【作者】 郭娜；

【导师】 王健； Tor Iversen；

【作者基本信息】 山东大学 ， 社会医学与卫生事业管理， 2014， 博士

【摘要】 研究背景我国新型农村合作医疗制度(以下简称新农合)从2003年开始试点,试点工作启动以来,在各级政府的支持下发展迅速,进展顺利。随着新农合的逐步开展,新农合政策也在不断调整和变化,2008年实现了基本覆盖后,2009年开始实行新医改,虽然新农合的实施效果得到了一定程度的认可,但是国内外仍然存在一些不同的声音。其中一个主要的不同意见就是医疗保障覆盖虽然扩大了,但医疗费用上升快。住院服务利用会产生较高的医疗费用,在一定程度上使患者家庭致贫,尽管2009年实际住院补偿比例达到了41.5%,但是不断提高的新农合覆盖率并没有降低疾病经济风险的影响。目前存在很多关于新农合的研究,从内容上看,部分文献关注对新农合体系的描述,部分研究关注对新农合实施效果的评价；关于分析方法,多数文献采用描述性分析的方法进行研究,也有一些文献采用更为严格的方法评价新农合实施效果,如倾向得分匹配法(PSM)、Logistic回归、倍差法(DID)等,在对公平性的分析上,多数文献采用集中指数(CI)、洛伦兹曲线、不平等斜率指数(SII)等方法,在运用Oaxaca分解法分析公平性纵向变化方面的相关研究不多；从研究时间段来看,缺乏对新农合实施后广覆盖阶段,新医改实施前后的几个关键时间段的系统分析；关于变量的纳入,新农合相关变量多为是否参合,也有文献将新农合政策纳入模型,但缺乏对补偿比、起付线和封顶线变化的系统分析；从数据类型上看,多数研究采用横截面数据或者混合横截面数据进行分析,也有少数采用家庭层面,村庄层面的面板数据进行分析,在利用相同个体层面的面板数据分析方面仍然存在着不足。2006年、2008年与2011年,本研究所调查的三县、区(章丘市、长清区、平阴县)住院费用补偿比例均呈现上升的趋势,并且补偿比例随着医院级别的升高而降低；住院封顶线也呈现大幅度上升的趋势；平阴县和长清区2006年和2008年均没有设置起付线,2011年各医疗机构均设置了起付线,章丘市2011年降低了县级和市级住院起付线,乡镇卫生院设置了起付线。从2008年之前的广覆盖阶段,再到2011年新医改实施后的阶段,新农合政策中住院补偿比、封顶线的提高以及起付线的设置是否增加了住院服务可及性?是否改善了公平性?这是本研究主要关注的问题。本研究采用个体水平(即相同个体)的面板数据,以健康需求理论为基础,从保险的收入效应、价格效应和风险规避作用出发,分析新农合补偿政策中补偿比、封顶线的提高以及起付线的设置对住院服务的可及性(包括可及性的总体水平和可负担性)以及公平性的影响,为完善新农合政策,提高住院服务的可及性及其公平性提供依据。研究目的本研究的总体目标是分析新农合补偿政策变化中补偿比、封顶线的提高和起付线的设置对住院服务可及性及其公平性的影响,揭示其相关影响因素,为完善新农合制度,提高可及性水平,降低不公平性提供政策参考。具体目的包括：研究新农合补偿政策中补偿比、封顶线的提高和起付线的设置对住院服务可及性的影响；分析新农合补偿政策中补偿比、封顶线的提高和起付线的设置对公平性的影响,明确新农合补偿政策对公平性的贡献；分析公平性的纵向变化趋势以及新农合补偿政策的变化对公平性纵向变化的贡献,进一步揭示公平性的变化是归因于弹性的贡献还是集中指数的贡献,为提高住院服务的可及性及其公平性提供政策完善的依据。研究方法本研究利用济南市三县(市、区)章丘市、长清区、平阴县三年家庭入户调查资料共2600户,9020人,主要研究调查时间上一年(2006年、2008年和2011年)住院服务情况。本研究以2008年的调查数据为基础,选取三年个体水平上的非均衡面板数据为研究样本,确保三年相同个体至少调查两次。本研究采用的分析方法主要有描述性统计分析,单因素统计推断和多因素统计推断。多因素统计推断包括：采用二元选择面板数据随机效应模型分析新农合补偿政策对住院服务利用的影响；采用随机效应负二项分布模型分析新农合补偿政策对住院次数的影响；采用广义线性模型分析新农合对住院自付费用和补偿费用的影响；采用集中指数衡量公平性,利用集中指数分解法分析新农合补偿政策变量对公平性的贡献；利用Oaxaca分解法分析公平性的时间变化,并将时间变化分解为集中指数的变化和弹性的变化。研究结果(1)本研究发现2006年、2008年和2011年住院率(以人次数计)分别为4.77%,4.40%和4.79%,差异没有统计学意义。新农合补偿水平与住院服务利用和住院次数成正比；县级医院设置起付线、乡镇卫生院没有起付线相对于均不设置起付线能够降低就医次数；封顶线对住院服务利用和住院次数的影响没有统计学意义；经济状况能够增加住院服务利用和住院次数。三年住院服务利用的集中指数分别为0.255,0.326和0.244,说明三年高收入人群利用住院服务高于低收入人群。三年新农合补偿比的提高促使高收入人群获得更多的住院服务,但是贡献较小。2006年与2008年乡镇卫生院没有起付线,县医院有起付线,降低了住院服务利用的不公平性,相对增加了低收入人群的住院服务利用。2006年至2008年集中指数变化为0.071,2008年至2011年集中指数变化为-0.082。新农合政策中的补偿比和起付线的贡献能够导致2006年至2008年住院服务利用向高收入人群方向转移。2008年至2011年新农合政策中补偿比的贡献与总变化的方向相反,说明补偿比不是导致这期间住院服务利用向有利于低收入人群的方向发展的原因,主要原因是经济状况的改善。(2)三年住院实际补偿比分别为12.15%,12.46%,33.40%。名义补偿比和实际补偿比与住院自付费用呈反比,与补偿费用呈正比,设置起付线能够增加患者的自付费用。住院天数两周及两周以上、住院机构选择县级及县级以上医院能够增加自付费用和补偿费用。三年受益程度的集中指数分别为0.035,0.020,0.033,说明高收入人群获得较多的新农合补偿。2006年和2008年住院实际补偿比对集中指数的贡献为负,2011年补偿比和封顶线的贡献为正。2006年至2008年集中指数变化为-0.015,2008至2011年集中指数变化为0.013。新农合政策补偿比,起付线和封顶线的贡献能够使2006年至2008年新农合受益程度向有利于低收入人群的方向发展,占总变化的11.33%。新农合政策中补偿比和封顶线的提高促使2008年至2011年新农合受益程度向有利于高收入人群的方向发展,分别占两年集中指数变化的32%和17%。(3)三年新农合补偿后住院导致的灾难性卫生支出发生率、平均差距和相对差距明显低于补偿前,呈现下降的趋势,下降率基本呈现上升的趋势,说明新农合降低疾病经济风险的能力在提高。从模型结果看出,实际补偿比与灾难性卫生支出的发生成反比；住院天数两周以上,住院机构级别越高,灾难性卫生支出的发生率越高；健康自评状况和经济状况较好的家庭灾难性卫生支出发生率较低；并且有65岁及以上成员的家庭更容易发生灾难性卫生支出。在40%阈值下,新农合补偿后三年灾难性卫生支出发生率的集中指数分别为0.155,0.096和0.034。三年住院费用补偿比的贡献促使灾难性卫生支出较多的发生在高收入家庭。2006年至2008年集中指数变化为-0.019,2008至2011年集中指数变化为-0.062。说明三年灾难性卫生支出的发生虽然集中在高收入人群,但有向低收入人群方向发展的趋势。2006年至2008年补偿比和封顶线的提高促使2008年低收入人群的灾难性卫生支出发生高于2006年,补偿比主要归因于弹性的变化,封顶线归因于集中指数的变化。2008年至2011年补偿比和封顶线的贡献促使2011年高收入人群灾难性卫生支出发生高于2008年,归因于弹性和集中指数的共同作用。结论与政策含义根据研究结果,本研究主要结论有：(1)新农合补偿比的提高,能够增加住院服务的可及性。新农合补偿比的提高能够增加农村居民的住院服务利用和住院次数,降低自付费用,增加受益程度,并且随着补偿比和封顶线的提高,新农合风险规避的能力在提高。(2)住院服务利用和受益程度存在利于富人的不公平,灾难性卫生支出主要集中在高收入家庭。住院服务利用的不公平主要是由于收入不公平导致,补偿比增加对不公平性的贡献不大,三年低收入人群的住院服务需要相对于高收入人群仍然没有得到满足。2006年和2008年补偿比促使低收入人群受益程度较高,2011年补偿比和封顶线的贡献促使高收入人群受益程度较高。三年经济状况、住院天数、住院机构、补偿比是增加受益程度不公平的主要因素。住院天数长、住院机构层级高是导致富裕家庭发生灾难性卫生支出主要原因,家庭规模和高补偿比是贫困家庭发生灾难性卫生支出的保护因素。(3)2006年-2011年补偿比的提高相对促使高收入人群利用更多的住院服务,2006年-2008年与2008年-2011年期间对受益程度和灾难性卫生支出公平性变化的影响方向相反。2006年-2008年新农合广覆盖阶段,新农合补偿比的提高促使高收入人群获得更多的住院服务,补偿比、封顶线的提高和起付线的设置促使受益程度向利于低收入人群方向发展,虽然低收入者的补偿比得到了提高,但是由于住院天数和住院机构层级的提高,相对增加了低收入家庭灾难性卫生支出的发生。2008年-2011年新医改实施前后,新农合补偿比的提高促使高收入人群获得更多的住院服务,2011年高收入家庭的补偿比和封顶线得到了提高,受益程度向有利于高收入人群方向发展,新医改虽然增加了新农合的保障力度,但是较多的补偿给了高收入人群,低收入人群的受益程度并没有得到提高。高收入人群由于平均住院天数和住院机构层级的提高,相对增加了灾难性卫生支出的发生。根据上述结论,本研究具有以下政策含义：1)完善适宜的新农合补偿水平,在提高可及性整体水平的同时,提高可负担性,防止农村居民因住院总费用的增长超过补偿水平的增长而导致疾病经济负担；2)降低乡镇卫生院住院起付线,提高县医院住院起付线,合理引导就医流向,降低住院服务利用的不公平；3)缩短住院天数,提高服务质量,减少不必要的医疗支出,降低灾难性卫生支出的风险；4)继续健全医疗救助制度,适当增加补助,提高低收入人群住院服务利用和抵御疾病经济风险的能力。更多还原

