2 17 ..

2556

The Impacts of Households Insurance Enrollment on Health Care Financial Protection in Vietnam
Le Hong Chung1

Abstract
The Vietnamese Government implemented the Law on Health Insurance in 2009, aimed at achieving Universal Health Coverage by 2014. However, we only achieved 60% coverage of population in 2010 and evidences of health insurance impacts were reported weakly. The main objective of this study is to determine the impacts of health insurance enrollment rate within a household on health care financial protection. We used the data from the Vietnam Household Living Standard Survey (VHLSS) in 2010. Logit model was employed for modeling the determinants of catastrophic health spending. The main findings of this study were: Higher health insurance enrollment rate within a household was found associated with lower probability of catastrophic health expenditure. More member in the households, living in the rural areas and higher asset index score were associated with lower rate of catastrophic encountering; while households with more elderly and children were more likely facing catastrophic health spending. Key words: Household, Insurance Enrollment, Health Care, Financial Protection

Introduction
In 2008, the Government of Vietnam implemented the Law on Health Insurance, taking effect on 1st July 2009. The Law aimed at achieving Universal health coverage (full coverage) in 2014 by mandatory health insurance (Government of Vietnam, 2008). However, after nearly 2 year implemented, we only achieved 60% of population coverage by national health insurance (Tien, Phuong, Mathauer, & Phuong, 2011). By the law, the main groups of individuals under the mandatory health insurance in 2010 were the employees, employers, civil servants, the poor, the children under 6 years old and social protection groups and the students; while others (such as the informal sector, farmers and dependents of employees) were still under the voluntary health insurance scheme, up to 2014. However, due to low compliance to the law, the
1

Center for Health System Research, Hanoi Medical University. Candidate of MSc. Health Economics and Health Care Management, Faculty of Economics, Chulalongkorn University; Email: hongchung_chsr@hmu.edu.vn

[91]

 2 17 ..2556
remaining groups now are the near poor, students, farmers, informal sector and their dependents. As the matter of fact, the out-of-pocket health expenditure (OOP) was remaining quite high (51%) (Van Minh, Kim Phuong, Saksena, James, & Xu, 2012). Moreover, the financing budget of national health insurance was still not sustainable due to the adverse selection problem (some people who havent been mandatory yet may choose to buy health insurance only when having diseases). For instance, in the past, the Vietnam Social Security (VSS) faced to the deficit funding problem for the budget of voluntary health insurance (Tien et al., 2011). How to expand the health insurance coverage was head-aching with all policy-makers and stakeholders. Individual-based health insurance was the only administration procedure up to now; nevertheless, the disadvantage and less effectiveness of that application gradually exposed. According to the Law, there were 25 categories of health insurance membership classification. The overlap among the beneficiary groups of the Government subsidy policies, for example: free health insurance card for the poor and children under 6 years old; the near poor and students with 50% and 30% of premium subsidy. People also didnt fully understand their benefits and interests of health insurance due to complicated administration system like this. As the result, this approach is hard to continue to be used so the Government is trying to roll out the universal coverage at the household level by the next coming years. It means that the classification of health insurance membership founded on individual level may turn into the household level. Household-based insurance may help reduce the administration and management workload of the VSS. It also helps achieve more population coverage of national health insurance if with a reasonable and attractive premium rate. Lack of evidences about the financial protection of households health insurance enrollment now is the most concern. Thats why we would like to conduct this research to provide critica l evidences for the Government and all stakeholders.

Objective and Scope

Objective The main objectives of this paper are answering these following questions: 1. What is the current health insurance enrollment rate among households, classified by different socioeconomic characteristics? 2. What are the impacts of household insurance status on health care financial protection (e.g. from catastrophic spending)? Scope In this paper, we analyzed the financial protection of health insurance from the households perspective about insurance enrollment rate within a household together with other socioeconomic factors. [92]

 2 17 ..2556
Secondary data used in this study was selected from the latest nationally representative household survey in 2010 - Vietnam Household Living standard surveys (VHLSSs).

Conceptual Framework
Information of individuals health insurance status is transferred into household level by taking the rate of number of people have health insurance divided by the household size. Logit model is used to estimate the impacts of health insurance from household perspective on financial protection. Explainable variables of catastrophic are health insurance enrollment rate within a household, number of the elderly and children in household, household size, location (urban/rural), the expenditure quintile; the head of households characteristics (e.g. Gender, education, work in informal sector) and the asset index score.

