【作者】 谢慧玲；

【导师】 程晓明；

【作者基本信息】 复旦大学 ， 社会医学与卫生事业管理， 2010， 博士

【摘要】 一、研究背景合作医疗制度在上个世纪80年代以前是我国农村居民医疗保障制度的主要模式,对维护农村居民的健康做出了不可磨灭的贡献。随着农村经济体制的变革,合作医疗制度在全国的覆盖率锐减。到2003年79.1%农村居民没有任何医疗保险,农村居民因病致贫、因病返贫的问题非常突出。为了解决我国农村居民的医疗保障问题,中央政府提出建立“以大病统筹为主”的新型农村合作医疗制度,以降低农村居民的疾病经济风险,解决因病致贫、因病返贫的问题,改善农村居民的健康状况。自2003年试点以来,虽然新型农村合作医疗的覆盖面、受益面和受益程度不断扩大,农村居民的潜在卫生服务需求在一定程度上得到了释放,但是新型农村合作医疗的实际住院补偿比例还比较低,农民的住院经济负担还比较重,到2008年仍需自付65%左右的住院费用,新型农村合作医疗尚不足以解决农民因病致贫,因病返贫的问题。已有关于新型农村合作医疗实际住院补偿比例的研究中,针对住院补偿比例影响因素的研究较少,也少有研究从全局的角度、系统的探讨导致实际住院补偿水平偏低的根本原因,通过实证研究分析相关因素对实际住院补偿比例影响程度的研究也比较少。为推动新型农村合作医疗的健康可持续发展,提高新型农村合作医疗的实际住院补偿比例,减轻参合农民的住院疾病经济负担,本研究选取开展新型农村合作医疗四年以上的县(市、区)作为研究对象,在对新型农村合作医疗实施情况进行分析的基础上,探索导致实际住院补偿比例偏低的原因,探讨影响新型农村合作医疗实际住院补偿比例的因素及其影响程度,为进一步发展和完善新型农村合作医疗制度提供参考。二、研究目的本研究的目的是通过对样本县(市、区)新型农村合作医疗制度运行状况的实证分析和研究,探讨导致实际住院补偿比例偏低的根本原因,从参合者、住院补偿方案和新型农村合作医疗运行状况三个方面定量的分析影响新型农村合作医疗住院补偿比例的因素及其影响程度,为提高实际住院补偿比例,促进新型农村合作医疗制度的可持续发展提供依据。具体目标为：1、了解样本县(市、区)新型农村合作医疗制度在运行过程中存在的问题。2、探讨导致新型农村合作医疗实际住院补偿比例偏低的原因。3、探讨影响和制约参合者实际住院补偿比例的因素及其影响程度。4、探讨影响样本地区实际住院补偿比例的关键因素。5、通过典型案例研究分析住院补偿制度、筹资标准、住院费用对实际住院补偿比例的影响。6、提出提高实际住院补偿比例的政策建议。三、研究方法(一)资料来源1、调查现场根据我国各省(市)的地理位置和经济发展水平,选取河南省、浙江省、辽宁省、重庆市和云南省五省,结合新型农村合作医疗运行状况在每个省各选2个新型农村合作医疗开展4年以上的县(市、区)作为调查现场,并最终确定河南省的长葛市；浙江省的安吉县、杭州市萧山区；辽宁省的法库县；重庆市的渝北区、荣昌县；云南省的江川县和玉龙县为调查现场。2、调查方法调查分两个阶段进行。2007年8-10月为第一阶段,收集各样本县(市、区)2004-2006年新型农村合作医疗制度运行的数据资料和文件。2009年10-11月为第二阶段,补充收集各样本县(市、区)2007-2008年的有关数据资料及相关文件。(1)机构调查。通过向各地新型农村合作医疗管理办公室发放“新型农村合作医疗运行状况调查表”,收集2004-2008年当地社会经济、人口、卫生资料状况和新型农村合作医疗运行状况资料,同时收集2004-2008年新型农村合作医疗住院补偿数据库资料。每个样本县(市、区)调查2-3所县级医院、4-5所乡镇卫生院、14-15所村卫生室,共计118所村卫生室、38所乡镇卫生院和20所县级医疗机构。了解2003-2007年县、乡两级医疗机构收支状况以及2007年县、乡、村三级医疗机构药品使用状况资料。(2)个人问卷调查于2007年利用自行设计的问卷,采用系统抽样的方法,在四个样本地区对2006年获得新型农村合作医疗住院补偿的参合农民进行问卷调查,了解其个人及家庭的基本情况、住院服务利用及新型农村合作医疗补偿情况、家庭收支状况以及对新型农村合作医疗的看法和意见。共收集有效问卷1034份。(3)定性访谈于2007通过关键知情人访谈和焦点组访谈,了解各县(市、区)新型农村合作医疗制度和住院补偿中存在的问题,以及各方对提高实际住院补偿比例的意见和建议。(4)二手资料收集系统收集国内外关于新型农村合作医疗的研究文献；中央政府自2002-2009年颁布的有关新型农村合作医疗制度的政策文件。(二)分析方法1、文献归纳法。对收集的国内外文献及各级政府的政策文件进行归纳和总结。2、灰色系统理论。运用灰色关联分析和灰色GM(1,N)模型技术探讨影响样本地区新型农村合作医疗实际住院补偿比例的关键因素及其影响的程度和方向；利用灰色GM(1,1)模型法预测次均住院费用,并在此基础上测算预测期内一定住院补偿比例下的人均保费,探讨住院费用、筹资标准和实际住院补偿比例之间的关系。3、医疗保险费用粗估法。通过医疗保险费得粗估法,测算不同新型农村合作医疗住院补偿方案下的人均保费,探讨住院补偿方案、筹资标准和实际住院补偿比例的关系。4、卫生系统诊断树分析法。通过卫生系统诊断树分析,由“果”溯“因”,探寻导致新型农村合作医疗实际住院补偿比偏低的根本原因。5、灾难性卫生支出测量方法。通过计算新型农村合作医疗补偿前后参合者因住院导致家庭灾难性支出发生的频率,分析新型农村合作医疗制度对减轻参合者住院经济负担的作用。6、因病致贫的测量方法。通过计算参合者住院前后及新型农村合作医疗补偿后住院费用导致家庭贫困的比例,分析新型农村合作医疗制度缓解或减轻因病致贫的能力。7、路径分析方法。利用路径分析的技术建立关于新型农村合作医疗实际住院补偿比例影响因素的模型,分析各因素之间的关系及相互影响程度的大小和方向。(三)分析工具新型农村合作医疗管理机构和医疗机构调查表资料录入Excel 2003并进行统计分析。新型农村合作医疗住院补偿数据库资料利用SPSS16.0进行统计分析。定性访谈的录音资料导入Mamqda2进行分析。个人问卷调查资料在录入Epidata2.1后利用SPSS16.0、AMOS16.0进行分析。灰色关联分析、灰色GM(1,1)模型和GM(1,N)模型等技术方法利用灰色建模系统软件和MATLAB6.0对资料进行分析处理。四、研究结果(一)样本县(市、区)社会经济、人口及卫生资源状况2003-2008年样本地区社会经济发展状况良好,农民收入稳步提高,农业人口的比例逐年减少。但是中、西部样本地区农民收入还比较低,绝大多数地方政府财政收入有限,收不抵支出现赤字。在卫生领域中,地方政府对医疗卫生事业的投入不足,到2008年大多数样本地区卫生事业费占财政支出的比例不足4%。农村基层卫生资源总量不足的问题依然比较严峻,到2008年绝大多数样本地区每千人医生数、每千人护士数、每千人床位数及医护比低于全国平均水平。(二)样本县(市、区)新型农村合作医疗制度运行状况及存在的问题2004-2008年样本地区新型农村合作医疗覆盖面逐年提高,到2008年部分地区参合率接近100%,提前实现了基本覆盖农村居民的目标；基金运行平稳,除个别地区基金略有超支外,大多数地区新型农村合作医疗基金结余均有结余且结余率逐年降低。研究期间,样本地区新型农村合作医疗制度的受益面和受益程度逐年提高,到2008年实际门诊补偿比例最高达100%,实际住院补偿比例最高达44.17%；大多数样本地区乡、村两级医疗机构门诊补偿人次占门诊补偿总人次的比例超过90%；所有样本地区县、乡两级医疗机构的住院补偿人次占住院补偿总人次的比例均在65%以上。