□ 曾祥嫚 ZENG Xiang-man 吴姗姗 WU Shan-shan 茅月存 MAO Yue-cun 杨毓芳 YANG Yu-fang

Objective To evaluate the surgical inpatient service capacity of a general hospital using DRGs indexes, so as to provide data support for refine management and continuous improvement of medical service performance. Methods The 21 surgical departments of a tertiary general hospital in Ningbo City were taken as the research subjects. The inpatient service capacity of the departments was evaluated by TOPSIS and RSR method using the indicators of 2018 including the number of Diagnosis Related Groups(DRGs), the total weight of DRGs, the case mix index (CMI), the number of difficult cases (RW≥2), the number of third and fourth grade operation cases, time consumption index and cost consumption index. The combination of TOPSIS and RSR methods were used to comprehensively evaluate the inpatient medical service capacity of each surgical department in 2018. Results Among the 21 departments, department A13, A18 and A19 were evaluated as good; department A1, A2, A3, A4, A5, A8, A9, A11, A12, A14, A15, A16, A17 and A20 were evaluated as middle; department A6, A7, A10, A21 were evaluated as bad. Conclusion Combination of TOPSIS and RSR method based on DRGs indicators is scientific for evaluating the surgical inpatient service capacity of general hospitals.

诊断相关组(Diagnosis-related groups, DRGs)作为医疗服务的管理工具，广泛应用于医疗绩效管理和医疗费用管理[1]。为综合了解宁波市某三甲综合医院外科住院医疗服务能力水平，本研究选取DRGs组数、DRGs总权重数、CMI值、疑难病例(RW≥2)例数、三四级手术例数、时间消耗指数、费用消耗指数等DRG指标，应用TOPSIS (Technique for Order Preference by Similarity to an Ideal Solution)法结合RSR(Rank Sum Ratio)法对2018年各外科科室住院医疗服务能力进行综合评价，为医院和科室精细化管理、持续改善医疗服务绩效提供数据支持。

资料与方法

1.资料来源。资料来源于2018年宁波市某三甲综合医院住院病例病案首页数据上传至该院DRGs绩效考核系统后，经上海联众分组器(2016版)分组后的结果，研究对象为该院21个外科科室(A1—A21)。

2.评价指标。根据分组器分组后的结果，计算代表各外科科室医疗服务能力的指标：DRGs组数、DRGs总权重数、CMI值、疑难病例(RW≥2)例数、三四级手术例数，这5项指标均为正向指标，其值越大越优；计算代表各外科科室医疗服务效率的指标：时间消耗指数、费用消耗指数，这两项指标均为负向指标，其值越小越优；由于2018年各外科科室“低风险死亡率”均为0，本次评价研究不采用此指标。

3.研究方法。应用TOPSIS法结合RSR法对各外科科室住院医疗服务能力进行综合评价[2]。其主要步骤如下：(1)对指标进行同趋化与归一化处理：将原始资料中的低优指标取负值，使其转化为高优指标，然后对同趋化后的原始数据矩阵进行归一化处理，以此消除指标计量单位的影响，即：表示第i个评价对象在第j个指标上的取值，aij为归一化处理后第i个评价对象在第j个指标上的取值；……