58579 Out-of-pocket spending and health service utilization in Lao P.D.R. Evidence from the Lao Expenditure and Consumption Surveys The World Bank, November 2010 Out-of-pocket spending and health service utilization in Lao P.D.R.: Evidence from the Lao Expenditure and Consumption Surveys World Bank, November 2010 Summary This briefing note provides an overview of trends in health service utilization and out-of-pocket expenditures in Lao PDR. It is based on an analysis of household data from two rounds of the nationally representative Lao Expenditure and Consumption Survey (LECS), conducted in 2002/03 and 2007/08 respectively. The findings indicate that (unconditional) utilisation of outpatient care has not risen over time and, if anything, has fallen slightly. Inequalities in the use of outpatient care appear to have widened slightly. Individuals have also reduced utilisation of inpatient care over the same five year period, although again the change is relatively small. Out-of-pocket spending on health care has risen in real terms, explained largely by an increase in spending on user fees, transport and health-related items other than drugs. However, it remains the case that drugs account for the majority of health spending by households. In 2002/03 there were enormous disparities in health spending across wealth quintiles and inequality has widened subsequently. Expenditure specifically on inpatient care has fallen ­ the result of a fall in utilisation rather than any change in the cost of care. Findings suggest the incidence of catastrophic has fallen marginally over the five years, but it is difficult to interpret this finding and should not be regarded as evidence of financial protection policies working. Finally, there is little evidence that out-of-pocket spending on health has increased impoverishment. When spending on health is considered involuntary, the percentage of households in poverty does not rise by a substantial amount. The briefing note finds little support for the proposition that the various risk-protection mechanisms in Lao PDR have contributed to a substantial increase in health care utilisation or a reduction in out-of-pocket spending. It argues that for Lao PDR to make strides towards universal coverage, general tax financing of health must increase substantially. In this regard, new opportunities are emerging as the tax base becomes stronger but actors in the health sector must be prepared to put forward a convincing case for increased investment in health. This note was prepared as part of a World Bank program of analytic work on health financing in Lao PDR. The note was drafted by Timothy Powell- Jackson (LSHTM) and Magnus Lindelow (WB). Inputs and comments from Nina Fenton, Bart Jacobs, Jean Marc Tomé, Adam Wagstaff and Christoph Kurowski are gratefully acknowledged. Background to briefing note to be saved at childbirth. While estimates of the maternal mortality ratio in Lao PDR differ [4, 5], it is unacceptably In the past decade, Lao PDR has made substantial progress high, indicating significant health system challenges. in reducing poverty and child mortality. It is well placed to reach the child health Millennium Development Goal One of the key determinants of whether people can access [1]. However, not only do health outcomes remain among care when they need it, and whether they suffer economic the poorest in South-East Asia and far below international consequences from ill health, is how the health system is averages, progress in expanding coverage of priority health financed. Financing arrangements must guarantee a interventions known to reduce mortality, has been patchy. sustainable and equitable flow of funds, predictable Proxy indicators of the performance of the primary health external finance, risk pooling across the population and care system show few signs of improvement in recent effective purchasing of high-priority health services [6]. An years. Skilled birth attendance has increased from 19 over-reliance on out-of-pocket payments to finance the percent in 2000 to 20 percent in 2006 [2, 3]. Over the health system can contribute to low utilisation of services same period, the proportion of children seeking care for and risks exposing households to financial catastrophe. In suspected pneumonia has fallen from 36 percent to 32 Lao PDR, as in other low-income countries, a key percent [2, 3]. challenge is to provide adequate financial risk protection to those in need and thus ensure equitable access to essential Maternal mortality is perhaps the best indicator of the health services. While developing its strategy to achieve state of the health system because many of the building universal coverage, important decisions will need to made blocks (eg. human resources, service delivery, health regarding how to raise funds, how to pool them and how financing etc.) must be in place if the lives of women are to use them to purchase health services. 1 Figure 1. Private health spending been the driving force behind The purpose of this note is to provide an overview of increases in total health expenditure health service utilization and out-of-pocket expenditures 120,000 Private expenditure on health in Lao PDR. It is primarily based on analysis of household General government expenditure on health data from two rounds of the Lao Expenditure and 100,000 Consumption Survey (LECS), conducted in 2002/03 and National health spending in kip 2007/08 respectively. The data are representative of the 80,000 entire country and thus useful for understanding of health (constant 1995) financing from the perspective of households. Before 60,000 discussing results from the LECS surveys, it is useful to provide some context by looking at the broad health 40,000 financing picture in Lao PDR over the past decade. 