco control to illustrate problems in current public health programs and suggest broad directions for change. It concludes that public health programs need considerable strengthening in China--in terms of public finance, program strategies, and content. Structure and Finance of Public Health Programs 3.3 The Ministry of Health (MOH), under the authority of the State Council, provides technical leadership and sets guidelines on public health activities as part of its overall leadership in the health sector. The Epidemic Prevention Service (EPS) is the backbone of public health programs in China. At the county level, the EPS is made up of the epidemic prevention stations and affiliated staff. Many other agencies also carry out some public health activities, or public health research, including the General Office of the National Patriotic Health Campaign Commission and its local branches, the Chinese Academy of Preventive Medicine, the Center for Health Statistics and Information, and the National Institute for Health Education. 3.4 The EPS employs a quarter of a million workers and extends disease control programs throughout rural China. EPS maintains the cold chain (refrigeration equipment) for immunizations, makes field visits to assure water quality, is responsible for the control of diarrheal diseases, and runs specific endemic disease control programs in many areas (such as those against malaria and schistosomiasis). For several decades EPS was fully funded, both in terms of staff and operating costs, from provincial budget outlays that drew on general revenues from each province. Public health activities until the 1980s were delivered in a highly organized fashion under EPS supervision, drawing on the village doctor, township health center (THC), and county hospital as needed. 25  3.5 China's fiscal decentralization in the early 1980s weakened both the financing and the coordinated operation of public health activities. The fiscal decentralization gave much more budgetary autonomy to local governments. Provincial health bureaus develop their own programs according to national guidelines. County hospitals, epidemic prevention stations, maternal and child health (MCH) centers, and THCs continue to receive some public subsidies for salaries, but they, too, are now required to generate substantial additional revenue from user fees. With the fiscal decentralization, poorer counties now have the least capacity to develop and maintain public health programs. Coordination has weakened between the epidemic prevention stations, THCs, and the village doctor. The epidemic prevention stations have difficulty supervising and influencing the activities of village doctors, who now operate as independent practitioners and generate income from fee-for-service medicine. 3.6 In absolute terms, public financing of EPS has Table 3.1. Government Finance of EPS and as a Share of remained at about 1.5 billion GDP, Selected Years yuan since 1986. As a share of GDP, however, it has fallen Public Finance from 0.11 percent in 1978 to 0.04 percent in 1993 (Table Year 3.1). The EPS faces funding EPS EPS EPS EPS difficulties due to rising input Budget" Budget Budget as Revenue costs and the fact that the (1993 per a share of from User personnel budget needed to billion capita GDP Fees cover retired health workers is yuan) (yuan) (percent) (1993 increasing. In order to cover the billion costs of its services, the EPS has yuan) had to rely increasingly on revenue generation. Over the 1978 0.94 1.0 0.11 N.A. past several years, ancillary 1982 1.13 1.1 0.10 N.A. services whose costs cannot be 1986 1.54 1.4 0.09 N.A. recouped through user fees have been cut. For example, in 1990 1.58 1.4 0.06 N.A. Shuoyang county, Shanxi, 80 1993 1.53 1.3 0.04 1.6 percent of the total EPS budget Source: Wei Ying 1995 and 1996. went to staff salaries in 1993. N.A. = not available The number of days assigned to 1/ These numbers differ from those presented in Table 2. 1, Chapter 2, field work in villages in 1994 because the numbers here include both the recurrent and capital budget. was less than a quarter of what it had been five years earlier. 3.7 EPS generated an estimated 1.6 billion yuan in revenue from fees paid by individuals and institutions in 1993 over and above the 1.53 billion yuan it received from the government budget. These user fees have had two detrimental effects. First, EPS staff have reallocated their attention to services for which fees can most easily be charged, such as food inspections, although these services are not necessarily the highest priority activities. Second, fees have reduced demand, particularly among the poor, for such services as tuberculosis control and preventive health services. 26  Immunizations: Improve Coordination and Funding 3.8 The Expanded Program on Immunization (EPI) is a key public health program in China. Since the late 1970s, China has provided political and financial support for immunization against tuberculosis, diphtheria, pertussis, tetanus, and polio, with measles coverage introduced later. Coverage of the four basic vaccines (BCG, DPT, polio, and measles) reached at least 80 percent in all provinces in 1988, and nearly 80 percent in all counties by 1990--with dramatic benefits. The incidence of pertussis decreased from 126 to I per 100,000 between 1978 and 1993, and of measles from 250 to 10 per 100,000. China recently expanded its immunization goals. The government hopes to eradicate polio from China soon and hepatitis B was added to the EPI program, although it is not yet clear how widespread the coverage of newborn infants will actually be. Given the high levels of post-natal transmission of Hepatitis B in China, the addition of immunization at birth against Hepatitis B provides a cost-effective way to prevent liver cancer and cirrhosis of the liver. 3.9 But the immunization program also faces major challenges. With the decline in funding for EPS, transport and refrigeration facilities are breaking down, and many local programs face shortages in supplies. Major system breakdowns occurred in 1993 and 1994 in many poor areas, which led to declines in immunization coverage. Village doctors operate now as independent practitioners, and no longer coordinate closely with EPS. Although village doctors are provided with vaccines from EPS and are paid a flat fee for carrying out immunizations, the fee is not adequate to cover their operational costs (for needles, syringes, sterilization, and operation of the cold chain). This can lead to inappropriate sterilization practices and loss of vaccine potency. EPS and village doctors, in turn, charge for immunizations in many areas, presenting a financial obstacle for poor households. To generate additional revenue, some EPS staff perform and charge for unnecessary antibody tests before providing immunizations. Table 3.2. Immunization Coverage in China, 1993 Finally, coverage differentials are still (percent) significant between provinces and BCG Polio DPT Measles between urban and rural areas (Table National 90 88 86 85 3.2). Measles coverage in rural Shanxi Urban 95 94 93 85 and Guizhou is as low as that of many Rural 85 86 84 73 Sub-Saharan African countries, while in Shanxi 65 66 62 49 the cities the increasing unregistered Guizhou 72 74 67 53 urban or "floating" population lacks Source: World Bank 1995e, p.3. Note: National, urban, and access to these public health services. rural figures for measles are inconsistent. 3.10 Additional funding is needed to consolidate and expand China's gains in immunizations. This requires changes in other priority public health programs as well: providing for adequate salaries and supplies, appropriate training and supervision, performance-related incentives--both for EPS staff and village doctors who coordinate with them--and good working conditions to maintain the commitment of public health workers. Tuberculosis Control: Expand and Subsidize the New Tuberculosis Program 3.11 The tuberculosis control situation in China illustrates both what can be achieved with a well-run public health program and what can go wrong. Although the death rates are decreasing, tuberculosis still remains a major health problem in China, accounting for an estimated 3 percent 27  of all deaths in 1990 (Annex 3, Table A3.1). Active tuberculosis is a highly infectious disease. Public health programs throughout the world try actively to identify and treat infectious persons early in the course of the disease to interrupt transmission to others. Without appropriate treatment, 60 percent of TB patients will die. Well-run programs can cure 80 to 90 percent of patients; poorly administered programs cure 30 percent or less. Most programs around the world try to avoid any financial barriers to tuberculosis treatment. Treatment is provided free of charge, and some programs even pay patients to comply with tuberculosis treatment. China made substantial progress in controlling tuberculosis during the 1960s and 1970s, using standard antibiotic treatment that was essentially free of charge. 3.12 Changes in China's health financing in the 1980s, however, had a detrimental impact on tuberculosis control programs and their effectiveness. As public subsidies were reduced, public facilities were encouraged to recoup their expenses from user fees. This resulted in many distortions. When doctors and hospitals expected to be reimbursed by GIS or LIS, they provided excessive diagnostic tests and examinations during treatment and dispensed higher-cost antibiotics that should have been reserved for more resistant cases. Daunted by the costs, many low income people infected with tuberculosis failed to enter treatment or dropped out early. There were no incentives to ensure that patients completed their treatment or were cured. As a result, many cases of tuberculosis remained infectious, spreading the disease to others. The spread of drug-resistant strains has also accelerated in China since the 1980s as a direct result of poor treatment practices. 3.13 Recognizing the problems that result from charging for tuberculosis therapy, China has launched a new tuberculosis control program that is already operational in certain areas of the country. The WHO-recommended program of directly observed short-course chemotherapy provides subsidies for treatment and appropriate incentives for providers of care. In the new tuberculosis program, patients with symptoms suggestive of tuberculosis are referred by health providers to the tuberculosis county dispensary under the EPS for physical examination and fluoroscopy. If a smear of the sputum tests positive for tuberculosis, drugs are to be provided free of charge. All smear-positive patients are to be treated with short-course chemotherapy and every dose is observed by the village doctor. The village doctor is paid for performance--an initial payment when the patient is enrolled, a further payment at two months, and a final payment after completion of treatment. Tuberculosis cases are to be closely monitored at the county level. Supervision is emphasized: at the beginning of each treatment, county tuberculosis dispensary staff and township disease control officers meet each patient and the village doctor managing the patient to review the treatment plan. Other aspects of the program are also supervised, including the laboratory protocol and county registry. In the first four years, almost 1.6 million patients with suspected tuberculosis were referred to the program. The cure rate is now 90 percent among new cases, compared to about 50 percent previously. The failure rate in previously treated cases, which is an indicator of drug resistance, fell from 18 percent in 1991 to 6 percent in early 1994 (China Tuberculosis Control Collaboration 1996). This experience shows that careful supervision, adequate funding, and appropriate financial incentives to providers can make a dramatic difference for this major public health problem. Unfortunately, many patients (those outside the project area, and those within the project area who are not referred to the tuberculosis program for treatment) are still being charged for treatment and their cure rates are low. The cost for drugs plus all other services, including unnecessary laboratory exams, x-rays, and traditional medicines, may total more than 1,000 yuan. 28  Anti-Tobacco Efforts: A Multisectoral Approach 3.14 Smoking is a major health problem in China today. If current smoking patterns persist, about 50 million of the Chinese who are now aged 0-19 will eventually die as a result of smoking (Peto 1986). According to a 1984 nationwide survey, 61 percent of men and 7 percent of women in China smoked manufactured cigarettes (Table 3.3). Chinese men account for about 10 percent of the adults in the world, but consume 30 percent of the world's cigarettes. Smoking prevalence is negatively related to educational status. More recent surveys indicate that smoking is becoming even more widespread. Table 3.3. Characteristics of Smokers in China, 1984 Male Female Regular Occasional Regular Occasional Smokers Smokers Smokers Smokers Ig.tal 55.5 5.5 6.4 0.7 Ag= <20 14.1 5.0 0.2 0.1 20 -60 63.3 6.0 5.8 0.6 >60 63.6 3.0 16.5 1.4 Selected Worker 60.7 5.0 6.6 0.7 Peasant 58.1 5.7 6.0 0.6 Cadre 54.9 4.3 5.6 1.0 Teacher 44.8 5.3 2.6 0.3 Doctor 49.1 7.6 2.2 0.2 Education College 39.6 5.2 2.8 0.8 Middle School 47.5 6.3 1.7 0.2 Primary 61.2 5.3 5.5 0.6 Illiterate 63.9 4.0 10.6 1.0 Source: National Survey on Smoking in China, 1984: See Teh-Wei Hu 1995, Ministry of Health, 1991. 3.15 Over 800,000 deaths were attributable to smoking in China in 1990, including deaths from coronary heart disease, chronic obstructive lung disease, and lung cancer (Annex 3, Table A3.7). Given present patterns of smoking, this number is projected to increase dramatically over the next thirty years to at least 2 million deaths per year. The economic loss in terms of health care expenditures related to smoking was estimated at 6.9 billion yuan in 1989 (Jin 1995c). 3.16 Effective public health programs to reduce tobacco use are based on two complementary strategies: tobacco taxation, to provide an economic deterrent to consumption, and measures such as health education, banning of smoking in public places, and banning of cigarette advertising and promotion, particularly by foreign tobacco companies. One important legislative option would be a complete ban on all direct or indirect advertising and promotion of any tobacco goods or trademarks. If advertising and promotion are not prohibited, then not only will the high prevalence of smoking among males be perpetuated, but also a substantial proportion of 29  females may be induced to smoke, further aggravating the growing epidemic of death from tobacco use in China. China already has cigarette taxes in place. In fact, cigarette taxes are a major source--31 billion yuan, or 9.5 percent--of government revenue (1992 figures; see Teh- Wei Hu 1995). The effective tax rate on cigarettes was about 38 percent in 1991. This compares to 60 percent in Japan and Korea and 85 percent in Denmark. Several Chinese cities have already taken steps to ban cigarette advertising and smoking in public places, but these restrictions are being circumvented by sport sponsorship and other forms of tobacco promotion, and more action is needed. The government is considering increasing the cigarette tax and using the revenue to finance anti-tobacco campaigns and other public health efforts. 3.17 The amount of revenue generated from an Table 3.4.aEstimates of De bacco increase in tobacco taxes depends upon the price Parices Elastricit ofDemand,elasticity of demand in China. In other countries, the Various Countries price elasticity of demand has been estimated at Country Elasticity between -.35 and -.74 (Table 3.4). Assuming a rough average elasticity of -.5 in China, a 10 percent USA -0.50 increase in tobacco prices would reduce demand by Ireland -0.38 an estimated 5 percent and would generate an United Kingdom -0.39 additional 5 billion yuan of revenue per year. A 20 percent incremental tobacco tax would generate an Switzerland -0.50 additional 10 billion yuan in revenue. An ad valorem Austria -0.54 tax (a tax based on a fixed proportion of the retail price) would be preferable, so that the tax would Finland -0.35 adjust to changes in tobacco prices. It is also Canada -0.74 preferable to impose the tax on all tobacco products, Median -0.45 so as to reduce the possibility of substitution. Such a SoreZmiganeso95 pptax would be more progressive, because high-income Source: Zimring and Nelson 1995, pp525 smokers who purchase more expensive brands would be willing to pay more. A 10 or 20 percent incremental tax on tobacco products, accompanied by further efforts in health education, bans on tobacco advertising and sports sponsorships, limits on the tar and nicotine content of manufactured cigarettes, and media campaigns against tobacco, would generate important health benefits in China. 3.18 Given the rationale for raising the tobacco tax to reduce tobacco-related illness now and in the future, there is an argument for targeting all or part of the revenue from this tax for complementary anti-smoking activities, such as health education programs. China should consider using the additional revenue generated to support both anti-smoking and other priority public health programs. Conclusions 3.19 China needs a strong publicly-funded public health service adaptable to the changing patterns of risk factors and disease burden in China. This may be achieved by a restructured and reformed EPS in coordination with other agencies. Central and provincial governments should fully finance high priority public health services since these services must be financed by the government if they are to be provided at socially optimal levels. Central and provincial funding is also necessary to ensure that the poorest counties, with some of the biggest public health 30  problems, can pay for their programs. Funding for activities that are now the responsibility of EPS, for example, needs to be considerably increased over the medium term from its present level of 1.3 billion yuan (in 1993) to at least 6.5 billion yuan in 2001, based on current cost estimates of fully funding tuberculosis control, the expanded EPI (including universal hepatitis B immunization for infants), endemic disease control, health education, and other important programs, as well as projected cost increases to the year 2001. (As income is projected to double in real terms, EPS costs will also rise, particularly due to wage increases). The new tuberculosis program needs to be expanded throughout the country, with tuberculosis treatment provided free of charge. Special outreach measures are needed for the unregistered urban population. 3.20 At the same time, the government must ensure that public health programs are implemented efficiently. This requires: (i) providing for an adequate salary structure, appropriate training, performance-related incentives, and good working conditions to maintain the commitment of public health workers, (ii) financing and developing new disease control programs that address the changing pattern of disease in China, and (iii) ensuring that China's highly effective health surveillance system is maintained and adapted to the changing pattern of disease burden and risk factors. Inter-sectoral cooperation and collaboration will increasingly be needed since it is beyond the scope of the health sector alone to deal effectively with many public health problems such as lead in air, indoor air pollution, drug abuse, traffic accidents, suicides, and HIV/AIDS (Box 3. 1). 3.21 Given China's looming burden of costly illness and premature death from tobacco- related diseases, a substantial increase in tobacco taxes is another recommended public health action, accompanied by stronger additional measures to reduce smoking. This would help reduce eventual illness and death from smoking-related diseases and, if the incremental revenue were directed to public health (as in Australia and California), would also help finance public health programs. The tobacco tax recommendation is discussed further in Chapter 8. 31  Box 3.1 The Chaflenge of STDs and HIV in China The prevention and treatment of sexually-transmitted diseases (STDs) and the Human Immunodeficiency Virus (HIV), the virus that causes AIDS, pose particular public health challenges in China today. WHO estimates that China had 100,000 cases of HIV infection at the end of 1995. Despite this relatively low number,. China faces the risk of rapid spread of HIV and an explosive epidemic. Several factors contribute to.this.risk. With economic modernization, increasing mobility of the population can be a vehicle for spreading the disease. Rapidly growing migrant populations are another problem. Migrants tend to have little access to medical services, little knowledge about HIV/AIDS, and migrant women from rural areas are more likely to engage in commercial sex work to supplement their income or as their main source of income. More broadly, many vulnerable subgroups in the population have almost no knowledge about the HIV virus and how to prevent its transmission.: Finally, the blood supply in China is poorly screened and overly commercialized, posing a threat to blood recipients. The areas where HIV is most prevalent in China and the modes of transmission are changing. About 80 percent of the reported cases of HIV infection were previously in the southwestern province of Yunnan, where drug trafficking and prostitution are major problems. But the number of cases in coastal areas and large municipalities is now rising. While about 64 percent of current HIV infections in China result from injection drug use (Ministry of Health 1995b), the govemment believes that the major mode of HIV transmission in the coming years will be sexual transmission, as a result of increasing rates of high-risk sexual behavior. Evidence for this comes, in part, from China's rising STD rates. STDs have reemerged as a significant public health problem in China, after major efforts to eradicate them in the 1950s and 1960s. Reported STD cases rose from 5,838 in 1985 to 237,573 in 1993 (Ministry of Health 1995b). Concerted prevention efforts now, while the prevalence of HIV is still relatively low in China, can be expected.to have a huge pay-off. The cost effectiveness of interventions drops sharply, however, when infections cross from high-risk groups to the general population. Moreover, increased prevalence of HIV is likely to generate many additional health problems. It would: greatly exacerbate the already large TB problem in China, for example, since TB is one of the major opportunistic infections of HIV. Prevention efforts should include disseminating information on how to avoid infection, promoting condom use, treating other sexually- transmitted diseases, and reducing blood-borne transmission. A growing body of research shows that STD prevention and treatment has a significant impact on averting HIV transmission. To be effective, a multisectoral approach to controlling the spread of HIV and STDs will be needed. The government's National Strategic Plan for the Prevention of AIDS and STDs (Ministry of Health, 1995b) lays out a plan that would involve twenty-two government agencies and groups in society in prevention activities aimed at different subgroups. These subgroups include drug users, prostitutes and clients, STD patients, migrant populations, transport workers, overseas laborers and businessmen, and men having sex with men. The Ministry of Railroads, for example, could help in condom promotion and in educating migrant groups about AIDS. With the support of the World Bank-financed Disease Prevention Project, the Chinese government has recently embarked upon programs to help prevent and control the spread of STDs and HIV. These prograrns, which will strengthen the planning and implementation capacity at the center, are being implemented in seven major cities and in Yunnan Province. They focus on intersectoral coordination, policy reform measures (for example, to promote anonymity of treatment, improved availability and use of condoms, and sex education in. schools), behavior risk factor surveillance, training of health workers, improved management of STD cases,. and other health promotion measures. 