Is Thailand's Health System Recovering from Economic Crisis? Developing Indicators to Monitor Equity

Is Thailand's Health System Recovering from Economic Crisis? Developing Indicators to Monitor Equity Executive summary Siriwan GRISURAPONG Thailand is a country facing with high inequity in income distribution. Since wealth and health are strongly related, this difference in income distribution has been considered as a leading factor in health gaps between the rich and the poor. In order to monitor progress towards inequity reduction, some indicators are needed as tools for this monitoring process. Most of reports and statistics on health status use average rates rather than rate ratio or rate difference by socioeconomic subgroups of population, therefore, inequity in health and health care have been concealed. Moreover, equity is something abstract for stakeholders implementing program. Prior to monitor equity by using some tools, training or sensitization of these related stakeholders on equity concept and measurement must be considered. Establishment of equity indicators from secondary source data showed that: There are difference in health status between male and female and people living in urban and rural areas. When health status have been considered across population in socioeconomic subgroups, people with lower education have worse health compared to people with higher education. When morbidity rates have been used to classify provinces with lower level of health of population, some certain provinces in the Northeast (the poorest region) such as Kalasin appeared in the high rank. More resource should be allocated to these provinces if we want to reduce these disparities. Consideration of health budget allocation by provinces, it was found that provinces in the Northeast region also received budget in the lowest rank whilst the Central region which is the most prosperous area received in the higher rank. When household resource allocation has been considered, the poor spend higher proportion of their income to health care in 1986 but the trend was changed. In 1999, the rich spend higher proportion of income to health care which mean that health intervention programs are quite effective and equity-oriented. However, the poor had to spend more on food compared to the rich. When consumption of alcohol and tobacco has been considered, the poor spend less on alcohol but more on tobacco. If utilization/accessibility to health care will be focussed, people with no formal education or lower education tend to do nothing, use alternative care, self treatment or use primary level of health facilities when they are ill whereas those with higher education tend to go to tertiary health facilities and private hospitals. After the economic crisis, choice of treatment of people with higher education in do nothing and alternative care seems to be increased. When accessibility to health care has been considered in terms of insurance coverage, people with lower education had higher coverage in insurance schemes compared to those with higher education although the proportion of people covered by health insurance schemes seems to be reduced after economic crisis in both groups When determinants of health have been considered to identify inequity of these variables, it was found that provinces in the Northeast, generally appear in the worse rank of these indicators. These indicators such as per cent of population attending secondary school, therefore, can use to target provinces for poverty reduction and it will 1

result in reduction of inequity gap. When safe water for use has been considered, the poor also are in worse accessibility to safe water compared to the rich. Is Thailand s Health System Recovering from Economic Crisis? Developing Indicators to Monitor Equity Author: Siriwan GRISURAPONG Affiliation: Faculty of Social Sciences andd Humanities Mahidol University, Salaya, Puthamonthon 4 Nakhonpathom 73170 Thailand Tel: (662) 0-2441-0220-3 ext. 1261, 1201, 1102 Fax: (662) 0-2441-9738 e-mail: shsgs@mahidol.ac.th Abstract: Although health status of Thai population has steadily improved but there are health gaps between the rich and the poor and people living in urban and rural areas. Simple and effective tools are needed for policy makers and local health care staff to monitor progress towards equity. This study aims to assess inequity in resource allocation, utilization/accessibility and health status at national and provincial level and identify indicators to monitor progress towards equity. Data from secondary sources have been analysed to establish indicators for monitoring of inequity in health and health care. Results demonstrated that there are inequity in health status, utilization/accessibility and resource allocation whatever indicators were used. Simple indicators can also be used to monitor progress towards inequity but indicators established from population-based data can be better classified inequities in health of population by socioeconomic subgroups. At the provincial level, data are only available for classification of provinces which are in disparities compared to other provinces but not adequate for monitoring inequity within province. Before facilitating policy makers and local health care staff to use indicators for monitoring of inequity in health, sensitization or training of equity concept and measurement is needed. Keywords: Equity, monitoring, indicators Background Health status of Thai population has gradually been improved as can be seen from life expectancy which increased from 66.4 to 66.9 for male and from 70.8 to 71.7 for female during the period 1990-1995. Infant mortality rate has also declined from 125 per 1,000 live births in 1960 to 30.5 in 1995. Although health status has been improved but there are some disparities. In 1994, IMR was 27 per 1,000 in urban and 41 in rural areas. Higher IMR in mother with lower level of education has also been reported. Distribution of health care personnel in Bangkok was 5 times higher than other regions of Thailand(Pannarunothai, 2000). Inequalities in expenditure and health action has also been reported by Makinen et al (2000). 2