20,000 Health financing in Lao PDR 0 According to the most recent figures, per capita health 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 spending in Lao PDR was around $28 in 2007,1 placing the country slightly below the South-East Asia regional Note: All data are presented in real kip with a base year of 1995. Data average of $31 per capita [7]. Trend data show that the come from WHO's National Health Accounts database [7] and are volume of financing for health in Lao PDR has increased expressed in terms of financing agent. markedly over the past decade (Figure 1). In real terms, total health expenditure more than doubled between 1996 These data give rise to a number of initial concerns. First, and 2006, outpacing gross domestic product which in the increasing share of funds managed through private comparison grew by 69 percent. arrangements, particularly out-of-pocket contributions from households, is likely to lead to a situation in which However, over the same period (1996-2006), general households must bear a greater risk of financial catastrophe government health expenditure has remained fairly static if they are to access care. Second, the volatility in donor in real terms, and it is the growth in private health aid is likely to make strategic planning in the health sector expenditures that have been the driving force behind the more difficult since donor funding for health represents increase in the overall resource envelope (Figure 1). around 30 percent of total health spending. Greater out-of-pocket spending from households seems to explain most of the rise in private health spending.2 Figure 2. Government health spending relative to the size of the External resources from donors have remained stable in economy is the lowest in the region real terms over the past decade, although the year on year 3.0% estimates are highly erratic, illustrating what is commonly General government health expenditure 2.5% observed in other low-income countries ­ that aid can be highly volatile. 2.0% as a share of GDP 1.5% 1.0% 0.5% 1 0.0% The reliability of this estimate is questionable given Indonesia Philippines Vietnam Bangladesh Cambodia Laos Malaysia Thailand uncertainties in external financing for health and private health spending, particularly by households. However, this should not distract from the fact that health spending in Laos is undoubtedly low. 2 Source: WHO NHA database 2007 Out-of-pocket spending by households represents 62 percent of total health expenditure in 2007, a share that is higher than other countries in the region. The same ratio is 26% in Thailand, A look at the government finances shows that the share of 60% in Bangladesh, and 55% in Vietnam [7]. Recent studies in GDP devoted to government health spending is low Lao PDR suggest that out-of-pocket spending in official NHA relative to other countries in the region (Figure 2). It has figures may be under-estimated. 2 been falling gradually over the past decade in Lao PDR, purchasing medicines without a prescription also fell, from mainly as a result of a stagnant share of total government 9.6 percent to 6.4 percent. expenditure allocated to health. Figure 4. Use of health services over full sample (unconditional) Utilisation of health services: Outpatient care 12% LECS3 (2002-03) The two rounds of the LECS provide a good opportunity 9.6% % of individuals seeking care 10% LECS4 (2007-08) to explore how utilisation of health services has changed in the five year period between the two surveys, conducted 8% in 2002/03 and 2007/08 respectively. Individuals demand 6.4% health services when they perceive there to be a medical 6% need. So before examining utilisation of health care, it is worth noting that the proportion of individuals who 4% reported being ill in the 4 weeks preceding the survey fell 2.1% 1.8% from 14 percent in 2003/04 to 10 percent in 2007/08.3 2% Over the five-year period between the two household 0.4% 0.3% surveys, the likelihood of an individual seeking care from a 0% modern health provider when ill has risen from 13 percent Modern provider Traditional Purchased medicines practitioner to 18 percent (Figure 3). A closer look at utilisation of outpatient care by various Figure 3. Use of health services, conditional on reporting illness background characteristics sheds light on the types of 80% individuals who have changed their health seeking 70% LECS3 (2002-03) 68.7% behaviour (Figure 5). Utilisation of outpatient services 63.2% % of individuals seeking care LECS4 (2007-08) from modern providers has risen for those aged 10 years 60% and under, but fallen in all other age groups. The largest 50% decrease in utilisation has been amongst adults and the elderly. More detailed information on utilisation by 40% background characteristics is shown in Table A1 in the 30% appendix. 18.1% 20% 13.2% Figure 5. Outpatient care seeking at modern providers by age 10% 2.9% 3.5% 4.5% 0% 4.0% LECS3 (2002-03) % of individuals seeking care Modern provider Traditional Purchased medicines 3.5% LECS4 (2007-08) when ill practitioner when ill when ill 3.0% Note: Modern provider includes a visit to a health facility, or an individual private health provider (doctor, nurse, midwife). The purchase of medicines 2.5% can be interpreted as self-medication since the purchase was made without a 2.0% prescription. 1.5% 1.0% Owing to the unreliability of self-reported measures of illness, it is more common to show utilisation estimates 0.5% across the entire population, irrespective of whether the 0.0% individual reported being ill. Using unconditional 0-5 years 6-10 years 11-15 16-20 21-65 65+ years estimates of utilisation, it seems that the proportion of old old years old years old years old old individuals who sought care in the past four weeks has Utilisation has fallen in every consumption quintile except fallen from 2.1 percent to 1.8 percent over the period the richest, which has seen a small increase in the (Figure 4). Likewise, the proportion of individuals proportion of individuals using outpatient care between 2002/03 and 2007/08 (Figure 6). Inequalities in utilisation thus appear to have widened. More detailed 3 There are well-known biases in self-reported measures of information on utilisation by background characteristics is illness [8]. shown in Table A1 in the appendix. 