Introduction
Since the collapse of the old Cooperative Medical Scheme in China after the economic reforms of the early 1980s, rural residents have been excluded from the public social security system[
1]. Financial barriers, among other factors, have become the most important contributor to impede the poor in trying to access medical services[
1]. In order to address this problem, the Chinese government initiated New Cooperative Medical Scheme (NCMS) in 2003 to reduce the financial burden of rural residents and to achieve universal coverage.
The NCMS is organized, guided and supported by the central government but has voluntary involvement[
2,
3]. Unlike its predecessor (the old Cooperative Medical Scheme), it operates at county rather than village level and variations existed in design and implementation across counties[
3]. The central government takes the overall responsibility to manage and supervise the scheme while the policy implementation responsibilities are decentralized to county level governments[
2]. Specifically, the central and provincial governments designed the essential or basic rules about implementation of NCMS, such as the minimum level of NCMS premium and policy reimbursement rate, the priority for the reimbursement of essential drugs or Chinese traditional medicine, etc., while the county governments are responsible for specific operations, such as defining benefit packages, designating participating providers, pooling risk across the local rural population and experimenting health policy innovations like payment reform. Under this context, the benefit package was usually the same for the participants in one county or province while it may differentiate across counties or provinces, which was closely related to the varied financing levels of NCMS in different areas.
The NCMS is heavily subsidized by central, provincial and county governments and also partly financed privately from individual farmers[
3]. Coverage of inpatient care is a reimbursement priority in the NCMS but also a relatively slight compensation for outpatient care, which depends on specific benefit packages in different counties. By now the scheme is also extended to cover other catastrophic diseases, such as chronic diseases, leukemia, cancer, etc. By 2011, 97.5% (around 832 million) of the rural population have been covered by the NCMS in China, meanwhile, the total NCMS revenues per capita increased from 30 RMB in 2003 to 250 RMB in 2011 (equivalent to 194 RMB in 2003 year’s price[
4]), as the subsidies from governments in central, provincial and county levels rose from 20 RMB per enrollee in 2003 to 200 RMB in 2011 (equivalent to 155 RMB in 2003 year’s price[
4]) (Ministry of Health, China). The rapid expansion of the NCMS inevitably raises challenging issues like any other health insurance systems as escalating healthcare costs, health care quality and the equity issue.
The socioeconomic equality in healthcare is one of the most important issues of concern in both developed and developing countries. For a specific health insurance system, it means that all enrollees should have equal access to utilize medical services, get equal reimbursement benefits and finally afford equally proportional self-payment, irrespective of their socioeconomic status, especially not dependent on the financial status[
5,
6]. In reality, the poor, who frequently are in need of more services, are often the least able to pay, while the wealthy utilize disproportionately more services although they have less need[
7,
8]. Moreover, it is also a disadvantageous factor for the poor enrollees that all individual farmers, regardless of their economic status, would pay the same contribution to be enrolled. Considering its rapid expansion and flat-rate personal contribution, it is necessary and meaningful to analyze and discuss the income related inequality situation of the utilization of medical service, reimbursement and self-payments in the current context of NCMS and more importantly, to examine the inequality changes during the evolution of NCMS.
Some studies have demonstrated the inequality status of varied aspects in the NCMS, such as benefit rate, reimbursement, medical service utilization and self-payments by enrollees[
2,
3,
9‐
14]. However, contradicting results were shown across the studies and limited evidence focused on the income related inequality in outpatient care in the NCMS by using concentration index. Nevertheless, relatively fewer studies could consider the inequality changes in the NCMS by time-series data while most only concerned one or two years between 2003 and 2009. Actually, due to the new health care reform in 2009, the Chinese government stimulated more funding to be invested in the NCMS, which greatly improved its financing and reimbursement capacity. As a result, the inequality status in the current context of NCMS was unknown and rigorous empirical studies with a longer time span are strongly needed.
The aim of this study was to analyze the degree and consecutive changes of income related inequalities for both inpatient and outpatient care in the NCMS from 2007 to 2011 by using five years’ continuous data in Junan County in Shandong province, China.
The paper is organized as follows. The next section presents a comprehensive review of the most relevant literature in this field. The subsequent sections include: the description of data and methods used; the illustration of the results obtained; and the last section offers the discussions of the key findings and principle conclusions.
Literature review
A review of the present literature concerning studies of NCMS equality could be divided into two types; one is to analyze the income related inequality among NCMS enrollees mainly by concentration index[
10‐
14] while the others are comparing equality status between NCMS-members and non-members[
2,
3,
9].