Literature Review
There were a lot of literatures that examined the effects of health insurance on financial protection: catastrophic health care spending and impoverishment. Generally, most researches that tested for specific types of health insurance or individual health insurance status but not the households insurance enrollment rate. Nevertheless evidences of health insurance effects in Vietnam was reported modestly in various studies. Adam W. and Eddy v D. investigated the financial burden for health care in Vietnam, using data from VLSSs in 1992-93 and 1997-98 (Wagstaff & van Doorslaer, 2003). Due to have no available information about health insurance at that time, Wagstaff just found that catastrophic at threshold 10% happen more among the rich rather than the poor. In a study of 6 Asian countries, discussion about catastrophic of threshold 10% and its determinants had been performed deeply. For health insurance status, information was only available in Hong Kong, Thailand and Vietnam and with individual insurance status. A research using data from 5 VHLSSs in the whole period of 2002-2010 studied about financing load in Vietnamese households for health care (Van Minh et al., 2012). They used the independent variable as household had at least one member with health insurance. There were a large number of studies evaluated the effects of particular health insurance programs or health policies in Vietnam, such as Voluntary Health Insurance (VHI), health insurance for Children (Nguyen, 2012a, 2012b). Looking at a systematic review reported about the impact of health insurance for the poor and informal sector in developing countries in 2012, 34 studies were analyzed and evaluated. But almost studies [93]

 2 17 ..2556
paid attention to the specific type of health insurance or on an individual level, rather than the rate of insured member within a household (Acharya et al., 2012). As the existing knowledge above, its obvious to state that studying about the impacts of the households insurance status on health care financial protection is really necessary and valuable.

Research Methods
As evidently defined in some books and articles, we use the definition and construction method of catastrophic in World Health Organization guideline (Wagstaff, 2008; World Health Organization, 2005; Xu et al., 2003) below: Out-of-pocket health payments: Out-of-pocket health payments bear on the direct spending by households every time when they acquire health services, include doctors fees, spending for tests, medicines and hospital bills. OOP also are net of insurance reimbursement. Households consumption expenditure: Household consumption expenditure embraces in term of either monetary or in-kind on all goods and services, and the money value of the consumption of home-based products. Household subsistence spending: A subsistence spending (minimum spending require for basic life) can be explored through a poverty line. Poverty line is defined from households spending on food as a share of total household expenditure, at the 50th of the country. Here to avoid measurement bias, its necessary to use the figure within the 45th and 55th percentile of the total sample. Considering the economy scale of household consumption, the household equivalence scale is used rather than actual household size. The value of the parameter has been estimated from prior studies derived from 59 countries household survey data, and it equals 0.56: The equivalized food for household is calculated by the equation: Then Poverty line is defined as below: So that, subsistence expenditure for each household is equal:

[94]

 2 17 ..2556
Households capacity to pay: A households capacity to pay is derived as effective income remaining after fundamental subsistence needs have been met. Effective income is taken to be the total consumption expenditure of the household. However, some households may report food expenditure that is lower than subsistence spending (SE > food). Such a situation could also be due to the fact that the reported food expenditure in the survey does not consider food subsidies, coupons, self-production and other non-cash means of food consumption. In this particular case the non-food expenditure is used as non-subsistence spending. Catastrophic health expenditure: Catastrophic heath expenditure crops up when a households total out-of-pocket payments for health equal or transcend 40% of households capacity to pay (We also used the threshold point of 10%, 20% and 30%). The dependent variables are dummy variables on incurring catastrophic health expenditure CATA40, CATA30, CATA20 and CATA10 (out-of-pocket spending for health as proportion of household capacity to pay at cut-off points 40%, 30%, 20% and 10%, respectively). The independent variables are: Households enrolled status: HI = Total number of insured members/ household size. Head of household characteristic: The head of household is a man (Manhead=1); the head of households has low education (highest education is less than high school Loweduhead=1) and the head of household works in informal sector (Infohead =1). Here informal sector is defined as: agriculture households and self business. And other socioeconomic indicators such as household size, number of elderly people in the household (>65 years old), number of children under 6 years old in the household, living area (urban/rural), household expenditure quintiles and asset index score. Asset index score is derived by Principal Component Analysis with 21 main types of durable assets and 6 housing conditions. We have the function: CATA_n = f (Hi, Hhsize, Number of children<6, Number of elderly, Urban, Quintile_i, Asset index score, Head of household characteristics) Or:

[95]

 2 17 ..2556

+ Where, CATA_n = Dummy variables of household facing with Catastrophic health expenditure at 10%, 20%, 30%, 40% of households capacity to pay. HI = Health insurance enrollment rate within a household Hhsize = Household size Age6_n = Number of children under 6 years old in family Age65_n = Number of elderly (over 65 years old) in family Urban = Household living in urban areas (Dummy variable, Urban=1) Qunitile_i = Expenditure quintiles of household (Dummy variables) Z_pcwi = Asset index score Manhead = The head of household is a man (Dummy, Manhead=1) Loweduhead = The head of household had low education (Dummy, Loweduhead=1) Infohead = The head of household working in informal sector (Dummy, Infohead=1) The coefficients and marginal effects are used to interpret the likelihood of incurring catastrophic health spending in households.