3 Figure 7. Individual private and traditional providers are being Figure 6. Outpatient care seeking at modern providers by age used less often and socioeconomic status 0.70 Outpatient vistis per capita per year LECS3 (2002-03) 3.0% LECS4 (2007-08) LECS3 (2002-03) 0.60 LECS4 (2007-08) 2.5% 0.50 % of individuals seeking care 0.40 2.0% 0.30 1.5% 0.20 1.0% 0.10 0.5% 0.00 Health facility Individual Traditional Total 0.0% private provider providers Poorest Q2 Q3 Q4 Richest Note: Health facility includes both public and private health institutions. Note: Modern provider includes a visit to a health facility, or an individual Private providers refer to individual doctors, nurses and midwives who provide private health provider (doctor, nurse, midwife). health services. An increase in the utilisation of health services can be When individuals seek care at a health facility, the most regarded as encouraging as long as individuals seek care at common option is the provincial or district hospital, appropriate health providers. In this regard, it seems followed by private health centres (Figure 8). There has reasonable to assume that health facilities are of a higher been no great change in this pattern over the period quality than traditional practitioners ­ that is, individuals between the two surveys, except that individuals are even who seek care at modern health facilities are more likely to more likely to choose provincial or district hospitals when make a full recovery. The fall in outpatient utilisation is seeking care at a health facility in 2007/08. least marked in health facilities; there is a more substantial fall in the number of outpatient visits per capita per year at Figure 8. The main providers of outpatient care are increasingly private and traditional providers (Figure 7).4 provincial and district hospitals (2007/8) Oth. or There is considerable variation in the change in utilisation missing, Cent. hosp., of health facilities according to the characteristics of the 11.1% 9.6% Hosp. individual. Many groups have actually increased utilisation abroad, 3.5% of health facilities, such as individuals under 15 years of age and the poorest and richest one-fifth of the population (see Table A2 in the appendix). Priv. prov., 23.7% Prov. /dist. Pub. HC, hosp., 39.8% 6.0% At least to some extent, utilization of outpatient service is expected to be related to physical access to health services. In the 2007/8 survey, 70% of households lived within 10km of the nearest HC, while just under 50% lived 4 The number of outpatient visits per year is calculated as the within 10km of the nearest hospital (see Figure 9). There product of the number of outpatient visits in the previous month have not been significant changes in access to health and the number of months in a year. facilities since the 2002/3 survey. 4 proportion of the population who were hospitalised at Figure 9. Distance to the nearest health center and hospital least once in the year preceding the survey fell from 2.3 percent in 2002/03 to 1.9 percent in 2007/08 (Table A3 30% Distance to nearest hosp. in the appendix). As expected, the elderly and richest Distance to nearest HC people in the population are more likely to access inpatient 25% care (Figure 11). The number of inpatient visits per person 20% per year also indicate a fall in the utilisation rate, from 32 inpatient visits per 1,000 people in 2002/03 to 26 visits in 15% 2007/08, with an average inpatient stay of 6.6 days. 10% Figure 11. The better-off use hospital services more 3.5% 5% HHs w. hospitalization last 12 months LECS3 (2002-03) 3.0% LECS4 (2007-08) 0% 2.5% 0-1 km 1-2.5 km 2.5-5 km 5-10 km 10-25 km > 25km 2.0% In fact, there is a clear relationship between utilization of 1.5% health services and distance to the nearest provider (Figure 10). However, utilization of public facilities for outpatient 1.0% services is low (approx. 0.4 visits per capital per year) 0.5% even for households with very good access, and the more significant reduction in utilization is only seen in 0.0% households that live more than 10km from the closest Poorest Q2 Q3 Q4 Richest health center. Similarly, the relationship between outpatient visits and distance to the nearest hospital is also quite weak (not shown). The majority of patients are hospitalised at the provincial / district hospital, with 75 percent of hospital service being Figure 10. Access to health centers and outpatient visits to provided at this level in 2007/08 (Figure 12). Just over a different types of providers (2007/8) tenth of inpatient cases are dealt with in central level hospitals. These patterns have not changed substantially in 0.50 the 5 years since the 2002/3 survey. Govt. health facility 0.45 OP care sought (visits p.c. / year) Private provider 0.40 Traditional healer Figure 12. The main providers of inpatient care 0.35 Hosp. Other, 5% 0.30 abroad, 4% Cent. hosp., 0.25 Priv. prov., 11% 4% 0.20 0.15 Pub. HC, 1% 0.10 0.05 0.00 0-1 km 1-2.5 km 2.5-5 km 5-10 km 10-25 km > 25km Distance to nearest health center Prov./dist. hosp., 75% Utilization of health services: Inpatient care Inpatient care refers to cases when an individual is hospitalised. It thus reflects more serious health Out-of-pocket spending on health care complaints than those treated as outpatient cases. The likelihood of a person being hospitalised in the past year Out-of-pocket payments not only deter households from provides one measure of utilisation of inpatient care. The seeking health care, they also have the potential to cause 5 financial distress and impoverish families. The unpredictability of many illnesses means that households survey and limits in the ability of respondents to give accurate answers to survey questions. For