32  4. MEETING THE NEEDS OF THE POOR 4.1 Poverty is both an important cause and consequence of ill health. The poor are more likely to suffer from ill health, and their health problems can contribute to keeping them in poverty. There is a strong justification on poverty assistance (or equity) grounds for government subsidies aimed at improving the poor's access to important health services. The Chinese government's role at present is minimal in this area. Resources need to be redirected or expanded to assure key services for the poor. At the same time, given scarce public resources, subsidies need to be carefully targeted to the very poor and to those services that provide the greatest impact on health for the monies spent. 4.2 China has experienced an Table 4.1. Incidence of Absolute Poverty, 1978-90 enormous reduction in poverty since 1978 1985 1990 economic reforms began. But Total Population 963 1059 1143 ~ progress has been unsteady. (millions) Improvements in the quality of life (millions)______ ________ for the poor proved rapid in the early Urban 172 251 302 1980s, driven by fast rural growth. Rural 790 808 841 The pace then stalled as the locus of economic expansion shifted to urban Incidence of Absolute Poverty (millions) and coastal regions. The majority of the poor in China are now Total Poor 270 97 98 concentrated in resource-deficient (% of population) (28.0%) (9.2%) (8.6%) rural areas, and comprise entire Urban Poor 10 I 1 communities located mostly in (% of population) (4.4%) (0.4%) (0.4%) upland sections of the interior provinces of northern, northwestern, Rural Poor 260 96 97 and southwestern China. Their (% of population) (33.0%) (11.9%) (11.5%) health status is bleak. Infant and Source: World Bank, 1992b. maternal mortality rates in very poor Note: The urban population figures underestimate those actually counties are at least 50 to 100 residing in urban areas. By convention, the Chinese statistical system treats unregistered urban residents as rural residents. percent greater than the national average, and are much greater yet in the poorest townships and villages. 4.3 There are signs that progress in poverty reduction resumed in the early 1990s. While the trends are clear, the number of Chinese considered to be poor depends on the poverty line. The national definition of poverty yields some 80-100 million poor (Table 4.1), whereas the figure jumps to 350 million (in 1993) if international norms of minimum acceptable living standards ($1 per person per day) are used (World Bank 1996c). 4.4 While absolute poverty is largely a rural problem in China, large numbers of urban poor, particularly unregistered migrant populations, also have little access to public health programs and clinical services. 4.5 A broad range of measures is needed to reduce poverty in China, including policies to improve labor mobility and foster rural enterprise and agriculture development in poor areas. 33  Targeted subsidies to improve the health status of the poor will also improve their welfare and economic productivity. This chapter discusses how targeted health assistance might best be achieved and the likely costs of such assistance. The bulk of the discussion is on the rural poor, since that is the locus of most absolute poverty in China. Health Problems of the Poor 4.6 The poor face a greater disease burden than the nonpoor. According to the findings of the World Bank's World Development Report 1993 (World Bank 1993a), selected childhood diseases, tuberculosis, maternal problems, micronutrient deficiencies, sexually-transmitted diseases, and a few others, account for much of the incremental disease burden borne by the poor relative to the middle classes. The Rural Poor 4.7 This pattern holds true in China Table 4.2. Health Status by Income, Rural China, as well. Much of China's infectious and 1993 parasitic disease burden, including Income I II I IV tuberculosis, diarrheal disease, and iodine Quartile deficiency disorders, is concentrated in Average Per 927 677 561 441 poor and remote areas. Roughly 50 Capita percent of children in households at or Income below the absolute poverty line are at (yuan/year) least mildly malnourished (stunted) while iron, vitamin A, and other micronutrient Infectious deficiencies remain common among the Disease/1000 poor. As many as 90 percent of poor children suffer chronic worm infections. Infant 29 34 44 72 Table 4.2 shows that the poorest quartile Mortality of the rural population reports an Rate/1000 infectious disease rate three times that of Life 71 69 68 64 the richest quartile of the rural population, Expectancy _ I and an infant mortality rate over twice Source: The 1993 National Health Services Survey and the that of the richest quartile. 1990 Population Census. 4.8 The rural poor not only have a larger disease burden, but they use health services less than higher income rural residents. One-third of low-income households sought no health care according to the Study of Thirty Poor Counties (China Network for Training and Research in Health Economics and Financing 1996), whereas only 16 percent of the high-income rural households sought no health care in the previous year. The number of visits per person for the lowest quartile income group was only 60 percent of the number of visits made by the highest income quartile; the number of inpatient days per person of the lowest income quartile was just 50 percent that of the highest quartile. Data from the Study of Thirty Poor Counties also show that the poor spend a high relative share of their cash income on medical care: 23 percent for those with household cash incomes under 250 yuan per month compared to 11 percent for those with incomes between 430 and 690 yuan per month (Luo Wujin as cited in Hammer 1996). 34  The Urban Poor 4.9 Much less is known about the health problems of the urban poor, notably the unregistered urban poor. The urban migrant population is much more likely to live in crowded, unsanitary conditions, and face higher risks of contracting communicable diseases, such as tuberculosis, than other urban residents. Migrant women are more likely to engage in commercial sex work on a full-time or part-time basis, with its concomitant serious health risks. China's cities have few community-based health service facilities, and the outreach of public health programs to unregistered urban populations is weak, at best. The Role of Government Spending 4.10 Government spending on health has not effectively reached China's poorest residents. China's 1981 public finance reforms decentralized public finance to the provincial and county level, limiting the central government's ability to redistribute funds from richer to poorer areas of the country. An analysis of public expenditure over eleven years shows that the allocation of public expenditure is skewed towards richer regions and, within regions, to those provinces that are growing fastest (Hammer 1996). Within provinces, government spending is concentrated on government employee health insurance expenses and hospital inpatient and outpatient spending. Services that disproportionately benefit the poor, such as the Maternal and Child Health program and the Epidemic Prevention Service, have been constrained and increasingly forced to rely on revenue from user fees. Not surprisingly, traditional public health activities such as immunizations have the greatest coverage in the highest income provinces. In China's poorest provinces, less than half the children receive their third dose of DPT, compared to more than 90 percent in the richest provinces (Hammer 1996). Bringing Key Health Services to the Poor The Rural Poor 4.11 The State Council Leading Group Office of Poverty Alleviation and Development has identified 592 counties with the greatest incidence of rural poverty (Table 4.3). These counties, with a total population of about 210 million people, make up about one-quarter of all Chinese counties. In these counties, 58 percent of the population had incomes below 500 yuan and about 26 percent of the population had incomes below 300 yuan per year in 1992-93. Experts disagree as to exactly how many of China's absolute poor reside in the 592 poor counties, but it is generally agreed that at least half are in these counties. Thus, the 592 counties constitute a useful beginning reference point for subsidies to improve the poor's health. 4.12 The key issue here is targeting: how to reach the poor (and only the poor) with the appropriate transfers without incurring excessive administrative or political costs. There are several ways to target the poor: * Geographical targeting to areas where the poor are concentrated. For example, poor administrative villages in China's 592 officially designated poor counties could be targeted for subsidized services. 35  * Individual or household targeting, by identifying and certifying the poor for subsidized services. (China may be one of the only developing countries in the world where this is actually feasible, given good government records, but it is administratively costly.) * Program targeting to health services that particularly benefit the poor, such as deworming and management of acute infections in children. (This would be additional to the public health programs discussed in the previous section.) 4.13 This report recommends a blend of geographic targeting (probably most practical at the administrative village level) in China's poor counties and universal finance of a limited number of services that particularly benefit the poor. Program targeting could be for the entire population in China's poor counties (which would provide an element of geographic targeting as well) or for the nation. Costs would be higher if targeting is done at the national level, but the one-third of the absolute poor who do not live in officially designated poor counties could be reached in this way. The financing for subsidies would have to come largely from the central and provincial level, since poor counties have little means to finance such programs. 4.14 Under a program of geographic targeting, the government might direct subsidies at health care providers in poor administrative villages (supply-side subsidies). Providers would be responsible for offering services free or at low cost to village residents. The government might further define what services would be provided for free, including highly cost-effective services to ensure maximum health impact from the program. If the administrative village was part of a viable community-financing scheme, the government could instead channel the subsidies to the community-financing scheme for health care of the residents (demand-side subsidies). Community-financing schemes are defined and discussed in the Chapter 6. 4.15 Under a program of service targeting, the government might fully fund certain health programs that disproportionately benefit the poor. The major disease conditions in China that are almost entirely associated with poverty are listed in Table 4.4. All told, these conditions accounted for an estimated 23 percent of China's disease burden in 1990. These conditions can be largely addressed by cost-effective public health and clinical interventions, such as immunizations, short-course chemotherapy for tuberculosis, iodized salt treatments, deworming, and prenatal and delivery care. Since these interventions largely benefit the poor, one option for targeting public subsidies to the poor is to focus public finance on these services. 4.16 Per capita health spending in rural areas is currently about 60 yuan per year, including public and private spending. For the sake of illustration, the public subsidy for a package of services for poor villages in poor counties might be 25 yuan per capita in 1993 in order to finance a 30 yuan package (with the other 5 yuan from other sources). The services that could be provided under a 30 yuan-per-person package are described in Box 6.1 in Chapter 6. With income growth projected to double between now and the year 2001, the per capita costs could be assumed to double as well. If the population of approximately 75 million in poor villages in poor counties were targeted, the costs would be about 3.75 billion yuan in the year 2001. An additional 1 billion yuan or more might be directed at program subsidies, either to the entire population in poor counties (210 million) or nationwide. These calculations could be altered depending upon assumptions about the size of the population to be targeted and the size of the per capita subsidy by the year 2001. 36  4.17 These poverty subsidies should be accompanied by systematic monitoring and evaluation to determine whether these programs indeed help the poor as intended and to guide policy improvements over time. The Urban Poor 4.18 While poverty is concentrated in rural areas, measures also need to be taken to improve access to essential health services for the urban poor, who are outside the coverage of the LIS and GIS urban insurance schemes. As a first step, fully funding priority public health programs with public finance would help the urban poor (Chapter 3). For clinical care, most urban health services are provided by hospitals. China could consider measures to revitalize community care, especially in poor neighborhoods, as a way of reaching the urban poor. Perhaps some of the current hospital subsidies could be redirected, over time, to subsidies for services for the urban poor, either at community clinics or hospitals. If some urban areas form a strong enough community, with sufficient stability and cohesiveness to enter into a social contract with each other, community-financing schemes might be a feasible channel for financing and organizing services for low-income urban residents (Hsiao 1995d). -Public subsidies for interventions that particularly benefit the poor could also help the urban poor. 37  Table 4.3. China's 592 Poor Counties: Comparison of Average Income, 1992-93 Total Population Population in designated poor counties of Poor with per capita incomes under: Counties 500 yuan 300 yuan 200 yuan Total (millions) 210.65 121.4 55 18.4 Percent 58% 26% 9% By province (percent): Yunnan 83 52 20 Guizhou 75 48 23 Zhejiang 74 26 Sichuan 74 29 7 Gansu 68 27 10 Shaanxi 66 29 6 Ningxia 64 34 9 Hebei 64 24 10 Shanxi 63 16 4 Hunan 60 29 11 Guangxi 58 35 11 Henan 58 22 6 Hubei 57 23 3 Qinghai 51 21 6 Xinjiang 42 19 8 Anhui 38 9 1 NeiMenggu 37 15 2 Jilin 33 20 12 Hainan 30 14 3 Guangdong 26 2 Jiangxi 23 5 1 Shandong ______22 4 2 Liaoning 17 4 Heilongjiang 15 8 3 Fujian 6 1 Xizang N.A. N.A. N.A. * indicates less than 0.5%. N.A. indicates not available. Source: Chinese authorities. 38  Table 4.4. Poverty-Related Disease Conditions, China, 1990 1/ Cost-effective Percent of Condition intervention available 2/ total disease burden, 1990 (DALYs) Infectious and parasitic diseases Tuberculosis Short-course chemotherapy 2.0 Diarrheal diseases Monitored oral rehydration treatment 1.8 Measles, polio, pertussis diphtheria, tetanus Immunizations 1.1 Intestinal nematode infections and anemia School-based deworming 0.7 Other infectious and parasitic diseases Acute chemotherapy 1.2 Lower respiratory infections Early identification and antibiotic 5.7 treatment Maternal conditions Emergency obstetric services with 1.2 prenatal care Perinatal conditions Obstetric care from trained personnel 4.9 Nutritional deficiencies Protein-energy malnutrition Breastfeeding; prevention and 1.0 treatment of childhood infections Vitamin A and iodine deficiency Supplementation/food fortification 0.4 Anemia Supplementation/food fortification 3.2 Total 23.1 1/ This list is intended to highlight only the main disease conditions that are almost entirely conditions of the poor. In addition to these conditions, the poor also suffer disproportionately from many others, such as diabetes, chronic obstructive pulmonary disease, sexually-transmitted diseases, and HIV infection. (Although existing projections of the STD/HIV complex's importance suggest little problem in China (Annex 3, Table A3.6), increasing clinical and public health reports suggest that these epidemics could become serious threats.) 2/ Cost-effective interventions refer to interventions costing less than $1 50 per disability-adjusted life year (DALY). Source: Annex 3, Table A3.4; World Bank 1993a. 39   5. IMPLEMENTING REFORMS IN PRICING AND PLANNING 5.1 Efficiency is another basic dimension of performance in the health sector. Government policies can have a major impact on whether health spending is allocated to the "best buys," whether services are produced at least cost, and whether care is clinically appropriate. Well- designed incentives are key to efficiency. The present pricing structure in China produces major distortions and inefficiencies by providing incentives for excessive and inappropriate care. In addition, the parallel public delivery systems produce considerable waste in tenns of excess capacity and idle resources. This chapter focuses on pricing issues and on planning and coordination of infrastructure and manpower. Major reforms in these areas could significantly improve the overall efficiency of the health sector in both rural and urban areas. Chapters 6 and 7 then turn to the issue of developing and strengthening health insurance or risk-pooling mechanisms in rural and urban areas, respectively. Price Distortions in the Health Sector 5.2 Prices for health services in China are set under guidelines established by the Price Commission, often at levels well below costs. Price Commission officials aim to make the prices high enough to protect and develop the services provided, yet low enough to assure affordability to the users. During the 1 960s and 1 970s, the government tried to increase access to health care by reducing the prices of medical visits and hospital days to levels that a poor farmer could afford. Prices for most medical services in China are still below costs, especially for services with a large labor input. The prices of high technology diagnostic tests, however, have been set far above costs to offset losses on other services. The system thus has two pricing extremes: most services are priced too low, leading to under-the-table payments to physicians and other problems. At the same time, high profit margins on drugs and diagnostic tests encourage overprovision in those areas. For example, now that village doctors' income depends on fees charged for drugs, injections, and diagnostic tests, these doctors have strong incentives to overprescribe drugs and tests. This drives up health spending without improving health. 5.3 At the heart of the issue of pricing policy, as it affects goods and services sold by the health sector to patients, is China's "Yellow Book" price list--a detailed listing of thousands of medical procedures, services, and diagnostic tests, which sets the price of each. Despite double- digit inflation in recent years, these listed prices are rarely updated in many provinces, thus fixing most health service prices at extremely low levels. 5.4 The difficulties caused by these pricing policies can be clearly seen in the hospital setting. Most Chinese hospitals charge patients (or the GIS and LIS insurance programs they belong to) for each item of service rendered and drug given; about 85 percent of revenues come from these charges, which are fixed and published in the Yellow Book. Hospitals also receive a provincial government subsidy for personnel wages (excluding supplementary wages and bonuses), and some funds for capital investment. 41  5.5 Faced with this irrational price structure, hospital managers use profitable products to cross- Box 5.1. The Leveraging Effect ..i subsidize under-priced products. Unfortunately, hospitals have to over-sell a high volume of profitable When hospitals have to generate services to generate a small profit. This "leveraging profits to cross-subsidize services: effect" is described in Box 5.1. Besides distorting the operating at a loss, the amount dtht patterns of medical services, the leveraging effect can be generated depends. on the. increases total health costs. profit margin of particular services, For example, drugs have an allowed mark-up of 15 percent. hi order t Pricing of Health Services. : E : generate 1 00 yuan of revenues above cost, a hospital has to increase sales 5.6 Table 5.1 shows the average costs for fairly of its drug prescriptions by 666 yuan. routine hospital procedures in Shanghai, using modern This "leveraging effect" contribues cost accounting methods, and the official prices for to rapid inflation of health: i insured and uninsured patients. The actual costs for the expendituresi procedures are two to four times greater than the allowed fees for patients paying out-of-pocket, and two to three times the allowed fees for insured patients. 5.7 In contrast to these routine Table 5.1. Comparison of Production Costs and procedures, the price of newly- Fees for Selected Procedures in Shanghai Hospitals, introduced diagnostic tests has been 1989 set at a level high enough to (in 1993 yuan) encourage their rapid adoption. A Allowable Fees comparison of the cost of a CT scan Procedure Average Insured Self-Pay with allowed fees in Shanghai and Costs Tianjin shows the large profit from Appendectomy 118 56 28 this test (Table 5.2). Cataract removal 142 44 22 5.8 An unintended consequence Gastrectomy 326 292 97 of this pricing policy has been that urban hospitals, and even county and Cholecystectomy 199 111 56 township hospitals and health Exploratory 167 56 28 centers, have come to see high laparotomy technology equipment as their financial salvation (Box 5.2). These organizations now routinely Cesarean section 140 Ill 56 organize investor groups to buy such Source: The Chinese Journal of Hospital Management equipment. They borrow from 1993;9:55 banks and often sell investment shares to staff members to purchase such equipment. Sometimes hospitals lease equipment from international equipment suppliers, with the lease payment set at a percentage of gross revenues generated from the use of the equipment. A hospital's reputation now is said to depend on possession of the latest equipment, be it computerized tomography, magnetic resonance imaging, fetal monitoring, upgraded intensive care units, burn units, or any of the wide variety of diagnostic techniques that manufacturers have successfully brought to market. China is in the midst of a diagnostic equipment race involving most of its hospitals and many health centers. 42  Box 5.2. The High Technology Equipment Race In Jurong County Hospital in Jiangsu province, the government subsidy now accounts for only a small part of hospital income: it declined from 17 percent of revenues in 1985 to only 2.2 percent by 1993. As a result, the hospital has been forced to collect funds from other sources such as drug sales and tests using high technology equipment. In 1993, the hospital received capital investments of 0.2 million yuan from the local government, borrowed 2.2 million yuan from local banks, and raised another 0.6 million yuan from hospital staff to buy a CT scanner and other new equipment. About 1,000 patients were diagnosed by CT in 1994, bringing the hospital 0.35 million yuan in revenue that year. The income from high technology diagnostic testing (mainly CT) more than doubled from 1993 to 1994. Although patient volume grew only 2.1 percent, the number using high technology equipment increased 50 percent. Total outpatient fees increased 168 percent (Jin Shuigao 1995a, p. 21). These stories are being duplicated throughout China--first in the somewhat richer coastal provinces and, more recently, in the interior. Even Guizhou province, which has the lowest per capita income of all of China's provinces, has seen an increase in high technology medicine. Table 5.2. Comparison of Production Costs and Fees for Body CT Pricing of Scan, 1988 (1993 yuan) Pharmaceuticals Cost Allowable Fees 5.9 D prices at Fixed Variable Total Self-Pay Insured the retail level are Patient controlled by the State Shanghai 109 47 156 181 362 Pharmaceutical Agency. Tianjin 80 32 113 181 362 The price structure Tianjin 80 32 113 181 362 ~~~allows mark-ups of 15 Source: Chen Jie 1994, pp.4-5. allows ar-pof 15 percent at both the wholesale and retail level. Hospitals have an incentive not only to overprescribe drugs they purchase from manufacturers, especially expensive drugs, but also to manufacture their own products to maximize the permitted margin. Patients are frequently treated with intravenous drip solutions of glucose, vitamins, antibiotics, and other drugs. In almost all cases, this does not constitute justified medical practice, but is done to maximize profit. Village doctors earn almost all of their health-related income from drug sales and injections, with injections preferred over oral prescriptions in order to maximize revenue. 