摘要目的:将疾病诊断相关分组(DRG)指标引入医院效率评价，分析DRG对效率值的调整效果，并对比分析数据包络分析(DEA)效率值与DRG效率指标的异同。方法:利用湖北省三级医院2017年和2018年的DRG试点数据及其他报表的相关数据，分别采用实际开放床位数、在职员工数、实际占用总床日和医疗收入为投入指标，采用总诊疗人次、出院人次、DRG权重和DRG组数为产出指标，组合不同指标建立7个模型，以Bootstrap-DEA效率评价方法计算医院效率。结果:以传统方法计算的效率值存在医院级别越高效率越低的现象，县属医院、市属医院、省属医院的效率值分别为0.83、0.74、0.71( P<0.01)。将DRG权重和组数作为产出指标进行DEA分析，县属医院、市属医院、省属医院的效率值分别为0.95、0.94、0.96( P＝0.20)。县属医院、市属医院、省属医院的DRG效率指数分别为1.42、1.11、1.00( P<0.01)，级别越高效率越低，而采用DEA计算的效率值则分别为0.84、0.82、0.86( P＝0.58)。 结论:以传统方法计算的效率值存在系统性偏差，会出现医院级别越高效率越低的现象。将DRG权重和组数作为产出指标进行DEA分析，可以有效校正这一偏差。研究结果为开展DRG管理下的医院效率评价提供了新的思路。

abstractsObjective:With Diagnosis related groups(DRG)indicator introduced into the hospital efficiency evaluation, to analyze the adjustment effect of DRG on the efficiency value, and to compare the similarities and differences between the DEA efficiency value and the DRG efficiency indicator.Methods:The DRG pilot data in 2017 and 2018 of tertiary hospitals in Hubei province and related data from other reports were collected. Indicators including the actual number of beds, the number of employees, the actual total bed days occupied, and medical income were used as input indicators, and total admissions, number of discharges, weight of DRG and number of DRG groups were used as output indicators. 7 models were constructed using different indicator combinations. The efficiency of the hospital was calculated by the Bootstrap-DEA efficiency evaluation method. The relationship between hospital attributes and efficiency was analyzed by a generalized linear model.Results:There was an abnormal phenomenon that the higher the hospital level, the lower the efficiency in the efficiency value calculated by the traditional method. The efficiency values of county hospitals, city hospitals and provincial hospitals were 0.83, 0.74 and 0.71, respectively( P<0.01). DRG weight and group number were used as output for DEA analysis, and the efficiency values of county hospitals, city hospitals, and provincial hospitals were 0.95, 0.95 and 0.96( P=0.20)respectively. The DRG efficiency indications of county hospitals, city hospitals and provincial hospitals were 1.42, 1.11, 1.00 respectively( P<0.01). The higher the level, the lower the efficiency, while the efficiency values calculated by DEA were 0.84, 0.82, 0.86, respectively( P=0.58). Conclusions:The efficiency value calculated by the traditional method presents a systematic bias, which could be corrected effectively when DEA analysis using DRG weights is used as output indicator. The results developed new ideas of efficiency analysis for hospitals practicing DRG management.