Report No. 2221 0-VN Vietnam Growing Healthy: A Review of Vietnam's Health Sector June 29, 2001 Human Development Sector Unit Vietnam Country Unit East Asia and Pacific Region Document of the World Bank CURRENCY EQUIVALENTS Currency Unit = Dong Exchange rate as of May 2001: US$1 = VND 14,710 GOVERNMENT FISCAL YEAR January 1 to December 31 Acknowledgments This report is the outcome of a cooperative effort between the Ministry of Health and a number of donors in the health sector. The Ministry made available a great deal of in- formation and data and commented extensively on drafts of the report. The donors are grateful to the Ministry for making this cooperation so productive. Richard Meyers (Task Team Leader) and Nguyen Thi Mai (Operations Officer) of the World Bank managed the preparation process on the donor side. Professor Anil Deola- likar was the primary consultant and directed the analysis of the background data and the writing of the report. Funding for the production of the report was provided by Sida, AusAID, the Royal Netherlands Embassy, and the World Bank. WHO, UNICEF, UNFPA, Sida, the Royal Netherlands Embassy, AusAID and the Asian Development Bank pro- vided other support and valuable comments on its contents. In addition, a number of gov- ernment agencies, NGOs and civil-society groups were consulted throughout the prepa- ration of the report. ACRONYMS AND ABBREVIATIONS ADB Asian Development Bank AFAP Australian Foundation for the Peoples of Asia and the Pacific Limited APC Aspirin, phenacetin, and caffeine ARI Acute respiratory infection ASEAN Association of Southeast Asian Nations AusAID Australian Agency for International Development BCG Bacillus Calmette-Guerin (tuberculosis vaccine) BfdW Brot fir die Welt CBPHC Community Based Primary Health Care CDD Control of Diarrheal Disease CHC Commune health center CIDSE International Cooperation for Development and Solidarity CPCC Committee for the Protection and Care of Children CPR Contraceptive prevalence rate DAV Drug Administration of Vietnam DOTS Directly Observed Treatment Short Course DPT Vaccine against diphtheria, pertussis, and tetanus EC European Community ENT Ear-nose-throat EPI Expanded Program of Immunization EU European Union GDP Gross domestic product GFR General fertility rate GMP Good Manufacturing Practice (ASEAN) GSO General Statistical Office HHTK Phenobarbital, phenacetin, aspirin HHW Hamlet health worker ICDS Intercensal Demographic Survey ICP Intercommunal polyclinic IDS Institute of Development Studies IEC Information, Education and Communication IMPE Malariology, Parasitology and Entomology IMR Infant mortality rate MCH Maternal and Child Health MCNV Medical Committee Netherlands-Vietnam MMR Maternal mortality rate MoET Ministry of Education and Training MoH Ministry of Health MPI Ministry of Planning and Investment MR Menstrual regulation NCHS National Center for Health Statistics NCPFP National Committee for Population and Family Planning of Viet Nam NGO Nongovernmental organization NIN National Institute of Nutrition ODA Overseas development assistance iii OPV Oral polio vaccine PCD Project Coordination Department PHB Provincial Health Bureau PPC Provincial People's Committee RTI Reproductive tract infection SCC Short Course Chemotherapy SDA Support for Disadvantaged Areas Sida Swedish International Development Cooperation Agency STD Sexually transmitted disease TB Tuberculosis TFR Total fertility rate UNDP United Nations Development Programme UNICEF United Nations Children's Fund UNFPA United Nations Population Fund VHI Vietnam Health Insurance VHIA Vietnam Health Insurance Authority VHW Village health worker VLSS Vietnam Living Standards Surveys VNDHS-II Vietnam Demographic and Health Survey VNDP Vietnam National Drug Policy WFP World Food Programme WHO World Health Organization Iv CONTENTS I. INTRODUCTION............................................... A. Background and Motivation....................................... I B. Importance of Health to the Economy ..........2........ ............2 C. Scope of the Report ............................................. 3 D. Structure of the Report.................4.......... .............4 It. STATUS OF AND TRENDS IN HEALTH INDICATORS ...................5 A. Morbidity and Mortality.........................................5 1. Infant and Child Mortality ..................................... 5 2. Main Causes of Morbidity and Mortality .....................11 3. Major Communicable and Noncommunicable Health Problems .... ...... 12 4. Smoking as a Behavioral Health Risk Factor....................... 16 B. Reproductive Health and Fertility Outcomes ......... ................. 21 1. Maternal Mortality ............................................ 21 2. Fertility................... ......................... .... 22 3. Other Reproductive Health Indicators ............................ 25 C. Nutritional Outcomes ..........................................27 1. Child Nutritional Status ............................... ..27 2. Micronutrient Deficiencies.............................. 32 Ill. DEMAND FOR HEALTH SERVICES ..................... ..........33 A. Levels, Trends and Disparities in Utilization............................ 33 1. Facility-Based Estimates of Overall Utilization........ ... .......... 33 2. Survey-Based Estimates of Overall Utilization........... ................35 3. Demographic and Regional Variations in Total Utilization . ............. 36 4. Health-Seeking Behavior of Individuals.......................... 38 5. Who Uses Public Health Facilities? .............................41 6. Hospitalization............................................. 43 B. Utilization of Preventive Health Services ............................. 45 C. Determinants of Utilization............ ................... ......46 D. The Private Cost of Health Care ...................................49 1. Out-of-Pocket Expenditure on Health Services Contacts in 1993 and 1998.... 49 2. Health Care Costs in Relation to Discretionary Income................53 3. Budget Share of Health Expenditure ................... ..........55 V 4. Regional Variations ......................................... 56 5. International Comparisons .................................... 57 6. Private Costs of Hospitalization...........................58 7. Health Care Costs: A Qualitative Picture ....................... 59 CHAPTER III ANNEX TABLES ............................... .....61 IV. THE PUBLIC PROVISION OF HEALTH SERVICES....................65 A. Structure of the Public Health Care System ................................ 65 B. Operations of Commune-Level Health Activities .................. ..... 68 1. Number and Regional Distribution of Commune Health Centers................ 68 2. Utilization of Commune Health Centers...................... 69 3. Reasons for Low Number of Curative Care Consultations..........................70 4. Management of Commune Health Activities.....................72 5. Evolution of Primary Health Network.............. ..............73 C. National Public Health Programs...................................74 1. National Malaria Control Program..............................74 2. National Tuberculosis Control Program...........................75 3. Expanded Program of Immunization.............................76 4. Leprosy Elimination Program...................................77 D. Hospitals................................................... 77 1. Distribution and Number..................................... 77 2. Utilization Rates........................................... 79 3. Financial Structure of Hospitals ....................... ......... 80 4. Hospital Efficiency and Marginal Costs .................. ......... 82 5. Differences in Performance Among Categories of Hospitals..............85 CHAPTER IV ANNEX TABLES ........................... ......... 87 V. THE ROLE OF PRIVATE FOR-PROFIT AND NONPROFIT PROVIDERS..........93 A. Legal Framework for Private For-Profit Health Services................... 93 B. Private Clinics and Providers ..................................... 94 1. Classification of Private Health Providers ............... .......... 94 2. Size and Growth of the Private Sector in Health Services Provision...........95 3. Regional and Provincial Distribution .................... ......... 97 4. Utilization of Private Health Clinics............................. 98 5. Why Do Patients Use Private Providers?............... ............ 99 Vi C. Private Inpatient Services..................... ..................101 1. Private Hospitals in Vietnam.................. ...............101 2. Semipublic Health Services in Public Hospitals ....................102 3. Contracting Out of Services..................................103 D. Monitoring and Regulation of the Private Sector..................104 E. Pharmaceuticals and Drugs......................................106 1. Drug Production and Imports................................. 106 2. Distribution of Drugs through Pharmacies and Drug Vendors................ 107 3. Health Effects of Drug Use: The Case of Antibiotic Resistance .........110 4. Recent Policy Actions........................................ 1 F. The Role of Nongovernmental Organizations in Health Services Delivery....... 112 1. Number and Distribution.................................... 112 2. History of NGOs in the Vietnam Health Sector ....................113 3. Role and Impact of NGOs on Health.......... .................. 114 4. Changes in Scope and Approach of NGOs .................... .... 115 3. Examples of NGO Initiatives ............................ ..... 116 6. NGO Advocacy .................................... ...... 117 CHAPTER V ANNEX TABLES Organizations.in. Hey.................................... 119 V. UMAN RESOURCES IN EALTH... ........................ ....... 12 A. Supply of Health Workers ................................ ...... 121 B. Productivity of Health Workers ........................... ............ 126 C. Salaries of Health Workers ............................. ......... 127 D. Training of Health Workers .......................... .............128 1 . Preservice Training: Planning and Development .............................. 128 2. Preservice Training: Selection and Quality Monitoring of Education......130 3. Postgraduate Training ............................................................ 132 4. Training in Traditional Medicine ................................................ 132 5. Village Health Workers.....................................133 6. In-Service Training .........................................134 7. Future Plans for Training of Health Personnel .....................136 E. Human Resource Management................................... 136 CHAPTER VI ANNEX TABLES................................... 138 VII. FINANCING OF EALTH EXPENDITURE .................................. 142 vii A. Sources of Financing ........... 142 1. Structure of Health Financing in Vietnam ........................ 142 2. Public-Private Financing of Health ........................ .....1 44 3. Sources of Financing Public Health Services ................. .....1 45 B. Cost Recovery ....................................... ........ 146 1. Policy ............................................ ..... 146 2. Cost Recovery Rates....................................... 147 3. Impact of Cost Recovery on Utilization of Health Services ...... ...... 148 4. Burden of User Fees on the Poor .................................... 149 5. Discrepancy between Official User Fees Collected and User Fees Paid by Households.............................................. .... 152 6. Provincial Distribution of User Fees ....................... ..... 152 C. Health Insurance ........................................ .....1 54 1. Description of the Program ........................... ........ 154 2. Population Coverage ................................. ........ 155 3. Role of Insurance in Financing Public Health Spending ........ ....... 158 4. Equity in Health Insurance ................................... 158 5. Level of Utilization of Health Insurance Fund Across Provinces.................. 160 6. Financial Viability.......................... ............. 160 D. Commune Health Center Finance ................................. 164 CHAPTER VII ANNEX TABLES .............................. ..... 166 VIII. PUBLIC EXPENDITURE ON HEALTH..................... ....... 170 A. Overall Health Expenditure.....................................170 1. Public Spending..........................................170 2. Total (Public and Private) Spending.............. ...............171 B. Cost-Effectiveness of Public Health Spending.........................172 C. Composition of Public Health Expenditure ...........................173 1. Functional Composition.......................... ................173 2. Economic Composition............................ ...............176 3. Composition by Service Level................................178 4. Provincial Distribution...................................... 179 D. Overseas Development Assistance to the Health Sector ..................181 1. Distribution of ODA by Donor................................ 181 viii 2. Provincial Distribution of ODA ........................ .....1 83 3. Donor Coordination ......................................... 184 IX. SUMMARY OF MAJOR FINDINGS ....................... ...... 186 A. Past and Present Trends ......................................... 186 B. Attaining Health Goals in a Market Economy ....................... 189 C. Summary of Major Challenges for Vietnam's Health Sector ..... ........ 189 1. Important Risks to Future Health ................................... 189 2. Problems Relating to Essential Health Services ............... ..... 190 3. Issues in Health Care Financing................. .................. 191 4. Health Sector Capacity Issues........................ ........ 192 BIBLIOGRAPHY ............................................ 195 List of Tables 2.1: Infant mortality, 1983 2.2: Infant, child and under-five mortality rates, 1979-83 to 1992-96 2.3: Infant and child mortality, by residence and region, 1984-93 2.4: Infant mortality decline by region, 1989-93 2.5: Infant and under-five mortality, by wealth quintiles, 1997 2.6: Infant and under-five mortality by sex, 1984-93 and 1992-97 2.7: Infant, child and under-five mortality rates, by education and demographic characteristics, 1987-96 2.8: Leading causes of morbidity and mortality in hospitals, 1998 2.9: Morbidity and mortality rate per 100,000 inhabitants of vaccine-preventable deaths 2.10: Tobacco use by urban/rural residence and educational level, 1998 (percent) 2.11: Tobacco use by per capita expenditure quintiles, 1998 (percent) 2.12: Tobacco use by region, 1998 (percent) 2.13: Changes in tobacco use between 1993 and 1998 2.14: Maternal mortality rate in selected provinces, 1991-93 2.15: Trends in age-specific and total fertility rates from various sources, 1987-96 2.16: Total fertility rates, by residence and region, 1992-96 2.17: Decline in general fertility rate by region, 1988-89 to 1993-94 2.18: Total fertility rates, by women's schooling level, 1997 2.19: Weight of newborns in seven provinces, 1999 2.20: Rates of malnutrition by per capita expenditure quintile, 1993 and 1998 3.1: Annualized health services contact rates, by provider and per capita expenditure quintile, 1993 and 1998 3.2: Annualized health services contact rates per capita, by sex and age group, 1998 3.3: Annualized health services contact rates per capita, by region, 1998 3.4: Distribution of health services contacts by providers and per capita expenditure quintiles, 1993 and 1998 ix 3.5: Distribution of inpatient and outpatient health services contacts by providers and per capita expenditure quintiles, 1993 and 1998 3.6: Inpatient admission rate and average length of hospital stay, by per capita expenditure quintile, 1993 and 1998 3.7: Immunization coverage of children under 1 year of age, 1986-96 3.8: Average out-of-pocket expenditure on fees and drugs per health services contact (in '000 constant 1998 VND), by provider and per capita expenditure quintile, 1993 and 1998 3.9: Average out-of-pocket expenditure on fees per health services contact (in '000 constant 1998 VND), by provider and per capita expenditure quintile, 1993 and 1998 3.10: Average out-of-pocket expenditure on drugs per health services contact (in '000 constant 1998 VND), by provider and per capita expenditure quintile, 1993 and 1998 3.11: Affordability ratios for health costs at public facilities (per-contact out-of-pocket expenditure on fees and drugs as % of annual nonfood consumption expenditure per capita), by provider and by per capita expenditure quintile, 1993 and 1998 3.12: Affordability ratios for health costs at public facilities (per-contact out-of-pocket expenditure on fees and drugs as % of annual nonfood consumption expenditure per capita), by region and per capita expenditure quintile, 1998 A3.1: Distribution of health services contacts at public and private health facilities, 1993 &nd 1998 A3.2: Cumulative distribution of health services contacts at public and private health facilities, 1993 and 1998 A3.3: Immunization coverage among children under 1 year of age, 1997 A3.4: Affordability ratios for health services, Cambodia (out-of-pocket expenditure per contact as % of nonfood consumption expenditure per capita), 1997 4.1: Population per commune health center, 1997 4.2: Mean number of daily service contacts at commune health centers, by region, 1997 4.3: Percentage of communes without a health center and staffing at commune health centers, by region, 1997 4.4: Immunization coverage of children under 1 year of age, 1986-96 4.5: Morbidity and mortality rate per 100,000 inhabitants of vaccine-preventable deaths 4.6: Number of hospitals by level and specialty, 1997 4.7: Total number of provincial and district hospitals and hospital beds, by region, 1997 4.8: Summary statistics on inpatients, by type of hospital, 1996 4.9: Average annual real growth rate in mean revenue per hospital (%), by source and by type/level of hospital, 1994-96 4.10: Mean revenue per hospital (millions of VND), by source and by type/level of hospital, 1996 4.11: Average annual real growth rate in mean expenditure per hospital (%), by source and by type/level of hospital, 1994-96 4.12: Summary of marginal cost, economies of scale, and economies of scope measures for 656 hospitals in Vietnam, 1996 4.13: Hospital performance indicators across types and levels of hospitals, 1996 X A4. 1: Number of commune health center facilities and beds, by province, 1997 A4.2: Annual health service contact rates at commune health centers, by province, 1997 A4.3: Total number of provincial and district hospitals and hospital beds, by province, 1997 5.1: Number of licensed private health personnel, by category, as of October 1996 5.2: Number of private health facilities, by type (as of October 1998) 5.3: Licensed and unlicensed private health facilities in 44 of 61 provinces, by ru- ral/urban areas, 1999 5.4: Average number of public and private sector health workers per commune, by re- gion and worker category, 1995 5.5: Number of private health facilities found violating regulations and laws, by cate- gory of facility, 1999 5.6: The most commonly dispensed drugs ranked by number of times sold in a two- week period, Hanoi, 1994 5.7: Antibiotic resistance of Salmonella typhii in Vietnam, 1993-98 5.8: Changes in NGO focus over time A5. 1: Number of private health facilities, by category and rural/urban, 1999 A5.2: Private health facilities (excluding traditional healers) in 44 surveyed provinces, 1999 6.1: Number of commune, district and provincial-level health workers per commune, by worker category and region, 1997 6.2: Staffing at commune health centers, by region, 1997 6.3: Average number of public and private sector health workers per commune, by re- gion and worker category, 1996-97 6.4: Average monthly salaries of health workers, 1992-97 A6. 1: Numbers of health personnel in the public sector, by category, 1997 A6.2: Number of commune, district and provincial-level health workers per commune, by category and province, 1997 A6.3: Total monthly compensation of provincial, district and commune-level health workers in Vinh Long province, 1992 and 1997 (in '000 constant 1997 VND) 7.1: Real spending on public health services, by financing source, 1991-98 7.2: Percent of users who reported paying nothing for a visit to a government health facility, 1993 and 1998 7.3: Percentage of fee exemptions and reductions in selected hospitals, 1993-97 7.4: Share of target population covered by health insurance, 1993-98 7.5: Composition of voluntary health insurance scheme membership, 1993-97 (%) 7.6: Service use and cost of services to VHIA, 1997 7.7: Shares of 10 provinces in national health insurance expenditures, 1997 7.8: Service Use and Health Expenditures of Health Insurance Enrollees, 1993-97 7.9: Annual health service contacts, by insurance status and per capita expenditure quintile, 1998 7.10: Inpatient admissions rate and average length of stay by insurance status and per capita expenditure quintile, 1998 XI A7. 1: Part A. Price scale for consultation and physical examination (in VND) A7.2: Part Bl. Price scale for one day of stay (inpatient) (in '000 VND) A7.3: Part B2. Maximum price scale for one inpatient treatment day (in '000 VND) A7.4: Part C. Price scale for technical services and biology and biochemistry lab tests and x-rays 8.1: Public spending on health, latest year available (1991-96), Asia 8.2: Functional composition of public expenditure on health, by financing source, 1991-98 8.3: Ministry of Health Expenditure on Health Programs, 1993-97 8.4: Distribution of public health spending by source and by economic type, 1991-98 8.5: Recurrent Health Sector Expenditures, 1991-98 8.6: ODA commitments and disbursements in the health sector through the end of 1998, by country 8.7: Overseas development assistance to the health sector (annual disbursements in '000 US$), 1991-98 8.8: Distribution of total ODA disbursements to the health sector over the period 1991-98, by province List of Figures 2.1: Infant mortality rate, 1960-99 2.2: Infant mortality rates and real GDP per capita across Asian countries, 1997 2.3: Decline in infant mortality rates between 1970 and 1997 in relation to real GDP per capita, selected Asian countries 2.4: Major causes of morbidity, 1976-97 2.5: Major causes of mortality, 1976-97 2.6: Morbidity and mortality rates from malaria, 1976-97 2.7: Morbidity and mortality rates from dengue fever, 1991-98 2.8: Number of HIV/AIDS cases detected and number of HIV/AIDS affected provinces, 1990-99 2.9: Average annual number of cigarettes smoked per adult (>15 years), Asia, 1990-98 2.10: Main causes of maternal mortality, 1997 2.11: Decline in total fertility rate between 1975 and 1997 in relation to real GDP per capita, selected Asian countries 2.12: Total fertility rates and real GDP per capita across Asian countries, 1997 2.13: Contraception prevalence rates and real GDP per capita across Asian countries, 1997 2.14: Age-specific prevalence of stunting, underweight and wasting among children aged 0-59 months, 1998 2.15: Prevalence of underweight, stunting and wasting among children under age 60 months, 1982-85 to 1998 2.16: Heights and weights of children aged 1-5 years, by real per capita consumption expenditure, 1993 and 1998 2.17: Prevalence of underweight, stunting and wasting among children under age 60 months, by sex, 1998 Xii 2.18: Prevalence of underweight, stunting and wasting among children under age 60 months, by region, 1998 2.19: Malnutrition and infant mortality across countries in Asia, 1995-97 2.20: Prevalence of underweight children aged 0-59 months, selected countries in Asia, 1975-95 3.1: Utilization rates of public facilities, 1976-98 3.2: Annual public sector health contacts per capita, Asia, 1990-95 3.3: Hospital bed occupancy rate and inpatient days per capita, 1976-98 3.4: Annual inpatient admissions per 1,000 persons, Asia, 1990-95 3.5: Average annual health service contacts per capita, by per capita consumption expenditure, 1993 and 1998 3.6: Economic composition of users of different health providers, 1993 3.7: Economic composition of users of different health providers, 1998 3.8: Lorenz distributions of service contacts with public hospitals, 1993 and 1998 3.9: Lorenz distributions of service contacts with public hospitals and commune health centers in comparison to Lorenz distribution of aggregate consumption expendi- ture, 1998 3.10: Average length of hospital inpatient stay (days), Asia, 1990-95 3.11: Lorenz distributions of hospital inpatient days, 1993 and 1998 3.12: Food, nonfood and medicine price indices, 1992-98 3.13: Donor spending on drug supply, 1995-98 3.14: Share of health in total consumption and nonfood expenditure, 1993 and 1998 3.15: Out-of-pocket expenditure per hospital admission (in '000 1998 VND), 1993 and 1998 3.16: Out-of-pocket expenditure on hospitalization per night spent in hospital (in '000 1998 VND), 1993 and 1998 3.17: Household financing of treatment expenditures, 32 communes in Northern Viet- nam, 1995 4.1: Morbidity and mortality rates from malaria, 1976-97 4.2: Number of hospital beds per 10,000 population, 1975-97 4.3: Persons per hospital bed, selected Asian countries, 1995 5.1: Number of private health facilities and GDP per capita across provinces, 1998 5.2: Health service contacts with private clinics and private providers as percentage of all health service contacts in the last 4 weeks, by per capita expenditure quintile, 1993 and 1998 5.3: Average out-of-pocket expenditure per health service contact, by per capita expenditure quintile and by provider, 1998 5.4: Average out-of-pocket expenditure on drugs per health service contact (across all health providers), by per capita expenditure quintile, 1998 5.5: Food, nonfood and medicine price indices, 1992-98 5.6: Average annual number of health service contacts per capita with drug vendors and pharmacy shops, by per capita expenditure quintile, 1993 and 1998 5.7: NGO assistance, 1992-98 Xiii 6.1: Number of health workers per 100,000 population, by worker category, 1976-96 6.2: Doctors per 100,000 population, selected Asian countries, 1993 6.3: Nurses per 100,000 population, selected Asian countries, 1993 6.4: Ratio of nurses to doctors, selected Asian countries, 1993 6.5: Annual number of public sector health service contacts per public health worker, 1976-98 6.6: Average annual number of service contacts at the commune level per commune health worker, by region, 1997 7.1: Public-private shares in financing total health expenditure, 1993 and 1998 7.2: Public spending on health as a percentage of aggregate health spending, selected Asian countries, 1990s 7.3: Sources of funding public expenditure on health, 1991-98 7.4: User fee collections, 1991-97 7.5: The share of hospital user fees and state budget subsidies in total hospital revenue, 1994-98 7.6: Utilization rates at public facilities, 1976-98 7.7: User fee collections per capita and provincial domestic product per capita, Vietnam provinces, 1997 7.8: Real growth in user fee collections over 1993-97 in relation to the level of user fees per capita in 1993, Vietnam provinces 7.9: Per capita public health spending out of user fees and out of the state budget, 1997, Vietnam provinces 7.10: Enrollment in health insurance, 1993-98 7.11: Percentage of total population covered by health insurance, by region, 1997 7.12: Annual health service contacts, by age group, 1998 7.13: Share of insurance and user fees in the financing of total public expenditure on health, 1992-97 7.14: The share of health insurance expenditure and state budget subsidies in total hospital revenue, 1994-98 7.15: Health insurance coverage by per capita expenditure quintile, 1998 8.1: Total public spending on health, Vietnam, 1991-98 8.2: Total (public plus private) health spending per capita, selected Asian countries, lat- est year (1991-98) 8.3: Productivity of public health spending, 1992-98 8.4: Life expectancy and real public spending on health per capita across low- and middle-income countries, 1995 8.5: Total public spending on health per capita in the provinces and provincial GDP per capita, 1998 8.6: Public and private health spending per capita, by province, 1998 8.7: Percentage annual increase in real public spending on health per capita across provinces, 1991-98, as related to the level of real public spending on health per capita in 1991 Xiv 8.8: ODA health sector disbursements per capita (US$) over the period 1991-98, by province 9.1: Distribution of population by age, 1994-2024 xv xvi I. INTRODUCTION A. Background and Motivation Despite being one of the poorest countries in Asia, Vietnam's overall state of health, as measured by conventional indicators, is much better than would be expected for a country at its level of income per capita. This is probably a consequence of its socialist character, which prompted it to invest considerable resources in establishing a vast net- work of primary health facilities throughout the country and in developing a number of effective categorical health programs to deal with priority health problems, such as ma- laria, diarrheal diseases, and immunizable diseases. Other contributing factors to the gen- erally impressive health indicators were the high rate of adult literacy achieved by the country early on and a relatively egalitarian distribution of income. In 1986, the government initiated a wide-ranging economic reform program, known as dci moi. The program put Vietnam firmly on the path of transforming itself from a planned economy to a market economy. There was a return to household-based farming in agriculture, removal of restrictions on private sector activities in commerce and industry, and decentralization of decision-making to managers of state-owned enter- prises. Initial progress on the reforms was slow, with many of the comprehensive macro- economic reforms being adopted forcefully only in 1989. These included a devaluation of the official exchange rate to the parallel market rate, decontrol of prices, and an increase in real interest rates to positive levels. Many of the doi moi reforms touched the health sector. Among the most impor- tant health sector reforms were the introduction of user fees for health services at higher- level public health facilities (viz., hospitals), legalization of private medical practice, lib- eralization of the pharmaceutical industry, and deregulation of the retail trade in drugs and medicines. All of these reforms had profound effects on the health sector and on household health-seeking behavior. In the immediate years of the reform program, there was considerable dislocation in the economy and in the health sector. With the advent of doi moi, the agriculture work brigades disappeared, and with them the production brigade nurse, whose responsibility was to support the work of the medical staff in the commune health center through out- reach. The commune health centers, which used to depend for their financing from the agricultural work brigades, had to rely instead on the largess of the local People's Com- mittees, which was insufficient and erratic. As a result, funds for infrastructure, equip- ment, drugs, and training were drastically reduced and the quality of health services at the local level deteriorated significantly. Even the salaries of commune health workers were paid irregularly, if at all. Thus, the country's extensive network of grassroots health facilities and community-financed health workers, which formed the backbone of the na- tional health system since the 1960s, faced imminent collapse. Meanwhile, there was a rapid growth in the number of private providers, espe- cially drug outlets and private pharmacies. The government was not well prepared to regulate and monitor the quality of care and drugs supplied by the private sector. This combination of an unregulated private sector and a seriously underfunded public sector led to unhealthy outcomes. It was against this backdrop that the first analytical review on the health, popula- tion and nutrition sectors in Vietnam was undertaken by the World Bank in 1992.' Be- cause of its timing, that review was necessarily focused on the difficulties of the transi- tion process from a state-financed and state-controlled health sector to one where private financing and private provision of health services were becoming increasingly important. Since then, the Government has initiated many actions to strengthen and further develop the public health services. In 1994, for instance, the government took over the responsibility of paying the salaries of commune health center workers out of the state health budget. In addition, most commune health centers in the country began purchasing drugs from the district and provincial pharmacies and selling them at a small mark-up to customers, which supplemented their meager budgets. Another major development was the establishment of a social health insurance program in 1993, which expanded very rapidly and now accounts for about 9 percent of total public spending on health and covers 12 percent of the population. The health insur- ance and hospital user fee programs have had important effects in generating additional resources for the health sector, especially at the secondary and tertiary level of curative care, and have allowed these facilities to provide services beyond what state subsidies can support. Even more importantly, the macroeconomic situation in the country stabilized in the early 1990s, and the Vietnamese economy managed to grow at impressive annual real rates of 8-9 percent between 1992 and 1997. Additionally, there was a substantial expan- sion of donor assistance during the 1990s. These are just some examples of important changes that have occurred in the health sector during the last 10 years. The need to take stock of the situation in the health sector in light of these changes has motivated this review of the Vietnam health sector, which is, to a large extent, based on the Vietnam Living Standards Surveys (VLSS) of 1993 and 1998 and on other studies conducted by the Ministry of Health and other or- ganizations.2 B. Importance of Health to the Economy It is by now widely accepted that besides being important in and of itself, health improvements are an integral part of a country's human capital. There is a large literature Viet Nam: Population, Health and Nutrition Sector Review, Report No. 10289-VN, Country Department I, East Asia Regional Office, The World Bank, October 1992. 2 The VLSS was first conducted over the period October 1992 to October 1993 with a sample size of 23,839 individuals drawn from 4,800 households. A second round of this survey was undertaken from De- cember 1997 to December 1998 with some changes in questionnaire design and a slightly larger sample size of 28,518 individuals from 6,000 households. About 4,300 of the 1997-98 sample households and 17,780 of the 1997-98 sample individuals are the same households and individuals interviewed in 1992-93, thus constituting a substantial longitudinal data set. The survey was conducted by the State Planning Com- mittee (now Ministry of Planning and Investment) and the General Statistical Office (GSO) in 1992-93 and by the General Statistical Office in 1997-98. The sample was selected using three-staged random stratified cluster sampling. With sampling weights, the data yield unbiased population estimates at the national level as well as at the regional and urban/rural levels. 2 on the large, positive impacts of health improvements on the cognitive achievements of children and on the labor productivity of adults. While quantifying and measuring these impacts for Vietnam is not straightforward and is beyond the scope of this report, some quick calculations suggest that the loss to the economy in terms of wages lost due to ill health is very large-nearly of the order of 8-9 percent of GDP. These are very large costs, considering that they do not even include the opportunity cost of premature mortal- ity, nor the direct cost of treating poor health. C. Scope of the Report Recently, the Ministry of Health (MoH) has formulated its strategic directions for the period 1998-2000. The three key objectives for the health sector that have been iden- tified in this strategy are: * improved health status, as reflected in morbidity and mortality reduction, * greater access to public health services, especially for the poor, and * increased quality and cost-effectiveness of health services. As the Ministry of Health fine-tunes its strategy for achieving these objectives, a number of policy questions will need to be addressed. Some of these are: * In furthering these three objectives, which priorities and approaches are likely to be most propoor? For instance, which financing options are most likely to generate the additional revenue needed for improving the quality of care, while at the same time increasing the access of the poor to the public health care system? * What are the emerging roles of government, households and donors/NGOs in the fi- nancing of health services? What is the optimal balance between the public and pri- vate sectors in the provision of health services? * What is the level of resources that is currently available to the health sector? How does the use of these resources match the agenda of the Ministry of Health? How ef- fectively are these resources being used to achieve the objectives of access and equity in the sector? * How can public spending be oriented to meet the strategic objectives of the Ministry of Health? How can public spending be leveraged so as to maximize its effect on ac- cess, equity and efficiency in the health sector? These questions can be addressed by analyzing empirical trends in and patterns of health services utilization, health outcomes, public health expenditures, and provision of health services. This report attempts to conduct such analysis and discuss the empirical findings of this analysis. Thus, the broad objectives of this report are to: (i) document changes in health outcomes and health services utilization, public and private health spending, and provi- sion of public and private health services during the last seven years, and (ii) identify the causes and factors influencing these changes. Policy responses to the analysis are not provided in this report. Rather, the information is provided in the expectation that inter- ested parties will find it useful for future policy deliberations. 3 One limitation of this report, which needs to be mentioned at the outset, is that it does not dwell on the biomedical causes and consequences of diseases. Instead, it is con- cerned with the socioeconomic and public policy determinants of health-seeking behavior and health outcomes. D. Structure of the Report The remainder of the report is organized as follows. Chapter II reviews the levels of and trends in health, nutrition and fertility indicators in Vietnam, and compares them to those of other countries in the Asia-Pacific region. The chapter also includes a discus- sion of smoking as an emerging health problem in Vietnam. Chapter III discusses the utilization of health services, focusing especially on the changes in utilization patterns and provider choice since 1993. It also presents qualitative information on health-seeking behavior, and reviews the cost of health services to individuals over time and in relation to their discretionary incomes. Chapter IV is concerned with the supply of publicly provided health services, fo- cusing on the delivery of health services by commune health centers and hospitals as well as on the role of NGOs in the health sector. Chapter V considers the presence of the pri- vate sector in the provision of health services, including the supply of drugs by private pharmacies. Chapter VI deals with issues related to human resources in the health sector, while Chapter VII analyzes the financing of public spending on health in the country. Chapter VIII considers trends in public expenditure on health, paying special attention to the issues of efficiency and equity in public health spending. 4 H1. STATUS OF AND TRENDS IN HEALTH INDICATORS A. Morbidity and Mortality 1. Infant and Child Mortality There is some disagreement on the current rates of infant and child mortality in Vietnam. The two most recent sources offer very different estimates of infant mortality- 44.2 per 1,000 live births for the 5-year period 1989-93 [estimated by the 1994 Intercen- sal Demographic Survey (ICDS)] and 28.2 per 1,000 live births for the period 1992-96 [estimated by the 1997 Demographic and Health Survey (DHS)]. It is extremely unlikely that infant mortality fell by 36 percent in the three years between 1989-93 and 1992-96, so the differences between the two estimates must be attributed largely to measurement error.3 The estimate from the General Population Census of 1999 lies between these two figures-viz., 36.7 per 1,000 live births (GSO 1999). Child mortality (1-5 years) and under-five mortality (0-5 years) were estimated to be 17.5 and 61.6 per 1,000, respectively, by the ICDS 1994. Estimates from the DHS are, as expected, nearly 50 percent lower-viz., 9.8 and 37.7 per 1,000. Trends. There are few re- Figure 2.1: Infant.mortality rate, 1960-99 liable estimates of the infant 1 mortality rate (IMR) before 140 1989. However, the various data that do exist indicate a steep de- 120 cline in the IMR during the last - 100 four decades-from a level of L about 160 per 1,000 live births in o 1960 to 90 in 1975-80, 75 in 3 1983, 45 in 1989, and to 44 in 60 1989-93. This suggests that the 40 IMR fell very rapidly in the first three decades since 1960, but has 20 1960 1975-90 1963 1989 1989-93 1999 leveled off during the 1990s (Figure 2.1). The fact that the IMR has leveled off since the late 1980s does not necessarily in- dicate that health gains have stalled. Indeed, there is some evidence to suggest that de- clines in child mortality have continued to be strong even as the decline in the IMR has leveled off.4 The experience of many other countries also suggests that the Vietnamese 3 The DHS figure is likely to be estimated more imprecisely because of its smaller sample size relative to the ICDS sample (7,001 households versus 13,202 households in the ICDS sample). Indeed, the DHS re- ports the 95 percent confidence interval for the infant mortality rate (IMR) estimate to be 21-36 infant deaths per 1,000 live births. Interestingly, the IMR estimated by the (first) Vietnam Demographic and Health Survey of 1989 was also considerably lower than that reported by the 1989 Census (33.5 versus 45 per 1,000 live births). 4 For instance, between 1979-83 and 1989-93, infant mortality declined by 19 percent but mortality among children 1-5 years of age declined by 62.5 percent (GSO 1995). 5 experience is not unique; infant mortality decline is usually rapid in the initial stages of decline and slows down as the IMR falls below 50. It is important to note that Ministry of Health (MoH) estimates of the IMR were significantly underestimated in the 1970s and early 1980s. For instance, the Ministry of Health reported an IMR of 34.2 per 1,000 for 1976, 34.7 for 1980 and 33.5 per 1,000 in 1983. If correct, these rates would have placed Vietnam at the very top of the list of all developing countries in terms of health outcomes during that period-ahead of even China, Sri Lanka and Thailand.5 While this is reason enough to doubt the reliability of these figures, there is other evidence that makes these numbers questionable. The IMR estimates of several international agencies working in the country, as well as of the State Planning Commission of Vietnam, were two to three times as large as the Ministry of Health estimates (Table 2.1). A difference in the IMR between 33.5 and 90.3, which is the extent of variation seen in Table 2.1, implies a variation of 19-20 years in average life expectancy at birth. Table 2.: Infant mortality, 1983 Infant mortality rae Source per 1,000 live btrrhs) MAinis) of Health 33.5 State Planning Commission 68.8 UNICEF '50 LNFPA. 1975-80 90 3 Source Vogel 1987 The ICDS estimated infant and child mortality rates for three five-year periods preceding the survey. These data, shown in Table 2.2, indicate that while infant mortality declined by 19 percent in the 15 years between 1979-83 and 1989-93, mortality among children 1-5 years of age declined by more than three times as much (62.5 percent).6 The much larger decline in child relative to infant mortality probably reflects the success of the government's immunization program, which was expanded considerably during this period. 5 The IMR in 1983 was 39 per 1,000 live births in Sri Lanka and China, 48 in Thailand, and 101 in Myan- mar (UNICEF 1985, as quoted in Vogel 1987). 6 The 1997 DHS shows a similar pattern-a decline of 11.3 percent in infant mortality but 59 percent in child mortality between 1982-86 and 1992-96. 6 Table 2.2: Infant, child and under-five mortali rates, 1979-43 to 1992-6 Period Infant mortality Child mortality Under-five mortality 1079-N3 i ICiDS 1______ 4 S' 82 1 I~~i K DS 4m Tu 14') 68.7 93 OCDS)1 44 2 -- 10 554 % decline, 1979-83 to 1989- 93 1 193 62.5 32.5 Source: GSO 1995. Regional variations. A national infant mortality rate of 45.1 for the 10-year period 1984-93 masks considerable variation in infant mortality across regions. Infant mortality in the rural areas of the country is nearly 80 percent greater than that in the urban areas. In addition, there are large disparities across ecological zones, with the IMR in the Cen- tral Highlands being more than two times as high as that in the Southeast and the Red River Delta (Table 2.3). The disparity in child mortality rates is even greater, with the child mortality rate in the Central Highlands being more than five times as high as that in the Central Coast and the Red River Delta. These differences are indicative of the large regional disparities in access to health services and in household living standards that ex- ist across regions. Table 2.3: Infant and child mortality, by residence and region, 1984-93 Reside'? Rte7vm.. 11-nnt morialr [Child nortality Lnder-five morrahrv uirban 27 0! 9 1 359 Rural 48.2 IS 8 66.1 Victnam 45.1 17 5 61.6 Northern Lplands 62 2 21 2 82 1 Red Ri%er Deta 3S 9 5 43 7 North Central 3- 13.7 50.8 Central (oa.4t 7 2 8.6 460 Centril Highl,inds 1 b 44 9 108.1 Soutle.j:t 31.2 IT 0 47.7 %ikun! Riawr Delia %48 2 2 I 68.9 S 'frr" GSO(.9q4 In addition, what is even more troubling is that regional disparities in infant mor- tality appear to have widened in the recent past. For instance, infant mortality rates in three regions-Northern Uplands, Central Highlands and Mekong River Delta-actually increased between the 1989 Census and the 1994 ICDS-even though the average infant mortality rate for the country did not change (Table 2.4). The ratio of infant mortality in the region having the highest TMR to that in the region having the lowest IMR increased from 1.7 in 1989 to 2.3 in 1994. 7 Table 2.4: Infant mortality deaine by region, 1989-93 Infant mortalin* rates 1989 Popula- 1994 tion Census lCDS % change Northern Uplands 45.4 62.2 37.0 Red River Delta 37.0 35 3 -4.6 North Central 46.5 37 7 -189 Central Coast 47.5 372 -21.7 Central Highlands 56.4 71 6 27.0 Southeast 33.9 31.2 -8.0 Mekong River Delta 44.1 48.2 9.3 Vietnam 45.0 45.1 0.2 Source- GSO 1990. 1995 Economic Differentials. The DHS 1997 data permit comparison of infant and un- der-five mortality rates across quintiles of households ranked by wealth. Although, as noted earlier, the overall estimates of infant and under-five mortality are biased down- wards in the DHS, the quintile-specific mortality estimates can be used to obtain an idea of relative mortality variations across economic groups. These data, shown in Table 2.5, indicate significant disparities in infant and under-five mortality across wealth quintiles, with the top 40 percent of households having an infant mortality rate that is about one- half of that experienced by the poorest 60 percent of households. Interestingly, there ap- pear to be few differences in infant mortality among the lowest three quintiles and among the two highest quintiles. The differences in infant and child mortality across wealth groups are most likely the result of socioeconomic differences in child nutrition, utiliza- tion of health services, and environmental hygiene (such as availability of safe drinking water and sanitation). Table 2.5: Infant and under-fie mortality, by wealth quintiles, 1997 Wealth Quintiles Infant mortalit Under-f ve mortality Bottom quintile 34.9 50.9 Second 20% 33.8 42.8 Third 20% 34.6 42.0 Fourth 20% 17.0 23.9 Top gintile 17.4 18.9 Total 34.8 45.9 Source NCPFP (forthcoming) 7 It should be noted, however, that income and wealth inequality is still moderate in Vietnam by interna- tional standards. For example, in 1998, the Gini coefficient of consumption expenditure (often used as a proxy for permanent income) was 0.35 in Vietnam--considerably smaller than that for Thailand (0.41). Thus, when the report refers to the highest or top quintile in Vietnam, it is referring to the relatively better- off in Vietnamese society-individuals who probably are not very wealthy or affluent by international-or even Southeast Asian-standards. 8 Gender Disparities. Excess mortality of female children relative to male children, commonly observed in some parts of Asia, does not appear to be a problem in Vietnam. Indeed, both the ICDS and DHS estimates indicate significantly lower infant and under- five mortality among females relative to males (Table 2.6). While this is commonly ob- served in many countries, the size of the gender disparity favoring females appears un- usually large, especially in the DHS estimates, suggesting that deaths of female infants and girls were probably under-reported by respondent households. Table 2.6: Infant and under-five mortality by sex. 1984-93 and 1992-97 Sex Infant mortalii Under-five mortalin Males. 1984-93 50 3 66.7 Females. 198-1-93 39.1 56.2 Males. 1992-96 35.6 42 1 Females. 1992-96 19.2 29.2 Source. GSO 1995. NCPFP (fonhcoming) International Compari- Figure 2.2: Infant mortaity rates and real GDP per capita sons. There is no question that across Asian countries, 1997 Vietnam's achievements in re- no ducing infant mortality are im- - -- pressive and have few parallels in the world. Vietnam's infant mortality rate is already among 6o the lowest in the Asia-Pacific -o region and comparable to coun- 40 - tries that have substantially 30 higher levels of per capita in- 20 - - --- - come. Indeed, as Figure 2.2 to - ------- --- -_-- - indicates, if Vietnam were to 0 follow the experience of other 1,000 3,000 5,000 7,000 9,000 11,000 13,000 Asian countries, it would be Rcal GDP Wr capita (PPPS), 1997 expected to have an IMR that is roughly two times as large as it currently has, given its per capita income level. 9 Not only is the level of Figure 23: Decline in infant mortality rates between 1970 and 1997 the infant mortality rate in Viet- GDPper capita, selected Asian countries nam unusually low in relation to its per capita income, the rate at which its IMR has declined in 7 the last 27 years is also among the highest in the region. Figure 2.3 shows Vietnam to be a sig- nificant positive outlier in the relationship between the rate of IMR decline and per capita GDP. In this sense, compared to other countries in the region 3, (and indeed in the world), Viet- Rea0 G,00 per 7apit 9,00S) L997 nam an b cosideed a anReal GDP per capita (PPP$), 1997 nam can be considered as an overachiever in bringing down its infant mortality rate. Factors influencing Infant Mortality. DHS data show four important risk factors for infant and child mortality (Table 2.7).8 First, infant and child mortality are strongly related to maternal education. For instance, women with lower secondary schooling have an infant mortality rate that is 46 percent lower and a child mortality rate that is 75 per- cent lower than that for women with no schooling. Second, maternal age at the time of a child's birth is an important risk factor for infant-but not child-mortality. Infant mor- tality rates are highest for mothers giving birth while under the age of 20 years, while child mortality rates increase with the mother's age at birth. Third, birth order plays an important role in influencing mortality, with first- and higher-order births experiencing higher mortality than births of orders 2-6. For example, the IMR is more than two times as high among births of order 7 and higher than among births of orders 2 and 3. The probability of a seventh-order child dying between the ages of 1 and 5 years is nearly three times as large as that of a first-order child dying during the same period. Fourth and finally, mortality levels are higher among children born within two years of a previous birth. For example, infant mortality is 58 per 1,000 for this group compared with 21 per 1,000 for children born after an interval of 4 years or more. 8 What are presented here are merely statistical associations that do not necessarily indicate causal relation- ships between mortality and demographic or socioeconomic characteristics. 10 Table 2.7: Infant, child and under-five mortality rates, by education and demographic characteristics, 1987-96 Soctoeconomic or denographic char- Infant Child UnderInec acteristic Morialitr mortalin mortalitv Mother's education None 49.7 27.8 76 1 Some pnmary 43 3 21 2 63 6 Completed pnmar% 34 7 9 1 43 5 Completed lower secondar 26 9 6.9 33.7 Completed higher secondary or more 29.6 5.0 34 4 Age of mother at birth <-20 years 44.0 6.3 50.0 20--29 Nears 33 3 11.2 44 1 30--39 ycars 35.2 15 8 504 Birth order S35.2 7 3 42 2 2-3 30.6 10.8 41.0 4-6 36.1 18.8 54.2 7+ 65.5 20.9 85.0 Previous birth interval <2 years 57.7 23.1 79.5 2-3 years 30.9 11.2 41.7 4+ years 21.0 10.2 31.0 Total 345 11.9 46.0 Source: NCPFP fbrhcoming). 2. Main Causes of Morbidity and Mortality Vietnam has ex- Figure 2. MajorCauses of Morbidity,1976-97 perienced an epidemiol- ogical transition during the - last 20 years, with a major 43% 39/ decline in the share of 7&/' communicable diseases in mortality and morbidity. s% This is seen in Figures 2.4 40/. and 2.5, which show the 30% major causes of morbidity 20% and mortality, respec- tively, in Vietnam from 1976 to 1997. While 1976 1986 1997 communicable diseases ICommunicable diseases -Non-cornmunicable diseases accounted for 50-56 per- EDAccidents, injuuies and poisoning cent of morbidity and mortality cases in 1976, the corresponding shares had fallen to 27 percent by 1997. Much of this decline was achieved during the last ten years. Most dra- matic has been the increased incidence of accidents, injuries and poisoning, whose share in mortality went up from 2 percent in 1976 to 22 percent in 1997. The share of non- communicable diseases also increased considerably over this period. 11 These shifts re- Figure 2.5: Major Causes of Mortality, 1976-97 flect the success of 2 communicable disease .t 80% control programs, espe- - 0, cially the expanded pro- 3 60% gram of immunization, a52% a50 which has dramatically reduced the incidence of 4 vaccine-preventable dis- 3o% eases in the country. Despite the de- o% cine in their incidence, 1976 1986 1997 infectious diseases con- . Commun-cable diseases O Non-comrmunicable diseases tinue to remain major O Accidents, injuries and poisoning public health problems in the country. In 1998, for instance, diarrhea, dengue fever, ma- laria and respiratory tuberculosis were among the ten leading causes of admissions at public hospitals, together accounting for nearly 600,000 cases (Table 2.8). In contrast, noncommunicable diseases now account for the largest share of hospi- tal deaths. In 1998, intracerebral hemorrhage, pneumonia, acute myocardial infarction and heart failure, hypertension, and suicides were among the six leading causes of mor- bidity, together accounting for nearly two-thirds of all hospital deaths. Table 2.8: Leading causes of morbidi and mortality in bospitals, 1998 % of total Cause of hospital cases % of total Cause of hospiral deaths below Diarrhea of infectiou. origin 17.70 Intraccrebral hemorrhage 20.54 Pneumonia 16 83 Pneumonia 19.00 Acute bronchitis 14 I Respiratory bronchitis 12.76 Slow fetal growth, fetal malnutrition and disor- ders related to short gestation and low birth Dengue fever 13.19 weight 9.91 Medical abortion 10.95 Acute myocardial infarction 7.76 Essential hypertension 6.60 Essential hypertension 6 98 Malaria 6 22 Deniue fever 6.83 Gastric and duodenal ulcer 4 79 Heart failure 5.44 Respiratory tuberculosis 4 6S Viral encephalitis 5.41 Appendicitis 4 24 Suicide 5 37 Total 100.00 Tctal 100 00 Source MoH 1998c 3. Major Communicable and Noncommunicable Health Problems The disease pattern in Vietnam is strongly related to geography, climate and so- cioeconomic characteristics. For instance, malaria is most serious in the Central High- lands and mountainous districts of the Central Coast provinces; the prevalence of dengue fever is highest in the Mekong River Delta; bubonic plague often appears in the Central provinces and Central Highlands; and Japanese encephalitis mainly occurs in the North 12 during the wet season. Malaria is often serious in the remote disadvantaged areas, while dengue fever typically breaks out in crowded urban and delta areas. Despite the decline in their incidence, communicable and infectious diseases con- tinue to be major public health problems in the country. The incidence of and trends in some of the most prominent diseases are discussed below. Diarrhea. Diarrhea continues to be the leading cause of morbidity in Vietnam, with over 250,000 hospital cases in 1998, despite a large decline in its incidence during the last 10-20 years. Because of much better case management, however, the case- fatality rate in diarrhea has dropped significantly. In 1976, diarrhea was the leading cause of mortality in public hospitals. However, since 1995, it is not even among the top 10 leading causes of mortality (Ministry of Health 1999a). An improvement in safe water supply, improved general nutritional status of children, greater use of oral dehydration therapy, and better health education among parents on how to deal with diarrhea have all contributed to a decline in diarrheal deaths. Acute Respiratory Infections. Acute respiratory infections (ARIs) are among the leading causes of morbidity in Vietnam. Indeed, there were a total of nearly 450,000 hos- pital cases of pneumonia and acute bronchitis together in 1998, which would put ARIs ahead of malaria as the leading cause of morbidity in Vietnam. Unlike diarrhea, ARIs-- particularly, pneumonia and respiratory bronchitis-are a major killer in the country, ac- counting for 2,195 hospital deaths in 1998 (Ministry of Health 1998c). The increase in ARIs is the direct result of an increase in air pollution and unhealthy dwelling conditions in poor urban areas. In addition, the increase in drug resistance-in turn the result of irra- tional antibiotic use-has compromised successful treatment of ARIs and contributed to the rise in ARI-related mortality. Malaria. Morbidity and mortality from malaria have fallen dramatically in recent years, thanks to one of the most successful malaria control programs in the world. Ma- laria control measures in the past 6-7 years have been particularly successful, with the malaria incidence rate falling from 16 cases per 1,000 persons in 1991 to merely 6 in 1998 (Figure 2.6). At the same time, the fatality rate for malaria has dropped to nearly one-tenth of what it was in 1991, with fewer than three deaths out of 10,000 malaria cases. Many localities in the country have reported no malaria outbreaks for 3-4 consecu- tive years. Vietnam is recognized as one of the world leaders in the use of impregnated bed-nets, which currently protect over 10 million persons. 13 Despite these impressive Figure 2.6: Morbidity and mortalitx rates from malaria, 191&97 achievements, malaria remains a major public health concern in 4 the mountainous and ethnic mi- 20 nority areas. Because of migra- 030 tion of ethnic people, malaria epidemics can still break out in other parts of the country. For 20 example, in recent years, epi- - demic outbreaks have occurred in Daklak and Tuyen Quang ow- a .Z P ing to migration of the H'mong -,0 ethnic people from the North. . Also, in areas where malaria has decreased, the immunity to the disease has also declined after 3-5 years, so epidemic outbreaks and resurgence of ma- laria are possible if preventive measures are not sustained. Denpue Fever. As noted Figure 2.7: Morbidity and mortali"y rates from dengue fever, 1991-98 earlier, dengue fever is the fourth 35 most important cause of morbid- ity (as recorded in public hospi- . . 4 tals). There has been an alarming 2s increase in the incidence of den- gue fever in the last 4-5 years (but particularly since 1997), 15 with the infection rate rising al- most fivefold between 1994 and C 1998 (Figure 2.7). Of course, be- cause of better case management, . the fatality rate has been falling 9 99 193 1q I 9S since the early 1990s, but dengue fever continues to be among the top 10 causes of mortality in the country. Vaccine-Preventable Diseases. One of Vietnam's greatest achievements in public health during the past two decades has been the dramatic decline in child mortality from vaccine-preventable diseases. Much of this success has been due to an impressive Ex- panded Program of Immunization (EPI), which was implemented on a pilot scale since 1982 and expanded to the whole country in 1985. The prevalence rate of polio fell from 2.6 per 100,000 persons in 1986 to 0.6 in 1996 (Table 2.9). In fact, the country expects to completely eradicate polio by 2000. The prevalence of diphtheria by 1996 had dropped to 5 percent of its level in 1986. Neonatal tetanus has been virtually eliminated in 591 out of 610 districts in the country (Ministry of Health 1999b). Deaths from measles have also declined dramatically during the past 15 years. 14 Table 2.9: Morbidity and mortality rate per 100,000 inhabitants of vaccine- preventable deaths lipe Cases per 100 000 pop Deaihs per 100.000 pop 1985 1990 1996 1 o.k 1990 /996 Diphthena 3 7 0.77 0 19 0.40 009 002 Perrussis 68.6 6.18 18.03 0.08 0.04 0 01 Polio 26 0 59 0.61 0 15 002 0J2 Tetanus 50 0.48 053 040 031 0 27 Masles 12 8 13 15 6.86 045 0.04 001 Source MoH 1996b Tuberculosis. Tuberculosis (TB) is another disease that has seen a sharp rise in re- cent years. For instance, the number of pulmonary smear-positive TB cases increased ninefold-from 6,645 to 54,897-between 1986 and 1998 (Ministry of Health 1999a). Of course, much of the increase in reported TB cases is probably the result of a large expan- sion in the TB control program, which covered only 23 percent of the population in 1985 but had 99 percent coverage by 1998. This means that many TB cases that went unde- tected and untreated in the past are being correctly diagnosed and treated today by a tu- berculosis control program that is recognized as one of the most successful in the world. However, the annual risk of TB infection rates has also increased during recent years, in- dicating a rise in the real incidence of TB in Vietnam. There are a number of factors that are contributing to the current rise in TB incidence, including the aging of the population, rising levels of air pollution, overcrowding in the urban areas, and high levels of cigarette smoking. Additionally, the high rate of drug resistance is making tuberculosis control more difficult. The Ministry of Health estimates that TB incidence will possibly rise to as many as 120,000 cases and 3,500 deaths per year by 2010 (Ministry of Health 1999a). HIV/AIDS. Since the first HIV infection was detected in Ho Chi Minh City in 1990, the disease has spread rapidly in the country. The number of new HIV cases has increased more than fourfold between 1993 and 1999, with nearly 5,000 new cases being reported in 1999 (Figure 2.8). Figure 2.8: Number of HIV/AIDS cases detected and number of By 1998, 61 provinces had an m1V/AIDS - affected provinces, 1990-99 HIV/AIDS case. Even then, it is estimated that there are s,00 only a total of 16,175 cumula- 4,500 INo. of cases detected tive cases of HIV infection 4,000 -*- No. of affected provinces 50 and 2,907 cases of AIDS 3soo (1,512 of which have proved 40 fatal already) (Ministry of 2,500 Health 1999a). The incidence 2o of HIV infection is estimated 1,50020 at about 0.1 percent- 1,500 considerably lower than the 2 1,000 10 percent rate in Thailand. In 500 this sense, Vietnam does not 0 0 yet have the kind of 1990 1992 1993 1994 1995 1996 1997 1998 1999 15 HIV/AIDS epidemic that is observed in neighboring Cambodia and Thailand. However, the potential remains for HIV/AIDS to escalate further in Vietnam if active measures to control it are not sustained. Even though 65 percent of the HIV infections reported so far are associated with intravenous drug use, heterosexual transmission among young people is increasingly be- coming more common. While the disease predominantly affects males (only 17.2 percent of HIV-infected persons were female in 1998), the incidence among women has been in- creasing rapidly (women constituted only 7 percent of HIV-infected persons in 1993) (Ministry of Health 1999a). Accidents, Injuries and Poisoning are another major cause of death in hospitals. In 1997, there were 475,000 cases of and 5,424 deaths from accidents, injuries and poison- ing. As noted earlier, this represents a dramatic increase from the situation twenty years ago. Many of these deaths are caused by traffic accidents. This is a trend that is character- istic of many developing countries experiencing economic expansion, rapid urbanization, and increased traffic. Heart Disease. With increasing affluence, the incidence of 'diseases of affluence', such as cancer and heart disease, has increased considerably over time. Indeed, acute myocardial infarction, essential hypertension, and heart failure together accounted for 20.2 percent of hospital deaths in 1998 (Table 2.8), making heart-related ailments the second leading cause of hospital-based morbidity in the country (just behind intracerebral hemorrhage). These diseases are much more prevalent in urban centers, such as Hanoi and Ho Chi Minh City, than in rural areas, and their incidence, at least as of now, is likely to be greater among better-off than among poor individuals. Yet, if the experience of other countries is any guide, the incidence of these noncommunicable diseases is likely to increase rapidly in the future with economic growth and increasing affluence. 4. Smoking as a Behavioral Health Risk Factor Smoking has been well established as a contributing risk factor in several dis- eases, such as respiratory infections, lung cancer, heart disease, and tuberculosis. Viet- nam has among the highest prevalence of smoking in the world (WHO 1997). However, annual consumption of cigarettes per adult (age 15 and over), which is estimated at about 909 from production figures and estimates of smuggled imports of cigarettes, is relatively low in comparison to estimates for other countries in Asia (Figure 2.9). Further, there are signs that smoking prevalence is declining. In addition, women have not yet begun smok- ing in large numbers, so there is still hope that an epidemic of tobacco addiction in Viet- nam can be prevented. 