example, individuals may be unable to remember are often poorly prepared to meet the financial cost correctly the number of times they visited a health centre or the amount associated with obtaining health care. Health care that was paid. The use of proxy respondents who provide information on payments that exceed a certain proportion of a family's behalf of other household members, typically children, can also be income are said to be `catastrophic,' indicating that the problematic because they are not able to recall correctly events that they household faces a significant expenditure burden and is at themselves did not actually experience. For this reason, mothers are often risk of having to reduce consumption of other goods and more reliable than the husband in recalling health events relating to the services as a direct result of the health expenditure. One child. Finally, errors may occur because inaccurate information is given of the main goals of a country's health system is to provide deliberately, due to embarrassment or a wish to get the survey finished financial protection against such a risk as a means of quickly. ensuring equitable access to health care [9]. Research on health accounting suggests that the following biases are likely to exist: Although a number of methodological issues arise in Inpatient events can be underreported by as much as 30-50% measuring out-of-pocket health expenditures by when using a recall period of 12 months households (see Box 1), the LECS provides valuable Outpatient events can be underreported by around 20% when the perspectives on trends and patterns in spending. Over the recall period extends beyond three days 5 year period between the two LECS, out-of-pocket Less salient events are more likely to be forgotten spending has risen in real terms, from 63,408 kip (US$ 6.6) per capita per year in 2002/03 to 81,947 kip (US$ Proxy respondents may underreport events by as much as 20%. 8.5) in 2007/08 (Figure 13).5 The increase is mainly the result of a rise in expenditure on user fees, transport and Figure 13. Total out-of-pocket health spending has risen health-related items other than medicines. Information on 90,000 out-of-pocket spending on health care from the Out-of-pocket health spending per consumption diaries of households suggests that the vast 80,000 majority of health spending is on medicines. Expenditure 70,000 24,914 on this item rose only marginally in real terms over the capita per year 60,000 9,933 five year period. 50,000 40,000 Box 1. Issues in the measurement of out-of-pocket spending 30,000 53,475 57,033 20,000 Problems can arise when using representative household surveys to 10,000 estimate private out-of-pocket health care spending [11]. These mostly fall under two headings: 1) sampling error; and 2) biases from non- 0 sampling errors. Sampling error is a result of variation between LECS3 (2002-03) LECS4 (2007-08) individuals and is thus a problem if the sample size is not sufficiently Spending on user fees, transport, etc Spending on medicines large. Sampling error is likely to be larger for expenditures that are less frequent, more variable between individuals and when the recall period is shorter. Note: Data are presented in real terms (constant 2007/08 kip). The adjustment accounts for spatial and temporal variation in the cost of living, Non-sampling errors refer to problems in the implementation of the and is based on the same information that is used to construct the national poverty lines in each round of the LECS. The graph only includes health expenditures reported by households in their monthly diaries. See footnote 6. 5 Household expenditures on health care is calculated using Meanwhile, expenditure on inpatient care has fallen over information contained in the household diary, consistent with the same period ­ the result of a fall in utilisation rather Approach 1 in a technical review of national household spending in Lao PDR [10]. In the LECS, households fill out a diary than any change in the cost of care (Figure 14). In fact, if containing information on consumption over a period of four we focus on specific episodes, inpatient care has become weeks. This diary provides the information to calculate spending more expensive for households, rising 20 percent in real on health related items (codes 263 to 268). All tables and terms from 1,144,699 kip per inpatient stay in 2002/03 to figures, except Figure 9, are based on health spending calculated 1,390,223 kip in 2007/08. This increase in the price may in this way. Annual health expenditure is calculated as the product of out-of-pocket spending in the previous month and the number of months in a year. 6 explain why fewer individuals are accessing inpatient Figure 15. Richer individuals spend many times more on health care.6 care than the poorest and the gap is widening 300,000 Figure 14. Out-of-pocket health spending on inpatient care has Out-of-pocket spending on health care LECS3 (2002-03) fallen slightly over the five year study period 250,000 LECS4 (2007-08) 60,000 per capita per year 200,000 Out-of-pocket health spending per 50,000 14,756 150,000 40,000 6,111 capita per year 100,000 30,000 20,000 50,000 35,380 33,912 10,000 0 Poorest Q2 Q3 Q4 Richest 0 LECS3 (2002-03) LECS4 (2007-08) Note: Data are presented in real terms (constant 2007/08 kip). The adjustment accounts for spatial and temporal variation in the cost of living, and is based on Spending on transport to seek care the same information that is used to construct the national poverty lines in each Spending on hospitalisation round of the LECS. The graph only includes health expenditures reported by households in their monthly diaries. Note: Data are presented in real terms (constant 2007/08 kip). The adjustment accounts for spatial and temporal variation in the cost of living, and is based on The vast majority of out-of-pocket health spending is made the same information that is used to construct