Although the Ministry of Public Health is a major provider of health services in Thailand but private providers also plays a major role in provision of health services. Health services system has been divided into: primary, secondary and tertiary level. People in the rural area will access first to the primary level and goes up to the higher level by the referral system. However, quality of care has been perceived as lower in the lower level of health service system. Resource allocation also more concentrated in health facilities in the cities due to use of high-technology equipment and highly specialized care of these facilities. Thailand face a situation of inequity in income distribution for several decades. The Gini coefficient has increased from 0.41 in 1962 to 0.45 in 1975, 0.49 in 1988 and 0.52 in 1999. This difference in income distribution has been noted as a leading factor in health gaps between the rich and the poor. The economic crisis in 1997 affected the poor more than the rich as has been reported by the Ministry of Public Health of the increase number of people in the low income group and number of their service utilization. Although health care reform has been initiated but equity has been left off in favor of efficiency. People in the low income group, low social class and rural areas receive less resource allocation and have lower access to health care than people in the higher groups and urban areas resulting in worse health status. Tools are also not available to monitor inequity in health and health care. Objectives of this study were to identify sets of equity indicators for policy makers and local health care staff to monitor inequities in health in terms of resource allocation, utilization/accessibility to health care and health status at national and provincial level. Methods: In order to establish indicators on inequity in health, analyses stratified by socioeconomic distribution were conducted using data from 2 surveys based on nationally representative samples of Thai population and administrative data or routinely collected data and vital statistics from several government agencies. In order to look at the views of health care staff, local government administrative staff and community leaders towards equity in health so that determination of equity indicators can go along with their preference, panel and group discussions of these related stakcholders from Nakhonpathom province, purposively selected, were conducted. Measurement: The measures of inequity have been classified into inequity in health status, resource allocation and accessibility/utilization of health care. 1. Health status Indicators established from administrative data can be classified only by gender or place of residence (Rural or urban). These indicators are: - Life expectancy at birth - Crude death rate - IMR - Malnutrition rate of children under 5 years Indicators demonstrated difference in morbidity classified by socioeconomic characteristics have to be drawn from population-based surveys. These indicators are: - Report of illness - Rate of admission to hospital 3

At the provincial level, difference in morbidity rate of population by province have been classifed by: - Rate of admission with heart diseases - Rate of illness with diarrhea 2. Resource allocation Indicator in this categary analysed from routinely collected data is: - Average health budget per capita by province Indicators which demonstrated resource allocation within household for health and determinants of health have been classified as: - Health care expenditure - Food expenditure - Alcohol & Tobacco expenditure 3. Accessibility/Utilization of health care Analyses of population-based data have been performed to identify indicators in this category. These indicators are: - Utilization rate of health services - Coverage of health insurance Studies to explore inequity in health and health care always include indicators on determinants of health since inequity in these determinanats led to inequity in health. Indicators which have been included are: - Income per capita - Proportion of population with secondary school education - Proportion of population in urban area - Proportion of population participating in economic works - Proportion of population using contraceptive method - Accessibility to clean water Data sources: Data from secondary sources have been employed in this study. These were data from the Socioeconomic Survey aggregated every 2 years and Health and Welfare Survey aggregated every 5 years by the National Statistical Office. These 2 surveys are nationally representative and collects data from heads of households for more than 10,000 households each round. Allthough these data can be a good representation of population in different socioeconomic groups at national and regional level but they just have been recently adjusted to represent population at provincial level. Besides these 2 surveys, population census aggregated every 10 years was used for analyses of difference in socioeconomic level of population by province. Data which have been routinely collected or used for administrative function reported or aggregated by National Economic and Social Development Board, Ministry of Public Health, Ministry of Finance and Ministry of Interior have also been included to represent the income, resource allocation and mortality & morbidity rates and ratios by province. Analysis of data Data were analysed using the Microsoft excel program. Descriptive analysis was used to describe indicators demonstrating inequity in health status, resource allocation and accessibility/utilization of health care. Results 4

