基于决策树模型的心力衰竭患者住院费用分析

Analysis of hospitalisation costs for heart failure patients based on a decision tree model

目的探索符合心力衰竭患者住院费用的病例组合,制定心力衰竭患者住院费用参考标准,为提高医院的管理效率和医疗质量,控制住院费用提供参考。方法运用单因素分析和多元线性回归分析对住院费用的影响因素进行筛选,并将关键影响因素作为分类节点变量,基于穷举卡方自动交互算法的决策树模型构建心力衰竭患者病例组合方案,计算各组标准费用和费用上限。结果将转科情况、住院天数、治疗方式、转归情况、心力衰竭类型、入院途径、合并症或并发症数量作为关键影响因素,最终组建19个心力衰竭住院费用病例分组;分析显示可能存在费用不合理现象。结论心力衰竭病例分组合理,基于分组方案得到的住院费用标准、病种权重及费用上限标准,可为控制医疗资源消耗、规范医生诊疗行为、提高医疗质量和医院管理效率提供依据,有利于控制住院费用,减轻心力衰竭患者的疾病负担,为进一步完善并推进疾病诊断相关分组预付制提供较高质量的数据支持。

Objective To explore the case mix of hospitalization costs of heart failure patients,to develop a reference standard of hospitalization costs for heart failure patients,and to provide a reference for improving the management efficiency and medical quality of hospitals and controlling hospitalization costs.Methods The influencing factors of hospitalization costs were screened using single factor analysis and multiple linear regression analysis,and the key influencing factors were used as categorical node variables to construct a case-mix scheme for heart failure patients based on the E-CHAID decision tree model to calculate the standard costs and cost ceilings for each group.Results Transfer status,hospitalization days,treatment modalities,regression status,heart failure type,admission route,and number of comorbidities or complications were taken as the key influencing factors;19 case groupings of heart failure hospitalization costs were finally formed,and the analysis showed that there might be cost unreasonableness.Conclusion The grouping of heart failure cases is reasonable;based on the grouping scheme,we obtained the hospitalization cost standard,disease weight and cost ceiling standard,which provides a basis for controlling the consumption of medical resources,standardizing the diagnosis and treatment behaviors of doctors,improving the quality of medical treatment and the efficiency of hospital management,which is conducive to the control of hospitalization costs,reducing the burden of disease of patients with heart failure,and providing higher-quality data support for the further improvement and advancement of the prepayment system of DRGs.