5.10 Given this incentive structure, it is not surprising to find that spending on drugs accounts for 52 percent of all health spending in China, compared to an average of 14 percent in OECD countries and 15-40 percent in most other developing countries.I This is not only inefficient, but the inappropriate use of drugs can generate high health costs (for example, from the development of antibiotic resistance). The high share of spending on pharmaceuticals in China is in part due to the fact that the price of pharmaceuticals is relatively high compared to many other health inputs, such as labor. Nonetheless, there is considerable evidence of overprescribing. 43  5.11 There are several ways to combat overuse and misuse of pharmaceuticals: * Take the profit out of prescribing. If hospital payments were case-based, they would be independent of the number of drugs and tests prescribed. * Incorporate high copayment levels for outpatient drug prescriptions in insurance benefit packages to reduce excess demand. * Educate providers and consumers. Some countries give providers prescribing handbooks and educate the public on the appropriate use of drugs and the negative health impact from their overuse. * Develop and enforce hospital formularies/essential drug lists to guide cost-effective prescribing. * Monitor prescribing patterns and provide feedback to physicians. Core indicators of appropriate prescribing patterns include the average number of drugs prescribed per patient encounter, the percentage of drugs prescribed by generic name, the percentage of patient encounters for which an antibiotic is prescribed, the percentage of encounters in which the patient receives one or more injections, and the percentage of drugs prescribed that are not included on the relevant essential drug list or local formulary for that level of care. 5.12 On the issue of prices more generally, previous analyses emphasized the health sector problems caused by pricing policies that distort supply and demand (World Bank 1992a). The standard advice in such situations is to move toward marginal-cost pricing, even if fees continue to be fixed. This step has been given extensive consideration by the authorities in China, but has not yet been adopted (see background papers prepared for this report, especially the papers of Zhao Yuxin, Cai Renhua, Chen Xiaoming, Meng Jianguo, Hu Haobo, Hu Shanlian, and Hou Yan and Zhou Heyu, all prepared in 1995). Senior policymakers in China seem to accept price reform in principle, but fear that jumps in price could set off extremely negative reactions. 5.13 A major study of costs and prices is needed to lay the basis for reform. Such a study should avoid attempting to revise thousands of prices in detail. Instead, it should focus first on gradually bringing prices closer to costs. Some prices, especially those for services with high labor content, would be raised, while others would be lowered. Second, the study should analyze methods of defining and pricing broad packages of care, in line with provider payment reforms discussed in the following chapters on risk pooling in rural and urban areas. Price reforms must include a mechanism, perhaps on an annual basis, to respond to inflation. Pricing Reforms and Reallocation of Government Spending 5.14 Chinese government spending on health is both inequitable and inefficient. Most government spending on health is directed at the hospital sector through MOH hospitals, including township health centers (4.5 billion yuan in 1993) and traditional Chinese medicine facilities (.8 billion yuan). As discussed earlier in this report, relatively little is devoted to more cost-effective public health (1.3 billion yuan) and maternal and child health services (0.3 billion yuan). The government insurance system, in addition, spends almost all of its monies on the hospital sector (11.5 billion in 1993). Government subsidies to hospitals are a small share--no more than 15 percent--of total hospital operating costs. The rest of hospital operating revenue 44  comes from out-of-pocket payments by individuals or insurance payments. The government currently subsidizes the basic salaries of MOH employees, which are supplemented by bonuses paid by the hospital. 5.15 The Ministry of Health has recently argued for more public subsidies to the hospital sector. This would only exacerbate the existing misallocation of resources, however. Those who benefit from the present subsidies are middle-income urban residents and those insured under the GIS and LIS systems. The first priority for government resources should be public health and cost-effective care for the poor. 5.16 Over the medium term, reallocating public spending away from hospitals towards other priorities would be facilitated by the reforms proposed in this report. Price reform would better align allowable fees with marginal costs so that hospitals would have less need to provide excessive diagnostic tests to cover loss-making activities. Ongoing urban insurance experiments are already trying alternative provider payment approaches. Provider payment methods need to encourage efficiency, and ensure that efficient hospitals can recover their costs. This is discussed in more detail in Chapter 7. In addition, public hospitals need to gain greater control over their personnel so they can operate efficiently. Public hospitals in China already have considerable autonomy over their budgets, investments, and fee collection, but not their personnel. Instead, the government assigns new medical graduates to public hospitals without sufficient regard for staffing needs. As a result, some hospitals have too many doctors and too few nurses, while others have too many personnel in total. Planning and Coordination of Infrastructure, Manpower, and Health Services 5.17 China has at least three separate vertical systems involved in the planning, financing, and organization of urban hospital facilities--the public health system, the state-owned enterprise system, and the traditional Chinese medicine system. Each vertical system protects its own institutional interests, and has little incentive to coordinate with others. Moreover, the public health hospitals are owned and managed by different levels of government--national, provincial, and county. In a major city, three major hospitals might be located quite close to each other, one operated by the central government, one by the provincial government, and one by an SOE. A few kilometers away, across the city line, there could well be a county hospital offering similar services. 5.18 In rural areas, the collapse of the CMS weakened the referral and supervision chains that had existed between the village, township, and county-level health organizations. There is duplication of supply between the Maternal and Child Health (MCH) and Family Planning (FP) facilities as well as between Western medicine and traditional Chinese medicine facilities. This duplication strains the government's health budgets and misuses scarce resources. This problem is likely to worsen as the Family Planning Commission extends its facilities to the township level, creating more duplication and a greater drain on government funds. Some of these problems are illustrated in Box 5.3 using the example of maternal and child health services. 45  Box 5.3. Problems of Rural Health Delivery: A Case Study of Maternal and Child Health Programs A look at problems with matemal and child health services in rural China provides insights into overall problems facing the rural health system.. Several health facility and household surveys, a cost- effectiveness study, and a program review conducted by the Ministry of Health (MOM) have all documented the need to improve the quality of maternal and child health services. Quality and Cost-Effectiveness of MCH Services Issues in the delivery of maternal and child health services mirror those of the rural health system overall. Because Maternal and Child Health (MCH) centers at the county level, like many other health providers in China, are now forced to generate a high share of their income from user fees, they have shifted the focus of their work from preventive public health measures to more curative functions for which they can charge fees. This effectively denies health care to a portion of the rural population, since most rural families no longer have health insurance and many cannot afford fee-for-service care. This has even, reduced the coverage of such. essential preventive programs as immunizations, while creating duplication of services and quality problems as well. In order to generate fees, MCH centers are now performing functions that might best be referred to the county hospital. Many MCH centers are investing in equipment to perform cesareaji deliveries, for instance, instead of referring patients who need this; surgery to nearby county hospitals. Not only does this duplicate the hospitals' services but such surgical procedures: should be concentrated in high volume centers, to the extent possible, where professional skills are more easily kept up to date. Throughout China, hospitals themselves have become increasingly dependent on profits from drugs and other services, particularly capital-intensive inpatient care for which public subsidies are highest. Treatmnent of childhood diarrhea--still a common problem in some areas of China--illustrates this problem. Anti-diarrheal treatment in China now often involves unnecessary hospital stays, routine intravenous fluids, and potentially dangerous drug treatrnents. Most cases would be better handled using basic oral rehydration, and appropriate antimicrobial treatments at outpatient facilities when necessary, The costs of such inefficient service are borne by three groups: the patients' families, .who pay more than necessary for treatment; the public sector budget; and patients who do not receive treatrnent because their families cannot afford it. . Organization and Coordination of MCH Services Another basic problem is the lack of functional coordination among the many different groups providing maternal and child health services. An organization chart of the MOH shows a number of parallel national programs at the central, provincial, prefectural, and county levels, converging at tha township hospital and village level. The Maternal and Child Health Department is generally responsible for maternal and child.. health policies, although the Medical Administration. Department develops hospital care policies, inctuding A MCH referral services at the provincial level and below. EPS under the Departmrent of Disease Control. develops policies on childhood immunizations and diarrheal diseases as well as training sites and materialsf for the management of childhood diarrhea. The Medical Science and Education Depart ent handles health workers' training, including MCH training, although MCH in-service training is designed and implemented by the MCH Department. The existence of these overlapping departnents and functions at all levels seriously constrains the development of consistent policies and guidelines, efficient training,. and quality MCH preventive and curative services. 46  5.19 China seems to have an adequate supply of total hospital beds, based on international standards. There appears to be overcrowding in tertiary (level III) hospitals, but the average length-of-stay of 15 days for patients in these facilities is excessively long by international standards. With policy changes, many of these patients might be adequately managed at lower levels and the average length-of-stay might be reduced. Currently, secondary hospitals and township health centers have low occupancy rates. China needs a rational hospital referral system to make appropriate use of its hospital resources. 5.20 Regional planning could improve coordination between services. Regional planning can be used to develop approaches to disease control, to establish referral systems, and to guide capital investments in order to avoid duplicative and fragmented systems. Regional planning efforts should involve all relevant actors, including GIS, LIS, traditional Chinese medicine facilities, medical schools, and the MOH. As urban insurance centers are expanded, they would also be a major stakeholder in regional planning efforts. Plans should cover a large enough population base to support comprehensive health services. Box 5.4 describes how regional health planning has worked in three prefectures. 5.21 Regional planning bodies might be complemented by central and provincial government efforts to better control investments in facilities, expensive equipment, and the mix and number of health personnel. Once expensive and sophisticated facilities are established, they have to generate sufficient revenues to fund their operations, which raises health care costs in the long term. Many countries have realized too late that they have too many hospital beds or too many physicians, especially specialists, which also puts upward pressure on health costs. Other countries are struggling to reform their health system without sufficient numbers of well-trained hospital managers. 5.22 Many countries have found that governments need to play an active role in controlling the growth of health infrastructure, both capital and human, since investments in infrastructure have such long term and powerful impacts on health service delivery and costs. In the Netherlands, for example, hospitals must apply and receive approval to purchase specific types of equipment and technology or to provide certain specialized medical services. This regulation has effectively prevented an oversupply of medical technology. Of course these policies also have drawbacks. Supply-side controls reduce innovation and restrict market entry of, for example, lower-cost providers of specialized medical services. The Netherlands' success has been largely attributed to its severe sanctions: hospitals may be fined, the service may be closed down, and/or insurance companies may refuse to reimburse for the service. Many countries have limited the number of health specialists, either by limiting training opportunities or by restructuring physician payment, to lessen the financial incentives to specialize. China might start first with efforts to control the proliferation of high technology diagnostic equipment, both through controls on the amount of equipment and through provider payment incentives. 47  Box 5.4. Regional Planning for Health Regional health planning began in China through the Integrated Regional Health Development Project. Each of three prefectures--Baoji in Shaanxi, Jiujiang in Jiangxi, and Jinhua in Zhejiang provinces--agreed on a five-year development plan. Each regional health committee includes government representatives involved in finance, planning, civil works, and health. These regional planning efforts have led to better disease surveillance and a new TB control strategy. An:anti- smoking campaign in schools and improved training and outreach for MCH services are in place. Ambulance services now function with radio networks in the provincial and regional cities. Managers accepted centralization of the high technology diagnostics in one hospital to serve others. Some resources moved down from hospital facilities to county and township levels. Patient: physical rehabilitation and care for the mentally ill improved. There is better in-service training for..... medical and health workers, and stronger networks for assessment and maintenance of equipment. The experiments in these three regions have attracted increasing attention from the senior leaders. One of the three prefectures, Jiujiang, was also chosen as one of two sites for the demonstration. insurance project described in Chapter 7. 48  6. OPTIONS FOR EFFICIENT RISK POOLING IN RURAL AREAS 6.1 In addition to strengthening funding for public health (Chapter 3), improving the poor's access to health services (Chapter 4), and introducing system-wide measures to improve efficiency (Chapter 5), China needs to improve risk-pooling mechanisms in both rural and urban areas. Pooling the risks of medical expenditures is desirable because in China, as in other countries around the world, health care expenditures are highly skewed: about two-thirds of medical costs each year are concentrated on the 10 percent of the population with the highest health expenses. In China's rural areas, the main insurance issue is how to best provide some form of risk pooling for the perhaps 700 million rural residents who have lost their access to prepaid health care since the collapse of the rural Cooperative Medical System (CMS) in the early 1980s. Nearly 90 percent of farming households now pay out-of-pocket for almost all of their health services. Publicly-run health institutions receive minimal government subsidies and so charge high user fees, even for emergency services. If China allows present trends to continue, the majority of the rural population will have to continue to finance health services, including catastrophic care, out-of-pocket and pay on a fee-for-service basis. These fee-for- service payment systems will also tend to escalate health spending. 6.2 The government can play an important role in the development of health insurance by providing a policy framework that would, ideally, (1) encourage the establishment of risk- pooling mechanisms to meet the population's demand for insurance; (2) minimize the well- known market failures that occur in insurance markets, and (3) promote forms of provider payment and health care organization and delivery that contain costs and promote efficiency over the long term. 6.3 How can China best encourage risk pooling for the general rural population? There are several possible approaches to the collective finance of a health benefits package that includes catastrophic care. * General revenue.financing is an approach taken in many countries, from Sri Lanka to Sweden. Under this approach, the publicly-financed health system, organized either nationally (e.g. Britain) or subnationally (e.g. Canadian provinces) finances almost all health services. In OECD countries with general revenue-financed systems, some of the population typically purchases supplemental private insurance for coverage of additional benefits such as private hospital rooms. While the Chinese central and provincial governments do provide some small subsidies for publicly-operated health facilities, increasing these subsidies to a level sufficient to effectively provide insurance coverage to the 800 million rural Chinese appears prohibitive, at least over the short to medium term. * Mandatory social insurance, funded by a wage tax or premium, is another approach that has been adopted by many countries, and is used in urban areas in China (the LIS system). But this has limited practical feasibility in rural areas in China today because social insurance relies on employers to enroll beneficiaries and collect contributions, and peasants are largely self- employed. In addition, China lacks the institutional and organizational capacity to manage large social insurance programs that would cover hundreds of millions of beneficiaries in rural areas. 49  * Private voluntary insurance is another approach to risk-pooling, but it presents several problems. International experience has demonstrated that the private insurance market does not emerge to supply health insurance in developing countries, except for the most affluent urban populations (Musgrove 1996). Even if that were not the case, private insurance is not an equitable or efficient approach to insuring basic health benefits because of information asymmetry and selection bias. Risk selection by insurance companies leaves the disabled, elderly, and less healthy population uninsured. Countries such as the United States are developing regulatory mechanisms to partially address these problems, but that regulation requires sophisticated administrative and institutional capacity. * Community-financing schemes provide collective health financing for entire rural communities. In community-financing schemes, funds generally come from three sources: households, government, and local industries. The local community--not the government-- organizes and manages the financing and delivery of services on behalf of the consumer (Hsiao 1995d). Analysis of ongoing community-financing schemes in China shows that they can reduce inappropriate use of drugs, improve quality of services, and reduce overall service costs--in part by encouraging service provision at lower-level health facilities wherever possible. (Study of Thirty Poor Counties--see China Network 1996; Jin 1995a; Liu Yuanli and others, 1996). Because community financing can promote universal or near universal coverage at the local level and efficient service delivery without being a major drain on government funds, community financing appears to be a more promising option for risk pooling in rural China over the medium term than the others options reviewed above. 6.4 This chapter looks at the ongoing and past experience of risk-pooling schemes in rural China, and examines options for reestablishing some type of risk-pooling arrangements under current economic conditions, given financial and institutional capacity in China. While China has much experience with community financing, particularly under the CMS, it is unclear how well community financing will work on a large scale today, given the reforms that have changed the economic organization of the countryside and given the size and heterogeneity of China. Therefore, this approach would need to be phased in and systematically monitored and evaluated. As noted in Chapter 4, such community-financing schemes, if they prove feasible, could also be used to channel subsidies for health care to the poorest rural residents. The Rural Cooperative Medical System in China: 1960 to 1983 6.5 China pioneered rural community-based health financing with the rural Cooperative Medical System (CMS), which operated under the agricultural commune system in the 1960s and 1970s. Under the communal system of agriculture, communes took in all farm revenues and paid them out to individuals and households on the basis of work points provided to the commune. Barefoot Doctors received work points for their medical work and, in that way, public health services were financed by the townships and villages with little or no additional subsidy from higher levels of government. Barefoot Doctors delivered free preventive and primary care services at the village level. Patients typically paid a coinsurance fee for drugs, secondary services, and hospital services, which was higher for inpatient services. Since CMS was largely financed by a village's communal welfare fund, the benefit package varied depending on the wealth of each community. The poorer communities often could only afford to cover primary care services and did not cover inpatient services at county hospitals. 