Table 1 shows level of health status of Thai population using conventional indicators: life expectancy at birth, crude death rate, infant mortality rate and malnutrition rate. These indicators, although, demonstrate change of level of health status by time but they can represent difference only by gender and place of residence which make it hard to identify target-population. When population-based data have been use for analyses of morbidity indicators: report of illness and rate of admission to hospital, classified by educational level (Table 2); these indicators show difference in morbidity rate of people with lower and higher education. Those with no formal education reported more illness and admitted to hospitals than those with primary, secondary, vocational school and university. Consideration of health status of population at the provincial level, table 3 presents the 5 highest rank of rate of illness with diarrhea and rate of admission with heart disease. Illness with diarrhea has been considered to be related to low level of socioeconomic development whilst illness with heart diseases mean higher level of these development. Data at the provincial level are not adequate to show difference in health status of population by socioeconomic subgroup but identification of provinces which are in disparity with others in health and health care can be made. From 1989-1999, although the high rate of illness with diarrhea tend to scatter through out the country but provinces in the Northeast region (Burirum and Kalasin) which has been considered as the poorest region seems to appear more in the rank. When rate of admission with heart diseases has been considered, provinces in the central region (the most prosperous region): Singburi, Utradit and Angthong have been in the rank in 1993-1999. When resource allocation in terms of health budget per capita by province have been ranked to explore whether poorer provinces received more or not. Table 4 demonstrate that provinces in the Northeast seem not to appear in the highest rank except Mukdaharn province which received health budget as the second rank in 1993. Consideration of provinces in the lowest rank, provinces in the Northeast were in the rank through out the whole period, for example, Surin and Kalasin which have been known as poor provinces have been in the lower rank. Difference in health care expenditure and expenditure on other determinants of health among population in each socioeconomic subgroup can be good indicators presenting inequity in health and health care. Table 5 shows proportion of expenditure spending for health care, food, alcohol and tobacco by income quintiles. The poor spend more on health care, food and tobacco compared to the rich but less on alcohol expenditure. The gap of health care expenditure between the rich and the poor seems to be reduced and the trend shows that the rich spend more to health care in proportion to their income. This imply that Thai health policy and development plan led to more equity in health care in terms of private expenditure if proportion of health expenditure will be used as indicators. Although the gap in health care expenditure seems to be reduced but the inequity in determinants of health represented by higher proportion on food expenditure of the poor means that interventions to improve standard of living of the poor must be emphasized. Most of the total income of the poor are spent on food although this proportion were reduced but after the economic crisis in 1997, this proportion seems to rising again. The lower proportion of spending on alcohol of the poor may not mean that the poor consume less but it may mean that the rich consume more expensive ones. What should be a serious warning is the higher proportion of spending on tobacco of the poor. This may lead to more inequity in health between the rich and the poor. Inequities in utilization/accessibility to health care are important indicators for interventions of health care program. Table 6 shows choices of health care utilization of 5

Thai population classified by educational level. Those with no formal or less education tend to do nothing, self treatment or use alternative care or lower level of health facilities compared to the higher education. In 1999, after economic crisis although the proportion of the higher education who do nothing or self treatment seem to increase but data still demonstrate that those with higher education use more services at the tertiary care and private hospitals whereas those with lower education use more primary care. Consideration of population classified by educational level who are not under coverage of any health insurance schemes (Table 7), it was found that those with lower education are less likely to be covered by any schemes compared to the higher. Although increase in their coverage was found in both groups but difference between these groups which pointed to inequities in accessibility still appeared. When some determinants of health were used to be indicators to classify poor provinces such as income per capita (Table 8), it was found that from 1989-1999, all province in the lowest rank are in the Northeast region. Since poverty is the major determinant of health, population in these provinces may face lower level of health status compared to other provinces which higher income per capita. When other variables have been considered to use as proxy for indicators on determinants of health such as per cent of population with secondary school education, living in urban areas, participating in economic works and using contraceptive method (Table 9), it was found that most of the provinces fell in the lowest rank of attaining secondary school education and living in urban area are provinces in the Northeast. It means that if development of and educational level of population in these provinces will be improved, people in these provinces may have higher standard of living leading to better health and reduce disparity in health status. If more participating in economic works mean higher income gain, most of provinces fell in the lowest rank are provinces in the central region. Only one province in the Northeast (Surin) is in this category. It was showed that people in the Northeast although participate more in economic works but they still have lower income compared to people from provinces in other regions. If per cent of population using contraceptive method will be considered as indicators for accessibility to health care by province, it may not be sensitive enough for Thailand since all province fell in this category are provinces in the South. The reason of lower rate of using contraceptive method may not demonstrate lower accessibility but because the majority of them are Moslem. Consideration of determinants of health at the national level using household with their own pipe-line water as indicators (Table 10), data demonstrate that those with lower income having their own pipe-line water less than those with higher income. Although the number of household with their own pipe-line water in all group seems to be increased, but the inequity in different income groups still appeared. When views of local health care staff towards equity in health have been explored through panel and group discussion, it was found that equity for them pinpoint to equality in service provision to population in every subgroup and receiving fair resource allocation according to the amount of services provided. Therefore, if we want to pursue for more equity by working with local health staff; sensitizing them or training them with the equity concept, measurement and participation may be necessary. Discussion At the national level, conventional indicators on health status such as IMR, Life expectancy at birth can demonstrate inequities in health by only gender and place of residence. Although these indicators may be useful for monitoring of inequity in health but they can not be used to target population that who should be focused to reduce level of 6