【Abstract】 BackgroundThe Chinese New Rural Cooperative Medical System (NCMS) was carried out from2003. It has a great process with the support of governments at all levels since the pilot was initiated. With the development of New Rural Cooperative Medical System, the policy was also adjusted and changed constantly. The basic coverage was realized in2008and the new healthcare reform was carried out in2009. Although the effect of NCMS has got a certain degree of recognition, but there are still some different voices about the actual effect of the NCMS. One of the main different opinions is although the medical coverage was expanded, the medical costs increased faster. The utilization of hospitalization service can generate higher medical costs, it can induce the households into poverty to a certain extent, although the actual hospitalization reimbursement ratio reached41.5%in2009, but the increasing coverage of NCMS did not decrease the economic risk of disease.There are many studies about NCMS, from the view of study content, some focused on the description of NCMS, some focused on the evaluation of implementation effects of NCMS; about the analysis methods, some used the descriptive analysis, some adopted more strict methods, such as propensity score matching (PSM), logistic regression, difference in difference method (DID), about the analysis of the fairness, most literatures used Concentration Index (CI), Lorenz curve, slope index of inequity (SII) and other methods, there still exist deficiencies by applying Oaxaca decomposition method to analyze changes of equity; about the study period, there still lack system analysis during the wide coverage period after the implementation of NCMS and the period before and after the healthcare reform; about the variables, the variables about NCMS were whether to participate the NCMS, some literatures also incorporate the NCMS policies into the analysis model, there still lack system analysis on the changes of reimbursement rate, deductible and ceiling; about the type of data, most of the studies used cross-sectional data or pooled cross-sectional data, there are only a little studies used household level or village level panel data, there is still lack of study using individual level(same individual) panel data.In2006,2008and2011, the hospitalization reimbursement rate of the three counties (Zhangqiu City, Changqing District, and Pingyin County) showed an increasing trend, and the reimbursement rate decreased with the increase of hospital level. The ceiling also showed a great increasing trend. Pingyin and Changqing didn’t set up deductible in2006and2008, but all medical institutions set up deductible in2011, ZhangQiu reduced deductible of county and municipal hospital and set up deductible in township level hospital in2011. From the wide coverage period before2008, to the period that after the first stage of the healthcare reform in2011, whether the improvement of hospitalization reimbursement rate, the increase of ceiling and the set of deductible have increased the accessibility of inpatient service? Whether improved equity? These are the main concerns of the study.This study used individual level panel data, based on the health needs theory, starting from income effect of the insurance, price effect and risk aversion effect, analyzing whether the improvement of the compensation level and ceiling, and the set of deductible has increased the accessibility of inpatient services (including the overall level of accessibility and affordability), and whether improved the equity, to provide the policy basis of improving accessibility and equity of