50  6.6 The CMS system has since collapsed in 90 percent of Chinese villages. The demise of China's CMS system can be explained by several factors. First, China shifted from a system of agricultural communes to an individual household responsibility system beginning in the late 1970s. The communal welfare fund disappeared, and with it went the major source of CMS financing. The source of support for the Barefoot Doctors was gone, as were CMS funds to pay for drugs and other health services. 6.7 As a result of the disbanding of China's communes, just 40-45 percent of China's villages were still covered by CMS by 1983. At about that time, an ideological shift prompted some high government officials to declare that the remaining CMS programs should be abolished. Thus, most communities that still had CMS in 1983 subsequently disbanded their programs by the mid-I 980s. 6.8 Even before its demise, patronage and corruption had weakened China's CMS. The CMS was controlled and managed by local officials, some of whom used their power for selfish gains. As a result, farmers lost confidence in the government-run CMS and refused to pay premiums once the system became voluntary after the late 1970s. This experience underscores the importance of effective organization and management of any new community-financing schemes, and the need for an adequate financial base. 6.9 The government did not replace the CMS with a new health-financing structure, but instead adopted a laissez-faire policy. In response, many communities voluntarily designed their own new funding mechanisms. Many villages fell back on a system of fee-for-service payments. In some villages, the poorest farmers, who could not afford to pay for services, could ask for support from the village welfare fund. That fund was maintained by a specific tax ofjust under 5 percent of the village's net output for farm production. One-fifth of this agricultural tax was designated for welfare assistance and to defray health care costs of those in need. China's Recent Experience with Community Financing 6.10 China can also look to its more recent experiences with relatively small-scale community-financing schemes to identify sound policy options in this area. These include the completed Sichuan Rural Health Insurance Experiment (see Mao 1995; Sine 1994; Cretin, Williams, and Sine 1995) and the ongoing World Health Organization (WHO) Fourteen Counties Study of Community Financing (Yang 1995). In addition, the Study of Thirty Poor Counties produced a wealth of data on rural health finance frequently cited in this report (Luo 1995 and China Network 1996). Surveys done in preparation for the World Bank-supported Rural Health Workers Development Project also revealed important information on the prevalence and operation of community-financing schemes in China today. These experiments and studies are described below. Sichuan Rural Health Insurance Experiment 6.11 The Sichuan Rural Health Insurance Experiment, which took place in 1989-90, assessed the potential impact of insurance and coinsurance on the demand for health care and on the likely costs of providing services under an insurance regime. This experiment involved a sample of twenty-six villages from two counties and 40,443 individuals (Sine 1994). 6.12 Three insurance benefit plans were implemented with varying reimbursement rates for inpatient and outpatient services. Premiums were set at 1.5 percent of average income. Insured 51  individuals could visit facilities at the village and township level, but could only visit county hospitals in an emergency or with the approval of the township health center. The experiment showed that households were willing to join such a scheme--over 90 percent of households in the test areas voluntarily joined the program and 95 percent voluntarily re-enrolled after the first year. It also showed that administrative costs could be kept low (8 percent of total reimbursements). In addition, the study found that: * Coinsurance (the requirement that the patient pay part of the cost of health services at the point of service) exerted a significant negative effect on demand for care across different population subgroups. No interactions were found between the effect of coinsurance and age, income, or health status. * With the exception of one village, users surveyed stated a high degree of satisfaction with the insurance arrangement (Mao 1995, p. 16); * Services were used less when there was no functioning village health station, underlining the importance of an adequate supply of basic services (Mao 1995; Sine 1994); and, * As in other countries, a small proportion of the population accounts for a large share of total health expenditures, underscoring the need for catastrophic insurance. Approximately 11.5 percent of the covered population incurred 70 percent of the total health expenditures. WHO Study of Fourteen Counties 6.13 An ongoing WHO study is assessing China's existing community-financing schemes to learn how to improve forns of organization, financing, and service delivery. The fourteen participating counties are located in Beijing, Henan, Jiangsu, Zhejiang, Jiangxi, Hubei, and Ningxia. In each county, a research team interviewed 540 households and surveyed health services. The study is in an intermediate phase. 6.14 The study has found that a typical community fund might collect 5 yuan per person from families, I yuan per person from the village's social welfare fund, and I yuan per person from the township. Coverage typically involves a deductible (for example, 100 yuan) paid fully by the patient, with copayment on expenditures above the deductible. Coverage of drugs is limited to 120 kinds of medicine, including traditional Chinese medicines, with a maximum reimbursement limit set for diagnostic tests. Study of Thirty Poor Counties 6.15 The Study of Thirty Poor Counties was conducted in 1993-1995 by a network of Chinese universities and Harvard University. It found that 16.5 percent of villages surveyed still maintained some type of community-based health finance schemes, covering 11.6 percent of the sampled population. About two-thirds of the schemes covered only primary care services at the village level and one-third covered comprehensive services ranging from primary care to inpatient services. All of the benefit structures incorporated coinsurance features and often set very high copayment rates for inpatient services. Table 6.1 summarizes the prevalence and benefits of community financing in the counties surveyed. 52  Table 6.1. Prevalence and Benefits of Community Financing in Thirty Poor Counties, 1993 Number of Percentage of Percentage of Type of Benefit Villages Covered Villages Covered Population Covered Comprehensive 29 5.1% 4.4% Primary Care Services Only 59 11.4% 7.2% Total 88 16.5% 11.6% Source: Study of Thirty Poor Counties (China Network 1996). Note: Comprehensive benefit refers to cooperative medical schemes that reimburse 30 to 100 percent of hospitalization fees for township and county-level hospitals, as well as 50 to 100 percent of outpatient fees. Primary care services refers to coverage of fees (or discounted prices) for most village-level services, with fees at the township and county levels paid out-of-pocket by patients. 6.16 The study found that the most prevalent type of community fund management was by village committee or by the village and township jointly. Table 6.2 presents the existing management arrangements in the thirty counties surveyed. Table 6.2. Management of Community Financing, 1993 Benefit Type Management Form Comprehensive Primary Care Services Only Township Government 17.2% 3.4% Township Health Center 20.7% 6.8% Village and Township Jointly 20.7% 10.2% Village Committee 34.5% 47.5% Village and Township Doctors 6.9% 32.1% Source: Study of Thirty Poor Counties (China Network 1996). 6.17 The study appears to provide strong support for the reestablishment of community- financed health care. As part of the study, 11,044 randomly selected households were asked about their preferences for such organized financing schemes. Seventy percent of the households who were not covered by community financing or child immunization and MCH prepayment schemes responded that they would like to see an improved CMS-type scheme reestablished, while 88 percent of those covered by community-financing schemes stated they would like them to continue. Among the 70 percent of the uncovered population that favored reestablishing community- financing schemes, about one-fourth preferred to see such a scheme managed by the village, one- fourth preferred it to be managed by the township, and the remainder preferred joint management by the township and village or by the health facility. 6.18 Another indication of management preferences was obtained in the course of preparing the World Bank-financed Rural Health Workers Development Project. More than 1,000 village cadres in five provinces were asked about their preference for the management of community- 53  financed health schemes. Close to 60 percent of the village cadres preferred management by a village committee and roughly one-fourth preferred joint management by township and village. 6.19 In most rural areas, particularly poor areas, it is not possible to derive adequate revenues for any organized financing scheme solely from households. Funding must come from multiple channels. Both the Study of Thirty Poor Counties and the preparation work for the Rural Health Workers Development Project obtained information on the source of financing for existing community-financed health plans. The two studies found that approximately one-half of the revenue cane from household contributions, about one-fourth came from village social welfare funds, and approximately 10 percent came from the government. The detailed results are shown in Table 6.3. Table 6.3. Percentage of Community Financing by Source Village Sources of Financing Government Social Househol Other Welfare d Fund Funds surveyed in 30 County Study (1993) 16.1% 20.3% 48.1% 15.5% Funds surveyed in 5 Province Survey 8.0% 30.3% 58.7/o 3.0% (1991) I _ _ _ I I_ _ I_ _ I Source: Study of Thirty Poor Counties (China Network 1996) and World Bank Rural Health Workers Development Project Preparation. 6.20 Table 6.4 presents information on the prevalence of community-financed health schemes and their benefit structures in Hebei, Shanxi, Fujian, Guizhou, and Henan provinces. The relatively poor province of Shanxi has the highest percentage of coverage--close to two-thirds of the villages maintained some form of community financing. But another poor province, Guizhou, has very few villages with community financing (only 0.8 percent). In these poor provinces, the schemes were largely financed by household contributions; benefits only covered primary care services because of the relatively small contributions that poor households were able to make. Policy Changes in 1994 6.21 The Chinese government announced a new policy direction for the financing of rural health care in a front-page article in the People's Daily, on July 2, 1994. This new policy appeared to draw in part on lessons from China's earlier experience with CMS and its other experiences in community financing. The government called for the development of community-based schemes to fund and organize health care for the rural population, guided by the following principles: * The government's role is to establish policy and provide leadership; * Each community organizes its own collective financing for basic health care; * Funding will be derived from multiple sources (government, collectives, and individuals); * Priority should be given to covering preventive services; 54  Table 6.4. Prevalence and Benefits of Community Financing in Five Provinces, 1991 Of which Province Number of Percentage of Comprehensive Primary Care Villages with Villages with Coverage Servkes Only Community Community Financing Financing Hebei 3992 13.1% 42% 58% Shanxi 4727 65.6% 15% 85% Fujian 512 6.3% 25% 75% Guizhou 160 0.8% 6% 94% Henan 1590 6.2% 7% 93% Total 10981 12.2% 24% 76% Source: World Bank Rural Health Workers Development Project Preparation (World Bank 1993b) Note: Data from the Study of Thirty Poor Counties indicates that almost 80 percent of the "services and drugs coverage" category is comprehensive coverage. Therefore, for the purposes of this table, the Five Province survey data was recategorized with "services and drugs" plans counted as comprehensive coverage and the remaining categories (services only, drugs only, other) counted as coverage of primary care services only. * The schemes and benefit package should vaiy according to community conditions and economic capacity; and * The schemes should be supervised by, and accountable to, the people. 6.22 The 1994 policy announcement emphasized that community financing should be voluntary rather than compulsory at the individual level. The government also suggested three management models with varying degrees of community control. 6.23 Although these health funding principles are already supported by the State Council, there is wide variation in the interpretation of the policy in different provinces. There is some confusion, for example, over the Ministry of Agriculture's decree that the tax burden on farmers must be reduced--some provinces view this decree as constraining any community-financed schemes that involve household contributions. As a result, the government's support of community-financed health care has had little impact to date. 6.24 China's villages and townships would have stronger incentives to develop community- based collective financing if the central government clarified and elaborated its priorities. The government could also provide technical assistance and perhaps a subsidy for the establishment and operation of community financing. There are strong economic arguments for such a subsidy, in order to encourage equitable and efficient risk-pooling mechanisms in rural areas. The basic elements needed to establish such equitable and efficient community financing are set forth below. 55  Making Community Financing Work 6.25 An affordable and acceptable benefits package. The first major issue involves deternining what benefit package and coinsurance levels are acceptable and affordable to most community residents. While economic principles such as cost effectiveness should help guide the design of the benefit package, the end product must have popular support, and people must be willing to pay their share of the required premium. The Chinese Network of Health Economic Institutions developed several potential benefit packages to test consumer demand and costs. The prototype test packages cover the full range of services, but with higher coinsurance rates for drugs, demand-elastic services, and inpatient hospital care. In aggregate, the proposed coinsurance rate was about 30 percent. The estimated cost for these illustrative benefit packages was 30 yuan per person per year in 1993 prices, which amounts to 5 percent of the disposable income of "modest income" rural households (see Box 6. 1). ' ' 6'1' Estimating theaCost and Content of H- l- ii0-g00|t siao and iLu 0(1996) jused data from household surveys, financial records of health ii0 0 -0 faciiies andresults fof thie Sichuan Rutal Health Experiment to model the likely utilization iand costsi ;0 0---of.a basic p:ackg of services under a: community-financing scheme. Since the majority of Sthe: rural "opulation}s curently uninsured, he provisiont of a basic benefit package would increase the 'Equanityoftservices! demanded. T o estimate the likely impact of this, Hsiao and Liu used demand: elasticity assuniptions Xfrots estimates from the Sichuan Rural Health Insurance experiment-that a 10 percen inl:0 crease int insurance coverage would increase outpatient expenditures by 6 percent and ; l-0inpatiet -exeditures: by: 4- percent. Thley: assumed that supply is perfectly Eelastic because China has:E l; an excess Ssupply of personnel and beds at thie village Xand township level. They also assumed thiat thei cormimuity-fiancinggscheme wvould pay inpatient services Eon a case-based payment, with payenelt - leveldetenined prospectively, rather than on a fee for service basis. This was assumed to generate 15 5percent savings eon inpatient costs.: The ll00 il;000 lebasic Xbenefits package modeled by Hsiao and Liu would cover services at the level of ; - vWage, tQwnship health center, and county ihospital. This package includes catastrophic care, but thei :0 defini{tioni of catastrophic care is limnited by whiat thie county hospital can provide in terms of its 0;: ' 'eq''ipment, 4rugs; and supplies, Xand personnel. (A county hospital, for example, is not equipped and i0 : -staffed to catrtyout open- heart- surgery.) Two prototype packages were developed, withi different E 0 Joev ofcoiitsurance (mediumr and high). The mediium level required a 30 percent copayment for " wnship health center outpatient fees, a 40 ipercent copaymnent for drugs, and a 3 5 percent - ioamn ffor inpatient chargs at thie county hospitals.i Thle high level required a 30 percenti " cpaymetfotwnship: health center outpatient fees, 50 percent fosr drugs, and 45 percent for: E: lll inpatitChares gat: the county hospitals. Ai :istop-loss level for coinsurance was built in tat 500 yuan. 1 The package wZith ahigh level of coinsurance was estimated to cost 28 yuan per person. The; i pacag t 4lwith a mediumi level of coinsurance was estimated to cost 31 yuan per person. i; iii ' -Some ecoomists argue thiat health insurnce should onl cover catastrophic expenses. In praci0e :0one raely finds health tinsurance Xarrangements that cover only catastrophic care, While It E -i' jes' notv make sensef to have Arisk pooling thiat excludes catastrophic expenses, there are sound0 i ; argumients for covering both cata;strophic and Xsonic noncattrphic expenses in a risk-pooling 0 airl-; gems8ent.0 For ex'ampl1e, if only gcatastrophic care gis covered under tinsurance, there are Xincentives ; 0 ti providcare in a hospital setting (t4 be eligible for insurance reimbursement) when less costly but ote --0onequally effective outpatient care might be available.a In addition, if only catastrophic care is 0 l '0proided,,consumners may 0have Wdisincentives for seekcing ealy treatment, although early treatmentibg0 Is tpicalymore ost-effective. f 000Dj:00t;if03:; 56  The Sichuan Rural Health Insurance Experiment and household survey responses indicate that people want coverage for a wide range of products and services, from drugs to village doctors to county hospitals. Furthermore, rural residents seem to accept coinsurance payments. Field tests could reveal whether households are willing to make voluntary payments that will cover the expense of providing these services. 6.26 Choice ofproviders. To promote consumer satisfaction and competition on the supply side, beneficiaries should be given a choice of primary care provider wherever possible. If beneficiaries seek services from qualified providers outside of those included in the scheme, the collective financing scheme could reimburse them at a reduced rate. 6.27 Universal coverage and adequate size of risk-pools. If possible, enrollment should be mandated at the local level to minimize adverse selection. In designing schemes, consideration should be given to an adequate population size for risk pooling at the village and township level. The population of a village generally averages 1,000 people--a relatively small size for pooling hospital expenses. A township generally averages 12,000 people, which is adequate for pooling the risk of hospital expenses. The size of the optimal risk pool needs to be weighed against declining accountability as the size of the pool increases. 6.28 Referral and supervision. Effective referral systems need to be established between village doctors, township health centers, and county-level hospitals. Provisions need to be included for monitoring quality of care. 6.29 Provider payment. The method of paying health workers affects overall costs and quality of care. Strong consideration should be given to replacing fee-for-service payments with a salary plus performance bonus for village doctors. Village doctors would be responsible for providing public health and basic preventive and curative services. This payment method would minimize incentives to overprescribe tests and drugs. It would also reduce administrative expenses by eliminating costs for claims processing. Capitated payment contracts might be made for hospital services so that hospitals would have a stronger incentive to provide services efficiently to the covered population. 6.30 Organization and management. Misuse of funds, favoritism, nepotism, and, in the worst cases, corruption will destroy the solidarity necessary to make community financing work. Ideally, community-based financing schemes would be established as non-governmental entities, with directors elected by those enrolled in the schemes. There must be frequent and periodic financial and quality-of-service reports. Under such a system, residents of a community would have a strong sense of control over their community's financing scheme. 6.31 Adequate financing. Sources of financing for such schemes can include households, rural collective enterprises, village social welfare funds, and central, provincial, and local government subsidies. In the Sichuan Rural Health Insurance Experiment, farmers paid 1.5 percent of their incomes as premiums to buy health insurance, in addition to making copayments. Rural collective enterprises voluntarily fund community health in some richer villages. Currently, about one-fifth of the village social welfare fund is spent on health. Local government subsidies would vary with the local conditions and with the priority given to health care by the local government. 6.32 Government Support. To provide an incentive for the establishment and operation of community financing without resorting to politically unpopular mandates, the central government 57  should consider providing technical assistance in the establishment of community financing. It could also consider providing a matching grant to supplement the contributions of households, rural collective enterprises, village social welfare funds, and local governments. The subsidy might only go to townships in which collective-financing schemes meet certain basic government guidelines. There is an efficiency argument for such a government subsidy, in that the incentive would promote the development of efficient risk-pooling mechanisms. For illustrative purposes, the central government financing might initially be 5 yuan-per-person for a 30 yuan-per-person benefits package for nonpoor townships. (Poor villages might receive considerably more as discussed in the previous chapter.) This would grow over time with income growth. If, by the year 2001, 120 million people in rural areas received the matching subsidy, the costs to the government might be 1.2 billion yuan (120 million people times a 10 yuan per capita subsidy in the year 2001). 6.33 China's experience with community financing indicates that it may be a promising approach to reestablishing risk-pooling arrangements for catastrophic medical expenses in China's rural areas. Depending on the design of the scheme, community financing can also promote efficient service delivery. Nonetheless, China is a large and heterogeneous country. While community financing has many advantages in theory, it may prove difficult to administer on a widespread basis. It may, for instance, be difficult for the community-financing scheme to collect premiums, administer contracts, and remain solvent, especially in poor communities without much administrative capacity. This approach, therefore, needs to be phased in with technical assistance and systematically monitored and evaluated. While some general guidelines are important, as discussed earlier, local experimentation and adaptation should also be encouraged. Key design elements that could be tested include: - Alternative benefit packages, including coinsurance levels, to test consumer demand and costs; -- Alternative provider payment arrangements (testing payment of primary care providers with a salary plus performance bonus and capitated arrangements for hospital stays and visits); e Alternative management models; and * Simple methods for monitoring quality and appropriateness of care, including pharmaceutical prescribing. 