2006年新型农村合作医疗制度补偿后参合者的贫困率降低5.49%,住院费用构成参合者家庭灾难性支出的比例降低了7.5%。参合农民对新型农村合作医疗的满意度较高,为94.9%。灰色关联分析结果显示研究期间样本地区新型农村合作医疗制度综合运行状况逐年提高；玉龙县新型农村合作医疗综合运行状况在所有样本地区中最好。2004-2008年样本地区新型农村合作医疗运行中存在的主要问题是：①筹资水平偏低,筹资标准不公平；②补偿水平偏低；③基金结余较多与赤字并存；④基层医疗机构药品不足,新型农村合作医疗基本药品目录窄。(三)导致新型农村合作医疗实际住院补偿比例偏低的根源分析利用卫生系统诊断树分析方法本研究归纳出导致新型农村合作医疗实际住院补偿比例偏低的原因为：①新型农村合作医疗管理者缺乏专业技能,管理能力有限；②个人参合费收缴方式不合理；③对新型农村合作医疗制度的宣传不够；④县级财政财力有限；⑤农民收入水平还比较低；⑥农民风险意识淡薄；⑦对医疗机构投入不足而导致的服务能力有限等七个维度,并最终将其归纳为新型农村合作医疗制度的组织与管理、社会经济状况、卫生资源的配置状况、卫生服务体系的构建以及农民自身因素等五个方面。(四)样本县(市、区)新型农村合作医疗实际住院补偿水平影响因素分析利用多元回归技术对问卷调查资料进行分析的结果表明,影响参合者实际住院补偿比例的因素包括：参合者的文化程度、就诊医院的级别、住院天数。除参合者文化程度外,其他各因素与实际住院补偿比例具有负相关关系,且参合者就诊医院的级别对实际住院补偿比例的影响程度最大。结合路径分析模型的结果,参合者就诊医院的级别在进入方程的变量中影响范围也最广,是控制住院费用,提高实际住院补偿比例的关键环节。但是由于上述因素都直接或间接的通过影响住院费用而影响实际住院补偿比例。因此,在提高新农合实际住院补偿比例的过程中,控制参合者的住院费用是关键。另外,随着参合者收入的提高,引导参合者根据健康需求合理选择就诊医疗机构,规范医疗机构的诊疗行为,对控制医疗费用,提高实际住院补偿比例也具有积极的意义。灰色关联分析的结果显示,2004-2008年影响样本地区新型农村合作医疗实际住院补偿比例的前三位关键因素是住院受益率、参合率和基金到位率。即在筹资标准一定的情况下,各地在制定提高实际住院补偿比例的措施时应优先从保证适度的住院受益率、提高参合率等方面着手。同时还要保证各级政府的财政补贴能按时、全额的到位。(五)新型农村合作医疗实际住院补偿水平影响因素的典型案例研究运用多元线性回归分析技术对江川县2005-2008年住院补偿数据库资料进行分析的结果显示,影响参合者实际住院补偿比例的因素包括参合者的年龄、就诊医院的级别和住院费用中可补偿部分的比例。其中参合者就诊医院的级别对实际住院补偿比例的影响程度对大,且与实际住院补偿比例具有负相关关系。随着我国农村人口老龄化的不断加剧,要使参合者的住院受益水平不断提高并维持在一个较高的水平,必须不断提高新农合筹资水平。为控制住院费用,缩短参合者的住院天数也非常重要。灰色关联分析和灰色GM(1，N)模型的结果显示,2004-2008年影响江川县新型农村合作医疗实际住院补偿比例的关键因素是住院统筹基金使用率、人均筹资标准和次均住院费用。因此,江川县在提高实际住院补偿比例的过程中应在提高筹资标准的基础上,优先从提高住院统筹基金使用率,降低住院费用的措施着手。对江川县住院补偿方案、筹资标准、住院费用、实际住院补偿比例之间关系的研究结果显示,①2005-2009年江川县住院补偿方案设计不合理。2005-2008年乡镇及县外医疗机构住院补偿起付线偏低,县、乡两级医疗机构封顶线偏高；2005-2008年住院补偿标准偏低,2009年偏高。结果提示,在筹资标准一定的情况下,新型农村合作医疗住院补偿方案对实际住院补偿比例、基金平衡的影响较大。②在其他因素不变的情况下,提高住院补偿起付线、降低封顶线或单独提高封顶线对新型农村合作医疗实际住院补偿比例、人均保费和基金结余率的影响都比较小。提高实际住院补偿比例应在提高筹资水平的基础上主要通过提高住院补偿标准来实现,而设置合理的起付线和封顶线也是非常有必要的。③扩大补偿范围,严格贯彻执行基本药品目录和诊疗目录对提高实际住院补偿比例有积极的影响。④在住院补偿起付线、封顶线、补偿范围和住院费用一定的情况下,江川县新型农村合作医疗实际住院补偿比例每提高1个百分点,人均筹资额平均需增加5元。⑤若使2009年江川县新型农村合作医疗实际住院补偿比例达到40%,人均筹资标准达到120元较为适宜。相应的住院补偿方案为：各级医疗机构的起付线分别为150元、300元和600元,年累计最高补偿限额为20000元,住院补偿标准分别为乡镇55%,县级45%,省级25%,非定地点15%。⑥当次均住院费用和住院补偿水平同时提高时,筹资标准的增长速度应快于次均住院费用,前者平均约为后者的8倍。本研究比较了医疗保险费粗估法和灰色GM(1,1)模型在测算人均保费时的差异和优缺点。结果表明,灰色GM(1,1)模型解决了少数据建模预测的问题,但是由于对影响因素考虑的不充分可能导致测算结果误差较大。而医疗保险费粗估法具有细致、准确的特点。五、政策建议第一，提高筹资标准。加大政府对新型农村合作医疗的投入,建立财政补偿的长效机制,确保各级政府补助资金及时、足额到位；提高农民个人的缴费水平；拓宽筹资渠道。第二,科学设计和调整补偿方案。结合筹资标准和医疗费用水平合理设计住院补偿起付线、封顶线和补偿标准；补偿方案的调整应以提高住院补偿标准尤其是提高县、乡医疗机构的住院补偿标准为主,并体现向基层倾斜的原则；同时逐步扩大补偿范围。第三,改革对供方的支付制度,加强对医疗机构的监管。推广和完善单病种付费制度；探索其他形式的支付方式；加强对医疗机构的监管,强化医院的内部管理。第四,加强新型农村合作医疗管理机构的建设。加大对新型农村合作医疗管理机构的投入,加强和完善新型农村信息网络系统建设；增加管理人员编制,加强对管理人员的培训,优化专业结构。第五,增加对基层医疗机构的投入。加大对乡镇卫生院的财政扶持力度；加强对基层卫生技术人员的培训,提高基层医疗机构的服务提供能力。更多还原

【Abstract】 BackgroundBefore 1980s, Cooperative Medical Scheme in China was an main medical insurance of rural residents. It made a great contribution to the safeguarding of health. Following the innovation of rural economic system, coverage rate of cooperative medical scheme in China reduced quickly. In 2003,79.1% of rural residents were not covered by any form of health insurance. Many of them had no access to health care services because of economic reason and impoverished due to disease.In order to solve this problem, the central government determined to build up a new cooperative medical scheme (NCMS), the aim is to effectively relieve rural residents’burden of