ICU患者医院感染logistic回归模型预测  

Prediction of nosocomial infection among ICU patients with model of logisticregression

作　　者：谢多双[1,2] 来瑞平[2] 符湘云[2] 王慧芳[2] 聂绍发[1] 

机构地区：[1]华中科技大学同济医学院公共卫生学院,湖北武汉430030 [2]湖北医药学院附属太和医院,湖北十堰442000

出　　处：《中华医院感染学杂志》2011年第12期2424-2426,共3页Chinese Journal of Nosocomiology

基　　金：十堰市科学技术研究与开发项目资助(2010st16)

摘　　要：目的构建ICU患者医院感染logistic回归预测模型,并对模型进行评价。方法以入住ICU>48 h的患者为研究对象,构建医院感染logistic回归模型,对模型进行拟和优度检验、ROC曲线下面积分析。结果入住ICU天数、气管插管、前列腺肥大、动静脉插管、基础疾病(肿瘤)等变量进入logistic回归方程,模型ROC曲线下面积为0.856。结论 logistic回归模型对ICU患者医院感染预测拟合度较好。OBJECTIVE To develop a predictive model for the healthcare-associated hospital infection in ICU and to evaluate the model′s effectiveness.METHODS Data of patients who had been hospitalized in the ICU for more than 48h were collected to construct the model of logistic regression.The fitness was tested and the area under ROC was calculated.RESULTS The follow variables were enter the logistic regression equation： the period of the patients in ICU,tracheal intubation,BPH,artery and venous cannula,tumor as the foundation disease,et al.The area under ROC was 0.856.CONCLUSION The model of logistic regression is effective in predicting the healthcare-associated infection among ICU patient.

关 键 词：LOGISTIC回归 医院感染 预测 重症监护病房 

分 类 号：R181.32[医药卫生—流行病学]