16 Patterns of Smoking In- 500 Figure 2.9: Average annual number of cigarettes smoked per adult (>15 years), cidence. Data from the two Asia,1990-98 3,240 rounds of the VLSS indicate that 3,000 3.10o tobacco use in Vietnam remains primarily a habit among men, with slightly over half of men aged 15 and over being regular 610 63 tobacco users, compared to only 2 500- 180 about 4 percent for women in 1 990 the same ages. This compares favorably to data from the Min- i [[[II istry of Health which indicate that 50 percent of men and 3.4 0 percent of women smoke (Le Ngoc Trong and others 1998). N Tobacco use in Vietnam consists of smoking cigarettes, pipe tobacco (including water pipes), and chewing tobacco, sometimes in conjunction with chewing areca nut and betel leaves, the latter being more common among women. Indeed, a much larger propor- tion of female tobacco users (relative to male tobacco users) are likely to use pipe or chewing tobacco. Among male tobacco users, cigarette smoking is relatively more com- mon than using pipe or chewing tobacco. Use of tobacco products is more prevalent in rural than in urban areas (Table 2.10). However, urban residents tend to smoke cigarettes more than rural residents, while rural residents are much more likely to use pipe or chewing tobacco. Rural and urban women tend to smoke cigarettes at the same rates, but rural women are three times as likely to use chewing tobacco or pipe tobacco as urban women. Table 2.10: Tobacco use by urbanL/rural residence and educational level. 1998 (percent) Alen Women Use tobacco Pipe chew- Use robacco Pipe.-che'%- Category products Cigarettes ing tobacco products Cigarettes irg tobacco Overall 50.76 34.55 20.01 3.64 1 12 2.59 Urban 46.28 4084 8 14 2 17 1.16 1 03 Rural 52 22 32.51 23.85 3.92 0 92 3.06 Education level Less than pnmary 58 75 3409 27 41 994 1.96 8.12 Primazy 55.12 3853 20 52 3 30 1 05 2.29 Lower seondarv 4569 28.13 21.74 1.19 029 0.97 Upper secondary 37.88 2979 1082 0.45 024 0 22 Technical school 4962 33.80 20.24 2.29 1 50 0 87 University and above 46 08 39.75 6 95 0 68 0.68 000 Source- VLSS 1998 Education levels have a strong inverse relationship with pipe/chewing tobacco use for both men and women (Table 2.10). However, for cigarettes, the highest rates of smok- ing cigarettes are found among individuals with university education and second highest 17 among those with only primary education. The relationship between education and smok- ing is likely confounded by income - e.g., people with high education are wealthier and therefore more able to purchase cigarettes. Table 2.11: Tobacco use by per capita expenditure quintiles, 1998 (percent) Men Women Per capita exp. Use tobacco Pipe/chew- Use tobacco Pipe/chew- Quintile products Cigarettes ing tobacco products Cigarettes ing tobacco Bottom 58A6 30.02 33.22 5.55 1.33 4.26 Second 54.19 32.12 26.45 4.53 0.91 3.73 Third 52.60 34.94 21.53 3.02 0.71 2.38 Fourth 47.70 36.34 15.35 2.72 .095 1.83 Top 43.02 38.06 7.17 2.08 1.06 1,02 Source: VLSS 1998. Table 2.11 indicates that the prevalence of cigarette smoking among men does increase with economic status, while the rate of pipe/chewing tobacco decreases. For women, the relationship for cigarettes is not clear, but as standard of living increases, the prevalence of pipe/chewing tobacco use declines. Another interesting, but expected, find- ing is that the cigarettes consumed by higher quintiles are higher-priced cigarettes. This does seem to indicate that ability to buy cigarettes has a strong influence on smoking be- havior, which suggests that, as incomes rise over time, the number of people smoking cigarettes will also rise unless strong policy interventions make smoking less attractive. Region of residence also has a strong relationship with smoking (Table 2.12). Among men, use of pipe or chewing tobacco is much more prevalent in the North than in the South, while cigarette smoking is much higher in the South than in the North. For women, tobacco use is especially high in the central regions, especially the Central High- lands. Table 2.12: Tobacco use by region, 1998 (percent) Men Women Use tobacco Pipe/chew- Use tobacco Pipe/chew- Region products Cigarettes ing tobacco products Cigaretnes ing tobacco Northern Up- lands 48.42 19.96 34.46 3.09 0.21 3.01 Red River Delta 45.69 24.33 26.43 1.78 0.58 120 North Central 49.93 25.53 27.98 7.68 1.77 6.03 Central Coast 46.83 44.32 3.43 4.44 1.69 2.76 Central High- lands 56.98 46.09 15.86 9.46 2.37 7.61 Southeast 49.61 45.62 5.64 2.75 1.50 1.25 Mekong Delta 59.54 46.82 16.35 2.08 0.63 1.45 Source: VLS9 1998. A relatively small percent of Vietnamese men aged 15-20 years use tobacco products (14 percent). However, among men, there is a'large increase in the incidence of 18 smoking between ages 15-20 and 20-25 years. This coincides with the age group that would have recently been demobilized from mandatory military service. It is widely known that smoking prevalence in the military is extremely high, and the army is some- times referred to by antismoking advocates as the "main training center for smokers." Younger women have very low incidence of smoking or chewing tobacco use in Vietnam, and rates for women increase gradually till about age 50, where a peak in ciga- rette consumption and the beginning of a large increase in use of tobacco mixed with ar- eca nut and betel leaves is observed. Changes in Smoking Incidence, 1993-98. Over the five-year period between 1993 and 1998, the prevalence of tobacco use declined from 63 percent to 51 percent among men and from 5 to 4 percent among women (Table 2.13). The percent of men smoking cigarettes declined from around 44 percent to 35 percent over this period, while the per- centage of women smoking cigarettes more than halved, falling from 3 to 1 percent. However, there was a slight decline in the overall proportion of adult men using pipe or chewing tobacco. Women also increased their use of these alternative tobacco types slightly. The number of cigarettes consumed seems to have declined only slightly over the period between the surveys. All regions saw a decline in cigarette smoking among men, but the strongest de- clines were in the Mekong Delta and the South Central Coast and the weakest in the Red River Delta. For women, the largest declines were in the Central regions, and there were increases in the incidence of smoking in the Northern Uplands and the Red River Delta. The relative decline in the incidence of smoking was more or less similar among men of different consumption quintiles (4-5 percent), while among women the percent- age declines were largest for those in the bottom quintile. The declines in smoking prevalence, the relatively low number of cigarettes con- sumed per smoker, and continued low numbers of women smoking are positive signs that a tobacco epidemic requiring expensive medical interventions in 10 to 20 years may be avoided in Vietnam. However, rising rural incomes combined with high rates of tobacco use among farmers suggest that stronger measures will need to be taken in the future to ward off future increases in smoking, and to reduce the current high rates. Importantly, the new tobacco control strategy of the Committee for Tobacco Control in Vietnam pre- sents a much more coordinated and comprehensive government policy than in the past. 19 Table 2.13: Changes in tobacco use between 1993 and 1998 Use tobacco prod- Sex, region, quintile ucts Cigarettes Cigarettes per day 1993 1998 1993 1998 1993 1998 Total 32,24 25.72 22.12 16.80 11.37 11.06 Men 63.11 50.74 4393 34.58 11.43 11.1i Region Northern Uplands 59.79 48.42 25.00 19.96 9.88 9.72 Red River Delta 60.87 45.69 25.86 24.33 8.41 8.33 North Central 65.74 49.78 30.70 26.02 10.88 10.90 Central Coast 62.76 46.83 61.73 44.32 13.32 13.50 Central Highlands 68.94 56.98 62.12 46.09 14.18 14:03 Southeast 61.25 49.61 54.66 45.62 11.77 11.61 Mekong Delta 67.09 59.54 67.09 46.82 11.10 11.09 Per capiua expenditure quintile Bottom 67.23 58,46 38.90 30.02 11.95 12.36 Second 67.30 54.19 40.27 3212 12.04 11.49 Third 64.77 52.63 42.67 34.99 11.77 11.44 Fourth 60.76 47.67 46.00 36.35 11.47 10.61 Top 57.06 43.02 50.14 38.08 10.99 10.25 Women 4.73 3.47 2.69 0.99 9.94 9.54 Region Northern Uplands 2.25 3.09 0.11 0.21 10,00 13.58 Red River Delta 3.10 1.78 0.17 0.58 9.73 9.70 North Central 6.13 7.54 4 17 1.78 4.80 4.95 Central Coast 12 73 4.44 844 1 69 9.48 10.32 Central Highlands 17.76 9.46 13.16 2.37 12.81 12.74 Southeast 2.33 2.75 1.89 1.50 12.74 12.81 Mekong Delta 261 208 2.10 0.63 8.69 8.75 Per capita expenditure quinile Bottom 9 78 5 55 6.28 1.33 8.95 9.82 Second 393 4.53 1.34 0.91 14.85 9.73 Third 4.30 3.02 2 47 0.71 9.43 8.63 Fourth 3 89 2.71 2 25 0.95 8.45 10.95 Top 2.60 2.09 1 62 1 08 8.71 8.49 Source- VLSS 1998. 20 B. Reproductive Health and Fertility Outcomes 1. Maternal Mortality Maternal mortality in Vietnam has declined from 400 or more in the 1950s to 200 in the 1980s and to about 160 per 100,000 live births at present (UNICEF 1999). There are large differences in the maternal mortality rate across regions, with the Central High- lands and the Northern Highlands having the highest levels of maternal mortality rate and the Red River Delta and the Southeast having the lowest levels (Table 2.14). However, all regions have experienced very large declines in the maternal mortality rate during the early 1990s. It should be noted, however, that these are all hospital-reported maternal mortality statistics, which are often not accurate as women who die while giving birth at home or before they get to a hospital are typically not reported. Additionally, deaths caused by abortion or miscarriage may not be reported either. Table 2.14: Maternal moftalith rate in selected provinces. 1991-93 Materna! deaths per 100.000 Y dechne in live births MMR, Region Prounce 1990 1991 1993 1Y91-93 Northern Highiands Cac Bang 326 298 150 53 9 Red River Delta HaiPhong 120 107 50 58 3 North Central Coast Nghe An 281 260 120 57 3 Central Coast Quang Tn 230 206 120 47.8 Central Highlands Gia Las 412 418 180 56.3 Southeast Baria-Vungt3u 146 130 90 38.3 Mekong Rj,,cr Delta Soc Trang 232 211 100 56.9 Source MoH and LNFPA 1999 The main causes of hospital-reported maternal mortality are hemorrhage (49 per- cent), followed by tetanus (27 percent) and eclampsia (9 percent) (Figure 2.10). Most ma- ternal morbidity in Vietnam Figure 2.10: Main causes of maternal mortality, 1997 is preventable, and related to the following factors: 0 Haemorrhage ClInfection w Eclampsia i Uterine rupture * Anemia, which is esti- [* New bom tetanus mated to affect 63 per- cent of Vietnamese women; 49% * Vitamin A deficiency, which is itself closely re- lated to anemia, with almost one-half ofo women showing low breast-milk retinol lev- 6% els; 21 * Poor maternal nutrition, as 40 percent of women suffered from chronic energy defi- ciency in 1994; * Reproductive Tract Infections (RTIs), which can cause premature deliveries, thereby putting women at higher risk of postpartum infections; * Induced abortions: Vietnam has one of the highest abortion rates in the world, with an average woman having 22 abortions during her lifetime. It is estimated that 1.3 mil- lion pregnancies are aborted annually in the country; and * Poor obstetric care and absent emergency obstetric care, especially in the remote and mountainous areas. 2. Fertility Levels and Trends. The total fertility rate (TFR) has been estimated at 2.3 by the Population Census of 1999 (GSO 2000). This estimate lies considerably below the esti- mates of 2.67 produced by the VNDHS-II for the period 1989-93 and of 3.25 projected by the ICDS for the period 1992-96. Few estimates of fertility are available for Vietnam before 1987. Using reverse survival methods, the 1989 Census suggested that fertility rates were flat between 1965 and 1974, but declined steeply after that. The reverse survival-estimated total fertility rate (TFR) fell from 5.9 in 1970-74 to 3.9 in 1985-89-a remarkable decline of over 33 per- cent in 15 years. More recent estimates suggest that fertility has continued to decline at the same pace during the last decade-from 3.98 in 1987 to 3.25 in 1989-93 to 2.67 in 1992-96, which represents a decline of 33 percent from 1987 to 1996 (Table 2.15). This experiences puts Vietnam at the forefront of countries that have experienced rapid fertil- ity declines. Table 2.15: Trends'in age-specific and total fertility rates from various sources, 1987-96 VNDHIS- Census ICDS-94 KNDHS-II Age group ears) 1987 1988-89 1989-93 1992-96 15-19 20 35 38 39 20-24 235 197 196 178 25-29 243 209 189 148 30-34 151 155 124 95 35-39 85 100 69 52 40-44 51 49 31 20 45-49 11 14 02 04 TFR 3.98 3.8 3.25 2.67 Source: VNDHS-1 rates are from NCPFP 1990, 1989 census rates are from GSO 1990, ICDS-94 rates are from GSO 1995, and VNDHS-11 rates are from NCPFP (forthcoming). Table 2.15 indicates that fertility rates declined most for women aged 40 years and over, moderately for women aged 20-39 years, and actually increased for women aged 15-20 years. This is a common pattern among populations experiencing fertility transitions, and occurs as older women, who are more likely to have reached their desired 22 family size, limit their births to a greater extent than younger women, who are likely not to have yet achieved their desired family size. Regional Variations. A national total fertility rate of 2.67 for the 5-year period 1992-96 masks considerable variation in fertility across regions. Fertility in the rural ar- eas of the country is more than 60 percent greater than that in the urban areas (2.9 versus 1.8) (Table 2.16). In addition, there are large variations across ecological zones, with the TFR in the Central Highlands being more than two times as high as that in the Southeast. Fertility rates are lowest in the Southeast (1.87), followed by the Red River Delta (2.28) and the Mekong River Delta (2.31). These differences are indicative of the large regional disparities in access to health and family planning services and in living standards that exist across regions. Table 2.16: Total fertility rates, by residence and region. 1992-96 Residence -Region Total fertility rate Urban 1.84 Rural 2.90 Vietnam 2.67 Northern Uplands 3.14 Red River Delta 2 28 North Central 3.26 Central Coast 3.39 Central Highlands 4 28 Southeast 1 87 Mekong River Delta 2 31 Source NCPFP (forthcoming) It is possible to compare the decline in the general fertility rate (GFR) between 1988-89 and 1993-94 across ecological regions.9 The data indicate that fertility decline has been uneven in the country. For instance, fertility decline was most rapid in the Me- kong River Delta (46 percent), followed by the Red River Delta (28 percent) and the Southeast (27 percent) (Table 2.17). On the other hand, the North Central region experi- enced only a 9 percent decline in fertility. The ratio of the general fertility rate in the re- gion having the highest GFR to that in the region having the lowest GFR increased from 1.76 in 1988-89 to 2.02 in 1993-94. 9 The general fertility rate is calculated by dividing the number of births occurring during a specific period by the number of women of reproductive age (15-49 years) and multiplying the result by 1,000. 23 Table 2.17: Decline in general fertility rate by regiog, 198849 to 1993-94 General fertili(y rare 1988-9 (1989 1993-94 (ICDS Region Census) 1994) % change Northern Uplands 141.0 123.2 -126 Red River Delta 104.3 74.9 -28.2 North Central 137.2 1249 -8.9 Central Coast 136.0 106.0 -22 1 Central Highlands 183.1 151 2 -17.4 Southeast 107.9 79.1 -26.7 Mekong River Delta 141.6 76.3 -46.1 Vietnam 1207 100.5 -16.7 Source. GSO 1995 Socioeconomic Differentials. The DHS 1997 indicates that fertility differentials are strongly related to differences in the educational attainment of women. Women who have completed primary schooling have a total fertility rate that is 31 percent lower than that of women with no schooling (Table 2.18). The TFR for women who have completed higher secondary schooling is 53 percent lower than that for women with no schooling. Since educational attainment is positively related to household income, it is likely that fertility and income are also inversely related to each other. Table 2.18: Total fertility rates, by women's schooling level, 1997 Women s choolng level Ferility rate None 4.03 Some primary 3.13 Completed primary 2.79 Completed lower secondary 2.53 Completed higher secondary or more 1.91 Total 2.67 Source: NCPFP (foibtoming). International Comparisons. Vietnam has experienced a major fertility transition during the past two decades. While Vietnam's fertility rate is not the lowest in the Asia- Pacific region, it has a rate much lower than a country with its per capita income. Indeed, as Figure 2.11 indicates, based on its per capita income and on the relationship between per capita income and total fertility rates observed across selected countries in Asia, Viet- nam could be expected to have a total fertility rate of 3.6-not 2.7 that it currently enjoys. This again confirms the remarkable success of Vietnam in bringing down its fertility rates. 24 Not only is the level of fertility in Vietnam unusually low in relation to its per capita income, it has experienced a faster rate of decline in fertility during the last two decades than most other countries in the region. Figure 2.12 indicates that, based on a positive relationship between the Figure 2.11: Delne in toal fertility ate beMeen 1975 And 1997 rate of fertility decline and per In te---n t r capita GDP observed across a set of Asian countries, Vietnam .1 could have been experienced a 41 fertility decline of about 30 per- 3 5- cent over the period 1975-97- 25 not the decline of 55 percent that 1 it actually accomplished. 3. Other Reproductive Health Indicators 1,000 ,000 5,000 7.oo00 9,000 I000 13,000 Birth Intervals. There is Real GDP pertpna (PPPS), 1997 compelling evidence from around the world that short birth intervals, typically of less than 24 months, are harmful to the health of infants and young children. Birth intervals are generally long in Vietnam; almost half of all second or higher-order births occur three or more years after the previous birth. Only 19 percent of births occur after an interval of less than 24 months. There is also some evidence that the median birth interval has in- creased from 32 months in the period 1989-93 to 36 months in the period 1992-96. There are, of course, re- Figure 2.12: Total fertility rates and real GDP per capita gional variations in birth across Asian countries, l97 intervals, with the Central Highlands and the Northern Uplands having a larger s.0 percentage of births occur- _ _ 4.5 ring after an interval of _ fewer than 24 months than other regions. 'A Incidence of Low Birth Weight. Low birth 254 weight-the percentage of 2.0 -- ----- infants born with a birth 1.5---------- weight of less than 2,500 1.000 3,000 5,000 7,000 9.000 11,000 13.000 grams-is a commonly R GDP p capita (PPP$) used indicator of the qual- ity of pregnancy and mFternal nutrition. The incidence of low birth weight was estimated to be about 7.3 percent in 2000 (GSO, 2000). However, a survey in seven provinces in 1999 indicates that the percentage of low-birth weight infants varies from 3.2 percent in Dac Lak to 11.6 percent in Nghe An (Table 2.19). 25 Table 2.19: Wei ht of newborns in seven provinces, 199 of which: % of newborns % weighing 2,500 % weighing more % whose weight Province weighed after birth Rm or less than 2,500 gm was not known Thai Nguyen 84.0 7.2 88.0 4.8 Thai Binh 97.9 8.9 88.9 2.1 Nge An 98.2 11.6 86.6 1.8 Khanh Hoa 95.3 7.9 89.4 2.6 Dac Lak 84.2 3.2 87.1 97 Tay Nioh 97.4 11.3 86.1 2.6 Can Tho 85.5 11.0 78.5 10.5 Source: MoH 1999a. Age at First Birth. The age at which a woman has her first child has important implications for her health and the health of her child, as well as for her work opportuni- ties in the labor force. In addition, age at marriage is strongly related to total fertility; women who postpone their first births are more likely to have fewer children than women who have early first births. Median age at first birth in Vietnam is 23.1 years, with some variations across regions. It is relatively high in the Southeast region (24.9 years), the Central Highlands (23.7 years) and the Red River Delta (23.5 years). It is lowest in the Northern Uplands (22.2 years) and the Mekong River Delta (22.7 years). Contraceptive Prevalence. The igure 2.13 Contraception previlence rates nd real GDP per capim contraceptive prevalence rate (CPR), or the current use of con- traception by currently married women in reproductive ages, is an indicator that measures the success of family planning programs. Ob- 50 viously, higher contraceptive 2 4 prevalence rates (CPRs) will be associated with lower fertility ' rates. The CPR is estimated to be 2 -- ---- 75 percent for currently married women aged 15-49 years in Viet- nan-up from 53 percent in 1988 RA GDP per cVita (PPPS) and 65 percent in 1994. A CPR of 75 percent is extraordinarily high in comparison to most other developing countries, and is the second highest in the region after the Republic of Korea. As Figure 2.13 suggests, a country at the level of Vietnam's per capita GDP would be expected to have a CPR of about 40 percent-not the 75 percent that was achieved.10 Contraceptive Method Mix. The most common currently used contraceptive method in Vietnam is the IUD (used by 39 percent of currently married women aged 15- 10 Of course, not all of the married women currently using contraception use modem methods. Approxi- mately, one-quarter of all women currently using contraception use traditional contraception methods, such as periodic abstinence and withdrawal. 26 49 years), followed by withdrawal (12 percent). The predominance of IUDs in the method mix was something that was observed even in the 1988 VNDHS-I. At that time, nearly 87 percent of married women aged 15-49 currently using modem contraceptives relied on IUDs. That ratio has declined to 69 percent in 1997. The popularity of condoms and pills have also increased relative to 1988, but is still very low. A significant factor behind the fertility decline in Vietnam is the extensive use of menstrual regulation (pregnancy termination within five weeks of conception) and in- duced abortion (termination of pregnancies of up to 12 weeks) to terminate pregnancies, even though these are not considered as contraceptive methods by the Government. Ex- tensive use of pregnancy termination can be seen as a measure of lack of accessibility to safe and affordable family planning services. Induced abortion has been legal in the North since the late 1960s and in the South since 1975. It used to be performed only in district and provincial hospitals, although the service is now provided even in some commune health centers. According to the 1988 VNDHS-I, about 7 percent of the mar- ried women in the reproductive age groups had sought recourse to either abortion or men- strual regulation. The VNDHS-II indicates that this ratio has more than doubled to 15 percent. Both of these numbers are likely to be serious underestimates, as Ministry of Health data suggest a total abortion rate per woman of 2.5-significantly higher than the rate of 0.5 pregnancy terminations per woman as estimated by the VNDHS-II (Goodkind 1995; NCPFP forthcoming). The high rates of abortion pose substantial health risks to women, with more than a third of women reporting a health problem following a preg- nancy termination. Pregnancy termination is much more common among better-educated women. While only 5.5 percent of women with no schooling have experienced a pregnancy ter- mination, the proportion is as high as 22 percent for women with completed higher sec- ondary education. There are also wide regional variations in the pregnancy termination rate, with the rates being very low in the Central Coast (2.1 percent), Central Highlands (6.6 percent) and the North Central region (7.5 percent). The highest rates of abortion and menstrual regulation are in the Northern Uplands (24.1 percent) and the Red River Delta (21.4 percent). C. Nutritional Outcomes 1. Child Nutritional Status Levels and Age-Specific Patterns. While Vietnam seems to have far superior lev- els of infant mortality and fertility than most other countries at its level of per capita GDP in the region, its performance in the area of child malnutrition leaves much to be desired. Indeed, child malnutrition rates in Vietnam are among the highest in the region. A recent national survey of nutrition conducted by the National Institute of Nutrition and UNICEF indicates that 39 percent of children under the age of 5 years are malnourished in terms of weight-for-age and 34 percent are undernourished in terms of height-for-age.II 11 Another national survey-the Vietnam Living Standards Survey of 1998-indicates an underweight rate of 36 percent and a stunting rate of 35 percent. 27 FIgure 2.14: Age-specific prevalence of stunting, underweight and wausting a nhong Malnutrition rates are children aged 0-59 months, 1998 much lower in the first year 5 of life, but increase sharply in the second year of life 40 (Figure 2.14). Underweight rates remain more or less 30 constant beyond age 2, but - . L ndcmc.X' stunting rates continues to increase, peaking at 52 per- cent at ages 54-59 months. 12 Wasting rates are highest during the second year of 0 life, and fall by as much as 0-5 6-11 12-17 18-23 24-29 30-35 36-41 42-47 48-53 54-59 50 percent in the third year. Age (months) The cumulative deficits of underweight child malnutrition begin in the first year of life for a significant proportion of children (one in five). Reasons for this may be low birth weights, sustained and nurtured by inadequate breast-feeding and complementary feeding practices. In addition, the high rates of maternal Vitamin A deficiency may mean that the quality of breast milk that infants obtain is inadequate. There is some evidence that the proportion of lactating women with low Vitamin A levels in breast milk (i.e., < 1.05 mol/L) has actually increased over time-from 41.1 percent in 1995 to 48.5 percent in 1997 and 56.3 percent in 1998 (NIN and UNICEF 1999). For many children, malnutri- tion sets in during weaning when breast milk intakes decline sharply and adequate com- plementary feeding is crucial for growth. That problem may be further complicated by premature introduction of weaning foods. Trends. As the National Institute of Nutrition has conducted several nutrition sur- veys over the last two decades, some idea can be obtained about trends in child malnutri- tion over time. These surveys suggest a mixed picture. Underweight malnutrition has de- clined, but at a sluggish rate of only 25 percent during the last 15 years (Figure 2.15). The rate of stunting has fallen much more impressively-at the rate of 43 percent over the same time period. Because stunting has declined faster than underweight malnutrition, the rate of wasting (i.e., deficiency of weight in comparison to height) has actually in- creased-from 7 percent in 1982-85 to 11 percent in 1998. 12 These trends are similar to those found in malnourished children elsewhere, although height-for-age usu- ally levels off after 3 years. 28 Socioeconomic Varia- Figure 2.15: Prevalence of underwaght stunting and wasling anong chldren under age 60 months, 1982-85 to 1998 tions. Another national survey- 70 the VLSS 1998-provides rates 60 of child malnutrition by the 60 .7 M92.85 child's socioeconomic status. 49 47 1987-89 These data, shown in Table 2.20, 45 45 f1995 42 F1998 indicate large differences in the 40 rates of child malnutrition across economic groups. For instance, U 3 while only 13 percent and 16 percent of children aged 0-59 20 months belonging to the top . quintile were stunted and un- derweight, respectively, in 1998, O the corresponding proportions Underveght Stuntsng Wasting for children in the bottom quintile were 42 percent. Table 2.20 also suggests that the rate of decline in child malnutrition was much greater among the top quintile than among the poorer quintiles. Table 2.20: Rates of malnutrition by per capita expenditure quintile, 1993 and 1998 (% of children aged 0-59 months below relevant NCHS standards) Underweight Stunted Wasted Per cap exp quintle 1993 1998 1993 1998 1993 1998 Bonom 58.71 42.16 68.46 41.55 4.49 13.06 Second 5251 41.33 58.19 42.45 6.58 12.68 Third 48.93 32 17 54.96 33.08 5.23 10.18 Fourth 42.24 31 83 45.00 28.18 5 52 5.09 Top 33.68 16 11 34.03 12.90 5.58 9.47 Total 48.80 35.62 54.27 34.94 5.47 10.97 Source- VLSS 1993 and VLSS 1998 Figure 2.16 plots the age- and sex-adjusted weights and heights of children aged 0-5 years in 1993 and 1998 for 11 real per capita expenditure groups.' Since expenditure is defined in constant 1998 dong, the expenditure groups are comparable across the two periods. Figure 2.16 shows that there was a definite increase in both heights and weights between 1993 and 1998 across all expenditure groups. The larger increase in height and weight for higher expenditure groups than for lower expenditure groups is confirmed by the figure. For instance, children living in households whose annual per capita expendi- ture was greater than VND 10,000,000 experienced a 12 percent increase in weights and 13 Age and sex adjustments were done on the basis of an estimated regression of height and weight on age and age squared. In addition, a dichotomous variable, representing the method of height measurement (i.e., whether height was measured standing up or lying down), was included in the height equation. All heights and weights shown in the figure are calculated for a 'reference' child of age 33 months and represent a simple average of male and female values. 29 a 7 percent increase in heights, but those in households with less than VND 75,000 an- nual per capita expenditure experienced a weight gain of only 2 percent and a height gain of 4 percent. Thus the economic disparity in child nutrition appears to have widened over the last five years. Figure 2.16: Height and weightx of children aged 0-5 years. hoy real per capita Sex Dfferetials Theconsumption expenditure, 199 and 1998 Sex Differentials. The1 NIN-UNICEF survey indicates slightly higher rates of under- weight and stunting among girls - 9 than among boys, but lower a5 rates of wasting (Figure 2.17). 80 Since the differences are small in magnitude, it is probably 1- unlikely that it represents paren- it - 0 tal discrimination against girls in the intrahousehold allocation of -7 60 f o .0-750 751- ]-(X)I- ",1 4,001- 1501- 33501- 4,501- 6,fX)- 7,0sn 1,0 food. 1( X -Y S 3 150 6() 3 Q Anro..I per ctpini cons. expenditlure (in *O00 1998 VND) Regional Variations. Al- figure 2.17.Prevatence of underweight, stunting and wousting among children though there are some variations in the child malnutrition rate 40 across ecological regions, the 36 variations are not as large as in the case of other social indica- tors, such as infant mortality or fertility. Even the regions with 20 the highest living standards-the E Southeast Region and the Red River Delta-have unusually high rates of malnutrition (32- 40 percent underweight and 25- U..-ight stunting Wasting 34 percent stunting) (Figure 2.18. Tis rter sggets hatFigure2.18: Prevalencendf underweight,stunting and wasting among childreo 2.18). This further suggests that udrae6 ots yrWo " cultural and social factors have 0 an important role to play in de- 4 C stig termining child malnutrition in 40 cosupto exediue 199 an39 Vietnam.E3534L,2 ~20 '01 37 1 Annus elal prCaponsu Cot expend lure i s R00ve998VND 306 International Comparisons. Vietnam has one of the highest child malnutrition rates in the region. In 1995, for example, it had the highest rate of underweight children in Southeast and East Asia-higher than Laos, Myanmar and Indonesia.14 Only the coun- tries of South Asia-India, Bangladesh and Nepal-had higher malnutrition rates than Vietnam. Vietnam's poor per- Figure 2.19. Malnutrition and Infant mortality across countries in Asia, 1995-97 formance on child malnutrition is especially surprising in view of its stellar performance on in-Z 70 fant mortality. Figure 2.19, 00 - which plots the relationship be- Z tween underweight rates and infant mortality for a sample of 5 countries in Asia, highlights this 40 point; Vietnam is observed to have child underweight rates -- that are at least 10 percentage 20 ..... .......... points higher than it would be expected to, based on the ob- to 20 30 40 50 60 70 80 90 100 served relationship between Infant mortity rate, 1997 child underweight rates and in- fant mortality rates across Asia. Figure 2.20 shows changes in the child underweight rate from 1975 to 1995 for Thailand, Indonesia and Vietnam, as well as the regional average rates of child malnutri- tion for these years. While Vietnam's malnutrition rates were well above the regional av- erage and above the rates Figure 220: Prevalence of underweight children aged 0-59 months, selected in the other two countries 6 . in 1975, they continued to remain so even until 5- 1995. (More recent data 0 Indoneda from 1998 suggest that 40 this picture has not Re0onalAvrW changed.) It is thus clear that Vietnam is a regional underperformer in the E20 area of child malnutrition. - ... It is somewhat 2 to -- - ---- puzzling that the factors that drove declines in in- 7 fant and child mortality 1975 1980 1985 1990 1995 and fertility in Vietnam have not operated to similarly reduce child malnutrition rates. This suggests that child malnutrition rates respond to something more than health services coverage and income 14 Of course, Vietnam's malnutrition rates did decline substantially from 1995 to 1998, but so must have those of the other countries. Unfortunately, more recent data for the other countries are not available. 31 growth. It is likely that child malnutrition is intimately linked to child rearing practices and cultural beliefs on child feeding that have been more difficult to change in Vietnam. 2. Micronutrient Deficiencies Micronutrient deficiencies, particularly those of iron, Vitamin A and iodine, are widespread in Vietnam. Iron deficiency or anemia affects all age groups, but particularly pregnant women and children under two years of age. It is estimated that 53 percent of pregnant women, 45 percent of children aged 5 and under, and 60 percent of children un- der the age of one are anemic. Anemia prevalence is particularly severe is the Central Highlands (where 61 percent of the vulnerable population is anemic) and the Mekong River Delta (52 percent). As would be expected, urban residents have lower prevalence of anemia than rural residents (36 percent versus 45 percent). There is also some evidence that between the ages of 11 and 14, boys are at much greater risk of anemia than girls. A study of school children found that, while iron defi- ciency rates were broadly comparable across male and female children aged 8-10 years, boys had significantly higher prevalence of iron deficiency at ages 11-14 years. For in- stance, in the urban areas, while only 1.3 percent of girls aged 14 experienced iron defi- ciency, the rate was as high as 13 percent among boys aged 14 (Khanh 1995). Deficiency of iodine is another major public health problem, especially at high elevations where iodine-depleted soils prevail. However, thanks to an effective salt iodi- zation program, rates of iodine deficiency have declined dramatically in recent years. The percentage of population with low urinary iodine has declined from 84 percent in 1993 to 43.5 percent in 1998 (NIN and UNICEF 1999). Indeed, in 1998, the only regions where median urinary iodine levels were below the 10 mcg/dl level were the Southeast region and the Mekong River Delta. 15About 61 percent of salt samples in the country had high iodine content. 32 III. DEMAND FOR HEALTH SERVICES A. Levels, Trends and Disparities in Utilization 1. Facility-Based Estimates of Overall Utilization A measure of overall Figure 3.1: Utilization rates at public faciffies, 1976-98 utilization of health services that 3.0 14 is often used in the literature is P . w h ern the annual per capita number of 25 medical contacts with the health A 1 services. Based on administra- tive data from health facilities 80 the Ministry of Health has esti- 460 mated an average annual health 1 contact rate of 1.7 per capita and an inpatient admission rate of 68 0 per 1,000 persons for Vietnam. Historical data on utilization, 0.5 0 shown in Figure 3.1, indicate that annual health contacts per capita increased sharply from 1976 to 1984 (rising from a level of 0.7 to 2.3), then re- versed direction to decline a the way back to a level of 1.0 in 1990. After bottoming out in 1990, annual contacts per capita increased steadily during the 1990s to reach a level of 1.7 by 1998. The decline in the annual inpatient rate began even earlier (1980), and by 1990, the rate was about one-half of the rate in 1980. Unlike the annual per capita contact rate, however, the inpatient rate has been declining in the 1990s (except for a brief period of rise in 1992-93). There are indications though that the decline in the inpatient rate may be bottoming out. Dating the decline in utilization rates is important because it is often argued that declining utilization of health services were the direct result of the health sector reforms introduced in 1989. Figure 3.1 makes it very clear that the decline in utilization set in several years before (indeed, a decade before, in the case of inpatient utilization rates) the reforms were introduced and implemented. The early 1980s was the start of the break- down of the health services system in Vietnam as a result of several factors, including the pressures of reunification, sharp reductions in external assistance from multilateral agen- cies, and a deteriorating macroeconomic situation. Indeed, it could be argued that reforms in the health sector were introduced to halt the precipitous decline in the sector. Certainly, the evidence on annual contact rates in the 1990s is one indicator that the health sector reforms may well have benefited overall utilization of health services in the country. 33 How do the utilization rates Figure 3.2: Annual pubtic-sector health contacts percaita, Asia, 1990-95 in Vietnam compare to those in other countries? Unfortunately, SriLanka --. comparable data on overall utiliza- tion of health services is available for very few countries in Asia. 4' Clia . However, an annual public sector contact rate of 1.7 per person com- Vietnam 1.7 pares favorably with those coun tries for which data are available. a. In Figure 3.2, Vietnam's utiliza- Ii, 0.8 tion rate is observed to be lower than that of Sri Lanka, China and Malaysia 0-7 Singapore-all countries that have Cambodia '*s substantially higher per capita in- come levels-but greater than that of Pakistan, India, Malaysia and Figure 33: Hospital bed occupancy rate and Inpatient days per capta, Cambodia. While the decline in the 9 . 85 1.0 inpatient admissions rate between . 5/'Rdocpnyrt 1993 and 1998 noted earlier may 8o \d p r be of concern, it is important to 75 -.9 note that two other measures of utilization of inpatient services- 6 -.7 viz., the number of inpatient (bed) ' 0 6. a 50 Average inpatient days per capita a days per capita and the bed occu- 4 0.5 pancy rate-both show an in- 40 4 crease in recent years (Figure 31 30 ) 3.3). The number of inpatient days per capita has increased by nearly 50 percent between 1993 Figure 3.4: Annual inpatient admissions per 1.000 persons, Asia, 190-95 and 1998, while the bed occu- pancy rate has increased from 73 percent to 91 percent over the 89,0 same period, indicating a fuller utilization of available hospital 68. beds. Additionally, even after the decline in the inpatient admis- sions rate, Vietnam still has 3 among the highest inpatient ad- missions rate in the region (Figure 34 .I d e ,teo l co nre in Indonesia Hongkong Canibodia China S Korea Vietnarn Japan Singapore 3.4). Indeed, the only countries in ans npiatsiaspr1Opeca Figure 3.2: Annual p tiet hatntss a s0 perscait,n s ia 9 9 Asia having a higher inpatient admission rate than Vietnam are Japan and Singapore. 34 2. Survey-Based Estimates of Overall Utilization The annual contact rate per capita discussed above is obtained from the adminis- trative statistics reported by public health facilities. It does not include health services contacts that individuals might have with private providers. In addition, its accuracy or reliability depends on the proper maintenance and reporting of utilization data by public health facilities. Fortunately, there have been two nationally representative household surveys in the last five years (Vietnam Living Standard Survey 1993 and VLSS 1998), data from which can be used to check the facility-based utilization statistics. Data from the VLSS surveys are shown in Table 3.1 below. The annualized pub- lic sector contact rate per capita based on the VLSS 1998 is 1."uite close to the Min- istry of Health's estimate of 1.7. For 1993, the discrepancy was somewhat wider-an es- timated annual contact rate per capita of 0.5 according to the VLSS 1993 and an estimate of 0.9 according to the Ministry of Health. Table 3.1 shows that there are large variations in total utilization across income groups. While the bottom quintile of individuals have an average of 2.4 (nondrug vendor) health services contacts per capita per annum, those in the top quintile have as many as 4.6 service contacts.16 The disparity in total utilization across income groups was present even in 1993. Table 3.1: Annualized health services contact rates, by provider and per capita expenditure quintile, 1993 and 1998 PEr capua expenditure quinrile Prot-ider Total Btae..-m Second 7hird Fourth Top 1993 Public hospital 0.32 0.16 0.22 0.30 0.35 0.56 Commune health center 0.19 0.24 0.19 0.22 0.17 0.12 Ofher goernment facility 0.03 0.02 0.03 0.02 0.04 0.07 Private clinic. doctor 0.66 0.46 0.52 0.78 0.64 0.88 Drug vendor 2.14 2.02 2.40 2.08 223 1.96 Traditional heiler 0.03 0.02 0.04 0.02 0.03 0.04 Other provider 0.01 0,00 0.00 0.00 0.01 0.02 1998 Public hospital 0.60 0.25 0.39 0.52 0.76 1.09 Commune health center 1.5 7 0 5 I )'2 078 0 59 020 Other go%emment Iaciiitv 0 25 0 l o (1 19 024 1 33 0 33 Pnvate clinic. doctor 1.76 I 1" 1 52 1.97 1 69 244 Drug vendor 6 7% 5 45 7 30 55 6S4 6.71 Traditional healer 0 36 025 035 0 27 t 34 0 5S Other pro% ider o 00 1. 01 0 0 0 00 (i 00 0.00 Source: VLSS 1993 and VLSS 1998 data. 16One serious shortcoming of the per capita health services contact measure is that it lumps together very different types of service contacts without weighting them. Thus, for example, a surgery at a public hospital for a life-threatening ailment is given the same weight in the per capita contact measure as a consultation at a commune health center healer for a common cold. 35 The data in Table 3.1 seem to indicate a very large increase in the overall utiliza- tion of health services between 1993 and 1998. However, the annual service contact rates are not strictly comparable across the two VLSS surveys for two reasons. First, the VLSS 1993 only asked individuals reporting an illness in the preceding four-week period whether they had visited a provider. On the other hand, the VLSS 1998 did not condition provider visits on an illness, and simply asked respondents if they had had a health ser- vices contact within the last 4 weeks. Second, the questionnaire design in the VLSS 1993 permitted individuals to report only a single health services contact for an illness episode in the preceding 4 weeks, while no such constraint was imposed in the VLSS 1998 ques- tionnaire. This might explain why the reported number of service contacts per person is so much lower in 1993 than in 1998, as it is often the case that individuals make multiple visits to the same or different providers for treatment of a single illness episode. It is, however, possible to impose ex post the same conditions on the 1998 data as were forced upon the 1993 data by the VLSS 1993 questionnaire design. This would have the effect of reducing the annual service contact rate for 1998, and making the utilization statistics from the two surveys more comparable. The adjusted (to be com- Figure 3.5: Aeage annual health iervice contacts per capita, parable to 1993) annualized 5 by per capita consumption expeoditure, 1993 aad 1998 health services contact rate for 1998 is estimated to be merely 4 4.35 per capita (instead of 10.33, as estimated earlier), which 9 represents a more modest in- crease of 29 percent in the ser- vice contact rate since 1993.7 3 0 Figure 3.5 plots the service con- - tact rates in 1993 and 1998 for 11 different per capita expendi- 2.0 ture groups. Since expenditure is 1,0 150 200 2,5 3 500 6,000 7,5 0 defined in constant 1998 dong, A p c . e (in'000 1998 VND) the expenditure groups are com- parable across the two periods. The figure shows that there was a widening disparity in utilization among different income groups between 1993 and 1998. The lowest income group (i.e., those whose per capita consumption expenditure was under VND 750,000 in each period) experienced no change in the utilization of health services, while middle- and upper-income groups experienced large increases in utilization. 3. Demographic and Regional Variations in Total Utilization There are also variations in the utilization of health services across demographic groups. For all ages, females have a higher annual contact rate per capita than males (11.3 versus 9.3) (Table 3.2). However, among children aged 0-4 years, the opposite is true, 17Note that the adjustment does not imply that the unadjusted annual contact rate for 1998 is overestimated relative to its true value. It is much more likely that the annual contact rate for 1993 is underestimated, as the 1993 survey design was more constrained. 36 indicating either that boys have a higher illness prevalence than girls or that boys are fa- vored over girls in the intrahousehold allocation of health services. At ages 5-14, boys and girls have nearly equal utilization rates, but beyond age 15, females have consistently higher rates of utilization than males. Table 3.2: Annualized health services contact rates per capita, by sex and age group, 1998 Age {years) Provider 0-4 5-14 15-29 30-44 45-59 60 & over All ages Public hospital 0 53 0 2' v 45 r 65 1 24 1.13 0.63 Commune health center 0 63 0 41 1 47 0 2 0 - j 1)4 0 62 Other public 0.20 0.09 0.15 0.31 0.74 0.63 0.30 Pnvate clinic. doctor 2.92 1.00 1.12 2.01 3.34 4.02 2.00 Drug vendor 6.34 4.90 4.90 8.76 9.77 13.77 7,39 Traditional healer 0.03 0.07 0.33 0.38 0.66 1.18 0.39 Other provider 0.00 0.00 0.00 0.02 0.00 0.00 0.00 Males Public hospital 0.96 0.30 0.26 0.48 1.28 1.44 0.58 Commune health center 1.19 0.48 0.25 0.47 0.72 0.76 0.52 Other public 0.19 0.11 0.15 0.10 0.24 0.83 0.20 Private clinic, doctor 3.50 1.06 0.60 1.74 1.96 2.71 1.50 Drug vendor 7.97 4,27 4.31 7.41 7.84 10.75 6.13 Traditional healer 0.14 0.05 0.36 0.38 0.32 1.06 0.32 Other provider 0 00 0 W) 0 00 00 0 00 0 00 0 00 Source VLSS 1993 and VLSS 190% The annual health contact rate varies significantly across regions as well. It is lowest in the Northern Uplands and the Central Highlands and highest in the Southeast and the Mekong River Delta (Table 3.3). Indeed, the annual per capita contact rate in the Southeast is nearly 22 times that in the Central Highlands. Table 3.3 shows that the pro- vider-specific contact rates also vary across regions. For instance, the contact rates for public hospitals are highest in the Southeast and Red River delta regions and lowest in the Northern Uplands and Central Highlands. On the other hand, the highest per capita contact rates with commune health centers are found in the Mekong Delta, Northern Up- lands and Central Highlands. Contact rates with private clinics are highest in the South- east and Mekong River Delta and lowest in the Northern Uplands. In addition, there are regional differences in the distribution of service contacts across providers. Public providers, particularly commune health centers, account for a much larger proportion of total service contacts, while private clinics and doctors account for a smaller share of service contacts, in the Northern Uplands than in other regions (Ta- ble 3.3). The Central Highlands region has the lowest share of drug vendors in total ser- vice contacts in the country. These patterns most likely reflect the relative scarcity of pri- vate providers and drug vendors in these two regions. 37 Table 33: Annualized bealth services contact rates per capita, by region, 1998 Northern Red River North Central Central South Mekong Provider Uplands Delta Central Coast Highland East Delta Public hospital 0.52 0.65 0.51 0.49 0.50 0.94 0.56 (8.94) (6.17) (4.72) (5.42) (8.69) (7.70) (4.05) Commune health center 0.69 0.51 0.48 0.38 0.53 0.27 0.88 (11.90) (4.83) (4.45) (4.25) (9.11) (2.20) (6.33) Other public 0.26 0.47 0.30 0.06 0.12 0.10 0.21 (4.54) (4.47) (2.74) (0.65) (2,05) (0.81) (1.54) Private clinic, doctor 0.56 1.99 1.70 1.08 1.46 2.37 2.61 (9.57) (18.90) (15.59) (12.03) (25.31) (19.32) (18.86) Drug vendor 3.56 6.57 7.26 6.56 2.96 8.34 9.18 (60.98) (62.38) (66.74) (73.26) (51.21) (67.96) (66.41) Traditional healer 0.23 034 0.63 0.39 0,21 0.25 0.39 (3.91) (3.24) (5.75) (4.39) (3.62) (2.01) (2.81) Other provider 0.01 0.00 0.00 0.00 0.00 0.00 0.00 (0.17) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) Note: Figures in parentheses are percentages of the total. Source: VLSS 1993 and VLSS 1998. 4. Health-Seeking Behavior of Individuals Which types of health providers do individuals belonging to different income groups visit when they are ill? And how has the health-seeking behavior of different in- come groups changed over time? These questions are addressed in two different ways below. First, the distribution of total health services contacts across providers are ana- lyzed and interpreted. Next, the provider-specific distributions of health services contacts for each type of curative health service-inpatient, outpatient and medicine/drugs-are discussed. Table 3.4 reports the distribution of total health services contacts across providers for each of five income groups in 1993 and 1998. These data reveal two important find- ings. First, there is very heavy reliance on drug vendors by all income groups in both pe- riods, with drug vendors accounting for roughly two-thirds of all service contacts." There was a small increase in the share of drug vendors in total contacts between 1993 and 1998. In both periods, drug vendors were somewhat more ubiquitous among the poor than among the better-off. 18 As noted earlier, only visits to drug vendors without a prescription from another medical provider are considered in this analysis. Visits to drug vendors to fill a prescription obtained from another health pro- vider are assumed not to constitute a separate health service contact. 38 Table 3.4: Distribution of health services contacts by providers and per capita expenditure quintiles, 1993 and 1998 Per capita eApendaurc quindie Pro Itdcr Bounrn SeconJ Third Fourth Top Total 1993 Public hospiial, 5 33 t 34 8 67 10 09 15.45 9.38 Commune helth center h 23 5 "O 6 52 4 89 3 23 5.61 Other public pro%iders 7 *'; 0 o 0 "2 1.03 1.80 1.04 Ali pubb prouders 14 31 124 1591 1601 20.48 16.03 Pn,ate clinics. doctors 15 81 15 41 2283 IS 53 24 16 19.52 Drug vendors 69 32 70 55 60 62 64 43 53 79 63.44 rradtiunal healer. 0 56 I 12 0.56 0 79 1.13 0 84 Other pr-. iders 0 0.1 0 OS 008 024 045 0.18 Toiji 100 i 1 00 00 10000 10000 100.00 100.00 199S Public hospitals 3 15 3 75 4.63 22 9 55 5.84 Commune health center 7.39 6 84 691 5.62 1 72 5.56 Other public pro. ider. 2 05 1 7t) 2 13 3.14 2.89 2.42 All public pro%iders 12 59 12 38 13 67 15.98 14.16 13.82 Pn%ate clinic.. doctors 1493 14 4o 17 38 16.00 21.38 17.02 Drui .endors 09 20 69 t 66 56 64.81 5937 65.66 Tradional healkrs 3 19 3 38 2 39 3.21 5.10 3.48 Other prov% ider 0 OU 0 C19 0 00 0 00 0 00 0.02 Total 100 UU 100 U I (0 0O 10000 100.00 10000 Source VLSS 1993 and \LSS 1998 dat Second, the share of hospitals in total health services contacts is very strongly as- sociated with economic status. Among the bottom 20 percent of the population, hospital visits constitute 3.2 percent of total service contacts, but this share increases to 9.6 per- cent for the top 20 percent of the population. This is a common finding in most develop- ing countries, and represents a shift to higher-quality providers (at least as perceived by users) with increasing affluence. However, there was an appreciable decline between 1993 and 1998 in the relative share of hospitals in total service contacts-from 9.4 per- cent to 5.8 percent. Unfortunately, the VLSS 1998 data do not permit a distinction be- tween inpatient and outpatient hospital visits, so it is not possible to tell whether the relative decline in the use of hospitals between 1992 and 1998 represents a relative reduc- tion in the demand for inpatient services or whether it represents a decline in the use of hospitals for outpatient care. Instead of lumping together health services contacts that are qualitatively very different from each other, it may be instructive to analyze the distribution of health ser- vices contacts across providers within each of three submarkets for curative health care: inpatient services, outpatient services, and medications. However, there are two formida- ble problems in undertaking such a disaggregated analysis. First, as noted earlier, the VLSS 1998 data do not distinguish between inpatient and outpatient hospital visits. Sec- 19 A separate module in the VLSS 1998 did obtain information on hospital admissions during the 12 months preceding the survey, but the 4-week recall data on utilization of health services did not distinguish between inpatient and outpatient hospital visits. 39 ond, the VLSS 1998 data do not report information on the ownership (whether public or private) of drug vendors visited by individual respondents. Table 3.5 shows the provider distribution of inpatient and outpatient health ser- vices contacts under the (strong) assumption that all hospital visits during the four-week reference period were for inpatient purposes. Four conclusions can be derived from this table. First, the public sector is the exclusive provider of inpatient services among all in- come groups. While there are a few private hospitals in the country, their share in total inpatient service contacts is negligible. Table 3.5: Distribution of inpatient and outpatient health services contacts by providers and per capita expenditure quindles, 1993 and 1998 Per capita expenditure quntle ProL ider Bottom Second Third Fourth Top Total 1993 Inpatient services Public hospital 10000 10000 100.00 10000 100.00 100.00 Pnsate hospital Q 00 0 00 0.00 000 0.00 0.00 Ouipaient sers ices Commune health center 3243 24 36 21 15 19.10 10.62 20.65 Other public provider 2.70 3.85 I 92 4.49 6.19 3.26 Priate clinic, doctor 62 16 66.67 7500 71.91 77.88 71.74 Traditional healer 2 70 5.13 1 92 3.37 3.54 3.26 Other pro1ider 000 0.00 000 1.12 1 77 1.09 All outpatient providers 100.00 100.00 10000 100.00 100.00 100.00 1998 Inpatient ser% ice-, Public hospital 100.00 100.00 10000 100.00 100.00 100.00 Private hospital 0 00 0.00 0 00 0.00 0.00 0.00 Outpatient services Commune health center 26.85 25 81 23.93 20.00 5.63 19.39 Other public provider 7.41 6.81 7 36 11.19 9.30 8.50 Pn ate clinic, doctor 54.17 54.48 6043 57.29 68.73 59.86 Tradional healer 11 57 12 54 8.28 11.53 16.34 12.24 Other pro% ider 000 036 0.00 0.00 000 0.00 All outpatient proViderS 100.00 100.00 100.00 100.00 100.00 100.00 Source VLSS 1993 and VLSS 1998 data. Second, the private clinics' share of all outpatient visits decreased significantly between 1993 and 1998-from 72 percent to 60 percent-despite an increase in the sup- ply of private-for-profit (outpatient) health services. Third, there was a large increase in the share of traditional healers in total outpa- tient service contacts. While visits to traditional healers constituted 3 percent of all outpa- tient visits in 1993, their share had increased to 12 percent by 1998. Fourth, the distribution of outpatient visits across health providers varies mark- edly across income groups, with a pronounced shift toward private clinics and doctors, and away from commune health centers, with increasing income. 40 5. W ho Uses Public Fgur 3.6: Economic composition of users of different health providers, 1993 Health Facilities? Al Another way of looking 0 at the data on provider distribu- tion of health services contacts presented in Tables 3.4 and 3.5 is to examine the economic 2 composition of all of the users of a particular health provider. 25 This information is important in 4 knowing which types of public Public hospitals Tradtional Private clinics Dragvendors health expenditures are likely to Per capita expenditure qintile be more propoor than others. R I * T Fout M Annex Tables A3.1I and A3.2 present the data in the form of Lorenz distributions of health services contacts for each type of provider. The data in Annex Table F1igure 3.7: Economic composition of users of different health providers, 1998 A3.1 are summarized for five main providers in Figures 3.6 and 3.7. These data clearly indi- cate that the top 20 percent of - the population is over- _ 6 represented among hospital us- ers relative to its population i share, while the bottom 20 per- 2 17 cent of the population is over- 2 2 16 represented among commune 0% health center users. For in- Public hospitals Traditional Private chnics Drug vendors CHCs healers stance, in 1998, 36 percent of all Per capita expenditure quintile public hospital users were to Second thirdl0Fo drawn from the top quintile and 61 percent were drawn from the top two quintiles. On the other hand, the bottom quintile was vastly under-represented among hospital users relative to their share in total popula- tion, constituting only 8 percent of all hospital users. Furthermore, the data suggest that the distribution of hospital users became more unequal between 1993 and 1998. This is clearly observed in the Lorenz curves for hospital contacts shown in Figure 3.8. The preponderance of more affluent individuals among users of hospitals is cer- tainly not the result of deliberate government policy. If anything, the intentions of the government are precisely the opposite-viz., to ensure access to hospitals by the poor and indigent. To achieve this goal, the government has put in place a number of programs to reduce and even eliminate the private costs of public hospitals for the poor. However, a number of other factors conspire to facilitate access to hospitals by the better-off and re- strict access by the poor. This is a common feature in the health sectors of many develop- ing countries. 41 The first of these factors is related to the physical location of hospitals. Since a large majority of hospitals are located in urban centers, geographical access to hospitals is much better for better-off urban dwellers than for poorer rural residents. The second fac- tor is related to the large private costs associated with hospital use. Since hospitals pro- vide more specialized and higher quality treatment, the cost to an individual of using hospital services is much greater than that of using services supplied by commune health centers. Third, the spread of health insurance between 1993 and 1998 may have uninten- tionally contributed to greater concentration among hospital users. As noted in chapter VII.C, individuals with health insurance coverage tend to be better off than those without insurance coverage, since coverage is mandatory for persons having certain salaried jobs (such as government employees and employees of large firms) but not mandatory for self-employed persons (including farmers) or casual day laborers, who typically earn lower wages. Since the health insurance agency largely reimburses hospital (but not commune health center) costs, there is a bias toward greater use of hospital services (and underutilization of primary health services) by health insurance enrollees. This is an addi- tional factor that may have contributed to the worsening distribution of hospital users ob- served in Figure 3.8. Publc s bsides o a ar-Figure 3.8: Loreni distributions of service contacts with public hospitals, Public subsidies to a par-193ad98 ticular level of health facility ioo (e.g., commune health centers) o will be weakly progressive if the share of the poor among all users Lefrcoul of that facility is greater than I their share in aggregate national consumption expenditure.20 In such a case, the distribution of 19+ commune health center users is 30 more progressive than the under- 20 lying distribution of personal consumption; hence, public sub- . ....... sidies to commune health centers 0 20 40 60 80 00 will be weakly efficient (in terms Curulativeshueofthepopulation of targeting). On the other hand, public subsidies will be strongly progressive if the share of the poor among all users of that facility is greater than their population share. Annex Table A3.2 and Figure 3.9 suggest that public subsidies to commune health centers and public hospitals are both weakly propoor, since the distribution of users of these public health facilities is more progressive than the underlying distribution of personal consumption expenditure. However, only public subsidies to commune health centers (but not to public hospitals) are strongly propoor, as the share of the poor among commune health center users exceeds their share in the population. This was true in both 1993 and 1998. 20Technically, from a public finance perspective, the right comparison is not with private consumption ex- penditure but with tax incidence, but data on this are rarely available. 42 The evidence thus clearly indicates that, despite government intentions and inter- ventions to improve access of the poor to public hospitals, the latter are used dispropor- tionately by the better-off in Vietnam. This means that the better-off capture a much lar- ger share of public subsidies to hospitals than the poor. 6. Hospitalization Figure 39: Lorenz distributions of serice contacts with public hospitals and P ~commusne health centers in comparison to losent distribution of aggregate consusmption expenditure, 1998 Since hospitalization is a 0 relatively rare occurrence in a 9_ population, the VLSS 1998 ob- R tained information on hospitali- 7 zations over a 12-month pe- riod.21 These data, shown below in Table 3.6, indicate an annual ' inpatient rate of 50 admissions 03 per 1,000 population- somewhat lower than the rate of 68 per 1,000 persons estimated by the Ministry of Health on the 0 20 0 basis of data reported by public Cumobs-f poputwn health facilities. The VLSS data confurn the decline in inpatient admissions per capita reported by the Ministry of Health. From 1993 to 1998, the inpatient admissions rate fell by more than 100 percent-from 105 to 50 per 1,000 persons. The government statistics indicate a decline of about 34 percent over the same period. The much larger decline ob- served in the VLSS data might be the result of differing recall periods for hospitalization in the two surveys. While the reference period for inpatient admissions was four weeks in the 1993 VLSS, it was 12 months in the 1998 survey. Since recall error is greater with longer recall periods, it is possible that the 1998 inpatient admissions rate might be un- derestimated relative to its true value, and this would overstate the decline between 1993 and 1998. Table 3,6: Inpatient admission rate and average length of hospital stay, by per capita, expenditure quintile, 1993 and 1998 .4 nnual inpaient rate per . %-erage length of Per LapIa %xpendure qinfie / 0o pi.rson.s per iear slai (davsi 1993 1991 1993 1998 Lws1 70 8 33 Q 7.6,j 10.26 Second 14 5 43 5 S48 10.91 Third 1Il9 493 790 1390 Founh 122.7 61.9 7.93 14.61 Highest 133,7 63.3 7.12 18.76 Total 104,7 50.4 7.77 14.26 Ministy of Health data 90.5 67.7 4.20 8.18 Source: VLSS 1993; VLSS 1998; and various issues of MoH Health Statistics Yearbooks. 21 The 1993 VLSS obtained information on hospitalization during the past 4 weeks. 43 Another measure of Igure3.10: Average length of hospital inpatient say(days),Asia, 1990-95 utilization of inpatient health S K _719.2 services is the average length of hospital stays. The VLSS data China 150 show a very high average length 1 1 of stay of 14 days in 1998. This 11anglades8 was up nearly 100 percent since ILI 1993. Interestingly, facility- based data from the Ministry of Health also show a 95 percent t0 increase in the average length of a hospital stay between 1993 Malaysia and 1998, although the mean SrLanka 42 length of hospital stay estimated by the Ministry is about 60 per- cent of that estimated by the VLSS. The VLSS estimate indeed appears too large, and is most likely an overestimate. 22 An average length of inpatient stay of even 8.2 days appears large in comparison with comparable rates in Sri Lanka, Malaysia and Indonesia (Figure 3. 10). This evidence, plus the fact that the average length of an inpatient stay nearly doubled between 1993 and 1998, seem to point to an overuse of hospitals in Vietnam.23The decline in the inpatient admission rate and the simultaneous increase in the average length of an inpatient stay is puzzling. It suggests that hospitals could be managed more efficiently. Both the rate of hospital Figure 3.11: Loenz distributions of hospital inpaienst days. Sr99 Lank 4992 admissions and the average 100..~ .. . . length of stay vary considerably across economic groups (Table ~ 3.6). For instance, the top qum- Th VLSSLes i e ofinded apearsltooi tile has an inpatient admission of an i rate and an average length of aipaiet_ta.i 3 50 - - ----- --- . hospital stay that are nearly two 0 times as large as comparable 90 rates for the bottom quin- .ii Since the poor, in fact, are likely so0 to bear a much larger burden of 0 illnesses than the better-off, the 001 4 6 0 8 0 100 data in Table 3.6 suggest that Cumulative share ofthe popuation 22The VLSS estimates of inpatient admission rate and average length of hospital stay would imply utiliza- tion of 56 million bed days during 1998--or 153,700 hospital beds each day of the year. In fact, however, there were only a total of 133,190 hospital beds available in the country in 1998. 23The length of inpatient stays observed in the VLSS could imply that (i) individuals wait until their ill- nesses are very serious before getting admitted to a hospital, and this necessitates the long hospital stays, or (ii) hospital authorities do not sufficiently monitor or regulate the length of inpatient stays, with the result that hospital stays are inefficiently long. 44 patients view longer hospital stays as synonymous with better-quality care, and the bet- ter-off are willing (and able) to pay for this better care. However, it is not clear that this is the best use of scarce hospital resources. The combination of higher inpatient admission rates and longer inpatient stays for the better-off relative to the poor means that aggregate hospital inpatient days are even more unequally distributed in the population than inpatient admissions or average length of stay. In 1993, the top 20 percent of the population accounted for 24 percent of aggre- gate inpatient days, while the bottom 20 percent accounted for merely 13 percent (Figure 3.11). By 1998, the distribution had worsened further, with the share of the top quintile in aggregate hospital days climbing to 33 percent and the share of the bottom quintile falling to 10 percent. B. Utilization of Preventive Health Services Immunization services. One of the Vietnam's greatest successes in the health sec- tor has been its Expanded Program of Immunization (EPI), which was implemented on a pilot scale since 1982 and expanded to the whole country in 1985. Since then, the pro- gram has grown impressively. Table 3.7 shows that the most significant achievements of the immunization program were achieved during the late 1980s-between 1986 and 1990-although coverage continued to expand beyond 1990. For example, coverage of measles increased from 39 percent in 1986 to 87 percent in 1990 and to 96 percent in 1996. Table 3.7: Iuinization coverage of children under 1 year of age, 1986-96 TyRe /986 /990 1996 BCG (tuberculosis) 54.5 89.9 95 4 Polio 447 865 94 5 DPT (diphthena. pertussis. and tetanus) 426 86 7 94 4 Measles 38.8 86.6 96 0 Fully vaccinated A7 0 95 1 Source. MoH 1996b As noted in chapter II.A.3, the increase in immunization coverage has been asso- ciated with a dramatic reduction in vaccine-preventable deaths. Vietnam is expected to completely eradicate polio this year. Likewise, neonatal tetanus has been virtually elimi- nated in 591 out of 610 districts in the country (MoH 1999b). Deaths from measles and tetanus have also declined considerably. The 1997 Vietnam Demographic and Health Survey (VNDHS-II) confirmed the generally high rates of immunization coverage reported by the Ministry of Health. Be- cause the reference child population was different in the VNDHS-II (viz., children aged 12-23 months as opposed to 0-12 months), the immunization rates were slightly differ- ent. For instance, the VNDHS-II found that the measles immunization coverage among 12-23 months was 77 percent-not 96 percent as reported by the Ministry of Health (Ta- ble 3.7). 