disease and to help alleviate the problem of impoverishment due to disease and improve health. Since 2003, NCMS has been built up, some potential health service demands of rural residents were released, coverage rate and benefit rate of NCMS increased continually. But effective reimbursement rate of NCMS were still low and burden of disease of rural residents was still heavy. In 2008, most of enrollees still paid about 65 percent of hospitalization cost by themselves. NCMS did not yet solve the problem of impoverishment due to disease.Previous studies on effective hospitalization reimbursement rate of NCMS were less concerning the influencing factors. The research related to systemically explore causes of low effective hospitalization reimbursement rate on NCMS and analyze the influence degree of those factors through empirical research were less reported. Therefore, in order to promote the sustainable development of NCMS and release the burden of disease of rural residents, this study chooses piloted counties which NCMS developed at least 4 years as objects, tries to explore the causes of low effective reimbursement rate in NCMS, then searches influencing factors, which may provide an evidence for improve and promote the development of NCMS.ObjectivesThe aim of this study is to explore causes of low hospitalization reimbursement rate on NCMS influencing factors and influence degree of them from enrollees, hospitalization reimbursement scheme and implementation of NCMS, which may provide evidence for improve NCMS effective hospitalization reimbursement rate and promote the development of NCMS.Detailed objectives are as follows:1. Finding out problems in the development of NCMS.2. Exploring causes of resulting in low hospitalization reimbursement rate of NCMS.3. Exploring determinants and influencing factors of hospitalization reimbursement level.4. Analyzing key factors of influencing effective hospitalization reimbursement level on NCMS of sampled counties.5. Exploring the relationship of hospitalization reimbursement scheme, premium, hospitalization cost and effective hospitalization reimbursement rate in case study.6. Providing political suggestion for improving hospitalization reimbursement scheme in NCMS and increasing effective hospitalization reimbursement rate.MethodologyData Sources1. Study sites. According to geographic and economic situation, five provinces Henan, Zhejiang, Liaoning, Chongqin and Yunnan were chosen. Two counties that their NCMS were built up above 4 years were selected from each province. As a result, Changge were chosen in Henan, Anji and Xiaoshan were chosen in Zhejiang, Faku were chosen in Liaoning, Yubei and Rongchan were chosen in Chongqin, Yulong and Jiangchuan were chosen in Yunnan.2. Data collection. The first stage was from August to October in 2007, and the second was from October to November in 2008.2.1 Institution investigation. Firstly, NCMS administration offices. A questionnaire was designed by researchers and filled out by NCMS officials in NCMS administration to know about the implementation of NCMS. Meanwhile, database related to hospitalization reimbursement were provided either. Secondly, hospitals. A questionnaire was designed by researchers and