58  7. OPTIONS FOR EFFICIENT RISK POOLING IN URBAN AREAS 7.1 Health insurance currently covers about half of China's urban population. This is in sharp contrast to rural areas, where the vast majority of the population has no access to any form of health insurance for catastrophic care and must rely on informal risk-pooling mechanisms such as the help of extended families. While coverage is better in urban areas, there are still urgent problems in the urban insurance systems. The two formal insurance systems, the government (GIS) and labor (LIS) insurance systems, cover just 15 percent of China's total population, yet they account for 36 percent of China's health spending. There are many problems in the operation of the GIS and LIS, particularly in terms of financial sustainability, labor market impact, and efficiency. The two urban insurance systems are urgently in need of reform to avoid becoming an obstacle to economic modernization and reform to the state-owned enterprise system. Change will be difficult, however, given the many powerful groups with vested interests in these insurance plans. Gradually expanding insurance coverage to uninsured groups in urban areas will be an additional challenge. 7.2 GIS beneficiaries, particularly retired veterans but also civil servants and university students, will oppose any reduction in their benefits. Enterprises or government units that have relatively young workforces or few retirees oppose pooling their risks with employers of primarily older workers. Government hospitals oppose payment reforms that may reduce their revenues or force their closure or downsizing. Various ministries represent and defend conflicting positions. The Ministry of Finance is mainly concerned with containing government outlays. The State Pharmaceutical Administration wants to maintain and increase drug sales. The Ministry of Health wants to protect its hospitals' revenues. The Ministry of Labor defends the interests of China's state-owned enterprise (SOEs), which themselves operate thousands of health facilities. Closing hospitals and firing staff is just as unwelcome as closing a steel plant or a coal mine. It is not a viable policy option except at the margin. 7.3 These constraints require a balancing act. The government needs to promote policies that control beneficiary demand for health care, improve efficiency and reduce waste in hospitals to keep costs down, yet allow hospitals to earn enough to cover their costs. The urban health finance experiments described below--in Shenzhen, Shanghai, and Zhenjiang/Jiujiang--are identifying strategies for national reforms of the GIS and LIS and, eventually, for broadening urban insurance coverage. These experiments focus on improving incentives in the system, particularly on provider payment reforms and coinsurance rates. The most recent experiments, in Zhenjiang and Jiujiang, appear to hold the greatest promise as national models. As a result, in January 1996, the government decided to replicate the Zhenjiang/Jiujiang reforms in two additional prefectures, or cities, in each province, for a total of roughly 50 sites. 7.4 This chapter reviews current experiments in reforming the GIS and LIS in selected Chinese cities. It then examines the issues involved in expanding these reforrns and in extending insurance to other urban residents outside the GIS and LIS systems. 59  Experiments in Reforming the GIS and LIS Systems Mixed results from the Shenzhen Experiment 7.5 Shenzhen is a newly established city of three million people. One million are permanent residents and two million are temporary residents who tend to be young contract workers. In 1992, Shenzhen implemented a new social health insurance program with the dual goals of extending insurance coverage and reducing health costs. 7.6 All employers were required to enroll their workers in the program. The insurance program provides a two-tiered benefit structure. The first tier is an individual medical savings account. The employer deposits 6-10 percent of a worker's average annual wage (varying by age) into this individual account, to be drawn upon only to pay for medical expenses. If expenditures exceed the amount available in the individual savings account, the patient is reimbursed from the second tier of financing--the social risk-pooling fund. For these expenses, the worker must pay a 10 percent coinsurance fee with a "stop-loss" ceiling. This ceiling for each worker was set at 8 percent of the wages that the worker earned in the previous year. Beyond this ceiling, the social risk-pooling fund pays 100 percent of the worker's medical bill. At the end of the year, 20 percent of any balance remaining in an individual account is carried over to the next year and 80 percent of the balance is transferred to the social risk-pooling fund. 7.7 Shenzhen finances its social health insurance by an 8-10 percent payroll tax on employers, varying by the age-sex composition of employees. Workers do not contribute directly. Benefits for retirees are financed from pension funds, which pay 10 percent of the monthly retirement benefits to the health insurance fund as a premium contribution. 7.8 The payment system can be characterized as fee-for-service with a cap. Providers bill the insurance fund for each item of service. Caps are set separately for outpatient visits and hospital bed-days. On a quarterly basis, the insurance fund reviews the bills submitted by providers and calculates payments that include a bonus for bills below the cap and a penalty for bills exceeding the cap. For quality assurance, 5 percent is withheld until the provider passes certain quality performance standards. 7.9 The Shenzhen scheme is managed by the Shenzhen Bureau of Health Insurance, which is controlled by the city's Bureau of Health. The deputy director of the Bureau of Health serves as the director of the Bureau of Health Insurance, which has created a conflict of interest, described in more detail below. 7.10 The Shenzhen program has had difficulties meeting its stated goals of expanded coverage and reduced costs. Enrollment and premium collection were problematic. Joint ventures and privately-owned companies refused to participate in the compulsory social insurance program, which was established through regulation but not through a new law. The Bureau of Health Insurance found it almost impossible to enforce the regulation. Consequently, only 5 percent (150,000) of thie 3 million target population participated. Almost one-third of these enrollees (46,000) were previously covered by GIS. The majority of the remainder had been previously covered by LIS. Only 27,000 contracted workers (18 percent 60  of enrollees) in the new program were previously uninsured. The new program also enrolled some 8,900 retirees. 7.11 Shenzhen also encountered financial difficulties due to both adverse selection in enrollment and fraud. Employers who refused to enroll in the program disproportionately employed younger workers, as shown by the fact that the average age of enrollees (36) exceeded the city-wide average (27) by 9 years. There is also evidence that many employers decreased their premium contributions by under-reporting wages. The magnitude of such under-reporting is conservatively estimated at 50 percent. 7.12 Shenzhen also had difficulty controlling hospital costs. Since the Shenzhen Bureau of Health Insurance is largely controlled by the Bureau of Health, the health bureaucracy both manages and controls the financing of health facilities. Even though the insured population does not constitute a majority of patients, the hospitals view insured patients as a major source of revenue. Given the Bureau of Health's dominance over the insurance operation, the conflict between providers and the insurance administration was resolved mostly in favor of the providers, at the expense of cost control. 7.13 In short, Shenzhen's social insurance program could not maintain financial solvency with the Bureau of Health Insurance acting in the role of both funder and provider. Because of this conflict of interest, a change is underway to set up the health insurance administration as an independent agency. The social insurance program has also had difficulty enforcing participation by joint ventures and privately-owned companies under the compulsory insurance regulation. The Use of Global Budgets in Shanghai 7.14 The city of Shanghai is a metropolis with approximately 14 million people. Seven million workers are covered in principle, if not in practice, under the GIS and LIS insurance plans. Perhaps another 3 million people are covered as dependents under these plans. Shanghai faces problems similar to those of other urban areas--high health cost inflation, poor quality of services, a high proportion of uninsured residents, and little risk pooling. 7.15 About four years ago, Shanghai developed a yet-to-be-implemented plan to overhaul its health care financing system. GIS and LIS would be combined into a single city-wide social health insurance plan. Copayments would increase: insured workers would pay 15 percent of outpatient charges and 8 percent of inpatient charges, up to a maximum of 1,000 yuan and 2,000 yuan, respectively. The plan includes a medical savings account for individual workers (4 percent of wages) and a social health insurance fund (15 percent of wages) with the majority of contributions from the employer. The social insurance program will be managed by the Shanghai Bureau of Health. This Bureau will establish a department of social insurance, which will also handle the pension scheme. The individual accounts will be managed by employers, and workers can withdraw funds equivalent to the coinsurance they pay for health services. 7.16 The long-terrn goal is to expand coverage to workers who are employed by joint enterprises, smaller collective industries, and private enterprises. However, workers in the "Big Eight" industries (including railroads, airlines, coal, steel, and the postal service) would be permitted to set up their own systems. 61  7.17 Before implementing this reform, Shanghai adopted a new payment system to control hospital cost inflation. From July 1, 1993, through June 30, 1994, hospital costs rose 53 percent over the previous year. In July 1994, Shanghai implemented a global hospital budget in which total costs were permitted to rise no more than 24 percent and drugs expenditures were permitted to rise no more than 15 percent. Prices were also reformed: fees for visits and surgeries were increased to reflect more accurately the labor costs for these services; fees for CT and MRI services were reduced by 12-15 percent. The hospital continued to be paid on a fee-for-service basis. At the end of September 1995, Shanghai found that its global budget approach had kept hospital costs within the target rates from July 1, 1994, through June 30, 1995. Promising Preliminary Results from the Jiujiang and Zhenjiang Experiments 7.18 A demonstration health insurance project began in December, 1994, in Jiujiang and Zhenjiang, two medium-sized cities on the Yangzi River in Jiangxi and Jiangsu provinces, respectively. The State Council sponsors these demonstrations. These cities, each with about 2.5 million inhabitants, form the "cutting edge" of urban health finance reform, and as such they have been visited by more than 600 officials from other provinces and municipalities. 7.19 In both cities, newly organized insurance centers collect Box 7.1. Wage Tax Flows From Contributors to insurance payments from enterprises Individual and Common Accounts and public sector agencies, then commit these funds to individual and group accounts by formula (Box 7.1). There are both individual worker and collective accounts. The Cnklbutwr: Employers 10% Emplyees 1% centers draw on these accounts to pay medical fees. The personal medical savings accounts plus large 11% copayments should encourage moderation of patient demand yet Fns provide stop-loss coverage to protect Funds: against catastrophic medical bills Coordinating Individual (Box 7.2). The insurance centers Fund A% also introduced an essential drug list 5% 6 of 1,100 Western and 500 traditional Chinese medicines, and they reimburse only medicines from the list. 7.20 The Zhenjiang model includes two additional innovations. First, the formerly separate GIS and LIS systems have been combined into a single insurance center. Second, packaged fees are set per outpatient visit and per inpatient admission. The payment rates are established prospectively. 7.21 All enterprises in Jiujiang and Zhenjiang that had GIS or LIS insurance were asked to join this pilot study. By the end of August 1995, 95 percent of the eligible population in Zhenjiang had joined and 93 percent of the contributions had been collected. In Jiujiang, 90 percent of the 62  ..-. -17.2.. The Three Tiers of Financing leialth Services in the Jiujiang and Zhenjaug- ,- Experimentj . . . . n Both the Jiujiang-and Zhenjiang experiments finance health expenditures: through three tiers of financing. The first tier :is the individual account. The amount deposited into ttis account dep,en,ds:on.a wker's annual salary, A sum equivalent to. approximately 6 percent of tfie '-wd'er' annual salary is deposited periodically into this individual account, with some variation by. age bracket. .Thesecond tier consists of out-of-pocket expenditures: for medical bills that exceed the m,o,unt in a. workees inidividual account. .When a worker's health expenditures exceed the funds ,in theindividual account, the worker pays. a deductible of up,.to 5 percent -of annual income --bef,ore-ceiving reimbursement froom the third tier of financing. The third tier is the social risk-pooling fund, which is designed to insure workers aga'inst .e .inC:ial burden of catatr,ophic Illess. The larger the expense beyond- the individual -d,dctible)j the- more the social risk -pooling flid pays, with slightly, more'generous coverage-in Zhenjiazig~ ~ ~ ~ s-Poiik eeruscvea , . , . , ,j , ;.., .,. .. . .- ... -. . -. --. . . . . - .. . -- - .. - - - : . . --.. Consider,-for example a. worker with -n 'veag anual wage::of 6,n00 uan. f e ''..,.,w'cer inrs medical exnses in-a given6year that'ttaiess ta6- per,ent ofhis or an,a-l L.'.' ess thlani36. yuan), then thosee expenses- are 'npl,,y paid from 'e 'ndivid:aa '. ''c,'' t.' If.i,n..that yar tota mne,dicalexenitue o:-f 56 -yua 'are: crred, te-.individ- ' ..'c"ou'ntpays exO.uananmd he'remain,ing 20 yuan come f,o.m current acome, paid uaa deduct,i..,>ble.. In,.neithe of- these twases does, te social ri5kooing- fud. c'e aypart of the:-- dical bil'.oL 1-lowemer, if me ical expenies:totalng .I 'y are ,ince , the social risk' :,, pooin fun'd:-wil,l pay ',lmo.st 90 perce,n:tof the,bill. Fo, -thii caaophic-medical' ex...............en........ se'--- .qialn t$ o two and .a half years' .inoe--'*e s'o'ial: rispoig fu-.d -liit t' wo'e' :f'na-nci,al -los,s'to: about a quarter of his or her antnual saiaiyt .---, ---. goverrnment units and 88 percent ofthe enterprises had paid their contributions. In total, 415,000 people in Zbenjiang and 370,000 people in Jiujiang are enrolled in the new plans. 7.22 Results from the first year of operation appear to be positive, as summarized below, with evidence of successful cost containment and reasonable control of administrative costs. (A thorough evaluation is still needed to confirm these results, however.) * Coverage gaps, especially for teachers and for workers in deficit-ridden enterprises, have been eliminated; * The rates of over-prescription and reliance on expensive diagnostic tests were cut significantly; * The annual growth rate of aggregate hospital expenditures declined by 23-28 percentage points from the previous average annual rate of 33 percent between 1991 and 1994; * The quantity of services declined 9 percent and the bed-occupancy rate by 2 percent, when compared to the previous year; * Both inpatient and outpatient utilization fell for enrollees (Cai 1995a; Zhenjiang 1995; Yip 1996). 63  In early January 1996 the State Council decided to expand the Jiujiang- and Zhenjiang- style experiments to additional cities throughout China, even prior to formal evaluation. The State Council will evaluate the results in 1996. 7.23 There appear to be clear benefits from separating the funding of health services from their provision, as done in Jiujiang and Zhenjiang. This separation can encourage cost containment, efficiency, and service quality because the funding agency represents the interests of the consumers, not the health providers, and can bargain with hospitals and monitor service quality. So far, the Zhenjiang insurance center has focused on innovative payment methods for hospital services in terms of packaged fees. In the future, it must also monitor quality and appropriateness of care. In Jiujiang, the insurance center continues to pay hospitals on a fee-for-service basis, and tries to monitor appropriateness of care with retrospective claim review. Issues in Expanding the Jiujiang and Zhenjiang Experiments 7.24 Further evaluation of the ongoing Jiujiang and Zhenjiang experiments is needed, and systematic monitoring and evaluation need to take place as they are replicated in all provinces, to maximize the learning from all these experiments. Evaluation should include analysis of: 1) the impact of the individual accounts and deductibles on patients' demand for services; 2) the impact of payment caps or global hospital budgets on the quality of medical services; 3) how hospitals alter their behavior under the reform, including changes in efficiency, cost shifting, risk selection, and capital investments; 4) the appropriate level of wages for the individual account; and, 5) organizational changes induced by reforms, such as changes in referrals between health facilities. 7.25 In extending the demonstration projects to fifty more cities and prefectures in 1996, program managers could consider these potential improvements to the demonstration projects: * Expand coverage to include dependents. * Protect the funds accumulated in the individual accounts and the reserves for the social risk pool by paying inflation-adjusted rates of return. * Establish modem scientific accounting and auditing methods, including public quarterly financial reports for the insurance centers. The operations of the insurance centers need to be as transparent as possible to win the confidence of beneficiaries, employers, and providers. * Include representatives of employers, workers, and providers in each city on a Board of Trustees to oversee the overall operations of the insurance program. * Direct insurance centers to pay hospitals on a case-based, packaged-fee basis categorized, perhaps, into trauma, surgery, medical, cancer treatment, pediatrics, and long-term infectious diseases. Adjust prices for labor inputs, drugs, and diagnostic tests to end distortions that now cause much of the obvious waste in the health system. 7.26 One aspect of the insurance experiments deserves rethinking. Benefit costs and personal health accounts in the Jiujiang and Zhenjiang experiments transfer income from the frail to the healthy worker relative to the earlier arrangements. Previously, all GIS and LIS 64  wage tax contributions were available for risk pooling. A person who used no health services received no benefits while a worker with major medical costs paid nothing. Under these experiments, about half of wage contributions is deposited to individual accounts--much of which will never be spent on health care because many people never use health services. Meanwhile, very ill workers will exhaust their individual accounts, paying another deductible equal to 5 percent of their wages before the risk pool will pay a portion of the excess expenditures. In its evaluation of the experiments, the government should consider the acceptability of such outcomes, along with the cost savings generated. Links Between Pension Reforms, Health Insurance Reforms, and Economic Modernization 7.27 Labor mobility has been extremely limited in China until recently. SOE workers, once hired, stayed with their employer for life. SOEs, in turn, met pension and health obligations to their workers on a pay-as-you-go basis. With lifetime employment, there was little need to make pensions and health care rights "portable" across employers, although problems of unfunded liabilities for pensions and retiree health benefits are already undermining the viability of pay-as- you-go finance. Further, linking pension and health benefits to employment makes workers dependent on their employers, not only for their jobs but in order to realize their pension rights and to have health insurance both while actively working and after retiring. Survival of the enterprise becomes, then, a political necessity even at the cost of open-ended public subsidy. To improve efficiency in the state enterprise sector, noncompetitive enterprises must be allowed to fail or restructure, and therefore labor must become more mobile. Delinking pension and health benefits from the enterprise helps to make this possible. The current system is now an obstacle to economic modernization, which requires labor mobility and state enterprise reform. Pension and health finance reforms are needed so that workers can transfer jobs without jeopardizing their pension and health benefits, to deal with problems of pay-as-you go finance and the cost of health care for retired workers, and to facilitate enterprise reform. 7.28 As for pensions, China appears to be moving away from a pay-as-you-go system operated by individual enterprises to a unified pension system, with a substantial part of retirement income coming from fully funded individual accounts. Benefits need to be reduced to make the new system financially sound. Coverage could be extended by gradually including all formal sector workers in urban areas and employees in large township enterprises. This reform would produce important benefits: it would delink pension administration from enterprise management, establish a level playing field for enterprise contributions, encourage labor mobility, contribute to capital accumulation, provide incentives for savings, provide poverty protection, and contribute to income security in old age (for covered workers). One issue in implementing such reforms is the cost of financing the transition, given the large outstanding unfunded liability of current pensioners in SOEs and the accrued pension rights of current workers. It appears unlikely that pension reform, in the absence of parallel reform of health finance, can meet its objectives of facilitating labor mobility and state enterprise reform. This report argues, therefore, that reform of enterprise health finance should proceed simultaneously with (and on similar principles as) pension reform. Options for Broadening Urban Risk Pooling 7.29 In addition to these ongoing experiments with GIS and LIS, health insurance coverage needs to be broadened to included the uninsured in urban areas. First, including dependents fully under GIS and LIS arrangements would significantly expand insurance coverage in urban areas. 65  Second, over the medium to long term, some form of risk pooling needs to be developed to cover the increasing proportion of the labor force that works in both the formal and informal sectors in urban China, as well as their dependents and the elderly. As the experience from the Shenzhen scheme indicates, getting joint ventures and privately-owned companies to participate in social health insurance will be difficult. It will take many years to set up well-functioning insurance systems that cover the majority of the urban population. In the meantime, government support to public health and to basic services for the poor will help to meet the needs of the urban residents who are not covered by insurance. 