Research Results
Table 1: Health insurance enrollment rate within a household Mean Total Type of Health Insurance 0.593 SD 0.362 0.391 0.272 0.387 0.37

Mandatory Voluntary
Expenditure Quintile

0.385 0.208

1st 2nd
[96]

0.682 0.532

 2 17 ..2556 3rd 4th 5th

Location 0.519 0.568 0.664 0.636 0.574 0.35 0.339 0.333 0.347 0.367

Urban Rural

In Table 1 and Table 2, the average coverage rate of health insurance within a household was 0.593. Only 34.51% of households had full health insurance enrollment and 14.83% of households have no health insurance. Among different expenditure quintile groups, the first quintile group had the highest average health insurance coverage rate and the 3rd quintile had the lowest rate. The 1st quintile group also had the highest percentage of households have full health insurance enrollment. The urban area had the higher health insurance rate than the rural. Our results point out that the Government free health insurance for the poor somehow targeted the right group. Table 2: Classification of household health insurance enrollment Full coverage Total Quintile 34.51% Partial coverage 50.66% 30.62% 51.55% 58.59% 60.09% 52.46% 51.50% 50.30% Non - coverage 14.83% 15.38% 18.59% 17.13% 13.75% 9.30% 11.67% 16.22% Chi square test -

1st 2nd 3rd 4th 5th

Location

54.00% 29.86% 24.28% 26.16% 38.24% 36.83% 33.49%

p < 0.05

Urban Rural

p < 0.05

The table 3 below shows the incidence of catastrophic health expenditure at 10%, 20%, 30% and 40% of household capacity to pay (27.45%, 13.61%, 7.93% and 4.7%, respectively).

[97]

 2 17 ..2556
Table 3: Probability of households facing catastrophic health payment Number of households (N) Cata10 Cata20 Cata30 Cata40 6,131,343 3,039,607 1,771,324 1,050,166

Percentage (%) 27.45% 13.61% 7.93% 4.70%

In the table 4 below, we present the picture of catastrophic with socioeconomic characteristics. The 2 quintile had the highest rate of catastrophic incurring while the 5th quintile had the lowest probability of catastrophic health spending. Household living in urban area and having at least one insurance had lower incidence of catastrophic.
nd

Table 4: Pattern of catastrophic with socioeconomic characteristics Cata10 Quintile Cata20 16.86% 16.24% 13.66% 13.24% 8.04% 15.30% 9.76% 17.00% 13.02% Cata30 9.36% 9.10% 8.67% 7.93% 4.60% 9.11% 5.23% 10.92% 7.41% Cata40 5.03% 5.77% 4.95% 4.65% 3.11% 5.41% 3.09% 7.29% 4.25%

1st 2nd 3rd 4th 5th

Location

32.76% 32.67% 29.32% 25.19% 17.30% 30.20% 21.18% 31.73% 26.70%

Rural Urban
Health insurance2

No Yes

Table 5: Determinants of catastrophic health expenditure at cut-off point 10%, 20% 30% and 40% Variables CATA10 CATA20 CATA30 CATA40

Yes = Household had at least one health insurance card No = Household had no insurance

[98]

 2 17 ..2556
Coef. Marginal effect Coef. Marginal effect Coef. Marginal effect -0.008 -0.013 -0.014 0.039 0.011 -0.030 Coef. -2.462* -0.261** -0.183 -0.360* 0.748* 0.303* -0.584* Marginal effect -0.008 -0.006 -0.012 0.024 0.010 -0.019

C -0.688* -1.454* -2.102* HI -0.199* -0.038 -0.232* -0.025 -0.127 Manhead -0.173* -0.033 -0.119 -0.013 -0.211* Hhsize -0.090* -0.017 -0.161* -0.017 -0.233* Age65_n 0.607* 0.117 0.610* 0.065 0.652* Age6_n 0.194* 0.037 0.205* 0.022 0.186* Urban -0.279* -0.054 -0.340* -0.036 -0.495* 1st quintile (Reference) 2nd quintile 0.225* 0.043 0.280* 0.030 0.439* 3rd quintile 0.194* 0.037 0.254* 0.027 0.639* 4th quintile 0.124 0.024 0.416* 0.044 0.806* 5th quintile -0.153 -0.029 0.125 0.013 0.576* Loweduhead 0.160* 0.031 0.098 0.010 0.093 Infohead -0.122 -0.024 -0.121 -0.013 -0.093 Z_pcwi -0.216* -0.042 -0.333* -0.035 -0.474* *: Coefficients are significantly at the 5% significance level (p<0.05). **: Coefficients are significantly at the 10% significance level (p<0.1).