filled out by hospitals to know about their revenue, expenditure and drug coverage.2.2 Household survey. To know about hospital service utilization, NCMS reimbursement and attitude towards NCMS among enrolled rural residents, household surveys were conducted in four sampled counties in 2007, with sample size as 1034.2.3 Interviews. To know about the implementation, problems in NCMS and stakeholders’attitudes, key-informant interview and focus-group interviews were conducted.2.4 Second-hand materials collection. Academic references, policy documents about NCMS in the central government and sampled counties from 2002 to 2009 were collected in order to understand research frontiers concerning NCMS.Analysis Methods1. References reviews. Summarizing second hand materials collected.2. Gray system theories. Gray relational analysis and gray GM (1, N) model were used for exploring key influencing factors of effective hospitalization reimbursement rate in NCMS and their influence degree. Gray GM (1,1) model was used for forecasting hospitalization cost, then calculating premium.3. Premium calculation. To calculation NCMS’s premium in different benefit packages.4 Diagnosis tree. Finding out causes of low effective hospitalization reimbursement rate in NCMS by using diagnosis tree in health sector.5. Catastrophic health expenditure measurement. To know about the effect of NCMS. on reducing burden of disease by measuring the frequency of catastrophic health expenditure.6. Incidence of poverty measurement. Analyzing the effect of NCMS on reducing the poor by measuring incidence of poverty before and after NCMS reimbursement.7. Path analysis. Setting up path analysis model and analyzing relationships of variables and influencing level with effect hospitalization reimbursement rate in NCMS.Analysis ToolsData from institution were input and analyzed by Excel 2003. NCMS inpatient reimbursement database were analyzed through SPSS16.0 and AMOS16.0. Interviews were recorded with informant consent and analyzed by Maxqda2. Data concerning household survey were double input to Epi Data2.1 and analyzed by SPSS16.0 and AMOS16.0. Gray relational analysis, gray GM (1,1) model and gray GM (1, N) model were through Gray model software and MATLAB 6.0.Results1. Social economic development, population and health resources.During 2003-2008 in sampled counties, social economy developed quickly, income of rural residents increased gradually and the proportion of rural residents in population decreased. But rural residents’income of counties in the central and west areas was still low. And Most of local governments were deficit spending. Financial input for health sector was low in many sampled counties. Rate of financial input for health sector in financial expenditure in many sampled counties was still under 4% until 2008. Health resources were inadequate in sampled counties. Till 2008, medical doctors, nurses and hospital beds per 1000 people in most sampled counties were under the national average.2. The state quo of NCMS implementation and problems.During 2004-2008, participation rate of NCMS increased gradually and approached 100% in some sampled counties. NCMS fund ran smoothly, deficit rate of fund was reduced gradually. Benefit