护理研究2025，Vol.39Issue(14)：2327-2335,9.DOI:10.12102/j.issn.1009-6493.2025.14.002
基于机器学习算法构建老年重症病人静脉血栓栓塞症风险预测模型
Development of risk prediction model for venous thromboembolism in elderly critically ill patients based on machine learning
摘要
Abstract
Objective:To develop risk prediction model for venous thromboembolism(VTE)in elderly critically ill patients based on machine learning.Methods:A total of 909 elderly critically ill patients admitted to the intensive care units(ICU)in 3 tertiary grade A comprehensive hospitals in Shandong province were selected as study subjects from January 2020 to June 2023.And clinical data were collected.The patients were randomly divided into training set(636 cases)and validation set(273 cases)at 7∶3 ratio.The occurrence of VTE during ICU hospitalization was used as the outcome variable.Prediction models were constructed using 4 machine learning,namely random forest,extreme gradient boosting,support vector machines,and gradient boosting decision tree.Model performance was evaluated using metrics such as area under the curve(AUC)of receiver operator characteristic and Brier score,and the optimal model was selected.Interpretability analysis of the best-performing model was conducted using the SHAP algorithm.Results:Among the 909 elderly critically ill patients,258 developed VTE,with incidence of 28.4%.Among the 4 models,the random forest achieved the higher AUC(0.803),accuracy(0.733),sensitivity(0.662),and specificity(0.760),along with the lowest Brier Score(0.171).Conclusions:The risk prediction model for VTE in elderly critically ill patients developed based on random forest demonstrated strong predictive performance.It could provide reference for optimizing VTE management in elderly critically ill patients.关键词
机器学习/老年人/重症/静脉血栓栓塞症/预测模型/随机森林Key words
machine learning/elderly/critical illness/venous thromboembolism/prediction model/random forest引用本文复制引用
金杰,徐清,卢洁,赵佳月,张晴,孔杨,许红梅..基于机器学习算法构建老年重症病人静脉血栓栓塞症风险预测模型[J].护理研究,2025,39(14):2327-2335,9.基金项目
山东省自然科学基金面上项目,编号:ZR2023MH378 （）
