Background
Although the Chinese health care system has initiated several reforms designed to eliminate access barriers to health services, millions of people are still prevented from seeking and obtaining needed health care owing to financial difficulty [
1]. This is especially true in western China where the per capita income is far behind the eastern areas of China [
2]. On the other side of the coin, people accessing health services may cause households to have no choice but to pay a large proportion of household effective earnings, then pushing households into financial hardship or even poverty [
3,
4]. Catastrophic health care expenditure (CHE) is a general term used to describe all kinds of health expenditures that pose a threat to the financial capacity of a household in order to maintain its subsistence needs [
5‐
8]. The World Health Organization (WHO) suggested that CHE occurs if out-of -pocket payments (OOP) are at or exceed 40 % of income remaining after household subsistence needs have been met in any year [
9]. CHE which could occur in both rich and poor areas is not always a byword of large health care spending [
10]. Lacking social support, even a small amount of health spending may be catastrophic to poor households. CHE could force households and their family members to cut back on other consumption, sell assets, make an overdraft on future life, and even be trapped in a long-term debt [
11]. Considering the serious consequences of CHE, protecting households from CHE has been widely considered as a desirable objective of many health policies [
12‐
15].
In the past few decades, many policy interventions aimed at reducing OOP health expenditures have been carried out to protect households against CHE in China. One of the most important interventions taken by the Chinese government is to take concrete steps towards the achievement of universal health insurance coverage. Three social health insurance schemes have been implemented in China: the Urban Employee Basic Medical Insurance (UEBMI) designed for the employed urban residents, the Urban Resident Basic Medical Insurance (URBMI) designed for urban residents without formal employment, and the New Rural Cooperative Medical Insurance (NRCMI) designed for rural residents [
16,
17]. Take the NRCMI, which was piloted in 2003 and comprehensively implemented in 2008, as an example, by 2013 over 800 million rural residents have participated in NRCMI in rural China. Per capita financing for the NRCMI amounts to around 340 yuan (RMB) (US$ 54.84, all US$ equivalents presented in this paper are calculated based on World Bank annual average exchange rate US$ 1 = RMB 6.20 in 2013), of which 280 yuan (82 %) is subsidised by central and local governments [
18]. The actual reimbursement ratio, however, is still at a relatively low level [
19]. Other interventions, such as promoting free health treatments to a targeted population, have also been carried out since 2009. By the year 2011, over 39 million elderly people over 65 years old had received health check-ups for free, and over 13 million pregnant and maternal women in the rural areas had received maternity allowance [
20].
Despite the fact that “average” catastrophic health spending could be reduced with policy interventions, inequalities in CHE will not simply be eliminated and inevitably exist across households due to geographic and economic factors. Therefore, measurement and regular monitoring of inequalities in CHE are of great concern for policy-decisions, epidemiologists and other health scientists [
21].
Previously studies mostly focused on incidence and intensity of CHE and its determinants based upon one-wave cross-sectional survey data [
4,
22]. So far, no study has been published to investigate the change of income-related inequality in CHE in western China along with the recent health insurance reform. Considering the unique social structure and economic profiles in local areas, the contributions to inequality in CHE may vary widely between provinces and countries.
The purpose of this study is to measure the overall proportion of households incurring CHE in 2008 and 2013 in Shaanxi rural areas, to compare the change of income-related inequality in CHE between two stages, and to analyse the contributions of determinants to socioeconomic inequality in 2013. The findings of this study will provide readers vital information about the severity of CHE and its inequality in Shaanxi Province. The results will also shed lights on policy suggestions regarding how to reduce the incidence of CHE and its inequality in developing countries.
Discussion
In the study, we estimated the overall proportion of households incurred CHE between 2008 and 2013 in Shaanxi Province with a conservative method. Firstly, we used a high threshold to define the occurrence of CHE. Secondly, indirect expenditure for seeking health services, such as transportation, food, accommodation, lost earnings due to illness, was not included in health expenditure [
9]. This conservative estimation method may lead to underestimating the financial consequences of household health expenditures [
9,
36]. Even though a conservative method was used to measure CHE, the proportion of households incurring CHE in Shaanxi Province was still considerably high in both years. This proportion was not only higher than that in most rural areas of other provinces in China, but also higher than that in most developing countries [
4,
9,
37,
38]. Several potential factors could be responsible for this phenomenon. Firstly, despite the fact that Shaanxi Province has achieved close to universal health insurance coverage, actual reimbursement ratio for NRCMI designed for rural residents is still at a low level. A high proportion of OOP in the total health expenditure remains a cause of households facing CHE. Secondly, health service access and use is also a very important determinant resulting in CHE. Economic growth generates higher demand for expanded health care provisions. In Shaanxi Province, the outpatient visits ratios within two weeks rose from 11.18 % in 2008 to 12.55 % in 2013, and the annual hospitalisation rate jumped from 5.39 % in 2008 to 10.06 % in 2013 (Yongjian Xu, Yanli Li, Sha Lai, unpublished observations). Increasing usage of health services and the lag of social institution’s development have a potential effect of putting households into economic catastrophe [
9]. Thirdly, the long-term split of basic health insurance schemes designed for different target populations—urban and rural, formal and informal—and the financial risk management at county or city-level, lead to the deficiencies of the risk pooling capabilities. Risk pooling shares medical costs with different profiles to prevent households from catastrophic expenditure due to unexpected diseases, and enables cross-subsidies from the advantaged to the disadvantaged [
39]. A fourth issue contributing to CHE is that the continuing shrinkage of household size reduces the household’s anti-risk ability to cope with catastrophic payments.