7.30 Various combinations of public and private financing arrangements could be developed for risk pooling in urban areas. For the increasingly large urban population working outside the state-owned sector, the government could encourage individual participation in an expanded LIS, complemented by public support for the indigent. The result would be akin to the social insurance systems for health care that now serve much of Europe. It will be a major challenge to fully insure GIS and LIS family members as well as employees and dependents of locally-based enterprises and the growing group of mixed ownership ventures. Insurers, be they public or private, will need to be established, like the insurance centers in the Jiujiang/Zhenjiang models, or other variants. 7.31 A number of intermediate steps could be taken to support the long-term goal of establishing broad-based, equitable, and efficient urban risk pooling of health care costs. Some of these intermediate steps are already being tested in the Jiujiang and Zhenjiang insurance experiments. - One option is to take a laissez-faire approach over the short to medium term and let private, voluntary insurance emerge, with some government regulation, to cover part of the gap in urban coverage. Employers could choose to include health insurance as part of their benefits and either self-insure or join a larger insurance pool. Individuals could join their employer's plan, if available, or voluntarily purchase private health insurance. There are problems with this model, however. Small employers would have difficulty purchasing private insurance. Workers would face the threat of losing insurance if they changed jobs, while individuals with high health risks would have difficulty purchasing insurance. * Another possible option is to require that firms offer health insurance to their workers, without requiring that employers contribute. This would at least provide the advantages of some risk pooling. * Alternatively, the government could mandate employer and employee contributions for health insurance. Initially, health care could be purchased by the firm. Ultimately, mandated contributions might be pooled in one fund, with the government contributing money to cover the indigent and elderly. 7.32 These various options for expanding risk pooling need to be carefully reviewed, especially in light of the experience in Jiujiang and Zhenjiang. Intermediate steps will be needed to phase in coverage and build up administrative and regulatory capacity. Implementing reforms will inevitably require some public financial support in terms of studies, technical assistance, and monitoring and evaluation. 66  8. RECOMMENDATIONS AND IMPLICATIONS FOR PUBLIC FINANCE 8.1 China had an enviable record in improving the health of its people between 1950 and 1980. The move toward a market economy that began in 1979 fundamentally altered China's financing and organization of health care, however. Without a coherent national health policy that appropriately defined the respective roles of the government and market in the reformed Chinese economy, the health system drifted--reacting on an ad-hoc basis to the nation's changing economic and social environment. Now China faces several major health care problems as summarized in Table 8.1. Table 8.1. Summary of Major Health Care Issues in China Issue Likely Causes Health status: While China's * Government spending for public health programs has declined. overall health status indicators are relatively high, child * Income inequality in China has grown. Some 100 million rural mortality rates have stagnated residents remain very poor. despite rapid increase in standards of living. * Access to necessary health care has declined due to inadequate public financing for the poor and lack of organized financing for the rural population. Access to health care in China * Rapid health cost inflation makes services less affordable to many. is inequitable * Due to lack of risk pooling after the collapse of the CMS, many poor and near-poor forego necessary medical services. * Insurance coverage is lacking for much of the urban population including migrant/unregistered workers in cities and workers in private or collective enterprises. Inefficiency and waste have * Misallocation of public resources favors hospital services over led to many problems public health spending. including a decline in clinical effectiveness and quality of * Distorted prices encourage overuse of drugs and high technology services. tests. * Duplication of facilities results from overlapping hospital systems and vertical health programs. * Epidemic prevention programs, clinics, and hospitals all rely on user fees to finance their operations, which encourages overprovision. Health costs increased by an * Distorted prices encourage overprovision of drugs and expensive average of 11 percent per diagnostic tests. person annually between 1986-1993, in real terms. * Fee-for-service payment encourages overprovision. * Aging of the population and increases in chronic diseases raise costs. 67  8.2 How the Chinese government addresses these health issues will strongly affect health conditions, equity and risk sharing, efficiency, and total health care costs. International experience shows that China's health care policy is at a critical juncture. If China continues on the present course, the problems will deepen and become more difficult to remedy. The United States, for example, failed to address its risk-sharing problems in the 1940s, leaving the worst health risks (the elderly, disabled, and poor) uninsured by the private market. Eventually, the government had to bear the burden for insuring the elderly, disabled, and poor, while 15 percent of the population remains uninsured today. Meanwhile, health costs escalated and consumed some 14 percent of GNP in 1995. South Korea and the Philippines did not correct their fee-for- service payment system and now face rapid inflationary pressure in health costs. Germany and Canada, on the other hand, took timely and effective action when problems appeared in the 1 970s and 1 980s. As a result, they were able to assure equitable access to health care for their citizens, relatively low health cost inflation, and steadily improving health conditions. Many OECD countries have undertaken health reforms in recent years that may contain some useful lessons for China, as summarized in Box 8.1. Priorities of Government Health Spending 8.3 As a general principle, governments should not spend their limited resources for health services that individuals will readily buy for themselves. Tax revenues are difficult to collect, and government spending is often less effective at meeting individual demand than spending done by individuals themselves. These considerations underline the importance of using public sector funds for important goods and services that would not--or cannot--be purchased privately, such as cost-effective public goods for all and priority clinical services for the poor. A related principle is that government health spending and regulation should seek to leverage effective and efficient use of health spending by other providers and buyers of health services. In China, direct government spending on health, including GIS, was no more than one-quarter of total sector spending. Both individual buyers of health services, and the SOEs taken as a group, spent far more on health than did the Chinese central, provincial, and local governments. 8.4 Following these principles, to address some of the health care issues enumerated in Table 8.1, the Chinese government will need to substantially increase funding for public health programs, i.e. those activities that the government must finance if they are to be provided at adequate levels. Full funding of priority public health programs is therefore the first priority for public spending on health, as shown in Table 8.2. This report recommends increasing public funding of EPS over the medium term from its present level of 1.3 billion (in 1993) to at least 6.5 billion in 2001. This calculation is based on current cost estimates of fully funding tuberculosis control, expanded immunization programs, endemic disease control, health education, and other important programs, as well as projected cost increases to the year 2001. (As income is projected to double in real terms, EPS costs will also rise, particularly due to wage increases.) 8.5 In addition to financing high priority public health programs for all of China's population, the government's next health spending priority should be to subsidize essential health services for the poor on poverty assistance, or equity, grounds. Most of the poor and near- poor in China live in rural areas, and are concentrated in China's 592 poorest counties. This report recommends a blend of geographic targeting (aimed at poor administrative villages in China's poor counties) and universal finance of a limited number of services that particularly benefit the poor. Geographic targeting could be implemented through community-financing 68  Box 8.1. Lessons From Reform of OECD Health Systems and Singapore What can China [earn from the incremental reforms that many OECD countries have undertaken in their health systems over the past fifteen years? OECD countries have employed a wide variety of financing, medical care provider payment, and delivery system arrangements. Reforms have been aimed at achieving universal or near- universal access to care, containing costs (and improving efficiency), and, more recently, improving consumer satisfaction. Lessons from OECD country health reforms can be summarized as follows: Extending coverage and risk-pooling arrangements * With the exceptions of Mexico, Turkey, and the U.S.A., all OECD countries have achieved universal access to health care, largely financed through national and local governments or mandated social insurance. * It is difficult to achieve universal coverage in systems that rely heavily on private, voluntary insurance. This is because private insurers will compete on the basis of risk selection, and sub-groups in the population with higher- than-average expected health expenditures will not be able to obtain insurance. Private voluntary insurance can also impede labor mobility if workers are reluctant to change jobs because of a possible loss of health insurance coverage. Some of these problems can be ameliorated through regulation. * An alternative to mandating health insurance is to mandate savings, with the further stipulation that the savings so generated can only be used to pay medical expenses--these are known as "medical savings accounts" (MSAs). Singapore set up a system that complemented MSAs with public finance for the indigent and catastrophic insurance to cover exceptionally high costs. One problem with MSAs is that they lose many of the efficiency advantages of risk pooling. At the same time, individuals with good health have an incentive to spend their medical savings account on health--perhaps unnecessarily--because the money is not fungible. Furthernore, MSAs and catastrophic insurance are most compatible with fee-for-service payment methods, with consequences for cost escalation and medically inappropriate care. MSAs have not been successful in controlling health care costs in Singapore--per capita health care costs have grown by 13 percent per annum since 1984-2 percent faster than before the introduction of the system. Cost containment and efficiency improvements * Increases in health expenditures have not necessarily brought about improved health outcomes. * Fragmented fee-for-service systems like those in the U.S. (and China), without a single payer or set of rules, have been the least successful in controlling expenditures and guaranteeing access. * Consumer expectations, new technologies, and aging populations continue to place upward pressure on health expenditures. * Most OECD countries now have surpluses of physicians, especially specialists, and hospital beds. These surpluses are a major factor in cost escalation. Once created, these surpluses are hard to reduce for political and other reasons, and are best not created in the first place. * Cost containment is a major issue in almost all OECD countries. Empowering consumers and having money follow patients within the context of an overall global budget appear to be the preferred strategies for controlling costs. Specific medical care provider payrnent approaches such as capitated payments and diagnosis-related groups (DRGs), if appropriately irnplemented, appear to control costs without compromising access and quality. Evidence is clear from OECD countries: fragmented systems that allow many individuals to remain uninsured and rely principally on fee-for-service payments, like China's, are inequitable, inefficient, and perform poorly in terms of health outcomes, Many countries are paying close attention to the refonn lessons from the OECD. Since the 1980s, China, however, has been moving toward less government support for priority public health activities for all and clinical services for the poor, and more reliance on a fee-for-service delivery system. If current trends persist, China will move into the 21 st century with a poorly performing health system that is nonetheless costly. 69  schemes or supply-side subsidies. If community-financing schemes can be established in poor areas, perhaps with technical assistance from the government, the government could channel subsidies for services for the poor through these schemes. China has considerable experience with community-financing schemes. They offer many benefits, in terms of providing risk pooling for the whole community and efficiencies in the delivery of care. Nonetheless, they are relatively difficult to administer, perhaps especially in poor areas with the least administrative capacity at the village level. An alternative or complementary approach would be to provide supply-side subsidies (subsidies to health programs or health facilities, in return for providing free care to the poor in the poorest counties). Existing subsidies to township health centers could, over time, be redirected to the poorest counties. It might be reasonable to target 25 yuan per person now, rising to 50 yuan per person in line with China's overall income growth by the year 2001. The costs of these subsidies by the year 2001 would depend on the number of people covered. Using a rough estimate of approximately 75 million people living in poor villages in China's 592 designated poor counties, the cost would be 3.75 billion yuan. This illustrative figure could be reduced accordingly if the target population group were smaller or if the per capita subsidy were reduced. These geographically-targeted subsidies could be complemented by program subsidies for clinical services that greatly benefit the poor, perhaps costing I billion yuan per year. 8.6 In urban areas, other measures are needed to improve the urban poor's access to priority health services. Better funding for public health programs will help, but the government could also consider redirecting some of the existing general hospital subsidies (4.5 billion yuan) to services for the urban poor. (Note that the Ministry of Health argues that it is committed to providing subsidies for basic salaries of hospital employees and that therefore reallocation is not possible.) These subsidies could focus on public facilities in poor areas of the cities, on programs that particularly benefit the urban poor, or on individual poor households, if individual targeting methods can be developed. Because there are considerably fewer urban poor than rural poor, subsidies to the urban poor would also be considerably lower, and are estimated here at 0.5 billion yuan. 8.7 A third priority for government finance is to support reforms to prices and provider payment mechanisms, and a fourth priority is to control investments and improve regional planning. These measures can generate sector-wide improvements in efficiency. The illustrative costs of these measures have not been estimated for this report but are of a much smaller magnitude than the first two priorities. 8.8 A fifth priority of public spending on health is to promote risk pooling among the broader population to provide protection, at a minimum, from catastrophic health expenses. Table 8.2 includes the cost of a small per capita subsidy to rural communities that operate community-financing schemes. Government incentives to rural communities to establish and operate community-financing schemes can be justified on efficiency grounds: the government has an interest in seeing that efficient risk-pooling mechanisms are established in rural areas. This support might be provided in the form of a matching grant to communities that agree to follow certain principles as discussed in Chapter 6. The subsidy might be 5 yuan per person now, growing to 10 yuan by the year 2001 in line with overall income growth. If 120 million members of the rural population received this subsidy by the year 2001, the total cost to the government would be roughly 1.2 billion yuan. Government support is also needed for reforms to improve the efficiency of urban insurance systems and support expanded coverage. As urban insurance experiments expand throughout China, the government will need to ensure that these 70  are systematically monitored and evaluated to guide medium- to long-term policy choices. Table 8.2 includes 0.1 billion yuan in the year 2001 for these expenditures. Table 8.2. Public Finance Implications of Proposed Health Initiatives: Illustrative Costs (billion 1993 yuan) 2001 1993 (Proposed or (Actual) Projected) GDP and Government Tax Revenue GDP 3,451 7,500 General Government Expenditures 450 1,400 Proposed Incremental Tobacco Tax 10 (20 percent, ad valorem) Central and Provincial Public Spending on Selected Priority Health Programs Fully fund priority public health programs 1.3 6.5 Clinical services for the poor Poor villages in poor counties 1/ 3.7 Specific disease program subsidies 1.0 Urban poor 0.5 Subsidies to promote community-financing schemes 2/ 1.2 Urban health finance initiatives 3/ 0.1 Total Expenditures, Selected Programs 1.3 13.0 as share of government expenditure 0.29 % 0.93% as share of GDP 0.04% 0.17% 1/ Subsidies to the rural poor could either go directly to health programs/facilities in poor rural areas for free clinical services for the poor, or be channeled through community-financing schemes in poor areas, if they exist. For illustrative purposes, a subsidy of 50 yuan per person for the 75 million people in poor villages in China's 592 poorest counties was used for this calculation for the year 2001. 2/ This report recommends that the govemment consider a small matching grant that provides an incentive for communities to operate community-financing schemes. This small matching grant might be 5 yuan per capita now, growing to 10 yuan by the year 2001 in line with income growth. If the subsidy was for community- financing schemes for 120 million people in rural areas by that year, the costs would be about 1.2 billion. 3/ This estimate includes the costs of major studies on pricing and provider payment, monitoring and evaluation of urban insurance experiments, and possibly some start-up funds for new initiatives. 8.9 In the long run, these financing policies will yield high returns. They will generate savings and improve health conditions, especially of the poor. They will also pool risks, improve efficiency, and reduce health cost inflation while keeping the percentage of GDP spent on health care at a relatively low level. While the needs are relatively clear, the crucial question becomes: does China have the resources and will to accomplish these goals in light of other pressing economic and social issues the country faces? Finding Funds for Public Spending on Health 8.10 Government revenues have declined in recent years as a share of GDP. To keep the budget deficit in check, the government's expenditures have been reduced from 33.8 percent of GDP in 1978 to 17.9 percent in 1994. Of the 17.9 percent of government expenditures in 1994, 71  14.1 percent were budgetary and 3.8 percent Table 8.3. China's Government Expenditures in were extrabudgetary. China's budgetary International Perspective expenditures as a percent of GDP are far below those in most other countries. The Government Central central government's share of all government expenditures as government expenditures is also unusually low in China as percent of GDP expenditures shown in Table 8.3. as % of all government 8.11 China has to increase public expenditures expenditures for several urgent needs: All Countries 39.1 72.3 poverty alleviation, health, education, infrastructure, environmental protection, Industrialized 47.6 65.9 pension reform, and unemployment insurance. Countries The World Bank has recently recommended Developing 31.7 77.8 public finance reforms in China to increase Countries government revenues and enable the government to better address these urgent (Budgetary, needs. The World Bank's recommended tax 1994) measures and their revenue effect in the year (Extra 3.8 - 2000 are shown in Table 8.4. This report budgetary) takes these recommendations into account and SurcetWrld k 1996a. then evaluates their implications for the public Note: Data are averages over three years ending in 1987 finance of health care. or 1988. Table 8.4. Revenue Implications of 8.12 With projected increases in government Recommended Tax Measures in the Year revenue from the tax measures shown in Table 8.4, 2000 the proposed priority health programs enumerated Tax Measure Incremental in Table 8.2 should be easily affordable. However, Revenue Effect government officials are reluctant to plan for by 2000 expanded public spending on health given fiscal (Percent of constraints in recent years. If the actual increase in GDP) revenues is lower than that projected in Table 8.4, VAT 2.1 these programs could be phased in more slowly. Iniiulicma. Public funds recommended for the priority health programs would only total about 13.0 billion yuan Enterprise income tax 1.2 in the year 2001, or 0.9 percent of projected public Taxes on pollutants 1.0 spending. Furthermore, as discussed in Chapter 3, Payroll taxes 0.9 there are strong public health reasons for increasing Total 6.0 the present level of taxation on tobacco products. If China does increase its tobacco tax, and directs Source: World Bank 1996a. much or all of the incremental revenue to priority health programs, the proposed health initiatives would be even more affordable. 72  Annexes Annex 1. China's Health Expenditure and Health Systems Data Annex 2. Outside Influences on Health Status Annex 3. Deaths and Disease Burden in China Annex 4. Child Mortality Trends in China Annex 5. Health Indicators from Disease Surveillance Points System   ANNEX 1: CHINA'S HEALTH EXPENDITURE AND HEALTH SYSTEM DATA 75  Annex 1 Table A1.1. China: GDP, Nominal Exchange Rates and Price Deflators, 1978-1994 Nominal official exchange rate GDP: billions of yuan GDP price deflator index (annual average) current prices constant 1993 prices 1990=1 1993=1 yuan/U.S.S 1978 362.4 880.9 0.541 0.411 1.7 1979 403.8 948.2 0.56 0.426 1.6 1980 451.8 1020.8 0.582 0.443 1.5 1981 486.2 1067.4 0.599 0.456 1.7 1982 529.5 1158.6 0.601 0.457 1.9 1983 593.5 1275.2 0.612 0.465 2 1984 717.1 1471.1 0.641 0.487 2.3 1985 896.4 1669.6 0.706 0.537 2.9 1986 1020.2 1815.4 0.739 0.562 3.4 1987 1196.3 2027.2 0.776 0.590 3.7 1988 1492.8 2256.4 0.87 0.662 3.7 1989 1690.9 2348.0 0.947 0.720 3.8 1990 1853.1 2436.8 1 0.760 4.8 1991 2161.8 2664.3 1.067 0.811 5.3 1992 2663.5 3040.4 1.152 0.876 5.5 1993 3451.5 34515 1.315 1.000 5.8 1994 4500.6 3860.6 1.533 1.166 8.6 Source: China Statistical Yearbook 1995, p. 32. .e7  Annex 1 Table A1.2. GDP, Health Spending and Population Growth, 1978-1993 (GDP and health spending in 1993 constant yuan) Year-end GDP Total Health Health GDP Population per Spending Spending (billions (10,000 capita (billions Per Capita of yuan) people) (yuan) of yuan) (yuan) 1978 880.9 96,259 915 25.8 27 1979 948.2 97,542 972 28.7 29 1980 1020.8 98,705 1,034 31.6 32 1981 1067.4 100,072 1,067 34.2 34 1982 1158.6 101,541 1,141 39.0 38 1983 1275.2 103,008 1,238 43.0 42 1984 1471.1 104,357 1,410 47.9 46 1985 1669.6 105,851 1,577 49.0 46 1986 1815.4 107,507 1,689 54.6 51 1987 2027.2 109,300 1,855 62.5 57 1988 2256.4 111,026 2,032 71.1 64 1989 2348.0 112,704 2,083 77.8 69 1990 2436.8 114,333 2,131 84.8 74 1991 2664.3 115,823 2,300 94.2 81 1992 3040.4 117,172 2,595 103.3 88 1993 3451.5 118,517 2,912 132.1 111 Annual Growth Rates: 1978-86 1986-93 1978-93. Population 1.4% 1.4% 1.4% GDP per capita 7.7% 7.8% 7.7% Health spending per capita 8.0% 11.2% 9.5% Source: China Statistical Yearbook 1986, p. 71; China Statistical Yearbook 1994, p. 59.; China Statistical Yearbook 1995, p. 32; Wei Ying, 1995. 