45 The VNDHS-II found that immunization rates were slightly greater for girls than for boys, and that higher-order children were much less likely to be immunized than firstborn or second-born children. Maternal education was also an important determinant of child immunization. For instance, while measles coverage was 65 percent among chil- dren of mothers with no schooling, it was as high as 93 percent among the children of mothers with higher secondary schooling. Both the Ministry of Health and VNDHS-II data confirm the relatively small in- terprovincial variation in immunization coverage rates (Annex Table A3.3). As expected, immunization rates are lower in the Northern Uplands and the Central Highlands, and are highest in the Southeast and Red River Delta. However, the differences are small, and child immunization rates are generally high in all regions. Maternal Care. Unlike immunization services, however, utilization of antenatal health services is more limited. About 72 percent of all live births in the period 1994-96 were attended by a doctor, nurse or midwife, with a quarter being attended by a doctor and 46 percent being attended by a nurse or midwife (NCPFP forthcoming). There are large regional variations in the proportion of births not attended by a doctor, nurse or midwife. In the Central Coast and the Central Highlands, 37-40 percent of all births are not attended by any health professional, while the corresponding propor- tion is as low as 12-18 percent in the Southeast and the Red River Delta. One of the strongest determinants of the use of antenatal services is a woman's education. While 70 percent of births of women with no schooling were not attended by a professional health worker, the corresponding proportion was as low as 9 percent in the case of births of women with higher secondary schooling (NCPFP forthcoming). There is almost a perfect correlation between the use of antenatal health services and the proportion of women receiving tetanus toxoid vaccinations during pregnancy. Approximately 28 percent of all pregnant women in 1994-96 did not receive any tetanus toxoid vaccination, while 17 percent received only one dose of the vaccine. More than half (55 percent) received two or more doses. C. Determinants of Utilization Why are the health services utilization rates of the poor lower than those of the nonpoor in Vietnam? The utilization of health services is influenced by a host of factors, such as geographical access (or lack thereof) to health facilities, the full private cost of using health services, the quality of health services, household income, education, age, sex and health insurance coverage. The full private cost of health services includes not only the official fees that may be charged by health facilities but also the costs of drugs, medical supplies, and informal (incentive) payments charged by providers. In addition, the opportunity cost of an individual's time associated with visiting a health facility may be an important factor in determining utilization in a developing country where distances to health facilities may be long. Recent evidence from a number of developing countries has shown that the de- mand for health services, particularly hospital-based and private health services, is highly income elastic. Controlling for other factors, individuals can afford more health care and 46 higher-priced health services as their incomes increase. The high income elasticity for secondary and tertiary care and for privately provided health services means that the dis- parity in utilization rates between the poor and the better-off will increase in a market- oriented health care system Another major determinant of the demand for health services is price. The higher the price of health services, the lower will be the demand for these services. This explains why the utilization of health services by the poor is less than that of the better-off-they face higher prices of health services in relation to their incomes. The perceived quality of health care may additionally influence its demand.24 Controlling for price, an improve- ment in the perceived quality of health services will typically increase utilization. Con- versely, utilization rates at health facilities perceived by the public to be of poor quality will be low. This factor is likely to work against the poor, as the poor in most countries typically have access to lower quality facilities relative to the better-off. This also means that the health care prices faced by the poor should be standardized for quality. A fourth determinant of health-services utilization is geographical access. Geo- graphical access is not likely to be an important contributing factor to utilization of health services in Vietnam, as the country has an extensive network of public health facilities. Of course, access is difficult in the remote and mountainous areas, where population den- sities are low and the terrain is difficult to traverse. Since the poorest people in the coun- try live in these areas (viz., the Northern Uplands and the Central Highlands), average distances to health facilities are on average greater for the poor than for the nonpoor. The longer distances raise the opportunity cost of seeking treatment from a health provider and thereby act as a deterrent to utilization. Education is another important determinant of utilization. There is a large body of empirical evidence which shows that, controlling for other factors, individuals with more schooling are more likely to utilize health services than individuals with no schooling. This is particularly true for the utilization of health services by children; children whose mothers have more schooling are more likely to obtain treatment for an illness than chil- dren whose mothers have little or no schooling. The fact that the poor typically have lower levels of schooling would thus help explain why they have lower levels of health- services utilization than the nonpoor. Health-Seeking Behavior. A number of qualitative studies have attempted to un- derstand the health-seeking behavior of poor people in the rural areas. While it is obvi- ously difficult to generalize from such studies, which tend to be based on very small samples owing to their open-ended questions, there are a number of findings that are con- 24 The notion of quality naturally covers several attributes, many of which may not necessarily be related to efficacy of medical treatment. These attributes could include courtesy shown to patients, shorter waiting times, cleaner and less-crowded facilities, and ample availability of medical supplies, drugs and health workers at the health facility. It is possible, although only in some cases, that patients incorrectly perceive that they are purchasing better quality of care because they are given saline injections or drips or unneces- sary drugs by the 'high-quality' providers. 47 sistent across many studies. These are summarized below to provide a qualitative feel for how the poor deal with an illness and how they make decisions on treatment.25 The large majority of focus group studies reveal that the poor in rural areas make use of these services only as a last resort. They have confidence in their health, and be- lieve that their illness will go away by itself Of course, a large part of this may be influ- enced by the cost of health care, which is perceived to be very high by the poor. But some of it may also be due to a lack of trust and confidence in health workers at the grass-roots level. There would appear to be differences in the knowledge and perceptions of the poor in the North and in the South. In the North, even the poor in the rural areas are gen- erally more knowledgeable about common health problems owing to the fact that educa- tional levels are higher and health programs at the grass-roots level have been around for longer in the North than in the South. The qualitative studies confirm the findings from household survey data regarding the extensive use of self-treatment in the rural areas. The similarity in how people deal with an illness is surprisingly similar across different income groups and across different regions. When medicines are supplied free of charge at commune health centers, as they are in some remote mountainous areas (under a government program to provide free or subsidized drugs to remote regions), ethnic minority people, who generally have little discretionary income, do visit the commune health centers, if only for obtaining these drugs. When the subsidized or free medicines are not enough-e.g., the free drug supply lasts for only three days in a five-day recommended course-people tend to stop their treatment halfway. It is common in a mountainous market to see a mother buying a single tablet for her child. That tablet is often the symbol of all her knowledge about and belief in modern treatment. It also is all that she can provide to her child, given her limited means. The concepts of drug efficacy and recommended drug courses are, for the most part, still limited among people in the mountainous areas (as they are among people in other locations as well). Accessibility of commune health centers is another important factor determining utilization, particularly in the remote mountainous areas. In these regions, users are very sensitive to distance, especially as it is difficult to traverse, and typically visit the nearest health facility, irrespective of what it is. Although the accessibility of commune health centers is generally very good in Vietnam-much better than in many other developing countries-a large proportion of the population, especially in the mountainous areas, lives more than 3-5 kilometers away from the nearest commune health center. Clients may prefer private providers as they are prepared to come to patients' homes at night to, say, assist in births and treat other problems. They reportedly treat cli- ents more respectfully, and, very importantly, are willing to extend credit that clients can 25 This section draws directly from a number of qualitative studies that were commissioned as part of the Vietnam health sector review. These were done in selected communes in the North, Central and South of the country. 48 repay in cash, in kind, or in labor. In the regions with large minority populations, lan- guage and cultural barriers also figure importantly in the reluctance of users to visit pro- fessional providers of health services who are unable to speak the local minority lan- guage. The qualitative studies indicate that people are not opposed to the idea of paying for health services. Indeed, most people prefer to pay something as they believe it is the only way to receive proper treatment. Of course, this does not mean that people would be willing to pay anything for any type of service. As soon as people pay for a service, they hold the provider of the service accountable for quality and timely service, and expect courtesy, competence, and attention from providers. In this sense, willingness to pay for health care is dependent on the quality of care. D. The Private Cost of Health Care Clearly, the cost of health care is not only a significant determinant of utilization and provider choice, it also has an important impact on the living standards and poverty of individuals. For this reason, it is important to know the full cost of public and private health care for individuals and to examine how onerous the financial burden of health care costs is on the poor. While all health care was available without charge before 1989, fees were intro- duced during that year in the three higher (district, provincial and national) levels of the health care delivery system. An explicit fee exemption mechanism was put in place, whereby certain classes of patients, such as the handicapped, orphans, individuals able to produce certification of indigency from their neighborhood or village People's Commit- tee, and patients suffering from mental problems, leprosy, and BK-positive tuberculosis, were exempted from all or part of the user fees. Health consultations (not including drugs) at commune health centers continued (and still continue) to be officially free. However, drugs are rarely provided free of charge at any facility. For the most part, indi- viduals are responsible for obtaining their own supplies of drugs, typically from private pharmacies and drug vendors or from a for-profit pharmacy run by the public health facil- ity. 1. Out-of-Pocket Expenditure on Health Services Contacts in 1993 and 1998 Average out-of-pocket expenditure per health services contact are shown for 1993 and 1998, by type of provider as well as by per capita expenditure quintile, in Table 3.8. Large variations in out-of-pocket expenditures per service contact are observed across both providers as well as across quintiles. There are three observations that can be made with these data. First, as would be expected, individuals with health insurance incur significantly lower out-of-pocket expenditures on public hospital contacts than individuals without in- surance. However, even for the insured, the out-of-pocket expenditure on a single visit to a public hospital is not trivial (VND 92,410). This reflects the fact that Vietnam Health Insurance (VHI) coverage benefits are not large enough to cover the full cost of a hospital visit. 49 Table 3.8: Average out-of-pocket expenditure on fees and drugs per health services contact (in '000 constant 1998 VND), by provider and per capita expenditure quintile, 1993 and 1998 Per capita expenditure quintile Total Provider Bottom Second Third Fourth Top 1993 Public hospitals 87.17 72,78 162.10 151.66 153.99 137.34 Commune health centers 24.23 40.28 44.40 49.31 58.14 41.01 Private clinics, doctors 29.75 38.03 52.53 64.65 70.31 54,14 Drug vendors 13.57 14.92 17.12 25.18 30.44 20.07 Traditional heaters . 42.43 47.97 126.72 61.88 76.21 1998 Public hospitals 82.50 143.53 169.10 186.17 193.29 168.57 Commune health centers 14.28 20,12 19.19 21.52 36.80 19.94 Private clinics 20.72 35.21 42.51 25.10 41.34 32.73 Drug vendors 9.04 9.23 11.50 15.78 27.31 14.58 Traditional healers 18.33 27.09 42.05 66.85 73.94 48.07 Public hospitals for uninsured 90.33 166.22 192.19 260.16 263,86 208.28 Public hospitals for insured 23.07 67.08 113.82 78.79 105.38 92.41 Prices for insured as % of those for uninsured . 40.36 59.22 30.29 39.94 44.37 Source: VLSS 1993 and VLSS 1998. Second, there are large variations in average out-of-pocket expenditure per ser- vice contact across expenditure quintiles, with better-off individuals paying significantly 26 more for a health visit to the same provider than poor individuals. This pattern holds for every single provider. For instance, an individual in the bottom quintile pays VND 82,500 for a public hospital contact, while a person in the top quintile pays more than two times as much (VND 193,290). The out-of-pocket expenditure on a visit to a drug vendor is three times as much for the top quintile of individuals as for the bottom quintile. The sharp economic differences in out-of-pocket expenditure reflect two facts: (i) some of the poor are exempted partially or fully from paying user fees, and thus end up incurring lower out-of-pocket expenditure, and (ii) better-off individuals purchase more and higher-priced (and presumably higher-quality) health care and drugs than poor individu- als. Indeed, the data in Table 3.8 suggest that the demand for the quality of health ser- vices (as proxied by the out-of-pocket expenditure on a single visit) is highly income- elastic. Third, and most surprising, out-of-pocket expenditure per visit actually fell in real terms between 1993 and 1998 for most providers. Some of these declines are very large. For instance, out-of-pocket expenditure on a visit to a commune health center fell by more than 50 percent in real terms over the five years-from VND 41,010 to VND 19,940. The only provider for which out-of-pocket expenditure per visit increased in real 26 In a 1995 survey of 32 communes in four districts in Northern Vietnam, Ensor and San (1996) found that the poor actually paid higher prices than the better-off at commune health centers. They argued that this might be the result of case-mix variations-viz., that the illness episodes of the poor might be more severe on average than those of the better-off. The latter could occur if the poor waited to obtain treatment until their illness episodes were severe. 50 terms was public hospitals. Overall, out-of-pocket expenditure on a visit to a public hos- pital increased by about 23 percent in real terms over the five years between 1993 and 1998. However, this average figure masks a large difference between the experience of insured and uninsured persons. For those with health insurance coverage in 1998,27 out- of-pocket expenditure on a service contact with a public hospital actually declined by 33 percent in real terms between 1993 and 1998, whereas real out-of-pocket expenditure on a public hospital contact increased by 52 percent for the uninsured. Thus, the only group for whom the cost of health care increased in real terms were public hospital users having no insurance. The real decline in out-of-pocket expenditure per service contact between is ex- tremely surprising and somewhat counterintuitive. It is not very often that medical costs decline in the real world! The mystery begins to unravel if the out-of-pocket expenditures shown in Table 3.8 above are disaggregated into those on user fees (Table 3.9) and those on drugs (Table 3.10). Table 3.9 shows that per-contact user fees increased in real terms across the board between 1993 and 1998. In some cases, the increases were very substan- tial (e.g., 1,400 percent for uninsured persons using public hospitals). User fees at com- mune health centers, which were negligible in 1993, increased significantly as well-to an average of VND 5,260. The latter finding suggests that commune health centers do not really provide free health services. Table 3.9: Average out-of-pocket expenditure on fees per health services contact (in '000 constant 1998 1ND). b pro,ider and per capita expenditure quintile, 1993 and 1998 Per capita exi?enditure quinnie ProJider Bollon Second Turd Fourth Top Total 1993 Pubic hospital% 34 2-6 9.73 11 57 736 7.69 Commune health centers 0 i 1) 0 17 0.02 0.09 0.77 0 20 Private clinics, doctors q 71 I 02 1 26 2.00 3.54 1 90 Drue %endors 0 (H) 00 0 00 001 0.00 0.00 Traditional healer-, 0 00 042 0 34 0 00 7 28 2.26 1998 Public hospitals 32.32 72 6t 73 33 99 63 115.32 88.89 Commune health center% 2 2 406 6 03 5 39 11 60 5 26 Private clinics 1740 2u 4. 3 36 13 00 17 59 12.82 Drug % endors (IUX ) 01? 0 00 0 00 000 0.00 Tradinonal healems 5.72 03 1439 31 05 41 04 21.38 Public hospitals for uninsured 35 94 8462 100(07 14424 16899 115.37 Public hospuals for insured 4 79 32 36 2S 36 34 92 4846 38.12 Prices for insured as ', o1 those for uninsured I3 33 38 24 28 34 2421 28 68 33.04 SOUrLCe VLSS 1993 and VLSS 199S 27 The health insurance program began only in 1993, with hardly any coverage during that year. By 1998, about 12 percent of the Vietnamese population had health insurance coverage. 51 However, the data in Table 3.10 indicate that there was a large decline in out-of- pocket expenditure on drugs across all types of providers. In the case of commune health centers, private clinics, and traditional healers, the declines were very substantial (more than 60 percent). Average out-of-pocket expenditure on drugs for a service contact with a public hospital declined by about 38 percent (28 percent for uninsured patients and 58 percent for insured patients), while out-of-pocket expenditure for a service contact with a drug vendor declined the least (27 percent). Furthermore, these declines were felt by vir- tually all income groups. In the case of private clinics, drug vendors and traditional heal- ers, the bottom quintile of persons experienced much larger declines in real out-of-pocket expenditure on drugs per service contact than the top quintile. However, for public hospi- tals and commune health centers, the size of the declines was roughly comparable across income groups. Table 3.10: Average out-of-pocket expenditure on drugs per health services contact (in '000 constant 1998 VNI), by provider and per capita expenditure qulatile, 1993 and 1998 Per capita expenditure quintile Provider Bottom Second Third Fourth Top Total 1993 Public hospitals 83.65 70.32 152.37 140.09 14613 129,65 Commune health cen- ters 24.04 40.11 44.38 49.22 57.37 40.81 Private clinics, doctors 29.04 37 01 51.27 62.65 66.77 52.24 Drug vendors 13.57 14.92 17.12 25.17 30.44 20.07 Traditional healers 139.62 42.01 47.63 126.72 54.60 73.95 1998 Public hospitals 50.19 70.87 95.76 86.53 77.97 79.67 Commune health cen- ters 11.46 15.16 13.16 16.13 25.20 14.68 Private clinics 3.32 14.73 39.16 12.11 23.75 19.91 Drug vendors 9.04 9.23 11.50 15.78 27.31 14.58 Traditional healers 1261 20.06 27.66 35.81 32.89 26.69 Public hospitals for uninsured 54.39 81.60 92,12 115.93 94.87 92.91 Public hospitals for insured 18.28 34.71 85.46 43.87 56.92 54.28 Prices for insured as % of those for uninsured . 42.54 92.77 37.84 60.00 58.42 Source: VLSS 1993 and VLSS 1998. The decline in out-of-pocket expenditure on drugs between 1993 and 1998 could reflect a decline in the utilization of drugs over the period or a decline in the price of drugs. The former possibility appears extremely unlikely, as casual empiricism and on- the-ground assessments suggest that the overall supply and availability of drugs and pharmaceuticals in the Vietnam market improved markedly between 1993 and 1998. This suggests that drug prices must have declined across the board. Data on food and nonfood prices collected by the General Statistical Office confirm this observation. The price in- dex for medicines was virtually flat between 1993 and 1998, in contrast to the overall CPI 52 and the food price index, Figure 3.12: Food, nonfood and medicine price indices, 1992-98 which grew at annual rates of 8 percent and 10.8 per- 180 P Indv cent, respectively, during 160 the same period (Figure 3.12). Consequently, the c M 0 real price of medicines de- clined by more than 30 120 percent over the period. Medne Pr Indcx What would explain this dramatic decline in %c,,nz Pn,c Indc% OaMc tL P1 medicine prices? Certainly, 0 deregulation of the market 60 for pharmaceuticals, which began in 1989, must be an % '0 important reason for the price decline.28 Not only was the volume of drug imports and domestic drug production significantly greater in 1998 than in 1993, but the distribution system for drugs was much more competitive. Likewise, the assistance of international donors in improving the sup- ply of drugs cannot also be discounted. Figure 3.13 shows that there was a large increase in donor spending on drugs in 1997 and 1998 relative to earlier years. While less than 2 percent of donor health spending in 1995 was devoted to drugs, this increased to 18 per- cent in 1997 and 15 percent in 1998. Anecdotal evidence shows that, perhaps as a result of these measures, the availability of drugs is much improved today relative to 1993, even at the smallest commune health centers in the country. 2. Health Care Costs in 5,000 Figure 3.13: Donor spending on drug supply, 1995-98 2.~ -elt ----e ------in 18 Relation to Discretionary In- .. ... , 50.000 -- -- -- - - -- - - - 1 come _1_ * - 45,000 14__ __ _ _ _ A decline in out-of- . - 40,000 - - - - - - -- -- pocket expenditure does not E 12 necessarily mean that private 35. ------- ----- - ------ --o health care costs are no longer a 30000 -- - ---- -- - 10 major financial burden on 25000 - - - - - ----- households. To assess the af- ,0 6 fordability of health services, it 2.,04 is important to look at health 15,0 care costs in relation to income. I0000 - 2 Table 3.11 shows the ratio of eco - - - 0 average health expenditure per 1995 1996 1997 1998 28 There is a fuller discussion of drug regulation and drug supply in chapter V.E. 53 service contact to annual household nonfood expenditure per capita, with the latter serv- ing as a proxy for discretionary household income.29 The cost/nonfood ratio is calculated using both quintile-specific out-of-pocket costs (which reflect the quality of services pur- chased) and national (quality-constant) costs in the numerator. Obviously, the latter will tend to be larger than the former for the poor (with the opposite being true for the better- off), as the poor typically purchase lower-quality services than the better-off. For exam- ple, the quintile-specific affordability ratio of 22 percent for public hospitals for the bot- tom quintile is calculated by dividing the bottom quintile's average public hospital cost of VND 82,500 per visit by its average annual nonfood expenditure per capita of VND 375,000. The corresponding quality-constant affordability ratio of 45 percent is calcu- lated by dividing the national average public hospital cost of VND 168,570 by the bottom quintile's nonfood expenditure per capita of VND 375,000. Table 3.11: Affordability ratios for health costs at public facilities (per-contact out-of-pocket expendi- rure on fees and drugs as %/ of annual nonfood consumption expenditure per capita). by pro% ider and b per capita expenditure quintile, 1993 and 1998 Qurntile-speci/ic equake. arrable; cos 1993 Per capita expenditure gantile Tra! P. r cai a1 Lpendaujifur d qT.;nu,: ProLider Bottom Second Third Fourth Top Bwtvm Se. onJ ThirJ Fourpi ToJ* Public hospi- tals 41,36 21.49 32.31 19.93 8.29 18.72 22,01 21,80 17.57 12.18 4.59 10.90 Commune health cen- ters 11.50 11.89 8.85 6.48 3.13 5.59 3.81 3.06 1.99 1.41 0.87 1.29 Polyclinics 5.53 5.35 4.42 1.64 0.98 2.12 Other public 19.55 17.53 12.68 7.08 2,27 6.83 6.77 4.92 1.21 1,27 1.09 1.80 National (quality-constanl cost Provider 1993 1998 Public hospi- tals 65.16 40.56 27.38 18.05 7.40 18.72 44.97 25.60 17.52 11.03 4.00 10.90 Commune health cen- ters 19.46 12.11 8.17 5.39 2.21 5.59 5.32 3.03 2.07 1.30 0.47 1 29 Polyclinics 8.73 4.97 3.40 2.14 0.78 2.12 Other public 23.79 14.81 9.99 6.59 2.70 6.83 7.41 4.22 2.89 1.82 0.66 1.80 Memo: Annual non- food cons. exp. per cap- ita 210 339 502 761 1,857 734 375 658 962 1,528 4,210 1,546 Source: VLSS 1993 and 1998. There are three important observations to be made from Table 3.11. First, not sur- prisingly from the discussion in the previous section, both the quality-variable and qual- ity-constant affordability ratios of health costs have fallen appreciably for all public pro- 29 It is assumed here that food expenditures, especially of the poor, are committed expenditures over which they have little discretion. Health care is thus largely financed out of nonfood expenditure. From a public finance perspective, this would be tantamount to saying that government health expenditures should be financed out of taxes on nonfood. 54 viders and all income groups between 1993 and 1998. For example, while a single hospi- tal contact constituted 41 percent of nonfood consumption expenditure per capita for the bottom quintile in 1993, the corresponding ratio was only 22 percent in 1998. The cost of a commune health center contact in relation to discretionary income for the bottom quin- tile fell even more-from 12 percent of nonfood expenditure per capita in 1993 to 4 per- cent in 1998. Second, despite the declines, health care costs at public hospitals are large in ab- solute terms for the poor. Even a single service contact with a public hospital takes up 22 percent of all nonfood expenditure for a year for a typical person in the bottom quintile. A service contact of standard (national average) quality would be even less affordable; it would constitute 45 percent of annual nonfood expenditure for an individual in the bot- tom quintile. Third, health care costs at public facilities are significantly less affordable for the poor than for the better-off. The cost of a public hospital contact constitutes 22 percent of nonfood consumption expenditure per capita for the bottom quintile, but only 5 percent for the top quintile. For services of identical quality, the bottom quintile pays 45 percent of its nonfood expenditure per capita for a single service contact with a public hospital, while the top quintile pays only 4 percent. The difference in affordability ratios across the bottom and top quintiles is equally large even at other public facilities and commune health centers-facilities that are intended primarily for the poor. These large differences indicate that the relative burden of user charges falls disproportionately on the poor in Vietnam. The evidence also suggests that the price discrimination currently practiced in public hospitals does not discriminate enough between the poor and the better-off. 3. Budget Share of Health Expenditure Although average out-of-pocket expenditure per contact may be higher for the poor relative to their nonfood consumption expenditure, the overall share of health in to- tal consumption expenditure could well be lower for the poor owing to their fewer num- ber of health contacts. Figure 3.14, which shows the share of health in both total expenditure and non- food expenditure for each consumption quintile, indicates that overall 11 percent of total consumption expenditure and 24 percent of nonfood expenditure were devoted to medical care in 1998-up 55 fro 7 ercnt nd 4 prcet,Figure 3.14: Share OhealthIn total consumiption and nonfood from 7 percent and 24 percent, epniue MadI respectively, in 1993. Although 3 - exp. there is no strong pattern to the 2 27 26 n49S-toalexp. share of medical care in total 25 *199-nonfoodexp. expenditure across income groups, there is a clear pattern of 130 medical care constituting a much 2 15 1 larger percentage of nonfood s 2 expenditure for the poor than for a 0 9 the better-off. For example, while the top quintile devoted 15 percent of their annual nonfood expenditure to health care in r 1998, the bottom quintile spent Per capita consumption cxpenditurc quintile 30 percent. This again highlights the fact that health care costs generally impose a larger burden on the poor than on the nonpoor. 4. Regional Variations The same affordability ratios calculated earlier for the entire country are reported in Table 3.12 for each of the seven regions. Large variations are observed in the afforda- bility of both public hospitals and commune health centers across regions, with public health services generally being most expensive (in relation to nonfood expenditure or dis- cretionary income) in the Central Highlands and the Northern Uplands and least expen- sive in the Southeast region. For instance, a single service contact with a public hospital costs an average of 17 percent of annual nonfood consumption expenditure in the Central Highlands, but only 5 percent in the Southeast. 56 Table 3.12: Affordabillty ratios for health costs at public facilities (per-contact out-of-pocket expenditure on fees and drugs as % of annual nonfood consumption expenditure per capita). by region and per capita expenditure quindle, 1998 Per capita expenditure quintile Region Bottom Second Third Fourth Top Total Public hospitals Northern Uplands . 24.25 13.71 3.60 7.87 11.69 Red River Delta 16.35 35.56 12.12 11.20 391 8.24 North Central 33.64 11.77 19.52 6.48 1.95 8.71 Central Coast . 12.56 29.08 5 29 2.92 6.87 Central Highlands 83.65 10.88 25.56 5.16 608 16.65 Southeast . 1654 15.43 14.46 3.58 5.33 Mekong River Delta 42.50 21.89 13.37 25.46 5 71 14.54 Vietnam 22.01 21.80 17 57 12.18 4.59 10.90 Commune health centers Northern Uplands 4.08 2.63 097 1.55 1.79 1 74 Red River Delta 3.17 2.83 1.90 1.18 0.91 1.05 North Central 7.44 6.82 2.37 2.00 1 64 2.55 Central Coast 1.17 1.02 1.30 1.13 0.42 0.56 Central Highlands 3.96 4.12 2 10 6.90 . 3.27 Southeast 3.14 4.36 0.36 0.97 0.48 0.57 Mekong River Delta 4.48 226 2.74 1.10 0.30 1.21 Vietnam 3.81 3.06 1.99 1.41 0.87 1.29 Note. Affordability ratios are calculated using quintile-specific costs (and thus do not control for quality). Source. VLSS 1998. The regional variations in affordability are even greater for individuals in the bot- tom quintile. For example, a person belonging to the bottom quintile in the Central High- lands spends an average of 84 percent of his/her annual nonfood consumption expendi- ture on a single contact with a public hospital. In contrast, the corresponding ratio for a person belonging to the bottom quintile in the Red River Delta is only 16 percent. Ensor and San (1996), in their study of 32 communes in four rural districts in Northern Vietnam, found that commune health centers in poorer communes charged more (in absolute terms) for their services than those in richer communes. This was the case because the commune health centers in the richer communes were better funded through community contributions and therefore needed to charge individuals less for their use. What all of this suggests is that the poor living in richer areas and in areas well served by public health facilities generally pay less for health care (in relation to their discretionary income). On the other hand, the poor living in the poorest areas and those underserved by public health facilities face a much greater relative burden of health ex- penses. Therefore, it may be necessary to place the highest priority for public assistance to the poor in the poorest areas. 5. International Comparisons How do the private costs of health care compare to those in other developing countries? Unfortunately, there are few studies that show out-of-pocket expenditures on health services contacts in relation to income or nonfood expenditure, so it is difficult to make comparisons. However, there are comparable data from a neighboring country, 57 Cambodia, that can be used to examine the relative burden of health care costs in Viet- nam. These data (shown in Annex Table A3.4) show that affordability ratios for health services in Vietnam are significantly lower than those in Cambodia for every type of pro- vider. For example, while out-of-pocket expenditure on a service contact with a com- mune health center in Vietnam in 1998 cost the equivalent of 4 percent of annual nonfood consumption expenditure for a person in the bottom quintile, the comparable figure for Cambodia was 33 percent in 1997. Of course, there are major differences in the health care systems of Cambodia and Vietnam. In comparison to Vietnam, Cambodia has virtu- ally no rural health infrastructure, very few commune health centers, and even fewer doc- tors and professional health workers. This makes delivery of health services much more expensive in Cambodia than in Vietnam. 6. Private Costs of Hospitalization The VLSS 1998 unfortunately does not distinguish between inpatient and outpa- tient health services contacts in the four-week recall data. However, information on hos- pitalization during the 12 months preceding the interview was collected separately in the survey. These data can be used to analyze the average cost of a hospitalization for differ- ent economic groups in the country. Figure3.15; Out-of-pocket expenditure per hospital admission (in '0010 1998 VND). 1993 and 1998 2000 11,907 1750 i On average, a hospital admission cost VND 862,000 in 1998-up almost 76 per " cent from VND 491,000 in I 1.038 1993. This is a major ex- -=9 pense-approximately i 600 587 US$66-constituting nearly 50 4 30 percent of annual con- 22 217 sumption expenditure per capita. ca ia Poet Second "nird Fourth Richest There are large differ- Pot capta consumption expenditure qui mie ences across consumption in toal ou-of-F igure 3.16: Out-of-pocket expenditure on hospitulizeatiun p er night spent i n quintiles in total out-of-'0001998 VN), 1993 and 199 pocket expenditures on a hos- 10 02 pital admission, especially in 1998, with the top quintile 82 spending nearly nine times as ii19 much as the bottom quintile 54 (Figure 3.15). In contrast, the ratio of hospital expenditure of the top quintile to that of 29 the bottom quintile in 1993 was only about 2.5. Some of the interquin- Poorest Second Third Fourth Richem tile variation in hospital Per capita consumption xpondisc quintise 58 spending comes from quintile differences in average length of hospital stay. As noted ear- lier in Section III.A.6 (Table 3.6), the average length of hospital stay for is nearly two times as long for a person in the top quintile as for a person in the bottom quintile. How- ever, as Figure 3.16 demonstrates, the quintile differences in total hospital spending are also the result of much greater out-of-pocket hospital expenditures per inpatient day by the higher consumption quintiles relative to the lower quintiles. 7. Health Care Costs: A Qualitative Picture Although the preceding discussion indicates that health care costs are substantial for the poor in Vietnam, it does not provide a full picture of how debilitating unantici- pated medical care costs can be for the poor. The peculiar feature of medical costs is that they are, for the most part, unanticipated. This makes them unlike other items of expendi- ture, such as food or children's schooling, for which households can plan in advance. Most illnesses (with the exception of chronic ailments) and injuries are sudden and un- foreseen, which means that the large expenditures incurred in treating them often have to be financed extraordinarily through such means as borrowing or distress sale of assets. A number of qualitative and quantitative studies confirm this fact. A study by En- sor and San (1996) found that about 5 percent of households in their sample of 32 com- munes in Northern Vietnam reporting an illness episode indicated that they were deterred from obtaining treatment be- Figure 3.17: Household financing of treatment expenditures, 32 communes in cause it was too expensive. 9 Northern Vietnar, 1995 For the bottom quartile of I 70 households, this figure was 10 percent. Interestingly, the proportion of the poor de- terred from obtaining treat- 32 1P" U Rich ment was significantly higher in the poorest communes 06 than in the richest communes 3 (13 percent versus 3 percent), in part because commune health centers in poorer communes had to charge higher prices to patients. Ensor and San also found that, while the poor were willing to pay for treatment (often they had no choice), this did not mean that they had the ability to pay. Unlike the better-off, who primarily fi- nanced their out-of-pocket expenditures on health out of their savings, the poor had to reduce spending on food, borrow from others (friends, relatives and moneylenders), or sell livestock to pay for treatment (Figure 3.17). This means that while the poor may be able to pay for current treatment, the large health costs compromise their ability to pay for unanticipated health expenses in the future. Some qualitative studies also indicate that the full cost of hospitalization are con- siderably greater than hospital user fees and drug costs. When a poor person from a vil- lage has to be hospitalized, it generally means that he or she has to be accompanied by 59 two other family members to the district or provincial hospital in town. The need for two family helpers arises because one of them has to be available near the patient's bedside round the clock and the other has to be available for fetching prescribed medicines (in- cluding supplies, such as saline drips) and food. One of the two companions or relatives have to cook meals for the patient and for the two companions in the hospital compound. The costs of boarding for all three patients can be sizeable for a hospital stay of 8-10 days.30 Indeed, the full cost of hospitalization for a serious illness can wipe out years, if not a lifetime, of savings for a middle-income family, and put a poor family in debt for a long time. 30 Lodging is generally free as the relatives typically sleep in the same ward as the patient or on the floor in a corridor in the hospital. 60 CHAPTER III ANNEX TABLES Table A3.1: Distribution of health services contacts at public and private health facilities, 1993 and 1998 Per capita expenditure quintile Providers Bottom Second Third Fourth Top Total 1993 Public hospitals 9.84 13.64 18.83 22.11 35.58 100.00 Commune health centers 25.43 20.52 23.70 17.92 12.43 100.00 Other public providers 12.50 15.62 14.06 20.31 37.50 100.00 Private clinics, doctors 14.02 15.93 23.82 19.50 26.72 100.00 Traditional healers 11.54 26.92 13.46 19.23 28.85 100.00 Drug vendors 18.92 22.45 19.46 20.86 18.31 100.00 Other providers 0.00 9.09 9.09 27.27 54.55 100.00 All providers 17.32 20.18 20.36 20.54 21.59 100.00 1998 Public hospitals 8.21 13.04 17.39 25.25 36.11 100.00 Commune health centers 20.25 24.99 27.29 20.64 6.83 100.00 Other public providers 12.88 14.99 19.32 26.46 26.35 100.00 Private clinics, doctors 13.36 17.26 22.43 19.19 27.75 100.00 Traditional healers 13.97 19.76 15.07 18.83 32.37 100.00 Drug vendors 16.07 21.54 22.26 20.15 19.97 100.00 Other providers 0.00 100.02 0.00 0.00 0.00 100.02 All providers 15.23 20.30 21.96 20.42 22.09 100.00 Source: VLSS 1993 and VLSS 1998 data. 61 Table A3.2: Cumulative distribution of health services contacts at public and private health facilities, 1993 and 1998 Per capita expenditure quintile Providers Bottom Second Third Fourth Top 1993 Public hospitals 9.84 23.48 42.31 64.42 100.00 Commune health centers 25.43 45.95 69.65 87.57 100.00 Other public providers 12.5 28.12 42.18 62.49 100.00 Private clinics, doctors 14.02 29.95 53.77 73.27 100.00 Traditional healers 11.54 38.46 51.92 71.15 100.00 Drug vendors 18.92 41.37 60.83 81.69 100.00 Other providers 0 9.09 18.18 45.45 100.00 All providers 17.32 37.5 57.86 78.4 100.00 Memo: Aggregate cons. exp. 8.43 20.74 36.73 58.30 100.00 Aggregate population 20.00 40.00 60.00 80.00 100.00 1998 Public hospitals 8.21 21.25 38.64 63.89 100.00 Commune health centers 20.25 45.24 72.53 93.17 100.00 Other public providers 12.88 27.87 47.19 73.65 100.00 Private clinics, doctors 13.36 30.62 53.05 72.24 100.00 Traditional healers 13.97 33.73 48.8 67.63 100.00 Drug vendors 16.07 37.61 59.87 80.02 100.00 Other providers 0 100.02 100.02 100.02 100.00 All providers 15.23 35.53 57.49 77.91 100.00 Memo: Aggregate cons. exp. 7.72 19.25 34.33 55.30 100.00 Aggregate population 20.00 40.00 60.00 80.00 100.00 Source: VLSS 1993 and VLSS 1998 data. 62 Table A3.3: Immunization coverage among children under 1 year of age, 1997 DPT (diph- theria, per- BCG (tuber- tussis, and Fully vacci- No Provinces culosis) Polio tetanus) Measles nated Whole country 96.4 95.1 94.9 96.0 95.4 Northern Uplands 94.5 94.4 93.8 93.8 92.7 1 HA GIANG 91.3 96.5 94.9 94.0 89.1 2 TUYEN QUANG 91.2 91.2 91.2 91.3 91.0 3 CAO BANG 97.5 94.5 94.2 91.9 90.2 4 LANG SON 93.0 98.0 94.0 95.0 92.5 5 LAI CHAU 94.7 89.8 88.8 92.9 87.7 6 LAO CAI 96.8 96.0 94.8 94.4 93.8 7 YEN BAI 88.2 88.5 88.2 88.2 88.2 8 THAINGUYEN 91.2 91.0 91.0 91.0 90.9 9 BAC KAN 88.0 88.0 86.5 86.8 86.3 10 SON LA 94.8 90.0 89.3 88,8 88.4 11 HOA BINH 93.2 93.6 93.5 93.3 92.2 12 PHU THO 99.5 99.1 98.8 98.8 98.6 13 VINH PHUC 98.4 98.5 98.3 98.3 98.3 14 BAC GIANG 96.0 96.0 96.1 96.1 96.0 15 BAC NINH 96.5 96.3 96.3 95.9 95.9 16 QUANG NINH 93.7 97.9 97.7 96.7 93.7 Red River Delta 99.4 99.1 99.1 99.0 98.9 17 HA NOI 100.0 100.0 100.0 100,0 100.0 18 HAl PHONG 100.0 100.0 100.0 100.0 100.0 19 HA TAY 98.8 99.0 99.0 98.8 98.8 20 HAIDUONG 100.0 100.0 100.0 100.0 100.0 21 HUNG YEN 100.0 99.9 99.9 99.9 99.9 22 THAI BINH 100.0 100.0 100.0 100.0 100.0 23 NAM DINH 98.5 96.3 96.3 96.0 95.5 24 HA NAM 99.3 98.7 98.6 98.6 98.6 25 NINH BINH 98.1 98.1 98.1 98.1 98.1 North Central 96.5 96.5 96.4 98.5 96.8 26 THANH HOA 99.2 99.2 99.2 99.2 99.2 27 NGHEAN 93.7 93.1 92.4 98.7 93.1 28 HA TINH 99.6 99.5 99.5 99.7 99.5 29 QUANG BINH 91.7 93.4 93.4 99.1 99.1 30 QUANG TRI 88.7 89.0 89.1 90.9 88.0 31 THUA THIEN HUE 99.6 99.5 99.5 98.6 98.6 Central Coast 95.8 95.6 95.1 97.4 97.2 32 TP DA NANG 93.7 94.9 94.9 100.0 100.0 33 QUANG NAM 95.3 97.3 94.5 100.0 100.0 34 QUANG NGAI 92.8 93.0 93.3 98.6 97.7 35 BINH DINH 96.4 96.9 96.9 98.5 98.5 36 PHU YEN 96.9 96.3 95.8 95.8 95.3 37 KHANH HOA 96.8 96.0 96.0 95.7 94.9 38 NINH THUAN 100.0 90.2 89.4 91.4 90.5 39 BINH THUAN 95.8 97.0 96.9 96.7 97.8 Central Highland 97.3 94.8 94.7 96.2 94.5 40 GIA LAI 97.2 93.3 93.3 95.4 93.2 41 CON TUM 97.2 90.3 90.4 88.3 87.1 42 DAC LAC 99.0 96.7 96.5 96.1 95.2 43 LAM DONG 94.1 94.3 94.3 100.0 97.1 63 Table A3.3: Immunization coverage among children under 1 year of age, 1997 (cont.) DPT (diph- theria, per- BCG (tuber- tussis, and Fully vacci- No Provinces culosis) Polio tetanus) Measles nated Southeast 93.3 88.5 89.2 91.0 90.8 44 HO CHI MINH 92.9 87.3 87.3 85.6 85.2 45 BINH DUONG 100.0 86.0 85.6 94.3 93.5 46 BINH PHUOC 84.4 84.3 93.5 95.4 94.9 47 TAY NINH 82.1 82.0 81.7 88.9 91.6 48 DONG NAI 97.7 92.5 92.5 97.5 97.0 49 BA RIA-VUNG TAU 97.7 96.2 96.2 95.1 94.8 Mekong River Delta 97.4 95.3 95.0 96.2 96.0 50 LONG AN 100.0 94.2 94.2 94.8 94.2 51 DONG THAP 93.5 91.9 91.9 98.6 98.3 52 AN GIANG 96.3 96.3 96.3 100.0 100.0 53 TIEN GIANG 99.2 99.8 99.8 100.0 100.0 54 BEN TRE 94.4 88.2 88.2 86.9 86.7 55 VINH LONG 96.5 90.1 90.1 92.2 92.2 56 TRA VINH 92.5 84.6 84.6 90.5 90.5 57 CAN THO 99.8 98.4 98.4 98.7 98.6 58 SOC TRANG 100.0 95.5 95.5 95.6 94.9 59 KIEN GIANG 100.0 98.1 98.3 100.0 100.0 60 BAC LIEU 93.4 100.0 94.0 92.7 91.5 61 CA MAU 100.0 100.0 100.0 92.9 93.6 Source: MoH 1997. Table A3.4: Affordability ratios for health services, Cambodia (out-of-pocket expenditure per contact as % of nonfood consumption expenditure per capita), 1997 Per capita expenditure quintile Provider Bottom Second Third Fourth Top Total Quintile average cost All public providers 41.98 62.47 46.02 34.92 19.56 33.55 Hospital inpatient 108.79 113.46 144.57 121.06 38.72 88.20 Hospital outpatient 36.42 49.12 41.89 25.11 18.34 33.49 Commune clinic/health center inpatient 53.44 445.03 58.71 81.40 17.99 66.66 Commune clinic/health center outpatient 33.49 35.93 23.11 17.45 9.02 16.05 Other inpatient . . . Other outpatient 43.08 9.27 134.85 2.77 10.72 17.84 Cambodia average cost All public providers 121.43 78.41 53.92 36.02 12.23 33.55 Hospital inpatient 319.27 206.17 141.78 94.72 32.15 88.21 Hospital outpatient 121.23 78.28 53.84 35.97 12.21 33.49 Commune clinic/health center inpatient24 1.30 155.82 107.15 71.59 24.30 66.66 Commune clinic/health center outpatient 58.08 37.51 25.79 17.23 5.85 16.05 Other inpatient 829.00 535.32 368.14 245.94 83.48 229.03 Other outpatient 64.58 41.70 28.68 19.16 6.50 17.84 Memo: Annual nonfood cons. exp. per capita (in 91.19 141.22 205.35 307.39 905.53 330.08 thousands of Riels) Source: World Bank 1999. 64 IV. THE PUBLIC PROVISION OF HEALTH SERVICES A. Structure of the Public Health Care System Health services in Vietnam are organized along a four-tiered pyramid. At the top of the pyramid is the Ministry of Health, which is the main national authority in the health sector and, together with the Provincial, District and Commune People's Commit- tees, formulates and executes the health policy and programs in the country. The Ministry manufactures and distributes pharmaceuticals, is involved (together with the Ministry of Education and Training) in physician training, coordinates medical research, sets prices in private health facilities, and is ultimately responsible for the provision of all preventive and a large part of the curative health services in the country. The Ministry of Health is a rather small ministry for Vietnam, with only about 300 staff at the Ministry offices in Hanoi (this number does not include the staff working in the National Institutes or other Ministry of Health-related facilities) compared to the Ministry of Planning and Investment (MPI), which has about 800 staff. The Ministry consists of 14 staff departments, either administrative, such as personnel or finances, or technical, such as planning, treatment, science and training, or medical equipment. The high-level "Committee on Health Strategies" (chaired by the Minister of Health) and its Secretariat within the Policy Unit constitute a forum for policy develop- ment and analysis as well as for weekly meetings of the leaders within the Ministry of Health (MoH). Long-term policies and strategies for the health sector are expressed in periodic 5- and 10-year plans as well as in specific decrees issued by the government and/or the Ministry of Health. The Minister of Health is supported by a number of Vice-Ministers, each with a portfolio of areas in health for which he is responsible. When the Government of Viet- nam finds it desirable to concentrate attention on a certain area related to health, for ex- ample, family planning, it can also create separate entities, which function outside the Ministry of Health. The National Committee for Population and Family Planning is one example, the chairman of which has a minister's status. This entity has established a structure parallel to the health system with the specific task of providing population and family planning services, all financed from a national budget. The Ministry is assisted in its activities by a number of central specialty institutes, which function as tertiary care referral centers and professional training and medical re- search hubs. Among these are the Institute of Malariology, Parasitology and Entomology (IMPE), National Institute of Tuberculosis and Respiratory Diseases, Institute for Protec- tion of Mothers and Newborn, the Institute for Hygiene and Epidemiology, the National Institute of Nutrition, and a number of other institutes responsible for research, training and patient care in the areas of cancer, pediatrics, otorhinolaryngology, traditional medi- cine, pharmacy, surgery and ophthalmology. At the second tier are the 61 Provincial Health Bureaus, each of which serves a population of 0.25-5 million (with a province having an average population of 1.2 mil- lion persons). A provincial health bureau typically has 4-8 departments within it. The provincial health bureaus have to follow the Ministry of Health policies, but are in fact an 65 organic unit of the provincial local governments under the Provincial People's Commit- tees (PPC). For example, although the Ministry of Health sent a circular in 1999 permit- ting provinces to add health staff according to need, the PPC in many provinces did not allow any change. This reflects the fact that the budgets of the provincial health bureaus are part of the local government budget, and in provinces where total revenues exceed their total expenditures, such as Hanoi and Ho Chi Minh City, the PPCs determine fund- ing for their respective provincial health bureaus (PHBs) and have greater influence on their practices. However, in provinces, such as Son La, where aggregate revenues do not exceed aggregate expenditure, the PHBs are dependent on the national government for a large part of their health expenditures, and therefore are more responsive to Ministry of Health directives. The PPC can also supply funds directly to the district or commune, by- passing the provincial health bureau, which means that the provincial health bureau may have little control, or even information, and cannot always ensure that Ministry of Health policies are implemented at lower levels. Planning of health services and programs is mainly the task of the provincial health bureau, including human resource management planning. The provinces prepare annual and multiyear plans on the basis of reports from the districts, including supervi- sion reports from both provinces and districts. In this, they have a certain amount of free- dom as to how it is done. For example, in the province of Quang Tri, the provincial health bureau has gradually adopted a community-based system of planning for health devel- opment, in which planning is done at village and commune level. These submit their plans to the districts, which try to fit the different plans into an overall district plan, which is submitted to the province for approval. This bottom-up approach, built up with the col- laboration and support of the Medical Committee Netherlands-Vietnam, includes contri- butions from the lower levels as well. Neighboring provinces that have visited Quang Tri and been impressed with the results have begun to introduce the same system in their dis- tricts. Although this approach is a new one, the implementation fits within the require- ments made by the Ministry of Health of the provinces. In each province, there is also at least one general hospital with 200-1,000 beds that typically has all seven departments: internal medicine, obstetrics and gynecology, surgery, pediatrics, infectious diseases, traditional medicine, and an emergency ward. The provincial hospitals are intended to be referral centers only, but the referral system does not always work well in practice. Relatively few of the patients cared for at these hospi- tals are referred from outer communities; most reside in the general vicinity of the hospi- tals. In addition to the general hospital, each province may also have one or more special- ized centers or hospitals (e.g., oncological hospitals, cardiology centers, psychiatric hos- pitals, traditional medicine hospitals or tuberculosis hospitals). At the third tier are the District Health Centers, each of which serves the popula- tion of their respective districts. The district health centers are in charge of health man- agement in the district. Each district health center has anywhere from 2-9 different de- partments. While the district health center, in principle, is in charge of all health activities down to the grass-roots level, in some cases it is only responsible for personnel and sala- nes. In each district, there is a district general hospital, including a laboratory and a post for hygiene, epidemiology and malariology. Typically, a unit for maternal and child 66 health (MCH) care and family planning is attached to the district general hospital. District hospitals are supposed to serve as referral institutions for all intercommunal polyclinics in the district. They also provide training facilities for health staff working in intercom- munal polyclinics and commune health centers (CHCs) in the district. Each district also has (i) brigades of hygiene and epidemiology, commanded by the regional branches of the central (national) specialty institutes, which move around the district providing sup- port to categorical health programs, and (ii) two or more intercommunal polyclinics, which are commune health centers that have been upgraded with selected laboratory and surgical equipment and 4-5 specialist doctors (usually an ophthalmologist, otorhi- nolaryngologist, dentist, internist, and clinical laboratory specialist). At the bottom of the pyramid are the commune health centers.31 Each of the 9,806 commune health centers in Vietnam is responsible for providing primary health care, in- cluding preventive, ambulatory and inpatient services, to between 2,000 and 10,000 per- sons, and for referring complicated cases to upper levels of care. They are expected to implement national health programs, such as MCH and family planning, acute respiratory infection (ARI), Expanded Program of Immunization (EPI), and Communicable Disease Control (CDD), and are generally responsible for the management of all health services at the commune level. A commune health center is supposed to have 3-5 health staff, under the leader- ship of the head of the commune health center, who may be an assistant doctor, or other staff such as a nurse.32 Often, one of the staff is a pharmacist responsible for dispensing drugs as well. Sometimes this team is complemented with an assistant doctor in tradi- tional medicine, a health worker responsible for immunizations and sanitation, and an auxiliary nurse. In exceptional cases, a full physician is also part of the health center staff. Ministry of Health staffing data, shown in Table 4.1, indicate a nationwide average of 4.2 staff members per commune health center, with a range among regions of 3.7 (in the Central Highlands) to 5.2 (in the Southeast). A large proportion (one-half to two-thirds) of the commune health center staff are permanent staff, with most of the others being contract workers. The head of a commune health center is selected by the Commune People's Committee and the district health center director. Mostly, they are local people. All commune health center staff are supposed to work eight hours per day as civil servants. However, as salaries are modest, some have other jobs or engage in private practice, which can lead to absenteeism from their commune health center jobs. During the past few years, the Government has revived and promoted the village health worker strategy of providing a minimum of health care to the inhabitants of the 31 There are also intercommunal polyclinics that belong to a level between district hospitals and commune health centers. Intercommunal polyclinics are located strategically in relation to the commune health cen- ters they serve, with each polyclinic serving about 80,000-90,000 people. 32 A study on the management capacity of heads of commune health centers carried out in 196 communes in 10 provinces around the country found that two-thirds of the heads were male, one-half were between 31 and 40 years of age, and most of the rest were over 40 years. Two- thirds of the commune health center heads were assistant doctors, one-quarter were doctors (often upgraded from assistant doctors by three years' extra training), and the others were nurses, midwives and pharmacists (REF, Hien thesis). 67 more remote areas. Members of the community receive training from the provincial health service, often at the district level, in a number of basic topics intended to enable them to cope with the most common medical needs of the population of the village or hamlet. Village health workers are supposed to mobilize and assist with immunization, antenatal care, and family planning programs, advise about clean water and sanitation, and offer simple treatments to people in remote villages. They are also trained to take ma- lana blood slides for analysis at the district health center. B. Operations of Commune-Level Health Activities 1. Number and Regional Distribution of Commune Health Centers Commune health centers form the first tier of the health care delivery network in Vietnam. One of Vietnam's greatest achievements over the last 30 years is the establish- ment of an extensive network of commune health centers throughout the country. In 1997, there were a total of 9,806 commune health centers in the country. The number of commune health centers has varied little over the last 15 years, which, in the context of a growing population, has meant that the population served by an average commune health center has been increasing. In 1997, a single commune health center served an average population of 7,019 persons-up from an average of 5,394 persons in 1980. Vietnam compares favorably to many other developing countries in the per capita availability of primary health facilities. For instance, in neighboring Cambodia, an aver- age commune health clinic serves more than 15,000 persons, while in Thailand-a coun- try that has considerably higher per capita income than Vietnam-a health center serves an average of 6,762 persons. The location of commune health centers is based on population norms, with re- gions having difficult terrain (such as mountainous areas) being allowed more commune health centers in relation to population than other regions. As a result, there are wide pro- vincial differences in the average number of commune health centers. Annex Table 4.1 shows the distribution of commune health center facilities and beds across provinces in 1997. The same data are shown by region in Table 4.1 below. A typical commune health Table 4.1: Population per tonmune health center. 1997 No of commune health Populaion Population Region centers (000) per CHC Facilines Beds Northern Uplands 2.620 10.101 13,020 4.969 Red river Delta 1,927 8.193 14,748 7,654 North Central 1,743 8.020 10,196 5,850 Central Coast 921 3.404 7,948 8,629 Central Highlands 557 2.365 3,315 5,951 Southeast 727 2.347 9,820 13,507 Mekong Ri%er Delta 1,311 6.753 16,619 12,677 Source MoH 1997a 68 center in the Northern Uplands or the Central Highlands is observed to serve considera- bly fewer persons than one in the Mekong River Delta and the Southeast (5,000-6, 000 persons versus about 13,000 persons), reflecting the very low population densities in the mountainous regions. It is unlikely, however, that the larger per capita supply of commune health cen- ters in the mountainous regions fully compensates residents of those regions for the diffi- culties of geographical access. Not only is population density very low in these regions. The terrain is difficult to traverse and the means of public transportation are very limited. In these areas, walking is often the only transportation option, and it can take several hours to reach the nearest commune health center on foot. 2. Utilization of Commune Health Centers The commune health centers have a dual purpose of being the first points of health service contact for the population as well as being responsible for a number of preventive health activities, such as immunization, ARI, CDD, vitamin A supplementa- tion, and health and nutrition education. As such, it is not possible to directly compare the multipurpose commune health centers with private-for-profit providers who focus exclu- sively on provision of curative medical services. There is some concern, however, that the number of consultations for curative care is quite low in the commune health centers, which contributes to the overutilization of secondary and tertiary facilities (due to the breakdown of the referral system). To as- certain this, facility-level data from the Ministry of Health are used to calculate the aver- age patient consultation workload of a commune health center for each province and each region in the country. Annex Table A4.2 reports the average number of daily consulta- tions per commune health center for each province, based on the assumption that a com- mune health center is open 312 days annually (6 days a week, 52 weeks annually). Table 4.2 below reports the same information by region. The data indicate that, on average, the average commune health center in Vietnam has 12 consultations daily. In some regions, such as the Central Highlands and the North- ern Uplands, mean utilization rates at commune health centers are only 7-8 consultations per day. The Mekong Delta is the only region where daily consultations are relatively large (31 outpatients daily). Annex Table A4.2 indicates that in three-quarters of all prov- inces, the mean number of daily consultations per commune health center is fewer than 15. While there are no norms for the number of daily consultations at a commune health center, the numbers shown in Appendix Table A4.2 and Table 4.2 seem small, given that a typical commune health center has 4-5 health workers. 69 Table 4.2: Mean number of daily service contacts at commune health centers, by region, 1997 Mean no. of Annual number daily consult- Average no. of Mean no. of ofconsultations ations per No. of commu- health workers daily consult- Number at the communal commune nal health per commune ations per Region of CHCs level health center workers health center health worker Northern Uplands 2,620 6,841,126 8.37 10,278 3.92 2.13 Red River Delta 1,927 6,534,871 10.87 8,042 4.17 2.60 North Central 1,743 4,783,642 8.80 6,591 3.78 2.33 Central Coast 921 2,490,741 8.67 4,120 4.47 1.94 Central Highlands 557 1,182,611 6.81 2,067 371 1.83 Southeast 727 2,913,805 12.85 3,791 5.21 2.46 Mekong River Delta 1,311 12,829,885 31.37 6,253 4.77 6.58 Vietnam 9,806 37,576,681 12.28 41,142 4.20 2.93 Source: MoH 1997a. Table 4.2 shows that commune health centers in the poorest regions in the coun- try, such as the Northern Uplands and the Central Highlands, have some of the lowest utilization rates in the country, while more developed regions like the Southeast and Red River Delta have significantly higher utilization rates. This could reflect a number of fac- tors, including the superior quality of commune health centers in the better-off regions relative to the poorer regions, higher living standards of the population in the Southeast and the Red River Delta, and the easier geographical access to commune health centers for residents of these regions. However, the above discussion misses the point that curative consultations are only one of many functions performed by commune health centers. In addition to curative patient consultations, commune health centers and their staff are responsible for imple- menting a number of preventive and primary health activities, including immunization, deliveries, family planning, tuberculosis treatment, vitamin A supplementation, and health and nutrition education.33 Indeed, Vietnam's national health programs, such as MCH and family planning, ARI, EPI, and CDD, owe much of their success to effective grass-roots implementation by commune health centers and commune health workers. If the health prevention and promotion activities of the commune health centers are taken into account, it would be difficult to argue that these grass-roots-level facilities are un- derutilized. 3. Reasons for Low Number of Curative Care Consultations Thus, commune health centers appear to function primarily as implementers of national health programs, but their use as outpatient clinics is relatively low. What could account for these low rates of curative care consultations? Based on the results of qualita- tive studies, it is possible to offer some reasons for the low utilization. n For instance, according to the DHS 1997, 42.6 percent of all contraceptive users in the country listed commune health centers as the source of obtaining their contraceptives. The corresponding proportions for different contraceptive methods were 53 percent for IUD users, 38.6 percent for pill users, and 24.6 percent for all condom users (NCPFP, forthcoming). 70 One important explanation for the low rates of utilization of commune health cen- ters for curative care has to do with the availability of alternative suppliers for curative health services. By all indications, commune health centers were heavily utilized for curative consultations and treatment up until the period when they were the exclusive grass-roots suppliers of preventive and curative health services. After the legalization of private medical practice and private sale of drugs in 1989, the number of alternative pro- viders of health services (including drug outlets) increased dramatically. A Ministry of Health survey found that, even as early as 1996, there were more private than public sec- tor health workers at the commune level in the North and in the South (see chapter V.B). District and provincial hospitals are also an alternative supplier of curative health services. While there were hospitals in the pre-doi moi days, geographical access to them has become a great deal easier in recent years with better road and transport infrastruc- ture. This makes it easier for commune residents, especially the more affluent ones, to bypass commune health centers and obtain treatment directly from district hospitals. This also explains why commune health centers in near proximity to district hospitals have particularly low outpatient consultation rates. In the mountainous and highland provinces, low rates of consultations are also the direct consequence of poor geographical access to commune health centers. In these provinces, population densities are very low, the terrain is difficult to traverse, and the means of transportation are limited. As a result, it can often take several hours-even a day-for people living in remote villages to reach the nearest commune health center. However, it is not practical to resolve this problem by providing more commune health centers in these regions. Often, the only practical way to provide health services to people in these remote areas is through outreach programs implemented via mobile health work- ers. Another problem, which is especially acute in the mountainous and highland provinces, is that of staffing. Although the majority of communes (74 percent) have a midwife or obstetric-pediatric assistant doctor, as required by Ministry of Health policy, the Central Highlands and the Northern Uplands have the largest proportion of com- munes without midwives or obstetric-pediatric assistant doctors of any region in the country (Table 4.3). Another important reason for low consultation/treatment rates at commune health centers is the public perception that, while they may be good at the provision of preven- tive and promotive health services, they offer low quality curative care in comparison with alternative providers. Many of the commune health centers are poorly equipped and often are unable to provide drugs and medicines free of charge. In addition, other than in the large commune health centers, health staff are often not present in the facility during a large part of the time that the facility is supposedly open to the public. (The health staff may be engaged in outreach activities or in their own private medical practice.) 