基于机器学习算法构建泌尿肿瘤外科老年患者术后谵妄风险预测模型

Construction of prediction model of postoperative delirium in elderly patients of urological oncology based on machine learning

目的基于机器学习算法建立泌尿肿瘤外科老年患者术后谵妄(POD)预测模型。方法选取2021年1月至2024年2月福建省肿瘤医院泌尿外科治疗的1180例老年患者作为研究对象,按照7∶3比例分为建模组和验证组。采用最小绝对收缩和选择算子(LASSO)筛选预测因子,建立6种预测模型并评估预测准确性。结果所有患者中POD发生率为9.4%。建模组中,LASSO回归筛选出5个非零系数变量,分别为年龄、糖尿病、美国麻醉师协会(ASA)分级、术前白蛋白、手术时间。建模组和验证组中随机森林模型的AUC值最高。结论基于随机森林模型建立老年泌尿外科患者POD的准确性最高,有助于医护人员筛选高危患者。

Objective Construction of prediction model of postoperative delirium(POD)in elderly patients of urological oncology based on machine learning.Methods A total of 1180 elderly patients treated in the Department of Urology in Fujian Cancer Hospital from January 2021 to February 2024 were selected as the research objects,and they were divided into modeling group and validation group according to the ratio of 7:3.The least absolute shrinkage and selection operator(LASSO)was used to screen the predictors,and 6 kinds of prediction models were established and the prediction accuracy was evaluated.Results POD incidence was 9.4%in all patients.In the modeling group,LASSO regression screened out 5 non-zero coefficient variables,namely age,diabetes,American Society of Anesthesiologists(ASA)grade,preoperative albumin,and operation time.The AUC values of random forest model were the highest in modeling group and validation group.Conclusion The accuracy of POD establishment in elderly urological patients based on random forest model is the highest,which is helpful for medical staff to screen high-risk patients.