rate and degree of enrollees were increased. In 2008, the highest effective outpatient reimbursement rate was 100%, and the highest effective hospitalization reimbursement rate was 44.17%. Above 90% of outpatients and above 70% inpatients were treated in basic hospitals. NCMS in sampled counties had a few effect in reducing enrollee’s burden of disease. It could protect 5.49% of enrollees off the occurrence of catastrophic health expenditure and reducing the severity of catastrophic health expenditure off 7.5%.94.9% of enrollees were satisfied with NCMS. The results of gray relational analysis showed NCMS in sampled counties implemented smoothly, and NCMS implementation in Yulong was the best in all sampled counties.There were some problems in NCMS implementation during 2004-2008. Firstly, the premium was low and unfired. Secondly, the effective reimbursement rate was low and enrollees still had heavy burden of disease. Thirdly, NCMS fund either had excessive surplus or deficit. Fourthly, basic medicine list and basic diagnosis and treatment list were narrow. And all of above problems finally led to low effective hospitalization reimbursement rate.3. Causation analysis of Low effective hospitalization reimbursement rate. Through diagnosis tree, causes for Low effective hospitalization reimbursement rat could be summarized as follows. Firstly, NCMS officials lacked knowledge on health insurance and administration abilities. Secondly, financial income of local government was limited. Thirdly, some rural residents’income was low. Fourthly, concept of health risk of most rural residents was generally weak. Fifthly, premium collection procedure was not reasonable. Sixthly, some of enrolled rural residents were misunderstanding of NCMS. Seventhly, low financing input of government led to low service abilities of hospitals.4. Influencing factors analysis of effective hospitalization reimbursement rate in NCMS.The result of path analysis showed hospital grade where enrollees hospitalized, hospital days, enrollees’education level were influencing factors of effective hospitalization reimbursement rate of NCMS for enrollees. The influence degree of hospital grade with effective hospitalization reimbursement rate was great and it could also influence other factors. Because all above factors could influence hospitalization cost directly or indirectly, hospitalization cost was key for increasing effective hospitalization reimbursement rate of NCMSThe results of gray relational analysis showed that benefit rate of hospitalization in NCMS, appropriation rate of fund and enrollment rate were the first three key factors with influencing effective hospitalization reimbursement rate of NCMS for sampled counties while premium of different counties was same or changed a little.5. Case study of influencing factors on effective hospitalization reimbursement rate in NCMSAccording to the results of multiple regression and correlation analysis, enrollee’s age, hospital grade where enrollees hospitalized and ratio of paying by themselves in hospitalization cost were influencing factors of their effective hospitalization reimbursement rate. And shortening hospital days was significative for controlling hospitalization