inpatient service.ObjectivesThe objective of this study is to analyze the effect of the evolution of NCMS compensation policies including increasing reimbursement rate and the ceiling, and setting the deductibles on the accessibility and equality of inpatient services, to reveal the related influencing factors, and to provide policy support to improve the accessibility of inpatient services and reduce inequity and economic burden. The specific objectives are as follows:to analyze the impact of the evolution of NCMS policies (increasing compensation rate and the ceiling, setting the deductibles) on the accessibility of inpatient services; to analyze the impact of the evolution of NCMS policie on the equality of inpatient servers, and explore the contribution of policy changes to the inequality; to clarify the longitudinal trends of inequality, the contribution of policies on the inequality, further reveal whether the changes of equality are due to the contribution of the elasticity or the contribution of concentration index, and provide basis for policy improvement.MethodsThis research mainly studied the inpatient services over the last year, using the household survey data which from three counties or districts of Jinan city (Zhangqiu county, Changqing district, Pingyin county) in the year2006,2008and2011, including a total sample of9020individuals from2600households. The unbalanced panel data of three years based on the year2008at individual level was selected as the study sample; the data ensured that the same individual was investigated at least twice in the three-year investigation.Analysis methods that were used in this research mainly included descriptive statistical analysis, single factor statistical inference and multivariate statistical inference. Multivariate statistical inference included the following five parts:(1) Using random effects of binary choice panel data model to analyze whether new rural cooperative medical system compensation policies affect the use of inpatient services or not.(2) Using negative binomial random effects model to analyze the impact of new rural cooperative medical system on the number of hospitalizations.(3) Using generalized linear models to analyze the impact of new rural cooperative medical system on hospitalization out of pocket expenses and reimbursement costs.(4)Using concentration index to measure the equity and using decomposition of concentration index to analyze the contribution of new rural cooperative medical system policy variables to equity.(5) Using Oaxaca decomposition method to analyze the time variation of equity, and then the time variation is decomposed into changes of concentration index and elasticity.Results(1) This study found that the hospitalization rates (in terms of the number of visitors) in2006,2008and2011were4.77%,4.40%and4.79%respectively, and there was no significant difference. The reimbursement level was positive proportional to utilization of hospitalization and frequency of hospitalization. County-level hospitals had deductible and township hospitals did not have deductible, which is inversely proportional to the frequency of hospitalization. There was no significant in the effect of payment ceiling on utilization of hospitalization and frequency of hospitalization. Economic status was a significant factor affecting the utilization of hospitalization.Concentration index of the utilization of hospital services was0.255,0.326and0.244in the three years respectively, indicating that the utilization of hospitalization of high income group was higher than low income group. Increasing reimbursement rate of NCMS prompted high income group to get more inpatient services, but the contribution was smaller. In the year of2006and2008, township hospitals did not have deductible, but county-level hospitals had deductible, which reduced the inequality of utilization of hospital services and prompted low income group to use more hospital services. The change of concentration index between2006and2008was0.071. The change of concentration index between2008and2011was-0.082. The contribution of reimbursement rate and deductible make high income group using more inpatient services in2006and2008. The contribution of new rural cooperative reimbursement rate is in opposite direction to the overall change, indicating that reimbursement rate was not the reason to make low income population use more hospital services in2008and2011. The main reason was the improvement of economic conditions.