the national poverty lines in each by the richest two-fifths of individuals in the country round of the LECS. The graph only includes expenditure on inpatient care reported in the health module of the survey and there is likely to be overlap (Figure 16). In 2002/03, the poorest three-fifths of the between this spending and that reported in the monthly diaries of households. population accounted for only 24 percent of all out-of- pocket health spending in Lao PDR and this imbalance has There are enormous discrepancies in out-of-pocket only become more pronounced over time. Most recently, spending by socioeconomic status (Figure 15). Individuals the poorest three-fifths of the population comprised only in the richest quintile of the population spent 13 times 17 percent of all out-of-pocket spending. more on outpatient care and self-medication than those in the poorest quintile in 2002/03 and, if anything, Figure 16. The majority of out-of-pocket health spending in inequality in health care spending has widened further Lao PDR is accounted for by the richest quintile of the population since the LECS3 in 2002/03. The widening in this 100% inequality might reflect one or a combination of a number 90% of possibilities, including: i) a greater rate of growth in the 80% utilisation of health services by the rich; ii) a shift towards Out-of-pocket health spending per 54% 70% higher quality / more expensive health providers by the 67% 60% rich; or iii) an increase in financial protection for the capita per year 50% poorest groups. 40% 22% 30% 16% 20% 10% 24% 17% 0% LECS3 (2002-03) LECS4 (2007-08) Richest 20% (Q5) Richer 20% (Q4) Poorest 60% (Q1, Q2, Q3) 6 Note that Figure 13 does also include some spending on Household health expenditures can be expressed as a inpatient care; health spending reported in the monthly diaries fraction of total household consumption to provide an of households (Figure 13) and that reported separately in the indication of the financial burden imposed by having to pay health module of the survey (Figure 14) overlap and should not be added together. 7 for health care (see Box 2).7 This ratio has fallen slightly Out-of-pocket spending on health care can also impoverish households from 1.8 percent in 2002/03 to 1.7 percent in 2007/08.8 [15]. Health care consumes resources that could have otherwise been There have been some interesting changes over time across used to buy essential items such as food. Conventional measures of geographical areas of the country (Figure17). The financial poverty do not take account of health spending. By calculating the burden of health care spending has increased in urban areas proportion of households below the poverty before and after out-of-pocket payments for health care are deducted from household resources, it is and in Vientiane. Elsewhere the OOP health spending possible to approximate the impoverishing effect of such spending [12, ratio has fallen. 13]. Figure 17. The financial burden of health care payments by The ratio has increased among Mon-Khmer and Chine geographical setting Tibet households, but fallen in the other ethnic groups 3% (Figure 18). The only consumption quintile to see a rise in LECS3 (2002-03) the financial burden of health care payments is the richest OOP health care payments as share of LECS4 (2007-08) 2% group, perhaps reflecting the fact that these households are total household consumption using health services more. 2% Figure 18. The financial burden of health care payments by 1% ethnic group and consumption quintile 3.5% OOP health care payments as share of 1% LECS3 (2002-03) total household consumption 3.0% LECS4 (2007-08) 0% 2.5% Urban Rur. w Rur. no Vientiane North Central South rd. acc. rd. acc. 2.0% 1.5% Box 2. Measuring the economic consequences of health care 1.0% payments 0.5% Households without any financial protection face having to pay for large 0.0% medical expenses when they fall ill and seek health care. Out-of-pocket Mon-Khmer Hmong-Mien Lao Tai Poorer (q2) Richest (q5) Chine Tibet Poorest (q1) Middle (q3) Richer (q4) health care payments have the potential to disrupt material living standards because the money they spend on health care might otherwise have been spent on items such as food and clothing [12]. In recent years, a number of methods to approximate the disruption of out-of-pockets health care payments on living standards have become popular. The proportion of households who spend more than 10 percent of their annual budget on health care fell from 4.2 When health care payments are large relative to the household budget, percent in 2002/03 to 3.8 percent in 2007/08, suggesting the disruption to living standards can be considered catastrophic. that the risk of catastrophic health care payments has fallen Specifically, when out-of-pocket health spending exceeds a certain marginally in recent years (Figure 19). The incidence of fraction of total household consumption ­ a 10% threshold is the most catastrophic spending rises with household consumption. common ­ it can be defined as catastrophic [13, 14]. The incidence of The low incidence among the poorer households partly catastrophic health spending is then the proportion of households that reflects the low utilisation of health services by these exceed this threshold. households. Thus, while they are less exposed to catastrophic health care payments, their health is more 7 likely to suffer as a consequence of not seeking health care. This ratio is calculated on the basis of information contained within the diary of household. It does not draw on data relating The richest quintile is the only group to have seen an to inpatient care expenditures in the health module, because the increase in catastrophic spending during the period denominator of total household consumption is calculated on the between the two surveys. It is perhaps encouraging to note basis of diary only. that the proportion of families that borrowed money to 8 The ratio of OOP health payments to total household pay for health care fell from 1.4 percent to 0.7 percent consumption in Lao PDR compares favorably with other over the same period, implying that they are experiencing countries in the region (eg. Vietnam 5.9%, Bangladesh 5.1%, less financial distress. Thailand 1.7%). But the surveys in these other countries are Figure 19. The incidence of catastrophic spending is higher more comprehensive in capturing the health expenditures of a among richer households household, so findings may not be directly comparable. 