Comparison of inequity in health-related quality of life among unemployed and employed in China


 Background

In China, achieving the health equity has been regarded as a key issue of health reforms and development in the current stage. It is well known that unemployment have a negative effect on health. However, few studies have addressed the association between unemployment and the inequity of health-related quality of life (HRQOL). The study aims to compare the inequality and inequity in HRQOL among the unemployed and employed in China.

Methods

The material regarding this study has been illustrated from the Chinese National Health Services Survey (NHSS) of Shaanxi Province for 2013. We have controlled the confounding factors by the utilization of coarsened exact matching method (CEM). Finally, 7,524 employed individuals and 283 unemployed individuals were aged 15 to 64 in urban area has been incorporated for this study. We use HRQOL as the outcome variable, which was evaluated by the Chinese version of EQ-5D-3L. Health concentration index, decomposition analysis based on the Tobit model and the horizontal inequity index were employed to compute the income-related equity and the contribution of factors among the unemployed and employed.

Results

After matching, unemployed people tended to express poorer EQ-5D utility than employed people. The horizontal inequity indices among employed and unemployed people were 0.0020 and 0.0077 respectively, demonstrating that the pro-rich health inequity of unemployed people was a great deal superior to employed people. Economic status, age, education, smoking and health insurances are the main impact factors that affect the inequality in HRQOL among employed and unemployed. Education status and basic health insurances reduce the pro-rich inequity in HRQOL for the unemployed.

Conclusion

It is suggested that unemployment intensifies the inequality and inequity in HRQOL. In view of the policy makers, the basic health insurances are still a critical health policy for improving health equity of the unemployed. Re-employment programs, the socialization of medical health insurances, initiatives to improve educational equity and the psychological counseling for the unemployed should be considered by government to attain health equity.

Health equity has gradually become a research hotspot in the field of health systems reform [1, 55 2] . Achieving health equity has been widely concerned about, supported and responded by all 56 countries of the world [3] . China also regards the realization of health equity as the key issue 57 of health reform and development in the current stage. Specifically, the planning outline of 58 "Healthy China 2030" proposed that we should focus on health problems of vulnerable 59 groups of people and to achieve the health equity [4] . Health inequalities is not only affected 60 by physiological conditions but also widely determined by socioeconomic characteristics, 61 inequalities may be further widened by unemployment [5] . The World Health Organization 62 proposed that each countries should set up health equity monitoring systems to reduce health 63 inequalities through collecting key indicators like employment status which can be 64 determined by the labor market [6] . Unlike retired people, most unemployed people quit the 65 labor force for non-physiological reasons and cannot sell their labor at a balanced price in the 66 market [7] . 67 There is a body of literature exploring the association between unemployment and 68 health and paying attention to the different dimensions of lifestyle behaviors (e.g. alcohol 69 4 of 25 consumption and smoking) [8,9] , mental health (e.g. depression, mental disorder and suicide 70 thoughts) [10][11][12][13] and self-reported health [14][15][16] . Experimental evidences has demonstrated 71 that unemployment has a severe depressing outcome on health [17,18] and may lead 72 households into a cycle of poverty [19] . There is also some evidence that unemployment had 73 a positive or no effect on health. Therefore, it is logical to start from the key groups and to 74 carry out the research on inequity in health-related quality of life among the unemployed. It 75 is with great consequence to prevent the unemployed from falling into long-term health and 76 poverty, to improve the precise poverty alleviation policy and to promote the construction 77 of healthy China. 78 Despite many health measurement being used to assess the effect of unemployment on 79 health, there is still not much knowledge about health-related quality of life [20] . 80 Health-Related Quality of Life (HRQOL) is generally considered as a key measurement 81 indicator of health care outcomes and is a multidimensional construct that relates to a 82 person's self-perceived health [21] . The EuroQol 5 dimensions (EQ-5D) is a standardized 83 instrument and the most commonly used instrument for measuring the quality of life in 84 public health research [22,23] . Some recent studies examine the correlates of unemployment 85 and HRQOL by using SF-8 instrument, SF-12 instrument and SF-36 instrument [24][25][26][27] , but 86 using EQ-5D instrument are lacking. In addition, few studies have addressed the 87 association between unemployment and the inequity of HRQOL. HRQOL of unemployed and employed by using EQ-5D-3L based on Chinese-preferences 95 tariff. Furthermore, we will offer well-informed estimates on the associations between 96 unemployment and income related inequality and inequity of Chinese HRQOL under 97 consideration. The third key strength is the findings of this investigation guarantee better 98 balance between the unemployed and employed groups by using the coarsened exact 99 matching method (CEM). were identified [28,29] .

110
The NHSS survey pay attention to the health status, health services need and utilization 111 of the Chinese residents covering a broad range of information on socio-economic 112 characteristics (e.g. age, gender, education status and economic level), health (e.g.  We used EQ-5D health utility as the outcome variable. HRQOL was measured by the 119 classical 3-level EQ-5D (EQ-5D-3L), which has been widely validated and utilized in the 120 world [30] . The EQ-5D is a self-report questionnaire that including five dimensions: (1) 121 mobility (2) self-care (3) usual activities (such as work, studies, housework and leisure 122 activities) (4) pain / discomfort (5) anxiety / depression. The three response alternatives to 123 above five mentioned dimensions are: (1) no problem (2) some problems (3) extreme 124 problems [31] . Finally, we used Chinese-preferences tariff which is applicable to Chinese 125 people to generate the score of EQ-5D utility among the unemployed and employed which 126 ranges from −0.1490 (stands for the worst health) to 1 (stands for the full health) [32] .