0.026 0.038 0.048 0.034 0.006 -0.006 -0.028

0.715* 0.855* 1.093* 1.099* 0.074 -0.174 -0.548*

0.023 0.028 0.035 0.035 0.002 -0.006 -0.018

The Table 5 above shows the results of Logistic regressions for determinants of Catastrophic 10%, 20%, 30% and 40%. The characteristics have the significant effects on almost Catastrophic thresholds at 95% of confidence interval are: Health insurance coverage rate within a household, household size, the number of elderly, the number of children under 6, living in urban area, the second and the third quintile group and asset index score (p<0.05). The main findings of the models are: 1) The households with higher rate of health insurance coverage were less likely facing catastrophic expenditure 2) Large household size was significantly associated with less likely incurring catastrophic; 3-4) Households with more number of elderly or children under six had significantly higher chance of catastrophic expenditure. 5) Households living in urban areas were less likely facing catastrophic health payment than ones living in rural area. 6-7) Households belonged to the second and the third expenditure quintiles were more likely to spend catastrophic payment for health than the first

[99]

 2 17 ..2556
quintile households and 8) Households with higher score of wealth index were less likely to face catastrophic health spending. The findings of our research are on par with other studies (Somkotra & Lagrada, 2009; Van Minh et al., 2012). Our results suggest that the Government policies should focus more on the near poor group and the rural area to attract them with health insurance. Also the higher coverage rate in a household, the better financial protection for household when suffer from catastrophic health cost.

Conclusion
Different from the results of other studies, our study point out the effects of health insurance coverage at household level significantly. It will be a critical evidence not only for policy-makers to implement the new health insurance management approach: household-based insurance, but also for increasing the knowledge about health insurance and compliance of Vietnamese people with the Law. Once the mandatory household-based insurance approach successfully implemented with the whole population, it may help not only increase the risk pooling within a household, for example children versus adults, old versus young, woman versus man, but also reduce the adverse selection because all members of a household will involve in the scheme whether having health problems or not. Our results also suggest that the group of people or households should be subsidized and focused more in the future are the elderly, the near poor households and those living in rural areas.

Reference
Acharya A., Vellakkal S., Taylor F., Masset E., Satija A., Burke M., & Ebrahim S. 2012. Impact of national health insurance for the poor and the informal sector in low- and middle-income countries: a systematic review (pp. 110 pp.). London, UK: EPPI-Centre, Social Science Research Unit, Institute of Education, University of London. Government of Vietnam. 2008. Law on Health Insurance 2008. Nguyen, C. V. 2012. The Impact of Health Insurance for Children: Evidence from Vietnam. DEPOCEN Working Paper Series. Nguyen, C. V. 2012. The impact of voluntary health insurance on health care utilization and out-of-pocket payments: new evidence for Vietnam. Health Econ 21(8) : 946-966. doi: 10.1002/hec.1768.

[100]

 2 17 ..2556
Somkotra, T., & Lagrada, L. P. 2009. Which households are at risk of catastrophic health spending: experience in Thailand after universal coverage. Health Aff (Millwood) 28(3) : 467-478. doi: 10.1377/hlthaff.28.3.w467. Tien, T. V., Phuong, H. T., Mathauer, I., & Phuong, N. T. K. 2011. A Health Financing review of Vietnam with a focus on social health insurance (Bottlenecks in institutional design and organizational practice of health financing and options to accelerate progress towards universal coverage) : 4-15. World Health Organization. Van Minh, H., Kim Phuong, N. T., Saksena, P., James, C. D., & Xu, K. 2012. Financial burden of household out-of pocket health expenditure in Viet Nam: Findings from the National Living Standard Survey 20022010. Social Science & Medicine : 1-6. doi: http://dx.doi.org/10.1016/j.socscimed.2012.11.028. Wagstaff, A. (Ed.). (2008). Measuring Financial Protection in Health : Worldbank. The Human Development and Public Services Team, Development Research Group. Wagstaff, A., & van Doorslaer, E. 2003. Catastrophe and impoverishment in paying for health care: with applications to Vietnam 1993-1998. Health Econ 12(11) : 921-934. doi: 10.1002/hec.776. World Health Organization. 2005. Distribution of health payments and catastrophic expenditures methodology. Geneva: World Health Organization. Xu, K., Evans, D. B., Kawabata, K., Zeramdini, R., Klavus, J., & Murray, C. J. 2003. Household catastrophic health expenditure: a multicountry analysis. Lancet 362 (9378) : 111-117. doi: S0140-6736(03)13861-5 [pii] 10.1016/S0140-6736(03)13861-5.

[101]