71 Table 43: Percentage of communes without a health center and staffing at commune health centers, by re- gion, 1997 Percentage ofcommunes without a: midwie or ob- commune health stetric-pediatric Region center assistant doctor Vietnam 6.76 25.59 Northern Uplands 9.43 34.78 Red River Delta 0.46 19.65 North Central 6.72 28.56 Central Coast 12.81 19.32 Central Highlands 16.34 43.45 Southeast 4.07 19.42 Mekong River Delta 3.51 11.61 Source: MoH 1997a. The low perceived quality of commune health centers in offering curative health services results in many individuals bypassing the primary-level health facilities and seeking medical care, even for fairly simple ailments, from secondary-level facilities, such as district and provincial hospitals. As a result, these facilities are congested, over- crowded, and overburdened with cases that often could be more effectively and inexpen- sively be treated at primary-level health facilities. 4. Management of Commune Health Activities A study on the management capacity of heads of commune health centers carried out in 196 communes in 10 provinces around the country found that, although most of them had several years of experience and knew something about each topic they were interviewed or observed on, in general the understanding of the role of each area (plan- ning, management, evaluation) in good management was unsatisfactory. For example, only 24 percent planned their budgets and activities on the basis of needs and resources available, while 34 percent just used the previous year's results. Only 9 percent used the above plus the norms provided by the district health center, as they are officially sup- posed to. When asked how they selected priority health problems to address, 68 percent of the commune health center heads said they simply implemented a plan prepared by the district health center. Nearly one-half did no supervision and of those who did, most had no records or reports to show for it. Only 7 percent had good practice with regard to evaluation, and 60 percent were rated as poor. Community involvement varied, and, al- though 77 percent of the communes had a primary health care committee, only 43 percent had regular activities to support primary health care. These were mostly through the Women's Union, and some through the Youth Union, and more than one- half of the ac- tivities consisted of health education. A training course for heads of commune health cen- ters was developed based on the needs identified in this survey, and the Ministry of Health is gradually offering the training to heads of commune health centers around the country to upgrade their knowledge and skills. 72 Of course, these results on the lack of planning by heads of commune health cen- ters may well reflect the fact that the heads do not have much control over their re- sources. If resources to the commune health center come earmarked for specific activities from higher levels, there might not be much for a commune health center head to man- age. Most commune health center heads are ultimately service providers-not manag- ers-and the bulk of their workload consists of implementing the national health pro- grams, such as EPI, family planning, malaria control, etc. In a 1994 report on a study of commune health center functioning in the Red River Delta, four factors were identified that could improve service: greater political and financial commitment by local government and the people's committees, a change in the incentives for commune health center workers away from drugs and towards reaching primary health care targets and improving service responsiveness, stronger mechanisms of accountability of the health station to users, and improved supervision and in-service training of commune health center workers by the district health center. That such in- vestments can improve quality of service and consequently utilization was illustrated by the success of a Save the Children U.K.-supported project in Hai Phong province, in which (in addition to inputs of drugs and equipment) retraining concentrated on ten basic conditions and a basic list of drugs, and in which supervision was a vital aspect of in- service training. 5. Evolution of Primary Health Network The primary health care network in Vietnam has experienced an interesting evolu- tion over the decades. Up until the mid- 1 980s, the agricultural work brigades financed the commune health centers and supplied the production brigade nurses, whose responsibility was to support the work of the medical staff in the commune health center through out- reach (e.g., simple first aid, surveillance of matemal health and family planning, educa- tion and propaganda on health matters, participation in immunization activities, and dis- tribution of oral dehydration solutions). The countrywide network of grass-roots health workers, who were in close and daily contact with the people, and community-financed commune health centers formed the backbone of the national health system in Vietnam. A large part of the credit for Vietnam's achievement of excellent health indicators for its population, even in the face of considerable hardship in the past, must go to this primary health network. With the advent of dci moi, however, the agriculture work brigades began to dis- appear, and with them the production brigade nurse. The commune health centers, which used to depend for their financing from the agricultural work brigades, became dependent on funding from the local People's Committees, which was insufficient to maintain the quantity and quality of services offered previously by the commune health centers. As a result, funds for infrastructure, equipment drugs and training were drastically reduced and the quality of curative health services offered by the commune health centers deteriorated significantly. In 1994, recognizing that the commune health centers were in danger of imminent collapse, the government took over the responsibility of paying for the salaries of ap- proved health workers in the commune health centers out of the state (provincial) health budget. In addition, since 1993, thanks in large part due to higher inflows of overseas de- 73 velopment assistance to the health sector, there has been a large increase in the resources allocated to the national health and family planning programs. Since these programs are implemented at the community level by the commune health centers, the latter have bene- fited considerably (in terms of equipment, training, and supplies) from these additional resources. These two factors have restored some of the vigor that was lost from the com- mune health network in the immediate aftermath of doi moi. C. National Public Health Programs As in other countries, the government is the exclusive provider of preventive health services and communicable disease control programs in Vietnam. There is a wide consensus among observers of Vietnam's health sector that the country's national (cate- gorical) health programs have been very successful in providing preventive health ser- vices to the Vietnamese population and in controlling the spread of communicable dis- eases. There are a total of ten national health programs at this time: malaria control, pre- vention of iodine deficiency, Expanded Program of Immunization, tuberculosis control, leprosy elimination, AIDS prevention, nutritional development, and a hospital equipment upgrade program. Four of the more prominent national health programs are discussed be- low. 1. National Malaria Control Program The Malaria Eradication Program was started in Northern Vietnam as early as 1958. Malaria was then endemic in large parts of the North, mainly the mountainous re- gions of the North and the Central Highlands, with the parasite rate ranging from 15 to 20 percent and even 50 percent in some localities. The main method employed to control the spread of malaria at that time was large-scale residual spraying, which was effective in bringing down the positive slide rate from 5.6 percent in 1957 to as low as 0.28 percent in 1964. After reunification, the program was extended to the South, which also saw a sharp decline in the rate of positive slides between 1975 and 1980. In the early 1980s, however, the program ran into serious technical and opera- tional problems, such as parasitic resistance to antimalarial, behavioral changes of vec- tors, development of resistance to spray chemicals such as DDT, and sharply reduced levels of foreign assistance to the Vietnam health sector. The situation became serious between 1985 and 1991, with deaths due to malaria rising from 1,431 in 1985 to 3,439 in 1989 and reaching a peak of 4,646 deaths in 1991 (MoH 1999b). 74 It was against Figure 4.1: Morbidity and mortality rates from malaria, 1976-97 this backdrop that the National Malaria Con- trol Program was 3.5 launched in 1991. The program has been very C2 successful, and has re- sulted in a dramatic de- 2.0 cline in malarial mor- ... .... 1.5. bidity rates as well as case fatality rates (Fig- ------ 1.0 ure 4.1). Many localities in Vietnam have re- M314ria per 1, malaria Cae ported no malarial out- a 7 Y breaks for 3grass-rootM4 years. One malaria prevention tool that has been popularized by the program is the use of impregnated bed-nets. Vietnam is one of the world's leaders in the use of these mosquito nets, with some 10 million persons throughout the country using these nets to protect themselves from contamination. 2. National Tuberculosis Control Program Although antituberculosis activities were started in the North as early as 1957 and extended to the South after the country's reunification in 1976, it was only since 1986 that the National Tuberculosis Control Program began applying the International TB Control Union's tuberculosis (TB) control model, which has been successful in many de- veloping countries and is in compliance with the recommendations of the WHO. This model has been deployed and fully integrated into the network of health services from the commune level to the district and provincial levels. Thus, TB diagnosis and treatment are carried out by commune and district health workers. The TB control program has very wide population coverage; by 1998, nearly 99 percent of the Vietnamese population in communes and wards were covered by the National Tuberculosis Control Program. In contrast, population coverage was only 23 percent in 1986. In addition to extensive cov- erage, the program has been successful in achieving an 85 percent cure rate among all detected cohorts of smear-positive cases and has thus met the objectives set by the WHO (MoH 1999b). Because of the huge increase in coverage of the National Tuberculosis Control Program, the number of reported TB cases has increased dramatically. For example, the number of BK-expectorated TB cases increased ninefold-from 6,645 to 54,897- between 1986 and 1998 (MoH 1999a), reflecting the fact that many TB cases that went undetected and untreated in the past are being correctly diagnosed and treated today. Vietnam's tuberculosis control program is recognized as one of the most successful in the world. Nevertheless, a number of factors could possibly contribute to a rise in TB inci- dence in the years to come: the aging of the population, increasing levels of air pollution, overcrowding in the urban areas, high levels of cigarette smoking, and the spread of HIV/AIDS. Additionally, the high rate of drug resistance is making tuberculosis control 75 more difficult. The Ministry of Health estimates that TB incidence will possibly rise to as many as 120,000 cases and 3,500 deaths per year by 2010 (MoH 1999a). 3. Expanded Program of Immunization Vietnam began implementing the Expanded Program of Immunization (EPI) on a pilot scale in 1982 and on a nationwide scale in 1985. Since 1993, the country has launched a campaign called the "National Immunization Day," when over 99 percent of children under 5 years of age are given two doses of oral polio vaccine (OPV). The EPI program has seen tremendous growth in immunization coverage since 1985, and cover- age is now estimated in excess of 90 percent of full immunization for children under one year of age against six vaccine-preventable diseases: diphtheria, tetanus, pertussis, po- liomyelitis, measles and tuberculosis. Table 4.4 shows that coverage rates against polio, diphtheria-pertussis-tetanus (DPT) and measles vaccinations more than doubled in the ten years between 1986 and 1996. In 1993, Vietnam made a commitment to the international community on eradicating polio, eliminating neonatal tetanus, and controlling measles by the year 2000, and, at this time, it appears that the country will be able to deliver on those commitments. Table 4.4: Immunization coverage of children under I year of age, 1986-96 Type 1986 1990 1996 Tuberculosis vaccine (BCG) 54.5 89.9 95.4 Polio 44.7 86.5 94.5 DPT 42.6 86.7 94.4 Measles 38.8 86.6 96.0 Fully vaccinated . 87.0 95.1 Source: MoH 1996. The increased immunization coverage has been associated with a dramatic decline in child mortality from vaccine-preventable diseases. For instance, the prevalence rate of polio fell from 2.6 per 100,000 persons in 1986 to 0.6 in 1996 (Table 4.5). By 1996, the prevalence of diphtheria had dropped to a mere 5 percent of its level in 1986. Neonatal tetanus has been virtually eliminated in 591 out of 610 districts in the country (MoH 1999b). Deaths from measles have also declined dramatically during the past 15 years. Table 4.5: Morbidity and mortality rate per 100,000 inhabitants of vaccine- preventable deaths Type Cases per 100,000 pop. Deaths per 100,000 pop. 1985 1990 1996 1985 1990 1996 Diphtheria 3.7 0.77 0.19 0.40 0.09 0.02 Pertussis 68.6 6.18 18.03 0.08 0.04 0.01 Polio 2.6 0.59 0.61 0.15 0.02 0.02 Tetanus 5.0 0.48 0.53 0.40 0.31 0.27 Measles 12.8 13.15 6.86 0.45 0.04 0.01 Source: MoH 1996b. 76 Immunization coverage against tetanus toxoid among women in child-bearing ages is also high, although estimates by the Ministry of Health differ from those provided by household surveys. The Ministry of Health estimates that 90 percent of women in re- productive ages and 80 percent of pregnant women have had at least two doses of tetanus toxoid (MoH 1999b). However, data from the DHS-1997 household survey (NCPFP forthcoming) indicate that only about 55 percent of pregnant women received two or more doses of tetanus toxoid. According to the DHS, approximately 28 percent of all pregnant women in 1994-96 did not receive any tetanus toxoid vaccination whatsoever. The main challenge to the program is expanding immunization coverage in the mountainous and remote border areas, where the difficult geographical terrain, low in- comes, and large ethnic populations all contribute to relatively low rates of vaccination coverage. 4. Leprosy Elimination Program Prior to 1982, Vietnam had one of the highest leprosy morbidity rates in the world. As in other countries, social stigma against leprosy patients was strong, and lep- rosy control was inadequate. In January 1982, the Central Institute of Venero- Dermatology was established, and in October of that year, the Ministry of Health launched a nationwide campaign for elimination of leprosy and application of multi-drug therapy regimen for leprosy patients, as recommended by WHO. Concomitantly, the ven- ero-dermatology network was established and consolidated right from the commune to the central levels. The Program for Elimination of Leprosy has been very successful in Vietnam, with 44,556 patients having been completely cured between 1983 and 1997. The program has achieved the goal of leprosy elimination in 46 of the country's 61 provinces. The lep- rosy prevalence rate has fallen from 6.87 per 10,000 inhabitants in 1983 to 0.61 per 10,000 inhabitants in 1997 (MoH 1999b). The main challenges facing the program in the future are elimination of leprosy in the remaining 15 Central and Southern provinces where leprosy prevalence continue to be high. Not only are these provinces poor, but they are located in remote and mountain- ous areas, are inhabited by ethnic minorities, and have bad roads and poor health service networks. D. Hospitals 1. Distribution and Number Vietnam has an extensive network of 817 public hospitals throughout the coun- try.34 Of these, the vast majority (64 percent) are district hospitals, while another 24 per- cent are provincial hospitals (Table 4.6). About 9 percent of all hospitals are specialty hospitals, while the remaining 91 percent are general (including traditional medicine) hospitals. Specialty hospitals include tuberculosis and leprosy hospitals, psychiatric hos- pitals, and a few high-tech centers for fields, such as cardiology, traumatology and oncol- There are also four private hospitals in the country, most of them very recently established, but informa- tion on these hospitals is not readily available. For more description, see Chapter V. 77 ogy. District hospitals are much smaller in size than provincial and central hospitals (71 beds per facility versus 241 and 479 beds, respectively), so that district hospitals account for only 38 percent of the total hospital beds in the country. As many as 78 hospitals and 4,965 beds belong to agencies and ministries other than the Ministry of Health (such as the police, the army, the railways, Ministry of Posts and Telecommunications, etc.). Table 4.6: Numnber of hospitals by level and specialty, 1997 Average size Facilities Beds (beds/facility) Type/level of hospual Number % of total Number on(of total Central 17 2.08 8.104 8-30 479 General 10 1.22 5.640 5.75 564 Specialty 7 0.86 2,500 2.55 357 Provincial 197 24.11 47,560 48.47 241 General 91 11.14 32,046 32.66 352 Specialty 65 7.96 11,794 12.02 181 Traditional Medicine 41 5.02 3,720 3.79 91 District (all general) 525 64.26 37,452 38.17 71 Non-Mol (all general) 78 9.S5 4,965 5.06 64 All hospitals 817 100.00 98,117 100.00 120 Source: MoH 1997a. There has been some, although not much, expansion Figure 4.2: Persons per hospital bed, selected Asian cuuntries, 1995 in the number of hospitals and hospital beds over time. However, this growth has not kept pace with population, with the result that the per capita availability of hospital beds has fallen sharply over time (Figure 4.2). q3 7" International Com- parisons. However, even after this sharp decline, the per F, Ll capita availability of hospital beds in Vietnam is among the Figure4.3: Numberofhospitil beds per 10,000popufation, 1975.97 highest in the region (Figure 40 4.3). The only countries haV- ing a lower ratio of popula- 36 tion to hospital beds than 3 Vietnam are Taiwan, Hong Kong, the Republic of Korea, Singapore and Sri Lanka, all of which have considerably larger GNP per capita than Vietnam. Indeed, even Ma- laysia, Thailand 22 20 1975 1976 1980 1996 1990 1993 1994 1996 1997 78 and the Philippines have higher population to hospital bed ratios. Prima facie this sug- gests that there may be further scope for reducing the per capita number of hospital beds in Vietnam, primarily by continuing to hold back any expansion in the aggregate supply of hospital beds and maybe even a consolidation in the number of hospitals. Provincial distribution. The provincial distribution of hospitals and hospital beds is shown in Annex Table A4.3, while the regional distribution is shown below in Table 4.7. As in the case of commune health centers, the per capita provision of hospitals and hospital beds also appears, at first glance, to favor the Northern and Central Highlands. However, since these regions have very low population densities, geographical access to hospitals (especially in terms of traveling time) is poor despite the higher per capita availability. Table 4.7: Total number of provincial and district hospitals and hospital beds, by region, 1997 Protincial & Ji srict Rean hoSp1uals Per IN 0f0perion.% fac11d1es beds fac 111ih Y h.Jd \ictnam 722 85,012 0 10 11 24 NorThem Uplands ],1 14,.567 0.14 II 19 Rcd Ri%er Delt. 126 10.861 009 II 43 North Central 91 9,905 4) 00 971 Central Coast 9.779 0 11 12.30 Central Highlands 52 -4.995 0 1, 1i.17 Southeast 69 14.7S5 0 W07 15 06 ickong River Delta 1 IS 14 120 0.07 8 50 Swurce Chinh 1998 2. Utilization Rates In 1996, the Ministry of Health conducted a survey of hospitals to collect infor- mation on hospital activities, revenues, and costs. The survey provided data on one of the largest samples of hospitals ever collected in a developing country, with information on 656 out of 817 public hospitals in Vietnam.35 The sample included four levels of hospi- tals (central, provincial, district and other ministry) and two types of hospitals at the cen- tral and provincial level (general and specialty), as well as a broad range of sizes of hos- pitals as measured by number of beds and number of patients. Table 4.8 shows inpatient statistics for admissions and beds by type and level of hospital. The occupancy rate varies from a low of 88 percent for district hospitals to a high of 103 percent for provincial specialty hospitals. This is consistent with field visits which indicate that district hospitals are relatively less utilized (in terms of their bed ca- pacity) than provincial and central hospitals. The average length of stay is 7.5 days for provincial general hospitals and 5.5 days for district hospitals, as compared to 9.4 or more days for the other categories of hospitals. 35 This survey is known as the Inventory of Hospitals Survey 1996. 79 Table 4.8: Summary statistics on inpatients, by type of hospital, 1996 Tpe of hospital Central Central Provincial Provincial Other min- All hospi- Variable general secialty general specialty District istrv tals Mean number of beds 632 238 374 192 79 94 143 Sample size 10 17 80 77 418 36 638 Total beds 6,320 4,046 29,920 14,784 33,022 3,384 91,234 Occupancy rate (%) 93 94 96 103 88 93 94 Average length of stay (days) 13.3 18.7 7,5 10.9 5.5 9.4 7.4 Source: MoH 1996a. Average length of stay can be an indicator of hospital efficiency, when it is possi- ble to adjust for differences in case mix and severity across hospitals. The longer average lengths of stay for specialty hospitals do not necessarily mean they are less efficient than general hospitals, because more of the beds in specialty hospitals are for infectious dis- ease and psychiatric patients who tend to need long stays. Comparisons with other countries provide some context for the Vietnamese re- sults, despite the absence of data to adjust for case mix and severity. In 1986, the average length of stay in China, which is known to have long lengths of stay, was 25.1 days for central, 17.9 days for provincial, and 13.3 days for district hospitals. In 1985, in Indone- sia, the average lengths of stay were 9.4, 8.7, and 5.9 days for central, provincial, and dis- trict hospitals, respectively (Barnum and Kutzin 1993). 3. Financial Structure of Hospitals36 The financial structure of hospitals in Vietnam has been rapidly transformed by the introduction of user fees and health insurance. Between 1994 and 1996, there were double-digit annual real growth rates for virtually every category of patient revenue in every category of hospital (Table 4.9). The growth rates in patient revenue for this period were impressive, even when taking into consideration that the growth stated from a small base (health insurance began in 1993). These increases in patient revenue occurred at the same time that the real growth rates in the state budget were low or negative. For some categories of hospitals, patient revenue by 1996 was almost as large a source of income as the state budget (Table 4.10). Since the Vietnam Living Standards Surveys (VLSS) data indicate that average out-of-pocket expenditure per public hospital contact increased from VND 137,000 in 1993 to VND 169,000 in 1998 (in real terms)-an implied annual growth rate of 4.3percent (see chapter 3), the extraordinarily high rate of annual growth of patient revenue (42 percent on average) seen in Table 4.9 must have occurred primar- ily via a large increase in the number of hospital contacts. 36 One limitation of the Inventory of Hospitals Survey is that it does not offer any information on informal (or "under-the-table") payments made by patients to hospitals. As a result, the analysis that follows is based on patient revenue as reported by hospital administrative staff. These may be smaller than the actual reve- nue collected by hospitals. However, it is possible that informal payments made by patients to hospital staff remain with the staff and do not find their way into hospital accounts in any case. 80 Table 4.9: Average annual real growth rate in mean revenue per hospital (%), by source and by type/level of hospital, 194-96 L eel and tpe of hospital Central Centra! Proincial Provincial Other Source veneral spec taln general specialt-v District panitr All ivel Outpatient revenue -40 51 T0 23 37 74 32 lnpatent revenue 47 81 41 27 62 10 40 VI 75 56 44 61 53 24 50 Domestic donations 6 n/a 15 49 53 20 31 Foreign donahions -97 54 271) -4 45 2.06" 35 State budget 6 -I I -10 2 9 ) Other 102 53 30 -8 12 35 40 Totalrevenue 39 10 23 3 12 14 14 Note. VHI stands for Vietnam Health Insurance. Source MoH 1996a On the expenditure side, the transformation is reflected in increased expenditures on bonuses for staff and drug and medical supplies. Hospitals retain most of the patient revenue to finance their recurrent costs, such as staff bonuses. Staff bonuses more-or-less doubled each year from 1994 to 1996 for every category of hospital, except district hospi- tals (Table 4.11). At the same time, the growth rates for salaries were negative, so that for some categories of hospitals bonuses were a larger type of expenditure for staff than sala- ries. By 1996, bonuses constituted an average of 30 percent of total staff income across all hospitals, with the proportion being as high as 46-47 percent for central general and provincial specialty hospitals. Table 4.10: Mean revenue per hospital (millions of VND). by source and by typellevel of hospital. 1996 Level and tipe of hospital Central Central Provincial Provncial Other nun- Source general specdaln general specially District ISt- .4l levels Patient revenue 10.804 1.870 3.611 I.95" 314 561 1,133 3 841 (24 67) (37 89) (42 41, (26 57) (35.06) (35.13) Domestic dona- tions 46 1 100 40 19 16 31 (0 17) (002) (I 05) (087) (1.59) k1 05) (097) Foreign donations 1.108 373 1.375 265 68 Q 274 3 98) 14 Q2) 114.43) (5 74) (5 74) (0 55) (849) State budger 12,138 4.513 4.183 2.253 765 958 1.651 (4363) 159 54) (43S8 (48 82) (6459) (61 38) (51 18) Other 3.722 b23 2t,2 100 18 17 137 (13.38', (10 85) (2 75) (2.17) (1.521 (1 06) (4.23) Total re%cnue 27.817 7.581 9.532 4.615 1.184 1,561 3,226 (10000) (100001 (10000) (11000 110000) (10000) 1100.00) Note Figures in parentheses are percentages of total Patient revenue include, fee collections from ouipa- rients and inpatients and health insurance reimbursementz Source MoH 1996a. 81 This has created significant differences in staff compensation levels across eco- nomically better-off areas and poor areas. The disparities in compensation mean that there are much greater economic incentives for a health worker to work in a better-off area, where the revenue from user fees is greater than in a poor area. This is, of course, contrary to stated government policies of providing stronger economic incentives for qualified medical staff to serve in poor remote regions. Furthermore, it must be recognized that in a system where public hospitals in- creasingly have to rely on private financing via direct user fees, there is a risk that public providers tend to act as private for-profit providers by devoting greater attention to more 'profitable' patients instead of providing services based on patient need. This behavior reinforces social inequities in access to and utilization of hospital services. Another trend is that the growth in drug expenditures has also been very large. If these growth rates persist, expenditures for drugs and medical supplies will soon exceed expenditures for staff at district and other ministry hospitals. Table 4.11: Average annual real growth rate in mean expenditure per hospital (%), by source and by type/level of hospital, 1994-96 Level and type of hospital Central Central Provincial Provincial Other minis- Source general specialty general specialty District try All levels Staff remuneration 55 65 41 30 54 22 42 of which: - Salaries 40 51 30 23 37 74 32 - Allowance 47 81 41 27 62 10 40 - Bonus 75 56 44 61 53 24 50 Drugs 6 nla 15 49 53 20 31 Maintenance -97 54 270 -4 45 2,067 85 Other 102 53 39 -8 12 35 40 Total expenditure 39 10 23 3 12 14 14 Source: MoH 1996a. The transformation in financial structure discussed above has not uniformly af- fected all categories of hospitals. The changes have occurred more rapidly in central hos- pitals and more slowly in district hospitals. 4. Hospital Efficiency and Marginal Costs One question that is inevitably raised about hospitals is whether or not they are an efficient way to provide health care. In most countries, hospitals consume the lion's share of health resources; in Vietnam, spending on hospitals accounts for anywhere from one- half to three-quarters of the state health budget (see chapter VIII.C). Many public health experts believe that some of those resources would be better spent on preventive and pri- mary care. Certainly, one way of reducing total spending on hospitals would be by im- proving the efficiency of their scale and scope and thereby reducing unit costs. The question about economies of scale is whether larger hospitals are more or less efficient than smaller ones. On one hand, hospitals require large investments in capital, such as in buildings, equipment and specialized staff, which may make it more efficient 82 to have one large hospital rather than two smaller ones. On the other hand, hospitals are complex organizations to manage, and at some point a smaller hospital may run more smoothly than a larger one. The question about scope is whether or not it is efficient to combine outpatient and inpatient care at the same facility. Physicians often need to see patients on both an inpatient and an outpatient basis. An outpatient who receives diagnostic exams may later need to be admitted to the hospital. An inpatient may later need follow-up care as an out- patient. In some cases, it may be more efficient for physicians to provide both types of care from a single office at the hospital. In other cases, it may be more efficient to reduce the daily flow of a large volume of outpatients at the hospital by having separate facili- ties. Using data on 656 hospitals from the Inventory of Hospitals Survey 1996, cost (expenditure) functions were estimated econometrically to measure economies of scale and scope in hospital services in Vietnam. Detailed results of the exercise are available upon request, while Table 4.12 summarizes the main results of the econometric analysis. Two sets of estimates are shown in Table 4.12, one using the total number of inpatient days as the relevant output indicator of inpatient care and the other using the total number of inpatient admissions as an alternative measure. The marginal cost of an inpatient day is estimated at VND 48,434 in central and provincial hospitals and at VND 13,391 at district hospitals.37 The marginal cost of an inpatient admission is significantly higher-VND 178,765 at central and provincial hos- pitals and VND 117,012 at district hospitals. Finally, the marginal cost of an outpatient visit is estimated to be VND 2,145 at central and provincial hospitals and VND 1,614 at district hospitals.38 Thus, the marginal costs of providing both inpatient and outpatient services at district hospitals are significantly lower-by 25-72 percent-than those at central and provincial hospitals. Note that these marginal costs are actual costs incurred by the hospital in providing health services. They may differ from the costs incurred by patients, as prices or fees charged by hospitals for a service may not reflect the true marginal cost of that service. 38 Note that the marginal cost of an outpatient visit in central and provincial hospitals, as estimated in the inpatient day model (column 1 of Table 42), is probably estimated incorrectly. The estimate is too low to be credible, and is therefore ignored in the discussion here. 83 Table 4.12: Summary of marginal cost, economies of scale, and economies of scope measures for 656 hospitals in Vietnam, 1996 Set Of estimates using: Subsample/Measure Inpatient days Admissions Central and provincial hospitals Marginal cost (in VNDI Inpatient days 48,434 Admissions 178,765 Outpatient care 380* 2,145 Economies of scale Short-run 0.83 1.52 Long-run 0.71 0.76 Economies of scope 0.07 0.36 District hospitals Marginal cost (in VND) Inpatient days 13.39I Admissions 117,012 Outpatient care 1,530 1,614 Economies of scale Short-run 1 86 1.53 Long-run 1 18 1.12 Economics of scope 0.27 0.16 Note: Economies of scale less than one mean that marginal unit costs increase with additional output. while economies of scale more than one mean that marginal unit costs actually decline with output. Source Weaver 1999. How do the estimated marginal costs compare to the user fees charged by hospi- tals for health services? Comparisons are complicated by the fact that the hospital user fee scale in Vietnam is by hospital class (or grade) rather than level. Central hospitals generally fall in classes I and 2, district hospitals are generally in classes 3 and 4, and provincial hospitals are distributed throughout all four classes. Despite these problems, a comparison of estimated marginal costs and user fees is revealing. The fee for a general outpatient exam in class I and class 2 hospitals ranges from VND 1,500 to 3,000, which is similar to the estimated marginal cost of an outpatient visit at central and provincial hospitals (i.e., VND 2,145). The corresponding fee in class 3 and class 4 hospitals ranges from VND 500 to 2,000, which is also similar to the estimated marginal cost of VND 1,600 at district hospitals. In contrast, the fee for an inpatient day in an internal medicine department of a class 1 or class 2 hospital ranges from VND 6,000 to 10,000, which is significantly less than the estimated marginal cost of VND 48,000 for central and provincial hospitals. The daily inpatient fee for a class 3 or class 4 hospital ranges from VND 2,000 to 5,000, which is also less than the estimated marginal cost of VND 13,000. 84 Thus, public hospitals appear to be recovering the marginal cost of providing out- patient care, but inpatient care is very heavily subsidized. The extent of the subsidy, even for patients who are charged the full fee, is large-about 73 percent at district hospitals and 83 percent at provincial and central hospitals. The estimated measures of economies of scale, analyzed at sample means, indi- cates short- run and long-run economies of scale at district hospitals and diseconomies of scale at central and provincial hospitals. For policy purposes, these results imply that there would be gains (in terms of lower costs per patient or patient day) from expanding output, but not necessarily capacity, at district hospitals, especially in the short run. In the long run, similar gains (in terms of lower costs per patient or patient day) would be real- ized by increasing the number of beds at existing district hospitals rather than building additional facilities. At the district level, there may also be justification for combining activities from two smaller hospitals at one larger one, especially in cases where such a combination would not reduce geographic access to the facilities. These results also indi- cate that there would be no efficiency gains from increasing the number of beds at central and provincial hospitals despite their high occupancy levels. Finally, the measures of scope, analyzed at sample means, show that there are economies of scope for district hospitals. These results suggest that there would be effi- ciency gains from combining inpatient and outpatient services at district hospitals. 5. Differences in Performance Among Categories of Hospitals Data from the Inventory of Hospitals Survey 1996 show significant variations in performance measures across categories of hospitals and across regions. For example, there is a more than four-fold difference in the percentage of deliveries that employ cae- sarian sections: 6 in other ministry hospitals and 28 in central specialty hospitals (Table 4.13).39 Likewise, there is a ten- fold difference in the mortality rate among inpatients: 0.2 percent in other ministry hospitals and 2.2 percent in central general hospitals. Statis- tical tests indicate that much of the variation in hospital performance measures is not due to chance, and represents real differences among hospitals. 3 Although these data are not shown in Table 4.13, there are equally large differences in the rate at which lab tests are done. For example, there is a threefold difference in the number of laboratory tests per inpa- tient between two levels: 35 at central general hospitals and 11 at provincial general hospitals. There is a fivefold difference in the same ratio across regions: 15 in the Southeast and 3 in the Central Highland. 85 Table 4.13: Hospital performance indicators across types and levels of hospitals, 1996 Level and type of hospital Central Central Provincial Provincial Other minis- Event general specialty general specialty. District try P-value Caesarian sections as a percentage of deliveries Value 21 28 16 22 9 6 <.001 (Sample size) (5) (1) (69) (6) (262) (13) Deaths per inpatient admitted Value 0,22 0.015 0.010 0.013 0.006 0.002 0.084 (Sample size) (10) (14) (79) (55) (380) (21) Deaths per oeration performed Value 0.14 0.14 0.09 0.61 0.18 0.04 0.06 (Sample size) (10) (7) (63) (20) (286) (19) Source: Weaver 1999. These differences raise the question of what causes them. In North America, dif- ferences across hospitals in performance indicators are correlated with provider variables, such as the supply of facilities and physicians, and community variables, such as socio- economic status and morbidity levels, but not with individual patient variables, such as income or attitudes towards health care (Paul-Shaheen, Clark, and Williams 1987). Other researchers suggest that the differences could be due to practice styles, which means the tendency for providers in a small area who informally communicate on a regular basis to develop a consensus about how to diagnose or treat specific illnesses (Wennberg and Git- tlesohn 1975). Yet another explanation might relate to the hospital referral system. For example, referral hospitals (at the central level) may have more sophisticated diagnostic equipment and may service disproportionately more patients requiring extensive labora- tory tests. The causes of the large differences in hospital performance observed in Table 4.13 require further investigation. To the extent that high rates of laboratory tests or cae- sarian sections represent overutilization of services and raise overall hospital costs, it may be important to encourage hospitals of different types and levels as well as hospitals in different provinces to follow more uniform diagnosis and treatment plans. Differences in physician practice styles could narrow as physicians across hospitals and across regions communicate formally and informally with each other on a regular basis to develop a consensus about how to diagnose or treat specific illnesses. 86 CHAPTER IV ANNEX TABLES Annex Table A4.1: Number of commune health center facilities and beds, by province, 1997 Number of commune Population Population health centers ('000) per CHC No. Provinces & cities Facilities Beds Vietnam 9,806 41,183 75,665 7,716 Northern Uplands 2,620 10,101 13,020 4,969 I HA GIANG 166 540 564 3,399 2 TUYEN QUANG 145 710 680 4,688 3 CAO BANG 122 88 566 4,637 4 LANG SON 149 447 729 4,891 5 LAI CHAU 134 600 556 4,147 6 LAO CAI 178 731 587 3,296 7 YEN BAI 165 933 687 4,161 8 THAI NGUYEN 177 885 996 5,629 9BACKAN 100 500 317 3,168 10 SON LA 193 165 847 4,388 11 HOA BINH 188 752 770 4,098 12 PHU THO 270 1,080 1,284 4,754 13 VINH PHUC 143 572 1,085 7,585 14 BAC GIANG 222 1,070 1,475 6,646 15 BAC NINH 123 592 940 7,639 16 QUANG NINH 145 436 938 6,472 Red River Delta 1,927 8,193 14,748 7,654 17 HA NOI 224 0 2,356 10,520 18 HAI PHONG 198 1,026 1,693 8,548 19 HA TAY 320 1,590 2,367 7,396 20 HAl DUONG 261 1,300 1,717 6,579 21 HUNG YEN 160 800 1,098 6,863 22 THAI BINH 285 1,082 1,853 6,500 23 NAM DINH 225 1,125 1,934 8,596 24 HA NAM 114 570 825 7,236 25 NINH BINH 140 700 906 6,471 North Central 1,743 8,020 10,196 5,850 26 THANH HOA 626 3,150 3,553 5,676 27 NGHE AN 463 2,316 2,874 6,207 28 HA TINH 262 1,310 1,359 5,185 29 QUANG BINH 148 444 806 5,449 30 QUANG TRI 93 400 562 6,046 31 THUA THIEN HUE 151 400 1,042 6,900 Central Coast 921 3,404 7,948 8,629 32 TP DA NANG 47 329 667 14,196 33 QUANG NAM 192 1,371 1,379 7,184 34 QUANG NGAI 169 268 1,234 7,299 35 BINH DINH 147 294 1,478 10,054 36 PHU YEN 72 240 770 10,689 37 KHANH HOA 130 92 994 7,642 38 NINH THUAN 54 260 483 8,952 39 BINH THUAN 110 550 943 8,573 87 Annex Table A4.1: Number of commune health center facilities and beds, by province, 1997 (cont.) Number of commune Population Population health centers ('000) per CHC No. Provinces & cities Facilities Beds Central Highlands 557 2,365 3,315 5,951 40 GIA LAI 161 780 844 5,245 41 CON TUM 76 345 269 3,539 42 DAC LAC 192 768 1,347 7,017 43 LAM DONG 128 472 854 6,673 Southeast 727 2,347 9,820 13,507 44 HO CHI MINH 282 600 4,990 17,694 45 BINH DUONG 77 231 650 8,436 46 BINH PHUOC 64 192 549 8,575 47 TAY NINH 86 430 932 10,836 48 DONG NAI 166 790 1,975 11,900 49 BA RIA-VUNG TAU 52 104 724 13,927 Mekong River Delta 1,311 6,753 16,619 12,677 50 LONG AN 182 825 1,300 7,143 51 DONG THAP 139 655 1,559 11,214 52 AN GIANG 129 1,290 2,056 15,934 53 TIEN GIANG 163 766 1,726 10,590 54 BEN TRE 150 655 1,394 9,293 55 VINH LONG 102 510 1,110 10,882 56 TRA VINH 61 305 1,003 16,448 57 CAN THO 94 356 1,905 20,262 58 SOC TRANG 98 490 1,255 12,801 59 KIEN GIANG 79 445 1,447 18,314 60 BAC LIEU 42 168 784 18,657 61 CA MAU 72 288 1,082 15,029 Source: MoH 1997a. 88 Annex Table A4.2: Annual health service contact rates at commune health centers, by province, 1997 Mean no. of daily No. of annual Mean no. No. of Mean no. service Mean no. service contacts of daily CHC of health contacts No. of No. of ofbeds at the commu- contacts health workers per health Province CHCs CHC beds per CHC nal level per CHC staff per CHC worker Northern Uplands 2,620 10,101 3.86 6,841,126 8.37 10,278 3.92 2.13 Ha Giangl6 6 540 3.25 300,283 5.80 580 3.49 1.66 Tuyen Quang 145 710 4.90 251,216 5.55 441 3.04 1.83 Cao Bang 122 88 0.72 216,532 5.69 570 4.67 1.22 Lang Son 149 447 3.00 228,976 4.93 633 4.25 1.16 Lai Chau 134 600 4.48 337,373 8.07 432 3.22 2.50 Lao Cai 178 731 4.11 281,319 5.07 531 2.98 1.70 Yen Bai 165 933 5.65 394,353 7.66 545 3.30 2.32 Thai Nguyen 177 885 5.00 456,897 8.27 761 4.30 1.92 Bac Kan 100 500 5.00 83,944 2.69 354 3.54 0.76 Son La 193 165 0.85 254,899 4.23 838 4.34 0.97 Hoa Binh 188 752 4.00 314,803 5.37 920 4.89 1.10 Phu Tho 270 1,080 4.00 1,003,323 11.91 1023 3.79 3.14 Vinh Phuc 143 572 4.00 642,325 14.40 608 4.25 3.39 Bac Giang 222 1,070 4.82 968,779 13.99 916 4.13 3.39 Bac Ninh 123 592 4.81 857,089 22.33 501 4.07 5.48 Quang Ninh 145 436 3.01 249,015 5.50 625 4.31 1.28 Red River Delta 1,927 8,193 4.25 6,534,871 10.87 8,042 4.17 2.60 Ha Noi 224 1106 4.94 0.00 HaiPhong 198 1,026 5.18 788,754 12.77 893 4.51 2.83 Ha Tay 320 1,590 4.97 951,658 9.53 1312 4.10 2.32 Hai Duong 261 1,300 4.98 899,354 11.04 920 3.52 3.13 Hung Yen 160 800 5.00 510,553 10.23 623 3.89 2.63 Thai Binh 285 1,082 3.80 1,162,846 13.08 1074 3.77 3.47 Nam Dinh 225 1,125 5.00 1,496,744 21.32 1015 4.51 4.73 Ha Nam 114 570 5.00 414,510 11.65 495 4.34 2.68 Ninh Binh 140 700 5.00 310,452 7.11 604 4.31 1.65 North Cen- tral 1,743 8,020 4.60 4,783,642 8.80 6,591 3.78 2.33 Thanh Hoa 626 3,150 5.03 1,251,215 6.41 2175 3.47 1.84 Nghe An 463 2,316 5.00 498,367 3.45 2001 4.32 0.80 Ha Tinh 262 1,310 5.00 1,184,130 14.49 826 3.15 4.59 Quang Binh 148 444 3.00 1,037,077 22.46 580 3.92 5.73 Quang Tri 93 400 4.30 98,318 3.39 421 4.53 0.75 Thua Thien Hue 151 400 2.65 714,535 15.17 588 3.89 3.89 Central Coast 921 3,404 3.70 2,490,741 8.67 4,120 4.47 1.94 Da Nang 47 329 7.00 314,631 21.46 266 5.66 3.79 Quang Nam 192 1,371 7.14 944,768 15.77 865 4.51 3.50 Quang Ngai 169 268 1.59 246,086 4.67 620 3.67 1.27 Binh Dinh 147 294 2.00 252,476 5.50 705 4.80 1.15 Phu Yen 72 240 3.33 228,966 10.19 378 5.25 1.94 Khanh Hoa 130 92 0.71 312,161 7.70 561 4.32 1.78 Ninh Thuan 54 260 4.81 191,653 11.38 258 4.78 2.38 Binh Thuan 110 550 5.00 0 0.00 467 4.25 0.00 89 Annex Table A4.2: Annual health service contact rates at commune health centers, by province, 1997 (cont.) Mean no. of daily No. of annual Mean no. No. of Mean no. service Mean no. service contacts of daily CHC of health contacts No. of No. of of beds at the commu- contacts health workers per health Province CHCs CHC beds per CHC nal level per CHC staff per CHC worker Central Highlands 557 2,365 4.25 1,182,611 6.81 2,067 3.71 1.83 Gia Lai 161 780 4.84 219,923 4.38 520 3.23 1.36 Con Tum 76 345 4.54 260,561 10.99 243 3.20 3.44 Dac Lac 192 768 4.00 474,602 7.92 794 4.14 1.92 Lam Dong 128 472 3.69 227,525 5.70 510 3.98 1.43 Southeast 727 2,347 3.23 2,913,805 12.85 3,791 5.21 2.46 Ho Chi Minh 282 600 2.13 0 0.00 1626 5.77 0.00 Binh Duong 77 231 3.00 1,300,952 54.15 334 4.34 12.48 Binh Phuoc 64 192 3.00 529,394 26.51 265 4.14 6.40 Tay Ninh 86 430 5.00 597,674 22.27 539 6.27 3.55 Dong Nai 166 790 4.76 246,343 4.76 760 4.58 1.04 Ba Ria Vung Tau 52 104 2.00 239,442 14.76 267 5.13 2.87 Mekong River Delta 1,311 6,753 5.15 12,829,885 31.37 6,253 4.77 6.58 Long An 182 825 4.53 416,084 7.33 596 3.27 2.24 Dong Thap 139 655 4.71 1,496,789 34.51 586 4.22 8.19 An Giang 129 1,290 10.00 5,522,558 137.21 848 6.57 20.87 Tien Giang 163 766 4.70 1,539,037 30.26 697 4.28 7.08 Ben Tre 150 655 4.37 353,496 7.55 712 4.75 1.59 Vinh Long 102 510 5.00 447,942 14.08 492 4.82 2.92 Tra Vinh 61 305 5.00 127,408 6.69 458 7.51 0.89 Can Tho 94 356 3.79 856,378 29.20 475 5.05 5.78 Soc Trang 98 490 5.00 929,499 30.40 455 4.64 6.55 Kien Giang 79 445 5.63 232,162 9.42 435 5.51 1.71 Bac Lieu 42 168 4.00 582,244 44.43 225 5.36 8.29 Ca Mau 72 288 4.00 326,288 14.52 274 3.81 3.82 Vietnam 9,806 41,183 4.20 37,576,681 12.28 41142 4.20 2.93 Source: MoH 1997a. 90 Annex Table A4.3: Total number of provincial and district hospitals and hospital beds, by province, 1997 Provincial & district No. Provinces & cities hospitals Per 10,000 persons No. offa- No. of No. offa- No. of cilities beds cilities beds Vietnam 722 85012 0.10 11.24 Northern Uplands 181 14567 0.14 11.19 1 HA GIANG 10 570 0.18 10.10 2 TUYEN QUANG 10 1010 0.15 14.86 3 CAO BANG 12 720 0.21 12.73 4 LANG SON 13 720 0.18 9.88 5 LAI CHAU 10 560 0.18 10.08 6 LAO CAI 11 560 0.19 9.55 7 YEN BAI 13 670 0.19 9.76 8 THAI NGUYEN 14 1357 0.14 13.62 9 BAC KAN 5 410 0.16 12.94 10 SON LA 12 1460 0.14 17.24 11 HOA BINH 10 645 0.13 8.37 12 PHU THO 16 1245 0.12 9.70 13 VINH PHUC 8 760 0.07 7.01 14 BAC GIANG 13 1520 0.09 10.30 15 BAC NINH 7 870 0.07 9.26 16 QUANG NINH 17 1490 0.18 15.88 Red River Delta 126 16861 0.09 11.43 17 HA NOI 14 2650 0.06 11,25 18 HAIPHONG 19 3196 0.11 18.88 19 HA TAY 16 1990 0.07 8.41 20 HAI DUONG 17 2120 0.10 12.35 21 HUNG YEN 11 1090 0.10 9.93 22 THAI BINH 16 1870 0.09 10.09 23 NAM DINH 16 1935 0.08 10.00 24 HANAM 9 1000 0.11 12.12 25 NINH BINH 8 1010 0.09 11.15 North Central 91 9905 0.09 9.71 26 THANH HOA 32 4250 0.09 11.96 27 NGHEAN 24 2640 0.08 9.19 28 HA TINH 11 1240 0.08 9.13 29 QUANG BINH 7 625 0.09 7.75 30 QUANG TRI 7 580 0.12 10.31 31 THUA THIEN HUE 10 570 0.10 5.47 Central Coast 85 9779 0.11 12.30 32 TP DA NANG 11 1640 0.16 24.58 33 QUANG NAM 15 1400 0.11 10.15 34 QUANG NGAI 14 1331 0.11 10.79 35 BINH DINH 15 1885 0.10 12.75 36 PITU YEN 9 810 0.12 10.52 37 KHANH HOA 8 1033 0.08 10.40 38 NINH THUAN 4 600 0.08 12.41 39 BINH THUAN 9 1080 0.10 11.45 Central Highland 52 4995 0.16 15.07 40 GIA LAI 14 1300 0.17 15.40 41 CON TUM 6 550 0.22 20.45 42 DACLAC 20 1970 0.15 14.62 43 LAM DONG 12 1175 0.14 13.76 91 Annex Table A4.3: Total number of provincial and district hospitals and hospital beds, by province, 1997 (cont.) Provincial & district No. Provinces & cities hospitals Per 10,000 persons No. offa- No. of No. offa- No. of cilities beds cilities beds Southeast Region 69 14785 0.07 15.06 44 HO CHI MINH 27 9815 0.05 19.67 45 BINH DUONG 5 660 0.08 10.16 46 BINH PHUOC 5 340 0.09 6.20 47 TAY NINH 12 1200 0.13 12.88 48 DONG NAl 14 2170 0.07 10.99 49 BA RIA - VUNG TAU 6 600 0.08 8.29 Mekong River Delta 118 14120 0.07 8.50 50 LONG AN 16 1700 0.12 13.08 51 DONG THAP 9 970 0.06 6.22 52 AN GIANG 12 1770 0.06 8.61 53 TIEN GIANG 9 1320 0.05 7.65 54 BEN TRE 10 1080 0.07 7.75 55 VINH LONG 8. 860 0.07 7.75 56 TRA VINH 8 810 0.08 8.07 57 CAN THO 13 1730 0.07 9.08 58 SOC TRANG 7 770 0.06 6.14 59 KIEN GIANG 13 1295 0.09 8.95 60 BAC LIEU 5 770 0.06 9.83 61 CA MAU 8 1045 0.07 9.66 Source: MoH 1997a. 92 V. THE ROLE OF PRIVATE FOR-PROFIT AND NONPROFIT PROVIDERS A. Legal Framework for Private For-Profit Health Services In 1986, following doi moi, the Government of Vietnam began allowing private provision of health services. The official rationale for allowing private provision was that this would meet (excess) demand that the public sector was unable to fulfill, and make fuller use of health workers who had retired from the public sector. The Government au- thorizes the Ministry of Health and the provincial health bureaus to manage private health activities (Baker and McKenzie 1999). Since 1989, the government has implemented numerous laws, ordinances, and decrees which govern the private health sector. Enacted in 1993, the Ordinance on Private Practices of Medicine and Pharmacy and its subsidi- ary regulations delineate specific rules and conditions regarding private health services. For example, the Ordinance lists the allowable private medical practices, stipulates the licensing and approval process, and specifies the professional requirements of the appli- cant. In addition, the Ordinance discusses ethical requirements and offers guidelines for the medical facility. The provincial health bureaus are the health authorities that can issue licenses for private pharmacies. These licenses are to be revoked if the licensed pharmacist is absent during three consecutive inspections or has seriously violated regulations. Licenses for private pharmacies are provided mainly to pharmacists who have a university degree, ex- cept in the case of remote areas, where assistant pharmacists can apply for licenses. The law stipulates that persons applying for a pharmacy license should have 2-5 years of ex- perience in public pharmaceutical facilities and, in the case of establishing a pharmaceu- tical enterprise or company, s/he should not be currently working for a public institution. Pharmacists currently working for public facilities can apply for licenses to operate a pri- vate pharmacy or retail outlet outside their official working hours. Private pharmacies (referred to as drug vendors or drug outlets in Chapter III) are allowed to sell drugs (local or imported) that are permitted by Ministry of Health and to sell medical equipment and supplies, veterinary drugs, and cosmetics. They are not al- lowed to sell pharmaceutical materials, chemicals for laboratories, toxic drugs, addicting drugs, counterfeit drugs, and other goods. They are not allowed to manufacture or pro- duce drugs themselves. Pharmacists can hire drug sellers who must be elementary phar- macists. The pharmacy is expected to clearly post its working time, and a pharmacist is supposed to be present in the pharmacy during its entire working hours. The decree on private practice of medicine allows only retired health personnel to have licenses for full-time private services. Working staff in public health facilities are not permitted to operate privately during working time; they can apply for license for part-time private services, e.g. evening clinics. The applicant for a license to run a private hospital or a private clinic needs to be a full physician who has at least five years of ex- perience in a public health facility. The applicant for a license to run a private maternity clinic needs to be an obstetrician or a midwife. The applicant for a license to run a diag- nostic center is required to be a specialist. Nurses are allowed to work in a team led by a 93 private physician or to apply for a license to run private health services, such as adminis- tering injections and massages. The law gives private facilities the same rights as public facilities in the purchase of medical equipment. Private physicians and facilities have the right to sign contracts with public specialists to assist them in the performance of advanced techniques, to de- velop paraclinical diagnostic laboratories to enhance care quality, and to keep and use some drugs for emergency cases on instruction of health authorities. They have the right to establish joint ventures with foreign investors according to existing government laws and regulations. They have the right to collect user charges based on approved and pub- licly displayed user-fee levels. Like public providers, private facilities are expected to operate according to exist- ing laws and regulations. One of these is that they cannot prescribe and sell drugs at the same time.40 They cannot advertise their services by providing incorrect information to users. Private providers can participate in the implementation of primary health care, health education, and curative and preventive activities. They are also required to respond without preconditions to the calling of health authorities for public service in the case of disasters and epidemics. They are required to report to health authorities cases of com- municable diseases and mass intoxication. Provincial health bureaus are health authorities responsible for issuing licenses to private providers. Provincial health authorities convene expert committees to assist them in the judgment of candidates who have applied for a private license. Regular supervi- sion, through random inspections, of private health facilities is also the responsibility of the provincial health office. Facilities violating laws and regulations can be punished with warnings, temporary suspension, criminal investigation, and license revocation. B. Private Clinics and Providers 1. Classification of Private Health Providers There are broadly two types of private for-profit health providers: (i) full-time providers who own private facilities, and (ii) part-time private providers who are on the staff of public health facilities but work during off-hours in private practice. At present, private health workers are not organized in any federation or professional organization. According to the Decree on Pharmaceutical and Medical Practices, there are ten types of private health facilities, four types of private traditional medicine facilities, and four types of private pharmaceutical facilities. These are: * Hospitals, maternity homes, polyclinics or specialized clinics, dentist's rooms, labora- tories and functional exploration facilities, X-ray rooms, cosmetic surgery centers, re- habilitation and nursing care centers, infirmaries, and family planning clinics. * Traditional medicine hospitals; traditional medicine clinics; traditional medicine re- habilitation and nursing care centers; and traditional medicine acupuncture clinics, massage facilities and sauna units. 40 Traditional healers are exempted from this rule. 94 * Private pharmacies, drug sale agents, pharmaceutical enterprises and limited compa- nies, and herbal medicine sellers. This classification supposes that all private health services are provided by pri- vately owned health facilities. It does not consider the provision of semipublic health care by public providers, which is emerging in some public hospitals in the form of semipub- lic wards. 2. Size and Growth of the Private Sector in Health Services Provision Little is known about the total size of the private sector in the health services market and its growth over the past ten years. In 1996, the Ministry of Health estimated the number of licensed private health personnel in the country to be 25,698-about a tenth of the 213,099 public health personnel. However, not all of these private health per- sonnel can be counted as separate workers, since nearly one-half of them are actually government staff, who work in the private sector during evening and weekend hours. Table 5.1 shows the number of private health personnel by category. Physicians constitute the largest proportion (37 percent) of private health workers, followed by pharmacists or drug vendors (24 percent) and traditional healers (13 percent). Nearly 83 percent of all private physicians are government employees, as compared to 45 percent of private pharmacists and only 2.5 percent of traditional healers. Table 5.1: Number of licensed private health personnel, by category, as of October 1996 Of which governMcnt % govern- No Caer, of Health Personnel Total sralT men[ staf I Ph%sicians 9.399 7.781 82 80 2 Pharmacists or drug vendors 6,146 2,763 45.00 3 Assistant physicians 1,562 643 41.20 4 Assistant pharmacists 2,182 570 26.10 5 Middle level nurses 214 103 48.10 6 Middle level midwifes 303 146 48.20 7 Middle level medical technicians 186 41 22.00 8 Elementary nurses 839 142 16.90 9 Elementary pharmacists 2,540 201 7,90 10 Traditional heaters 3,327 84 250 TOTAL 25,698 12,474 48.54 Source: MoH staff estimates. According to the Ministry of Health, there were a total of 34,018 private pharma- ceutical and medical facilities, including drug outlets and sales agents for pharmaceutical companies, in 1998 (Table 5.2). General practitioners' clinics constituted the largest pro- portion of private health facilities, followed by infirmaries and private pharmacies or drug outlets. There are at present only four private general hospitals in the country, al- though there are five health facilities with foreign investment and collaboration. 95 Table 5.2: Number of private health facilities, by type (as of October 1998) No. Type offaciliy Number Private health facilities I General hospitals 4 2 Maternity homes 264 3 Polyclinics 98 4 Consulting rooms of general practitioners 7,005 5 Specialized clinics 3,423 6 Dental clinics 2,305 7 Lab tests and functional exploration 240 8 X-ray examination 197 9 Cosmetic surgery 42 10 Infirmary 5,569 11 Rehabilitation and nursing care 125 12 Family planning services 550 13 Foreign-invested health facilities 5 Private pharMaceutical facilities 14 Private pharmacies or drug outlets 5,192 15 Sales agents for pharmaceutical companies 8,822 TOTAL 33,850 Source: MoH 1999b. In addition there are also private providers working without license. A mail sur- vey of private health providers sent to provincial health offices in every province in early 1999 yielded new insight into the size of the private sector.41 Unfortunately, only 44 of the 61 provinces contacted responded to the questionnaire, so the estimated magnitude of the private sector by this survey is likely too low. This survey indicated a total of 36,442 private health facilities (29,400 not including traditional healers) in the 44 provinces (Ta- ble 5.3). Approximately, 12.6 percent of private health facilities were unlicensed. Table 5.3: Licensed and unlicensed private health facilities in 44 of 61 provinces, by rural/urban areas, 1999 Licensed privaie heahh L n/censed pri aie health /t- Sector facities cihies Totalprivate health faciltes Number N Number %Number 9 Urban 22.007 86 5 3.445 13.5 25.452 100.0 Rural 9,835 895 1.155 10 5 10,990 1000 Total 31.842 87.4 4.600 126 36.442 100.0 Source. Dung 1999 Table 5.3 indicates that about 70 percent of the private health facilities are in the urban areas, with the remaining 30 percent in the rural areas. Since 80 percent of the population is rural, there are roughly nine times as many private health facilities per cap- ita in the urban areas as in the rural areas. Thus, the supply of private health facilities and health workers tends to reinforce the existing inequalities in the rural-urban distribution of public health facilities and public health workers. 41 The survey was targeted to the units responsible for administration and regulation of private health ser- vices in the provincial health offices. 96 Annex Table A5.1 shows the number of private health facilities enumerated in the 1999 survey by category of facility. Drug commission sellers constitute the largest pro- portion of private facilities (18 percent), followed by general practitioners' clinics (16 percent), traditional medicine clinics (14 percent), private pharmacies (14 percent), and nursing services (12 percent). The combined share of drug commission sellers and private pharmacies or drug outlets is nearly 32 percent. As a comparison, it may be recalled that VLSS 1998 data indicate that individual contacts with drug vendors and private pharmacies constituted nearly two-thirds of all health service contacts that individuals had (chapter 3). 3. Regional and Provincial Distribution There is little information about how many private health workers operate at the commune level. One of the only investigations in this area is a Ministry of Health survey of selected communes in the North, Center and South in 1996-97 (MoH 1998d). This survey indicated that, even as early as 1996, there were slightly more private than public sector health workers at the commune level in the North and the South (Table 5.4). Thus, a typical commune had 4-7 public health workers and 4-7 private health workers (pri- marily assistant doctors and nurses). Of course, it is not clear from the survey whether the public and private health workers were different persons, or whether there was overlap between the two categories of health workers. In the more remote and poorer areas of the central region, there were fewer of both public and private health workers in an average commune. Table 5.4: Average number of public and private sector health workers per commune, by region and worker category. 1995 North Center South Category Public Private Public Private Public Private Doctor 0.3 0.5 0.1 02 0.4 1.2 Assistant doctor 3.3 2.5 2.3 0 7 2.6 1.6 Nurse 1.1 1.7 1.1 1.4 1.1 1.8 Midwife 0.7 0.4 07 0.5 0.9 0.5 Traditional medicine doctor 0.1 0 6 0.1 0.6 0 1 0.5 Others 0.1 0.5 0.1 0.3 0.1 0.9 Total 5.5 6.0 44 37 5.3 6.6 Source. MoH 1998d The 1999 survey data on private health facilities is shown by province in Annex Table A5.2. There are large provincial differences in the distribution of private health fa- cilities. The province of Lai Chau only has 41 private health facilities, while Bac Kan and Ha Giang have fewer than 100 private facilities. On the other hand, Ho Chi Minh City alone has 6,212 private health facilities. Other areas with a large private sector presence in health are Hanoi (having 2,883 private health facilities), An Giang (2,278), and Dong Nai (1,360). Even after adjusting for population, there are wide differences in the avail- ability of private facilities. Lai Chau only has 100 private facilities per million persons, while Ho Chi Minh City and Hanoi have over 1,400. 97 How are these differ- Figure 5.1. Number ofprivate health facilities and GDP per capita across provinces, 1998 ences in the presence of the pri- 1,0 vate sector related to provincial income? Figure 5.1 indicates a 1 -. strong positive relationship, such ,40o that at higher levels of GDP per capita, the number of private health facilities per million I population is mtich greater. This is to be expected, as there is evi- dence from all over the world, as well as from Vietnam, that the demand for private health ser- vices (as expressed in the ability Provincial GDPpercapita(000VND) to pay) rises strongly with household income. Providers respond to this increased demand, and establish private fa- cilities in higher-income provinces and regions. However, the greater demand for private providers at higher incomes may be driven by aspects of health services that appeal to consumers (but do not have any clinical significance), such as air-conditioned waiting and examination rooms, pleasant surrounding, and nonessential amenities. 4. Utilization of Private Health Clinics The only nationally Figure 5.2: Health service contacts wnith private clinics and peivate providers as representative information o re rsnaie ifr ain on percentage of all health service contacts in the last 4 wseeks, by per the utilization of private 30 capita expenditure qaintile, 1993 and 1998 health services comes from Priate tacilites, 1993 the VLSS 1993 and the VLSS wriae faciftm, 1998 24 1998 household surveys. 9 20 aPublichspitls, 1990 1 These data indicate a definite positive relationship between 61 greater utilization of private health services and living standards (Figure 5.2). While 6 health service contacts with private facilities and providers_1 constituted 15 percent of all pcct Seond Thid Fouet P-h.4 health service contacts (during ics . "P."da- qaindpl r a four-week reference period) for the lowest quintile of individuals in 1998, the corre- sponding share was 21 percent for the highest quintile. However, the relationship between per capita consumption expenditure and the utilization of private health providers is not as strong as might be expected. Indeed, the use of public hospitals is associated much more strongly with income than the use of pri- vate health facilities (Figure 5.2). The ratio of public hospital usage between the highest and the lowest quintile is more than 3, while it is only 1.4 in the case of private health providers. What might explain this result? Since private clinics may be regarded as alter- 98 native providers to commune health centers and drug vendors, increased use of health services that accompanies higher incomes tends to be distributed across these alterna- tives. But the lack of alternatives to public hospitals means that higher demand associated with higher income is concentrated on public hospitals and leads to much more intensive use of public hospitals by the better-off Additional support for th A i a y oh sis-viz. that Figure5.3: Average out-of-pocket expenditure per ealth service contact.by per the earlier hypothesis-viz., thatand by provider, 1998 private clinics and commune 25 p health centers are competitive 0 C health centers alternatives-is found in VLSS * Pratinan ealers 1998 data on average out-of- pocket expenditure per service contact for different types of health providers (Figure 5.3). For all quintiles, the out-of- 1 pocket costs of private clinics 0 74 and commune health centers are comparable (with private clinics 21 1 tj costing slightly more). However, public hospitals are in a class of their own, demanding the highest out-of-pocket payments from all quintiles. 