inequity. When these simple indicators have been compared to complex indicators such as concentration index, index of dissimilarity, they can be used as simple tools for policy makers or local health care staff to monitor their own program implementation and health development policies. The point that should be focussed may not depend on what indicators but on distribution of data which can classify indicators by socioeconomic subgroups. Pannarunothhai & Rehnberg (1998) who test several complex indicators to measure level of inequity in health and health care found that use and interpretation of these indicators may be difficult for lay person or policy makers since concept of these indicators are quite abstract. When simple indicators such as report of illness and rate of admission have been classified by income or educational level to represent distribution of health across population in different socioeconomic subgroups, they can present inequities in morbidity rates in these subgroups very clear. For the developing countries like Thailand where income data may be not easy to collect, distribution of population by education, occupation and other socioeconomic characteristics can be good used to establish indicators to assess inequity level. We can advocate these simple indicators to policy makers to use for monitoring of inequity in health to solve the problem of complex and hard to understand indicators. However these indicators may be able to use to show that inequity exist or to assess progress towards equity but they may not be able to use for targeted population. Some simple indicators demonstrating disparity in health and determinants of health at the provincial level may be better to use for targeting provinces which are behind other provinces. Since health and poverty are things that strongly related (Leon, Walt and Gilson, 2001), targeting provinces which are in relative poverty can also reduce inequities in health. Reduction of poverty now be a focus of several international organizations (World development report 2000-2001). Inequity have also been related to this attempt. If using of simple indicators to monitor inequity in different population subgroups after economic crisis will be considered, it was found that these indicators can also be used for such monitoring. Findings from this study demonstrated that these indicators present some change in rate or per cent such as increasing of percent of people in lower education choosing self treatment. These findings show the same results as Wongkongkathep (2000) s study who found that self treatment and hospital visit were increase among the poor. Aungkasuvapala et al (2000) reported that after the economic crisis, the proportion of self-medication group has increased among the poor and changing of utilization of private health facilities to public. Recommendation Simple indicators established from administrative or routine data can be good used for monitoring of inequity in health by policy makers and local health care staff the same as complex indicators but data used in this establishment should be population-based surveys if presentation of inequity classified by population in socioeconomic subgroups is needed. Indicators for monitoring equity at national level may be difficult to use for targeting population who are in disparity in health and health care. Targeting by determination of poor people or provinces may be easier for policy makers to reduce poverty since poverty and health are strongly related. The World Bank has already determined 5 provinces in the Northeast to be the targets for poverty reduction. However, an attempt should be made to train or sensitize policy makers and local health care staff on equity concept and measurement and how to pursue for more equity before encouraging them to use simple indicators for monitoring of equity. 7

