基于决策树算法的心血管疾病住院患者医院感染发生风险预测模型构建

Construction of nosocomial infection risk prediction model for inpatients with cardiovascular disease based on decision tree algorithm

目的:构建心血管疾病住院患者医院感染发生风险预测模型。方法:选取2021年6月至2022年6月沈阳医学院2所附属医院收治的972例心血管疾病患者为研究对象,收集患者的社会人口学资料和疾病相关资料,采用决策树算法构建心血管疾病住院患者医院感染发生的风险预测模型。结果:医院感染发生率为29.6%(288/972);主要的感染部位为呼吸道(55.6%,160/288),泌尿系统(28.8%,83/288),胃肠道(13.9%,40/288),皮肤软组织(5.6%,16/288)。决策树算法分析结果显示:年龄≥65岁、饮酒史、手术史、高血压病史、糖尿病史是心血管疾病患者并发医院感染的独立危险因素。结论:应针对心血管疾病患者并发医院感染的影响因素采取有效的措施,以降低医院感染的感染率,提高患者的就医质量。

Objective:To construct a risk prediction model of nosocomial infection in inpatients with cardiovascular diseases.Methods:A total of 972 patients with cardiovascular diseases admitted to two affiliated hospitals of Shenyang Medical College from Jun 2021 to Jun 2022 were selected as the research objects,and the patients'socio-demographic data and disease-related data were collected.The risk prediction model of nosocomial infection was constrcted based on decision tree algorithm.Results:A total of 288 patients were complicated with nosocomial infection,the infection rate was 29.6%(288/972),and the main infection sites were respiratory tract(55.6%,160/288),urinary tract(28.8%,83/288),gastrointestinal tract(13.9%,40/288),skin and soft tissue(5.6%,16/288).The results of decision tree algorithm showed that age≥65,drinking history,surgical history,history of hypertension and history of diabetes were independent risk factors for nosocomial infection in patients with cardiovascular disease.Conclusion:Effective measures should be taken to reduce the rate of nosocomial infection in inpatients with cardiovascular disease and improve the quality of medical treatment.