128
In the light of the existing literature, we controlled variables for socioeconomic 129 characteristics and health behavior related to inhabitants, such as gender (0=Male, 130 1=Female), age (in years), per capita annual income (Yuan) (1= Lowest group, 0=other; 131 1=Lower group, 0=other; 1=Medium group, 0=other; 1=Higher group, 0=other; 1=Highest 132 group, 0=other), marital status (1= Single, 0=other; 1=Marriage, 0=other; 1=Widowed and 133 Divorced, 0=other), education status(1=Elementary school and below, 0=other; 1=Middle 134 school, 0=other; 1=Senior high school, 0=other; 1=College degree and above, 0=other), 135 health insurances (1=No, 0=other; 1=Basic medical insurances, 0=other; 1=commercial 136 insurances and other insurances,0=other;), smoking status (1=No smoking, 0=other; 137 1=Non-daily smoking, 0=other; 1=Daily smoking, 0=other;) and drinking status (0=don't 138 drink alcohol, 1= drink alcohol ). ignore the fact that there may be other potential confounding factors. Therefore, we adopted 143 the coarsened exact matching method (CEM) in this article, which is a new technique for 144 improving the assessment of causal inference between two groups by controlling potentially 145 confounding variables [33,34] .The original sample can be retained to the maximum extent and 146 the weighted variables generated to equalize samples within two groups during the matching 147 process [35] . The multivariate imbalance measure 1 L was employed to ensure the balance 148 before and after matching. 1 L ranges from 0 to 1, where 1 indicates that the data of two 149 comparison groups are completely unbalanced and a smaller value indicates the better 150 balance between comparison groups. The multivariate imbalance measured by the Eq.1 [34] : (1) Concentration index (CI) has been widely accepted as a standard method for measuring the 159 income related inequality of health status [36] . The CI value is between −1 and 1. The positive 160 of the CI indicates that the health is more concentrated among the members with higher per 161 capita household income and vice versa 0 indicates that there is no inequality [37] . The concentration index was computed by Eq.2 : Where C denotes concentration index, x refers to HRQOL, µ is the average of EQ-5D 165 utility value, h symbolizes the ranking of per capita household income.

166
Decomposition of the concentration index 167 The decomposition analysis is to decompose the concentration index into the contribution of 168 every variables to the inequality in HRQOL. However, the EQ-5D utility value generally has 169 a ceiling effect (i.e. most residents had the full health of 1), the decomposition analysis based 170 on Tobit model [38] was commonly used as Eq.3: Where i y is the score of EQ-5D utility; x are the unavoidable determinants of HRQOL decomposition of the concentration index C could be written as: Where µ represents the mean of the EQ-5D utility, j C denotes the concentration index of    Table 2 reported the EQ-5D health utility among the employed and unemployed in China.

201
After matching, the results indicated that the mean of EQ-5D utility scores are also

209
The CIs of the 5 dimensions and EQ-5D utility scores among employed and unemployed 210 were presented in Table 3. Although, the concentration index of the five dimensions were  The overall decomposition analysis for the EQ-5D utility values among the employed 220 and unemployed was presented in   The horizontal inequity index of HRQOL is also presented in Table 5. After deduction of 236 the contributions of need variables in health (e.g. age and gender) from the concentration 237 index of EQ-5D utility value, the horizontal inequity index of the HRQOL among employed 238 and unemployed individuals were 0.0020 and 0.0077 respectively, which entails that there is 239 a pro-rich inequity in HRQOL among the unemployed and employed. In addition, the 240 horizontal inequity was higher in unemployed as compared to the employed. aspects of this study that should to be discussed. 253 Firstly, the most fascinating finding was that there was statistically lower EQ-5D utility 254 in employed compared with the unemployed, which was for the first time to assessed 255 HRQOL among the employed and unemployed by using EQ-5D-3L instrument in China.

256
This concluded that unemployment is associated with poor HRQOL. This is consistent with 257 several reports that unemployed people more possible to have poorer HRQOL than the 258 employed [20,25,40] . Specifically, it may be due to the fact that people experienced 259 unemployment are deprived of these benefits (e.g. income, social contact, status and index not only illustrated that there are the pro-rich inequity in HRQOL among two groups, 295 but also this inequity of unemployed was still higher than employed which may be explained 296 through the reduction in income associated with unemployment [19,27] . People have unequal and employed, the government should consider the contribution of education and basic 302 health insurances schemes to reduce the pro-rich inequity.

303
At the same time, we acknowledge that the present study has some limitations. Firstly, 304 the data derived from Shaanxi Province and our conclusion may not be generalized to the 305 whole of China. Moreover, due to the cross-sectional study, causal interpretations are 306 hazardous. Therefore, we refer to associations between unemployment and HRQOL.

307
Additionally, it is difficult to solve the endogenous problem between unemployment and 308 HRQOL. The present study was subject to possible unobserved confounding factors, such as 309 the disability status, access to healthy food, social interaction and so on.

311
In conclusion, the unemployment is linked with health related quality of life and inequality in 312 HRQOL. It appeared that unemployment intensified the inequality and inequity in HRQOL.

313
Furthermore, the major contributors of the inequality in HRQOL were economic status, inequity, but also the extensive social services such as psychological counseling and spiritual 322 care should also be provided [25] . We want to express our appreciation to the Health Department of Shaanxi Province for providing data. 328 We also express our gratitude to all participants in this study for their participation and co-operation in 329 the data collection.