5. Why Do Patients Use Private Providers? Located in the northeast of Vietnam, Hai Phong City is the third largest city in the country and the population of the province is over 1.7 million. The public system of the province includes one urban obstetrics and gynecology hospital, 12 district health centers, one general hospital, one military hospital, and 217 commune health stations-all of which provide reproductive health services. One of the earliest clinics opened in 1987 in Hai Phong as part of a public and private cooperation contract between the Vietnam Czech-Slovakia Friendship Hospital. In 1992, this hospital became the first private site to provide reproductive health services. Beginning in 1988, the Provincial Health Services Department registered its first private clinic. Of Hai Phong's 500 registered private pro- viders today, 28 are physicians and 11 midwives or physicians assistants who provide reproductive health services. Currently, the Hai Phong Health Services Department processes 20-30 permits per month, including renewals. The registration permit is granted for five years at a cost of VND 300,000 (US$21.58) per permit and each renewal. Physicians must have a mini- mum of five years of clinical experience in reproductive health (or an applicable spe- cialty) to open a private clinic. The registration permit specifically designates which pro- cedures the clinic is allowed to perform. For example, reproductive health clinics may be allowed to offer all or only a partial list of services, depending on the expertise of the physician. The policy authorizes private reproductive health services to offer antenatal, family planning and contraception, menstrual regulation (MR) or termination of preg- nancy of less than 8 weeks by curettage, treatment for reproductive tract and sexually transmitted infections (RTIs/STDs), fertility treatment, and other minor gynecological 99 complaints. Private clinic providers are not authorized to remove intrauterine devices (IUDs) or terminate any pregnancy over 8 weeks. Once granted a permit, the clinic managers or health care providers must submit quarterly reports. Monitoring and evaluation of the private health clinics is under the di- rection of the Hai Phong Health Services Department. A five-member inspection team consisting of members from the provincial and district health agencies makes both sched- uled and unscheduled on-site visits. The directors of the district health centers are in charge of supervising the private health sector in their respective districts, including pharmacies, clinics, and traditional medicine sites. Health Seeking Practices. Qualitative surveys of patients using private reproduc- tive health services indicated that one of the main reasons for choice of private providers was shorter waiting times. In addition, patients felt that doctors at private clinics had more time to spend with their patients as they did not have as heavy a case load as public doctors. Some consumers perceived the overall quality of private health services to be higher than that of public health services. They feel that private providers were more re- sponsive to their concerns and health needs. Although consumers have to pay extra for the additional (perceived) quality and lower waiting time, the price difference between public and private services is not that great. Many patients in the Hai Phong sample thought that the price of private reproduc- tive health care was affordable and reasonable compared to the price of care at a public health clinic. For example, a MR at private clinics is VND 25,000 plus medications and the VND 5,000 registration fee. At the public clinics, the price of a MR is around VND 15,000 plus medication. Provider Issues. The desire to make money, a perceived unmet demand for ser- vices, and a belief in their experiential efficacy motivate private providers and adminis- trators to establish private practices and clinics. Both working and retired physicians see private practice as an opportunity to increase their income. In the public sector a physi- cian's salary varies from VND 300,000-500,000 per month, depending on experience. On average, doctors who are employed in the private sector can earn VND 400,000- 800,000 dong per month. Both private and public providers define the unmet demand for services as higher-quality care. They defined higher-quality care as having more time to develop a patient/provider relationship. The initial capital investment needed to open a private clinic requires enough money for registration fees, taxes, equipment, supplies, hiring staff, and preparing the facility. Many private providers borrow money from friends and family and procure used equipment and supplies in order to open their private clinics. Standards of Practice. The standards of practice among the private reproductive health providers in the Hai Phong sample were surprisingly not all that varied. However, private sector providers reported that they had to be more accountable to their clients than the public providers. Specifically, they reported that: * The provider/client relationship was very important. The patient returns to the pro- vider who treats him/her well and, from their perception, the care is better. This is 100 true even if the selected provider neither follows the best hygienic technique nor has new and modem equipment. * The private provider is seen as having to be more responsible and accountable for his or her standard of practice. This accountability includes both the client/provider rela- tionship and financial accountability. Unlike providers practicing in the public clinic, private doctors are often forced to pay compensation to the patient and/or family for complications, which occur while the patient is under their care. * Prices for services in private clinics had to be reasonable and not significantly higher than public clinics. C. Private Inpatient Services 1. Private Hospitals in Vietnam The growth of private hospitals has been relatively limited in Vietnam. Ten years after liberalization, there were only four private hospitals in the country. Informal discus- sions with administrators at private hospitals and private physicians suggest a number of factors that together have resulted in limited private sector investment in the hospital sec- tor in Vietnam. Changing Tax Policies. Government policies with respect to private hospitals are frequently changed, with the result that private investors face considerable uncertainty over the regulatory environment. For example, the administrator of a private hospital mentioned that his hospital was initially granted a five-year tax exemption by the gov- ernment. However, after one year, the exemption period was shortened, much to the det- riment of the hospital. The private sector may fear that after having invested large amounts of capital in the establishment of a private hospital, the government might sud- denly change the regulatory or tax framework, putting the large private capital investment at risk. Certainly, this is an important factor contributing to insufficient private invest- ment in the hospital sector. Absence of Clear Policies on Liability. Another factor affecting private invest- ment in the hospital sector has to do with the absence of a clear policy on liability result- ing from patient deaths. Physicians in private hospitals may fear that if a patient under their care were to develop serious complications and die, this might be grounds for crimi- nal investigation. In addition, because malpractice insurance coverage is not yet readily available to physicians in Vietnam, individual physicians or hospitals are potentially ex- posed to unlimited financial liability. A clear government policy on the extent of physi- cian or hospital liability for accidental or illness-related patient deaths would enable pri- vate hospitals to better assess and manage their risks. Subsidized Public Hospitals. The presence of subsidized public hospitals means that as long as public hospitals offer inpatient services well below cost to all users, espe- cially the better-off patients (see section D.4 in chapter IV), it will be difficult for private hospitals to become financially viable. This may partly explain why many of the private hospitals in the country (such as the International Hospital in Hanoi, for example) have been experiencing severe financial problems. Limited Size of Market for Private Hospitals. Many of the private for-profit hos- pitals in the country have experienced financial problems as the commercial market for 101 expensive high-tech health services is limited. People cannot afford the very high prices charged, or prefer to go abroad to Singapore or Bangkok for specialized treatment. 2. Semipublic Health Services in Public Hospitals One phenomenon that is becoming increasingly common is the growth of semi- public wards in public hospitals. Typically, this involves setting up a separate ward with additional amenities and charging higher rates for patients who have the ability to pay. For example, the Viet Due hospital-one of the oldest and most prestigious central spe- cialty hospitals in Hanoi--set up a semipublic ward with 31 beds in February 1999. The (fully) public wards have a total of 450 beds. In addition to beds, the semipublic ward has its own operating theater and post-operation recovery rooms. The semipublic ward has air-conditioned rooms with attached bathrooms and other amenities which are occupied only by two patients. The nurse/patient ratio is higher in the semipublic ward than in the fully public ward. However, costs to patients are also considerably higher. For instance, a surgery using laparoscopy methods that costs VND 500,000 in the public ward can cost as much as VND 2,000,000 in the semipublic ward. The daily bed rate is VND 10,000- 20,000 in the public ward but VND 200,000 in the semipublic ward. In addition, patients, both in public and semipublic wards, have to pay for all consumables, such as sutures, needles, and blood, and for other special items that may be needed in a surgery (e.g., prosthesis, heart valves, replacement hips, etc.). The head of Viet Duc Hospital mentioned several reasons for establishing semi- public wards in public hospitals: * Rehabilitation of the hospital. The establishment of the semipublic ward is generating revenue for the hospital that can be used to rehabilitate equipment in the entire hospi- tal, not just in the semipublic ward. * Development of a model for the public ward. Public wards in hospitals are typically overcrowded, unclean and offer few services to patients. The higher-quality semipub- lic ward can thus serve as a model for emulation for the public ward. * Improving worker morale by providing them financial incentives. A proportion of the resources generated by the semipublic ward is used to reward all health workers in the hospital, irrespective of whether they work in the public or semipublic wards. This improves health worker morale and productivity. * Satisfying excess demand. Many public hospitals, especially in the large towns and cities, face excess demand for their services. This is manifested in long waiting lists (sometimes extending to months or even a year) for patients needing elective or nonemergency surgery or treatment. The development of the semipublic ward can re- duce these waiting lists by enabling those consumers who can pay additional charges to be accommodated in the semipublic ward. * Cross-subsidizing the poor. A portion of the revenues generated from the semipublic ward can be used to provide free or subsidized treatment to the poor. * Cost savings for the government. The head of Viet Due hospital estimated that each bed in his hospital's semipublic ward saves the government VND 30 million per year, which is the average annual cost to the state budget of a bed in Viet Due hospital. 102 Thus, the establishment of 31 beds in the semipublic ward saves the government a to- tal of nearly one billion dong annually. There are arguments for and against the development of semipublic wards in pub- lic hospitals. On the one hand, there is a real danger that hospitals will divert resources and their best health workers from the low-paying public wards to the high-paying semi- public wards, thus worsening the state of the public wards. This may create a two-tier system within public hospitals. On the other hand, the fact that public hospitals can cross- subsidize public wards from the revenues earned by the semipublic wards is an attractive option, although it is not clear that they engage in such cross-subsidization. The estab- lishment of semipublic wards can also reduce the financial dependence of public hospi- tals on the state budget, thus freeing resources needed for more important causes, such as primary health care and preventive activities. Informal discussions with employees and administrators of public hospitals hav- ing semipublic wards suggested that the establishment of semipublic wards in these hos- pitals had actually improved the quality of care and service standards in the public wards. However, it was not possible to confirm these assertions with empirical data. In addition, there are other questions which need to be addressed before determin- ing the financial viability and validity of this approach. For example, are the semipublic wards in public hospitals subsidized or not? In other words, do they generate revenues that exceed their full costs? Where do the investments for these semipublic wards come from? If from public funds, are these investments paid back from revenues? Some developing countries have successfully used this approach to increase the financial autonomy of public hospitals and to wean them away from public subsidies. Additional research in Vietnam on the impact of semipublic wards in public hospitals on the quality of care in public wards, financial autonomy of hospitals, and cross- subsidization would help inform discussions of these topics. 3. Contracting Out of Services Another form of privatization, which is popular in the health sectors of many de- veloping countries, is the contracting out of selected services in public health facilities to private contractors. The idea behind contracting out is that it will improve both service delivery as well as reduce costs. Both clinical and nonclinical services (e.g., cleaning, laundry, accounting) can be contracted out. In the last few years, many public hospitals in Vietnam have begun contracting out selected services. The most popular services for contracting out are cleaning and din- ing. Cho Ray hospital has contracted out its food preparation activities and eliminated its Department of Nutrition. Some second-level general hospitals and third-level private hospitals have also been contracting out specialized surgeries to take advantage of high- technology centers and skilled surgeons elsewhere. Contracting out is still in its infancy in Vietnam. There is scope for contracting out services and diagnostics, such as laundry, dining, maintenance of medical equipment and laboratory and X- ray services. However, a number of legal, contractual and eco- nomic feasibility issues relating to contracting out need to be addressed first. For exam- ple, what exactly will be the cost savings from contracting out different clinical and non- 103 clinical services? Is there clear evidence that efficiency in service delivery will be im- proved as a result of contracting out? What are the legal constraints, if any, in contracting out services at public health facilities? Should private contractors be required to retain the existing government health staff at the selected facilities, and, if so, how much authority will the contractor have over them? Can public hospitals establish an open and transpar- ent process for evaluating and selecting bid proposals from competing contractors? D. Monitoring and Regulation of the Private Sector There is considerable variation across private providers in the quality of care, in- cluding infection control, hygienic practices, and standards of practice protocols.42 Diag- nostic methods are also often inconsistent. For instance, in the Hai Phong study of private reproductive health providers, it was found that many providers treated RTIs by syn- dromic management without laboratory confirmation. Patients treated for known STDs were not routinely referred for HIV testing and counseling. Information, education and communication (IEC) about partnership notification was infrequently given to the patient. This indicates the need for better monitoring of quality of care and standards of practice observed by private providers. The mail survey of private health facilities in 44 provinces revealed that 14 per- cent of the private health facilities investigated by provincial authorities were found to be violating regulations (Table 5.5). The highest rate of regulation violations were found among traditional medicine rehabilitation centers (67 percent), polyclinics (39 percent), sellers of traditional medicine (30 percent), acupuncture rooms (27 percent), and mater- nity homes (27 percent). The data did not indicate the seriousness of these violations and if the violations compromised the quality of care at these facilities. Only 8 of 21,486 fa- cilities were found in violation of laws. 42 Of course, the same could be said of public providers. 104 Table 5.5: Number of private health facilities found violating regulations and laws, by category of facility, 1999 Investigaed Normal i, Plating rcgu- CaICgorv ofichn facilities overanons lnor Penal case. General hospiials i 1 0 0 N1atemi1y homes 112 82 30 (27%) 0 Polyclinics 69 41 27(39%) 1 GP's clinics 4,416 3,906 510 (12%) 0 Specialist's clinics 1,865 1,666 199 (11%) 0 Dentist's rooms 1,397 1,137 260(19%) 0 Laboratories 132 116 16(12% 0 X Ray rooms 33 31 2 (6.) 0 Aesthetic surgerN rooms 24 23 1 (4%) 0 Rehabilitation centers 98 84 14 (14%) 0 Traditional medicine rehabilitation centers 109 3h 73 (67%) 0 Nursing ser"ices 2,464 1.940 524 (210.%) 0 Family planning services 320 269 JI (160.' 0 Traditional medicine clinics 2,632 2.527 104 (4%) 1 Acupuncture rooms 86 63 23 (27%) 0 Private pharmacics 2.823 2.527 295 (10)0 1 Drug commission sellers 3.574 2.995 576 (16%) 3 Pharmaceutical companie,. 104 103 1 ( 1%) (1 Sellers of traditional medicines 979 684 293 (300) 2 Others 288 239 49 (17%) 0 TOT A L 21.486 17.669 3.048(14%) 8 Note Figures in parentheses are percentages of imestigated faciinies Source, Dung 1999, While there are many regulations and decrees governing the minimum quality standards and practice protocols expected of private providers, the implementation of these laws through regular inspections of facilities is less than satisfactory because in- spections of private facilities are the responsibility of the provincial health authorities, who often lack the human and other resources needed to undertake regular inspections of all private facilities. Some countries have established new and independent regulatory units within the Department of Health Services that are partially financed by registration fees. These unit need to be independent and their members should not be practicing in the private sector, but should be selected for their professional experience. In addition, they could provide and organize ongoing training and mentoring programs for practitioners in both sectors. They could also strengthen the referral and information systems by developing a public database about providers and their areas of expertise. In many ways, consumers are currently the main regulators of the private health care system in Vietnam, and effectively vote with their feet. Yet, consumers may not be that well informed about what constitutes quality of services and there is a compelling need for more professional supervision and control of this sector. 105 E. Pharmaceuticals and Drugs 1. Drug Production and Imports43 Until 1989, pharmaceutical factories, like all other industrial units in Vietnam, had little control over decisions involving production and financial matters, production, input procurement and pricing (including employee wage levels), sale and pricing of out- puts were all determined by state planners. In practice, pharmaceutical factories sold drugs and medicines to public hospitals and health centers at subsidized prices fixed by the state. After 1989, pharmaceutical factories (and factories producing or repairing medical equipment) were given the freedom to make their own decisions concerning: (a) use of inputs, including wage levels, (b) production levels, and (c) selling prices of drugs and medicines. Pharmaceutical factories were also informed that they could no longer rely on government subsidies. Until 1989, drugs and medicines were dispensed free through the public health network to all patients. In 1989, the Government began to allow pharmaceutical factories to open retail pharmacies and sell drugs and medicines to individuals, hospitals and health centers directly. Qualified pharmacists could also apply for licenses to open pri- vate pharmacies. The annual market in pharmaceuticals was estimated at about VND 5,526 billion (US$425 million) in 1997-or about US$5.20 per capita in sales. Of this, only VND 1,385 billion or 25 percent represents the market for locally produced drugs, the remain- ing 75 percent of the market is accounted for by imports. It is estimated that essential drugs account for 60 percent of the total turnover of drugs in the country. There are a total of 8,000 registered drugs in the market, 5,000 of which are pro- duced domestically and the remaining 3,000 imported.44 In addition, another 1,200 tradi- tional medicines are registered for sale in the Vietnam market. Most raw materials for the domestically produced drugs are imported into the country. Domestic production of pharmaceuticals has increased phenomenally over the past decade. In 1997, domestic production was approximately US$107 million-up from a level of merely US$777,000 in 1989. There was a sharp decline in the output of phar- maceutical products during the 1980s, primarily due to a general lack of financial re- sources and the consequent inability to import raw materials and spare parts for equip- ment (World Bank 1992). There are a total of 18 central government pharmaceutical factories and 126 pro- vincial government factories producing drugs. However, only five of these drug produc- tion units have been upgraded to attain ASEAN Good Manufacturing Practice (GMP) standards. Nearly all of the pharmaceutical factories are state owned, with little private 43 This section draws heavily from Tornquist 1999. 44 The government retains an important position in the importation of drugs via its import and trading com- panies. 106 ownership in drug production.45 However, the private sector plays an important-indeed, dominant-role in drug retailing. The Drug Administration of Vietnam (DAV) is the highest authority for drug management in the country, and was established only four years ago. It replaced the De- partment of Pharmacy in the Ministry of Health. It has two support agencies-the Drug Quality Control Institute and the Drug Inspectorate. In each province, there is a pharmaceutical department and a Drug Quality Control station that falls under the management of the Provincial Health Authority. At the central level, VINAPHARM is a state-owned corporation responsible for supplying drugs to the entire country. Its members are trading enterprises at the central and provincial levels. The remaining provincial pharmaceutical enterprises are presently under the control of the Provincial Health Authority. Thus, despite the major changes that have occurred in the sector, the government continues to retain a significant commercial interest in the production, importation and distribution of drugs in Vietnam. 2. Distribution of Drugs through Pharmacies and Drug Vendors At the level of the consumer, there have been three major impacts of the deregula- tion of drugs and pharmaceuticals. The first impact has been on drug availability. With deregulation, the drug supply situation has improved dramatically, and shortages of drugs, which were endemic throughout the Vietnamese health care system until the early 1990s, have disappeared. Even commune health centers in the rural areas, which used to be chronically short of drugs, have adequate supplies of drugs for sale to patients. In part, the distribution of drugs has improved because of the ubiquity of drug vendors and phar- macy shops throughout the country. It is estimated that there as many as 20,000 or more drug retail outlets in the country. Even the smallest village has its own drug shop. Second, the deregulation Figure 5.4: Average out-of-pocket expenditure on drugs per health service contact (across ail health providers), by per capita expenditure has lowered prices of mainly quintile, 1993 and 1998 domestic drugs for the con- sumer. Higher levels of domes- tic production, greater imports, 19944 and a more competitive distribu- tion system all have contributed ' to a decline in real drug prices. 30 This is observed in the VLSS 23 survey data in the form of a real 920 1 decline in out-of-pocket expen- diture on drugs per health ser- 9I vice contact between 1993 and 1998 for all income groups Poorest Second Thrd Fourth Richest (Figure 5.4). It is also observed Per capita expenditure quintile 45 The state-owned companies have entered into partnership with foreign firms for the manufacture of drugs, but the government has a majority stake in these joint ventures. 107 in the behavior of the medicine price index relative to the overall consumer price index over the 1993-98 period (Figure 5.5). While the price index for medicines was virtually flat between 1993 and 1998, the overall CPI and the food price index grew at annual rates of 8 percent and 10.8 percent, respectively, implying a decline of more than 30 percent in the real price of medicines. Figure 5.5: Food, nonfood and medicine price indices, 1992-98 In addition to 10. these two positive developments, there 160 - - ----- has been a third, alarming develop- 1 _40 ment from the de- regulation of the pharmaceuticals in- .i P Iode,c dustry. With easier availability and M tndex rclatve to CM lower prices, con- sumer purchases of drugs, especially for self-medication, V -O'0t" have increased dra- matically. Data from the VLSS show a large increase in the utilization of drug vendors (from whom individuals typically obtain drugs without prescription). While individuals made an average of 2.1 annual service contacts per capita with drug vendors and phar- macy shops in 1993, the number had increased to 6.8 annual contacts per capita by 1998. Interestingly, there were no major variations in the annual number of drug vendor con- tacts per capita across income groups (Figure 5.6). Indeed, drug vendors are the most fre- quently used health provider in Vietnam, accounting for two-thirds of all health service contacts. W hat is disturbing is Figure 5.6: Average annual number of health service contacts per capita with drug vendors and pharmacy shops, by per capika expenditure that most of the drug ven- quintile 1993 and l998 dor contacts represent pur- 1993 7 chases of drugs without a 71 7 prescription. The VLSS 6 1998 data indicate that 93 percent of all drug vendor contacts were for obtaining medicines without a c 2 prescription, with not much " 2 r 2 variation across economic groups. This effectively means that consumers self- o Rottom Second ThIrd Fourth Top treat their illnesses, perhaps Per capita expenditure quntile with some advice from the drug vendor. 108 It is very likely that this unbridled increase in self-treatment of illnesses is related to the easier availability of drugs and the real decline in drug prices over time. The con- sequences of this reliance on self-treatment are alarming and are discussed later in this chapter. A study of 1,833 patients in two pharmacies in Hanoi during a two-week period in 1994 confirmed that the vast majority of patients visiting pharmacies even in an urban center like Hanoi do not have medical prescriptions. In the Hanoi study, only 0.8 percent of patients had a medical prescription. Almost all (94.9 percent) had already decided which drug they would be purchasing before they arrived at the pharmacy, while a few patients (4.3 percent) asked for advice from the salesperson at the pharmacy. The average number of drugs sold per service contact was 1.4. Of the total number of patients, nearly 17 percent bought antibiotics. About 20 percent of the patients buying antibiotics were children under the age of 15 years. Table 5.6 shows the most commonly dispensed drugs in the two pharmacies, ranked by the number of times these drugs were sold in the two-week period. The most commonly dispensed drugs were antibiotics, vitamins, analgesics, tranquillizers and drugs for eye infections. Table 5.6: The most commonly dispensed drugs ranked b- number of times sold in a two-,^eek period, Hanoi, 1994 Name of drug Frequenc v Ampicillin 138 Vitamin C 132 Vitamin B 128 Chloramphenicol 4% eye drops 95 Naphazolin 87 Metamizon (Analgin) 63 Diazepam (Sed&xen) 50 Co-tmmoxazol 49 Becbenn 4 Tetracycline 42 Anticoldness oil 36 Mentol (Bachap 36 Erythromycin 36 Ephedrine + Sunphanlamid (Sulfarin) 33 Chloramphenicol + Dexamethason - Naphazolin (PoI- dexan) 33 Phenobarbital + Phenacetn + Aspinn (HHTK) 32 Bo Phe (traditional anticough medicine) 31 Paracetamol + Cinchona bark (Ankitamol) 31 Chloramphenicol (Clorocid) 30 Paraceramol + Chlorpheniramol (Pamin) 29 Dimenhydranate 29 Sao .ang Balm (traditional balm of mixed essences) 29 Vitamin B6 29 Note. Brand names within parentheses. Source. Chuc and Tomson 1996. 109 Nearly one-half of all individuals buying antibiotics purchased an antibiotic course of only 22 or fewer days. Another 46 percent purchased antibiotics for 212--5 days, while only 2 percent purchased antibiotics to last the recommended 10-day course. A higher proportion of short antibiotic courses were observed for children aged 0-1 years. Combination drugs also were also found to be popular among customers visiting the two Hanoi pharmacies. Indeed, the proportion of combination products is higher in Vietnam (22.2 percent) than in countries with a comparatively rational pharmaceutical sector, such as Norway and Sweden, where they represent fewer than 5 percent of all drugs in the market (Chuc and Tomson 1996). Combination drugs can have dangerous side effects. Two combination drugs that were among the most commonly dispensed drugs in the Hanoi study-APC (aspirin, phenacetin, and caffeine) and HHTK (pheno- barbital, phenacetin, aspirin)-are known to have dangerous side effects, including kid- ney damage and drug dependence, and have been withdrawn in several countries.46 3. Health Effects of Drug Use: The Case of Antibiotic Resistance Antibiotic resistance levels in Vietnam have increased and reached epidemic lev- els. This is an extremely serious problem that threatens to derail the significant achieve- ments in the health sector, as Vietnam loses the ability to control and prevent the spread of many infectious diseases. The spread of antibiotic resistance is directly the result of overuse (unnecessary consumption), irrational use (broad instead of narrow spectrum) and ineffective use (short course instead of full course) of antibiotic drugs by individuals self-treating themselves. The problem is compounded by low levels of competence in clinical pharmacology and clinical pharmacy among pharmacists, drug vendors and the public. Even when drugs are prescribed by doctors, there is low compliance by patients with rational treatment guidelines. Often, poor patients might limit themselves to a two- day course of antibiotics (instead of the recommended ten-day course) as they do not have enough money to buy the full course. Antibiotic resistance is manifested in the increasing difficulty in the last few years of treating typhoid fever and containing its spread in Vietnam (Tomquist 1999). This is the direct result of the typhoid-causing microbe, Salmonella typhii, developing resistance to chloramphenicol, ampicillin and trimetoprim/sulphamethoxazol. Salmonella typhii has by now developed complete resistance to these three drugs. In this situation, the fever of the infected person does not respond to treatment with these drugs, nor does the epidemic decline when these drugs are used in the battle against typhoid. Table 5.7, which shows the results of clinical trials, indicates that antibiotic resistance levels of Salmonella typhii have increased from about 40-50 percent in 1993 to nearly 90 percent in 1998. Fortu- nately, more potent antibiotics, such as quinolones and third-generation cephalosporins, to which the microbe has not yet developed resistance, are available in the fight against typhoid. However, the Salmonella typhii microbe is capable of developing resistance to these drugs, too, if these drugs are used excessively and irrationally. 46 In Vietnam, phenacetin was withdrawn (for import) from 1996 onwards, and was forbidden in any for- mulation since 1998 (Letter No. 3098). 110 Table 5.7: Antibiotic resistance or Salmonella tphii in Vietnam. 1993-98 1998 1996 1994 1993 .%v of' Rew.%iant \'o of Re.%ivant No of' Resistant No at Resistant Dngsqmpl,, i-111 Samples.! r%) sumples~ ("o sornpiN ff% Ampicillin 161 88 534 87 9LI 44 43 Cefiriaxone 17 0 304 1 45 0 Gentamicmtl 90 39 0 Chloramphenicol 200 S) 632 91 90 9-' 45 Cotnmo.uzol 196 91 6 11 5 90 98 4.1 47 Nortloucn 60 0 443 0 90 0 Tobramycin 90 Tctracvchnc 34 59 Source Tornguis 1999 Antibiotic resistance is not limited to the Salmonella typhii, indeed, there is com- pelling evidence that antibiotic resistance is now affecting all infectious diseases in Viet- nam. The prevalence rate for macrolide resistance is 40 percent in isolates of pneumo- cocci, staphylococci and in Group A beta-haemolytical streptococci (Tornquist 1999). This situation is extremely worrisome as it compromises the use of macrolides, which is the drug of choice in the treatment of respiratory tract infections in patients allergic to penicillin. 4. Recent Policy Actions The case of the pharmaceutical sector is a telling example of the consequences of deregulation that is not accompanied by proper monitoring, supervision, and regulation. Fortunately, the Government has understood the seriousness of the situation and is taking concrete steps to control the irrational and indiscriminate use of drugs. The Government has recently formulated a Vietnam National Drug Policy (VNDP), which is being piloted in seven provinces and cities throughout the country. The eventual aim is for nationwide implementation during the period 2001-2015. The VNDP is an attempt at developing a holistic and comprehensive policy on drugs. Its objectives are twofold: (i) assuring a sufficient supply of quality drugs to serve the health needs of the population, and (ii) ensuring rational and safe use of drugs. To achieve these objec- tives, the VNDP will support a number of initiatives, including: * consolidation of a nationwide quality assurance and quality control system, * upgrading of quality control laboratories, * strong penalties for illegal marketing of counterfeit and substandard drugs, * enforcing ASEAN GMP standards on all Vietnamese drug production units, * strengthening professional competence in the staff of the central drug management authority through training, * development of a drug information system for health workers, pharmacists and drug vendors, * strengthening Drug and Therapy Committees in hospitals, 111 * education and training for key personnel in district and provincial pharmaceutical of- fices in the area of drug management, * control of advertisements for pharmaceuticals, * preparation and dissemination of a Vietnamese national drug formulary-a clinical drug use handbook, and * revising training for health workers in clinical pharmacy and pharmacology. In addition, a new drug law will soon be developed, implemented and enforced. The new drug law will include detailed regulations on prescribing, over-the-counter sales, and division of medical and pharmaceutical practice. Already, some of the initiatives are paying off. Thanks to the quality control on drugs, the quality of drugs on the market has improved impressively within the past few years. The proportion of counterfeit drugs on the market dropped from 7.1 percent in 1990-91 to 0.2 percent in 1998. However, the proportion of substandard drugs has re- mained at 3.5 percent. Five pharmaceutical factories have met the ASEAN GMP stan- dards, however, these standards will be difficult to achieve for the majority of factories without a major upgrade of production technology. It will, however, take some time be- fore the spread of antibiotic resistance can be halted by an education and communication campaign that restricts the overuse of antibiotic drugs by consumers, pharmacists, and health workers. A dilemma facing the government is that the Ministry of Health is both a major owner and manager of drug production in the country as well as the agency in charge of regulating drug production and importation. This can at times lead to a potential conflict of interest between the commercial interests of government in drugs and the public health benefits from drug use. F. The Role of Nongovernmental Organizations in Health Services De- livery 1. Number and Distribution There are about 400 international NGOs active in Vietnam. Out of these, about a hundred are active in health. According to the Project Coordinating Department of the MoH, more than US$7 million out of US$74 million overseas development assistance (ODA) support in 1998 came from NGOs (Figure 5.7) (MoH 1998e). However, these are rough figures, and it is estimated that the total contribution of NGOs to the health sector may be three times higher (about US$ 25 million), constituting roughly one-quarter of all disbursed ODA to the health sector. The reason for the dis- crepancy 112 between official and actual es- igure .7: NGOauistance, 1"2- timates is that many NGOs not 8WO. 1. only work with health-related 7.000 ' "0 o governmental bodies but also 6 channel funds through mass 9 organizations, such as the 0 Women's Union, Red Cross, 4.0.0 Peoples' Committees, and the -. CPCC. Additionally, spending 3,00 - - 0-. - data are not always disaggre- 6 ,0--- ---- ------ - - - gated by sector, and many . - NGOs support integrated activi- 1,00 ties. -- 4 1992 1993 1994 1995 1997 1998 NGOs work in a large number of provinces in the country. There are 27 provinces that host three or four NGO activities on average, which adds up to more than 90 separate NGO activities in health. There are 42 provinces that have at least one primary health care project run by an NGO. Of course, in any one of these provinces, the primary health care activities may reach only a small number of districts or even communes. There is a large concentration of NGO health activity in Hanoi, Ho Chi Minh City and Hue, which together have at least 94 NGOs. 2. History of NGOs in the Vietnam Health Sector There were very few NGOs in Vietnam-about 10-during the period from the late 1960s to the late 1980s. All of these NGOs had a health component and concentrated on relief and reconstruction efforts and the supply of medicines and hospital equipment, all of which were badly needed at that time. During this period, NGOs could only work with the central government as a counterpart. At least until the start of doi moi in 1986, many NGOs were still mainly providing supplies to hospitals and supporting the Ministry of Health directly. However, some were also sending Vietnamese health staff abroad for training, so there was some attention to structural development and institution building. The total support in 1988 by all NGOs (including those from socialist countries) amounted to about US$70 million, and it was the health sector that received, by far, the greatest proportion of that amount. With the opening up of the economy in the late 1980s, international NGOs began coming into Vietnam in large numbers. Many opened representative offices in the coun- try. What the NGOs found in Vietnam was a difficult situation. Although the reforms had done much to stimulate the market economy, the health and education sectors were in a state of crisis. One social consequence of reforms was the breakdown in the commune health system. This made it imperative for NGOs to concentrate their work at the district and commune levels. NGOs working in the health sector have long realized that partici- pation of the community in identifying and addressing their own health needs is an essen- tial element for sustainable health care. In addition, the Government encouraged NGOs to focus their efforts on strengthening the capacity of commune health workers and volun- teers. 113 At the provincial, district and commune levels, the direct counterparts are often the local administrative bodies (People's Committees), the CPCC, and the health ser- vices. The local branches of mass organizations (notably the Women's Union and the Red Cross) provide an additional network to reach the communities. The lack of an inde- pendently developed "civil society" (and local NGOs) was a new situation for interna- tional NGOs with experience in other countries. Although this still hinders the operation of some NGOs that insist on working with local NGOs, several of the most experienced NGOs in Vietnam have managed to work well with the government.47 3. Role and Impact of NGOs on Health NGOs in Vietnam are active in a wide array of health-related fields and at very different levels. They are active in helping to build national health programs,48 cooperat- ing with central research institutes and universities,49 training government personnel at all levels, and working with communities in integrated development projects. The range of projects supported by NGOs include: * Community development, as integrated projects including health * Primary health care, projects having intersectoral collaboration but with the main fo- cus on health * Sectoral health development, especially in training, research, prevention, curative and clinical work * Advocacy. By their nature, the community development and primary health care projects in- clude a wide range of different activities, based on the localities and their characteristics, and could therefore better be subdivided on a geographical basis. Sectoral health devel- opment projects include: * Infectious disease control (HIV/Aids, TB, malaria, dengue, vaccination) * Reproductive Health * Water and sanitation * Mother and child care * Traditional Medicine * Rehabilitation of disabled people, special corrective surgery, etc. 47 See for instance McCall: "Partnership with Government for Poverty Alleviation". 48 Like MCNV with the National Tuberculosis Program; AFAP and MCNV in the emerging Dengue Pro- gram. 49 Like Australian Foundation for the Peoples of Asia and the Pacific Limited (AFAP) and Medical Com- mittee Netherlands-Vietnam (MCNV) in dengue and malaria research; International Cooperation for De- velopment and Solidarity (CIDSE) in Hanoi Medical School; Brot fur die Welt (BfdW) with Can Tho Uni- versity and many more. 114 * Commodity support, language training for health professionals, etc. A main strength of NGOs lies in introducing participatory, integrated approaches and in working with the communities and health staff in the communes and districts. Whether starting from a primary health care perspective or when directly addressing the needs as the community identifies them, there is always a health component embedded in an integrated whole, which includes agriculture, small-scale business and credit pro- grams, gender issues, social development, and local infrastructure (e.g., roads, bridges, electricity supply). But also when working at the central level, where they are being increasingly dis- placed by international donors, NGOs have made important contributions, many of which have already been institutionalized. Among some of the important NGO contributions in Vietnam in the past are: * early warning of the emerging HIV/AIDS pandemic * social marketing of contraceptives * the concept of community-based rehabilitation * the concept of "inclusive education" * introduction of participatory needs assessment methods (PRA, for instance) * the "positive deviance inquiry approach" to alleviate malnutrition among children50 * poverty alleviation by way of local credit and saving systems51 * the establishment of the Community Health Research Unit in Hanoi Medical School52 * the nationwide introduction of Short Course Chemotherapy (SCC) and Directly Ob- served Treatment Short Course (DOTS) in the tuberculosis (TB) control program * the introduction of impregnated mosquito nets for malaria control * community-based methods (including biological control) for dengue control * the introduction of participatory approaches in traditionally top-down managed verti- cal health programs.53 4. Changes in Scope and Approach of NGOs Over time, the focus of NGO work in the health sector has changed. Table 5.8 summarizes some of these changes. For instance, those NGOs that were in the past work- ing in the vertical health programs, the centralized and large-scale provision of commodi- ties, or relief aid have largely redirected their focus to other levels and other fields of at- tention. This was made possible by changes in Vietnamese policy and the entrance of of- 50 Pioneered by Safe the Children US and replicated through the "Living University" concept. 51 Among others: CIDSE. 52 By CIDSE. 53 At present taking place in the TB program, for instance. 115 ficial donors. Likewise, NGOs that originally worked in more selective fields have, over time, gravitated to an integrated, community-managed approach in which health is de- fined as broadly as possible and the local potential for self-reliance can be mobilized. Table 5.8: Changes in NGO focus over time From To I. Mainly central level Mainly district and community Level Commodity support as a complementary element of the 2. Relief, construction & commodity support main activities Stronger focus on qualitative research using commut- 3. Quantitative research nity based methodologies Training for health professionals and villagers TechrLical training for health professionals Training on participatory methods, capacity building. 4. (equipment & maintenance) communication skills and technical matters. 5. Mainly curative Stronger focus on pre,eniive care 6. Easily-accessed areas Remote areas Pnmarv health care and integraed programs (commu- 7. Selective care nity development) Because of changes in the priority policies of the Vietnamese Government and the relaxation of security regulations, some NGOs have shifted to work in increasingly re- mote areas of the country and with the poorest and most disadvantaged groups. Training contents of NGO programs have changed from merely technical to capacity building in community skills for health education and participatory methods. Likewise, the nature of research (e.g., for needs assessments) has changed in the sense that the aims, methods and implementation are determined with and by the community, rather than the community being studied as a passive subject by outsiders. The same is true in the monitoring and evaluation of projects. 5. Examples of NGO Initiatives Some observers regard the small scale of NGOs as a weakness. For example, it is unsatisfactory when the success of an NGO remains limited to its original target or inter- vention area, since this increases inequality between districts or communes. When NGOs operate at a relatively small scale, they are unable to replicate successful projects by themselves. This is, however, a misleading argumnent, since one of the main functions of NGOs is to play an advocacy role and set up models for emulation. The Vietnamese health system has shown an ability in the past to readily absorb and replicate innovations pioneered by NGOs. Indeed, it is one of the advantages of the Vietnamese system that there is a strong internal monitoring built into the system, which functions to pick up any successful developments and replicates them. An example of a small NGO project that has had broad impact is the action-based training course called Work Improvement and Neighborhood Development (WIND). WIND was started by a group of health professionals in Can Tho. The WIND training methodology is used to train community health volunteers and health personnel at the grassroots level. Although the program was started as a pilot project on a very small scale in the early 1990s, it has expanded and now has a national profile. The methodology is highly participatory, involving villagers directly in assessment and then in the setting up 116 and implementing practical and low-cost action plans related to health and environment issues at the household and community levels.54 Likewise, the approach of 'Community-Managed (Health) Development,' as pio- neered in Quang Tri province,55 is also increasingly being replicated by other provinces in the Central Region. This occurs mainly through Vietnamese contacts and without much "advocacy" from the NGO-in fact, the demand generated in other provinces can- not be met by the NGO and in several of these areas, the Vietnamese health staff started a similar program on their own, helped in training by their colleagues in Quang Tri. Yet another example of an NGO health project that was replicated is the monthly immunization drive in the Phu Luong district CBPHC program. The system (cold chain and all the organization this takes) was introduced very early in the program and has been operating in Phu Luong ever since. Monthly immunization now also occurs in every other district of what was Bac Thai province (now Thai Nguyen and Bac Kan), thanks to the emulation effect. The emulation effect is also observed in the approach originally pio- neered by AFAP and MCNV but now incorporated in the national Dengue Control Pro- ject.56 This suggests that the disadvantages of operating at a small scale are often exag- gerated. If an activity is well thought out and implemented, it tends to get replicated at a broader level throughout the system.17 Governmental organization partners do absorb and learn from working with NGOs, and are able to later apply these skills in their own work. 6. NGO Advocacy Advocacy is an activity in which NGOs use their experiences to influence gov- ernmental policies. Some NGOs spend much effort in advocacy it the provincial and even central levels,5 but most others do not have the human resources to spend much time and energy outside their direct project work. In general, it is felt that advocacy is a useful function of NGOs, although it is preferably done at the intermediate, provincial level and not at the central level. As discussed before, it is best if the information about new approaches percolates naturally upwards, instead of being directed downwards in the form of nationwide decrees. The unique experiences of NGOs at the grassroots level must sometimes be fed back to the central levels to create "political space" for projects to innovate and take risks. This calls for an improvement in the professional competence within NGOs to cre- ate and maintain productive relationships with all levels of the government.59 NGOs are often in a good position to help local government organizations in experiments to develop 54 From BfdW. 55 By the Provincial Health Service of Quang Tri in collaboration with MCNV. 56 From AFAP. 57 From CIDSE. 58 For instance, Oxfam GB. 59 E. McCall, "Partnership with Government." 117 appropriate solutions. Once it is understood why some approaches work and others fail, it is helpful to use this knowledge to build models as examples for other areas. While there have been successful examples of emulation and 'copying', some NGO workers also warn against the pitfalls of a 'model approach.' If models are actively promoted for repli- cation in other situations, they turn into another blanket approach and may lose their ap- propriateness. At some point in the future, the inputs of NGOs may no longer be needed, the ex- plicit objective of many NGO projects is to transfer all know-how into Vietnamese hands. Nevertheless, for the foreseeable future, their specific role, which is complementary to donors and the government, will still be needed. 118 CHAPTER V ANNEX TABLES Table A5.1: Number of private health facilities, by category and rural/urban, 1999 Licensed Nonlicensed All facilities Urban Rural Urban Rural No. Category No. % No. % No. % No. % No. % 1 General hospitals 2 0.0 2 0.0 0 0.0 0 0.0 0 0.0 2 Specialized hospitals 1 0.0 1 0.0 0 0.0 0 0.0 0 0.0 3 Traditional med. hospitals 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 4 Maternity 152 0.4 103 0.5 49 0.5 0 0.0 0 0.0 5 Polyclinics 311 0.9 89 0.4 4 0.0 15 1.3 203 5.9 6 GP's clinics 5894 16.2 4231 19.2 1046 10.6 143 12.4 474 13.8 7 Specialist's clinics 3251 8.9 2693 12.2 537 5.5 8 0.7 13 0.4 8 Dentist's rooms 1971 5.4 1477 6.7 447 4.5 26 2.3 21 0.6 9 Laboratories 182 0.5 174 0.8 7 0.1 1 0.1 0 0.0 10 X Rays' rooms 183 0.5 160 0.7 17 0.2 5 0.4 1 0.0 11 Aesthetic surgery's rooms 30 0.1 30 0.1 0 0.0 0 0.0 0 0.0 12 Rehabilitation centers 125 0.3 123 0.6 2 0.0 0 0.0 0 0.0 Traditional medicine reha- 13 bilitation's centers 111 0.3 110 0.5 1 0.0 0 0.0 0 0.0 14 Nursing services 4213 11.6 1932 8.8 1483 15.1 216 18.7 582 16.9 15 Family planning services 495 1.4 349 1.6 134 1.4 5 0.4 7 0.2 Traditional medicine's clin- 16 ics 4962 13.6 1817 8.3 2028 20.6 158 13.7 959 27.8 17 Acupuncture's rooms 361 1.0 185 0.8 102 1.0 24 2.1 50 1.5 18 Private pharmacies 4906 13.5 4336 19.7 457 4.6 92 8.0 21 0.6 19 Drug commission sellers 6621 18.2 2439 11.1 3206 32.6 326 28.2 650 18.9 20 Pharmaceutical companies 220 0.6 219 1.0 1 0.0 0 0.0 0 0.0 Sellers of traditional medi- 21 cines 1596 4.4 924 4.2 305 3.1 97 8.4 270 7.8 22 Joint ventures 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 23 Others 855 2.3 613 2.8 9 0.1 39 3.4 94 2.7 TOTAL 36442 100.0 22007 100.0 9835 100.0 1155 100.0 3445 97.1 Source: Dung 1999. 119 Table A5.2: Private health facilities (excluding traditional healers) in 44 surveyed provinces, 1999 Licensed Nonlicensed No. Province Total Urban Rural Urban Rural I Tuyen Qunag 829 109 323 20 377 2 Ha Giang 80 77 3 3 Cao Bang 166 166 4 Lai Chau 41 41 5 Yen Bai 182 106 75 1 6 LaoCai 108 55 32 3 18 7 Thai Nguyen 318 153 23 127 15 8 Bac Kan 64 10 40 9 5 9 QuangNinh 304 227 77 10 VinhPhuc 253 58 116 26 53 11 BacNinh 162 83 79 12 Bac Giang 463 35 84 67 277 13 HaNoi 2883 2237 646 14 HaiPhong 872 581 112 53 126 15 Hao Binh 176 60 116 16 HaTay 442 205 237 17 HungYen 138 114 24 18 Thai Binh 828 109 322 20 377 19 NamDinh 335 88 227 11 9 20 HaNam 210 39 171 21 NinhBinh 268 194 74 22 ThanhHoa 889 181 533 80 95 23 Nghe An 742 226 298 14 204 24 Thua Thien Hue 719 223 111 83 302 25 DaNang 379 379 26 QuangNam 489 203 186 2 98 27 QuangNgai 327 149 178 0 28 BinhDinh 598 132 189 172 105 29 Phu Yen 275 103 172 0 0 30 BinhThuan 318 152 164 1 1 31 Dac Lac 501 183 318 32 Lam Dong 377 257 120 33 TPHoChiMinh 6212 6212 0 34 BinhDuong 418 406 7 5 35 TayNinh 715 496 219 36 DongNai 1360 352 1008 0 37 Ba Ria-Vung Tau 466 290 176 0 38 An Giang 2278 2278 39 BenTre 689 161 475 53 40 Tra Vinh 465 147 318 41 Soc Trang 275 91 182 2 42 Kien Giang 758 312 397 49 43 Bac Lieu 435 433 2 44 Ca Mau 608 246 362 TOTAL 29400 18359 7999 876 2166 Source: Dung 1999. 120 VI. HUMAN RESOURCES IN HEALTH As in the case of health facilities, Vietnam has been successful in establishing an extensive network and plentiful supply of health personnel. In 1997, there were a total of 213,099 health workers in the public sector, with the vast majority (84 percent) employed at the local level. As Annex Table A6.1 shows, there are several categories of health workers, although a few types predominate-viz., assistant doctors (constituting 23 per- cent of all health workers), doctors (15 percent), second-degree nurses (11 percent), and elementary nurses (10 percent). There is data on private sector workers but this is incom- plete, as only those with a license are known and shown. A. Supply of Health Workers Trends over Time. There has been a major expansion in the number of doctors since 1976, with the availability of doctors per 100,000 population increasing from 19 in 1976 to 44 in 1996 (Figure Rguie 6.1: Numberoflhasthwoikenper]0,000population,byworker 6.1). However, the change category, 1976-96 in the number of assistant 1413 doctors and pharmacists has not been as impressive, and 1 the number of nurses60 and . 1] 119% midwives has actually fallen i in relation to population 8 size. Of particular concern 8 is the 57 percent decline in the number of nurses per 4 100,000 persons between 19 th ube fnusspe 2 0 17 1986 and 1996. The fact that the Doctors Ass't doctors Pharmacists; Nurses MidvAves numbers of health workers in different categories have changed at very different rates over the last two decades im- plies that mix of workers in the health sector has altered qualitatively. The ratio of nurses to doctors, which was 7 in 1976 dropped to 4 in 1986 and has dropped further to 1.3 in 1996. Likewise, the ratio of assistant doctors to doctors has fallen from 2.9 to 1.4 over the same period. As doctors are significantly more expensive to train than assistant doctors and nurses, it is not clear that the evolving mix is the most appropriate and cost-effective mix of health workers for the country. 60 The number of nurses includes second-degree nurses and elementary nurses. 121 International Compari- Hgure 6.2: Doctors perl10,000 population, sected As a countries, 1993 sons. The adequacy of the total 147 number of health workers in 12 Vietnam can be gauged by com- 1 paring its position with that of other countries in the region. Fig- ure 6.2, which shows the number of medical doctors per 100,000 persons in selected Asian coun- tries, indicates that Vietnam 2 compares very favorably with5 other countries in the region in terms of physician availability. c 01P Indeed, even countries that have a significantly higher per capita Hgum 6.3: Nums per10,000 popul6r; selected Asian countres, 1993 GNP than Vietnam, such as Ma- laysia, Thailand and the Philip- 416 pines, have fewer doctors per 100,000 population. Vietnam appears to be less favorably placed with re- spect to the per capita availabil- ity of nurses (Figure 6.3). With 136 58 nurses per 100,000 persons, 8 Vietnam has fewer nurses in re- - 32 lation to its population than Ma- laysia (160), Thailand (99) and China (88). But Vietnam's ratio of nurses to population is greater than that of the Philippines (43), Rgure 64 tio oftrumsto dactors, selcted Min countrs, 1993 Pakistan (32) and Bangladesh (5). The relatively large sup- ply of doctors and smaller avail- ability of nurses results in a low nurse/doctor ratio. Vietnam has one of the lowest ratios of nurses to doctors in the region-lower 1 3 than that in Indonesia, Sri Lanka, Thailand, the Philippines, Malay- L5 sia, Cambodia and Myanmar (Figure 6.4). ii 13 Provincial Distribution. The distribution of health person- V. 'W nel around the country is not ex- actly as planned. More doctors 122 and other higher-level staff are available in the cities, where there are generally enough to fit the plan, while in the rural areas the numbers are too few to fill the planned positions. In provinces with a medical school there are more or less sufficient staff, but in poorer provinces, especially the newer ones with no secondary medical school of their own, there are not enough (Annex Table A6.2). Table 6.1 indicates the regional disparity in the availability of health workers, especially doctors. While a commune in the Southeast Re- gion has an average of 6.8 doctors, there are only 1.7 doctors per commune in the North Highlands and the North Central Coast. Table 6.1: Number of commune, district and pro%incial-level health %orkers per commune, by worker category and region. 1997 Rano of Assistant .411 Pharma- to medi- health Medical .4ssstant Pharm- Nur.Wes to YIs to cal doc- Region workers doctors dcior% acsts Nurses dctors doctors tors Whole country 17 3 2 7 4.5 1 6 3 5 1.29 0.59 1 67 North Highland 108 1.7 3.7 0 5 24 1 39 0.29 2 17 Red River Delta 147 2 8 3.7 08 3.4 1 20 027 132 North Central Coast 11.4 1 7 3 9 06 2 4 I43 0.34 2.33 Central Coast 18.2 2 9 5 5 (9 3S LA.3 0 31 1 85 Central Highland 13.8 2.4 3 6 0.7 3.2 1 36 030 1 52 Southeast Region 37.3 6.8 5 7 I 7 90 1 32 0725 084 Mekong Rbver Delta 199 3.2 6.7 1.1 40 1 23 036 2 10 Source- MoH 1997a In 1997, 67 percent of health workers were paid by provincial budgets, 28 percent by district budgets and 5 percent by commune budgets. The Ministry of Health has a tar- get for the year 2000 of 40 percent of all commune health centers in the country having a doctor. The present ratio varies greatly among provinces, obviously higher where the conditions are more attractive and the supply of staff greater, such as the southern prov- inces. In the Southeast Region, for example, nearly 60 percent of communes have a doc- tor, while in the Northern Highlands the proportion is only 10 percent (Table 6.2). Most of these doctors are the 'irregular' or 'upgraded' doctors, who first were assistant doctors, then after some years of working followed a further three-year study to become doctors. There are very few 'regular' doctors who followed the six-year curriculum working at commune health center level. This is not necessarily a bad thing, since the upgraded doc- tors may have a more appropriate training and experience to do the work well at that level. 123 Table 6.2: Staffing at commune health centers, by region, 1997 Percentage of communes without a- midwife or obstetric- pe- diotric assis- assistant doc- Region fant doctor medical doctor tor Vietnam 25.59 76 74 3.63 Northern Highlands 34.78 9088 5.53 Red River Delta 19.65 68.99 0.51 North Central Coast 28.56 84.49 3.53 Central Coast 19.32 81 61 4.86 Central Highlands 43.45 70.38 12.93 Southeast Region 19.42 41.73 1.05 Mekong River Delta 11.61 66.05 I 07 Source MoH 1997a. Two factors play a role in staffing in the periphery: whether the service can pay enough staff to fit the Ministry of Health formula, and whether they can obtain the staff, regardless of ability to pay. Poor provinces and remote areas find it difficult to recruit enough staff to meet Ministry of Health guidelines even if they have funds, as health workers from outside are often unwilling to move to these areas. There are generally adequate (in relation to Ministry of Health guidelines) num- bers of health workers at the provincial level, although some provinces do not have enough specialists. At the district level, there are generally not enough doctors with spe- cialization in priority areas, such as obstetrics and gynecology or emergency surgery. At the commune level, there are not enough doctors (nearly all of those at the commune health center level are the upgraded former assistant doctors) and often not enough staff with other training either, except in the densely populated delta areas and near the cities. For example, in 1997, 26 percent of communes were lacking an obstetric-pediatric assis- tant doctor or a midwife (Table 6.2). Ministry of Health policy requires one of these in all communes owing to the high priority accorded to maternal and child care at the commu- nal level. Private sector health workers. The lack of detailed information on the size and activities of the private sector makes it difficult for the Ministry of Health to consider that aspect in its planning. A Ministry of Health survey of selected communes in 1996-97 showed that there were, in fact, more private than public sector health workers at the commune level in the North and the South (Table 6.3). Of course, some of the private health workers could have been public health workers engaged in private medical practice outside of their of- fice hours. But to the extent that some of the private health workers were full-time private providers, the data in Table 6.3 suggest that the North and the South have more of both types of workers, while the poorer and more remote Center has fewer of both types of workers. This in turn implies that the public sector does not take account of the supply of private health workers in making its own staff allocation decisions. 124 Table 6.3: Average number of public and private sector health workers per commune, by region and worker category, 1996-97 North Center South Category Public Private Public Private Public Private Doctor 0.3 0.5 0.1 0.2 0.4 1.2 Assistant doctor 3.3 2.5 2.3 0.7 2.6 1.6 Nurse 1.1 1.7 1.1 1.4 1.1 1.8 Midwife 0.7 04 0.7 0.5 0.9 0.5 Traditional medicine 0.1 0.6 0 1 06 0.1 0 5 doctor Others 0.1 0.5 0.1 0.3 0.1 09 Total 5.5 6.0 4.4 3.7 5.3 6.6 Source: MoH 1998d. Gender Distribution of Health Workers. In Vietnam, 50 percent of the labor force is female. In the health sector, the ratio of females is higher, viz., 63 percent. Still, the proportion of women in leading positions in the health service is rather low, in part be- cause there are fewer women than man with higher academic degrees. In institutes such as the national or specialized research institutes, many of the staff are female. For example, in the Institute of Hygiene and Epidemiology of Tay Nguyen, just over 50 percent are female, and among heads and vice-heads of depart- ments, half are female. A few national institutes are even headed by women but there are more women in vice-director positions. Fore example, the Pasteur Institute in Ho Chi Minh City has two female vice-directors. The report from Vietnam at the Knowhow Con- ference in Amsterdam in August 1998 describes the information and communication to support women in the National Assembly elections. It notes the decline in the numbers of women in the past two decades from the relatively high level at the end of the war in 1975. Ethnic Considerations in Health Personnel. Ministry of Health policies for health development in the mountainous regions where most ethnic groups live are guided by the policies for minority groups of the Party (Ministry of Health Strategy for People's Health Care, June 1996). The representation of ethnic minorities among health personnel varies greatly in different provinces. In some provinces there are very few people belonging to ethnic mi- norities, such as some of the delta provinces. In others, there are some ethnic groups but in small numbers. In Ninh Thuan Province, for example, there are four ethnic groups, with three minorities comprising about 20 percent of the total population. The Cham group (10 percent of total) is concentrated in lowland districts and is well represented among health staff in those districts; in fact, is in the majority. There are also 12 Cham doctors among the 80 in the provincial hospital. The other two groups are in the more remote areas, have lower educational levels and are very few among the health workers. In other provinces the same sort of picture emerges: some ethnic groups seem to be better integrated into the Kinh society and to be represented among health personnel, while oth- ers are highly marginalized. Although there is concern among Kinh health workers and leaders for the health problems of the socially disadvantaged minorities, communication and education with 125 minorities remain difficult within the constraints of the health and education sectors in all provinces (generally poor ones) in which minority ethnic groups live. B. Productivity of Health Workers Since health workers Figure65: Annual nmber ofpubic-sector health service contacts per public typically perform many health wocker, 1976-98 functions-preventive, deliveries, outpatient 550 curative, and inpatient curative-it is not easy to 4 calculate their productiv- ity. However, to obtain a - - rough idea of how the productivity of health workers in the country has '3 evolved over time, one < 250 . . . can look at the average 2 annual number of service consultations that each 150--------------- S ~eat wokr 1976-98 .3 & worker makes with the population. The results indicate that although worker productivity declined sharply from 1984 to 1990, it has made a remarkable recovery since then. Indeed, worker productivity is at the highest level it has ever been in the last 22 years (Figure 6.5). Between 1990 and 1998 alone, productivity has doubled. These findings reflect the fact that the overall number of public health workers has declined since the early 1990s, but the overall number of service consultations with public health facilities has increased along with the population. Of course, these produc- tivity numbers do not include the preventive and promotive work that public health workers also perform. There also appear FIgure 6.6: Average annual number of service contacts at the commune level per t o b e l a r g e w o r k e p r o -c o m m u n e h ealth w o rk r, b y re g io n , 199 72 . 