8 8.0% LECS3 (2002-03) 7.0% LECS4 (2007-08) Figure 15. Health spending does not appear to impoverish Incidence of catastrophic spending households to any great degree 6.0% 40% (using a 10% threshold) 5.0% 33.5% 34.9% % of population in poverty 35% (national poverty line) 27.6% 28.4% 4.0% 30% 3.0% 25% 20% 2.0% 15% 1.0% 10% 0.0% 5% Poorest Q2 Q3 Q4 Richest Total 0% LECS3 (2002-03) LECS4 (2007-08) It is difficult to compare estimates of catastrophic Prepayment poverty headcount Postpayment poverty headcount expenditure and impoverishment due to health care payments across countries because of differences in how national surveys gather information on health spending by households.9 However, recognizing this limitation, the Conclusion estimates for Lao PDR are low relative to other low- This note has provided an overview of evidence on trends income countries in the region. For example, the and patterns in health service utilization and out-of-pocket incidence of catastrophic spending in Vietnam and expenditures from two rounds of the Lao Expenditure and Bangladesh is 15.1 percent and 15.6 percent respectively, Consumption Surveys (2002/03 and 2007/08). There has while the corresponding figures for middle-income been a fall in utilisation of both outpatient and inpatient countries in the region ranges from 2.0 percent for care (based on unconditional estimates). Outpatient Malaysia to 4.6 percent for the Philippines. The low utilisation of modern health providers has declined in incidence of catastrophic spending in Lao PDR may be due every wealth quintile, except the richest one-fifth. to how health expenditure data are collected, but is also Inequalities thus appear to have widened slightly. likely to be due to the very low levels of utilization. Utilisation of inpatient care has fallen for households in every wealth quintile. In Figure 20, the prepayment poverty headcount is the standard estimate of the proportion of households in Lao Meanwhile health care spending has risen in real terms PDR that are in poverty in each survey year. When out-of- over the five year period, explained largely by an increase pocket payments on health care are excluded from the in spending on user fees, transport and health-related household consumption total (ie. they are assumed to be items other than drugs. But it remains the case that drugs involuntary) the incidence of poverty increases by 1.4 account for the majority of health spending by households. percentage points (relative difference 4.1%) in 2002/03 In 2002/03 there were enormous disparities in health and 0.8 percentage points (relative difference 2.9%) in spending across wealth quintiles and, if anything, 2007/08. These findings suggest the impoverishing effect inequality has widened over the five year period. The of out-of-pocket payments is not substantial. incidence of catastrophic spending has fallen slightly over the five years between the two surveys, but it is difficult to Using an international poverty line of $1.08 per capita to interpret this finding. The poorest households, while less facilitate country comparisons, the difference in the exposed than richer households to catastrophic health care poverty headcount in Lao PDR pre and postpayment is 1.3 payments, are likely to have suffered the health percentage points in 2002/03. This estimate compares consequences of not seeking health care. Finally, findings with 1.1 percentage points and 3.8 percentage points for suggest that health spending does not contribute Vietnam and Bangladesh respectively. Note the limitations substantially to increased levels of poverty. above of such international comparisons. Lao PDR is currently engaged in implementing four risk- 9 The main differences are in the recall period (most surveys use protection mechanisms, all of which seek to increase one month for outpatient expenditures and one year for utilisation of health care services and provide financial inpatient expenditures) and items included in the definition of protection to families. These include health equity funds health spending. 9 (HEF), community based health insurance (CBHI), the Such a policy could be targeted towards specific high Civil Servants' Scheme (CSS), and Social Health Insurance priority health services (e.g. maternal and child health (SHI) for private sector employees and state-owned services), a level of health care (e.g. primary health) or companies. While it is beyond this briefing note to geographical areas (e.g. poorest regions). Lao PDR is examine rigorously whether these various initiatives have currently piloting a scheme in which delivery care is had an impact on their intended objectives, it is sobering provided free and it will be important to learn from this to note that utilisation of health services has fallen between experience. Key steps in the implementation of free care 2002/03 and 2007/08, and total out-of-pocket spending have been clearly articulated on the basis of international on health care has risen in real terms. Hence, it seems that experience [16]. In Lao PDR, particular consideration these risk protection mechanisms have not contributed would need to be given to the financing and procurement significantly to greater financial protection of the of drugs, which represents the most important driver of population, most likely because the coverage of these out-of-pocket health spending. programs remains very low (approx. 