Acknowledgements: The project was supported by the Alliance for Health Policy and Systems Research, an initiative of the Global Forum for Health Research in Collaboration with the World Health Organization. References: 1. Gwatkin, D.R., & Guillot, M., (2000). The Burden of Disease among the Global Poor: Current Situation, Future Trends, and Implications for strategy. The Global Forum for Health Research and the Health, Nutrition, and Population Department; The World Bank, Washington DC. 2. Leon D.A., Walt, G., and Gilson L. International perspectives on health inequalities and policy BMJ, 322, 2001: 591-4 3. Makinen, M., et al (2000) Inequalities in health care use and expenditures: empirical data from eight developing countries and countries in transition, Bulletin of the World Health Organization, 78(1):55-65 4. Ministry of Public Health, Thailand Health Profile 1997-1998, 1998 5. Ministry of Public Health, Report of In-patient Utilization in Public Health Facilities classified by Diseases, 1990 6. Ministry of Public Health, Report of In-patient Utilization in Public Health Facilities classified by Diseases, 1993 7. Ministry of Public Health, Report of In-patient Utilization in Public Health Facilities classified by Diseases, 1996 8. Narongsakdi Aungkasuvapala, Panbuadee Ekachampaka & Suthisarn Watanamano, 2000 Change in the Health System after Economic Crisis Journal of Health Plan and Policy (3) 2, April-June 2000 9. Pannarunothai, S. & Mills, A., (1997) "The poor pay more: health related inequity in Thailand" Social Science and Medicine, 44 (12) 1781-90. 10. Pannarunothai, S. and Rehnberg C., Equity in the Delivery of Health Care in Thailand, Research Report, June 1998 11. Pannarunothai, S., (2000) Equity in Health, Naresuen University. 12. Sunee Wongkongkathep, 2000 Illness and Health Seeking Behavior of the Poor before and after Economic crisis in 8 Provinces Journal of Health Plan and Policy (3) 2, April-June 2000 8

Appendix: Table 1 Difference of health status of Thai population between gender and place of residence at the national level Indicators 1995-1996 1995-1996 Male Female Urban Rural - Life expectancey at birth 69.9 74.9 - - - Crude death rate (/100,000 population) - - 414 650 - Infant mortality rate (/1,000 live birth) - - 15.2 28.2 - Malnutrition rate of children under 5 years - - 5.3 9.0 Table 2 Report of illness and rate of admission to hospital of Thai population classified by educational level Year 1996 Level of education Report of illness Rate of admission - No formal education 9.98 10.10 - Elementary school 5.80 6.49 - Secondary school 3.08 4.84 - Vocational school 2.04 5.16 - University 1.57 4.02 9

Table 3 The 5 highest rank of rate of illness with diarrhea (/1,000 population) and rate of admission with heart disease (/100,000 population) by province year 1989, 1993, 1999 Year Rank 1989 1993 1999 Rate of illness with diarrhea 1 Mae Hong Sorn 37.4 Samutprakarn 44.9 Burirum 52.0 2 Mukdaharn 32.3 Maehongsorn 44.8 Satul 46.8 3 Phuket 28.7 Pathumthani 41.0 Pathumthani 45.5 4 Nonthaburi 28.5 Kalasin 39.2 Samutprakarn 42.3 5 Tak 27.5 Yala 38.6 Kalasin 40.1 Rate of admission with heart disease 1 Nakhonnayok 316.0 Nan 371.5 Singburi 644.6 2 Nan 292.8 Singburi 329.7 Utradit 382.17 3 Samutsongkram 228.3 Phang Nga 244.4 Nan 367.1 4 Yala 208.2 Utradit 236.9 Angthong 336.0 5 Ratchaburi 178.2 Phuket 211.7 Phang Nga 331.99 10

Table 4 The 5 highest and lowest rank of health budget per capita by province year 1989, 1993, 1999 (Baht/capita) Year Rank 1989 1993 1999 Highest rank 1 Mae Hong Sorn 403.1 Uthaithani 2,147.0 Sukhothai 3,259.3 2 Phang Nga 325.4 Mukdaharn 1,836.0 Nonthaburi 1,338.9 3 Ranong 318.3 Yala 1,504.2 Krabi 1,123.1 4 Singburi 312.4 Phijit 1,350.5 Mae Hong Sorn 915.9 5 Samutsongkram 310.5 Pattani 956.9 Ratchaburi 824.4 Lowest rank 1 Chumporn 74.1 Khonkaen 199.8 Samutsakorn 201.6 2 Ayuthaya 53.5 Chiengmai 173.6 Surin 192.8 3 Udornthani 44.5 Chaiyaphum 167.6 Prajuabkirikhun 183.9 4 Roi-et 37.6 Kanjanaburi 162.5 Kalasin 177.6 5 Kalasin 13.4 Surin 122.9 Phetchabun 157.7 11