5 to be large worker pro-2.5 ductivity differences across regions. Figure 6.6 5 1,900 shows the number of 1,700 health service contacts at 1,500 the commune level per commune health worker 1 0 for the seven regions in the country. Productivity 726 7 of health workers is low- est in the Central High- lands, and highest in the Mekong River Delta. In- C' deed, on average, a com- mune health worker in the 126 Mekong River Delta makes nearly four times as many service contacts annually with the population as does a commune health worker in the Central Highlands. The low produc- tivity of health workers in the Central and North Highlands is likely related to low popu- lation density and poor geographical access to commune health centers in these regions. It may also be related to demand, since a large proportion of communes in these two re- gions lack an obstetric-pediatric assistant doctor or a midwife (Table 6.2). If the local commune health center is poorly staffed, patients may see little utility in visiting the fa- cility for treatment. C. Salaries of Health Workers Like all civil servants, health workers are paid very low salaries. In most cases, the salary paid is barely a living wage, with the result that most health workers have to augment their salaries by practicing privately or taking up a second job in another sector. It was estimated, for instance, that 49 percent of all private health workers and 83 percent of all private physicians in 1996 were in fact government employees (see Chapter V.B). Table 6.4 constructs average monthly salaries (including allowances but excluding any bonuses) for health workers over the period 1992-98 from data on total employment in the sector and aggregate expenditure on salaries.61 It is seen that, other than sharp in- creases in 1992 and 1993, average monthly salaries have remained essentially unchanged in real terms since 1994. In 1998, the average monthly salary of a health worker was merely US$29. Table 6.4: Average monthly salaries of health workers. 1992-97 % increase in salary over pre- .4 verage monthly salary vious year No e4"f Aninual nsage constant constant 'mploy- bill (million current 1989 current 1989 )car ecs VND) PND VND US5 VND ; ,VD L'SS 1992 1%2.901 136,254 05 62,080 1.094 5.57 1993 193.933 4I1.436 5 176.795 47.526 16.62 18479 162.66 19844 1994 195.687 649.50z.36 276.503 6S.009 25.20 5645 43 10 51 63 1995 204.t64 771.315.39 314.094 66.111 2830 13.56 -2.79 12 31 1996 212.103 828.97t 71 325.697 64,919 28.32 3.69 -1.80 0.09 1997 213.099 955.051 89 373.477 72.114 28.87 14.67 11.08 1.93 Source Ministry of Healkh and Ministry of Finance estimates Annex Table A6.3 shows the total monthly compensation of 16 types of health workers at the provincial, district and commune level for 1992 and 1997 in Vinh Long province. There are four major findings from these data. First, in the same job classifica- tion, provincial-level workers earn more than district-level workers, who in turn earn more than commune-level workers. The differences in compensation are, however, small. For instance, in 1997, a pharmacist earned VND 337,000 at the provincial level, VND 323,000 at the district level, and VND 308,000 at the commune level. Second, medical Note that these are official salaries and, therefore, do not include any informal (or "under- the-table") payments that public health workers might receive from patients. 127 doctors earn more than pharmacists and assistant doctors, who in turn earn more than nurses and traditional medicine practitioners. Nurses typically earn more than midwives, but not by very much. Third, allowances and other earnings constitute a small proportion of the compen- sation package. In most cases, basic salary constitutes 80-90 percent of total compensa- tion. Fourth, compensation of all health workers increased significantly in real terms be- tween 1992 and 1997. For example, the average monthly compensation of a doctor in- creased nearly threefold in real terms. The compensation of elementary nurses increased by more than two times. However, despite these salary increases, average compensation remains small. For instance, a provincial-level doctor was paid an average of VND 477,000 or US$37 per month in 1997, while a commune-level traditional practitioner was paid an average of only US$18 per month.62 The one exception where bonuses constitute a large share of total compensation is staff of hospitals at central level and hospitals in the more affluent areas. Hospitals are allowed to retain 30 percent of patient revenue to finance their recurrent costs, including staff bonuses, so that bonuses constitute a large share of total staff compensation across central general and provincial specialty hospitals (see section D.3 in chapter IV). D. Training of Health Workers 1. Preservice Training: Planning and Development Most of the training for health personnel concentrates on medical doctors (pro- duced by the eight medical schools), pharmacists (from specialized colleges), assistant doctors, nurses and midwives, produced by the three national and about 50 provincial secondary medical schools. Originally, the plan was to have a secondary medical school in each province, but many new provinces have only a training center or not even that yet. Such provinces send their trainees to neighboring provinces for training. The numbers of doctors, pharmacists and dentists to be trained are determined by the Ministry of Health, based on the needs they perceive, and implemented by the medi- cal schools. The schools do, however, also have some say in how many students they admit each year. The numbers of students admitted to the secondary medical school is determined mainly by the province, according to perceived needs and to available funds. These numbers fluctuate more than do those in the medical schools. Some of the training programs are short (6-9 months, e.g., primary nurses) while others may be up to three years (e.g., for assistant doctors). The provinces make plans and predictions for such staff which are submitted to Ministry of Health for approval. The Ministry of Health also bases its decisions on training health workers on in- formation received through international contacts (conferences in Vietnam and outside, visits to other countries, to WHO and other organizations), advice given by higher au- 62 It is possible that these low official salaries are supplemented by other sources of income, such as accep- tance of 'informal' payments from patients. However, there is virtually no information available on the nature and magnitude of such payments. 128 thorities in the government, opinions of the Party structure from the local authorities up, and suggestions from mass organizations such as the Women's Union or Farmer's Union. For example, one reason for the decreasing utilization of commune health centers in the last decade was thought, partly as a result of research, to be the lack of confidence in the health staff there. One solution to this was the policy to try to have doctors in the commune health centers, with a higher level of training and therefore better capacity to provide the service people want. The difficulty of persuading graduate doctors to spend time in the commune health centers led to the plan to offer upgraded training for assistant doctors. These staff, with a three-year training plus some years of experience, can attend a further three-year training given by seven of the medical schools to become a 'commu- nity' doctor. They are expected to return to their commune health centers, and many of them do. In some provinces most of the commune health centers are already staffed by such doctors, and there is evidence that utilization of those commune health centers does increase. Now there is a policy to reduce the training of assistant doctors, and to replace that with more advanced training for nurses. Both hospital and CHC or public health nurses are needed. At present, the status and training of nurses does not attract enough candidates for training, but change is underway. This is to bring Vietnam more in line with international practices and staffing trends. A need for assistant doctors is foreseen, however, for the coming years in the remoter areas. An example of a policy proposed by the government in the interests of equity (particularly the provision of adequate health care to those in remote areas) is the new program to improve the network of village or hamlet health workers (VHWs or HHWs). There is a need in nearly every province, certainly in those with mountainous and remote areas, for village or hamlet health workers. The tasks of these workers will vary accord- ing to the actual needs in each situation. At present, the training of hamlet health workers is similar to that for primary nurses-viz., a nine-month largely theoretical curriculum; this is recognized to be inappropriate. There is a great need for a curriculum that is suit- able and training methods appropriate to adult education. Trainers are needed who can appreciate the life experience of the trainees and adapt the training to fit with them in a stimulating and constructive way. A system for supervision is also needed, including training of the commune health center workers to supervise village health workers in a supportive manner. The Ministry of Health supports the new move towards health science universi- ties, which include faculties of different health-related fields. In Ho Chi Minh City, the university already exists, and has seven faculties: medicine, nursing and technology, tra- ditional medicine, pharmacy, dentistry, public health and basic sciences. In Hanoi, the Medical College is moving more slowly in this direction, but is now forming a Faculty of Public Health. Also in Hanoi, graduates of the three-year assistant doctor training in a secondary medical school can now follow a new supplementary training to become Bachelor's in Nutrition, which should lead them to relatively attractive jobs such as hos- pital dieticians or food safety experts for food processing companies. Job Opportunities for Medical Graduates. Detailed data on medical graduates, many of whom are working outside the medical sector, are not available, as most schools 129 have not done tracer studies to find out where graduates are working. One such study done informally by Ho Chi Minh City Faculty of Medicine and Pharmacy found that 60 percent of the graduate doctors were in fact working in the provinces they had come from, but the other 40 percent were not traced. They may have been working in the City itself, in other provinces or cities, or in other fields. A certain number of health graduates work under other ministries or industries, which have their own hospitals and health care activities, such as Transport or Mines, not under the auspices of the Ministry of Health but still in the medical sector. In 1997, more than 10,000 of a total of around 190,000 health workers fell under other ministries. The Armed Forces have their own health staff training center and their own facilities. In a 1996 report on the Netherlands' project to strengthen teaching in medical faculties, Abbatt (1996) notes that in the past all graduates of medical faculties were taken into medical employment, mainly in district or provincial hospitals. Young gradu- ates were asked to do service in communes and remote areas for a few years before tak- ing up such posts and opportunities for private practice were limited. By 1995-96, how- ever, the state no longer provided employment for all graduates and service in remote ar- eas had disappeared. The estimates of employment of graduates cited are pessimistic: about 10 percent of graduates would go to government hospitals, a further 10 percent to commune health centers, while a large proportion (40 percent) would be employed by drug companies or in pharmacies. Ten percent would find unpaid work in hospitals in the hope of future employment, 10 percent might do completely different work, and the re- maining 20 percent practice private medicine, perhaps illegally, or be unemployed. These figures were thought to be more valid for the big urban faculties like Hanoi, Ho Chi Minh City and Hue, and less so for the faculties providing for the less developed areas, such as Can Tho, Tay Nguyen and Thai Nguyen. Good data to better describe the fate of graduate medical doctors are not available, although the information given for Ho Chi Minh City and for the Ho Chi Minh City municipal training center shows that it may be less gloomy than the picture painted by Abbatt. It remains true, however, that there are many unem- ployed or underemployed doctors in the urban areas and too few medical professionals, including doctors, in less developed areas, as is the case in almost all countries. 2. Preservice Training: Selection and Quality Monitoring of Education The Ministry of Health and the training institutes themselves are responsible for monitoring the quality of the preservice training, although the Ministry of Education and Training (MoET) also plays a role for the universities. This is done by spot-checking the questions and answers on examinations, from entrance to subject to graduating examina- tions. Site visits by DST staff also help them to identify problems and correct them. The training institutes, in particular the universities, try to improve their selection and exami- nation procedures but are hampered in modernization by the requirements of MoET. For example, the entrance examinations to the Ho Chi Minh City Medical Faculty consist of long-answer written examinations. Since there are thousands of applications for about 300 places, the perusal of the examinations takes weeks, and is delegated to con- tracted staff because the teaching staff simply cannot do it in the time allowed. They would like to replace the present system with a multiple-choice system that could be marked by computer but the validity of that system does not yet satisfy the Ministry. 130 At a seminar on the medical training in Hanoi Medical School in 1992, the con- clusion was that more modification was needed to bring the curriculum into line with the objectives. The main deficiencies in the curriculum were listed as: lack of practical ex- perience for clinical and traditional medicine; overload and irrational distribution of the subjects; outdated, 'teacher-oriented' methods; and outdated assessment methods. In response to this need, the Netherlands began a project to assist four medical faculties in Vietnam (Hanoi, Hue, Ho Chi Minh City and Can Tho) in reorienting the training more to the community, and in improving the teaching and assessment methods used. As noted by one of the participants at the meeting in 1992, " many obstacles are foreseen", including "lack of means, money, poor knowledge of foreign languages, and the inertia among us all to changing the old methods." In spite of this some progress has been made, aided by additional projects in Hanoi and Ho Chi Minh City supported by Sida, to introduce active teaching methods, and by Sasakawa Foundation, to support field teaching. In Can Tho, the medical faculty is involved in a far-reaching process of change to a student-oriented, block system curriculum, as part of a universitywide project sup- ported by the Netherlands. In addition to the products of the MoH/MoET schools, the City Health Bureau of Ho Chi Minh City also has a training center, which accepts students only from the city and its suburbs. They are trained either as medical doctors or as other support staff. This college has more flexibility in its curriculum and materials and attempts to make the training more practical and community-oriented while at the same time technical up-to- date. Many of the teachers in the MoH University of Health Sciences also teach in the municipal training center, and find that in the latter they are more free to develop new methods and materials and, for example, to require the students to work in the rural dis- tricts. Most of the graduates are employed after graduation through the City Health Bu- reau in the hospitals and health centers of the city, those with the best results having first choice at the favorite posts. Because of this relationship with the city, the students can work in city hospitals and can be almost assured of employment as medical doctors upon graduation, in contrast to the graduates of other schools. All agree that too much of the teaching for most health professionals is theoretical and that graduates have too little practical experience. A major limitation for the medical schools is the lack of teaching hospitals; some schools have an arrangement with one or two hospitals for teaching but the relationships depend very much on personal contacts. The secondary medical schools generally have a better working relationship with the pro- vincial hospitals, which also provide the clinical teachers, but the time and the methods are still limited. Sida has also invested in equipment and training courses to modernize the teaching methods in many of the secondary medical schools. The quality of the graduates is monitored by DST and MoET, who spot-check both examinations and results, but they are not really in a position to solve the main prob- lem of lack of practical experience of graduates. Most of the faculties are still developing test banks and just beginning to use modern methods such as structured observation to assess the students. Methods to assess behavior and attitudes are not yet on the horizon as long as assessment of practical skills is still so limited. 131 3. Postgraduate Training Most postgraduate training is reserved for those who graduate from medical facul- ties, although there are additional training opportunities for others, including assistant doctors but also graduates from other faculties such as biology or engineering, who work in medical fields. The traditional system was that doctors from the main faculties would graduate, work for two years, then return to work with the university doctors for one year to be- come 'specialist first degree'. After a further two years working in another hospital and a further one-year with the university, they could become 'specialist second degree'-these qualifications could be achieved in nonclinical fields as well and provided a sort of 'in- service' training. There is now an increasing trend towards Master's degrees in all fields, which seem gradually to be replacing the specialization as a recognized higher qualifica- tion. This could be to the disadvantage of those who had worked for the older type of de- gree, as the new Masters degree seems to have higher status. It may also be more difficult to do the Master's as an in-service degree, although in future they may also be offered as part-time courses. Relatively few go on to complete a PhD in a medical field, and those few are mostly working in the main universities or national research institutes. The regulations for obtaining the PhD in Vietnam have been changed at frequent intervals during the past several years, in an effort by the MoET to improve the quality of the degree. An optimal system has not yet been developed. In the past some were selected to do their higher de- grees in other countries, often Eastern Europe and the Soviet Union, but these opportuni- ties are no longer widely available. Now it is more common to obtain a Master's level degree in the region (the Philippines, Thailand, and Australia) and then do the PhD in Vietnam. Health research in Vietnam is mostly done by staff in universities and research institutes who are either studying for or supervising a post-graduate degree. Graduate students in universities also use the undergraduate students to carry out large-scale sur- veys among workers or in communities. Training in research is obligatory for those doing a higher degree but the content of the training, based on the standardized curriculum from MoET, emphasizes statistics and computer use. There is not necessarily any link between selection of research and the information needed for the development of health policies and strategies for the ministry. 4. Training in Traditional Medicine The health policy and strategies of the Government of Vietnam give recognition and emphasis to the role played by traditional medicine at different levels of the health care system in Vietnam. At the commune health center level, health workers are expected to use traditional herbal remedies and selected nonherbal methods of treatment for com- mon diseases. All health workers receive some training in traditional herbal medicine as well as some nonherbal methods such as acupuncture and massage during their "modern' training. For the six-year medical doctors the training in traditional medicine takes two months and is carried out by the Department of Traditional Medicine in each faculty. There are, however, training programs for traditional medicine itself. 132 For example, students in the regular six-year medical course can choose after the fourth year to specialize in selected areas, one.of which is traditional medicine. These students spend the final two years learning more details and skills in traditional medical practice. Up to 1995, 1348 such doctors were graduated. Traditional medicine is also a recognized subject for specialization and there are more than 300 with the first degree of specialization in traditional medicine and about 50 with the second degree, as well as more than 10 master's and 20 PhDs in this topic. Twenty- two professors (associate and full) have been nominated in traditional medicine. In 1981, a college of traditional medi- cine was established at Thue Tinh, where the secondary medical school had a long tradi- tion of training in this field, training more than 1500 assistant doctors to use traditional herbal preparations and acupuncture. The National Institute of Traditional Medicine and the National Institute of Acu- puncture are the reference centers for research and training in this field, and offers short- and long-term training to medical and traditional practitioners. They have also offered training for traditional medical practitioners in modem medicine. Now in Ho Chi Minh City, the former Department of Traditional Medicine, sited within the Institute for Tradi- tional Medicine, has become a full-fledged faculty within the new University of Health Sciences. Some but not all secondary medical schools also offer training in traditional medicine in some provinces, in collaboration with the Provincial Associations of Tradi- tional Medicine. Still, though, there are not enough trained practitioners of traditional medicine to staff the grass-roots levels of the health services as foreseen by the Ministry of Health. The training in the different medical and secondary medical schools is not harmo- nized and suffers from the same limitations as the other medical training: old fashioned teaching methods, too little practical experience and inadequate assessment tools. The relationship between the modem and traditional medical streams is not always harmoni- ous. Often, in both training and practice, the two systems are parallel and not integrated, although the policy is one of integration at all levels. 5. Village Health Workers Until recently, the training of hamlet health workers was similar to that of primary nurses, namely, a nine-month largely theoretical curriculum offered by the secondary medical schools. It has been recognized that such training may not be the most appropri- ate and that there is a need for a curriculum that is more suitable, using training methods appropriate to adult education. Also, trainers are needed who can appreciate the life ex- perience of the trainees and adapt the training to fit in with them in a stimulating and con- structive way. A system for supervision is also needed, including training of the com- mune health center workers to supervise village health workers in a supportive manner. The experience of the Support for Disadvantaged Areas (SDA) project suggests that selection of the candidates for village/hamlet health worker positions is still difficult. The Ministry of Health ideally would like people who are literate and have a basic educa- tion, but in the most needy areas few of these are available and others, with less educa- tion, are likely able to serve the population well if prepared properly. Their training would, however, have to be adapted to low-literate trainees but that should be possible. 133 Low and illiterate village health workers have been trained in many countries whose rate of literacy is lower than Vietnam's. The need for such appropriate training, with respect to both content and methods, has been emphasized by district and provincial health lead- ers from provinces with large remote areas and ethnic minority populations. It is also helpful to break up the training of village health workers into smaller periods that are short enough that the trainees are not away from home too long and so that the amount of information presented is not too much for them to digest at one time. When they spend time at home between training periods, they can apply what they have learned and identify questions or problems that can then be taught when they return for the next training period. This adds some costs for travel but the improvement in outcome (what they actually learn) makes it worthwhile. The training should not only be technical, but also include how to work with the community, that is, communication and motivation skills. That means that the trainers should also have these skills. There is still a shortage of trainers with appropriate experience and expertise for the training of village health workers. Sometimes, trainers brought in from other areas find it difficult to adapt their ideas about skills and knowledge to the very limited situation in which the village health worker has to work. The lessons from the SDA project as well as from other government and NGO efforts are being used by the MOH to address a number of the problems cited above and to institute a major new program of training and retraining for hamlet health workers. It will be of much shorter overall duration and the training will be broken up into several sessions depending on the needs and wishes of localities. CHC workers will also be trained in methods of supporting the work of hamlet health workers. New training mate- rials have been drawn up and new supervision tools and a supervision program will be put in place after training. These steps should go a long way toward addressing the defi- ciencies in previous hamlet health worker training. 6. In-Service Training There are no regulations for continuing education, although the importance of that and refresher training are recognized. Refresher courses for government staff are usually offered when a national program or a donor can provide funding. Some areas are over- trained, with some individuals spending more time on training courses than at their jobs in a given year. Other areas and staff have not been offered training in decades. A UNICEF survey to assess medical equipment in 392 UNICEF-supported commune health centers found that between 10 and 30 percent of the equipment was not being used. The principal reason given was that the commune health center workers had not been trained to use it, or had received inadequate training (MoH and UNICEF 1996). Considering the changes in the system and the practices during the past ten years in Vietnam, refresher training is essential if quality of services is to be improved. Not only technical training but especially training in management, such as planning and evaluation, are greatly needed at all levels. Many studies have identified the need for continued training of health workers, and many projects address some of these needs to a greater or lesser extent. National pro- grams such as Control of Diarrheal Diseases (CDD) and EPI have provided consistent training to meet their needs, and some of the topics, such as supervision, or monitoring 134 using local data, are common with those of the health system in general. They have also been training for many years and have covered most of the country. Still, systematic training to improve the capacity of all health workers is an enormous undertaking, and can hardly be done, even by large projects, in the way that would bring the best results. Budgetary and time constraints and performance requirements force most donor-funded projects to spend too little time on preparation of trainers and training materials, and too little time on the training itself. One-off courses are most common but without supportive supervision (which also requires preparatory training as well as supervision by the next level) probably have little effect. Again, partly because many Vietnamese trainers are not experienced in practical and interactive teaching, most in-service training is too theoreti- cal and classroom-based and therefore not appropriate for the target groups. A review of the training in one project noted this as a general phenomenon in many projects (Bergstrom 1997). An additional problem is that there is a relatively small pool of trainers with suffi- cient experience and expertise to carry out the training required by many projects. This pool is also mainly based in Hanoi. Support for a continuous education project on interac- tive and participatory training methods could increase the size of the pool and help to de- velop more trainers in other areas of the countries, so that the input into training would be more effective. Since 1976, the Ministry of Health has had its own School of Health Administra- tion, now known as the Hanoi School of Public Health. In the past it mainly provided in- service training in administration to heads of health services and hospitals, mainly in the north of the country. In recent years it has been shifting its aims and its own human re- source development to provide more modem management and research training in public health and now offers one of the few courses leading to a Masters in Public Health in Vietnam. The first class of around 20 students graduated early in 1999 after two years of study. Plans for refresher training of government health workers can be prepared but the funding must come from the local levels, provincial, district, or commune, and these funds are extremely limited. Mostly they prefer to spend any funds they have for health on new facilities or equipment, which are also needed. District and provincial level lead- ers said that they usually did not make training plans because anyway there wouldn't be money to carry them out. The local authorities do contribute to the costs of training assis- tant doctors to upgrade them to doctors, which plays a role in the return of these doctors to their communes and districts. There is a need for structured plans and requirements for refresher training of government health staff, but until funds can be allocated to carry it out the exercise would not be productive. There are no structured programs, requirements or plans for training of private practitioners. Especially in the south, where private practice has been in existence for a longer period, the need to reach these health workers begins to be recognized. For exam- ple, the Pham Ngoc Thach Institute, in the context of the National Tuberculosis Control Program, has begun to invite private doctors to seminars about the new treatment regimes for tuberculosis, and about the need to adhere to them to avoid the increasing problem of drug resistant infections. This approach will become ever more important, both to control 135 diseases of public health importance and to maintain standards of medical care for the public. 7. Future Plans for Training of Health Personnel The Department of Science and Training together with the Departments of Plan- ning and of Personnel and Organization are responsible for planning for the future needs for health personnel. They do this mainly on the basis of their own observations and on the plans submitted by the provinces for multi-year periods. There has been some col- laboration with Sida to use computer-generated projections to predict future needs, but so far the Ministry of Health has not made systematic use of that approach.. According to Ministry of Health plans, every intercommunal polyclinic (ICP) in the Northern Uplands and Central Highland regions should have at least one to two doc- tors and that the number of health workers engaged in preventive health, malaria control and family planning at the district level should be 1.5 times greater than that in the plains districts with equal population (MoH 2000). High-tech medical centers are to be estab- lished in Son La, Thai Nguyen and Quang Ninh provinces, with sufficient staff. Although the plans in these strategy notes are admirable, the time is short to achieve these goals. It is suggested in the document that 'the health sector should come up with feasible solu- tions for sending health staff to difficult areas, so that there will be sufficient health workers in the highland regions by the year 2010." However, there is no discussion of what these feasible solutions could be. The lack of detailed information on the size and activities of the private sector makes it difficult for the Ministry of Health to consider that aspect in its planning. There are health policies to let the private sector develop and do its best in the urban and densely-populated areas and for the Ministry of Health to concentrate its policies and ef- forts on the less privileged areas. For the more remote areas, the Ministry of Health Strat- egy document recommends that private health practitioners be mobilized to do commu- nity health work, which would include curative, preventive and family planning activi- ties. The growing role of health programs and projects funded by international donors has created a need for more management training, for example, in MPH courses. Many staff of health departments, Ministry of Health, universities were sent abroad, but now two courses are already running in Hanoi and others will follow. Both Hanoi and Ho Chi Minh City Faculties of Medicine have now become Universities of Health Sciences, with Faculties of Public Health, showing the increased importance of public health as opposed to medicine in recent years. E. Human Resource Management Staffing. Strategies for providing health care at the different levels include details on the planned staffing at each level. The job descriptions are not, however, fixed, and there are no fixed targets for service provision. One post may be filled by different types of staff, for example, at a commune health center the head may be a doctor, an assistant doctor, or a nurse, or in some cases another kind of health worker. Although there are common features in the job description of CHC head for each of them, there are differ- ences in their functions because of their different training, so the job description is not 136 consistent or uniform in different locations. There are regulations as to the numbers and types of staff at each health facility but in many cases there are neither the funds nor the personnel available to meet the prescribed numbers. If the numbers are less than planned, then the activities will also be less than planned and will be carried out by those avail- able. Although the Ministry of Health recognizes the need to strengthen the number and quality of staff providing care in the context of the official policy of Primary Health Care, it is still searching for strategies within its own financial and human resource capacity that can achieve what is desired. An official circular provides the basis, but has no power to assure the implementation (MoH circular 07BYTTT). Performance monitoring, reward and punishment, and motivation. Monitoring of performance and service delivery is mainly by reporting (written reports sent up through the system) and by supervision carried out by higher levels, with their reports. The Peo- ple's Committees play an important role in monitoring the quantity and quality of care delivered especially at commune and district levels, but also at provincial level. As there are no targets or quotas, the evaluation is based on standards of behavior, presence, and appearance of serious working attitude, among other things. Reward and punishment according to performance are not very important in the health services. As in many countries, only really bad acts can lead to dismissal, for ex- ample stealing drugs or money. Dismissal is a rare occurrence. Rewards for good per- formance are also limited. Small amounts of money can be awarded, small raises in sal- ary or raises earlier than what is automatic, for example. Certificates presented by the People's Committees are often given for recognition of good service under difficult con- ditions; in the past these were more highly regarded than at present. Opportunities for training may also serve as rewards. Eventually, good performance can, for the minority, lead to promotion, for example, to director of a hospital or department head, and to a sen- ior position in the province. Working for the state also makes it easier to obtain a license for private practice, which acts as a motivator. Most health leaders note that the quality of the work delivered by the health workers is lower than what is desired, and that the quality of the training is also less than what they need. The low salaries and poor working conditions are not motivating. Virtu- ally all health workers have to have additional income and many health staff at local lev- els have small market gardens, keep pigs or chickens, and find other ways to supplement their incomes. 137 CHAPTER VI ANNEX TABLES Table A6.1: Numbers of health personnel in the public sector, by category, 1997 of which: Central Other Category of health worker Total level Local level branches Total 213,099 24,587 178,477 10,035 1 Medical PhDs 483 395 85 3 2 Pharmacy PhDs 123 113 10 - 3 Medical Doctors 33,518 4,099 27,624 1,795 4 Pharmacists 5,283 1,441 3,597 245 5 Assistant Doctors 48,459 423 46,047 1,989 6 2nd Degree Medical Technicians 6,221 1,263 4,891 67 7 Assistant Pharmacists 6,485 657 5,575 253 8 2nd Degree Pharm. Technicians 1,655 541 969 145 9 2nd Degree Nurses 22,672 2,809 16,364 3,499 10 2nd Degree Midwives 8,563 224 8,257 82 11 Elementary Nurses 20,768 475 19,350 943 12 Elementary Midwives 4,479 18 4,442 19 13 Laboratory Technicians 1,760 79 1,608 73 14 Traditional Medicine Practitioners 415 105 278 32 15 Elementary Pharmacists 8,067 651 7,101 315 16 Other Bachelor Degrees 4,224 1,875 2,247 102 17 Other 2nd Degree Level 6,038 1,110 4,830 98 18 Others 33,886 8,309 25,202 375 Source: MoH 1997a. 138 Table A6.2: Number of commune, district and provincial-level health workers per commune, by category and province, 1997 Ratio of: Assistant All Nurses Pharma- to medi- health Medical Assistant Pharma- to doc- cists to cal doc- No. Region workers doctors doctors cists Nurses tors doctors tors Whole country 17.3 2.7 4.5 1.6 3.5 1.29 0.59 1.67 I Northern Uplands 10.8 1.7 3.7 0.5 2.4 1.39 0.29 2.17 1 CAOBANG 9.5 1.5 3.0 0.5 2.0 1.30 0.32 1.99 2 LANG SON 8.6 1.4 2.7 0.3 1.7 1.26 0.25 2.02 3 SON LA 12.5 1.1 3.9 0.8 3.4 3.06 0.68 3.49 4 LAICHAU 10.7 1.1 3.8 0.4 2.8 2.51 0.32 3.49 5 LAO CAI 9.4 1.2 2.4 0.5 3.3 2.72 0.46 2.00 6 YEN BAI 11.2 1.9 3.2 0.6 2.8 1.47 0.32 1.73 7 QUANG NINH 13.8 3.0 3.3 0.5 3.2 1.08 0.16 1.10 8 HA GIANG 8.5 0.8 3.5 0.3 2.0 2.52 0.42 4.47 9 TUYEN QUANG 11.0 2.0 4.5 0.6 1.4 0.71 0.30 2.22 10 BACKAN 7.5 1.0 3.0 0.3 1.5 1.40 0.25 2.90 11 THAINGUYEN 12.7 3.1 3.9 0.5 2.6 0.82 0.16 1.23 12 PHU THO 10.2 1.8 4.0 0.4 2.1 1.19 0.23 2.22 13 VINI PHUC 10.8 1.8 4.1 0.6 2.0 1.10 0.34 2.26 14 BAC GIANG 12.5 2.0 4.6 0.5 2.5 1.22 0.27 2.26 15 BAC NINH 13.9 2.7 4.5 0.7 2.8 1.06 0.26 1.69 16 HOA BINH 10.2 0.6 4.7 0.5 2.0 3.18 0.84 7.61 II Red River Delta 14.7 2.8 3.7 0.8 3.4 1.20 0.27 1.32 17 HA NOI 24.1 6.0 2.8 1.2 6.7 1.13 0.19 0.46 18 HA TAY 13.5 2.1 5.5 0.5 2.3 1.13 0.27 2.65 19 HAIDUONG 12.7 2.2 3.5 0.6 2.3 1.04 0.29 1.61 20 HUNG YEN 11.6 1.9 4.1 0.5 2.4 1.23 0.26 2.16 21 HAIPHONG 30.9 6.0 5.3 1.7 8.6 1.43 0.29 0.88 22 THAI BINH 12.5 3.2 2.8 0.8 2.5 0.78 0.24 0.89 23 HA NAM 7.6 1.2 2.2 0.7 1.8 1.48 0.54 1.76 24 NAM DINH 32.0 7.0 7.4 1.9 8.2 1.18 0.27 1.07 25 NINH BINH 14.1 2.4 4.8 0.9 2.6 1.08 0.35 1.99 III North Central 11.4 1.7 3.9 0.6 2.4 1.43 0.34 2.33 26 THANH HOA 10.4 1.6 4.6 0.4 1.4 0.88 0.27 2.82 27 NGHEAN 13.3 1.7 3.7 0.8 3.8 2.29 0.45 2.24 28 HA TINH 11.1 1.3 3.8 0.5 2.4 1.75 0.36 2.82 29 QUANG BINH 10.5 1.0 3.6 0.8 2.2 2.16 0.81 3.59 30 QUANG TRI 11.0 2.0 2.4 0.6 2.8 1.44 0.29 1.21 31 THUA THIEN HUE 12.1 2.8 3.2 0.5 1.9 0.68 0.18 1.13 IV Central Coast 18.2 2.9 5.5 0.9 3.8 1.30 0.31 1.85 32 DA NANG CITY 46.1 11.7 7.0 2.5 10.0 0.85 0.21 0.60 33 QUANG NAM 15.4 2.3 4.7 0.8 3.6 1.54 0.35 2.02 34 QUANG NGAI 13.8 1.8 4.5 0.6 3.2 1.80 0.32 2.50 35 BINH DINH 18.4 2.8 5.5 0.8 3.8 1.34 0.28 1.94 36 PHU YEN 17.4 3.0 5.3 1.0 3.4 1.16 0.34 1.79 37 KHANH HOA 16.6 2.6 4.6 0.9 3.6 1.36 0.34 1.75 38 NINH THUAN 20.9 2.6 7.6 1.4 4.0 1.52 0.51 2.86 39 BINH THUAN 19.4 2.8 7.7 0.9 3.2 1.12 0.31 2.72 V Central Highlands 13.8 2.4 3.6 0.7 3.2 1.36 0.30 1.52 40 GIA LAI 12.3 1.4 2.6 0.6 3.7 2.62 0.43 1.83 41 KON TUM 12.6 1.6 2.8 0.7 4.5 2.88 0.42 1.78 42 DACLAK 13.6 3.1 3.9 0.6 2.6 0.84 0.19 1.27 139 Table A6.2: Number of commune, district and provincial-level health workers per commune, by category and province, 1997 (cont.) Ratio of Assistant All Nurses Pharma- to medi- health Medical Assistant Pharma- to doc- cists to cal doc- No. Region workers doctors doctors cists Nurses tors doctors tors 43 LAM DONG 16.8 3.0 4.8 1.0 2.7 0.92 0.34 1.63 VI Southeast 37.3 6.8 5.7 1.7 9.0 1.32 0.25 0.84 44 HO CHI MINH CITY 62.2 11.8 6.0 2.5 15.8 1.34 0.21 0.51 45 BINH PHUOC 14.6 2.0 4.9 0.7 3.0 1.46 0.35 2.43 46 BINH DUONG 20.4 3.1 5.5 1.7 3.4 1.12 0.57 1.78 47 TAY NINH 25.8 4.1 8.7 1.4 5.4 1.31 0.34 2.11 48 DONGNAI 21.6 3.6 4.8 1.1 5.2 1.45 0.31 1.33 49 BA RIA VUNG TAU 20.7 4.8 4.0 0.8 4.6 0.96 0.16 0,83 VII Mekong River Delta 19.9 3.2 6.7 1.1 4.0 1.23 0.36 2,10 50 LONG AN 14.0 2.1 4.6 0.7 2.8 1.36 0.32 2.25 51 DONG THAP 16.7 3.3 5.4 0.8 3.2 0.96 0.24 1.60 52 AN GIANG 26.2 4.7 8.1 2.0 5.7 1.22 0.43 1.73 53 TIEN GIANG 16.4 2.6 6.0 0.6 3.4 1.30 0.24 2.27 54 BEN TRE 15.3 2.3 5.1 1.0 3.6 1.54 0.44 2.20 55 TRAVINH 19.2 2.4 6.7 1.2 4.5 1.90 0.52 2.86 56 VINH LONG 15.7 2.4 6.2 0.7 2.7 1.13 0.29 2.56 57 CAN THO 32.8 5.9 8.7 2.1 7.1 1.19 0.36 1.47 58 SOC TRANG 18.7 2.5 6.2 1.3 4.1 1.65 0.54 2.48 59 KIEN GIANG 23.6 3.9 8.7 1.3 3.9 1,00 0.34 2.22 60 BAC LIEU 28.3 4.2 11,2 1.3 4.7 1.13 0.30 2.68 61 CA MAU 27.7 4.3 10.5 1.5 3.5 0.81 0.35 2.44 Source: MoH 1997a. 140 Table A6.3: Total monthly compensation of provincial, district and commune-level health workers in Vinh Long province, 1992 and 1997 (in '000 constant 1997 VND) Province level District level Commune level Allow- Allow- Allow- Type of health worker Salary ance Other Total Salary ance Other Total Salary ance Other Total 1992 Medical doctors 53 69 27 149 53 69 27 149 53 69 27 149 Pharmacists 54 71 27 152 Assistant doctors 47 60 27 134 45 59 27 131 42 54 21 118 Second degree medi- caltechnicians 42 54 21 118 42 54 21 118 Degree assistant phar- macists 47 60 27 134 42 54 21 118 39 51 21 112 Second degree phar- macy technicians 39 51 21 112 Second degree 47 60 27 134 Second degree mid- wives 39 51 21 112 42 54 21 118 42 54 21 118 Elementarynurses 42 56 21 119 41 53 21 115 39 51 21 112 Elementary midwives 41 53 21 115 44 57 21 122 44 57 21 122 Lab. technicians 38 50 18 106 38 50 18 106 Traditional medicine practitioners 51 66 27 145 41 53 21 115 Elementary pharma- cists 47 60 27 134 39 51 21 112 38 50 21 109 Other bachelor degree 56 72 27 155 Other second degree level 44 56 21 121 44 57 21 122 Other 42 54 21 118 44 57 21 122 1997 Medical doctors 477 0 0 477 372 0 0 372 Pharmacists 308 29 0 337 301 22 0 323 289 19 0 308 Assistant doctors 319 36 0 355 314 36 0 350 301 22 0 323 Second degree Medi- cal technicians 322 36 0 358 284 36 0 320 Degree assistant phar- macists Second degree phar- macy technicians 302 0 0 302 278 0 0 278 226 0 0 226 Second degree 263 29 0 292 252 29 0 281 238 20 0 258 Second degree mid- wives 238 0 0 238 266 0 0 266 255 17 0 272 Elementarynurses 266 0 0 266 311 0 0 311 262 14 0 276 Elementary midwives 250 0 0 250 Lab. technicians Traditional medicine practitioners 302 0 0 302 244 0 0 244 231 0 0 231 Elementary pharma- cists 352 0 0 352 Other bachelor degree 287 0 0 287 292 0 0 292 Other second degree level 285 0 0 285 274 0 0 274 Source: Data provided by the Ministry of Health. 141 VII. FINANCING OF HEALTH EXPENDITURE A. Sources of Financing 1. Structure of Health Financing in Vietnam In Vietnam, the public health budget contains four fiscal transactions by levels of government: the central and the three levels of local governments: provincial, district and commune. Since communes are not required to report their health spending to the Minis- try of Health, a comprehensive database which aggregates the four level transactions into a consolidated account of sectoral expenditure is lacking. However, central, provincial and district governments account for about 90 percent of the national budget. The Ministry of Health has little role in deciding the overall budgetary allocation to the health sector. Rather, it is the Ministry of Finance (MoF) and the Ministry of Plan and Investment (MPI) that play deciding roles in this process, albeit in discussion with the Ministry of Health. The overall level of recurrent spending on health is determined by projected growth rates of total revenue and of total recurrent expenditures (after subtract- ing estimated expenditures for new policies, such as increased salaries, etc.) and the share of health in total recurrent expenditures. This type of incremental budgeting pays insuffi- cient attention to health goals and health priorities set by the Ministry of Health. The overall budgetary resources to the health sector are allocated on the basis of certain norms across two dimensions: functional (curative versus preventive) and sec- toral/level (Ministry of Health, other sector ministries, and the provinces). The norms are based either on the number of hospital beds belonging to a ministry or province (for cura- tive expenditures) or on the population of the province (for preventive expenditures). These amounts are supposed to cover salaries for a centrally determined number of health workers in the provinces, fees and drug exemptions for the poor, and other recurrent eK- penses. The allocation norms vary across five geographical regions, presumably to cap- ture variations in the need for health services in different geographical regions. In addi- tion, Hanoi and Ho Chi Minh City enjoy a norm that exceeds those of other cities by 50 percent as they have regional and national functions as well. In general, the norm for curative expenditures is about 3 times as high as that for preventive expenditures, pre- sumably reflecting the higher unit cost of curative interventions. In the annual budget allocation document sent to provinces, only an aggregate re- current budget is provided-not its composition across sectors.63 Thus, provinces allocate budget to each sector, including health, based on the norms issued by Ministry of Fi- nance. For poor province, in general, this is minimum budget that has to be allocated to health. The provinces also can spend on health out of their own revenues, although this 63In fact, provinces are provided with two minimum amounts for spending on education and science and technology and a maximum for spending on administration. Other than these, no sectoral breakdowns are provided. 142 generally happens only in the better-off provinces. 64 The financing of health expenditure out of their own revenues in the better-off provinces is what causes large disparities in government health expenditure per capita across provinces, as better-off provinces are able to spend more on health (and other sectors) than poorer provinces. Although, in principle, provinces have to inform the Ministry of Finance about how they allocated their aggregate recurrent budget, this is an ex post formality. It is not clear how strongly, if at all, the Ministry of Finance can influence provincial spending on health out of the aggregate budget already approved for the provinces. The Ministry of Health has an even smaller role in this process. In fact, provinces are not required by law to even notify the Ministry of Health of their budget allocation for health. As a conse- quence, the Ministry of Health has little information on the provincial health budget. In addition to the budgeted amounts for health, the provinces receive funds from the various national health programs (such as the Expanded Program of Immunization, the National Tuberculosis Program, etc.). Until the year 2000, specific amounts were earmarked for each program, and provinces had no authority to reallocate funds across programs. This resulted in underutilization of some program funds and excess demand for other program funds in the same province. Starting with FY 2000, a new mechanism of fund allocation has been introduced whereby provinces are given block grants for all targets of the national health program and have discretion in allocating the funds across the different programs. In principle, the national health programs should serve as vehicles for providing additional funds for disadvantaged areas and addressing various disease patterns specific to each province. A recent study by a group of Ministry of Health experts, however, found that the allocation of program funds is more or less equally distributed across prov- inces and fails to take into account the specific disease profile of a province. Districts receive most of their support from the provinces, although they, too, may have their own revenue sources. Provinces have almost complete discretion in the re- sources they pass on to district health services. While some provinces split their alloca- tions evenly (i.e., 50 percent for the province and 50 percent for all the districts in the province), other provinces may spend two-thirds of their allocation on nationally deter- mined staff salaries, which leaves little room for discretion. This discretion on part of the province may be the source of even greater variation in per capita resource availability at the district level than that existing at the provincial level. Since the communes are expected to be basically self-sufficient, they are virtually excluded from the integrated state budget. Since 1994-95, however, the government has taken over the payment of salaries to approved staff of commune health centers out of the state (provincial) budget. Virtually all communes supplement their meager budgets through the sale of drugs. They buy drugs from the province or district pharmacy and re- sell it with a markup (typically between 1 and 10 percent). In better-off communes, households are sometimes assessed a "tax," the revenues of which are used to defray the 64A province can retain revenue that it collects in excess of its assigned target. It can choose to spend this retained revenue in any way it wishes, including on health. Likewise, provinces have to reduce spending on health if their revenue performance falls short of the assigned target. 143 operational or capital expenses of the commune health center. This is the reason why commune health centers in better-off communes typically offer better-quality services than those in poorer communes. 2. Public-Private Financing of Health As in most other de- Figure 7.1: Public-private shares in financing total health expenditure, veloping countries, the share 1993 and 1998 of households in total health So r expenditure is high in Viet- nam. In 1993, aggregate 70 public spending on health 60 (by various levels of gov- ernment, donors, and the " Vietnam Health Insurance) 40 was VND 2,167 billion, but 30 private (household) spend- - ing was more than two 20 times this amount (viz., VND 5,051 billion). Thus, _o households accounted for 71 0 - percent of total health 1993 1998 spending, with government, donors and insurance accounting for the remaining 29 per- cent (Figure 7.1). By 1998, although aggregate public spending had more than doubled in nominal terms (to VND 5,620 billion), private spending had increased nearly fourfold to VND 23,153 billion, so that the share of households in aggregate health spending had fur- ther increased to 80.5 per- Figure 7.2: Public spending on health as a percentage of aggregate health cent. spending, selected Asian countries, 1990s Even a public share 80 76 of 29 percent in total health P 60 expenditure in 1993 would 60 54 5 have placed Vietnam among 46 48 49 the six countries in the re- A 37 gion with the lowest public 2 2 sector share in total health spending (Figure 7.2). 2 However, at 19.5 percent, 10 the public sector share in o 4 -T aggregate health spending J in Vietnam is the lowest in the region after Cambodia. The large private share in aggregate health spending comes about not only from the ex- tensive use of private health services for curative treatment by individuals but also from payment of user fees and drug costs at public health facilities. One consequence of this heavy reliance on private financing is that a large number of the poor (those who are un- able to obtain fee exemptions) have to bear the burden of high user charges at public health facilities. 144 3. Sources of Financing Public Health Services Table 7.1, which reports the spending on public health by source over the period 1991-97, indicates large real increases in public spending on health from every source.65 Total public spending on health increased about threefold in real terms over this period. This rate of increase was, however, exceeded by user fee-financed expenditure, which increased fourfold, and central government health expenditure, which increased by more than fivefold. The most impressive increase, however, was recorded by health insurance- financed expenditure, which increased from nil in 1991 to VND 65 billion in 1993 and VND 443 billion in 1997 (in constant 1994 VND). Provincial government and donor health expenditures approximately doubled over the 1991-98 period. Table 7.1: Real spending on public health services. by financing source, 1991-98 (millions of constant 1994 VND) Oersea. Development .4 stysance Total Tpnd- fOD.) ing on puble Central gov- Provincial Health rn- grants and Commune health ser- Year crnmen* governmens surance L.'ser kes loans budeer ViceN 1991 213.391 1.041.160 - 91.278 229.329 60,699 1.635.858 1992 380.140 1,218.594 - 109.422 252,266 71.044 2.031.466 1993 487.217 1,369.229 64,619 114.350 313.864 79.82o 2.429.105 1994 706.191 1.530.958 150.080 110.000 248.305 89.255 2.S34,789 1995 968.612 1.589.498 236,158 192.502 332.425 92.671 3.411.866 1996 825.461 1,808.414 359,904 240.577 246,308 109.864 3.590,528 1997 870.370 2.095.448 383.528 280,443 367,137 118.081 4,115,006 1998 504,156 2,220,244 442.812 304.878 4-0.961 169,377 4.112.427 Mte *ODA grants and loans that are directly provided to the central government have been excluded from the figures in this column. Source Minisir) of Finance estimaten Figure 7.3 shows the Figure 73: Sources of health financing N oftotal expenditure on sources of health financing within the public sector. In 1998, the bulk uhold, M Prmncial BCentral goVt of the public spending on health mDon. M insurance Comnunes was undertaken by provincial gov- ernments, followed by the central government, health insurance, ex- ternal donors and communes. Fig- ure 7.4, which shows the evolving 8u5 a96 shares of each financing source in total public health spending, indi- cates that the share of provincial governments in total financing has fallen, while that of 65 spending on health is defined here as government spending on public health programs. 145 the central government has risen correspondingly, between 1991 and 1998. The most dramatic change has been the rising share of health insurance in financing public health expenditure. Starting from a small base in 1993, it has come to contribute nearly 11 per- cent of public spending on health currently. The share of donor financing of health de- clined from 14 percent in 1991 to only 7 percent in 1996, but has since increased to 11.4 percent in 1998. B. Cost Recovery 1. Policy Basic health care was Figure 7. Sources of funding public expenditure on health, 19"1-98 formerly free for all. In 1989, a 0 F fee system was introduced in the three higher (district, provincial 7 9 and national) levels of the health 66 care delivery system, requinng E those patients who are able to do so to pay at least a minimal part of their health care. The handi- 4 capped, orphans, families of 44 health personnel, individuals able 2 471 to produce certification of indi- gency from the Commune Peo- ple's Committee, and patients suf- 999 fering from mental problems, lep- I Local bade& 0 Cental budget ] Heals rosy, and BK-positive tuberculo- a Userfees 0 ODA grants and am U Commune budget sis are to be treated free of charge. In 1995, the Ministry of Health issued user fee schedules for each type of consul- tation and each type of diagnostic test and procedure performed in clinics and hospitals. For inpatient services, there is an additional daily bed charge. The user fees indicated in the Schedule vary by hospital level (i.e., first-class hospital, second-class hospital, fourth- class hospital and polyclinic, etc.). In addition, the schedule specifies a range of charges-not a single fixed charge-for each type of service. (See the Annex Tables for an illustrative list of user fees.) For example, a full medical examination for issuing an employment eligibility certificate to an individual costs VND 25-50,000 at a Class I hos- pital, VND 25-40,000 at a Class ss hospital, and VND 18-35,000 at a Class III hospital. Most fees are required to be paid in advance by noninsured patients and patients not eligible for fee exemption. For the most part, patients are responsible for the purchase of drugs themselves, either from a private pharmacy or a pharmacy run by the public fa- cility on the public premises. There are two important things to note with regard to user fees. First, because the price schedule issued by the government only indicates a range of fees that can be 66 Primary health care at commune health centers continues mostly to be free, but fees are paid at inter- communal polyclinics. 146 charged, hospitals have some discretion over the user fees they can charge, especially for technical services and laboratory tests. Central hospitals normally apply the ceiling price indicated in the price schedule, while lower-level hospitals charge the floor price. There are also differences in fee levels across regions. In the South, user fees tend to be higher than in the North for comparable services. Second, the user fee Figure 75: User fee ections, 1"1-97 schedule that is in use today dates back to 1995, even .... -- though the general price level has increased by about L' . ... .. I5. r.-'IU ..I.r o 23 percent since then. Total .... user fee collections have grown rapidly over time (Figure 7.5), in part because . of an increase in the number of patients seen at public facilities but also in part be- .P , cause of an increase in the ..-.-. ...----. - - - - number of laboratory and 1991 1992 1993 1994 1995 1996 1997 radiology tests performed per patient. Because hospitals have greater discretion in charging private patients for tests, it is likely that the number of superfluous tests performed on patients has expanded. There are guidelines on how the user fees can be used, especially in hospitals. Of the total user fee collection, 70 percent is supposed to be used for improving the supply of consumables (drugs, blood supply, chemicals, X-ray materials, etc.) and medical equipment in the facility, 25-28 percent should be used to reward health workers in the form of bonuses, and 2-5 percent should be returned to upper management authorities to build a supporting fund for hospitals. Hospitals are not supposed to use any of the user fee proceeds for construction activities. 2. Cost Recovery Rates User fee collections have increased dramatically over time, but the acceleration has been particularly rapid since 1994 (Figure 7.6). While user fee collections increased at a compounded rate of 14.4 percent from 1991 to 1994, they increased at a rate of 39.5 percent between 1994 and 1997. The increase in the ratio of user fees to total public spending on health has been less meteoric. Indeed, the share of user fees in total public spending on health declined from 1991 to 1994, but has since increased (Figure 7.5). In 1998, user fees accounted for 7.4 percent of total spending on public health services. 147 Since user fees are Figure 7.6: The share of hospital user fees and state budget subsidies in largely collected by hospitals, tam hospital revenue, 1"4-9 76 the cost recovery rate of 7.4 per- cent, as calculated above, under- 70 Userfec estimates the true role of user 1 54 55 fees in financing hospital activi- 52 ties. In the last five years, user 47 fees have become a very impor- 0 - tant source of hospital financing, 30 accounting for nearly a third of 23 all hospital revenue in 1998 20 (Figure 7.6). This has allowed hospitals to significantly reduce their reliance on the state budget 1994 995 1996 1997 1998 (from 76 percent of total revenue in 1994 to 47 percent in 1998). 3. Impact of Cost Recovery on Utilization of Health Services The debate on cost recovery in the health sectors of developing countries has been ongoing for more than two decades. Proponents of greater cost recovery base their rec- onmendations on the findings of several empirical studies that suggest that the demand for health care in developing countries is price inelastic. It is also argued that, unless user fees are charged in secondary and tertiary hospitals, people will overuse these facilities and underutilize primary health centers that offer care for free or for low prices. Another argument is that user fees generate much-needed revenue for public facilities, thereby allowing these facilities to improve the quality of their services. The introduction of user fees also imposes greater accountability to consumers on the part of governent health facilities. On the other hand, opponents of the cost recovery argument contend that raising fees will reduce access to care, especially by the poor, and consequently affect health status adversely. Opponents also argue that health is a public good with positive external- ities (i.e., the benefits of improved health extend beyond the immediate consumer to soci- ety at large), so that pricing health care leads to suboptimal utilization of health services from a societal perspective. Unfortunately, the existing literature on the demand for health care in developing countries does not offer clear evidence regarding the price response of user fees on health services utilization. One strand of literature suggests that prices are not important deter- minants of health care utilization. Akin and others (1984, 1986), Schwartz, Akin, and Popking (1988), Birdsall and Chuhan (1986), and Heller (1982) all report very small and sometimes positive price effects, most of which are statistically insignificant. Another strand of work by Mwabu (1986), Gertler, Locay, and Sanderson (1987), Alderman and Gertler (1988), Gertler and van der Gaag (1990), and Deolalikar (1997) concludes that prices are important. This suggests that the findings of the studies are country-specific and cannot be generalized. 148 Data on annual health Figure 7.7: tilization rates at public facilities, 1976-98 3.0 140 service contacts per capita-a commonly used measure of I Nr.'r -1 Lt. 120 overall utilization of health ser- 25t vices-show that utilization .00 rates in Vietnam began a steep 20 A / pa 0 80 decline in 1984-several years before user fees were promul- gated, let alone implemented. Indeed, annual per capita con- <40 tact rates have improved sig- nificantly since the early 1990s, 20 which is when user fees began to be implemented (Figure 7.7). However, inpatient utilization rates have continued to fall since 1980, except for a brief period of recovery in 1992-93. Figure 7.7 does not indicate any acceleration of the decline in inpatient utilization rates since 1989, suggesting that the decline in the utilization of inpatient services probably has other causes. The debate on the appropriateness of cost recovery in the health sector is some- what academic in the Vietnamese context. User fees in public health facilities have been in place for 10 years now, and any adverse effects of these fees on utilization, if they did at all occur, have most likely passed. 4. Burden of User Fees on the Poor An important element of any sound cost recovery scheme is the protection of the poor and the vulnerable from the burden of high user fees for health care. Exempting the poor from user fees in public health facilities can help improve access of the poor to health services. It is, therefore, important to examine how user fees have affected the poor in Vietnam. The private cost of health care in Vietnam has already been discussed at length in chapter III.D. Therefore, only the major findings relating to the cost of publicly provided health services are summarized here. Out-of-Pocket Costs. Quite surprisingly, average out-of-pocket expenditure per health service contact declined in real terms for most public providers between 1993 and 1998. The main reason for this was that the prices of drugs, which account for the bulk of out-of-pocket costs at health facilities, fell over this period (primarily owing to deregula- tion of the pharmaceutical industry, which in turn improved the supply of drugs). Despite this decline, average out-of-pocket costs for a health service contact, especially for hospi- tal care, remain large and unaffordable for the poor. For instance, even a single visit to a public hospital takes up 22 percent of all nonfood expenditure for a year for a typical per- son in the lowest quintile. A visit of standard (national average) quality is even less af- 149 fordable; it constitutes 45 percent of annual nonfood expenditure for an individual in the lowest quintile.67 The analysis in chapter III.D indicates that health care costs at high-quality facili- ties (such as hospitals) relative to discretionary income are significantly greater for the poor than for the better-off. For instance, for services of identical quality, the lowest quintile pays 45 percent of its nonfood expenditure per capita in a single visit to a public hospital, while the highest quintile pays only 4 percent. The difference in affordability across the lowest and highest quintiles is equally large even at commune health centers- facilities that are primarily intended for the poor. These large differences indicate that the burden of user fees falls unevenly on the poor in Vietnam. Fee exemptions. An important element of cost recovery is to make sure that the poor and vulnerable are exempted from the burden of user fees. As in other countries, there is a formal mechanism in Vietnam for exempting certain classes of people, such as the poor, handicapped, war veterans, orphans, and individuals suffering from certain ail- ments (such as tuberculosis and leprosy), from user fees. In addition, children are sup- posed to receive free of charge services provided through the vertical programs, funded directly by the central government or aid agencies. There are two types of exemptions that are typically given: exemption from com- mune health center charges (typically for drugs) and exemption from district hospital charges. The former are handed out very sparingly, if at all. In their survey of 32 com- munes in 1995, Ensor and San (1996) found that only six communes reported giving ex- emptions to individuals at the commune health center. One reason for this is that the commune has to bear the cost of exemptions itself. The second type of exemption that is requested by the commune is exemption from charges at hospitals. This is not an entitlement as such, but only a referral letter from the Commune People's Committee (CPC) to hospitals that the individual concerned deserves an exemption because of special considerations. Since the cost of these exemp- tions is borne by the hospital, these exemptions are given out more generously than ex- emptions from commune health center charges. Exemptions at hospitals are also more valuable to the individual recipient than those at the commune health center level, since services at hospitals cost considerably more than those at commune health centers. How well do the formal exemption mechanisms work in actual practice? And how well are they targeted to the poor? Ensor and San (1996) found virtually no correlation between fee exemptions and household income, nor between the number of fee exemp- tions given and the general level of affluence of the commune. The VLSS 1993 and 1998 data are used below to calculate the percent of users from different economic back- grounds who were exempted from user fees and drug charges at public hospitals and commune health centers in 1993 and 1998. There are several observations that can be made from this table (Table 7.2). First, the vast majority (82 percent) of individuals do not pay any user fees for a visit to a commune health center. At public hospitals, only 42 67 The poor typically pay less for a health service contact than the rich, because the average quality of health services they purchase is lower. Using a national average level of out-of-pocket expenditure controls for quality variations across income groups. 