脓毒症患者预后的分类决策树分析  

Classification decision tree analysis of prognosis in patients with sepsis

作　　者：彭田英 黄华勇 邹文洁 于紫英[2] 彭正良[2] PENG Tianying;HUANG Huayong;ZOU Wenjie;YU Ziying;PENG Zhengliang(Department of Emergency,the Second Affiliated Hospital of University of South China,Hengyang 421001,Hunan,China;Department of Emergency,the First Affiliated Hospital of University of South China,Hengyang 421001,Hunan,China)

机构地区：[1]南华大学附属第二医院急诊科,湖南衡阳421001 [2]南华大学附属第一医院急诊科,湖南衡阳421001

出　　处：《中南医学科学杂志》2020年第5期544-547,共4页Medical Science Journal of Central South China

基　　金：湖南省卫生健康委课题(20201927)。

摘　　要：应用机器学习的分类决策树方法对脓毒症患者的预后进行分析,建立评估脓毒症预后的简易模型。收集并分析急诊科收治的167名脓毒症患者的临床资料,根据入院30天的生存状态将患者分为生存组和死亡组。结果显示,心率和收缩压两个自变量组成的精简模型能够对72.5%的病例正确分类,心率大于118次/分是脓毒症30天死亡重要的分类指标。本研究说明,心率和收缩压结合分类决策树方法可以对脓毒症患者的预后进行高效的分类。To classify the outcome of sepsis,and establish a simple model of sepsis prognosis by using the classification decision tree method of machine learning.The clinical data of 167 patients with sepsis admitted in the emergency department were collected and analyzed,and the patients were divided into survival group and death group according to the 30-day survival status.Results showed,a simplified model consisting of two independent variables,heart rate and systolic blood pressure,can correctly classify 72.5%cases.A heart rate greater than 118 beats per minute is an important classification indicator for the 30-day death.Therefore,heart rate and systolic blood pressure combined with classification decision tree method may efficiently classify the prognosis of sepsis.

关 键 词：脓毒症 分类决策树 心率 收缩压 预后 

分 类 号：R631[医药卫生—外科学]