基于决策树模型脑内出血患者DRGs分组分析  

DRGs Grouping Analysis of Patients with Intracerebral Hemorrhage Based on Decision tree Model

作　　者：曹丙艳 房振胜 刘钢 朱锋 唐欣元 赵旭 Cao Bingyan;Fang Zhensheng;Liu Gang;Zhu Feng;Tang Xingyuan;Zhao Xu(The Second Affiliated Hospital of Shandong First Medical University,Taian 271000,Shandong Province,China;不详)

机构地区：[1]山东第一医科大学第二附属医院,山东省泰安市271000

出　　处：《中国病案》2025年第3期47-52,共6页Chinese Medical Record

基　　金：泰安市2023年科学技术发展计划(2023NS467)。

摘　　要：目的通过研究脑内出血患者的DRGs病例组合分组及支付标准,为本地区相关部门优化该病种住院费用分组及医院合理控制住院费用提供参考依据。方法收集某三甲综合医院2022年1月1日—2024年5月31日脑内出血患者(ICD-10编码I61)住院病案首页信息共计1004例。利用单因素分析和多元线性逐步回归方法,筛选出住院费用分类节点变量,运用决策树(CHAID)算法建立脑内出血患者DRGs病例组合模型。结果脑内出血患者住院费用范围4398.59~213091.06元,平均值为39624.35,中位数为22823.01元,四分位数间距为43073.20元。以是否使用呼吸机治疗、住院天数、是否转科、是否合并肺炎、是否气管切开、是否手术为分类节点,最终产生8个病例组合,各组内CV值均小于1,组内同质性较好,不同组内费用差异有统计学意义,组间异质性较好。其中DRGs3组(未使用呼吸机,住院天数≤14天、发生转科)超额患者占比最多,DRGs2组(未使用呼吸机、住院天数>14天、未发生转科)超额费用占比最高,应重点监控。结论利用CHAID决策树模型构建的脑内出血患者住院费用分组合理,该分组方案可为本地区相关部门优化分组方案及医院控制住院费用、优化资源配置提供一定参考。Objectives This study aims to explore the DRGs case combination grouping and payment standards for patients with intracerebral hemorrhage,providing a reference for relevant departments in the region to optimize the grouping of hospitalization expenses for this disease and for hospitals to reasonably control hospitalization expenses.Methods The data on the front page of medical record of patients with intracerebral hemorrhage(ICD-10 code:I61)in a tertiary comprehensive hospital were collected from January 1,2022 to May 31,2024,with a total of 1004 cases.Univariate analysis and multiple linear stepwise regression methods were used to obtain the influencing factors of hospitalization costs,and the decision tree(Chi-squared Automatic Interaction Detection,CHAID)algorithm was employed to establish a case combination model of DRGs in patients with intracerebral hemorrhage.Results The hospitalization cost of patients with intracerebral hemorrhage ranged from 4,398.59 to 213,091.06 yuan,with an average value of 39,624.35 yuan,a median of 22,823.01 yuan,and a quartile range of 43,073.20 yuan.The classification nodes were whether the patient used a ventilator,the number of days in hospital,whether the patient was transferred to another department,whether the patient had pneumonia,whether the patient had a tracheotomy,and whether the patient had surgery.8 case combinations were generated.The CV values within each group were all less than 1.The intra-group homogeneity was good,the cost differences within different groups were statistically significant,and the heterogeneity between groups was good.Among the case combinations,the DRGs3 group(no use of ventilator,hospitalization days≤14 days,and transfer of departments)had the largest proportion of excess patients,and the DRGs2 group(no use of ventilator,hospitalization days>14 days,and no transfer of departments)had the highest proportion of excess costs and should be monitored closely.Conclusions The grouping of hospitalization expenses of patients with intracerebral hemorr

关 键 词：脑内出血 病例组合 住院费用 CHAID算法 

分 类 号：R197.3[医药卫生—卫生事业管理]