77  Annex I Table AI.3. China: National Health Expenditures, 1978, 1986, and 1993 (millions of 1993 yuan) 1978 1986 1993 Health Percent Health Percent Health Percent Expenditure health Expenditure health Expenditure health Funding Source expenditure expenditure expenditure Government Budget 7,292 28 17,288 32 18,878 14 (excluding GIS) LIS and GIS 7,689 30 18,274 33 46,108 36 Out-of-Pocket Payment 5,268 20 14,185 26 56,106 42 Rural Cooperative 5,109 20 2,918 5 2,243 2 Medical System Other Sources 428 2 1,956 4 8,713 6 TOTAL 25,786 100 50,621 100 132,048 100 Total Health 3.0 3.2 3.8 Expenditure as Percentage of GDP Source: MOH (data provided in May 1995). HEI re-estimated 1993 out-of-pocket payments in March 1996 but did not review earlier years. Table AI.4. China: Government Spending on Health, Selected Years, 1978-93 (millions of 1993 yuan) Spending Category 1978 1986 1990 1991 1992 1993 Recurrent Health Budget 5,294 10,598 10,686 10,658 10,965 10,786 Recurrent Budget: 0 643 869 901 951 917 Traditional Chinese Medicine Recurrent Budget: 0 1,409 1,730 1,965 2,211 2,292 Family Planning Fund Research 160 197 132 134 177 251 Higher Education 0 0 0 899 857 899 Capital Investment 628 2,107 1,400 895 877 1,144 Other Ministries 1,209 2,335 2,113 1,226 1,406 1,415 Government Spending Subtotal .291 17288 16.930 16.680 17.443 18.878 Government Employees' 1,178 3,360 5,834 6,234 6,667 11,667 Insurance (GIS) Government: TOTAL 8469 20.648 22.764 22.914 24.1 25323 Source: MOH (provided in July 1995). 78  Annex I Table A1.S. Cbina: Components of Recurrent Health Budget, Selected Years, 1978-93 (millions of 1993 yuan) Spending Category 1978 1986 1990 1991 1992 1993/ Hospital Operating Expenses 1,7945 4,488 4,254 4,072 4,290 4,183 Subsidies for Health Centers 1,421 1,903 1,968 2,040 2,233 2,145 Epidemic Prevention Fund 944 1,544 1,583 1,655 1,948 1,305 (*) Maternal and Child Care Fund 0 295 401 429 536 324 (*) Pharmaceutical Control Fund 0 144 147 164 194 229 Professional Middle School Fund 264 511 503 517 590 626 Training Fund 09 100 0 108 142 156 Rural Cooperative Medical System 89 43 34 33 29 27 Fund__ _ _ _ _ _ __ _ ___ _ _ Kindergarten Fund 0 7 0 7 9 9 Indigent Patients' Hospital Fund 0 18 0 17 18 15 Other 783 1,662 1,789 1,610 1,856 1,919 Recurrent Health Budget: TOTAL 5,295 10,722 10,684 10,658 11,850 10,939 Source: MOH/DPF, May 1995; revisions of Epidemic Prevention Fund and Matemal and Child Care Fund for 1993 by HEI, March 1996. (*) Prior to 1990, govemrnment budget data excluded revenues from user fees; from 1991 onwards, govemment data included user fees with budget figures. The budget data for Epidemic Prevention Fund and Maternal and Child Care Fund for 1993 were revised to exclude net user fees, but 1991 and 1992 figures have not yet been revised. Discrepancies in the total recurrent budget figures between Table A1.4 and Table Al.5 arise as a result of this inconsistency in accounting for sources of funds. Table A1.6. China: Health Insurance Coverage and Spending, Rural Population, 1993 Health Total Health GROUP Population Income per Percent Services Services (millions) capita insured Spending Spending by (1993 yuan) within the per capita Group (billion group (1993 yuan) 1993 yuan) RURAL 900 750 10 60 54 Upper quartile 225 920 40 98 22 Upper-Mid quartile 225 668 0 67 15 Low-Mid quartile 225 489 0 44 10 Bottom quartile 225 361 0 31 7 Poverty Group (70) <300 0 <10 (1) TOTAL 1,200 2100 110 132 Source: Estimates based on data from Wei 1995. HEI made revised estimates of national health expenditures in January 1996, adding out-of-pocket spending to total health expenditure. Other sources are Zhao Zhuyan and Lusheng Wang 1995 (Iah 12, for rural income). Statistical Yearbook of China 1994 provides data for urban income. HEI staff adjusted the 1993 National Health Services Survey data to estimate spending by sub-groups. 79  Annex I Table A1.7. China: Revenues and Government Subsidies of Health Institutions Providing Hospital Services, 1993 (billions of 1993 yuan) Source of Revenue Hospitals Township Health Total Centers Revenuesrom: Medical Treatment 18.0 3.2 21.2 Sale of Pharmaceuticals and Others 26.8 9.2 36.0 Subtotal 44.8 12.4 57.2 Government Subsidies 4.6 2.0 6.6 TOTAL REVENUE FROM ALL 49.4 14.3 63.8 SOURCES Source. HEI, based on MOH Statement of Health Budget and Expenditure: (Berman and others, 1995). These data differ slightly from those presented in Meng 1995. Table A1.8. China: Average Number of Hospital Beds and Health Care Personnel, 1993 and 1994 1993 1994 Type of Hospital No. of No. of No. of Beds No. of Medical/ No. of Beds No. of Medical Personnel Technical Personnel Technical Personnel Personnel Hospitals at and above County level 137 189 145 139 192 148 Rural Township 16 21 18 114 120 17 Hospitals Other Hospitals 45 59 47 42 56 44 Source: Statistical Yearbook of China, 1995, Table 19-15, p. 667. 80  Annex I Table A1.9. China: Utilization of Hospital Beds at County Level and Above, Selected Years, 1985-94 (Unit) 1985 1989 1990 1992 1993 1994 A. Government Hospitals Bed Turn-Over Rate Turn over per year 19.9 19.9 19.9 18.9 17.9 17.9 Bed-days in Use Days 320.9 315.9 313.9 303.9 267.9 263.9 Bed Occupancy Rate Percent 87.9 86.9 85.9 83.9 75.9 72.9 Average Length of Stay Days 15.9 15.9 15.9 15.9 15.9 14.9 B. SOE-run Hospitals Bed Turn-Over Rate Turn over per year 14.9 14.9 13.9 12.9 12.9 12.9 Bed-days in Use Days 253.9 259.9 255.9 245.9 221.9 226.9 Bed Occupancy Rate Percent 69.9 70.9 69.9 67.9 60.9 62.9 Average Length of Stay Days 16.9 16.9 17.9 17.9 16.9 16.9 C. All Hospitals Bed Turn-Over Rate Turn over per year 18.9 18.9 17.9 16.9 15.9 15.9 Bed-days in Use Days 302.9 299.9 296.9 286.9 259.9 251.9 Bed Occupancy Rate Percent 82.9 81.9 80.9 78.9 71.9 69.9 Average Length of Stay Days 15.9 15.9 15.9 16.9 15.9 15.9 Source: Statistical Yearbook of China, 1995, Table 19-20, p. 671. 81   Annex 2 ANNEX 2: OUTSIDE INFLUENCES ON HEALTH STATUS 2.1 As discussed in Chapter 1, health policy is not the only important determinant of a nation's health outcomes. This annex discusses several key outside influences on a nation's health status, with specific reference to China. The Effect of Income 2.2 The effect of per capita income on a country's health status becomes progressively weaker as incomes rise. In 1990, a doubling of per capita income from an initial level of $1,000 (adjusted for purchasing power parity) corresponded on average to a gain of approximately eleven years in life expectancy. A doubling from an initial level of $4,000 per capita, however, led to a smaller (but still substantial) increase of four years (World Bank 1993a, pp. 59-72). In China in the mid-1970s, provincial level data indicated that a 10 percent increase in income resulted in a gain of about eight months in life expectancy (Prescott and Jamison 1985). Studies undertaken at the household level similarly show a strong but declining effect of income on health indicators. This evidence points to the importance of reducing poverty as well as raising overall per capita incomes for improving health outcomes. 2.3 Rapid economic growth in China has led to rising incomes and contributed substantially to health improvements--particularly since, until recently, poverty reduction has accompanied growth. In the period following the beginning of economic reforms in China (1978-1985), the percent of the rural population living in absolute poverty declined from about 33 percent to about 12 percent (See Table A2.1). In the subsequent five years of continued high GDP per capita growth rates, however, the rural population in poverty declined by only about half of a percent overall and may have risen in some areas. In a reassessment of the 1985-90 data from four southerm provinces, Chen and Ravallion (1996) conclude that the poorest were better off than earlier analyses had indicated, although they Table A2.1. Income and Poverty in China, 1978-94 agreed progress in Percent of reducing the percentage of GDP Per Rural the population in absolute Capita GDP Per Capita Population poverty virtually ceased in Year (in 1993 Growth Rate in Absolute the period 1985-90. yuan) Poverty 1978 916 33.0 2.4 Continued rapid 7.8% (1978-85) income growth in China 1985 1577 11.9 therefore can be expected 6.0% (1985-90) to provide a base for real 1990 2133 11.5 (but modest) health 10.3% (1990-94) improvements for the 1994 3221 n.a. majority of China's population. But unless the Source: Statistical Yearbook of China, 1995, p. 32. next phase of economic World Bank (1992b p. ix) provides data on absolute property. grothaces every growth reaches the very 83  Annex 2 poor, they will not have the means to Table A2.2. Illiterates and Semi-literates, Selected escape from the sickness and Provinces, undernutrition that help keep them in 1982 and 1990. poverty. High overall economic growth rates can, however, help provide the GDP Per Illiterates and resources needed to meet the priority Capita, Semi-literates as a resources needed tof meet the p riori.y (constant Percentage of the health needs of the poor. 1993 Total Population' The E.ffect of Education yuan) Region 1990 1982 1990 2.5 Education complements income Nationwide 2133 23% 18% in enabling households to improve their *** ___ own health. Indeed, provincial data on Beijing 4707 13% 10% determinants of life expectancy in China Zbeijang 2259 24% 19% in the mid-1970s suggested that Guangdong 2424 17% 13% provincial literacy rates were more *** ___ directly associated with changes in life Gansu 1235 35% 31% expectancy than were provincial income Guizhou 849 33% 29% levels (Prescott and Jamison 1985). Shaanxi 1224 25% 20% Source: Statistical Yearbook of China, 1985, 1992, and 1993 2.6 Here, too, China's performance 'For 1982: Aged 12 and over. has contributed to progress. Table A2.2 shows the clear progress China has made in reducing the percentage of the population classified as illiterate or semi-literate between 1982 and 1990, although regional disparities remain. As with income growth, gains in education account for much past health improvement and establish a foundation for future gains. Water, Sanitation, and Food 2.7 Cleanliness and quality of water relate directly to a broad range of diseases, while access to sanitation may be even more important to health. Relative to other countries in Asia with per capita income near that of China, China's levels of access to both water supply and sanitation are good, although differences in access between urban and rural areas persist. (Table A2.3 provides data on China and comparators.) 2.8 The quantity and Table A2.3. Access to Safe Drinking Water and Sanitation, quality of the food supply China and Comparators, 1990 qaiyo h odspl also have a direct impact on Safe Drinking Water Sanitation health. In extreme cases of food shortage, the lack of Country Urban Rural Total Urban Rural Total nutrients, or infections, can China 87 68 72 100 81 85 overwhelm the immune India 86 69 73 44 3 14 systems of the _____ _____ ______ ~~~~malnourished. In the period Indonesia 35 33 34 79 30 45 1959-61 China underwent Sri Lanka 80 55 60 68 45 50 such a famine, with extremely adverse Japan 100 85 96 100 100 100 consequences for mortality Source: World Bank 1994b, Table A.2. rates and health status more Note: Numbers in the table are expressed as a percent of total population. 84  Annex 2 generally. Short of famine, low levels of Table A2.4. Per Capita Nutrient Availability nutrient intake can combine with infection in China, 1950-95 to result in malnutrition, e.g. low levels of height-for-age or high levels of anemia. National Average Daily per Year Capita Nutrient Availability 2.9 Table A2.4 shows China's progress Energy Protein Fat in increasing nutrient availability to its (Keal) (gm) Fat population in the period since 1949, both in (Kcal) (gm) (gm) terms of the quantity of the diet and in 1950 1742 49 22 increasing dietary quality (protein and fat). 1960' 1578 41 17 1970 2192 56 26 Fertility decline and demographic change 1980 2473 64 35 2.10 China's recent demographic history 1990 2679 66 51 compresses into a third of a century the 1995 2727 67 51 changes that took more than a century to ] 960 was the middle year of a 3-year famine in China that unfold in most of today's high-income pushed the infant mortality rate up to around 330 per countries. China's rapid mortality decline, thousand from the level of 200 achieved in 1957 (World combined with rapid growth in income and Bank 1984, p. 113). Energy availability per capita in 1957 educational opportunity, created the was about 2217 kcal. precondu tional opportasunity eaied f ithe Source: Piazza 1986. After 1980: FAO 1994. preconditions for a sustained fertility decline. Strong family planning policies then lowered fertility rates to levels well below those actually desired by most of the population. Table A2.5 provides an overview of changes in China's key demographic variables. China's total fertility rate (TFR) declined from 5.8 to 1.9 between 1970 and 1995. The total number of births and the crude birth rate declined sharply as well, although less dramatically than TFR because of a concentration of population in the child- bearing years. Table A2.5. Key Demographic Indicators for China: 1970,1995, and Projections to 2020 Year Variable 1970 1995 2020 1. Mid-year population total (millions) 818.3 1201.4 1437.0 2. Life expectancy at birth (years) 61.4 69.4 73.7 3. Median age at death (years) 39.8 67.4 70.2 4. Total deaths (million) 6.2 8.9 12.3 5. Crude death rate (per 1,000 population) 7.6 7.5 8.6 6. Total fertility rate 5.8 1.9 2.05 7. Total births (million) 27.3 21.4 20.3 8. Crude birth rate (per 1,000 population) 33.4 17.9 13.7 9. Percent of population over age 60 6.8 9.6 15.6 10. Percent urban population 17.5 30.3 50.8 Source: For percent urban, UN 1995; for 1970 indicators, Hill 1988; for 1995 and 2020 indicators, World Bank projections from official estimates. 85  Annex 2 2.11 Declining TFR and birth rates have several important consequences for the health sector. In the short run, greater spacing between births improves the health of both mother and child. A reduction in the actual number of births--e.g. the decline in China from 27.3 million in 1970 to 21.4 million in 1995--reduces the need for obstetric care, immunization, and other child health interventions (although there is still a high unmet need for good quality obstetric and child health care in rural areas). These often constitute a significant fraction of the public effort on health in low-income countries. Finally, in the long run, declining fertility redistributes the age distribution of the population away from younger ages and toward middle and older ages (Figure A2. 1). The change in the age distribution of the population and the age distribution of death profoundly affects the pattern of demand for health services. 2.12 The rapid increase in rural to urban migration in China is another demographic factor with a major impact on health status. Rapid, sustained economic growth in coastal China since 1979 has created a huge demand for labor, mostly in urban areas. Between 1970 and 1995 the urban percentage of the population increased rapidly from 17.5 to 30 percent (Table A2.5). Although urban dwellers are required to be officially registered in the city where they live, in practice large numbers of rural people are gaining access to the cities. Most are not registered urban residents, however, and, therefore, they are not eligible for the food and medical subsidies, state sector employment, and insurance arrangements available to official urban residents. 2.13 This large unregistered urban population poses several problems for the health system. Their sheer numbers put a strain on urban health facilities. Moreover, their dense living conditions, poor hygiene, and inadequate access to medical care are conducive to the spread of infection. Large numbers of male migrant workers living away from their families create strong demand for prostitution, which generates circumstances particularly conducive to the spread of sexually-transmitted diseases and HIV infection. The living conditions of the transient urban population provide an environment for incubating a range of infections affecting that population and capable of spreading beyond it. 86  Annex 2 Figure A2.1. Evolving Patterns of Age Distribution and Mortality In China, 1970,1995, and 2020 China 1970 75+ 70-74 65-69 60-64 55.59 50-54 45-49 25-29 20-24 15-19 UW4 10-14 w 11 5-9 Eaw-i I I I 0_ =0f" I I I __ _ 20 15 10 5 0 5 10 15 20 25 30 35 40 45 50 55 Percentage of total population Percentage of total deaths China 1995 75+ 70-74 65-69 60-64 55-59 50-54 - 45-49 40-44 35-39 i 30-34 525 25-29! C0-24 15-19,f 10-141 5-9 0-4 20 15 10 5 0 5 10 15 20 25 30 35 40 45 50 55 Percentage of total population Percentage of total deaths China 2020 75 1 1 1 1 ,v1 1 1 1 8 70S74u W B = a proecton 60-64 _ 55-591 . 50-54 45S49 f 40-44 11|f 35-39 : 30-34 25-29 20-24 1 15-19 |||llll|ll 10-14 | 20 15 10 5 0 5 10 15 20 25 30 35 40 45 50 55 Percentage of total population Percentage of total deaths Source: Wodd Bank estimates and projections. 87   Anmex 3 ANNEX 3: DEATHS AND DISEASE BURDEN IN CHINA 3.1 This annex presents the results from China of a forthcoming assessment of the burden of disease for major world regions in 1990 with projections to 2020 (Murray and Lopez eds., 1996). Most assessments of the relative importance of different diseases are based on how many deaths they cause. This convention has certain merits: death is an unambiguous event, and the statistical systems of many countries routinely produce the data required. There are, however, many diseases or conditions that are not fatal but that are responsible for great loss of healthy life: examples include chronic depression and paralysis caused by polio. These conditions are common, can last a long time, and frequently place significant demands on health systems. 3.2 To quantify the full loss of healthy life, the World Bank and the World Health Organization undertook a joint exercise as background to preparation of the World Bank's World Development Report 1993. Investing in Health (World Bank 1993a). ' Diseases were classified into 109 categories on the basis of the WrHO's International Classification of Diseases (ninth revision). These categories cover all possible causes of death and about 95 percent of the possible causes of disability. Using the recorded cause of death when available, and expert judgment when records were not available, the study assigned all deaths in 1990 to these categories by age, sex, and demographic region. For each death, the number of years of life lost was defined as the difference between the actual age at death and the expectation of life at that age in a low-mortality population. For disability, the incidence of cases by age, sex, and demographic region was estimated on the basis of community surveys or, failing that, expert opinion; the number of years of healthy life lost was then obtained by multiplying the expected duration of the condition (to remission or to death) by a severity weight that measured the severity of the disability in comparison with loss of life. Diseases were grouped into six classes of severity of disability. Class 2, for example, which includes most cases of leprosy and half the cases of pelvic inflammatory disease, was given a severity weight of 0.22, while class 4, which includes 30 percent of cases of dementia and 50 percent of blindness, was assigned a severity weight of 0.6. (The approach to disability weighting in the 1993 World Development Report was revised and updated with more extensive inputs for estimates reported here.) The death and disability losses were then combined, and allowance was made for a discount rate of 3 percent (so that future years of healthy life were valued at progressively lower levels) and for age weights (so that years of life lost at different ages were given different relative values). The value for each year of life lost rises steeply from zero at birth to a peak at age 25, and then declines gradually with increasing age. These age weights reflect a consensus judgment, but other patterns could be used--for example, uniform age weights, with each year of life having the same value, which would increase the relative importance of childhood diseases. 3.3 The combination of discounting and age weights produces the pattern of disability- adjusted life years (DALYs) lost by a death at each age. The death of a newborn baby girl represents a loss of 32.5 DALYs; a female death at age 30 means the loss of 29 DALYs; and a female death at age 60 represents 12 lost DALYs. (Values are slightly lower for males due to slightly shorter life expectancy.) The sum across all ages, conditions, and regions is referred to as the global burden of disease (GBD). The text in this annex draws in substantial part directly on pp. 26-27 of this 1993 study. 89  Annex 3 3.4 This annex reports the values for China of estimates of the numbers of deaths by cause and of disease burden (in DALYs) for the sixth (and final) iteration of disease burden estimates for 1990. The 1993 World Development Report reported the third iteration while the fourth (Murray, Lopez and Jamison 1994) appeared in a WHO-published compendium of background papers for the 1993 report. That compendium includes a detailed discussion of the methods and assumptions underlying the DALY as a measure and underlying the construction of the empirical estimates. In this annex, Tables A3.1 and A3.2 report the numbers of deaths in China by age, sex, and cause (in absolute terms and in percentage terms). Tables A3.3 and A3.4 report the disease burden (in DALYs) by age, sex, and cause. This work draws fully on Chinese death registration, sample surveys, and, particularly, the invaluable data from the Disease Surveillance Points (DSPs). Table A3.5 compares the distribution of DALYs by cause in China with other regions; Table A3.6 projects those comparisons forward to 2020 on the basis of one set of plausible assumptions about the future course of the main determinants of health. Disease burden can also be decomposed by risk factor; Table A3.7 looks at the consequences of one such factor--tobacco use--over the coming years. 3.5 The method used to compute the burden of disease can also be used to track improvements in a nation's health over time by following changes in the national burden of disease. It is to be expected that an assessment of burden by cause for China in 1995--if it were assembled--would allow careful tracking of progress from 1990 in disease control and identification of specific areas where further effort was called for. Disease Burden Findings 3.6 Noncommunicable diseases and injuries make up a growing share of China's disease burden. The marked increase in prominence of noncommunicable diseases results mostly from the aging of the population that will result from the drop in fertility over past decades; indeed age-specific rates from most NCDs (except the psychiatric conditions) are expected to decline although, based on current growth patterns in tobacco use, rates for tobacco-related diseases may increase. In 1990, an estimated 24 percent of China's disease burden was due to communicable, maternal, and perinatal conditions, compared to 49 percent for all developing countries (Table A3.5). By the year 2020, the share attributable to communicable, maternal, and perinatal conditions is projected to fall to just over 4 percent of the nation's total disease burden (Table A3.6) although, for this actually to occur, it will be necessary to reverse the recent pattern of increases in child mortality (documented in Annex 4). 