Regarding the first type of studies, where the equality situation among members of NCMS has been demonstrated, we mainly reviewed the studies with relatively more reliable data, larger sample and by using concentration index, the most frequently adopted method in the equality analysis. First, we concentrated on the studies using one-year data (cross-sectional studies). Considering the income related equality status in the inpatient care, the study in Mei County Shanxi Province in 2009 revealed a distinct pro-rich inequality in both inpatient benefit rate (the number of enrollees got reimbursed from the NCMS irrespective of the amount) and NCMS inpatient reimbursement[
10]. Another study with data sources from three cities (Wuxi, Shengde and Chishui) in 2009 showed conflicting result that more and more inpatient reimbursement was actually concentrated on the poorer enrollees by using household data[
13]. It also presented a clearly pro-rich inequality in total inpatient expenses, length of stay and self-payments[
13]. The study in Yunnan Province in 2006 supported that total inpatient expenses and length of stay were concentrated more on the rich[
14], in the meanwhile, it was worth paying attention in the same Yunnan study that the equality status of inpatient care utilization in the NCMS was much better than those counties without NCMS[
14]. Limited evidence has been found to focus on the income related inequality in the outpatient care in the NCMS by using concentration index. The Yunnan study showed pro-rich inequality in both the number of visits and total outpatient expenses, and moreover, the degree of equality of these two indicators in the NCMS was worse than patients in the non-NCMS counties[
14]. Contradictory equality result was also shown by another study that the utilization of outpatient services was concentrated more on the poorer participants, and it also revealed the pro-poor inequality in outpatient reimbursement and self-payments by using the data from three cities (Wuxi, Shengde and Chishui)[
13]. Second, relatively fewer evidence could be found to reveal the equality changes in the NCMS by time-series data. The study in Jiangxi Province showed the inequality of inpatient benefit rate changed from pro-rich in 2006 to pro-poor in 2008 while the equality in the NCMS reimbursement remained relatively stable with smaller positive values of concentration index around 0.04 in all the three years by using NCMS reimbursement claim data of 40 counties in this Province[
11]. In another study in Guangdong Province, the reduction of pro-rich inequality in the inpatient reimbursement among NCMS enrollees was presented by analyzing two years’ (2006–2007) NCMS reimbursement data, in the meantime, the study also showed the inequality of outpatient reimbursement changed from pro-rich in 2006 to pro-poor in 2007[
12].
In addition, three studies were found that focused on the impact of NCMS on income related inequality by comparisons between NCMS members and non-members. Particularly, by using both household survey data and routine health facility data from 15 counties in 12 provinces in 2003 and 2005, a vigorous study[
3] indicated that the poor experienced larger increase in outpatient care and the rich experienced a larger increase in the inpatient care. Another two studies also supported more NCMS members from the high income group used impatient services than non-members by using data from six counties in Shandong and Ningxia provinces in 2006[
2,
9].
In summary, more evidence supported the pro-rich inequality in the inpatient care while it is difficult to conclude the equality status in the outpatient care given that relatively rare studies could be found. However, we need to be cautious towards the findings revealed in these studies considering the limitations. First, the data sources used among the above studies, including both household survey and NCMS routine reimbursement data, were between 2003 and 2009 and no recent studies after 2009 related with the equality issue in the NCMS could be found according to our knowledge. Considering the development of NCMS, however, the policy priority in this study period from 2003 to 2008 (the initiation year varied among counties) was to achieve universal coverage through rapid expansion but also lower contributions. The NCMS policy priority has already changed since the new health care reform was implemented in 2009 when the Chinese government stimulated more funding to be invested in the NCMS, which greatly improved the financing and reimbursement abilities. Hence, it is necessary and more meaningful to explore the degree of income related inequality in the NCMS after 2009, more importantly, not only the equality extent at a certain time point but the consecutive changes of this inequality during a longer study period to see its evolution. Second, there were still lack of scientific evidence concerning with the inequality situation of self-payments and medical service utilization in both inpatient and outpatient care.
Given the research gap mentioned above, this paper aims to go a step further. The goal is to show the inequality status and its consecutive changes for both inpatient and outpatient care in the NCMS from 2007 to 2011 by using five years’ continuous data in Junan County in Shandong province, China. It hopes the findings could shed some light on the further evolution of NCMS and on similar health insurance systems in other developing countries.