Table 5 Health care expenditure, food expenditure, alcohol & tobacco expenditure to total expenditure distributed by income quintiles year 1986-1999 Years 1986 1992 1996 1998 1999 Income quintiles Health care Food Alcohol Tobacco Health care Food Alcohol Tobacco Health care Food Alcohol Tobacco Health care Food Alcohol Tobacco Health care Food Alcohol Tobacco First 8.36 102.92 1.80 2.29 7.97 92.29 4.30 4.68 6.61 88.56 3.79 4.39 5.59 99.12 2.22 2.57 5.02 93.82 2.86 3.11 Second 8.05 95.60 2.32 2.36 6.67 88.35 5.18 4.93 6.92 81.79 4.86 4.62 6.05 90.64 2.72 2.87 6.01 84.25 3.56 3.61 Third 7.33 89.30 2.30 2.54 6.25 83.49 6.43 5.01 6.34 77.24 5.77 4.83 5.85 84.19 3.49 3.28 5.74 80.96 4.50 3.47 Fourth 6.08 86.54 2.74 2.47 6.78 74.11 7.15 4.50 6.74 70.73 6.06 4.06 6.03 75.64 3.57 3.16 5.62 72.59 4.47 3.48 Fifth 6.84 70.41 3.21 1.75 7.73 54.08 5.85 2.01 10.58 51.54 5.25 2.38 6.59 59.93 3.35 1.90 7.29 54.78 4.43 2.10 12

Table 6 Choices of health care utilization of Thai population by educational level Year 1996 1999 Educational level Do nothing Alternative care Self treatment Private clinic Public primary care Public secondary care Public tertiary care Private hospital Do nothing Alternative care Self treatment Private clinic Public primary care Public secondary care Public tertiary care Private hospital - No formal education 5.59 5.37 26.56 14.50 17.44 19.87 19.43 4.05 6.59 3.68 16.34 14.63 14.63 19.59 21.13 2.22 - Elementary school 3.60 3.07 32.44 17.87 16.34 14.11 17.27 3.01 3.91 1.94 22.07 17.43 12.07 17.19 20.16 3.35 - Secondary school 3.71 1.02 36.98 22.15 10.01 9.27 15.85 5.28 4.00 0.95 26.96 19.92 8.33 10.84 18.22 6.30 - Vocational school 3.85 1.20 34.13 24.04 5.05 7.93 20.19 6.49 3.59 0.75 23.92 23.32 5.08 6.58 26.61 8.97 - University 2.86-34.86 27.43 1.14 0.57 23.43 11.43 7.73 1.10 24.13 20.72 2.21 4.70 22.65 14.92 13

Table 7 Thai population who are not under coverage of any health care insurance schemes classified by educational level Years 1996 1999 Educational level - No formal education 48.13 36.42 - Elementary school 56.08 40.05 - Secondary school 53.78 40.86 - Vocational school 37.89 30.42 - University 32.13 27.31 Table 8 The 5 lowest rank of income per capita by province year 1989, 1993, 1999 (baht) Year Rank 1989 1993 1999 1 Yasothon 10,741 Burirum 15,278 Surin 20,363 2 Burirum 10,687 Kalasin 15,219 Yasothon 20,304 3 Nakonpanom 10,562 Surin 14,542 Srisaket 20,283 4 Surin 10,001 Yasothon 14,471 Amnatcharoen 19,413 5 Srisaket 9,942 Srisaket 13,738 Nongbualumpoo 18,735 14

Table 9 The 5 lowest rank of per cent of per cent of population with secondary school education, living in urban area, participating in economic works and using contraceptive method by province in 1990 Province Rank Per cent Secondary school education 1 Nakhonratsima 4.1 2 Srisaket 5.7 3 Surin 5.7 4 Burirum 6.0 5 Chiengrai 6.3 Living in urban area 1 Roi-et 4.5 2 Surin 7.3 3 Srisaket 8.4 4 Mahasarakam 9.8 5 Sakonnakhon 10.4 Participating in economic works 1 Surin 64.1 2 Nonthaburi 65.3 3 Phuket 65.7 4 Samutprakarn 67.7 5 Narathivat 67.8 Using centraceptive method 1 Pattani 24.2 2 Yala 27.3 3 Narathivat 27.5 4 Songkhla 39.6 15

5 Satul 41.1 Table 10 Household with their own pipe-line water classified by income quintile year 1986-1999 Years Income quintiles 1986 1992 1996 1998 1999 1. First (Poorest) 7.0 13.0 33.60 42.0 47.11 2. Second 10.0 20.0 39.08 49.0 54.04 3. Third 19.0 33.0 47.78 58.0 62.58 4. Fourth 34.0 51.0 58.57 65.0 70.86 5. Fifth (Richest) 57.0 76.0 72.26 78.0 81.71 16