Data Envelopment Analysis With Interactive Variables And Its Application In Hospital Performance Evaluation

【Background】In recent years,as people’s living environment and living standard changing constantly,people demand of more medical services.although we have the radical reform of the medical service field,but we still have a long way on the way of reform,especially on how to reform the public hospitals.Now how to further reform the management system of public hospitals,establish a sound performance evaluation mechanism of input and output mechanism,standardize the performance evaluation of public medical and health institutions at various levels and of work,promote medical institutions to improve the service quality is still looming.Therefore,in order to better evaluate the hospital service efficiency and performance level we need a set of practical and feasible evaluation system.From the point of the present literature,there are two main types of hospital performance evaluation methods: parametric and non-parametric method.The parametric method represented by stochastic frontier analysis(SFA),but SFA needs to build a function model for single output variable,but hospital is multiple input and multiple output mechanism,so the non-parametric method data envelopment analysis(DEA)is more suitable for hospital performance evaluation,In all the classical DEA models,the combination form of the multiple inputs(or outputs)is a linear weighted sum,it needs an assumption that there is no interaction among the contributions from individual information sources so that the joint contribution is just the simple sum of contributions from individual information sources.But in practice,the variables(input or output)are usually strongly correlated,there are interactions among variables.In general,the inherent interaction cannot be ignored in the efficiency evaluation..So aiming at the disadvantages of classical DEA method,this research studies the variable interactions with DEA model,and carries on the empirical research,analysis of the results of the two models,to verify the feasibility of the new model.【Research purposes】(1)Using the cluster analysis,the correlation analysis and coefficient of variation analysis select suitable input and output indicators,to construct the index system of hospital performance evaluation.(2)On the basis of classic DEA model,we build up a new DEA model with interaction variables,and then explore the nature of the new model theorem and its advantages 【Research methods】(1)Index selection: using cluster analysis,correlation analysis and coefficient of variation analysis,select the appropriate input and output index.(2)Model choice: classical DEA model,DEA model with interaction variables and cross efficiency evaluation,to evaluate and rank the overall performance of the decision making units.(3)Analysis software: MaxDEA and Lingo12 software 【Research result】(1)The results of the classical DEA modelThe classical DEA model analysis result shows that there are 14 decision making units overall efficiency value is 1,which accounts for 45%.Illustrating that the 14 hospitals inputs is best combination in theory,the output results are also achieved the best,scale reward is in a state of constant.From 17 non DEA efficient DMU on behalf of the input and output indicators of hospital can be seen to improve value and target value,input indicators,improvement of the negative,to illustrate the 17 hospitals at different level of investment in redundant state,so these hospitals should reduce inventory,reasonable control the scale of hospitals and improve resource utilization,for getting the most output with the least amount of inputs.(2)The results of CH-CCR modelThe number of effective DMUs of this model is nine more than the classical DEA model,The hospital represent by DMU5,DMU7,DMU9,DMU12,DMU14,DMU16,DMU17,DMU24,DMU26,DMU31 are also effective in CH-CCR model,but the hospital DMU15 is effective in CCR model not in CH-CCR model.The efficient results of classical DEA model and CH-CCR model are relatively effective,rather than absolute.The results of CH-CCR model shows even DEA efficient decision making units,also has the improvement space,【Conclusion】(1)Model conclusionIn order to overcome the classic DEA model based on the analysis of input and output decision-making unit index which ignored the interaction effects on the accuracy evaluation results.Choquet integral is used as additive operator introduced a new data envelopment analysis(DEA)model CH-CCR model,this paper introduces the principle of CH-CCR model and the mathematical programming model and its property theorem,and join cross efficiency in the new model,which makes the results more accord with the actual situation.(2)Case study conclusionthe efficient results of the 31 hospital are different with the two model.the CH-CCR model calculation result is more practical significance and guiding value.using the CH-CCR model calculate the ultimate performance of value is less than 1,because the results of the DEA evaluation out is relative rather than absolute,even the relative efficiency of 1 hospital,the hospital service efficiency and performance level also has room for improvement.