Discussion
The NCMS have been playing a prominent role in the financial protection of rural residents by now. Over the past decade, undoubtedly, it made impressive advances towards universal health coverage with a stable participation remaining at a high level. Accompanied by huge increases in financing and reimbursement ability especially after the new health care reform in 2009, the mean values of pertinent inpatient and outpatient indicators grew rapidly as a whole, however, in the meantime, the income related equality status in certain aspects was not equalized, especially not for the inpatient care, as revealed in this study. Four key findings are worth considering further in depth.
First, the inpatient benefit rate has seen larger increases for all income groups while slight increase or even decrease was found in the outpatient benefit rate during the study period in the NCMS. The different policy reimbursement rates in the NCMS from 2007 to 2011 greatly contributed to this result. Taking the sample county Junan for example, inpatient reimbursement was priority of the government, in each year, around 70% of total NCMS contributions were allocated to the inpatient reimbursement funding while the outpatient funding only accounted for 30%. Therefore, the policy reimbursement rate for inpatient care was much higher than that in the outpatient care and moreover, from 2007 to 2011, the inpatient reimbursement rate at the town level health providers rose from 40% to 90% for the expenditure group (300–3000 RMB) while it only increased from 20% to 35% for outpatient services (irrespective of total expenditure) in the same health institutions during the same period. Consequently, more and more enrollees were stimulated to utilize inpatient services to substitute outpatient services to get higher reimbursement from NCMS, which could probably explain that the inpatient benefit rate increased dramatically from 2007 to 2011 and in the meantime no obvious changes for the outpatient benefit rate was observed. The substitution between inpatient and outpatient services was also found by Zhou[
19] using the data of 2003 and 2008 national health services survey. Although his study was done in the rural area and did not target NCMS enrollees only, it is still supportive since NCMS has been the largest health insurance system in rural China covering over 95% of the rural residents by 2008. The extent and induced results of this substitution need to be carefully studied in the future, especially for the influence on medical expenses.
Second, along with the overall increase in the utilization of inpatient services, however, the distinct pro-rich inequality in the inpatient care from 2007 to 2011 was revealed in this paper, which means the poorer enrollees would not get equal benefits from the big progress achieved by NCMS. In contrast, the inequality of outpatient services was always related to pro-poor or remained relatively stable around equality line. More evidence supported the rich enrollees utilized more inpatient services than the poor[
10,
12,
13]. Given the target of the whole rural area of China, strong pro-rich inequity of inpatient utilization still remained and income was the principal determinant of this inequality[
7]. The most possible explanations for the pro-rich inequality in the inpatient care are related with NCMS reimbursement deductible and much higher medical price for the inpatient service, both of which hindered the poorer enrollees to utilize more inpatient services compared with the richer ones. In Junan, the deductible existed only in the inpatient reimbursement and it would become higher with advanced health institutions. In 2011, the deductible for town level health providers was 150 RMB while 500 RMB for county level hospitals, with the meaning that at least 150 RMB at town level and 500 RMB at county level need to be paid by enrollees themselves first before receiving NCMS reimbursement. In the meanwhile, the higher medical price of inpatient service played a vital role. In Junan, the average expenditure per admission has reached 4171 RMB by 2011 while the average expenditure per outpatient visit was only 41 RMB at the same year. Although NCMS has already made great progress in the inpatient reimbursement gradually, including the improvements of reimbursement rate and ceilings as shown in Table
2, the actual co-payment rate was still very high for the poorer enrollees, with 72.4% in 2007 and improved to 60.9% in 2011. Compared with the richer enrollees, the poorer with limited financial ability had relatively higher price elasticity and consequently, more sensitive to medical prices, which possibly lead them not to seek expensive inpatient services when the diseases are not so severe to threaten their lives in general. In contrast, the inpatient reimbursement deductible and medical price were relatively much easier for the richer to afford. NCMS released their medical demanding further and stimulated them to utilize more inpatient services to get higher reimbursement. Besides, no deductible existed for the outpatient reimbursement and the majority of the drugs prescribed by town and village level health providers were covered by NCMS reimbursement list, as a result, the poorer preferred to seek outpatient services instead.
In addition, transportation costs could also be an adverse factor for the poorer enrollees to utilize inpatient care[
20,
21] and stimulated them to turn to outpatient services. Enrollees could access outpatient care relatively more easily because village clinics providing outpatient care is always within walking distance. In contrary, NCMS enrollees need to go to at least the township health centers to be hospitalized, which are usually located in the center of town and usually only one in each town. The transportation costs would be much higher for utilizing the inpatient care in county or county-above health providers. Compared with outpatient care, geographical access for inpatient care would be lower.