(2) In the year of2006,2008and2011, actual inpatient reimbursement rate was12.15%,12.46%and33.40%respectively. Nominal reimbursement rate and actual reimbursement rate was inversely proportional to inpatient out of pocket expense and was proportional to reimbursement cost. Setting deductible can significantly improve the patients’out of pocket payment. Choosing county level and above hospitals and two weeks or above length of stay can increase the out of pocket payment and reimbursement costs significantly.In the year of2006,2008and2011, Concentration index of reimbursement cost was0.035,0.020and0.033respectively. The contribution of real inpatient reimbursement rate on concentration index was negative in2006and2008. Contribution of reimbursement rate and deductible was positive in2011.The change of concentration index between2006and2008was-0.015.The change of concentration index between2008and2011was0.013. The contribution of NCMS policies of reimbursement rate, deductible and ceiling prompted concentration index of NCMS beneficial degree favoring low income group during2006and2008, accounting for11.33%of total change. The improvement of reimbursement rate and ceiling prompted concentration index of NCMS beneficial degree favoring high income group during2008and2011, accounting for32%and17%of the change of concentration index in the two years.(3) The incidence, mean gap and positive gap of catastrophic health expenditure after reimbursement of New Rural Cooperative Medical System were obviously lower than that before; the rate has showed a decreasing tendency after reimbursement of NCMS in three years; the scale of decrease was increased almostly. As we can see from the results of the model, the actual reimbursement rate was inversely proportional to the occurrence of catastrophic health expenditure, and staying in hospital for more than two weeks and visiting higher level hospital can significantly increase the catastrophic health expenditure, but the families which self-evaluate health status and economic status were better have a lower incidence of catastrophic health expenditure. Households that have members above65years old faced with catastrophic health expenditure more easily.At40%threshold, after the reimbursement of NCMS, the concentration index of incidence of catastrophic health expenditure were0.155,0.096and0.034respectively in the three years. The contribution of reimbursement rate leads to high income group occurred more catastrophic health expenditure. The change concentration index was-0.019during2006and2008, during2008and2011concentration index change was-0.062. Although the occurrence of catastrophic health expenditure for three years concentrates in high income households, it has a tendency to favoring low-income population. The rising reimbursement rate and ceiling induced the occurrence of catastrophic health expenditure in low income households in2008higher than2006. The reimbursement rate was mainly attributed to the change of elasticity, and ceiling was mainly attributed to the change of concentration index. The contribution of the reimbursement rate and ceiling in2008and2011induces the occurrence of catastrophic health expenditure in high income households in2011higher than2006and that is mainly attributed to the co-effect of elasticity and the concentration index.Conclusions and Policy ImplicationsAccording to the research results, the main conclusion of this research is:(1) The increase of NCMS reimbursement rate can increase the accessibility of inpatient service.The increase of NCMS reimbursement rate can increase the utilization and the number of hospital service, reduce out-of-pocket payments, increase benefit, and with the improvement of reimbursement rate and ceiling, the ability of new rural cooperative medical system in risk aversion is improving; the setting of the deductible will improve the patients’ out-of-pocket payments.