心脑血管病防治2025，Vol.25Issue(12)：39-44,58,7.DOI:10.3969/j.issn.1009-816x.2025.12.008
机器学习在心脏外科术后谵妄风险预测模型中的应用进展
Advances in the application of machine learning for risk prediction models of postoperative delirium after cardiac surgery
摘要
Abstract
The incidence of postoperative delirium(POD)after cardiac surgery is high(18%-42%),and is closely associated with adverse outcomes such as cognitive decline,prolonged hospitalisation and increased mortality,posing significant clinical risks.However,approximately 30%-40%of cases of delirium are preventable,which makes the early and precise identification of high-risk patients critically important.Traditional prediction models face issues such as limited predictive factors and incomplete performance evaluation.Machine learning offers a new approach to constructing high-performance POD prediction models,enabling the more effective identification of potential risk factors and enhancing prediction accuracy and sensitivity.This article systematically reviews the fundamental concepts and classification systems of machine learning,focuses on exploring algorithm selection strategies in developing POD prediction models for cardiac surgery and conducting multidimensional comparative analyses of the key stages in prediction models.The review also outlines future research directions in this field,aiming to provide references for clinical healthcare professionals in formulating risk prediction and management strategies for delirium in cardiac surgery patients.关键词
机器学习/心脏外科/谵妄/预测模型/综述Key words
Machine learning/Cardiac surgery/Delirium/Predictive modeling/Review引用本文复制引用
陈晶立,胡琼,范琴,武秀丽,耿丽..机器学习在心脏外科术后谵妄风险预测模型中的应用进展[J].心脑血管病防治,2025,25(12):39-44,58,7.基金项目
职业危害识别与控制湖北省重点实验室联合基金(JF2023-Y05) （JF2023-Y05）
武汉市卫生健康科研基金(WX21D13) （WX21D13）
