指标选择对医院效率评价的影响——以2010年省级数据DEA模型为例

Impact of indicator selection on hospital efficiency assessment using data envelopment analysis:An empirical application with 31 provincial data in 2010

目的:探讨在应用DEA方法评价医院效率时,指标选择对分析结果的影响。方法:(1)以31个省2010年的数据为例,建立一系列包含不同投入产出指标数量的数据包络分析模型,观察指标选择对分析结果的影响;(2)通过实例说明专题地图在展示效率地域分布特征中的作用。结果:随着投入产出指标数量的增加,DEA分析结果的区分度逐渐降低,有效决策单元数量由5个增加到17个,默认有效单元的数量由3个增加到5个,效率均值由0.867上升到1.056。各地区医院效率分布大致呈现南、中、北的分布特征。结论:投入产出指标的数量严重影响DEA模型的分析结果,在数据支持的情况下,应该通过比较多种投入产出指标的组合以检验DEA分析结论的可靠性。专题地图在展示效率地域分布特征时具有简单、直观的优势,是发现效率影响因素和进行实证研究的基础。

Objective：To explore the impact of indicator selection on hospital efficiency assessment using data envelopment analysis.Methods：A series of DEA models are constructed using different number of input and output indicators,and observes the impact of the selected number of inputs and outputs on the results.The hospital data of 31 provinces in 2010 are used.A thematic map is used to show the geographical distribution of the efficiency analysis results.Results：The discriminating power of DEA results is remarkably reduced in line with the number of inputs and outputs is increasing.For instance,the effective decision-making units are increased from 5 to 17 as well as the default effective decision-making units increased from 3 to 5 respectively.In addition,the scale mean efficiency is increased from 0.867 to 1.056.The geographical distribution of efficiency scores shows decreasing tendency from south to north.Conclusion：The DEA result is highly dependent on the number of inputs and outputs.Multiple comparison using different combinations of inputs and outputs should be made to validate the conclusions,if data are available.A thematic map is a useful tool to estimate influencing factors and solid empirical evidence as regarding to show the geographical distribution of hospital efficiency.