All of them contributed to the pro-rich inequality for inpatient service utilization and more pro-poor inequality for outpatient service utilization. Since the reimbursement from NCMS is closely related with the volume and types of medical service utilization, consequently, the richer enrollees usually got higher inpatient reimbursement and also bore larger OOP under such circumstances.
Third, regarding outpatient care, it seems contradicting to find that the inequality of outpatient medical expenses was kind of pro-rich although it was pro-poor for outpatient visits and reimbursement, which was relatively distinct from 2007 to 2009. To some extent, it indirectly reflected that the rich would like to seek outpatient services from the health providers at higher level (town or county) while the poor probably inclined to visit village clinics more frequently, the lowest level of health system in China. Generally, the quality of medical care in China was closely related with the level of health providers, ie., the quality in county hospitals was always regarded as the best subsequent by the township health centers while the quality in village clinics was usually considered as the most disadvantages. Mostly, the most important function for village clinics was considered as selling drugs not giving treatment to the patients while patients could get better examinations at township health centers and county hospitals due to more skilled health personnel and advanced equipment, especially at county hospitals. The medical expenses went up higher at town and county hospitals under the same health needs compared with village clinics. As a result, although the poorer visited more frequently at village clinics, the inequality of medical expenses was pro-rich. The variation in the level of health providers the NCMS enrollees sought became the most possible explanations for the contradictory. Additionally, the location of enrollees also played a vital role as the enrollee living near the town level hospital would probably visit it for convenience instead of the village clinic.
The two aspects also can be used to explain the inconsistent changes between outpatient reimbursement and outpatient expenses in the same period. In the Junan NCMS, the ceiling for outpatient reimbursement is relatively low, as described in Table
2, and there is no reimbursement for visits at county hospitals. It would be much easier for the enrollee to reach the ceiling when seeking medical services at higher level hospitals. Thereafter, the enrollee could not be reimbursed by the NCMS for next outpatient visits that’s why the inequality of outpatient reimbursement and outpatient expenses were not inconsistent sometimes. In this study, since the outpatient visits reimbursed were merged together for each enrollee in each year, we could not quantify the influence of level of health providers and geographic access here, reminding us the further study is needed to demonstrate the inequality situation of types of health providers among NCMS enrollees.
Lastly and more recently, the development of income related inequality in the NCMS has shown a reduction in the length of stay, total inpatient expenses and inpatient reimbursement from 2010 to 2011 after its increase from 2007 to 2009, although there is still pro-rich advantage. In addition, the overall inequality trend of outpatient reimbursement changed from pro-rich in 2007 to pro-poor in 2011 with some fluctuations during the study period. Both were favorable for the poorer enrollees as they started to enjoy more equal benefits of NCMS gradually in line with the health care reform and also implied the financial access of poorer enrollees have been improved step by step. The changes were actually consistent with policy priorities of NCMS. In the initial stage, the government concentrated on rapid expansion of NCMS coverage which was necessary for a government-oriented insurance system. Under the circumstances of high copayments of inpatient utilization and limited financial ability, seeking outpatient care, instead of inpatient, became the first choice of poorer participants and a very limited demanding got released in the hospitalized services. In the meantime, the richer enrollees per se have the ability to afford the disease burden so the implementation of NCMS improved their access further. Both stimulated the increasing pro-rich inequality from 2007 to 2009. After 2009, with the implementation of health care reform, special targets and actions for NCMS were realized by improving the coverage in depth with higher reimbursement and wider benefit package, the average contribution per capita reaching 250 RMB totally in 2011 compared with 30 RMB in 2003. These policy adjustments greatly improved the policy reimbursement rate and decreased the actual co-payment rate for inpatient services at the same time. Compared with previous stage, more and more medical demands from the poorer enrollees could be released. A problem is that no later evidence or data after 2010 could be found to compare with the changing trend indicated in this study.
Regarding the pro-rich inequality in the inpatient care, which already showed reduction in the recent two years as discussed above, how to improve the inpatient service utilization among poorer enrollees and reduce current pro-rich inequality situation further are still challenging issues for the policy makers in the following stage of NCMS development. Reforming the flat-rate personal financing systems, widening the benefit package and reducing cost sharing and deductibles[
21] could be probably prioritized in the NCMS policy agenda.
