广西乳腺癌患者疾病诊断相关分组的研究  

Study on diagnosis-related groups in breast cancer patients in Guangxi

作　　者：李艳 吕炜[2] 唐忠 LI Yan;LYU Wei;TANG Zhong(Statistics Department,Wuming Hospital of Guangxi Medical University,Nanning 530199,China;Guangxi Center of Disease Prevention and Control,Nanning 530021,China;School of Information and Management,Guangxi Medical University,Nanning 530021,China)

机构地区：[1]广西医科大学附属武鸣医院统计室,南宁市530199 [2]广西壮族自治区疾病预防控制中心,南宁市530021 [3]广西医科大学信息与管理学院,南宁市530021

出　　处：《广西医学》2020年第12期1556-1559,共4页Guangxi Medical Journal

基　　金：广西教育科学“十三五”规划一般课题(2017C368)。

摘　　要：目的研究广西乳腺癌的疾病诊断相关分组(DRGs)方案及分组效果。方法收集广西199家医院共35 162例病案数据。应用相关分析比较各变量与住院总费用间的相关性,并根据文献分析及专家咨询结果选择关键变量。利用卡方自动交互检测算法构建决策树,得到DRGs,利用变异系数和方差减少量对DRGs进行评价。结果以住院天数、医院级别、治疗方式、合并症或并发症为关键自变量,住院总费用为因变量,进行决策树分析,得到24个DRGs,且组内同质性强,组间异质性大,分组合理。结论利用决策树构建乳腺癌DRGs科学合理,可以为本地区相关部门制定乳腺癌及其他病种分组费用标准提供参考。Objective To study the regimens of diagnosis-related groups(DRGs) and DRGs effect in breast cancer in Guangxi. Methods The medical record data of 35162 cases were collected from 199 Guangxi hospitals. Correlation analysis was applied to the comparison of the correlation between variables and total hospitalization costs, and key variables were determined based on literature analysis and expert consultation results. Chi-squared automatic interaction detection algorithm was adopted to establish decision-making tree so as to obtain DRGs, coefficient of variation and reduction in variance were applied to DRGs assessment. Results The decision-making tree analysis was performed using hospital stay, hospital tier, treatment approach, comorbidities/complications as key independent variables while total hospitalization costs as dependent variable, 24 DRGs were obtained, exhibiting strong inter-group homogeneity, great inter-group heterogeneity and reasonable grouping. Conclusion Applying decision-making tree to DRGs construction in breast cancer is scientific and reasonable,and can provide the reference on formulating the DRGs expense standard of breast cancer and other diseases for local relevant departments.

关 键 词：乳腺癌 疾病诊断相关分组 决策树 住院总费用 住院天数 医院级别 治疗方式 合并症 并发症 

分 类 号：R737.9[医药卫生—肿瘤]