150 percent of individuals do not pay a user fee for a service contact. Second, while fee ex- emptions may be common, exemptions from payment for drugs are nonexistent in com- mune health centers and public hospitals. Third, a larger percentage of the poor than the better-off are exempted from user fees at both types of facilities. But the variations in the rate of exemptions across economic groups are not as large as one would expect a priori from a program targeted mainly to the poor. Table 7.2: Percent of users who reported paying nothing for a visit to a government health facilits. 1993 and 1998 Per capau consumption expendirure quintile Health provider LttLest Second Third Fourth -fighesp A erage 199S Government Hospital ".paying no fees 60 57 57 55 58 57 r. paying nothing for drugs 7 4 2 2 5 4 Commune Hcalth Centers a%paing no fees 94 Q0 98 90 77 91 "paying nothing for drugs 13 3 5 3 0 5 1998 Government HospitaIs O'D paying no fccs 50 42 42 41 38 42 0'. pa)inp nothing for drues 0 0 0 0 0 0 Commune Health Centers R'6 payng no fees 90 83 78 79 75 82 "6 paving nothing for drugs 0 0 0 0 0 0 Source VLSS 1993 and 1998 Table 7.2 also reveals that the overall rate of fee exemptions at both public hospi- tals and commune health centers has declined since 1993. While 57 percent of users were exempted from user fees at hospitals in 1993, only 42 percent were in 1998. Likewise, the exemption rate at commune health centers declined from 91 percent to 82 percent over the same period. The decline occurred across all economic groups. Table 7.3 show data reported by nine hospitals in the country on the percentage of fee exemptions they offered over the period 1993-97. The data show large interhospital variations in the of fee exemptions. In some hospitals, such as Bach Mai, Viet Tiep, Dong Da and Son Tra, fee exemptions and reductions are offered to only 2-5 percent of pa- tients. But in other hospitals, such as Khanh Hoa and Da Nang, fee exemptions appear to be given out liberally. It is unlikely that these large variations. reflect socioeconomic dif- ferences in the clientele of the various hospitals. More likely, the data suggest that the fee-exempting mechanism is not applied cqnsistently by different hospitals. 151 Table 7.3: Percentage of fee exemtions and reductioas in selected hospitals, 1993-97 Hospital 1993 1994 1995 1996 1997 Bach Mai - 4.0 4.0 45 5.0 Cho Ray 24.5 21.7 27.4 24.2 25.3 Thai Nguyen 40.3 30.7 22 5 27.8 16 3 Da Nang 28.0 28.5 29.9 28.1 298 Kanh Hoa - - - 38.9 30.2 Viet Tiep - 2.8 2.6 2.1 Dong Da 2.9 2.1 2.0 2.0 1.8 Hai Chau 13 4 14.4 10.8 12.8 14.6 Son Tra 2 9 2.8 4.3 6.9 7.7 Source- MoH 1999c. 5. Discrepancy between Official User Fees Collected and User Fees Paid by Households An important point to note is the large discrepancy between user fee collections at public health facilities as reported by the Ministry of Health and those reported by house- holds in the VLSS 1998 survey. Such discrepancies are not uncommon in developing countries, as some of the fees collected from individual patients presumably may be in- formal (or "under-the-table") payments that are not reported to the facility. In Vietnam, user fee collections as reported by the Ministry of Health were VND 450 billion in 1998. However, the VLSS 1998 survey reported aggregate household spending on user fees at government health facilities of VND 6,321 billion.68 The fact that households report pay- ing 14 times as much in user fees at public facilities as the government reports in user fee revenue collection, which implies that public health facilities are able to capture only 7 percent of user fees actually paid by patients, is troubling and warrants further explora- tion. 6. Provincial Distribution of User Fees How equitably are user fees distributed across provinces? There is sufficient lee- way in the user fee structure to allow poorer provinces to set lower user fees for health services than better-off provinces. However, the need for poorer provinces to raise reve- nues for health expenditure is also greater. 68Note that this amount merely reflects user fee payments at public facilities. It neither includes payment for drugs and medicines (VND 14,024 billion), nor payment of fees to nongovernment health providers (VND 2,643 billion). 152 Data on user fee reve- Figure 7.8: Userf collectionsper capita and provincial domestic nues cros proince indcateproduct per capita. Vietnam provinces, l997- nues across provinces indicate 1 large interprovincial variations - in user fee collections per capita. 8 In 1997, user fees per capita R ranged from VND 910 in Hung g Yen to a high of VND 8,960 in Ho Chi Minh City. Only 17 provinces out of 61 had user fee collections greater than VND 2 2,000 per capita. Figure 7.8 sug- gests that better-off provinces do 0- indeed collect more user fees per capita than do poorer provinces. Poica oetcpoutprcpt 00VD capia tan o pore prvines. 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 9,000 10,00 I1,0 Thisis nt a unepeced rla-Provincial domestic product per capita ('000 VND) This is not an unexpected rela- tionship; not only are individuals in better-off provinces able to pay higher fees per health service contact, they also have larger number of contacts with the public health services than individuals in poorer provinces. User fee collections have Figure 7.9: Real growth in user fee collections over 1993-97 in relation to the level of user fees per capita In 1993, Vietnam provinces grown very rapidly in the recent 350 r past. Indeed, in some provinces, such as Da Nang, Tien Giang, 300 and An Giang, user fee collec- 250 tions per capita grew by more than 300 percent in real terms between 1993 and 1997. The o --- ----0-- - - --- province with the smallest in- . 100 * - crease in user fees per capita - over the 1993-97 period was so ---------- Soc Trong (where user fees 2 grew in real terms by 31 per- 0.0 0.5 1.0 1.5 2.0 25 3.0 35 4.0 4.5 cent). Figure 7.9 suggests that Userfeespercapitain 1993 (000VND) provinces that had lower user fees per capita in 1993 experienced more rapid real growth in per capita user fee revenues from 1993 to 1997 than provinces that had higher levels of user fees in 1993. This indi- cates that the interprovincial disparity in user fee collections per capita narrowed some- what over the 1993-97 period. One important question for interprovincial equity is: how is the financing of health activities through the state budget related to user fee collections? In other words, are provinces that are unable to collect more user fees per capita (because of their lower living standards and higher rates of poverty) able to obtain higher levels of state budget financing for public health activities? As seen in Figure 7.9 above, the ability of prov- inces to generate user fee revenue is strongly related to their GDP per capita. Thus, the role of government policy should be presumably be to provide greater assistance to those provinces that are unable to mobilize user fees to the same extent as other, better-off provinces. 153 Unfortunately, the pro- Figure 7.10: Per capita public health spending out of user fees and out of vincial data indicate exactly the the state budget. 1997, Vietnam provinces opposite relationship (Figure 7.10). A very strong positive re- lationship is obtained between per capita public health spending 5 financed from user fees and that financed from the state budget. This indicates that the state g budget has not reduced the dis- parities in user fee financing. Thus, provinces that are able to collect lower levels of user fees 0v 5 15 2.5 3.5 4.5 5.5 6.5 7.5 8 5 per capita owing to their lower Per capita public health spending out of user fees ('000 VND) living standards are further dis- advantaged in the resources they are able to get from the state budget for financing health activities. C. Health Insurance 1. Description of the Program Formal health insurance began in Vietnam in 1993, following a decree issued in 1992 that set up two separate insurance schemes: (i) a compulsory scheme covering pri- marily current and retired civil servants and employees of state and large (i.e., having more than 10 employees) private enterprises, and (ii) a voluntary scheme aimed at the remainder of the population, viz., farmers, school students, and family members of com- pulsory scheme enrollees. Most of the enrollees in the voluntary scheme are students, since schools are often under pressure from local authorities to enroll all of their students in the scheme. So far, the scheme has had limited success in enrolling farmers and other rural residents. The coordinating agency for health insurance is the Vietnam Health Insurance Authority (VHIA). Each province has a provincial health insurance office, which until recently was under the control of the Provincial Health Bureau. However, since 1999, the provincial insurance offices have been placed under the direct control of VHIA, as sev- eral of the local schemes were in financial difficulties. The compulsory scheme provides both inatient and outpatient care benefits, and also pays for drugs used in inpatient treatment. The voluntary scheme has two pack- ages, a lower- priced package that covers only inpatient care and a higher-priced package that includes outpatient care and, in some situations, drugs as well. Most voluntary enrol- lees (including students) are enrolled in the lower-priced insurance scheme.70 69 For some beneficiaries, pharmaceuticals are covered even for outpatient visits. However, for most bene- ficiaries, this is not the case. 70 Although students can obtain outpatient care at a school clinic, which the VHI helps large schools estab- lish. 154 Premiums in the compulsory scheme are paid from a 3 percent payroll tax, which is shared by employers (two-thirds) and employees (one-third). Enrollees in the voluntary scheme pay fixed annual premia that vary across provinces and according to benefit cov- erage chosen. In some provinces, annual premia are as low as VND 10,000, while in oth- ers they are VND 50,000. Health facilities are reimbursed by the insurance scheme on a fee-for-service basis for both inpatient and outpatient care delivered to enrollees.71 Payment typically consists of a flat fee for accommodation (depending on the admitting department) and a fee for each medical test and procedure based on a standard price tariff. Each provincial health insurance office has (i) a department staffed by medical professionals who scrutinize the diagnosis and treatment given to each patient, (ii) an accounting department to verify the standard payment to be made to the providers, and (iii) an enrollment department to reg- ister and collect premia from enrollees. Until very recently, only government health facilities were accredited to deliver the benefits covered by the insurance program. Outpatient benefits are typically provided by hospital-based clinics or outpatient departments at hospitals, except in the case of em- ployees in special resource sectors, such as transportation, communication and mining, who tend to use health facilities operated by the ministries governing these sectors. 2. Population Coverage Membership in VHI grew at double-digit rates from 1993 to 1996, but has since slowed down considerably (Figure 7.11). In 1998, there were a total of 9.8 million enrol- lees, of whom 38 percent (3.7 million) were enrollees in the voluntary program. Indeed, there was a decline (albeit small) in membership in the voluntary scheme between 1997 and 1998, in part because of increases in premium charged and because of competition from the Bao Viet Insurance Figure 7.11: Enrollment in HeAth Insurance, 1"3-" Company for student health in- 6 5 surance. The slowdown in en- 45 rollment rates reflects the diffi- culties of expanding the present type of payroll-financed social - - health insurance beyond its cur- 5 2 rent level. 2 Compulsory 15 5 0 Voluntary -a - Voluntaryas%oftotal ...................... ...... ........ ........ ~ 5 1993 1994 1995 1996 1997 1998 71 At the time of inception, hospitals were reimbursed on a fixed daily allowance basis. Since this method of payment created incentives for patients to be kept in hospitals for longer periods, the system was changed in late 1995 to a fee-for-service basis. 155 Regional Variations. Vietnam's health insurance program currently covers about 12 percent of the country's population, with large variations across geographical regions. Coverage is only 6 percent in the Mekong River Delta and 9 percent in the Central High- lands, while it is 15-16 per- land , whle i is 5-16 per-Figure 7.12: Percentage of total population covered by health insurance, by cent in the Southeast and the reion, yb Red River Delta (Figure 18 7.12). There are even larger 6 15 differences across provinces. 5 4 Hai Phong, where the pro- 12 gram first started on an ex- perimental basis, has 38 per- -2 cent of its entire population s covered by health insurance. At the other extreme, the 8. province of Soc Trang in the Mekong River Delta has a coverage of only 3.7 percent. Northen d edRiver Nort) Central Central Soutrast Mckong ietnam Uplands DeIla Central Coast Hiehlands Rwer Delta Compulsory versus Voluntary Scheme Coverage. The compulsory scheme has been successful in terms of expanding coverage of its target population (Table 7.4). When it started in 1993, it covered less than 50 percent of its target population; by 1998, it had covered 77 percent of its target population. The voluntary program is still in its in- fancy, having covered a mere 5 percent of its target population. Thus, most of the growth of social health insurance in the future will have to come from expansion in the coverage of different types of voluntary and subsidized community-based social health insurance schemes. Table 7.4: Share of target population covered by health insurance, 1993-98 Compulsory scheme Voluntary scheme Enrolled Target popu- Enrolled as Enrolled Target popu- Enrolled as population lation (md- % of target population lation (md- % of target Year (mdlions) lions) population (millions) lions) population 1993 3.47 7.10 48.92 0.33 63.86 051 1994 3 72 724 51.38 0 54 65.15 0.82 1995 4.87 7.38 65.99 2.23 66.41 3.36 1996 5.56 7.52 73.92 3.32 67.65 4.91 1997 5.74 7.66 74.87 3.82 68,89 5.54 1998 6.05 7.87 76.82 3.74 70.13 5.34 Source: Dunlop 1999. Within the present voluntary scheme, over 90 percent of members are school children, who, as noted earlier, are quasi-compulsory, as parents and schools are under a great deal of pressure from local authorities to enroll them (Table 7.5). The number of humanitarian enrollees fluctuates a great deal from year to year, as insurance policies for these enrollees are provided by a variety of organizations, including government at the national and local levels, donors and charity organizations, such as the Red Cross. 156 Table 7.4: Composition of voluntary health insurance scheae membership, 1993-97 M%) Total voluntary Humanitarian Other members scheme member- Year School children (free cards) (e.g.. farmcrs) ship 1993 000 26.07 7393 100.00 1994 44.49 6.17 49.35 100.00 1996 91.79 6.15 207 10000 1997 9077 5.48 3.75 100.00 Note. Figures in parentheses are percentages of total. Source: Jowett and Thompson 1999. Demographic Composi- Figure 7.13: Anal health service contacts, by age group, 19 tion. There is an interesting dif- ference in the demographic composition of enrollees in the I compulsory and the voluntary health insurance schemes. While 5 i 46 percent of all enrollees in the 4 compulsory scheme are over the age of 60 years, the large major- ity (over 90 percent) of enrollees in the voluntary scheme are stu- - - - - - - dents, who are typically aged 5- - 19 years (VHIA 1999). It was 0-4 5-19 20-34 35-44 45-59 60 & over noted in Chapter III.A that Age goup (ye-rs) school-age children have the lowest utilization rates of health services in Vietnam, while those aged 60 and over have the highest rates. This is shown again in Figure 7.13, which shows the annual health ser- vices contact rate of different age groups from the VLSS 1998 data. It is seen that indi- viduals aged 60 years and over have two times the annual service contact rate of those aged 5-19 years. The annual service contact rate with public providers varies even more across the age groups, with the elderly (aged 60 and over) making three times as many public provider contacts as school-age children. This has important implications for the expenditures incurred by the health insurance program, as is discussed in the following section. 157 3. Role of Insurance in Financing Public Health Spending Despite its relatively figure 7.14: Share of insurance and user fees in the financingof total public recent start (1993), the health expenditure on health, 1992-97 to insurance program has come 9 to play an important role in generating additional re- User lecs, sources for the health sector 7 7 7 In 1993, its first year of op- eration, it financed about 2.7 percent of total public spend- ing on health in the country. By 1997, this share had in- creased to 9 percent (Figure 7.14). Meanwhile, the share of user fees in total public spending grew from only 5 1 9 9 19 196 99 percent in 1993 to 7 percent in 1997. However, these figures understate the very important and significant role that health insurance expenditure plays in financing hospitals in Vietnam. In 1998, for in- stance, health insurance reimbursements constituted 15 percent of total hospital revenue (Figure 7.15). 4. Equity in Health Insurance Economic Composition of~~~~~~~~ Enoles Tow a xetd re 7.15: The share of health insurance expenditure and state budget of E rolles.To w at xten dosubsidies in total hospital revenue, 1994-98 the poor benefit from the social 11 76 elh n=c health insurance program? The ]l £1c Useue Usererfees economic background of health 70EZ: State budget insurance enrollees can provide 60 154 552 an indication of which income4.7 groups benefit most from the social health insurance program. 40 - Because enrollees in the com- 30j f31 e 5 1 pulsory health insurance ~2 scheme are primarily civil ser- 20 0 13 15 i vants and salaried employees of 1to state and large private enter- 0 2 prises, many of whom are mid- 1994 1995 1996 1997 998 die-class urban residents, it is likely that enrollees in the compulsory insurance scheme are significantly better off than the general population, which is largely rural. It is difficult to guess a priori the relative economic situation of enrollees in the voluntary insurance scheme. As noted earlier, 90 percent of enrollees in the voluntary scheme are students. Although there is pressure on all parents to enroll their school-going 158 children in the scheme, at the present time the voluntary scheme covers only a fraction of all students in the country. It is likely that it is the better-off students who enroll in the voluntary scheme, as their parents are able to pay the annual health insurance premium. However, there is no data to confirm or refute this hypothesis. Figure 7.16 suggests ththelhi sur anc coverauget Figure 7.16: Health insurancre coverage by per capita expenditure quintile, 1998 that health insurance coverage 4 rates are much smaller among 37 the poorer consumption quin- 3 - q/i of individuals with insurance coverage tiles than among the better-off 4 tiles ~ ~ ~ ~ ~ ~ ~ - thnaogtebetrof%i of~all insurance enrollees drawn from qLuntile 29 quintiles. For instance, only 6 percent of the lowest quintile 7 25 0 have coverage, while 29 per- ' cent of the highest do. Looking 2 at the data another way, health 15 - insurance enrollees in Vietnam are drawn disproportionately 6 from the better-off groups in 5 society. For example, while only 8 percent of insurance en- o rollees are drawn from the low- Per capita consumption expenditure qsiile est consumption quintile, as many as 37 percent from the highest quintile are. Equity across Compulsory and Voluntary Enrollees. Another equity issue is the distribution of insurance benefits across enrollees in the compulsory and voluntary parts of the health insurance program. While voluntary enrollees constituted about 40 percent of total enrollees in 1997, they accounted for only 9 percent of the total premiumrs. This meant that the average premium is about 7 times as high for compulsory enrollees as for voluntary enrollees (Table 7.6). However, the service usage of compulsory enrollees is also proportionately greater. The average inpatient admission rate is more than three times as large for compulsory enrollees as for voluntary enrollees, while the outpatient contact rate is 10 times greater. 159 Table 7.6: Service use and cost of services to VHIA. 1997 Compulsorn oluntart Variable scheme scheme Both schemes Insured population (millions) 5-7'0 3 80 9.50 Averge premium per enrollee ('000 VND) 91 23 13.16 6000 Rate of inpatient admissions per 1,000 enrollees I7N 40 52 60 126.30 Number of outpatient service contacts per 1.000 enrollecb 2 II 0)21 1 35 A%erage expenditure per inpatient enrollee ('000 \ ND) 216 67 Average expenditure per outpatient sem ice contact b, enrol- lee ('000 VND) 18 75 Ratio of average premium per enrollee to average expendi- ture per inpatient enrollee 0.42 0.06 0.28 Ratio of average premium per enrollee to average expendi- ture per outpatient contact 4.87 0.70 3.20 Source: Dunlop 1999, 5. Level of Utilization of Health Insurance Fund Across Provinces Since civil servants and salaried employees of large firms derive most of the benefits of the health insurance program, one would expect its benefits to be highly con- centrated in a few provinces having large urban centers where large numbers of individu- als buy compulsory health insurance. This is observed to be the case in Table 7.7, where Hanoi and Ho Chi Minh City alone account for 29 percent of national health insurance expenditure, even though they account for 10 percent of the national population. The 10 provinces receiving the largest health insurance reimbursements account for more than one- half of aggregate national health insurance reimbursements but account for only 31% of the national population. Table 7.7: Shares of 10 provinces in national health insurance expenditures, 1997 Shore in national Cumulalive share in national national health national health insurance national insurance national Province expenditure population expenditure population Hanoi 14.67 3.05 14.67 3.05 Ho Chi Minh City 13.98 6.67 28.65 9.72 Haiphong 5.16 2.24 33.81 11.96 Thanb Hoa 3.29 4.70 37.10 16.66 Nghe An 3.25 3.80 40.35 20.45 Nam Dinh 2.76 2,56 43.12 23.01 Quang Ninh 2.18 1.24 45.29 24.25 Ha Tay 2.16 3.13 47.45 27.38 Thai Nguyen 2.13 1.28 49.58 28.66 Dong Nai 2.11 2.60 5169 31.25 Source: VHIA 1998. 6. Financial Viability The expansion of the present type of social health insurance program can run into financial difficulties because of problems of moral hazard and adverse selection. Moral hazard relates to the practice of excessive utilization of covered services by the insured. This can occur because (i) the insured demand more health services as the effective price 160 of care for them is lower (consumer moral hazard), and (ii) providers deliver more ser- vices to the insured because they have a financial incentive to do so through the fee-for- service system (provider moral hazard). Adverse selection refers to the process by which individuals with poor health (and therefore most likely to make use of insured health ser- vices) self-select themselves into a voluntary insurance program. Both moral hazard and adverse selection can result in uncontrolled growth of insurance reimbursements, thereby endangering the financial solvency of an insurance fund. Overuse of Health Insurance (Moral Hazard). There is compelling evidence that overuse of health insurance is a real problem in Vietnam. Table 7.8 shows that, while the annual outpatient service contact rates of health insurance enrollees were not significantly different from those of the general population in 1997, the inpatient admission rate was significantly higher (by about 70 percent). In addition, both the outpatient contact rate and the inpatient admission rate have grown much more rapidly over the period 1993-97 for the insured than for the general population. Indeed, the inpatient admission rate for the general population has been more or less stagnant over the entire period, while that of the insured has grown by about 41 percent. The growth rate of inpatient admissions among the insured is probably unsustainable. Table 7.8: Service Use and Health Expenditures of Health Insurance Enrollees, 1993--97 Health Inurance Program en- General Health Insurance Program en- rollees population rollees Ratio of Ratio o' average Ratio of inpaticnt exp per ALerage A,erage inpatient .4erage A Lerage admis- Ai erage 4 verage outpatient number of number o/ admissions number number oj sfons to exp per exp per contact to outpatient inpatient to ourpa- of outpa- inpatient outpa- outpatient inpatient average contacts admis- ien con- tient con. adinis- tient con- contact admssion exp per per enrol- sons per tacis per tacts per stons per tacts per ('000 ('000 inpatient Year lee enrollee enrollee capira capita capia i'ND) IVND) admission 1993 0.48 0.07 0 14 084 0 09 0 11 1167 103.06 8.83 1994 0.95 0.12 0 12 0.98 0.08 0.08 14.12 122 11 865 1995 097 0 11 0 11 1 1S 0.07 006 1763 169.27 960 1996 1.20 0 12 0.10 1 49 0.0' 0 05 1808 J93.95 10 3 1997 136 0.12 009 1.52 007 0.05 18.56 215.71 11.62 Source- Dunlop 1999. Table 7.8 also indicates that the average expenditure on an inpatient admission for insurance program enrollees has been increasing much more rapidly than the average ex- penditure for an outpatient service contact. In 1993, the cost of an inpatient admission for an enrollee was about 9 times the cost of an outpatient service contact. By 1997, the ratio had increased to a factor of 12. The increase in inpatient expenditures probably reflects provider moral hazard, as hospitals respond to reduced public subsidies by increasing the number of billable and reimbursable services to insured patients. One problem with a simple comparison of the health services utilization of indi- viduals with health insurance coverage and those without is that it does not control for the other characteristics that may be causing the difference. In particular, as noted earlier, the 161 insured as a group are more affluent than the noninsured, and this fact alone could be driving the differences in utilization. Fortunately, it is possible to control for income status when comparing the utilization behavior of the insured and noninsured, using the VLSS 1998 data. Such a comparison, shown in Table 7.9 below, is very revealing, be- cause it suggests that, even after controlling for economic status, the insured have signifi- cantly higher rates of service utilization of public providers, especially government hospi- tals, than the noninsured. The discrepancy in public sector utilization rates between the insured and noninsured is observed for every consumption quintile. Table 7.9: Annual bealth service contacts, by insurance status and per capita expenditure quintile, 1998 Annual service contacu per Annual service contacts per Annualpublic sector capita with public Per cap. capita service contacts per capita hospitals exp. quin- Nonin- Nonin- Nonin- tile sured Insured Total sured Insured Total sured Insured Total Lowest 8,21 8.12 8.21 0.93 1.96 0.99 0.24 0.40 0.25 Second 11.08 10.54 11.03 1.19 2.35 1.30 0.31 1.22 0.39 Third 12.16 11.28 12.04 1.47 2.06 1.55 0.44 1.10 0.52 Fourth 11.39 9.79 11.07 1.43 2.69 1.69 0.55 1,62 0.76 Highest 13.32 9.93 12.33 1.41 2.11 1.62 0.91 1.51 1.09 Total 11.10 10.06 10.94 1.27 2.27 1.43 0.46 1.34 0.60 Source: VLSS 1993 and 1998. Table 7.10, which reports statistics related to 12-month hospitalization by insur- ance and consumption quintile status, also clearly shows that the insured have signifi- cantly higher inpatient admissions rates and significantly longer lengths of hospital stay than the noninsured across virtually every consumption quintile. The longer average length of stay for the insured most likely reflects provider moral hazard, as hospitals try to increase the amount of insurance reimbursements they can obtain by extending the du- ration of hospitalization for insured patients. Table 7.10: Inpatient admissions rate and average length of stay by insurance statVs and per capita expenditure quintile, 1998 Inpatient admissions rate per Average length of hospital stay Per capita 1.000 persons (days) expenditure Nonin- Nonin- quintile sured Insured Total sured Insured Total Lowest 33.32 42,28 33.86 10.38 8.75 10.26 Second 36.70 107.32 43.48 9.52 15.29 10.91 Third 41.99 96.14 49.28 13.73 14.39 13.90 Fourth 53.79 93.99 61.86 11.91 20.84 14.61 Highest 59.23 73.23 63.35 18.31 19.66 18.76 Total 44.05 84.22 50.36 12.99 17.85 14.26 Source: VLSS 1993 and 1998. Adverse selection. Adverse selection is not yet a major problem for the social health insurance program simply because the program has not yet expanded its voluntary 162 scheme enrollment much. As noted earlier, 90 percent of enrollees in the voluntary scheme are students, who are more or less quasi-compulsory as their parents and schools are under pressure to enroll them in the health insurance program. In addition, as a demo- graphic group, school-age children are among the lowest users of health services-not just in Vietnam but in most other countries. A large survey of 2,751 individuals in three provinces-Hai Phong, Ninh Binh and Dong Thap-done in 1999 provides compelling evidence about the existence of ad- verse selection in the voluntary insurance scheme. Individual respondents in the survey were asked whether they perceived their own current health status (within the past 12 months) as generally good or not. Slightly more than a third of the sample (36 percent) were voluntary members, with 79 percent of the voluntary insurance enrollees being adults and the remaining 21 percent being children. A statistical analysis of the determi- nants of voluntary membership indicated that, even after controlling for a large number of characteristics, such as age, sex, marital status, household income, religion, occupation, and region of residence, individuals who considered their current health status to be 'good' were in fact less likely to purchase voluntary health insurance (Jowett, Vinh, and Martinsson 1999). The experience of other countries also suggests that, as the health insurance pro- gram makes a big push toward enrolling farmers and rural residents in its voluntary scheme, it will experience the problem of adverse selection. People who are prone to ill- nesses and who generally have poor health will have a stronger incentive to enroll in the program than healthy individuals. Financial viability. The problem of consumer and producer moral hazard has meant that utilization rates are too high for the available revenue of the VHIA. In 1996 and 1997, many of the provincial health insurance funds were overdrawn in comparison to the paid premiums during that year. In 1996, nearly 25 percent of provinces had insur- ance expenditures exceeding insurance revenues, while in 1997 about 10 percent of the provinces ran a deficit (Dunlop 1999). While the financial integrity of the entire insur- ance fund was not in jeopardy, such problems are likely to keep cropping up in the future. In response to these financial difficulties, the provincial health insurance offices have been placed under the direct control of VHIA since this year (1999), and VHIA has tightened the guidelines regarding the control and use of funds. Several changes to the health insurance package have been implemented, including: * a 20 percent copayment for those insured under the two schemes (with school chil- dren exempted), up to an annual maximum of six months' basic salary (about VND 860,000); * maternity benefits for only the first two children; * extension of waiting period (for benefits to commence) from one month to three months. In addition, many provinces have placed their own rules to limit reimbursements. For example, Ben Tre has limited outpatient reimbursements to health facilities to 55 per- cent of the total facility charges, effectively raising the copayment by individuals. The 163 province has also placed a ceiling of VND 102,000 per case on reimbursements for inpa- tient care. The new decree also provides for reimbursement by the insurance fund to private health care providers who meet minimum quality-of-care standards. Apparently, the VHIA has already signed contracts with private providers in selected locations, including in Phu Tho province, for high-tech diagnostic services. Since the latter are high-cost ser- vices typically used by the better-off and the elderly, it is not clear how this change will impact on the expenditures of VHIA and on the benefit incidence of social insurance in general. The new decree also states that the insured population should be able to obtain outpatient services at any facility, including commune health centers. While this could serve to increase the utilization of commune health centers for curative treatment, there is also the risk that extension of provider benefits to the commune health centers would re- sult in spiraling costs for the insurance agency, as the distance traveled to district hospi- tals is currently acting as a nonprice rationing device to limit utilization rates and insur- ance claims (Ensor 1999). Perhaps, the 20 percent copayment requirement might offset some of the increase in insurance expenditures that would occur as a result of extending provider benefits to the commune health centers. D. Commune Health Center Finance The issue of how commune health centers are financed is an interesting one. Until 1994-95, they were entirely off-budget and were expected to be self-sufficient, but the state budget now provides for the salaries of commune health center workers. Before 1988, the agriculture work brigades used to finance the commune health centers and the production brigade nurse, whose responsibility was to support the work of the commune health center workers through outreach. But after doi moi , support for commune health centers and the brigade nurses decreased. The commune health centers had to depend for their financing from the commune people's committee, and this funding was erratic and insufficient to maintain the quality of health services previously offered. As a result, the extensive commune health center network, which had been built up with great effort and resources, was weakened. The situation only improved after the government decided in 1994-95 to pay the salaries of commune health center workers out of the state (provin- cial) budget. Data on commune finance are not normally collected and available routinely through the health information system or the budgetary system. However, a longitudinal study of four communes in Quang Ninh province in 1991 and 1995 did collect such data from the records of each of the four commune health centers (MoH and IDS 1998). This study provides interesting insights into the financing of commune health centers before and after the government decision to finance commune health center worker salaries out of the state budget. The study found that commune health center expenditure per com- mune member nearly doubled in real terms between 1991 and 1995 in two of the com- munes, quadrupled in one commune, and was stagnant in one. The impact of the 1994-95 government decision is reflected in the sharp increase in the proportion of salaries and allowances in total recurrent expenditure from an average of 26 percent in 1991 to 40 164 percent in 1995. Real income of commune health center staff increased almost twofold during the four-year period. The responsibility for the payment of staff worker salaries shifted from the Com- mune People's Committees to the district health center (the unit through which the state budget monies were channeled). In addition, the fact that the government took over the responsibility of paying for commune health center worker salaries meant that salaries did not have to be financed to the same extent as in 1991 from the marked-up sale of drugs. In 1991, profits from drug sales financed an average of 13.5 percent of salaries; by 1995, this proportion had halved to 6.4 percent. The weakened financial incentive for commune health center workers to (over)prescribe drugs in turn resulted in a real de- crease in the value of commune health center drug turnover (observed in three of the four communes) between 1991 and 1995. Of course, drug sales at cost (i.e., with no mark-up) were still accounting for an average of 35-40 percent of total commune health center ex- penditure, reflecting the fact that few drugs at commune health centers are provided free of charge to patients. A more recent study of six communes in Ninh Binh, Phu Tho, Can Tho and Ben Tre provinces found that over three-quarters of total service costs at the commune level are financed by user fees (Dunlop 1999). The study also found that commune health cen- ter expenditure amounted to an average of VND 46,000 or 2.7 percent (with a range of 0.7 to 5.6 percent) of commune GDP. Interestingly, the study found that, in provinces where the provincial health insur- ance funds ran into severe financial difficulties in 1996 and 1997 (e.g., Ben Tre), the commune health centers were in fact providing reimbursable services to health insurance enrollees. This was not the case in provinces whose insurance funds had not experienced financial difficulties. The reason for this is that the cost per case treated at a commune health center is only a fraction of that treated at a district or higher level hospital (Dunlop 1999). 165 CHAPTER VII ANNEX TABLES HOSPITAL USER FEE SCALES Table A7. 1: Part A. Price scale for consultation and physical examination (in VND) Fourth class First class hos- Second class Third class hospital and Types of exam pital hospital hospital polyclinic 1 General examination; specialist examination 2,000-3,000 1,500-3,000 1,000-2,000 500-1,000 Examination according 2 patient's request (patient choose doctor) 10,000-30,000 10,000-20,000 10,000-20,000 Not applicable Examination for giving certificate of injury, medical assessment for the court (not including 3 biology & biochemistry Lab test, X-ray) 15,000-35,000 15,000-30,000 10,000-20,000 Not applicable Entire/full examination for recruiting staff, driver (not including biology & 4 biochemistry Lab test, X- ray) 25,000-50,000 25,000-40,000 18,000-35,000 Not applicable 166 Table A7.2: Part Bl. Price scale for one day of stay (inpatient) (in '000 VND) Second Fourth First class class hospi- Third class class hospi- Types of bed by department hospital tal hospital tal Fee per day for emergency/intensive care unit; delivery and 2 days after de- 1 livery 12-18 8-12 6-9 3-6 Fee per day for internal department Type 1: Departments: - Infectious disease - Respiratory - Hematology - Cancer - Cardiac-vascular - Nervous system - Pediatric - Intestinal disease - Kidney - The day from the third day after de- livery; the day for treatment after 10 2 days of surgical operation. 8-10 6-8 3-5 2-3 Type 2: for Departments: - Muscular, skeleton/bone; - Dermatology -Allergic dept - Ear-nose-throat; eye; teeth-jaw-face, surgery, and obstetrics-gynecology 3 without operation 6-8 4-6 2-5 1.5-2.5 Type 3: for Departments: -Traditional medicine 4 - Rehabilitation 4-6 2.5-4 1.5-3 1-1.5 Fee per day for Surgical department and burning department Type 1: After special operation; burn 5 degree 3-4, > 70% 15-20 10-16 Type 2: after the operation class 1.; and 6 burn degree 3-4: 25-70% 10-15 6-10 5-10 Type 3: after the operation class 2; and burn degree 2: >30%, burn degree 3-4: 7 < 25% 8-10 5-8 4-7 Type 4: after the operation class 3; and 8 burn degree 1, 2: < 30% 6-8 4-6 3-5 2-3 167 Table A7.3: Part B2. Maximum price scale for one inpatient treatment day (in '000 VND) Second Fourth First class class hospi- Third class class hospi- Department hospital tal hospital tal One treatment day in intensive care 1 unit 120,000 86,000 30,000 20,000 One treatment day in internal depart- 2 ment 2.1. Hematological diseases and cancer 50,000 50,000 - - 2.2. Pediatric, communicable diseases, respiratory, cardiac- vascular diseases, urinate diseases, bones, mental, derma- tological and diseases without opera- tion in surgical, obstetric- gynecologi- cal, eye, teeth- jaw-face, ear-nose- throat (ENT) diseases 40,000 40,000 20,000 10,000 2.3. Traditional medicine; rehabilita- tion department 30,000 20,000 15,000 10,000 One treatment day in surgical; burn 3 department 3.1. after the operation class 3; and burn degree 1, 2: < 30%; bum degree 3-4: < 25% 60,000 50,000 30,000 20,000 3.2.after the operation class 2; burn degree 2: > 30% 70,000 60,000 40,000 25,000 3.3. after the operation class 2; burn degree 2: > 30% 90,000 80,000 60,000 - 3.4. After special operation; burn de- gree 3-4: > 70% 120,000 100,000 - - 168 Table A7.4: Part C. Price scale for technical services and biology and biochemistry lab tests and x-rays Names of technique Minimum price (VND) Maximum price (VND) 1 Urinary relief 2,000 6,000 2 Stool relief 2,000 6,000 3 Injection to node of lymphocytes 3,500 10,500 4 Injection to Thyroxin-gland 4,000 12,000 Injection to abdominal/lung mem- 5 branes 3,500 10,5000 Injection to lung membranes/air relief 6 from lung membranes 15,000 45,000 7 Clean urine ball 7,000 21,000 Urinary tract relief, putting a pipe in 8 urine tract 5,000 15,000 9 Treatment for the penis (male) 5,000 15,000 10 Operation for the kidney (one time) 150,000 300,000 Stool absorption through abdominal I1 membrane 150,000 300,000 12 Skin biopsy 5,000 15,000 13 Muscular/lympho node biopsy 5,000 15,000 14 Skeleton core-biopsy 10,000 30,000 15 Lung membrane/abdominal biopsy 10,000 30,000 16 Intestinal tract biopsy 10,000 30,000 Male gland inside the penis biopsy 17 through urine-ball 15,000 45,000 Abdominal cavity otospectomy and 18 biopsy 10,000 30,000 19 Stomach otospectomy and biopsy 10,000 30,000 Sigma/big colon (big intestinal tract) 20 otospectomy and biopsy 15,000 45,000 Big intestinal tract after sigma/colon 21 otospectomy and biopsy 10,000 30000 169 VIII. PUBLIC EXPENDITURE ON HEALTH A. Overall Health Expenditure 1. Public Spending A consistent time series of Figure 8.1: Total public spending on health, Vietnam. 1991-98 public spending on health in Vietnam is available for the pe- 5 . .- riod 1991-98. These data indicate _0_000 a large increase in public spend- - so.o0o ing on health over time (Figure , - 0 8.1). Both total and per capita t.,000 public spending on health in- e 30000o creased more than twofold in real terms72-representing an annual - 20,000 rate of growth of around 12-14 2 Total o heath 4 mic%ofGDP percent, while health spending as - . 10,000 a share of GDP and as a share of 0 - 0 total government expenditure in- 1991 1992 1993 1994 1995 1996 1997 1998 creased at more modest rates. How does this level of public spending on health compare with spending levels in other countries in the region? Table 8.1 shows that Vietnam's level of per capita public spending on health places it roughly in the middle of 14 countries in Asia-ahead of Bangladesh, Nepal, Pakistan, India and Laos, but behind Indonesia, Cambodia, China, Sri Lanka, Philippines, Thailand, Malaysia and the Republic of Korea. However, its public health spending when expressed as a proportion of GDP or total government expenditure places it among the top five countries in the region. All of the countries spending a higher proportion of their GDP or their total government expenditure on health (with the sole exception of Cambodia) have significantly higher levels of per capita GDP than Viet- nam.73 72 In this report, government health expenditure is defined as spending on health programs and services by public health facilities that is financed from the state budget, external donors, and from other sources such as user fees and health insurance reimbursements that public facilities receive for their services. The state budget includes both the central state budget (which in turn includes not only the budget of the Ministry of Health but also the health budgets of other ministries and government agencies, such as Posts and Tele- communications, the army, and police) and the local health budgets of provincial governments. While offi- cially posted user fees are collected by government facilities from payments by households (and might be considered a private source of finance), they are typically included as a component of public spending be- cause decisions on how to allocate and spend user fee revenues on various public programs and facilities are under the direct control of the government. Household expenditures on health that are controlled by the households themselves, such as spending at private health facilities and nonuser fee spending at public health facilities (e.g., informal payments), are included as private health expenditure. 7 The high level of public spending on health in Cambodia is largely supported by donors, who account for 78 percent of all public spending on health in that country (World Bank 1999). 170 Table 8.1: Public spending on health. latest year a%ailable (1991-96), Asia Pubbic venJine on hejih per capita (LS5 as' ,f,tdo l col I e m % of GDP Bangladesh 2.4 PakIS1n 35 M%anmar 0.4 Nepal 2 4 %1,annar 4 0 India 0 7 Pakistan 3 1 India 4 4 Indonesia 0 7 India 3 7 Indonusia 4 7 Pakistan 0 8 Lao PDR 4.3 Sri Lanka 5 4 Nepal 1 2 Vietnam 5 1 Malayia A 5 13.ingladush I 2 Indonesia 6 2 Philippines 6 g Malaysia 1 3 Cambodia 7 2 Nepal 7 I Philippines I 3 China X 1 icinam 5 Lao PDR 1.3 RepubliL %f Ko- Sri Lanka 9.2 rea 10 0 Sri Lanka 1 4 Philippines 12 3 Thailind 10 5 Vietnam 1.7 Republic. of Thailand 28 7 Cambodia 22 t Korea 1.9 Malaysia 40 7 China 2n 3 Thailand 2 0 Republic of Ko- rea 13( 7 China 2.1 Cambodia 22 Note. Data for Vietnam are ror loNs Source- World Bank (1999) and UNDP (1I9) At the same time, an annual (public) health spending of about $6 per capita is only one-half of the $12 recommended by the World Development Report 1993 as being needed to finance an essential package of health services, including public health, for a low-income country.74 To achieve this recommended level of public health spending, the Vietnamese government would need to allocate an additional US$468 million annually on health. 2. Total (Public and Private) Spending If private health spending, data on which are available from the VLSS 1993 and VLSS 1998 surveys, is added to public health spending,75 a total health spending figure of VND 28,773 billion in the aggregate or VND 383,226 per capita is obtained for 1998. This is equivalent to US$27.41 per capita or 8 percent of GDP. In comparison, total (pub- lic plus private) spending on health in 1993 was US$9.67 per capita or 5.2 percent of GDP. 74 World Bank 1993. 75Private spending is defined here as total household-reported expenditure on health visits and drugs less officially reported user fee collections. See footnote 1 of this chapter for a definition of public and private health expenditure. 171 A toal ealt sp ndin ofFigure 8.2: Total (public plus private) health spending per capita, selected Asian A total health spending of(1991-98) US$27 per capita places Vietnam 400 408 among those countries in Asia that spend the most on health care (Figure 8.2). Indeed, only Malaysia, Thailand and the Re- 250 public of Korea spend more on 200 health per capita than Vietnam. B. Cost-Effectiveness of Public Health Spending 5 12 Given the large social re- turns to health in low-income Oe 4 0 countries, more public spending on health is generally desirable. However, it is essential to ensure that public spending on health is used in a cost-effective manner. There are two ways of determining cost-effectiveness: first, by comparing over time measures of utilization against public spending on health, such as service contacts and inpatient days at public health facilities, and, second, by comparing across countries health outcomes produced by public health spending, such as life expectancy and infant mortality. Cost-Effectiveness over Time. Figure 8.3, which shows the number of public sec- tor service consultations and inpatient days 'produced' per million VND of public health spending (measured in constant 1994 VND) for the period 1992-98, indicates that the 'productivity' of public health spending declined precipitously from 1992 to 1995, but has since trended upwards. Nevertheless, the number of service consultations and inpa- tient days per million dong of public health spending in 1998 were well below their re- spective values in 1992. What the data in Figure Figure 83: Productivity of public health spending, 1992-98 8.3 essentially indicate is that 41 - -- ---- - ----- 20 health service consultations at = public facilities have not kept i pace with public health spending 37 ----- between 1992 and 1998. This _s could reflect a real decline in the 3 5 Conmukaonsper million VND productivity or efficiency of the j 13 ---- -- 14 public health system, or it could simply reflect other factors at Jt12 work. The latter could include I " an improvement in the quality of 27 npa n 10 public sector inpatient and out- ' patient contacts over time (e.g., 2 --- - - 1992 1993 1994 1995 1996 1997 1998 more drugs and medical sup- plies, better-trained health workers, etc.), or an increase in the number of preventive health activities performed within the public health system. Quality improvements would 172 serve to increase public spending on health without a commensurate increase in the num- ber of service consultations or inpatient days. Likewise, because the 'output' in Figure 8.3 reflects only curative activities of the health system, it does not account for the possi- bility that public health spending may have shifted to (more desirable) preventive health activities and preventive contacts. Cost-Effectiveness Figure 8.4: Life expency and real public spending on heaitb per capita across Countries. Countries that spend more public resources per capita on health tend to have 70 higher levels of life expectancy, 65 controlling for literacy. Adult e literacy needs to be controlled for in this relationship, as there i 55 is a strong positive relationship 7 5t- between life expectancy and adult literacy. Figure 8.4 plots the relationship between life 40 expectancy and public health spending per capita for a cross- 100 150 200 250 300 350 section of low- and middle- Real public s pending on health per capita (PP a r) income countries from around the world.76 Vietnam is observed to be almost on the re- gression line in the graph, which suggests that Vietnam more or less enjoys a life expec- tancy rate consistent with its high level of public spending on health. However, there are a number of countries that are well above the regression line, which suggests that there may be further scope for Vietnam to improve the cost-effectiveness of its public health spending. Thus, Vietnam's own experience in the past as well as the experience of other low- and middle-income countries suggests that it may be possible for Vietnam, with pol- icy changes and expenditure reallocations, to improve the effectiveness of its public health spending and thereby obtain superior health outcomes without commensurate in- creases in public health spending. C. Composition of Public Health Expenditure 1. Functional Composition One of the key messages in health economics is that preventive care interventions should have the first claim on public resources due to the substantial externalities they create for society and due to the resulting underspending on them by private agents. However, in most developing countries, the bulk of public spending on health goes into providing curative services, especially at the secondary and tertiary levels, with relatively few resources allocated to preventive care and primary health programs. Such a distribu- tion may be both inefficient and inequitable, as survey data from many countries, includ- 76 The sample of low- and middle-income countries in Figure 8.4 includes countries from all over the world-not just from Asia. The data are taken from UNDP (1998). 173 ing Vietnam, show that a disproportionately large share of hospital users tend to be drawn from the higher income quintiles. Unfortunately, the functional classification system used by the Ministry of Health and the Ministry of Finance does not readily permit disaggregation of public spending on health by service levels (i.e., primary, secondary and tertiary). However, the system does allow classification of health expenditures into preventive, curative, and family planning expenditures. Preventive expenditures include monies spent on communicable disease- control programs and on the training of village-level health workers, while curative inter- ventions reflect spending on hospitals, health centers and health facilities providing am- bulatory care.77 Table 8.2 shows the functional composition of health spending from each public source, as well as from all public sources together, for the period 1991 to 1998. It should be noted that central state budget spending includes not only the budget of the Ministry of Health but also the expenditures incurred by other ministries (e.g., Posts and Telecommunications, the army, police, etc.) on health. Although this disaggregation is not perfect, analysis of these data can provide some idea about the efficiency of public spending on health. There are five observations that stand out from an examination of Table 8.2. First, curative care takes up a disproportionately large share of public spending on health (70 percent in 1998). In contrast, preventive interventions, such as communicable disease- control programs, account for merely 11 percent of public spending on health. The share of family planning is about 10 percent. While this budgetary emphasis on curative activities is not unusual (and is ob- served in many developing countries), it is less efficient and equitable than it could be. As noted above, a key messages in health economics is that preventive care interventions should have the first claim on public resources due to the substantial externalities they create for society and due to the resulting underspending on them by private agents. This pattern of spending is inequitable, as most curative spending is focused on hospitals and survey data for Vietnam show that a disproportionately large share of hospital users tend to be drawn from the better-off sections of the population (see sections A.5 and A.6 in chapter III). 77 The separation of expenditures into preventive and curative is not perfect. Until recently, for example, one of the national programs included under preventive expenditures involved upgrading equipment at pub- lic hospitals. 174 Table 8.2: Functional composition of public expenditure on health. by financing source. 1991-98 Year Ceniral Sate Budler L Local (PrOl incial Siati BJe1t Fandv Famill Cura- Preen- plan- PrelsLn- plin- 01' ue nq-Zing Other Total Cura.,ii, ive nint! Oilier Totai' 1991 56.07 9 56 296 31 40 10000 19Q91 "' 29 1255 4 55 11.61 I M.170 1992 73 27 2073 5 90 0 10 10000 1992 69 85 13 79 496 11 41 1(10 00 1993 83.28 t 60 9.53 0 59 100.00 1993 6674 13.6S 551 14 US 101100 1994 529 Q 2462 20 159 10000 lo94 6 5 7 1399 427 1300 101100 1995 44.07 23.16 20.70 12.07 100.00 1995 70.26 13.17 3.29 13.27 100.00 1996 38.74 35.90 18.58 6.78 100.00 1996 70.97 11.74 4.14 13.15 100.00 1997 33.17 27.30 22.00 17.53 100.00 1997 79.82 10.59 3.04 6,55 100.00 1998 31,22 22.92 25,73 20.13 100.00 1998 81.53 9.82 3.09 5.55 100.00 All Public Spending (including User Fees and Year ODA Grants and Loans Year HLealth In.sturanL e Fa,nul Faul, Cira- Preven- plan- Prcven- plan- tive mve ning Other Totl/ Curailc IlLf nine Other Total 1991 8529 13 "4 097 000 100 00 1991 70 55 11 29 3 52 1464 10Ii 00 1992 51.28 4861 0 12 000 100(00 1992 7C0 12 1902 4 39 6.48 1 (J(I0 1993 65 38 34 26 A 36 0.00 10000 1993 -'3.S5 13 02 5.-6 7 37 1 (irno 1994 59 24 40.31 0.45 0.00 10000 1994 65.59 1,1 37 890 7.13 100.00 1995 75 76 15 17 0.39 %.68 100.00 1995 6b 22 14 RX S 10 10.80 10001) 1996 62.50 1.39 N 20 9.92 100.00 199t 65 x2 I' 30 7 90 8 99 100 011 1997 46.55 2089 13.98 18.59 10000 1997 '7 1420 8 47 9.47 100 00 1998 31.11 17.63 3460 16.66 100.00 1998 69.72 11 42 10 31 S.55 1000)0 Source Ministry of Finance data Second, the curative and preventive shares in total public health spending have not changed much over time. The curative share has remained in the range of 66-74 per- cent over the period 1991-98, while the preventive share has had a wider range-1I per- cent to 19 percent. Third, there has been a sharp increase in the share of family planning in total pub- lic health spending-from 3.5 percent in 1991 to 10.3 percent in 1998. Almost all of this increase has been the result of significantly increased funding of family planning activi- ties by the central government and international donors. Fourth, the relative constancy of the curative share in total public health spending masks important changes that have taken place in the curative bias of central versus local government expenditures. The proportion of central health spending on curative interven- tions fell very sharply between 1993 and 1998-from 83.3 percent to merely 31.2 per- cent. Correspondingly, the shares of preventive programs and family planning in- creased-from 6.6 percent to 22.9 percent in the case of preventive programs and 9.5 percent to 25.7 percent in the case of family planning. Indeed, over this period, central spending on hospitals and curative care declined even in nominal terms. In contrast, the proportion of local health spending on curative interventions increased from 66.7 percent in 1993 to 81.5 percent in 1998. What would explain the sharp decline in central government (budgetary) spending on curative health interventions? As noted in Chapter VII, health insurance and user fees have become important sources of financing in the health sector, especially since 1993. 175 Most of the health spending from user fees and health insurance revenues goes into cura- tive (hospital) activities. It is possible that the growth of these two alternative financing sources has allowed the central state budget to cut back its allocation to the hospital and curative sector. Fifth and finally, the data in Table 8.2 show wide fluctuations in the functional composition of ODA grants and loans over time. For instance, the share of preventive interventions in ODA health spending increased from 14 percent in 1991 to 49 percent in 1992, then declined all the way to 15 percent in 1995 and was up around 18 percent in 1998. This probably reflects the entry and exit of different donor programs, with varying priorities, in the health sector over this time period. Spending on Health Programs. Greater disaggregation of government expenditure on (vertical) health programs is obtained from Ministry of Health records. These (shown in Table 8.3) indicate that Ministry of Health spending on disease control programs in- creased by 15 percent annually in real terms over the 1993-97 period. Of course, some important reallocations took place within the health program budget, with funding to the malaria control program being cut in real terms by 40 percent and EPI, tuberculosis and leprosy control programs seeing major increases in funding. As noted earlier, the in- creased financing of hospitals by user fees and health insurance reimbursements has gen- erally permitted the Ministry of Health to increase funding of preventive programs. Table 83: Ministry of Health Expenditure on Health Programs, 1993-97 Iodine defi- Year Malaria ciency EPI AIDS TB Leprosy Nutrition Other* Total Expenditure in nominal terms (million VND): 1993 44,799 27,235 10,738 1,362 84,134 1994 46,693 17.518 32,961 9,880 107,052 1995 35,277 30,419 37,103 4,770 930 108,499 1996 43,232 28,796 40,690 6,766 22,073 2,689 1,904 14,606 160,756 1997 37,642 25,087 55,618 33,897 6,685 42,828 201,757 Expenditure in real terms (millions of constant 1989 VND): 1993 12,043 7,321 2,887 366 22,617 1994 11,481 4,307 8,105 2,429 26,322 1995 7,425 6,403 7,809 0 1,004 196 22,837 1996 8,617 5,740 8,111 1,349 4,400 536 380 2,911 32,042 1997 7,268 4,844 10,739 0 6,545 1,291 0 8,270 38,957 % share of each program intotal MoH expenditure on health programs: 1993 53.25 32.37 12.76 1.62 100.00 1994 43.62 16.36 30.79 9.23 100.00 1995 32.51 28.04 34.20 - 4.40 0.86 100.00 1996 26.89 17.91 25.31 4.21 13.73 1.67 1.18 9.09 100.00 1997 18.66 12.43 27.57 - 16.80 3.31 - 21.23 100.00 * Includes the hospital equipment upgrade program. Source: Ministry of Health estimates. 2. Economic Composition There are three important observations to be made on the economic composition of public spending on health. First, capital spending appears to comprise an unusually 176 large share of the state health budget, although this share has declined from 28 percent in 1991 to 21 percent in 1998 (Table 8.4). Nearly one-half of capital spending over the years (but rising to 82 percent in 1998) is on construction, which is again very large share in comparison to the health budgets of most developing countries. Unfortunately, more de- tailed disaggregation on the capital budget is not available, so it is not possible to isolate the items responsible for the large construction spending. Second, only about 29 percent of the recurrent health budget is spent on salaries and wages, while 59 percent is spent on goods and services. Since health insurance reim- bursements and user fees are typically not used to pay for salaries, the share of salaries and wages in total government health expenditure (i.e., state budget plus insurance plus user fees) is even smaller-23 percent.8 In this sense, Vietnam is unlike other develop- ing countries which spend the majority share of their recurrent health budget on salaries and wages, and have very little to spend on consumables such as medical supplies and drugs. Table 8.4: Distribution of public health spending by source and by economic type, 1991-98 Recurrent Capital 4% % of total capital erpendaure Ay % oftotal recurrent expenditure erpendaure exrpenditure as ' of Subsidies as % o Purchase Capital total Saluries Goods and and trans- tutal Ck- iflted construc- Year expendiure and 1,ages services k;rs Otheri pendaure a%sets lion State Budget on Health (Cenrlpus Local) 1991 71.70 22.67 7129 5 59 0 15 28.30 54 48 45.52 1992 70.73 1634 76.79 634 053 29.27 67.70 32.30 1993 76.22 27 45 69.57 2.56 043 23.78 59 19 40.81 1994 77.99 3292 62.32 4.74 003 22.01 53.07 46.93 1995 76.21 31 36 62 51 6.10 002 23.79 48 25 51.75 1996 77.51 2922 65.75 5 02 0.00 22.49 40 38 5962 1997 78.63 27.59 6'7.82 4 58 U 01 21.37 5260 4740 1998 79.08 29 38 59.38 402 722 20.92 1.73 82.27 Total Public Sriending on Health (including User Fees and insurance) 1991 73.34 20.88 '3.56 5.43 0.14 26.66 54.48 45.52 1992 72.42 1504 78.64 5.83 0.49 27.58 67.70 3230 1993 78.02 2499 72.00 2 31 070 21.98 59.19 40.81 1994 80.07 29.41 65.84 4 18 0 z7 19.93 5307 4693 1995 79.49 2633 67.99 5.04 0.64 20.51 48 25 51 75 1996 81.39 23 47 71 91 3 96 066 18.61 40.38 59.62 1997 82.31 22 19 73.63 3 62 0.56 17.69 52.60 47.40 1995 82.77 2294 67.80 3 08 6.18 17.23 17 25 8275 Aotes- Figures in bold are percentages of total health spending The remaining figure- are c%pressed as percentages of total recurrent or capital expenditure. Source Ministrm of Finance data Third, the amount spent on subsidies and transfers to the poor (via fee exemptions or free health cards) constitutes only 4 percent of the state recurrent health budget and 3 percent of recurrent public spending on health, in spite of the fact that high hospital user 78 User fees may be used, however, for paying bonuses to health workers. 177 fees reduce economic access to hospital services for the poor and low-income groups. Additionally, the share of subsidies and transfers appears to have declined over time. This trend is not consistent with the government's stated objective of improving equity in health and providing greater access to public health services for the poor. 3. Composition by Service Level Unfortunately, budgetary data in Vietnam are not organized by service levels, so it is difficult to know how much of the (central and provincial) state budget is devoted to primary, secondary and tertiary care. A rough decomposition of budgetary health spend- ing into spending by levels is shown in Table 8.5. This decomposition is not perfect, as it is not easy to allocate many items of expenditure, such as training and communicable disease programs, to specific service levels. Therefore, the estimates shown in Table 8.5 should be treated with caution and are only indicative. These suggest that hospitals take up a disproportionately large share (75-87 percent) of the recurrent state health budget. There are two additional caveats to these numbers. First, the figures shown in Table 8.5 only reflect recurrent health expenditures financed by the state budget (both central and provincial) and exclude recurrent expenditures financed by donors, user fees and health insurance. Since virtually all of the spending out of user fees and health insurance goes into hospitals, it is likely that the share of hospitals in total recurrent public health spend- ing is even greater than that shown in Table 8.5. Second, first referrals to district hospitals are very much part of an integrated primary care system. Therefore, if spending on dis- trict hospitals is excluded, the share of central and provincial hospitals in the total stage budget reduces to about 50 percent. In contrast to hospitals, commune health centers account for a much smaller share of the state health budget. However, the commune health center share in the recurrent state health budget increased considerably over time-from 7 percent in 1991 to 12 per- cent in 1996, reflecting the fact that commune health center worker salaries began to be paid out of the state budget from 1994-95. The large share of hospitals in the state health budget has important equity impli- cations. As noted in chapter III, the majority of users of hospital-based services in Viet- nam, as in many other developing countries, are individuals belonging to the higher in- come quintiles. In 1998, 36 percent of all hospital users were drawn from the top quintile of the population, while only 7 percent were drawn from the bottom quintile. The fact that more than three-quarters of the recurrent state health budget is spent on hospitals thus implies that public health subsidies disproportionately benefit the better-off segments of Vietnamese society. 