10 percent of the population. Second, existing financial protection mechanisms could be built upon. Government financing could be used to The context behind these findings is significant. The provide subsidies for the enrolment of households in CBHI government spends less than 1 percent of GDP on or expand HEFs. Given the low coverage of both types of health--very low by international standards. Moreover, schemes, it is reasonable to assume that they are unlikely government health spending has not risen in real terms to reach a substantial proportion of the rural population over the past decade. This situation may explain unless government subsidies are introduced (or increased enthusiasm for mechanisms that seek to mobilise resources in the case of HEFs). Countries that have expanded from private and non-governmental sources. Increased coverage of various types of insurance rapidly, such as financial protection and greater progress towards universal Vietnam, Thailand and China, have relied heavily on coverage are unlikely if there is not a substantial increase government tax funding combined with, in some cases, in general tax financing of health to subsidise health mandatory enrolment [17-19]. services, particularly for the poor. Such a source depends heavily on having a strong tax base and growing economy. Equity and efficiency should be key considerations in any In this regard, new opportunities are emerging. Revenues chosen strategy. The removal of user fees, for example, from hydropower and other natural resources are may be administratively less costly than CBHI or HEFs, increasing, and tax administration has improved. But there but there may be concerns about a potential lack of is no guarantee that the health sector will be the main targeting. The development of a new health financing beneficiary. With this opening of fiscal space, a strong case strategy provides an opportunity to lay out in more detail for increased government spending on health must be the pros and cons of various financing options in Lao PDR. advanced. This should prove useful in establishing a vision for how the Government of Lao PDR will move towards universal The case would be strengthened if there were clear plans coverage over the next decade. Financing reforms on the on how increased funds in the health sector should be scale needed will take many years to accomplish, and their used. A wide range of options are available, and while success will inevitably depend on the quality of international evidence provides a good guide to the kind of implementation. health interventions that should be prioritised, much less is known about the effectiveness of different financing References strategies to deliver these interventions. Experiences of developing countries that have made considerable progress 1. UNICEF, Tracking progress in maternal, newborn and in moving towards universal coverage suggest that strong child survival. 2008, UNICEF: New York. political leadership and substantial government subsidies 2. Department of Statistics and UNICEF, Lao PDR are both crucial whatever strategy is embarked upon. Multiple Indicator Cluster Survey 2000. 2000, Given recent experiences in Lao PDR, two options in Department of Statistics and UNICEF: Vientiane, particular are likely to be at the forefront of policymakers' Lao PDR. minds. 3. Department of Statistics and UNICEF, Lao PDR Multiple Indicator Cluster Survey 2006, Final Report. 2008, Department of Statistics and UNICEF: First, the public health system could be further subsidised so that at least some services are available free of charge. Vientiane, Lao PDR. 10 4. Hogan, M.C., et al., Maternal mortality for 181 countries, 1980-2008: a systematic analysis of progress towards Millennium Development Goal 5. Lancet, 2010. 375(9726): p. 1609-23. 5. UNICEF, Tracking progress in maternal, newborn and child survival. 2010, UNICEF: New York. 6. Mills, A. and J. Frenk, Constraints to scaling up and costs: Working group 1 report. 2009, Taskforce on Innovative Financing for Health Systems: London. 7. WHO, National Health Accounts Database. 2010, World Health Organization: Geneva. 8. Sen, A., Health: perception versus observation. BMJ, 2002. 324(7342): p. 860-1. 9. WHO, World Health Report 2000. Health systems: improving performance. 2000, Geneva: World Health Organization. 10. Patcharanarumol, W., et al., Assessment of the Lao Economic and Consumption Survey and the estimation of annual household spending on health. 2009, International Health Policy Program, Thailand: Vientiane. 11. Rannan-Eliya, R., National Health Accounts Estimation Methods: Household Out-of-pocket Spending in Private Expenditure. 2007, Institute for Health Policy: Colombo. 12. O'Donnell, O., et al., Analysing health equity using household survey data: a guide to techniques and their implementation. 2008, Washington DC: World Bank. 13. Wagstaff, A. and E. van Doorslaer, Catastrophe and impoverishment in paying for health care: with applications to Vietnam 1993-1998. Health Econ, 2003. 12(11): p. 921-34. 14. Xu, K., et al., Household catastrophic health expenditure: a multicountry analysis. Lancet, 2003. 362(9378): p. 111-7. 15. van Doorslaer, E., et al., Effect of payments for health care on poverty estimates in 11 countries in Asia: an analysis of household survey data. Lancet, 2006. 368(9544): p. 1357-64. 16. Gilson, L., The lessons of user fee experience in Africa. Health Policy Plan, 1997. 12(4): p. 273-85. 17. Ekman, B., et al., Health insurance reform in Vietnam: a review of recent developments and future challenges. Health Policy Plan, 2008. 23(4): p. 252-63. 18. Lieberman, S. and A. Wagstaff, Health Financing and Delivery in Vietnam: Looking Forward. 2009, Washington DC: World Bank. 19. Wagstaff, A., et al., Reforming China's Rural Health System. 