3.7 The major causes of death in China in 1990 were dominated by noncommunicable diseases and injuries. Chronic obstructive pulmonary disease accounted for 16.1 percent of deaths, followed by cerebrovascular disease at 14.3 percent, and ischemic heart disease at 8.6 percent of deaths. Acute lower respiratory infections, which are a major killer of young children and the elderly, are the fourth major cause of death (5.3 percent). Suicides (3.9 percent), stomach cancer (3.6 percent), liver cancer (3.3 percent), tuberculosis (3.1 percent), hypertension (3.1 percent), and lung cancer (2.5 percent), make up the rest of the top ten causes of death (Table A3. l). 3.8 This order changes considerably when the DALY measure is used for disease burden, incorporating disability and the extent to which mortality is premature. Chronic obstructive pulmonary disease remains the number one cause of disease burden. But it is followed by major affective disorders, acute lower respiratory infections, "other unintentional injuries," 90  Annex 3 cerebrovascular disease, perinatal conditions, suicides, anemia, "other digestive diseases," and ischemic heart disease (Table A3.3). 3.9 China's pattern of disease burden in 1990 is distinct from other countries, both developed and developing, in several ways: * Chronic obstructive pulmonary disease (COPD) accounts for 8.5 percent of the total disease burden in China in 1990. This is nearly twice the share in established market economies (4.8 percent) and developing countries (4.3 percent). The morbidity associated with this disease is large, with consequent demands on health care systems and loss of productivity. Smoking is the most important proven risk factor for chronic obstructive pulmonary disease in the high-income countries but, while it undoubtedly contributes in China, other factors are likely more important. Near equal rates of COPD (Table A3.4) in males and females (despite the fact that most smoking is by males) and a concentration of particularly high rates in the northeast suggest that outdoor (industrial) air pollution and indoor air pollution (from cooking and heating fires) probably dominate the determinants in China. On this substrata of exposure, though, smoking multiplies risk--so the smoking control policies discussed in Chapter 3 will play a useful role in control. L liver cancer is also strikingly high in China, accounting for 1.9 percent of disease burden in China compared to 0.3 and 0.5 percent, respectively, in established market economies and all developing countries (Table A3.5). This results from chronic lifelong infections with the hepatitis B virus and perhaps the consumption of aflatoxins. Hepatitis B also contributes to China's high levels of cirrhosis of the liver. China's immunization program, which now includes hepatitis B vaccine, can provide primary prevention against liver cancer and cirrhosis of the liver (see Chapter 3). Unfortunately, most individuals who will die of liver cancer in the next 40 years have already been infected, so even a successful program of immunization will be unable to avert the doubling projected for liver cancer's share of disease burden by 2020 (Table A3.6). * Relative to other countries, China is characterized by an extremely high suicide rate, particularly among rural females. In China, suicide accounts for an estimated 33 percent of all deaths of women aged 15-29, and 15 percent of all deaths of men in that same age group (Table A3.2). It accounts for a striking 5 percent of all DALYs lost among females at all ages (Table A3.4). * Unintentional injuries also make up an unusually large share of disease burden--13 percent. Drowning is one such cause. It is a serious cause of child death (6 percent of deaths in males aged.0-4 and 25 percent in males 5-14). * Undernutrition and intestinal nematode infections (hookworm, roundworm, and trichuris) remain unusually important in China given its level of economic development. These problems are interrelated; the nematode infections likely account for a substantial proportion of China's heavy burden from anemia and some of its burden from protein-energy malnutrition. Intervention to control the intestinal nematode infections costs little and is highly effective; continued expansion of areas receiving iodine supplementation could accelerate progress against the still major and debilitating problems of iodine deficiency disorders. 91  Tabe A3.1. Nwba of Deaths In Cline by Age, Sa. mand Cm. 1990 mm" Nam N.J.. maw NL Mai ms mats Fusuim Fwasu_ Frmals Fbais Finals Falrh F_rmals C,nilen Al D"t AN N.j. Al Fsmat 0-4 5-14 15-29 30-44 45-56 0-a 70+ 0-4 5-14 15-29 20-44 45-50 g0-r 70+ Total 0u _aSs 4.,2z 4,056 505 a m 347 746 1.061 lAOS 56 e3 231 233 462 e 1.807 X I. Ca?lwict. Mgnimd P5_1. 1,405 705 697 342 1a 19 35 71 74 151 m 17 s7 35 37 47 124 W Co_ndc A, NImS" A r Po 544 515 220 61 9 14 33 as 50 75 a 9 I1 23 31 5 54 1. TLbamcbst km 173 104 3 1 7 15 47 46 53 3 2 7 16 24 24 30 2 STDs eAgW 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 iL gyphift I 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b. CNmlyd 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 C Gonwrhs 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 d. OwSTDO 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3. MN 0 0 0 0 0 0 0 0 0 0 0 a 0 0 0 0 0 4. DantwIisse 93 44 46 22 1 2 1 2 5 11 26 1 1 1 1 2 14 S. Chlkhood istw 53 25 25 24 2 1 1 0 0 0 22 2 0 0 0 0 0 A Psrmds 17 9 a a I 0 0 0 0 0 a I 0 0 0 0 0 b. Poo 3 I 1 1 0 0 0 0 0 0 1 0 0 0 0 0 0 c. pM_aia 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 d. Mas_N 15 6 7 * 1 0 0 0 0 0 6 1 0 0 0 0 0 a. Tetai 1 10 6 a I 0 1 0 0 0 7 0 0 0 0 0 0 6.Msttngo 41 21 20 7 0 1 5 3 3 2 7 0 1 5 3 3 2 7. Hop"" 34 23 12 2 0 3 5 7 2 4 2 0 1 1 1 3 3 OMa.ba 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9.Tm catotaIsr 1 1 1 o0 0 0 0 0 0 0 0 0 0 0 0 0 a Triypwa.uw 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 b.Ch bgdsica 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 c diSensonisals 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 d. Ldbstmk 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 *. Lwnphdc Ows 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 t Onchomwdt 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 10.L _..y a 0 0 0 a 0 0 0 0 0 0 0 0 0 0 0 0 1 1. Doasi 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 12. Apanme 3 1 1 1 0 0 0 0 0 0 1 0 0 0 0 0 0 13. T,sdhmu 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 14. _is nnhlstodsa 7 4 3 0 3 0 0 0 0 0 0 3 0 0 0 0 0 aAseak 4 2 2 0 2 0 0 0 0 0 0 2 0 0 0 0 a b. T',tdgt 2 1 1 0 1 0 0 0 0 0 0 1 0 0 0 0 0 c Hootwn. 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 d. OUisturtds i 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 15. ois rnt.do. 33 19 14 3 0 1 2 5 3 5 3 0 1 1 2 2 6 B. aRssrsJ.yt dk 474 220 245 132 4 3 2 5 12 71 159 4 3 2 3 9 U4 1. Lowsnasordm 467 225 241 129 4 3 2 5 12 70 156 4 3 2 3 9 63 2. Uppaqumr 5 2 2 1 0 0 0 0 0 1 2 0 0 0 0 0 1 3 Caumsna 2 1 1 1 0 0 0 0 0 0 1 0 0 0 0 0 0 4.Othwrredy 0 0 0 0 0 0 0 0 0 0 0 0 0 0 o 0 0 0 Source: Murray and Lopez, 1996.  Table A3.1. Nwunbr of D_Ihs In CNn. by Ag. SB. and Ci, tsa1 mama m aim Males Mai ms M aIm FanWs Fwnaim Fsr,lim Finales Fwnsis Furake Finfish Conclon Ns DOsU Al Maim AN Finmals 0-4 5-14 15-29 30-44 45-50 0o-6s 70+ 0-4 5-14 15-20 30-44 45-569 o-w 70+ (codinuaM C.Malsmsl cordlom 30 0 30 0 0 0 0 0 0 0 0 0 20 a 2 0 0 1. Hsnonhfi 12 0 12 0 0 0 0 0 0 0 0 0 a 3 1 0 0 ZSpl 1 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 3.Edsnbd 2 0 2 0 0 0 0 0 0 0 0 0 1 0 0 0 0 4. Hypuimson 1 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 B. csbI @6imf 0 0 0 0 0 0 0 a a 0 0 0 0 0 0 0 0 S. Abodon 3 0 3 0 0 0 0 0 0 0 0 0 2 1 0 0 0 7. OhsmaWnml 12 0 12 0 0 0 0 0 0 0 0 0 8 a 1 0 0 D.P_lni co-re 276 133 144 133 0 0 0 0 0 0 144 0 0 0 0 0 0 1. P6gngaiIalsboi 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2. Low bk hwsilt 46 24 22 24 0 0 0 0 0 0 22 0 0 0 0 0 0 3. OM spIrasmdtra41 147 67 60 87 0 0 0 0 0 0 60 0 0 0 0 0 0 4. OUp _ paina 63 41 42 41 0 0 0 0 0 0 42 0 0 0 0 0 0 E. NudloinI 60 31 49 17 3 2 1 1 2 5 31 4 2 3 1 3 6 1. Prn-a_rgrmubillon 35 14 24 8 0 0 0 0 1 4 is 0 0 0 0 1 3 2 bd d5sn"dW a 4 4 3 1 0 0 0 0 0 3 1 0 a 0 0 0 3. 2wnAdnA 9 5 5 3 1 0 0 0 0 0 3 1 0 0 0 0 0 4. AIund 25 8 17 2 1 1 1 1 1 2 5 2 1 3 1 2 3 5. Oum, s _ul 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 N. NcoiwnaLed 6.400 3.531 2.920 66 29 106 199 504 933 1,583 95 19 60 130 376 612 1.606 A. _gens rmeopim 1.484 924 540 5 10 30 66 245 295 250 7 5 20 53 129 151 174 1. Mouth mid omphisaW 30 25 11 0 0 1 5 7 7 5 0 0 1 2 3 3 2 2.EsophapA 1es 129 60 0 0 1 a 31 47 44 0 0 1 1 12 25 23 3.Simwsdc 319 209 1O 0 0 1 9 55 78 o6 0 0 2 9 22 34 43 4. Colorectum 63 48 37 0 0 2 4 11 14 15 0 0 1 3 9 11 13 5 Uw 293 213 0o 0 I 6 41 77 56 34 0 1 2 10 23 22 24 BPe.P ro 32 20 13 0 0 1 1 4 7 7 0 0 0 0 2 5 6 7. Trdi ebondujUamV 216 152 66 0 0 2 3 35 60 51 0 0 1 2 15 20 26 8. heknamsocw slin I I 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9. esl 26 0 20 0 0 0 0 0 0 0 0 0 0 6 a 5 6 1oeCari 21 0 21 0 0 0 0 0 0 0 0 0 0 3 7 a a 11.Corps ulai 6 0 6 0 0 0 0 0 0 0 0 0 0 0 3 1 1 12. Owy 10 0 10 0 0 0 0 0 0 0 0 0 1 2 3 2 2 13. PFi_s 5 5 0 0 0 0 0 0 2 2 0 0 0 0 0 0 0 14.addw le 1s 4 0 0 0 0 2 6 6 0 0 0 0 0 2 2 15. LWq*mm 23 15 6 0 1 3 1 3 5 3 0 0 1 1 3 2 2 1e. Lokqia 66 34 32 3 5 6 a a 3 3 3 3 7 a 5 3 4 17. Ohsrman 115 61 65 2 2 6 9 17 12 13 3 1 3 7 14 12 14 Oth Nsopman 21 11 10 1 0 2 1 2 3 2 1 1 1 1 2 3 2 Rab@ssmsgus 60o 27 33 0 0 1 3 a a 9 0 0 1 2 a 11 12 D C. NubitonWncdoans 14 5 9 0 0 0 1 1 1 2 2 0 I 1 1 1 3 Source: Murray and Lopez, 1996.  TIb A3.1. NHmbet d Des In CNn. by Age, 5e, nnd Cme., IM NW" Mois Mmm Mai Mim Mam Mabs Finis Femab Feuis Fel_ Feuds. Fedsl. Feadma C.n.tn Al ONSO Al Mask Al FOrmais 0-4 8-14 15-26 20-44 45-56 40-` 70+ 0-4 5-14 15-21 30-44 45-5 00-40 70+4 (~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~( 0. Nssm-psdimic 06 53 46 2 2 14 10 5 5 14 2 1 7 7 5 I 1i s 1.Ms N q e a a 0 0 0 0 0 0 0 0 a 0 0 0 0 0 0 2.DOIKd _scvaiddm 2 1 2 0 0 0 0 0 0 0 0 0 0 0 0 0 1 lpa s5 10 5 0 0 a 4 1 1 2 0 0 1 t 1 1 1 4. Epla 12 7 S 0 0 s 1 1 0 1 0 0 2 1 1 0 1 5 Mm .p im. 1 5 4 0 0 0 0 2 1 0 0 0 0 0 0 0 0 0 .L AkVMbkue*u@ndhwdumi Z7 11 16 0 0 1 0 1 2 6 0 0 0 0 1 3 10 7. PuIim..t dmme a a 2 a 0 0 0 0 0 2 0 0 0 0 0 0 2 e.mpbsdmds 5 2 * a a 0 1 1 1 0 0 0 0 1 1 1 0 6. Daigd ndsn0 0 8 0 0 0 0 0 a 0 0 0 0 a a 0 0 10. PTlD 0 0 0 0 0 0 0 0 a O O 0 0 0 0 O a 11. 1s1 mme pa_d 2a 14 12 2 1 a 2 1 1 2 1 1 3 * 1 1 2 E. Sene oqm I 10 a 1 0 0 0 2 3 2 1 0 0 1 2 3 2 1OhAM a 4 3 0 0 0 0 1 1 1 0 0 0 0 1 I I 2z Cidweem 6 3 2 0 0 a 0 1 1 1 0 0 0 0 1 1 1 lOWahrm.nse 6 2 3 a a 0 0 1 1 1 0 0 0 0 1 1 1 F. C _dm.lr 2.U 1.322 1246 10 4 25 47 152 337 704 7 2 25 42 143 254 772 1.UlmAadch_1_deemme 1e3 6e 04 a 1 10 7 13 12 2e 1 1 6 11 19 17 37 2. Mh4c hudIimmm, 76 aes 377 0 0 5 15 53 96 217 o a 3 12 37 74 251 S s C __3.C.,mbeLnaadr 1,272 e72 6o0 2 2 7 16 96 16 3S 2 1 s 13 71 137 372 4. ffu mnycadm s 33 33 2 1 4 3 7 5 11 1 0 4 4 6 4 13 s Hypmauim n 277 151 126 1 0 1 3 20 30 69 1 0 1 2 7 20 66 6. Oeiwdsc 26 12 16 5 1 1 0 1 2 2 3 0 4 0 2 3 4 G.RPAurM. 1,530 789 741 10 1 5 10 67 204 401 14 1 2 10 47 156 5S0 1. COPD 1.432 73 696 4 0 3 7 Q6 195 465 10 0 0 a 44 131 8e0 2.MAIna 35 1 16 0 1 1 2 3 4 a 0 0 1 1 3 3 7 3.OUm ,hqrs.y 63 33 30 5 0 2 1 2 5 16 4 0 1 1 1 3 20 H. r4=mm 411 241 170 17 3 10 32 57 57 65 10 3 a 12 30 32 as 1. PsWw d.mm 33 is 14 0 0 1 1 3 4 0 0 0 0 0 2 a a 2 CkhodsotI9em 167 123 64 2 1 4 24 37 34 23 1 1 2 6 17 16 g0 3.Appmvd.s 12 7 5 0 1 3 2 1 0 0 0 1 2 1 0 0 0 4. Ow dgmm 179 02 87 15 1 3 5 16 le 33 1 1 4 5 11 12 57 I. GOao-mumy 124 73 s0 2 2 9 11 14 25 2 1 6 6 8 11 16 1. NeIftlaMpims gI a 7 42 1 2 9 a 10 t 1 17 2 1 5 7 7 6 12 Z 9 llgn pmsmic h lp*qt 7 7 0 0 0 0 0 0 0 6 0 0 0 0 0 0 0 3. O6 i dW-aisany 16 10 9 0 0 0 0 1 2 5 0 0 0 1 2 3 4 SU1n Crm 12 7 5 0 1 0 0 1 1 4 0 0 1 0 0 1 2 J. Musubcslo-a 36 17 10 0 0 1 0 1 5 10 0 0 1 2 2 2 11 1. PlunmdduhO 2 1 1 0 0 0 0 0 0 1 0 0 0 0 0 0 1 2. Oso.UtiUs 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 Ols1mumaMo-salsel 34 16 16 0 0 1 0 1 4 10 0 0 1 2 2 2 II Source: Murray and Lopez, 1996.  Tdble A&. Numbe d Dr_I In CM.. by Age. Sea. , Caus. IWO ma make Maim ma_m mums M_ai Ma Fram Fe_ma Famai FenaM Famal_a FPaaM PanaM Conlnea Al D0 Al Mae A l FPnl. 0-4 5-14 15-28 30-44 45-5 o0-es 70+ 0-4 5-14 15-2s 30-44 45-59 606 70+ (cen1num K CongudaMmmxUg_ 105 51 54 40 5 5 1 0 0 0 41 5 7 1 0 0 0 1.Abd 5 damd 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Z Anw=W 36 15 24 15 a 0 0 0 0 0 24 0 0 0 0 0 0 3.Anwau.bda0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4S Clullp 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 & CIO E@op d 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 G. EmaphIWaisda 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7.RPnmagu.uds I 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 G. Dom syncen Y 12 6 7 3 1 2 0 0 0 0 3 1 2 0 0 0 0 9. Congeddt tow d 20 16 13 11 2 3 0 0 0 0 7 2 4 1 0 0 0 10.SplnaMlda 12 7 5 a 1 0 0 0 0 0 4 1 0 0 0 0 0 L Oral hah 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1. O alnla 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2. Padodontaldine 0 0 0 0 0 0 0 0 0 0 0 a 0 0 0 0 0 3.Edmgtiy 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4.Ow orenbOM 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 IN. hnMW 1,020 s5c0 43 75 42 154 113 o1 54 71 71 27 114 5s 50 36 75 A. U didem a o2e 400 228 71 35 99 d0 51 30 42 62 24 36 21 22 17 43 %0 1. AMr d baaddwa 135 97 38 4 a 27 26 17 9 7 3 5 9 6 a 3 4 2. Pdeo*nl e5 38 27 5 2 7 7 9 3 a 2 2 9 5 2 3 4 l Fab e5 32 33 2 1 a 4 4 5 10 4 1 1 1 3 a 10 4. Rra 24 14 10 3 1 2 2 1 0 5 2 1 1 1 1 1 4 S. D rmm 147 91 50 30 22 20 6 4 3 5 20 11 7 3 2 2 5 6.0 oewrngonsitnIuJii 190 128 02 26 7 37 23 1e 9 10 27 4 o a a 3 9 IL hkawal 304 190 204 5 3 55 44 30 24 29 9 3 75 36 26 19 32 1.SuIncS-lrd 343 156 154 0 2 42 37 27 23 25 0 2 78 33 26 10 30 Z HwIdd.md 'n' 51 30 20 5 1 13 7 3 2 0 9 1 2 4 2 1 1 3. Ww 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Source: Murray and Lopez, 1996.  Tabm AI. Nui_w d Dgth in Chhba by Age. Sa. ad Come. IWb. by P_n:t _. _.. WM_ _d. Nd. Nd.. Md. F _md.. Funm.. F _nag Fund. F _nd.. F _nm.. Fu_d.. Comiton Al Deiw Nl Mg.. M Fund.. 0-4 5-14 15-2S 30-44 45-50 a0-4S 70+ 0-4 5-14 15--S 30-44 45-SB 00-eo 70+ Tobl D.gt ('000) e.ess 4,829 4.056 506 86 279 347 748 1.01 1,8o5 565 63 231 233 462 a96 1,807 X Petcwt 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 1W.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 1. Comnncmbb, Mdwaii and Peurlal Candflor 15.8% 14.7% 17.2% 67.7% 18.1% 7.0% 10.1% 9.5% 7.0% 8.4% 70.6% 27.1% 16.0% 15.4% 7.9% 6.8% 6.9% A lrEick & Pa,uik 6.1% 6.5% 5.7% 12.0% 10.1% 501% 9.4% 8.7% &6% 4.1% 11.6% 14.3% 4.9% 10.0% a7% 5.0% 3.0% 1. TLfukob 3.1% 3.6% 2.6% 0.5% 0.8% 2.5% 5.2% 5.3% 4.3% 2.9% 0.5% 2.9% 3.1% 6.3% 5.1% 3.5% 1.6% 2. STID ududlng HIV 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% a. SypNblll 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% b. Chanyda 0.0% 00% 0.0% 0.0% 0.0% 0.0% 0o0% 0.0% 0o0% 0.0% 0o0% 0.0% 0.0% 0o0% 0o0% 0o0% 00% c. Gononte 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% d. OUw STD. 0.0% 0o0% 0o0% 0.0% 0.0% 0.0% 0.0% 0o0% 0.0% 0o0% 0o0% 0o0% 0o0% 00% 00% 00% 00% 3 HIV 0.0% 00% 00% 00% 00% 00% 0.0% 00% 00% O0% O% 00% 00% 0.0% 00% 00% 0.0% 4. Diwrt*l dbes. 1.0% 0.9% 1.2% 4.3% 16.% 06% 0.4% 0.2% O0S% 06% 50% 1.9% 0.3% 0.5% 0.3% 0.3% 0.7% &. Childhood cdW 0.0% 0.6% 0.6% 4.7% 2.3% 0.2% 02% 0.0% 0.0% 00% 4.0% 2.6% 0.2% 0.1% 0.0% 0.0% 00% AL Peuauls 0.2% 0.2% 0.2% 1.0% 0.7% 00% 00% 00% 0.0% 00% 1.3% 0 9% 0.0% 00% 0.0% 00% 0.0% b. Polb 0.0% 0.0% 0.0% 0.3% 0.0% 00% 00% 00% 00% 00% 0.2% 00% 00% 00% 00% 00% 00% c. DlpNthmw 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 00% 0.0% 0.0% 0.0% 00% 00% 0.0% 00% d. Munl. 0.2% 0.2% 0.2% 1.3% 1.0% 0.1% 0.0% 0.0% 0.0% 0.0% 1.1% 1.3% 01% 00% 0.0% 0o0% 0.0% & TMwm 0.2% 0.2% 0.2% 1.5% 0.6% 0.2% 0.2% 0.0% 0.0% 0.0% 1.3% 0.4% 01% 0.1% 0.0% 0.0% 0.0% 6.O. wngtin 0.5% 0.4% 0.5% 1.4% 0.1% 0.3% 1.0% 0.4% 0.3% 0.1% 1.2% 0.1% 0.4% 2.1% 0.6% 0.4% 0.1% Cy, 7. Hupds 0.4% 0.5% 0.3% 0.4% 0.4% 0.9% 1.4% 10% 02% 0.2% 0.3% 0.5% 0.5% 0.5% 0.3% 0.4% 0.2% 8. mdu a 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 00% 0.0% 00% 0.0% 0.0% 0.0% 0.0% 0.0% o.0% 0.0% 9. TrqpbuI dier 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% b. TCph' oxrdeek 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0,0% 0.0% 0.0% 0.0% 0.0% 0.0% b.Chmg.s'dlemm 0.0% 0.0% 00% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% cSd. Lomo mis 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% d. LdgIrAIuh 0.0% 00% 0.0% 00% 00% 0.0% 00% 00% 00% 0.0% 00% 0.0% 00% 00% 00% 00% 00% . Lynphok fwadW 0.0_% 0o.% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% t. rOnchooaui 0.0% 00% 0.0% 0.0% 0.0% 010% 00% 0.0% 00% 0.0% 0.0% 0.0% 00% 00% 00% 0.0% 00% 11. LDpiM 0.0% 0.0% 0.0% 0.0% 00% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 1 1. Dungg 0.0% 0.0% 0.0% 0.0% 0.3% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.5% 0.0% 0.0% 0.0% 0.0% 0.0% 13. JT nda.ptuIl 0.0% 0.0% 0.0% 0.1% 0.5% 0.0% 0.0% 00% 0.0% 0.0% 0.1% 0.5% 0.0% 00% 0.0% 00% 0.0% 13.T,tnuch nu 0.0% 0.0% 0.0% 0.0% 0.0% 00% 0.0% 0.0% 0.0% 0.0% 0.0% 00% 0.0% 0.0% 0.0% 0.0% 00% 14.Irimdlnrmdodes 0.1% 0.1% 0.1% 0.0% 37% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 4.7% 0.0 00% 0.0% 0.0% 0.0% b. ATdch 0.0% 0.0% 0.0% 0.0% 13% 0.0% 0.0% 00% 0.0% 0.0% 0.0% 2.9% 0.0% 0.0% 0.0% 0.0% 0.0% b. iTr mhL 0.0% 0.0% 0.0% 0.0% 15.% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 1.6% 0.0% 0.0% 0.0% 0.0% 0.0% d Hoo_ ir" 00.% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0 0% 0.0% 0.0% 0.0% 0.0% 0 0% 0.0% 0.0% 0.0% d. OkhwrItri.in 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 15& omulrdiros 0.4% 0.4% 0.3% 0.5% 0.4% 0.3% 0.5% 0.7% 0.3% 0.3% 0.5% 0.6% 0.3% 0.4% 0.3% 0.2% 0.3% B.R*Wlrd kiwloi 5.3% 4.8% 6.0% 2t.1% 5.0% 1.1% 0.5% 0.7% 1.1% 3.9% 28.2% 6.9% 1.4% 0.8% 0.6% 1.4% 3.5% 1. Low mepkidy 5.3% 4.7% 5.9% 25.6% 4.9% 1.1% 0.5% 0.7% 1.1% 3.9% 27.7% 6.8% 1.4% 0.7% 0.6% 1.3% 3.5% ZLUpperrpkiray 0.1% 0.0% 0.1% 0.3% 0.0% 0.0% 00% 0.0% 00% 0.0% 03% 0,1% 00% 0.0% 00% 00% 0o0% Source: Murray and Lopez, 1996.  Tabl A32. Numw o Ded.h hI Ch. by Ago, Sm. aid Cme. 1994 by Pe_ Him Mi. Hi. mim Him Mi.. Mi.. F_nd.s Fumdin F_ndi F_mim F_nd. Fu_di Frndl. Cond_bn AN DmI A Allml.. AN Fwmis 0-4 5-14 15-29 30-44 45-50 0-89 70+ 0-4 5-14 15-29 30-44 4S-50 0-69 70+ 3. O, Ut medIs 0.0% 0.0% 0.0% 0.2% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.2% 0.1% 0.0% 0.0% 0.0% 0.0% 0.0% 4. ofM*" 0.0% 00% 0.0% 00% 00% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% C. Materrldcondll" 0.3% 0.0% 0.7% 0.0% 0.0% 0e0% 0.0% 0.0% 0.0% 00 % 0.0% MO% a.8% 3.5% 0.4% 0.0% 0.0% 1. Hontu 0.1% 0.0% 0.3% 0.0% 0.0% 00% 00% 00% 00% 0.0% 00% 0.0% 3.5% 1.4% 0.2% 00.% 0.0% Z S s 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.4% 0.2% 0.0% 0.0% 0.0% 3, Edwpmnn 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.5% 0.2% o.O% 0.0% 0.0% 4. H ^ ypwpt.rim 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 00% 0.0% 0.0% 0.0% 0.0% 0.0% 0.2% 0.1% 0.0% 0.0% 0.0% 5. b d kbr 0.0% 0.0% 0o0% 0.0% 00.% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.1% 0.0% 0.0% 0.0% 0.0% t Abon 0.0% 0.0% 0.1% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.7% 0.3% 0.0% 0.0% 0.0% 7. OUm nuw 0.1% 0.0% 0.3% 0.0% 0.0% 00% 0.0% 0.0% 0.0% 00% 0.0% 0.0% 3.4% 1.3% 0.1% 0.0% 0.0% D. Pulriouidiolq 3.1% 2.7% 3.5% 26.3% 0.0% 0.0% 0.0% 0.0% 00% 0.0% 25.4% 00% 0.0% 0.0% 0.0% 0.0% 0.0% 1. PW r INdlao 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 00.% 00% 0.0% 0.0% 0.0% 0.0% 0.0% 00% 2 Low bi,f ndgt 0.5% 0.5% 0.5% 4.8% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 3.9% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 3 BiEh phyio a-d trarna 1.7% 1.4% 2.0% 13.2% 0.0% 00% 0.0% 00% 0.0% 0.0% 14.2% 0.0% 0.0% 00% 0.0% 00.% 0.0% 4. 0U p.dril 0.9% 0.9% 1.0% 8.2% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 7.4% 0.0% 0.0% 0.0% 00% 0.0% 00% E. NlNk 0.9% 0.6% 1.2% 3.3% 3.0% 0.6% 0.2% 0.1% 0.2% 0.3% 5.4% 5.9% 0.9% 1.2% 0.2% 0.4% 0.3% 1. P,shn-ngr mdrujrib3n 0.4% 0.3% 0.6% 1.7% 0.1% 01% 0.0% 0.0% 0.1% 0.2% 3.4% 0.1% 0.1% 0.0% 0.0% 0.1% 0.2% 2 iodkudibdIry 0.1% 0.1% 0.1% 0.6% 0.7% 0.1% 0.0% 00% 0.0% 0.0% 0.5% 0.8% 0.1% 00% 00% 0.0% 00% 3.Vtaam Adibcy 0.1% 0.1% 0.1% 0.e% 1.5% 0.1% 0.0% 0.0% 0.0% 0.0% 0.6% 2.0% 0.1% 0.0% 00.% 0.0% 0.0% 4. Anwmbi 0.3% 0.2% 0.4% 0.3% 0.7% 0.5% 0.2% 0.1% 0.1% 0.1% 0.9% 2.9% 0.5% 1.1% 0.2% 0.3% 0.2% 5. CoUw r bhru 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0 .0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 00% 00% 11. N o,_anunbable 727% 73.1% 72.2% 17.4% 33.5% 37.8% 57.3% 79.6% 87.9% 87.7% 16.9% 30.4% 34.6% 5s.9% 81 3% ee 1% 69.0% AMiigr-t nwpim. 1a5% 19.1% 13.3% 10% 11.6% 107% 24.7% 332x 27.a% 13.9% 1.2% 79% 8.6% 227% 279% 21.% 9.6% 1. Mouihwid omp.ym 0.4% 0.5% 0.3% 0.0% 0.2% 0.4% 15% 0.9% 0.6% 0.3% 00% 0,0% 0.4% 0.8% 0.7% 0.4% 0.1% ZEgophmgLs 2.1% 2.7% 1.5% 0.0% 00% 0.2% 1.9% 4.2% 4.4% 2.4% 00% 0.0% 0.3% 0.4% 2.5% 3.5% 1.3% 3.8tomch 3&8% 4.3% 27% 00% 0.2% 0.4% 27% 7.4% 7.3% 3.7% 0.0% 0.3% 10% 3.9% 4.8% 4.9% 2.4% 4. CQkrdAn 0.9% 1.0% 0.9% 0.0% 0.2% 0.8% 1.3% 1.4% 1.3% 0.9% 0.0% 0.4% 0.4% 1.5% 1.9% 1.6% 0.7% 5. L1e 3.3% 4.4% 2.0% 00% 0.7% 2.0% 11.7% 10.3% 5.2% 1.9% 00% 0.9% 0.7% 4.1% 4.9% 3.1% 1.3% e.Parcm 0.4% 04% 0.3% 0.0% 0.0% 0.2% 0.3% 0.5% 0.e% 0.4% 0.0% 00% 01% 0.1% 0.4% 0.7% 0.3% 7. TmNw6BwrohmLwA" 2.5% 3.2% 1.6% 0.0% 0.5% 0.7% 0.9% 4.7% 5.7% 2.8% 0.0% 0.0% 0.3% 0.9% 3.3% 2.9% 1.5% 8. Miara and othrwIdn 0.0% 00% 0.0% 00% 0.0% 0.0% 0.0% 0.0% 00% 0.0% 00% 0,0% 00% 0.0% 00% 0.0% 0.0% 9. Bre 0.3% 0.0% 06.% 00% 0.0% 0.0% 0.0% 0.0% 0.0% 00% 00% 0.0% 01% 27% 1x% 07% 0.3% W.CerV13 0.2% 00% 0.5% 0.0% 00% 0.0% 0.0% 00% 00% 00% 00% 00% 0.2% 1.1% 1.4% 0.9% 0.3% 11.Co,pwiirl^ 0.1% 00% 0.2% 00% 00% 00% 0.0% oO% 0.0% 0.0% 00% 0.0% 0.1% 0.2% 06% 0.2% 01% 12 Owy 0.1% 00% 0.3% 0.0% 00% 00% 0.0% 0.0% 0.0% 0.0% 0.0 0.3% 04% 0.8% 0.7% 0.3% 0.1% I&Pfoie 0.1% 0.1% 0.0% 00% 00% 00% 00% 0.0% 0.2% 0.1% 00% 0.0% 00% 0.0% 0.0% 0.0% 0.0% 14. BIddw 0.2% 0.3% 01% 00% 00% 0.1% 01% 0.3% 0.6% 0.3% 0.0% 0.0% 0.0% 0.0% 0.1% 0.2% 0.1% 15. Lyrrpho. 0.3% 03% 0.2% 00% 1o0% 0.9% 0.2% 05% 0.5% 0.2% 0.0% 0.0% 0.3% 0.3% 0.6% 0.2% 0.1% 1. LukwnIa o7% 0.7% 0.8% 0.6% 5.% 2.9% 1.7% 0.6% 0.3% 0.2% 0.5% 4.1% 3.2% 2.8% 1.2% 0.5% 0.2% 17. Olucem 1.3% 1.3% 1.3% 0.3% 2.9% 2.0% 2.5% 2.3% 1.1% 0.7% 0.6% 1.9% 1s% 3.1% 3.0% 1.7% 0.8% OltwN.oplnwm 0.2% 02% 0.2% 0.2% 05.% 0.5% 0.3% 0.3% 0.3% 0.1% 0.2% 0.9% 0.3% 0.5% 0.3% 0.4% 0.1% x Source: Murray and Lopez, 1996.  T bl A3.2. Numbhc cd Doihs hi Chins by Age. Sax, and Ce. 1990. by P:mwu md.. -Nd Nd.. Md1_ mdim N Md. F1und.._ Femae Fendes Fn. Foud. Found. FondF Ckwnmi2 M Death Al Nd.. Al Foud. 0-4 5-14 t5-20 30-44 45-59 a0-89 70+ 0-4 5-14 15-20 30-44 45-59 60-89 70+ S. O1bi. melu. 0.7% 0.6% 0.e6 0.0% 0.4% 0.3% 0.8% 0.8% 0.7% 0.5% 0.0% 0.5% 0.5% 0.7% 1.4% 1.6% 0.7% X C. NulbonW ndsln. 0.2% 0.1% 0.2% 0O% 0.2% 0.1% 0.2% 0.1% 0.1% 0.1% 0.3% 0.7% 0.5% 0.5% 0.1% 0.1% 0.2% D. Nou-pyhNift 1.1% 1.1% 1.1% 0.5% 2.2% 4.8% 2.9% 0.7% 0.5% 0.5% 0.3% 1.4% 3.0% 2.9% 1.1% 0.9% 1.0% 1. MUsordNsdw diorder 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 00% 00% 0.0% 0.0% 0.0% 0.0% 2I d idwdIWmdiv dor 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 00% 0.0% 0.O% 0.0% 0.0% 0.0% 00% 0,0% 0.0% 0.0% 0o1% 3.Psyd 0.2% 0.2% 0.1% 0.0% 0.0% 0.9% 1.0% 01% 0.1% 0.1% 0.0% 0.0% 0.5% 0.6% 0.2% 0.1% 0.1% 4.Ephp 0.1% 0.1% 0.1% 0.0% 0.4% 1.1% 0.4% 0.1% 0.0% 0.0% 0.0% 0.2% 0.9% 0.% 0.1% 0.1% 0.0% 5*Abohddpmou 0.1% 0.1% 0.0% 0.0% 0.0% 0.2% 0.7% 0.1% 0.0% 0.0% 0.0% 0.0% 0.0% 01% 0.0% 0.0% 0.0% * AkhdnranddCwduiw1ia 0.3% 0.2% 0.4% 0.1% 0.3% 0.2% 0.1% 0.2% 0.2% 0.3% 0.1% 0.1% 0.2% 0.2% 0.3% 0.4% 0.6% 7.Poutdn's di..a. 0.1% 0.1% 0.1% 0.0% 0.0% 0.% 0.0% 0.0% 00% 01% 0.0% 00% 00% 00% 0.0% 0.0% 01%