178 Table 8.5: Recurrent Health Sector Expeeditures, 1991-98* Esimated actual expenditures Budg, i (plannedi (million VND) (nullion VND Program category 1991 1992 1993 1994 1995 1996 199" 199 Central admini- stration and man- agement 2.730 3.450 4.100 4.50 4.60. 5.0K 7.700 7.700 (0.40) (0 43) 10 41) t0.39) t0.34) (0.35) (0.44) (0.44) Commune health centers 50,000 50,000 70,000 70,000 150,000 200,000 200,000 240,000 (7.30) (6.22) (7.03) (6.02) (11.03) (11.79) (11.51) (13.83) Intercommunal polyclinics 14,942 21,347 22,185 29,974 24,798 23,565 24,000 (1.86) (2.14) (1.91) (2.20) (1.46) (1.36) (1.38) District health centers 86,100 46,100 36,960 83,880 156,200 273,643 231,143 (12.57) ((5.73) (3.71) (7.21) (11.48) (16.14) (13.31) All hospitals: 546,000 690,000 864,000 983,080 1,019,602 1,191,422 1,274,603 1,464,200 (79.73) (85.77) (86.71) (84,48) (74.95) (70.26) (73.38) (84.35) District hospitals 128,900 248,400 311,040 318,120 363,773 402,357 458,857 527,112 (18.82) (30.88) (31.22) (27.34) (26 741 (23 73) (26.42) (30.37) Provincial hospi- tals** 324,300 324,300 406,080 488,330 477,032 599,063 548,390 688,174 (47.35) (40.31) (40.75) (41.97) (35 0') (35.33) (31.57) (39.64) Other hospitals (including cen- tral)*** 92,800 117,300 146,880 176,630 178,797 190,002 267,356 248,914 (13 55) t14 58) (14 74) (15 IS) (13 14) (11 20.' 115 39) (14 34) Total 684.830 804.492 996.407 1.163.645 1.360.376 1.695.763 1.737.011 1,735.900 (100.00) (100.00) (100 001 (100.00) f100 00'1 (100.00) (100.00) 1100 0) Include, -tate budget (central and pro%incial) expenditures only. Does not include ODA sources Includes both general and specialty hospitals. *0 Includes teaching hospials. spccialts hospitals and hospitils for special populations Noic Numberm in parentheses are percentages of the total Source Calculationb by Chinh 1999. Obviously, this is not an intentional outcome of government policy. However, as noted earlier, hospitals throughout the developing world tend to attract more affluent pa- tients because of their location (typically urban), perceived high quality, and high costs. If public health subsidies are to be better targeted to the poor, there will probably need to be a further reduction in the share of the state health budget spent on secondary and tertiary hospitals. 4. Provincial Distribution Another important dimension of equity in public spending is captured in the pro- vincial distribution of public health expenditure. There are three important questions here: (i) how is the level of per capita public spending on health in a province related to its GDP per capita? (ii) does public spending on health compensate for or reinforce pro- vincial disparities in private (i.e., household) health spending? and (iii) have interprovin- cial disparities in per capita public spending on health widened or narrowed over time? 179 Eciutv n Povin ialAl-Figure 8.5: Total public spending on health per capita in the proAime and Equity in Provincial Al-pmiiaGDpecpt,1S locations. Provincial spending 120 on health per capita varies from a low of VND 20,000 in Nghe wo . An to a high of VND 313,000 in Da Nang in 1998, but with the vast majority of provinces hav- ing a public spending level of 60 VND 25,000-50,000. The high levels of public health spending in Ho Chi Minh City, Da Nang 40 and Hanoi might reflect the fact that these provinces are also 20 - ----. partly responsible for provision 1,000 2,000 3,000 4,000 5,000 6,000 7,000 000 of public health services in sur-Provincial doestic product per capita (000 VND) rounding provinces. Figure 8.5 shows that public health spending per capita in a province is strongly related to provincial domestic product per capita, even after Ho Chi Minh City, Da Nang and Hanoi are excluded owing to the special nature of their provincial health services. The positive association between government health expenditure and provincial GDP per capita is not entirely unexpected, since a portion of government health expenditure in a province is financed by provincial governments out of their own revenue sources. This allows local governments in better-off provinces to spend more on health than those in poorer provinces. It is also likely that central government health subsidies to provinces, especially in the form of central hospitals and research institutes located in the provinces, are biased in favor of better-off provinces because these provinces have higher absorption capacity for funds owing to their better infrastructure, facilities, and administrative re- sources. Public and Private Spending on Health. An important policy question is the rela- tionship between public and private (household) spending on health across provinces. Figure 8.6 shows that prov- Figure &6: Public aod priNate health spending per capita, by pTovinCe. inces having higher levels of 120 -- - -- private spending on health 110 per capita also have higher levels of government health 7 90 expenditure per capita. This F means that public spending IF on health is not being used K appropriately as a policy tool for reducing the inequi- ties inherent in private 2 40 health care spending. 31 0 1000 2,000 3,000 4,000 5,000 6,000 7.000 8,000 9,000 10,000 Private heali spending per capita ('000 VND) 180 Provincial Spending Disparities Over Time. Mean provincial government health expenditure per capita grew by about 14 percent per annum in real terms between 1991 and 1998. Provinces that Figure 8.7: Percentage annual increase in real public spending on health per had lower levels of public capita across provinces, 1991-98 as related to the level of real public spending on health per capita in 1991 health spending in 1991 ex- 30 perienced more rapid growth in public health 25 spending than provinces that had lower levels of public 2 spending in 1991 (Figure _ 8.7). As a result, provincial I disparities in per capita pub- M lic spending on health nar- rowed somewhat between 5 . 1991 and 1998. While this is a positive trend, it needs to 5 10 15 20 25 30 35 40 45 50 55 60 65 be accelerated further. Pubic spending on health per capita in 1991 ('O0 1994 VND) D. Overseas Development Assistance to the Health Sector Data on overseas development assistance (ODA) to the health sector are notori- ously inconsistent in many countries. In Vietnam, there are several different estimates of the total amount of ODA to the health sector-those reported by the Ministry of Finance, Ministry of Planning and Investment, Department of Planning in the Ministry of Health, Projects Coordination Department of the Ministry of Health, and the donors themselves, most of which do not agree with each other. One important reason for the inconsistency of estimates is that some of the ODA is channeled through the state health budget, while some of it is off-budget and goes directly to supporting local donor projects. Another rea- son is that different donors have different ways of calculating their contributions. The Projects Coordination Department of the Ministry of Health, which helps to coordinate donor and NGO assistance in the health sector, reported that, as of December 1998, there were a total of 179 ongoing ODA-assisted projects in the health sector, with a total ODA commitment of US$ 668 million (MoH 1998e). Vertical (disease) programs, family planning and primary health care each accounted for about a quarter of the total commitment. The remaining commitment was divided between hospitals and clinics (19 percent) and sector policy and planning (7 percent) (MoH 1998e). The 179 projects in the health sector, of which 99 were under the Ministry of Health, were financed by 19 bilateral donor countries, four U.N. agencies, the World Bank, Asian Development Bank and the European Community. In addition, as noted in chapter IV.D, there are 106 NGOs working in the health sector. 1. Distribution of ODA by Donor Of the 99 ODA projects in the Ministry of Health with a total ODA commitment of US$ 470 million, US$ 143 million had been disbursed through the end of 1998. The distribution of ODA by donors, which is shown in Table 8.6, indicates that the six largest contributors to the health sector account for over 85 percent of total ODA resources to the 181 sector. These are the World Bank, the European Community, Japan, UNICEF, Sweden and Australia. Other agencies-UNFPA, WHO, the Netherlands and other bilateral do- nors-have functions or subsectors within which they play a crucial role. Table 8.6: ODA commitments and disbursements in the health sector through the end of 1998, by country Total commament .4mount dahur.sed Donor agenc prqlects USS mill -6 of oral SS mdl % of total World Bank I 127.30 27 11 14.40 10.06 European Community Q 00 40 19 25 070 049 Japan 3 82.60 17.59 19 60 1369 LNICEF 8 42.20 8.99 1.00 12 58 Sweden 9 31.20 6.65 20.60 14.39 Australia 8 2600 5 54 1000 699 UNFPA 4 1200 2.56 11.90 8 31 WFP I 1059 2 2o 8.45 590 WHO 25 9.38 200 10.28 7.18 France 4 8 50 1 81 5.40 3.7- Netherland 8 7.80 1.66 8.70 6 ON Ital% 2 6.70 1.43 2.00 1.40 Belgium 7 5.90 1 26 4 30 3110 Spain 5.00 1 06 500 3.49 Thailand 1 2 10 045 2 10 1 47 Denmark I 080 0.17 0.80 0.56 LISA 1 0.41 0.09 0 29 0 20 CIDSE 1 0.30 006 030 0 21 United Kingdom I 0.29 006 0.22 0.15 Republic of Korea 1 0.03 0.01 0.09 006 Total 96 46950 100.00 143.13 100.00 Source MoH 1998e. .Vote Only data for Ministr-, of Health projects are captured in this table. Thus, the table excludes LISS 41 million of Asian Development Bank and USS 50 million or World Bank .ssisance prc%ided in 1996 for family planning and reproductive health activities under a project implemented by the National Committee for Population and Family Plan- ning The annual flow of resources going to the health sector has more than doubled between 1991 and 1998, with US$75 million of ODA being disbursed to the sector in 1998 (Table 8.7). However, as noted in Chapter VII, the government's own resources al- located to the health sector, as well as other sources of revenue such as user fees and health insurance, have increased at even faster rates, so that the share of ODA in total public health spending has actually declined between 1991 and 1998. Table 8.7 shows that in 1991, there were few bilateral donors. The World Bank had not begun its assistance program to the Vietnam health sector either. As a result, most of the ODA was from the U.N. agencies and from bilateral donors like Sweden and Australia. However, by 1998, there were considerably more bilateral donors, and the World Bank alone was disbursing US$20 million. 182 Table 8.7: Overseas develptent assistance to the health sector (annual disbursements in '000 US$) 1991-98 Donor 1998 Total 1991-9S '/, of 4genc 1991 1992 1993 1994 1995 1996 1997 Amount ?ofttol 1991-98 total UNICEF 4.174 5.124 6.811 6.939 7,888 3.604 3.668 3.013 4.03 41,220 932 LINFPA 3.530 1.121 4.863 1.783 2,774 4.115 4.605 5,040 6.74 27.839 6 29 1VFP 5.80S 5.808 5.80S 5,808 2.552 2,234 1,993 1.907 2 55 31.917 7 22 WHO 2.697 1.461 2.923 3.160 3,445 3.500 3.510 1.49 2.00 22.195 5.02 UNDP 700 700 0.16 EU/EC 158 1,643 3,489 589 2,330 865 1,750 12,158 I1 25 22.982 5.20 World Bank 4 3,700 20,000 26 73 23,704 5.36 Japan 2,000 14,900 19,300 14,100 3,600 16,500 13,800 18 44 84.200 19.04 Sida 6,713 2,902 5,804 5,524 8,973 7,240 4,253 6,800 9.09 48,208 10.90 Australia 4,438 5,634 5,615 7,205 6,734 6,436 5,105 3,607 4.82 44,774 10.12 France 998 1,024 1,234 2,435 2,131 4,198 3,799 735 098 16.553 3.74 Nether- lands 0 266 821 2,188 2,364 3,848 5,308 412 0.55 15,208 3.44 Germany 0 1,100 1,975 1,900 180 1,800 2,620 2f670 3.57 12,245 2.77 Belgium 270 80 773 1,054 1,292 1,885 1,724 464 0.62 7,541 1.71 Italy 8 41 440 2,634 1,507 577 600 0.80 5,806 1.31 Spain 750 2,125 2,125 2.84 5,000 1.13 Thailand 71 241 1,052 1,052 2,416 0,55 Canada 329 954 618 1,902 0.43 Republic of Korea 1,760 63 1,823 0.41 U.K. 147 217 217 174 104 860 0.19 NGOs 5,822 3,372 5,395 2,700 3,400 4,514 25,203 5.70 Total 34,607 31,697 60,739 61,482 63,060 48,458 67,421 74,830 100.00 442,295 100.00 Source: 19io 1998e. 2. Provincial Distribution of ODA ODA assistance to the health sector is highly concentrated among a few prov- inces. For example, nearly one-third of total disbursements of external health sector assis- tance over the period 1991-98 went to Ho Chi Minh City, and another 9 percent went to Da Nang. Table 8.8 shows that 15 of the largest recipient provinces in the country ac- counted for 82 percent of the total ODA to the health sector over the 1991-98 period. These provinces accounted for only 31 percent of the country's population, however. 183 Table 8.8: Distribution of total ODA disbursements to the health sector over the period 1991-98, by province Cumulative Share in Cumulative Share in share in national share in na- national national population tionalpopula- Province ODA (/) ODA (0) (4 lion I/) 1 Ho Chi Minh City 31.91 31.91 6.67 6.67 2 Da Nang 9.36 41.28 0.88 7.56 3 Bac Giang 5.56 46.84 1.95 9.51 4 Quang Ninh 4.32 51.16 1.24 10.75 5 Ha Noi 4.31 55.48 3.05 13.79 6 Nghe An 3.66 59.14 3.80 17.59 7 Binh Dinh 3.37 62.51 1.95 19.55 8 Lai Chau 3.29 65.80 0.73 20.28 9 Hoa Binh 2.89 68.69 1.02 21.30 10 Ha Tay 2.83 71.52 3.13 24.43 11 HaiPhong 2.72 74.24 2.24 26.66. 12 Lao Cai 2.25 76.49 0.77 27.44 13 Tra Vinh 2.00 78.49 1.33 28.76 14 Quang Nam 1.86 80.35 1.82 30.58 15 Tuyen Quang 1.82 82.17 0.90 31.48 Source: MoH 1998e. Targeting of ODA to a Figure 8.8: ODA health-securdisburserets perecpita (US$) over te period few provinces might make sense 1991.9, by province if the selected provinces are among the poorest in the coun- 10 try. However, Figure 8.8, which plots per capita ODA to each province against the province's per capita GDP, shows no sys- tematic relationship between the 4- two variables. This suggests that health sector external assistance is targeted to provinces on the basis of donor preferences and priorities rather than on the basis 5,10)0 7.000 9.000 1.000 of a province's income level. Provincial doinestic product per capita (000 VND), 1997 3. Donor Coordination There is concern in all developing countries receiving large amounts of ODA from several different donors and aid agencies about the need for coordinating the activi- ties of the various donors. This is particularly the case in Vietnam, where, up to the early 1990s, the government had limited experience with external assistance to the health sec- tor. Also, the Ministry of Health is a relatively small ministry and thus has limited man- agement and supervision capacity. The monitoring and coordination of more than 99 on- going ODA-assisted projects in the health sector, with a total commitment of US$470 million, places a significant administrative burden on the Ministry. The result has been that the Ministry has sometimes had too little involvement and participation in identifying and designing ODA-funded projects. Furthermore, it should be recalled that the potential 184 for coordinating support from different donors is limited by the fact that donors have their own specific objectives and policies, which may differ both within the donor community as well as with the polices of the Government. The above considerations have had four negative consequences: first, some do- nor-assisted health projects have not always reflected government priorities in the health sector. Second, the lack of coordination between individual donors and the government as well as among the donors has resulted in overlap, duplication and inconsistency among different donor and government programs. This has reduced the overall effectiveness of aid to the sector. Third, the lack of participation of the Ministry in project identification and design has meant that Ministry staff have missed out on valuable opportunities to de- velop their capacities in this area. Fourth and finally, implementation of these donor- financed projects has suffered as the local partners have not felt ownership of the pro- jects. The overall objective of donor coordination should be to facilitate and strengthen the ownership of the Government and Ministry of Health with regard to the policy devel- opment process within the health sector. Government and Ministry ownership should also be strengthened with respect to specific aid-supported programs and projects. The same situation applies to monitoring and evaluation of donor-funded projects. This is often left to the donors, partly because the need and the plan for evaluation are of most interest to the donors and partly because of overstretched human resources in the Ministry. As a result, little is known about the effectiveness of ODA assistance to the sec- tor. A comprehensive study of the effectiveness of ODA should be conducted in the near future. However, there is optimism that the situation may be changing. Indeed, this re- port-the Vietnam health sector review-is itself an outcome of joint Ministry of Health- donor-NGO collaboration. Increasingly, donors, NGOs, and the Ministry of Health have shown more interest in coordinating their activities and in leveraging their interventions and projects around a common set of objectives and priorities. 185 IX. SUMMARY OF MAJOR FINDINGS A. Past and Present Trends The Vietnam health sector review provides clear evidence that Vietnam has made very impressive gains in its health indicators during the last three to four decades. It has achieved what few low-income countries have been able to achieve-a reduction of in- fant, child and maternal mortality to levels that are typically observed in countries that have two or three times the per capita income of Vietnam. And it has done this, for the most part, without much external assistance. In addition, there is compelling evidence that the situation in the health sector, which was somewhat precarious at the start of the decade of the 1990s, has improved considerably during the past decade. Public spending on health has increased impres- sively in real terms as well as in relation to GDP and total government spending since 1991. The government has taken over the responsibility of paying the salaries of health workers at the commune level since 1994-95. While health worker salaries continue to remain low in absolute terms and in relation to salaries in the private sector, they have risen considerably in real terms. Shortages of drugs, which were endemic throughout the system but particularly severe in the commune health centers in the early 1990s, are largely a thing of the past. An improved supply and distribution system for drugs has meant that private out-of-pocket expenditures on health care, while still large for the poor in relation to their incomes, have declined substantially in real terms for every income group. Communicable disease programs are much better funded, and this is evident in the dramatic decline in the prevalence of malaria during the 1990s. Even in the area of child malnutrition, which had remained at a stubbornly high level for a long time, there is re- cent evidence of improvement; the latest data from a national survey indicate that there has been a significant decline in child malnutrition during the last five years. Developments Outside and Within the Health Sector. Of course, some of these positive developments are the result of events outside the health sector. The Vietnamese economy grew at a rapid rate of 8.4 percent per annum in real terms between 1991 and 1997 (although the growth has slowed considerably since 1997). The growth of the econ- omy has allowed real public revenues to grow at an annual rate of 17 percent per annum and overall public expenditure at an annual real rate of 16 percent. This has certainly made it easier for the government to increase public spending on health. At the same time, household incomes and living standards have increased across the board. The two rounds of the VLSS indicate that per capita consumption expenditure increased by 49 percent in real terms between 1993 and 1998. Even among the bottom 20 percent of indi- viduals in 1998, consumption expenditure per capita grew by 31 percent in real terms over the period.80 Evidence from around the world shows that improvement in utilization 80 Of course, the fact that the growth of mean consumption expenditure per capita was greater than the growth of per capita consumption among the poorest 20% of the population indicates that consumption inequalities increased over the period. 186 of health services, health status, and child nutritional status is associated with improve- ment in living standards. Yet it would be wrong to argue that improvements in the sector have occurred primarily due to events outside the sector. There are many positive steps that the govern- ment has taken to improve the functioning of the health sector, and some of these steps have had important impacts. For example, the decision to deregulate the production and distribution of drugs has had a profound effect on improving the availability of and re- ducing the price of drugs to consumers. Likewise, the decision to allow private practice has increased the population's access to health services and offered people with a choice of health providers. The government's decision to pay salaries of commune-level health workers out of the state budget was an important step in maintaining the viability of the commune health center network. In addition, the fact that the state health budget has in- creased at a faster rate than GDP or total government expenditure since 1991 suggests a strong political commitment to health. The introduction of user fees at hospitals and the establishment of a health insurance program have generated additional resources for the health sector, especially at the secondary and tertiary level of care. The above are only a few examples of specific policy interventions in the health sector that have contributed to the improved functioning of the sector and that have had important effects on consumer choice and welfare. Challenges. Despite the very significant achievements of the last few years, major challenges lie ahead. Indeed, it could be argued that the easy gains in the sector have al- ready been achieved, and further gains will be difficult and will require even greater ef- fort and resources. In addition, the period of reform and renovation, while generally suc- cessful, has led to a new set of problems. Two examples of such problems are increased inequality in health opportunities and dangerously escalated levels of antibiotic resistance in the population. The health sector is ultimately a microcosm of society; therefore, it is not surpris- ing that the overall increase in interpersonal and inter-regional income inequality in Viet- nam associated with economic reforms and liberalization has manifested itself in the health sector in the form of some widening disparities in access to health services and in health outcomes.8' The better-off sections of Vietnamese society have been more able to take advantage of opportunities in the health sector, while the poor have lagged behind. An example of this is the dominant use of public hospitals by the better-off segment of the population. Another example is the greater representation of the top income quintile among health insurance members in the country. This implies that the many benefits of health insurance as well as of subsidized public hospital services are available dispropor- tionately to the better-off and not enough to the poor. Both of these outcomes are not only antithetical to the intentions of the government, but have occurred despite the best inten- 81 See World Bank 1999b (Chapter 4, "Ensuring Equity") for a description of these societal trends. The report makes the important point that despite a modest increase in income inequality between 1993 and 1998, Vietnam remains a moderately egalitarian society by international standards. Its level of inequality, as measured by the Gini coefficient, is 0.35-lower than in other Southeast Asian countries (such as Indo- nesia, Thailand, Malaysia and the Philippines). 187 tions and interventions of the government to provide access to quality health services to the poor. The other example of a problem created by reforms in the health sector is that of increasing antibiotic resistance in the population. As drug prices have declined and the availability of drugs has improved, drug consumption, especially for self-medication, has spiraled out of control. For example, the average number of annual service contacts per capita with drug vendors and pharmacy shops has increased more than three times from 1993 to 1998. Indeed, drug vendors account for two-thirds of all health service contacts of individuals. The excessive and irrational use of broad-spectrum antibiotics, in particu- lar, has caused antibiotic resistance levels in Vietnam to reach epidemic levels. There is evidence that antibiotic resistance is now affecting the clinical management of all infec- tious diseases in Vietnam. This is a very serious problem that threatens to derail the sig- nificant achievements in the health sector, as Vietnam could lose the benefit of using cheaper early-generation antibiotics and may instead have to rely on more expensive later-generation drugs for the management and control of common infections. It is thus clear that future intervention efforts will need to focus both on improv- ing the poor's geographical and economic access to health services as well as on reducing excessive and irrational use of drugs in the population. In addition, as Vietnam has sig- nificantly reduced morbidity and mortality from infectious and communicable diseases, it finds itself facing a new generation of health problems associated with economic devel- opment, increasing affluence and aging of the population. New diseases, such as diabetes, cancer and heart ailments, are beginning to appear, especially in the urban areas, in part due to high rates of smoking and changes in the diet. HIV/AIDS has the potential of be- coming a major cause of morbidity and mortality in the future. And injuries and accidents are already the leading cause of mortality in the population. How Vietnam manages these emerging problems, while extending and consolidating its past gains in controlling his- torically prevalent communicable diseases, will affect the course of future developments in the health sector. Figure 9.1: Distribution of population by age, 1994-2024 Of particular importance 100 is the fact that Vietnam's popu- lation is aging at a rapid rate, 0 thanks to its success in bringing so 1860+ down fertility rates sharply over 70 15-59 the last two decades. For exam- o 00-14 ple, while the overall population - 50 will grow by about 31 percent in the next 25 years, the population aged 60 years and over will 30 - grow by 118 percent, while that 20 below the age of 15 years will 1o actually decline by 14 percent - (Figure 9.1). A larger proportion 1994 1999 2004 2009 2014 2019 2024 of the elderly in the population will mean increased morbidity rates, greater prevalence of diseases that afflict the elderly (such as diabetes, coronary disease, and cancer), and greater pressure on health services. 188 B. Attaining Health Goals in a Market Economy The market economy in Vietnam is real and robust. The Vietnamese government will have to pursue and attain its health goals within this market economy. In a market economy, the government is not the exclusive provider nor the only financier of health services. As such, it has a new role for itself-that of regulating the private sector. More specifically, the government needs to strengthen the legal framework for the private sec- tor, setting minimum quality-of-care standards, promulgating laws and decrees that gov- ern appropriate behavior of private and public providers, and implementing all of these regulations and laws through a system of regular inspections and monitoring. The regula- tory role of the government becomes particularly important in the context of a rapidly expanding private sector with considerable variation across private providers in the qual- ity of care (e.g., infection control, hygienic practices, and standards of practice protocols) and quality of drugs dispensed. In the transition to a market economy, out-of-pocket payments that individuals have to make for utilizing health services typically increase sharply. This often leaves the poor in a vulnerable position, unless there are well-functioning mechanisms for exempt- ing the poor from the increase in health care costs. It is therefore important for the gov- ernment to take on a direct role in financing health care expenditures of the poor by, say, providing them with prepurchased health cards that can be used to obtain free inpatient services and drugs at hospitals and clinics. Private health providers in a market economy are naturally limited to the provi- sion of curative health services. The government continues-and needs-to remain as the exclusive provider of preventive health services, such as immunization and control of communicable diseases.82 One of the key messages in health economics is that preventive care interventions should have the first claim on public resources due to the substantial externalities they create for society and due to the resulting underspending on them by individuals. C. Summary of Major Challenges for Vietnam's Health Sector There are a number of issues that this review has identified that will need to be addressed by policy. While a complete discussion of these issues and the specific inter- ventions needed to address these issues are beyond the scope of this report, the following is a summary of some of the most important challenges that lie ahead for the sector, based on the findings of the Review. 1. Important Risks to Future Health Relatively high rates of child malnutrition. Vietnam has been much more success- ful at reducing its infant and child mortality rates than in lowering its child malnutrition rates. Indeed, even with very recent declines, child malnutrition remains a major problem in Vietnam, with more than a third of children under the age of 5 years being under- weight and stunted. It is somewhat puzzling that the factors that drove declines in infant 82 While it is possible for the government to pay for these services and subcontract their provision to pri- vate parties, this is typically not done in most low- and middle-income countries. 189 and child mortality and fertility in Vietnam have not operated to similarly reduce child malnutrition rates. This suggests that child malnutrition rates respond to something more than health services coverage and income growth-they are likely linked to child rearing practices and cultural beliefs on child feeding that have been more difficult to change in Vietnam. High rates of induced abortion and menstrual regulation. The impressive rates of fertility decline in Vietnam have been associated with extensive use of menstrual regula- tion and induced abortion to terminate pregnancies, even though these are not considered as contraceptive methods by the Government. Some estimates suggest that a typical Viet- namese woman experiences 2.5 abortions during her reproductive age span-an ex- tremely high number by international standards. These high rates of abortion pose sub- stantial health risks to women. Extensive use of pregnancy termination is often inter- preted as indicating inadequate access to safe and affordable family planning services. High rates of smoking. Despite recent declines, Vietnam has one of the highest rates of male smoking prevalence in the world, with over one-half of men aged 15 and over being regular tobacco users. Given the association between smoking and several diseases, such as respiratory infections, lung cancer, heart disease, and tuberculosis, the high prevalence of smoking has serious implications for health status as well as the cost of health care in the country. Health problems of an aging population. Due in large part to its success in bring- ing down fertility rates sharply over the last two decades, Vietnam's population is aging at a rapid rate. It is estimated that, over the next 25 years, the elderly population (i.e., those aged 60 years and over) will increase more than five times as much as the overall population. A larger share of the elderly in the population will mean increased morbidity rates, greater prevalence of diseases that afflict the elderly (such as diabetes, coronary disease, and cancer), and greater pressure on health services. 2. Problems Relating to Essential Health Services Excessive reliance on self-medication and on drug vendors for treatment. While it is not uncommon for individuals in many developing countries to bypass the health care system and obtain drugs directly from pharmacies and drug vendors (i.e., self- medication), there appears to be excessive reliance on drug vendors as health care pro- viders in Vietnam, especially among the poor. Recent surveys suggest that drug vendors account for roughly two-thirds of all individual health service contacts (excluding visits to drug vendors for merely filling prescriptions obtained from another medical provider). Both self-medication and reliance on drug vendors for medical advice are dangerous, and have no doubt contributed to the high and rising rates of antibiotic resistance in the coun- try. Excessive and irrational use of drugs. The easier availability of drugs and a de- cline in the real prices of drugs over the last six years have resulted in an increase in the practice of self-medication. Self-medication has been associated, in particular, with the overuse and inappropriate use of antibiotic drugs, and this has caused antibiotic resistance levels in Vietnam to reach epidemic levels. This is an extremely serious problem that threatens to derail the significant achievements in the health sector, as Vietnam loses the ability to control and prevent the spread of many infectious diseases. The antibiotic resis- 190 tance problem is compounded by insufficient competence in clinical pharmacology and clinical pharmacy among pharmacists, drug vendors and the public. Even when drugs are prescribed by doctors, there is low compliance by patients with rational treatment guide- lines. Often, poor patients limit themselves to a two-day course of antibiotics (instead of the recommended ten-day course), as they do not have enough money to buy the full course. Heavy reliance on hospitals for treatment. The public health care system in Viet- nam is relatively hospital-intensive. Vietnam has more hospital beds in relation to its population than even Malaysia, Thailand and the Philippines-countries whose per capita GNP is several times larger than that of Vietnam. Not surprisingly, Vietnam has among the highest inpatient admission rates in the Asia-Pacific region. All of this suggests that hospitals, which are not the most cost-effective facilities to treat the types of diseases that prevail in Vietnam, are overutilized relative to primary health facilities, particularly among the better-off segment of the population. Unequal use of public hospitals. One of the major dilemmas facing the public health system in Vietnam is the inequality in use of public hospitals. In 1998, 36 percent of all hospital users were drawn from the top income quintile (top 20 percent) of the population, while only 8 percent were drawn from the bottom 20 percent. Further, the share of the bottom 20 percent in public hospital use has declined since 1993 (from 10 percent to 8 percent). The Government has instituted a number of programs to attempt to offset the inherent disadvantages faced by the poor, such as providing them with free health cards and exemption from hospital user fees. The empirical evidence suggests, however, that these important measures have not improved the poor's access to public hospitals to the extent intended.83 The dominance of the better-off among hospital users is of particular importance as more than three-quarters of the recurrent state health budget is spent on hospitals. 3. Issues in Health Care Financing High cost of health care, especially inpatient services, to the poor. Despite a de- cline in the real price of drugs in recent years, health care costs, especially user fees at public hospitals, are very large in relation to discretionary income for the poor. For ex- ample, a single service contact with a public hospital takes up 22 percent of all nonfood expenditure for a year for a typical person in the lowest quintile. Thus, a single catastro- phic illness involving extended hospitalization can wipe out years-indeed, a lifetime- of savings for the poor and drive them into debt. This compromises their ability to pay for unanticipated health expenses in the future. Large share of government health spending spent on hospitals, especially by the provinces. While hospitals take up a disproportionately large share of public subsidies for health in most developing countries, public expenditure data for Vietnam suggest that the public hospital share in the recurrent state health budget is exceptionally large-in the 831t is also the case that despite the large number of methods devised by the government to help the poor, the total amount of assistance provided to the poor to help defray medical costs remains low in absolute terms. For example, in 1997, only 3.6 percent of the total public spending on health in the country went toward subsidies and transfers. 191 range of 75-87 percent. Much of this is due to the large expenditure on public hospitals by local (provincial) governments-not the central government. Indeed, the share of the central government health budget spent on curative interventions (including hospitals) has declined sharply in recent years. However, since the local government health budget is more than two times as large as the central government health budget, the pattern of spending by the provinces has a profound effect on overall public expenditure patterns. Limited coverage of the health insurance program. While Vietnam's health insur- ance program expanded very rapidly in its first five years, starting from a zero base, health insurance coverage rates have leveled off in the last two years at about 12% of the population. Further, coverage is greater for the better-off segments of the population, as coverage is mandatory for civil servants and organized-sector employees, most of whom tend to belong to the upper income quintiles. For instance, only 6 percent of individuals in the lowest quintile have health insurance coverage, while 29 percent of individuals in the highest quintile do. In turn, this means that health insurance enrollees in Vietnam are drawn disproportionately from the better-off groups in society. In 1998, the share of the bottom 20 percent of the population among health insurance enrollees was merely 8 per- cent, while that of the top quintile was as high as 37 percent. Unequal spending on public health services across provinces. Spending on public health services is distributed unevenly across provinces, with better-off provinces spend- ing a great deal more per capita on health than poorer provinces, both because provincial health spending is financed by provincial governments out of their own tax revenues, which are strongly correlated with provincial incomes, and because better-off provinces are able to generate more funds from user fees and from the social health insurance. Moreover, the national health budget that falls under provincial control is allocated to the provinces by the Center on the basis of norms, such as the number of hospital beds in the province (for curative expenditures) and on the population of the province (for preventive expenditures), that do little to redress the inherent inequality in provincial health spend- ing per capita. Low wages of health workers. Health workers are one of the most important in- puts in the health care system. The quality of health services is thus synonymous with the morale and motivation of health workers. In Vietnam, the morale of public health work- ers is low because health workers, like all civil servants, are offered salaries that are very low. Average monthly salaries of health workers essentially have remained unchanged in real terms since 1994. In 1998, the average monthly salary of a government health worker was merely US$29. Public health workers thus sometimes seek additional sources of in- come, and this reduces their time, attention, and dedication to their work. 4. Health Sector Capacity Issues Large number of hospitals. As already noted earlier, Vietnam has more hospital beds in relation to its population than many other countries that have significantly higher per capita incomes. The analysis of hospital costs in this report indicates that district hos- pitals in Vietnam are characterized by economies of scale, implying that there probably are too many district hospitals in the country, with some being too small to provide ser- vices efficiently. Typically, in such a situation, there can be efficiency gains from con- solidating or combining smaller district hospitals, especially in cases where such a com- 192 bination would not reduce geographical access to the facilities. The Ministry of Health is already taking steps in this direction through its efforts to promote regional hospitals where appropriate. Mix of Health Workers. The number of doctors has increased much more rapidly during the last twenty years than the number of assistant doctors and pharmacists, while there has actually been a sharp decline (of about 57 percent between 1986 and 1996) in the number of nurses and midwives per capita. This gives Vietnam one of the highest doctor/nurse ratios in the Asia-Pacific region. As doctors are significantly more expen- sive to train than assistant doctors and nurses, the evolving mix of health workers needs to examined in light of Vietnam's disease profile and in light of cost-effectiveness con- siderations. More importantly, the mix of health workers may reduce access of the poor to health services, since many of the trained doctors are reluctant to move to the rural ar- eas, especially the poorest rural areas of the country. (Many doctors prefer to work in the urban areas, which offer better prospects for professional development and/or private practice, and thus end up serving the better-off sections of society.) This means that the poorest communes in the country not only miss out on doctors but also on the potential nurses, midwives and pharmacists who might have been trained with the scarce training resources currently allocated to physician training. Poor geographical access to health facilities in remote regions. While overall (geographical) access of the Vietnamese population to health facilities is generally excel- lent, there are a few regions where geographical access is inadequate. These are primarily the mountainous and remote regions of the Central Highlands and the Northern Uplands, which are home to 55 ethnic minority groups. An estimated 11 percent of the country's population-about 7-8 million people-live in these regions, which are among the poor- est and most isolated parts of the country. In these regions, population density is low and the terrain is difficult to navigate, so that traveling times to the nearest grass-roots health facilities are overly long (in some cases, a day or longer). Additionally, because of pov- erty, these regions find it difficult to attract health workers, and so access to health ser- vices, public or private, is very poor. Health worker training. While Vietnam has an abundant supply of health workers in relation to its population-indeed, it has more doctors per capita than Malaysia, Thai- land and the Philippines-there is substantial scope for improving the training and com- petence of health workers in the country. The Department of Science and Training of the Ministry of Health is responsible for quality control of training at all levels, but this de- partment has too few resources, both human and financial, to adequately monitor the quality of the training in the nine medical schools, 50 secondary medical schools, and the numerous projects and programs undertaking in-service training. Preservice training pro- grams in the health sector are sometimes quite theoretical, with little interaction and stimulation of students to learn how to use the information presented. Facilities and teachers for practical sessions and experience are limited; there are no official teaching hospitals; and most graduates begin practicing with too little practical experience. In ad- dition, there are too few resources allocated to retraining of staff, so the opportunities for health workers to keep up with new developments in the field are very limited. Use of information and data. Planning and management in the health sector in Vietnam are impeded by problems in the collection, analysis, and use of data and infor- 193 mation. There is a health information system in operation, but it is uneven in terms of its data collection. All facilities are expected to submit data on key indicators to the Ministry of Health at regular intervals. The Health Statistics and Informatics Division within the General Planning Department of the Ministry of Health consolidates some of these data from facilities. However, a good deal of information that could be readily obtained from facilities and administrative records is not collected or reported. For instance, data on fa- cility-level expenditures are not routinely collected or reported. While aggregate utiliza- tion data are collected, disaggregated data on utilization by key socioeconomic or demo- graphic groups (such as women, children, the elderly, and ethnic minorities) are not col- lected. Also, little information is collected on the number and type of private practitioners operating in communes, districts and provinces, with the result that not enough is known about the role of the private sector in health in the country. In addition, the information obtained by the health information system could be used more effectively for policy and planning purposes. For example, one possibility would be to calculate the utilization rates of commune health centers and to explore the reasons why some commune health centers consistently report lower utilization rates than other centers. Unit costs of health services could also be compared across hospitals to see which hospitals have additional scope for cutting costs and improving efficiency. Regulation of health providers. While there are many regulations and decrees governing the minimum quality standards and practice protocols expected of health pro- viders, the implementation of these laws through regular inspections of health facilities is less than satisfactory. In part, this is the result of a vast increase in the number of health providers, especially private health providers and drug vendors, in the last few years. The provincial health authorities, whose responsibility it is to undertake regular inspections of all private facilities, have neither enough inspectors nor an adequate budget for such in- spections. In addition, many provincial health bureaus do not even have a complete list- ing of all private providers in their province, with the result that only licensed providers are subject to inspections and those whose quality really needs to be monitored-the un- licensed providers-sometimes get away without inspections. There are also issues of conflicts of interest in the existing inspection process, as the members of the inspection team are often providers who work in the public sector and have their own private clinics. Donor Coordination and Aid Management. Donor assistance to the health sector in Vietnam has increased sharply (in absolute terms) from a very low base. As of the end of 1998, there were 179 ongoing ODA-assisted projects in the health sector, with a total commitment of US$668 million. Such a scale of donor assistance has placed a significant administrative burden on the already overstretched Ministry of Health. At the same time, donors sometimes tend to be reluctant to change their policies in the name of donor coor- dination. The result has been that the Ministry has sometimes had marginal involvement and participation in identifying and designing ODA-funded projects. This has meant that (i) some donor-assisted health projects have not always reflected government priorities in the health sector, (ii) there has sometimes been overlap, duplication, and inconsistency among different donor and government programs, and (iii) implementation of donor- financed projects has suffered as the local partners have not always felt ownership of the projects. 194 BIBLIOGRAPHY Abbatt, F. 1996. "An Analysis of Health and Health Manpower, Vietnam." Hanoi. Mimeo. Min- istry of Health. Akin, J. S., C. Griffen, D. K. Guilkey, and B. M. Popkin. 1984. The Demand for Primary Health Care in the Third World. Totowa, N.J.: Littlefield, Adams. . 1986. "The Demand for Primary Health Care in the Bicol Region of the Philippines." Economic Development and Cultural Change 34 (4):755-82. Alderman, Harold, and Paul Gertler. 1988. "The Substitutability of Public and Private Medical Care Providers for the Treatment of Children's Illnesses in Urban Pakistan." Living Standards Measurement Study Working Paper. World Bank. Washington, D.C. Barnum, H., and J. Kutzin. 1993. Public Hospitals in Developing Countries: Resource Use, Cost, and Financing. Baltimore, Md.: Johns Hopkins University Press. Birdsall, Nancy, and P. Chuhan. 1986. "Client Choice of Health Treatment in Rural Mali." World Bank, Population, Health and Nutrition Department, Washington, D.C. Bloom, Martin W., and Jonathan Gorstein. 1995. "Viet Nam: Xerophthalmia Free." National In- stitute of Nutrition (March). Brenzel, Logan. 1997. "Evaluation of the Unit Costs of Safe Motherhood Interventions in Viet Nam." Paper prepared for the Safe Motherhood Demonstration Project, Hanoi, Decem- ber. Brugemann, Ilona, and Barbara Franklin. 1997. Love and the Risk of AIDS for Women in Viet- nam: A Qualitative Study of Women in the Hanoi Area. CARE International in Vietnam and UNICEF. CARE Vietnam. 1997. An Overview of the Situation in Relation to HIV/AIDS in the Socialist Re- public of Vietnam. Hanoi, October. Chin, Le Due. Undated(a). "Hospital Activities in Vietnam." Ministry of Health, Hanoi. Unpub- lished working paper. Chinh, Le Duc. Undated(b). "Public-Private Mix System in Vietnam." Ministry of Health, Ha- noi. Unpublished working paper, Hanoi. Chuc, Nguyen Thi Kim, and Goran Tomson. 1996. "Doi Moi and Private Pharmacies: A Case Study of Dispensing and Financial Issues in Hanoi, Vietnam." European Journal of Clinical Pharmacology. Dai, Pham Xuan, Do Hong Ngoc, Truong Trong Hoang, and Christopher N. H. Jenkins. 1997. "Vietnam: A Tobacco Epidemic in the Making." National Center for Human and Social Sciences, Institute of Sociology, Hanoi. Deolalikar, Anil B. 1997. "The Demand for Health Services in Indonesia: The Role of Prices, Service Quality, and Reporting of Illnesses." In Aman Ullah and David E. A. Giles, eds., Handbook ofApplied Economic Statistics. New York: Marcel Dekker, Inc. 195 Dung, Pham Huy. 1997. "The Political Process to Increase the Private Health Sector's Role in Vietnam." In William Newbrander, Private Health Sector Growth in Asia: Issues and Implications. New York: John Wiley and Sons. Dung, Pham Huy. 1999. "Study on Private Health Sector Growth in Vietnam." Study commis- sioned for the Vietnam health sector review. Hanoi. Mimeo. Dung, Truong Viet, Nguyen Thanh Tam, Gill Tipping, and Malcolm Segall. 1994. "Quality of Public Health Services and Household Health Care Decisions in Four Rural Communes of Quang Ninh Province, Vietnam." Centre for Human Resources in Health, Ministry of Health, July. Dunlop, David W. 1999. "Situation Review of Health Financing and Health Insurance in Viet- nam." Paper prepared for the Asian Development Bank Rural Health Project TA, Hanoi, June. Economist Intelligence Unit (EIU). 1998. "Profile-Healthcare Provision in Vietnam." In Healthcare Asia (third quarter). Ensor, Tim. 1995. "Introducing Health Insurance in Vietnam." Health Policy and Planning 10(2):154-63. . 1996. "Health Charges and Exemptions in Viet Nam." Bamako Initiative Operations Research Programme, Research Paper No. 1, UNICEF, New York. . 1999a. "Developing Health Insurance in Transitional Asia." Social Science and Medi- cine 48:871-79. - . I999b. "A Preliminary Financial Analysis of Health Insurance in Viet Nam." Centre for Health Economics, University of York, York, U.K., June. Ensor, Tim, and Pham Bich San, 1996, "Access and Payment for Health Care: The Poor of Northern Vietnam." International Journal of Health Planning and Management 11:69- 83. Franklin, Barbara. 1993. The Risk of AIDS in Vietnam: An Audience Analysis of Urban Men and Sex Workers, with Guidelines for Prevention. CARE International Monograph Series No. 1, Hanoi, June. Fritzen, Scott. 1999. "Decentralization, Disparities, and Innovation in Vietnam's Health Sector." In J. I. Litvack and D. A. Rondinelli, eds., Market Reform in Vietnam: Building Institu- tions for Development. Westport, Conn.: Quorum Books. Gertler, Paul, and Jennie Litvack. 1998. "Access to Health Care during Transition: The Role of the Private Sector in Vietnam." In David Dollar, Paul Glewwe, and Jennie Litvack, eds., Household Welfare and Vietnam's Transition. Washington, D.C.: World Bank. Gertler, Paul, Luis Locay, and Warren Sanderson. 1987. "Are User Fees Regressive? The Wel- fare Implications of Health Care Financing Proposals in Peru." Journal of Econometrics 33: 67-88. Gertler, Paul, and Jacques van der Gaag. 1990. The Willingness to Pay for Medical Care: Evi- dence from Two Developing Countries. Baltimore: Johns Hopkins University Press. 196 Goodkind, Daniel M. 1995. "Vietnam's One- or Two-Child Policy in Action." Population and Development Review 21(1):85-111. GSO (General Statistical Office). 1990. General Population Census of VietNam 1989. Hanoi. 1995. Vietnam Intercensal Demographic Survey 1994: Major Findings. Hanoi. May. 1996a. Vietnam Intercensal Demographic Survey 1994: Abortion. Menstrual Regulation and Unwanted/Mistimed Pregnancies. Hanoi. October. . 1996b. Vietnam Intercensal Demographic Survey 1994: Birth Spacing and Child Mor- tality in Viet Nam. Hanoi. May. . 1996c. Vietnam Intercensal Demographic Survey 1994: Contraceptive Discontinuation and Failure. Hanoi. October. . 1996d. Vietnam Intercensal Demographic Survey 1994: Contraceptive Knowledge and Practice-Patterns and Differentials. Hanoi. May. . 1996e. Vietnam Intercensal Demographic Survey 1994: Education in Viet Nam. Hanoi. May. . 1996f. Vietnam Intercensal Demographic Survey 1994: Infant Feeding Practices in Viet Nam. Hanoi. May. . 1996g. Vietnam Intercensal Demographic Survey 1994: Knowledge and Use of Oral Contraception. Hanoi. October. . 1996h. Vietnam Intercensal Demographic Survey 1994: Study of Factors Related to Unmet Need for Contraception in Viet Nam. Hanoi. October. . 1997a. Vietnam Intercensal Demographic Survey 1994: Assessment of the 1994 VNICDS Data. Hanoi. January. . 1997b. Vietnam Intercensal Demographic Survey 1994: Population Structure and Household Composition. Hanoi. January. . 1997c. Vietnam Intercensal Demographic Survey 1994: Proximate Determinants ofFer- tility in Viet Nam. Hanoi. October. . 1999a. Population Projections of Vietnam: 1994-2024. Project report VIE/97/P 14. Ha- noi. May. 1999b. "Viet Nam Living Standards Survey 1997-98." Report No. VIE/95/043, August. 2000. General Population Census of Viet Nam 1999. Hanoi. Hannah, Joe. 1995. "Control of Diarrhoeal Disease Programme." Hanoi. November. Haughton, D., J. Haughton, S. Bales, Truong Thi Kim Chuyen, and Nguyen Nguyet Nga. 1999. Health and Wealth in Vietnam: An Analysis of Household Living Standards. Singapore: Institute of Southeast Asian Studies. Heller, Peter. 1982. "A Model of the Demand for Medical and Health Services in Peninsular Ma- laysia." Social Science and Medicine 16:267-84. 197 Hoi, Nguyen Dinh, Truong Dinh Kiet, Le Hoang Ninh, Truong Phi Hung, Nguyen Do Nguyen, Nguyen Bich Loan, Do Van Dung, and Bui Dai Lich. 1999. "Health Development in the Period of Renovation." Medical University of Ho Chi Minh City. Hospital of Endocrinology and Rainer Gutekunst. 1993. "Report on National Survey on Iodine Deficiency Disorders in Vietnam 1993." Hanoi. August. Hung, Pham Manh, I. Harry Minas, Yuanli Liu, Goran Dahlgren, and William C. Hsiao, editors. 2000. Efficient, Equity-Oriented Strategies for Health: International Perspectives- Focus on Vietnam. Melbourne: The Center for International Mental Health. Institute of Sociology and the Futures Group. 1997. "Family Planning and Reproductive Health among Ethnic Groups in Vietnam." Prepared for the NCPFP, November. Investment for Child Nutrition in Viet Nam. RETA document. April 1998. Jenkins, Christopher N. H., and others. 1997. "Tobacco Use in Vietnam-Prevalence, Predictors, and the Role of the Transnational Tobacco Corporations." The Journal of the American Medical Association 277 (June 4). Jimenez, Emmanuel. 1987. Pricing Policy in the Social Sectors: Cost Recovery for Education and Health in Developing Countries. Baltimore, Md.: Johns Hopkins for the World Bank. Jowett, Matthew, and Robin Thompson. 1999. "Paying for Health Care in Vietnam: Extending Voluntary Health Insurance Coverage." Centre for Health Economics, Discussion Paper No. 167, University of York, York, U.K., March. Jowett, Matthew, Nguyen Duc Vinh, and Peter Martinsson. 1999. "Voluntary Health Insurance in Vietnam: Preliminary Results from a Household Survey." Center for Health Econom- ics, York University, York, U.K. October 29. Mimeo. Khanh. 1995. The complete source was unavailable. Knodel, John, Phan Thuc Anh, and Dao Xuan Vinh. 1995. "Vietnam's Population and Family Planning Program as Viewed by Its Implementors." Regional Working Paper (South and East Asia) No. 2, Population Council, Bangkok. Malhotra, Kamal. 1999. "Basic Health Services for Vulnerable Families in Vietnam: Should User Fees Be the Way Forward?" UNICEF and Save the Children, Hanoi, April. McCall. E. Partnership with Government. The complete source was unavailable. MoH (Ministry of Health). 1991. Health Statistics of Viet Nam 1986-90. Department of Health Statistics and Informatics. December. . 1993. Health Statistics of Viet Nam 1990-92. Health Statistics and Informatics Centre. February. . 1994. Health Statistics Yearbook 1993. Bureau of Health Statistics and Informatics. February. . 1995a. Health Statistics Yearbook 1994. Health Statistics and Information Division. April. 1995b. Health Statistics Yearbook 1995. Health Statistics and Informatics Division. 1996a. Inventory of Hospitals Survey, Department of Therapy, Hanoi. 198 . 1996b. Health Statistics Yearbook 1996. Health Statistics and Informatics Division. - 1997a. Health Statistics Yearbook 1997. Health Statistics and Informatics Division. . 1997b. "A New Essential Drugs Policy Can Help Alleviate Six Key Problems of Pri- mary Health Care." Occasional Paper No. 5, PHC Unit, September. _. 1998a. "Economic Reform, Poverty and Equity in Access to Health Care: Case Studies in Viet Nam." Draft report of a collaborative research project between the Centre for So- cial Sciences in Health and the Institute of Development Studies, Hanoi, July. 1998b. "Health Development Plan." Hanoi. 1998c. Health Statistics Yearbook 1998. Health Statistics and Informatics Division. . 1998d. "Sentinel Surveillance Study." Planning Unit, Ministry of Health. Hanoi. Mimeo. . 1998e. "Overseas Development Assistance to the Health Sector." Project Coordination Department, Ministry of Health. Hanoi. Mimeo. . 1999a. "Evaluation of Twenty Years of Implementation of Primary Health Care in Viet- nam." Ministry of Health, Ministry of Defense, and Military Academy of Medicine, Ha- noi, December. 1999b. Health Service in Vietnam Today. Hanoi. 1999c. "The Mechanism of Collection and Spending of Hospital Fees." National Health Support Project, Component 3, Hanoi, April. . 1999d. Report of the National Workshop on Policy of Essential Obstetric Care. MCH/FP Department. March. . 1999e. "Review of Options for Securing Funds for the Health Care Needs for the Poor and Other Groups Exempted from User Fees." National Health Support Project, Hanoi, March. __ . 2000. "Strategy for People's Health Care during 2001-2010." Hanoi. Mimeo. MoH and IDS. 1998. "Study on Financing of CHCS before and after Government Decision." MoH and UNFPA. 1999. The complete source was unavailable. MoH and UNICEF. 1996. "A Study on Factors Influencing the Health Service Utilization by Households." Hanoi, August. MoH, WHO, and UNICEF. 1998. "Vietnam National Review for the Expanded Program of Im- munization." Hanoi, October. MoH in collaboration with the Rockefeller Foundation, Department of Health Services Victoria, and SIDA, eds. 1999. Conference on Efficient Equity Oriented Strategies for Health, 7- 10 April, Ha Long. Mwabu, Germano. 1986. "Health Care Decisions at the Household Level: Results of Health Sur- vey in Kenya." Social Science and Medicine 22 (3): 313-319. NCPFP (National Committee for Population and Family Planning of Viet Nam). 1990. Vietnam Demographic and Health Survey 1987. Hanoi. 199 . 1995. Reproductive Health Survey 1995 (VNRHS-95). Hanoi, June. . Forthcoming. Vietnam Demographic and Health Survey 1997. Hanoi. . 1996. "Promotion of Family Health in Five Provinces of Vietnam." Hanoi, November. Nga, Nguyen Nguyen. 1997. Health Sector Development and Economic Reform in a Transitional Economy: Vietnam, 1989-97. Thesis submitted to the University of Manchester for the degree of Doctor of Philosophy in the Faculty of Economic and Social Studies, Manches- ter, U.K., August. NIN (National Institute of Nutrition). 1995. "Report of the National Anemia and Nutrition Risk Factor Survey." Viet Nam. . 1997. "Results of the Vietnam National Iodine Deficiency Survey, Iodine Deficiency Control Program." May. NIN and UNICEF. 1999. Vietnam National Nutrition Survey. Hanoi. Paul-Shaheen, P., J. D. Clark, and D. Williams. 1987. "Small Area Analysis: A Review and Analysis of the North American Literature." Journal of Health Politics, Policy and Law 12:741-809. Phan, Vu Thi, Tran Quoc Tuy, Le Xuan Hung, and Ly Van Ngo. "Epidemiological Characteris- tics of Malaria in Vietnam, 1992--1997." Mekong Malaria Forum 1(December):22-33. Population Council. 1997. "A Study of Vietnamese Youth's Decision Making for Health and HIV/AIDS Prevention in Kien Giang and Quang Ninh Provinces." Hanoi, June. . 1998a. "Study on Sexuality in Vietnam: The Known and Unknown Issues." South and East Asia Regional Working Papers, No. 11, Hanoi. . 1998b. "Men's Attitudes towards Family Planning: A Pilot Study in Two Communes of Northern Vietnam." Research Report No. 8, Hanoi. . 1999. "A Situation Analysis of Private Reproductive Health Services in Five Provinces in Vietnam." Report for the NCPFP Population and Family Health Project, Hanoi, Sep- tember. Read, Paul, Harry Minas, and Steven Klimidis. 1999. "A Preliminary Exploration of Life Expec- tancy, Wealth and Economic Development." in Ministry of Health in collaboration with the Rockefeller Foundation, Department of Health Services Victoria, and SIDA, editors, Conference on Efficient Equity Oriented Strategies for Health, 7-10 April, Ha Long. Ron, Aviva, and Guy Carrin. 1996. "The Development of National Health Insurance in Viet Nam." Macroeconomics, Health and Development Series Paper No. 23, WHO, Geneva, October. Schwartz, J. B., J. S. Akin, and B. M. Popkin. 1988. "Price and Income Elasticities of Demand for Modem Health Care: The Case of Infant Delivery in Indonesia." The World Bank Economic Review 2 (1): 49-76. Segall, Malcolm, Gill Tipping, Henry Lucas, Truong Viet Dung, Nguyen Thanh Tam, Dao Xuan Vinh, and Dao Lan Huong. 2000. "Health Care Seeking by the Poor in Transitional Economies: The Case of Vietnam." IDS Research Report No. 43, Sussex, U.K., April. 200 Solon, Orville, and Tran Van Tien. 1997. "The Challenges of the Vietnam Health Insurance Pro- gram." Hanoi, June. Tornquist, S. 1999. "Vietnam-Sweden Health Cooperation in the Area of Drug Policy and Con- trol Project Document." Ministry of Health, Hanoi, April 24. Mimeo. Trong, Le Ngoc, and others. 1998. "Evaluation of the Smoking Situation in Vietnam (Danh gia Thuc Trang Tinh hinh Hut thuoc la o Viet Nam)." Ministry of Health, Hanoi. Mimeo. UNDP (United Nations Development Programme). 1996. "Viet Nam Public Expenditure Re- view." Report VIE/94/025. Hanoi, May. --. 1999. Human Development Report 1999. New York: Oxford University Press. UNFPA (United Nations Population Fund). 1995. "Viet Nam Reproductive Health/Family Plan- ning Sector Review and Analysis." August. UNICEF (United Nations Children's Fund). 1985. The State of the World's Children. New York: Oxford University Press. 1997. The State of the World's Children 1997. New York: Oxford University Press. - 1999. "Vietnam CCA." Hanoi, May. Vietnam Economic News. 1999. "Better Health for the Nation." Vol. IX, No. 16, April 19-25. Vietnam Health Insurance Authority (VHIA). 1998. Vietnam Health Insurance Statistics 1998. Hanoi. .1999. Vietnam Health Insurance Statistics 1999. Hanoi. Vietnam-Sweden Health Cooperation. 1999. "Project document in the Area of Drug Policy and Control." Hanoi, April 24. Vietnam-Sweden Mountain Rural Development Programme in collaboration with SIDA and the World Bank. 1999. Lao Cai Province Participatory Poverty Assessment, Part One: Main Report. Hanoi. VLSS (Vietnam Living Standards Survey). 1993 and 1998. Conducted by the General Statistical Office. Hanoi. Vogel, Ulrich. 1987. "The Whole of Vietnam can be considered as One Well-Designed Project: Some Reflections on Primary Health Care Experiences in Vietnam, 1945-85." Disserta- tion submitted in partial fulfillment of the requirements of the M.Sc. (Economics) in Tropical Epidemiology and Health Planning, University of Wales. Vu, Nguyen Hong, and Claudio Schuftan. 1995. "Summary of Recent Primary Health Care Indi- cators in Viet Nam." Primary Health Care Unit, Hanoi, June. Weaver, Marcia. 1999. "Three Studies on Hospital Finance from the Survey of Hospitals in Vietnam." Study commissioned for the Vietnam health sector review. Seattle. Mimeo. Wennberg, J., and A. Gittelsohn. 1973. "Health care Delivery in Maine: Patterns of Use of Common Surgical Procedures." Journal of the Maine Medical Association 66(5):123-49. White, Les. 1998. "Health Insurance in Vietnam: A Report on Social Health Insurance Devel- opment in Vietnam, 1993-mid- 1998." Deakin University, Victoria, Australia, July. 201 WHO (World Health Organization). 1995. "Women's Health Profile: Viet Nam." Women's Health Series Volume 8, Regional Office for the Western Pacific, Manila. 1997. "Tobacco or Health? A Global Status Report." Geneva. 1999. The World Health Report 1999. Geneva. Witter, Sophie. 1996. "Doi Moi and Health: The Effect of Economic Reforms on the Health Sys- tem in Vietnam." International Journal ofHealth Planning and Management 11:159-72. World Bank. 1992. Viet Nam: Population, Health and Nutrition Sector Review. Report No. 10289-VN, East Asia and Pacific Regional Office, Washington, D.C., September. . 1996. Viet Nam: Fiscal Decentralization and The Delivery of Rural Services. Country Operations Division, East Asia and Pacific Region, Report No. 15745-VN, Washington, D.C. . 1998. Vietnam: Rising to the Challenge. Report No. 18632-VN, Poverty Reduction and Economic Management Sector Unit, East Asia and Pacific Region, Washington, D.C., November. . I999a. Health, Nutrition and Population Indicators: A Statistical Handbook. Human Development Network, Health, Nutrition and Population Series, Washington, D.C. .1 999b. Vietnam: Attacking Poverty, Poverty Working Group, Hanoi. 202