Comparison of changes in 2015 to 2019 for socioeconomic inequalities and inequities in health services utilization among patients with hypertension in Pearl River Delta Region of China

Background: Assessing inequities in health services utilization contributes to build effective strategies for promotion of health equity. services in these two years (107.54% in 2015, 106.92% in 2019). Oaxaca decomposition revealed that factors such as residential location, registration, health insurance and time to the nearest health facilities, employment status and educational level, etc. made positive contributions to decline the inequalities. While factors pushed the equalities toward deterioration included health needs, economic status and household size. Conclusion: There were certain decline in the socioeconomic inequalities and inequities in health services utilization by hypertensive patients in Pearl River Delta Region of China by comparison of 2015 and 2019. Although the pro-rich inequities persists, it does suggest that government policies have improved health equity over time. .


Introduction
Inequality in health service utilization, it rises a constant concern of health system in the world [1], which means that the differences access to health service between social class, such that one group is better off than another [2][3][4]. When different demands have been adjusted, the existing inequality can be interpreted as inequity [4]. outpatient care due to hypertension in the last 2 weeks? (2) Have you visited a doctor for inpatient care due to hypertension in the last year?
The answers to questions were "yes" or "no".

Independent and control variables
In our study, age, gender and years of hypertension were classified as the need variables. Age was categorized into three year groups until 65 years and years of hypertension was also divided into three groups: 1-4, 5-9 and 10 or above.
There are economic status, educational level, employment status, marital status, household size, residential location, registration, health insurance and time to the nearest health facilities considered as socioeconomic variables (non-need variables). Economic status was constructed by using household income per household member, which was calculated by dividing the income for last year by household size, and then divided into five quintiles, which meant that the first and fifth quintile represent the poorest and richest wealth quintiles, respectively.
Educational level was categorized into four groups: primary or below education, middle school, high school, and college or above education.
Two employment categories and marital categories were employment and unemployment, married and single, respectively. Single includes unmarried, divorced, widowed and separated. Residential location and registration were divided into rural and urban, migrants and locals. Health insurance is based on whether the respondent is covered by the social health insurances.

Concentration index
The measurement of socioeconomic inequality in use of health service is based on a widely accepted index, appointed as the concentration index(CI). CI ranges from -1 to 1, with an index of 0 equivalents to perfect equality. A positive CI signifies that a health or health care variable is more concentrated among the richer population and vice versa [22]. The index can be calculated by employing the equation as follows: In equation (1), where y is the health variable (e.g. outpatient or inpatient service utilization in this study), is the fractional rank of the ℎ individual in the economic distribution, ranging from 0 to 1, is the mean of y.

Decomposition of inequality
In order to analyze the contribution of independent variables of the  (2): In equation (2) Then, CI for ycan be calculated as equation (3): In equation (3), where j  and k Z are the means of and ; and are the concentration indexes of and ; is the mean of y and GC is the generalized concentration index of ε. This equation reveals the total concentration index is made up of two components. One of that is the residual component ( ), which represents the inequality that is unexplained by the regressor. The other one is the explained component which is associated with two elements: (1) the impact of each determinant on health outcome, which is measured by its elasticity (ƞ= or ); (2) the extent of unequal distribution of each determinant across economic groups, which is measured by CI [24]. We then calculated the absolute contribution by multiplying the ƞ and CI with respect to that determinant. This value can be both positive and negative. We also calculated the percentage contribution of each regressor, which is equal to the absolute contribution divided by the total concentration index.
Notably, the positive and negative contributions may be offset in the aggregate and the sum of percentage contribution and error term is equal to 100%, therefore, the percentage contribution of several regressors may represent large positive and negative contributions, even exceeding 100%.

Horizontal inequity
Horizontal inequity (HI) is the concentration index that measures the need-standardized health service utilization. It reflects socioeconomic inequality in the use of health services after controlling for the impacts of biological needs, such as age and sex [25]. HI can be computed as follows: suggests that the health service is more concentrated among the richer groups and vice versa.

Decomposition change in inequality
At this stage, we used Oaxaca-type decomposition to determine the extent to which change in inequality in health service usage by hypertensive patients between 2015 and 2019 which was owing to change in inequality in the determinants [23,26,27]. The decomposition formula is as follows(equation (5)): In equation (5), where ƞ and ƞ All analyses were performed on STATA 14.0. Statistical significance level was set as 0.05.

Social demographic characteristics of respondents
The characteristics of the study population are displayed in Table 1. From 2015 to 2019, the utilization of inpatient service due to hypertension increased greatly, with the growth rate at 34.20%, whereas, a slight rise in the utilization of outpatient service, with a growth rate at 10.64%. Nearly half of the patients aged 65 or over. More than half of the patients were suffering from hypertension for less than 5 years. Most of the respondents finished middle school. Over half of the participants were unemployed and married, and the household size is no more than 4 people. Most of the respondents with hypertension were the urban and local residents. The health insurance coverage increased from 97.71% to 99.94% during this period. In terms of the access to care, it took most of them less than 15 min to get to the nearest health facilities.

Decomposition of the inequalities
A positive contribution to socioeconomic inequality means that the considered variable increases inequality. Table 2

Decomposition changes in inequalities between 2015 and 2019
As shown in Table 1, the CIs of outpatient and inpatient utilization by hypertensive patients reduced by 0.0300(20.03%) and 0.0334 (16.85%) from 2015 to 2019. Then, the reductions were decomposed to seek contributing factors following by Oaxaca-type decomposition. The results were presented in Table 4. The second and fourth columns show changes in the amount of inequality in determinants; and the third and seventh columns show changes in elasticities of determinants.
From the changed CIs and elasticities for utilization of health services, we found that the three major contributors to reduce the decrease inequalities including the educational level, registration and time to the nearest health facilities. What's more, changes in employment status, residential location, marital status and health insurance could explain the reduction of CIs to some extent.
However, the changes of economic status accounted for the biggest contributor for the pro-rich inequalities. In addition, the changes of household size and the need variables of age and years of hypertension also pushed these inequalities into deterioration, especially in the utilization of inpatient .

Discussion
Our study explored the socioeconomic inequalities and inequities in the health services utilization by hypertensive patients in Pearl River Delta region of China between 2015 and 2019, and further quantifies the contribution of selected factors toward the inequalities. In addition, we also assessed the changes in the inequalities during the survey-period.
The main findings were as follows: 1) obvious pro-rich inequalities and inequities in utilization of health services by hypertensive patients existed in Pearl River Delta region in both periods but they declined over time. 2) The changes in such inequalities were caused by the alteration in the interaction among the relevant determinants. This study has some limitations that must be mentioned. Firstly, recall biases could not be avoidable in questionnaire-based surveys, especially the self-reported utilization of health services. Secondly, the supply-side variables used in the decomposition only include the time accessibility of health facilities, but lack of other factors, such as the price of health services. Finally, since the decomposition analysis is a descriptive statistic, we were not able to carry out a causality analysis.
Despite the above limitations, this study has important policy implications for China towards reducing socioeconomic inequalities in health services utilization among patients suffered from NCDs.

Conclusion
Overall, the government's health strategies and policies have greatly