基于随机森林及Logistic回归模型的急性心肌梗死死亡影响因素分析  

Analysis of Influencing Factors of Death from Acute Myocardial Infarction Based on Random Forest and Logistic Regression Models

作　　者：苑思雨 罗艳侠 Yuan Siyu;Luo Yanxia(School of Public Health,Capital Medical University,Beijing 100069,China;不详)

机构地区：[1]首都医科大学公共卫生学院,北京市100069 [2]首都医科大学附属北京友谊医院,北京市100050

出　　处：《中国病案》2025年第3期34-37,共4页Chinese Medical Record

基　　金：国家自然科学基金资助项目(81773512)。

摘　　要：目的分析急性心肌梗死死亡的影响因素,为有效降低急性心肌梗死死亡率提供参考依据。方法收集某三级甲等综合医院2010年1月1日—2023年12月31日住院的急性心肌梗死患者住院病案首页信息,应用随机森林算法对急性心肌梗死死亡影响因素进行重要性排序及降维,将降维后的影响因素变量纳入Logistic回归模型进行分析。结果随机森林算法显示变量数为7时误分类率最低,筛选得到的变量依次是心肌梗死类型、年龄、合并心源性休克、合并高血压、合并糖尿病、性别以及是否行经皮冠状动脉(冠脉)介入(PCI)治疗。Logistic回归分析显示心肌梗死类型、年龄、合并心源性休克以及是否行PCI治疗是急性心肌梗死死亡的影响因素。与前壁心肌梗死相比,未特指部位的急性心肌梗死死亡风险更高(OR=3.977,95%CI:2.708~5.843),下壁(OR=0.587,95%CI:0.378~0.912)、心内膜下(OR=0.528,95%CI:0.389~0.715)的死亡风险降低。与<45岁年龄组比,65~79岁(OR=8.597,95%CI:2.457~30.087)、≥80岁(OR=14.023,95%CI:3.987~49.327)死亡风险更高。合并心源性休克的急性心肌梗死死亡风险较高(OR=80.528,95%CI:56.665~114.439),进行PCI治疗是急性心肌梗死死亡保护因素(OR=0.071,95%CI:0.050~0.101)。结论前壁心肌梗死、年龄≥65岁、合并心源性休克的人群可能是心肌梗死死亡的高危人群,医院应加强对此类人群的监管并采取有效的治疗,从而降低院内病死率。Objectives This study aims to analyze the influencing factors of death from acute myocardial infarction,and provide reference for effectively reducing the death rate from acute myocardial infarction.Methods The front-page information of the hospitalized medical records of patients with acute myocardial infarction who were hospitalized in a tertiary-level general hospital from January 1,2010 to December 31,2023 was collected.The random forest algorithm was used to rank the importance and reduce the dimension of the influencing factors of death from acute myocardial infarction,and the influencing factor variables after dimension reduction were incorporated into the Logistic regression model for analysis.Results The random forest algorithm showed that when the number of variables was 7,the misclassification rate was the lowest.The variables screened were type of myocardial infarction,age,concurrent cardiogenic shock,concurrent hypertension,concurrent diabetes,gender,and whether percutaneous coronary intervention(PCI)treatment was performed.Logistic regression analysis showed that myocardial infarction type,age,combined cardiogenic shock and whether PCI treatment was performed were influencing factors for death from acute myocardial infarction.Compared with anterior wall myocardial infarction,the risk of death in unspecified acute myocardial infarction was higher(OR=3.977,95%CI:2.708-5.843),while the risk of death in inferior wall(OR=0.587,95%CI:0.378-0.912)and subendocardial(OR=0.528,95%CI:0.389-0.715)myocardial infarction was lower.Compared with the age group<45 years old,the risk of death was higher in the age group of 65-79(OR=8.597,95%CI:2.457~30.087)and≥80 years old(OR=14.023,95%CI:3.987~49.327).The risk of death from acute myocardial infarction complicated with cardiogenic shock was higher(OR=80.528,95%CI:56.665-114.439),and PCI treatment was a protective factor for death from acute myocardial infarction(OR=0.071,95%CI:0.050-0.101).Conclusions People with forearm myocardial infarction,age≥65,and concomitan

关 键 词：急性心肌梗死 死亡 影响因素 随机森林 LOGISTIC回归 

分 类 号：R54[医药卫生—心血管疾病]