cost as well.According to the results of gray relational analysis and gray GM (1, N) model, utilization rate of inpatient fund, premium and hospitalization cost were the first three key factors for effective hospitalization reimbursement rate of NCMS in Jiangchuan.Results of research of pay scheme in NCMS in Jiangchuan showed could be summarized as follows. Firstly, hospitalization schemes of NCMS from 2005 to 2009 were not reasonable. When premium is stable, influence degree of hospitalization scheme with effective hospitalization reimbursement rate and fund balance of NCMS is great. Secondly, increasing deductible and decreasing ceiling had small influence to effective hospitalization reimbursement rate and fund balance. And based on increasing premium, increasing nominal hospitalization reimbursement rate was significance to increase effective hospitalization reimbursement rate. Thirdly, extending the scope of basic drug list and basic diagnosis and treatment list were significance to increase effective hospitalization reimbursement rate. Fourthly, when the other parameters were stable, effective hospitalization reimbursement rate raises 1 percentage point, premium should increase 5 Yuan RMB. Fifthly, in 2009, when effective hospitalization reimbursement rate of NCMS in Jiangchuan was 40%, premium should be 120 Yuan RMB. Sixthly, when hospitalization cost and effective hospitalization reimbursement rate of NCMS in Jiangchuan increased at the same time, the growth of premium should be higher then one of hospitalization cost.According to results of premium calculation and gray GM (1,1) model, Results of premium calculation are more precise and accurate than those of gray GM (1,1) model. Gray GM (1,1) model can use a little data to build up a model, but the result might have a big error.Political Recommendation1. Increasing the premium. Firstly, government at all levels should further increase input of NCMS, and subsidize NCMS in time and in full amount. Secondly, increasing personal premium of rural residents. Thirdly, broadening funding channels of NCMS.2. Scientifically designing and adjusting pay scheme. Based on premium and health cost, scientifically designing the pay scheme. It is impartment to increase normal reimbursement rate of county hospitals’and town hospitals’. At the same time, gradually extending the scope of basic drug list and basic diagnosis and treatment list.3. Innovating payment system for hospitals and strengthening cost containment. Firstly, developing and improving case-based payment system. Secondly, exploring other models of payment system. Strengthening supervision and management of hospitals.4. Intensifying the capacity building in NCMS administration. Government at all levels should take overhead expenditure of NCMS office into considerations, add equipment, network and position in NCMS office. Meanwhile,employing more officials, improving administration abilities and knowledge of NCMS administration officials through trainings.5. Government should increase input to basic hospitals especially rural township health centers in rural area and strengthening medical technical personnel’s training to improve health service abilities of hospitals in rural area.更多还原