(2) There exist inequality in hospital service utilization and reimbursement costs, favoring rich; catastrophic health expenditure was mainly concentrated in the high income households.Inequality of hospital service utilization was mainly due to income inequality; the increase of reimbursement rate contributed a little to the inequality, the need of hospital service is still not satisfied for low-income people relative to the high-income people in the three years.In2006and2008, reimbursement rate improved the beneficial degree in low-income group; reimbursement rate and ceiling contribute to high income group getting more reimbursement in2011. The economic status, length of stay, hospitalization institution, and reimbursement rate were the main factors to increase inequality of reimbursement costs. Length of stay, high level of hospitalization institution were the leading causes for catastrophic health expenditure happened in rich households; the household size and high reimbursement rate were the protection factors of catastrophic health expenditure in poor households group.(3) The improvement of reimbursement rate can prompt the rich group to use more hospital service during the period of2006-2011, but the orientation of influence on equality of beneficial degree and catastrophic health expenditure were opposite in2006-2008and2008-2011.During the broad coverage period of NCMS between2006and2008, the increased reimbursement rate improved the rich group using more inpatient services. The improvement of reimbursement rate and ceiling, and the setting of deductible induced beneficial degree favoring poor group. Although the reimbursement of low-income group has been improved, the length of hospitalization stay and high level of hospital institution have relatively increased the occurrence of catastrophic health expenditure in low-income households. During the new healthcare reform period between2008and2011, the increased reimbursement rate induces high income people to get more inpatient service. In2011, reimbursement rate and ceiling of high income household has improved, which benefit more high income group. Although the new healthcare reform increased the security intensity of new rural cooperative medical system, it compensated more for the high income groups and did not improve the benefit level of low-income group. But because of the increasing of hospitalization length of stay and hospitalization institution level, the incidence of catastrophic health expenditure of high-income households was increased relatively.Based on the above conclusions, this study has the following policy implications:1) improve appropriate compensation level, raise the overall level of accessibility, and improve the affordability, to prevent rural residents bearing high economic burden of disease due to the increase level of hospitalization expense exceeding the increase level of reimbursement;2) reduce the deductible in township hospitals, improve the deductible in county hospitals, guide patients to choose hospital reasonably, and reduce the equality of hospital service utilization;3) shorten hospitalization length of stay, improve service quality, reduce unnecessary medical expenditure and reduce the risk of catastrophic health expenditure;4) continue to improve medical assistance system, increase subsidies properly, and improve inpatient service utilization and the ability to resist economic risk of disease in low income group.更多还原