This study emphasised some key factors as determinants of catastrophic health care expenditure. Most were similarly reported in related studies [
40,
41]. As we expected, a lower economic status played an important role in increasing the risk of incurring CHE in 2013. However our study found that poorest households showed no statistically significant association with CHE in 2008. One possible explanation is that the poorest households forgo their needed health care due to high health care–related cost. Our unpublished analyses, using the same data, found that in 2008, 36.08 % members in poorest households refused inpatient treatment in the past year, whilst 53.50 % members in these households suffering from illness and injury in the past two weeks chose not to seek medical treatment. Consistent with most previous studies in China, our findings showed that absence of social health insurance increased the risk of households incurring CHE in the rural areas of Shaanxi Province [
41]. In addition to household economic status and absence of social health insurance, having elderly members, inpatient and outpatient health services usage, and lack of commercial health insurance were also key determinants of CHE. From 2008 to 2013, the percentage of households having a male head dropped from 82.71 % to 78.36 %, and our study found that the male household head increased the risk of households incurring CHE in 2013, but there was no significant association in 2008. Further studies are needed to explore the reasons. Unlike previous studies conducted in other countries, our study showed that there was no statistical association between CHE and households with a child or children below 5 years old in both years [
40]. One of the potential reasons is that an expanded program of immunisation for children, promoted in Shaanxi Province, has successfully prevented and controlled outbreak and spread of many childhood infectious diseases which may make households face large health expenditures [
42]. A small household with more inpatient treatments, having elderly members, lack of health insurance and illiteracy of the household head, had the higher risk of incurring CHE. Therefore, policy interventions aimed at reducing the probability of household incurring CHE should primarily consider the needs of vulnerable households.
The issue of CHE will not simply be solved with an increasing trend of income. Local health systems should be improved in several aspects to protect households from CHE. One key approach reducing economic catastrophe is to establish pre-payment mechanisms and move away from post-payment mechanisms in basic health insurance systems [
10]. Another approach is to enhance the actual reimbursement ratio of basic insurance (or reduce OOP payment ratios). A third, developing financial risk pooling should be placed on the agenda. In the present circumstances, an incremental approach, which starts with different pools designed for target populations and combines them over time, is a feasible way to risk pooling used by policy-decisions [
39].
From 2008 to 2013, local health systems were mainly dedicated to protecting households against catastrophic payments and eliminating access barriers to needed health care by carrying forward universal health coverage as well as other measures. However, it is worth noting that the inequality in facing CHE significantly increased over these years. Thus, further policy interventions should be taken to reform health systems to address the remaining inequality.
Our study revealed that some determinants, such as absence of commercial health insurance and households having elderly members, showed positive contributions to socioeconomic inequality in CHE, which meant CHE was greater among the poor. Furthermore, our study also indicated that households with chronic disease members had negative contributions, favouring the poor. A potential explanation is that most chronic diseases, such as hypertension, diabetes and coronary artery disease, is more prevalent among the advantaged and associated with a high risk of incurring CHE. Consistent with earlier studies, our study found that the contributions of health service usage variables to the probability of incurring CHE were in a pro-poor direction, reducing socioeconomic inequality in 2013 [
36,
43]. That is because the rich households are relatively more prone to use these services and accessing such services has a higher risk of incurring CHE, while conversely, poor households are less likely to be affected with little use of health services. From the decomposition analysis of concentration index, it is evident that most of the inequality in CHE is explained by household size and household economic status in 2013. Therefore, potential solutions to relieve socioeconomic inequalities should concern the degree of determinants’ contribution, and it is helpful to reduce inequalities in CHE by narrowing the gap of household economic status and improving anti-risk capability of small scale households.
There are some limitations in the study. Firstly, all of the demographic, socioeconomic and health services usage information were self-reported. Household health expenditure and health services usage data may be less accurate than medical records. Secondly, the presence of missing data may lead to biased estimates of regression parameters and weaken generalisability of our results. However, considering less than 1 % data were excluded owing to the missing value, this problem may not be serious. Thirdly, the way we calculated CHE has some limitations. Poor households choosing not to seek health care were excluded from CHE calculations. Also, since the economic status was measured by annual household expenditure, poor households spending catastrophic expenditure on health services increased their CTP and total expenditure, thus these households were categorised into a higher economic status in analyses. Fourthly, there was likely to be recall bias because of a long recall period.
Conclusions
Even though we used a conservative method to measure CHE, the overall proportion of households incurring CHE is still considerably high compared to other rural areas of China and most developing countries in the world. Furthermore, there exists a strong pro-rich inequality of CHE in the rural areas of Shaanxi Province. Many determinants, such as absence of health insurance, having elderly members and lower economic status, increased the probability of incurring CHE. Lower economic status and small household size were the main determinants contributing to inequality favouring the rich. This study suggests that in Shaanxi Province, which has almost achieved universal health insurance coverage, narrowing the gap of household economic status, establishing prepayment mechanisms in health insurance, strengthening the depth of reimbursement (reducing OOP ratio) and subsidising vulnerable households are helpful for both reducing the probability of incurring CHE and pro-rich inequality in Shaanxi Province.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.
The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
YX participated in the study design, data analysis and interpretation, and was the primary person responsible for drafting the manuscript. JG contributed to study design, data analysis and reviews. ZZ conceived of the study and participated in data analysis, writing and revision. QX and GC contributed to data analysis and revision. JY and HL participated in drafting the manuscript. YL participated in data management and data analysis. SL participated in revision. All authors read and approved the final manuscript.