Strengths and limitations
The strengths of this study could be summarized into four main aspects. In the first place, the study sample in this paper targeted the whole NCMS enrollees in one county, representing over 97% of rural residents for all five years (2007–2011) when analyzing the income related inequalities. This large sample could help make a comprehensive inequality result. Secondly, we provided the latest evidence and the continuous changing trends of income related inequality in the NCMS from 2007 to 2011, which were lacked by other studies but strongly required in practice. Thirdly, compared with other studies, more comprehensive indicators covering benefit rate, medical service utilization, NCMS reimbursement and self-payments in both inpatient and outpatient care were analyzed in this study, which could help policy makers and researchers get a better understanding of the equality status in the NCMS of China since fewer studies were found internationally published. Lastly, in this study we extracted the accurate individual utilization, reimbursement and out of pocket expenditures data from information system. Compared with other studies using household survey[
9,
10,
13,
14], the data here could be regarded as a much better accuracy because no recall bias or unconscious intension maximizing medical expenditures but minimizing reimbursement from respondents existed, which greatly contributed to a more objective conclusion. In addition, all of the data here were after age-sex standardized to avoid the confounding effects of other demographic factors besides income as much as possible, which were not mentioned in the domestic studies[
11‐
14,
22]. All of the strengths contributed to enrich our knowledge of income related inequality in NCMS.
However, the study is not without limitations. First, as we mentioned above, annual net income per rural resident at the town level was used as a proxy income variable, which would possibly underestimate the inequality in this study although the income difference within the same town for such a poor county is not so distinct. Referring to employee-based health insurance, the contribution of employees is also a useful surrogate index for actual household income because it is calculated on the income, property and private auto taxes of the employee[
23,
24]. For NCMS, the contribution of enrollees is flat so town-level income is finally adopted, but still income data at individual level would be preferred in the analysis of income related inequality if possible. Second, the information system included the utilizations finally got reimbursed in NCMS not all utilization information. The most possible reasons for the cases without NCMS reimbursement are that the enrollees may go to county hospitals for outpatient care or their inpatient expenses during the admission may not reach the reimbursement deductible. According to the NCMS principles, the outpatient expenses could not get reimbursed at county level and only the inpatient expenses exceeding the deductible (under ceiling) could be reimbursed. Under this condition, it could possibly make the study overestimate the average inpatient expenses and underestimate the average outpatient expenses while underestimate OOP because the excluded ones were completely paid by enrollee themselves. In addition, if the enrollee went to the non-NCMS health providers due to certain considerations (such as convenience or geographic access, etc.) when seeking medical services. They were also not reimbursed by NCMS. But the proportion of such cases should be very small due to the high coverage of NCMS health providers in the rural areas. Third, approximately 1.4% of total enrollees, 0.2% of total outpatients and 3.7% of total admissions in the five years were omitted because of lack of accurate or complete information according to the pre-determined exclusion criteria mentioned above, which was necessary for the study but also may result in certain subjective bias. In 2010, particularly, lack of admitted date or discharged date took account of 81.3% of its total missing cases. Most were hospitalizations outside the county. According to the interview of local officers, we learned that they updated their electronic information system in 2010, resulting in some merging problems which were the main reason for these missing cases. Besides, the medical service utilization and reimbursement of non-members were not recorded in the NCMS information system, so we only targeted NCMS members and did not compare the results with non-members in this study. Finally, although the paper has already standardized the age-sex distribution of NCMS enrollees to reduce the cofounding effects when analyzing income-related inequality, other important confounding factors, such as level of health providers, different deductibles and ceilings, or geographic access, also need to be considered. However, the concerning indicators were calculated for per enrollee but not per admission or visit since the goal was to analyze the inequality status from the perspective of NCMS enrollees in five years’ period. As a result, the value of health indicators was the sum of all the reimbursed visits and/or admissions occurred in each year for per enrollee. For enrollees seeking medical services more than once, the level of health providers may be different which made other confounding factors related to the choice of health providers very difficult to control. Under this context, we finally adjusted all the health indicators by age and sex.
Competing interest
All authors declare that they have no competing interests.
Authors’ contributions
SY collected, processed and analyzed the data, wrote and finalized the manuscript. CR provided guidance and advice on the analysis framework and also detailed comments on the manuscript. XS and XL actively negotiated with local officers and participated in the data collection and paper modification. QM directed the data collection, oversaw the interpretation of data and revised the paper. All authors read and approved the final manuscript.