2009, Washington DC: World Bank. 11 Table A1. Proportion of individuals who sought outpatient care in the past 4 weeks, according to background characteristics Outpatient care sought (%) Outpatient care sought when ill (%) LECS3 (2002-03) LECS4 (2007-08) LECS3 (2002-03) LECS4 (2007-08) Background characteristic Modern Traditional Modern Traditional Modern Traditional Modern Traditional provider practitioner provider practitioner provider practitioner provider practitioner Age 0-5 years old 2.1% 0.2% 2.2% 0.3% 10.6% 1.0% 16.7% 1.1% 6-10 years old 1.1% 0.2% 1.1% 0.3% 9.6% 1.7% 13.2% 2.1% 11-15 years old 0.8% 0.2% 0.7% 0.2% 9.2% 1.9% 12.4% 2.5% 16-20 years old 1.0% 0.3% 0.9% 0.1% 9.9% 2.7% 19.3% 2.1% 21-65 years old 3.0% 0.7% 2.4% 0.5% 16.2% 4.0% 21.0% 4.9% 65+ years old 4.0% 0.9% 3.4% 0.7% 12.4% 3.2% 11.8% 2.4% Type of residence Urban 2.5% 0.4% 2.3% 0.5% 17.6% 2.8% 22.0% 3.8% Rural with road access 2.4% 0.5% 1.6% 0.3% 15.6% 3.7% 17.2% 3.3% Rural without road access 1.3% 0.4% 1.8% 0.4% 7.9% 2.0% 13.6% 3.5% Priority districts 1st priority districts 1.4% 0.3% 1.4% 0.3% 8.5% 1.5% 14.3% 2.9% 2nd priority districts 1.6% 0.4% 1.2% 0.2% 10.3% 2.5% 13.3% 2.7% Non-priority districts 2.4% 0.5% 2.1% 0.4% 15.7% 3.6% 20.3% 3.8% Region Vientiane 2.7% 0.5% 2.7% 0.3% 18.9% 3.8% 33.9% 5.0% North 1.7% 0.3% 1.7% 0.2% 10.1% 1.9% 14.8% 1.7% Central 2.1% 0.4% 1.7% 0.4% 16.7% 3.7% 21.7% 4.5% South 2.2% 0.5% 1.9% 0.5% 11.2% 3.0% 13.6% 4.4% Ethnic group Lao Tai 2.3% 0.5% 2.0% 0.3% 15.7% 3.6% 20.7% 4.1% Mon-Khmer 1.5% 0.4% 1.8% 0.5% 7.7% 1.9% 13.1% 2.7% Chine Tibet 1.0% 0.1% 1.5% 0.2% 5.7% 0.8% 15.1% 2.0% Hmong-Mien 1.7% 0.2% 1.1% 0.1% 13.3% 1.5% 15.2% 1.3% Other 3.2% 0.5% 0.2% 0.2% 11.8% 1.4% 2.2% 2.2% Consumption quintile Poorest (q1) 1.1% 0.4% 1.0% 0.2% 6.7% 2.4% 10.4% 2.0% Poorer (q2) 1.6% 0.3% 1.3% 0.2% 10.4% 2.1% 13.7% 2.7% Middle (q3) 2.1% 0.2% 1.7% 0.5% 12.8% 1.7% 16.3% 2.8% Richer (q4) 2.5% 0.6% 1.9% 0.3% 16.3% 3.7% 19.6% 3.8% Richest (q5) 2.8% 0.6% 2.8% 0.5% 17.6% 4.2% 25.0% 5.1% Total 2.1% 0.4% 1.8% 0.3% 13.2% 2.9% 18.1% 3.5% Note: Note: Modern provider includes a visit to a health facility, or an individual private health provider (doctor, nurse, midwife). Data are weighted to account for survey design. 12 Table A2. Number of outpatient visits per capita per year by type of provider, according to background characteristics Health facility Private provider Traditional healer Background characteristic LECS3 (2002-03) LECS4 (2007-08) LECS3 (2002-03) LECS4 (2007-08) LECS3 (2002-03) LECS4 (2007-08) Age 0-5 years old 0.30 0.33 0.14 0.20 0.06 0.04 6-10 years old 0.14 0.17 0.06 0.15 0.04 0.04 11-15 years old 0.09 0.11 0.08 0.07 0.03 0.03 16-20 years old 0.17 0.10 0.08 0.04 0.05 0.01 21-65 years old 0.56 0.50 0.33 0.20 0.22 0.13 65+ years old 0.72 0.71 0.54 0.31 0.18 0.16 Type of residence Urban 0.52 0.46 0.23 0.25 0.18 0.09 Rural with road access 0.39 0.30 0.26 0.13 0.11 0.07 Rural without road access 0.20 0.30 0.12 0.15 0.09 0.11 Priority districts 1st priority districts 0.24 0.25 0.11 0.09 0.08 0.06 2nd priority districts 0.27 0.20 0.17 0.11 0.08 0.06 Non-priority districts 0.41 0.41 0.25 0.20 0.15 0.10 Region Vientiane 0.64 0.58 0.20 0.16 0.21 0.06 North 0.28 0.27 0.14 0.13 0.09 0.04 Central 0.33 0.32 0.21 0.20 0.10 0.07 South 0.35 0.37 0.30 0.16 0.16 0.18 Ethnic group Lao Tai 0.40 0.40 0.23 0.18 0.15 0.09 Mon-Khmer 0.25 0.25 0.15 0.18 0.08 0.10 Chine Tibet 0.14 0.21 0.12 0.14 0.01 0.05 Hmong-Mien 0.24 0.22 0.14 0.05 0.05 0.02 Consumption quintile Poorest (q1) 0.15 0.20 0.13 0.10 0.10 0.04 Poorer (q2) 0.27 0.21 0.20 0.11 0.12 0.06 Middle (q3) 0.31 0.25 0.18 0.18 0.05 0.07 Richer (q4) 0.42 0.34 0.24 0.16 0.13 0.10 Richest (q5) 0.56 0.62 0.26 0.24 0.20 0.13 Total 0.36 0.35 0.20 0.16 0.12 0.08 Note: Data are weighted to account for survey design. The number of outpatient visits per year is calculated as the product of the number of outpatient visits in the previous one month and the number of months in a year (ie. 12). 13 Table A3. Proportion of individuals who were hospitalised in the past year, according to background characteristics Inpatient care sought (%) Background characteristic LECS3 (2002-03) LECS4 (2007-08) Age 0-5 years old 1.8% 1.9% 6-10 years old 1.9% 1.0% 11-15 years old 1.2% 0.9% 16-20 years old 1.1% 1.4% 21-65 years old 3.0% 2.3% 65+ years old 4.7% 3.9% Type of residence Urban 2.5% 2.2% Rural with road access 2.6% 1.7% Rural without road access 1.6% 2.0% Priority districts 1st priority districts 2.2% 2.1% 2nd priority districts 1.7% 1.2% Non-priority districts 2.4% 1.9% Region Vientiane 2.3% 2.1% North 2.0% 2.0% Central 1.8% 1.6% South 3.3% 2.0% Ethnic group Lao Tai 2.4% 2.0% Mon-Khmer 2.1% 1.7% Chine Tibet 0.9% 1.8% Hmong-Mien 2.1% 1.1% Other 2.4% 2.4% Consumption quintile Poorest (q1) 1.3% 1.1% Poorer (q2) 1.7% 1.5% Middle (q3) 1.9% 1.6% Richer (q4) 2.8% 2.1% Richest (q5) 3.2% 2.6% Total 2.3% 1.9% Note: Data are weighted to account for survey design. 14 Table A4. Financial burden of health care payments and incidence of catastrophic spending, according to background characteristics OOP spending on health as share of total Incidence of catastrophic health care spending Background characteristic household consumption using 10% threshold LECS3 (2002-03) LECS4 (2007-08) LECS3 (2002-03) LECS4 (2007-08) Type of residence Urban 1.7% 2.1% 4.0% 4.4% Rural with road access 2.1% 1.6% 4.9% 3.7% Rural without road access 1.5% 1.1% 3.4% 2.3% Priority districts 1st priority districts 1.3% 1.3% 2.6% 2.8% 2nd priority districts 1.8% 1.1% 4.0% 2.3% Non-priority districts 2.1% 2.0% 4.8% 4.4% Region Vientiane 1.8% 1.9% 5.1% 4.7% North 1.6% 1.4% 3.6% 2.9% Central 1.8% 1.7% 4.1% 3.7% South 2.3% 1.9% 4.8% 4.7% Ethnic group Lao Tai 2.1% 1.9% 4.7% 4.2% Mon-Khmer 1.4% 1.5% 2.9% 3.3% Chine Tibet 0.4% 0.8% 0.6% 1.3% Hmong-Mien 1.9% 1.1% 4.9% 2.7% Other 1.6% 2.4% 1.3% 3.7% Consumption quintile Poorest (q1) 1.1% 0.6% 1.8% 1.2% Poorer (q2) 1.5% 1.2% 2.6% 2.4% Middle (q3) 1.8% 1.5% 4.5% 3.6% Richer (q4) 2.3% 2.0% 5.9% 4.6% Richest (q5) 2.6% 3.2% 6.5% 7.2% Total 1.8% 1.7% 4.2% 3.8% Note: Data are weighted to account for survey design. 15