基于多元线性回归与决策树模型的病种费用分析——以儿童支气管肺炎为例

Analysis on Cost of Disease Kind Based on Multiple Linear Regression and Decision Tree Model

目的对儿童支气管肺炎住院费用分组,旨在DIP付费下,为住院费用的控制提出更合理的建议。方法通过统计软件SPSS 22对某三甲医院2019—2022年10月3448名儿童支气管肺炎出院患者,进行单因素分析和多元线性回归分析,结果筛选出分节点变量,用决策树模型对费用进行组合,并制定各组标准住院费用与权重。结果该院儿童支气管肺炎住院费用逐年上升,但费用结构却在优化。通过多元线性回归分析,将平均住院日、入院病况与年龄三个显著性因素作为分类节点,使用决策树模型,共分为6个费用组,并提出各组费用控制标准与预警上线。结论通过住院费用的预测值与费用预警线能够为支气管肺炎费用管理提供更为精准的参考与建议,节约卫生资源。

Objective:To group the hospitalization expenses of the children with bronchopneumonia in order to put forward more reasonable suggestions for the control of hospitalization expenses under DIP payment.Method:The statistical software SPSS 22 was used to conduct the univariate analysis and the multiple linear regression analysis on 3,448 children discharged with bronchial pneumonia from a 3A hospital from October 2019 to 2022.Result:The results showed that the node variables were screened out;the costs were combined by decision tree model;And the standard hospitalization costs and weights for each group were established.The hospitalization cost of children with bronchopneumonia in this hospital is increasing year by year,but the cost structure is being optimized.Through multiple linear regression analysis,the three significant factors of average length of stay,condition of admission and age were taken as classification nodes,and the decision tree model was used to divide them into 6 cost groups.The cost control standard and the early warning line of each group were proposed.Conclusion:The predicted value of hospitalization cost and the cost pre-warning line can provide more accurate references and suggestions for bronchopneumonia cost management and save health resources.