老年病人缺血性脑卒中后衰弱预测模型的前瞻性研究  

A prospective study of prediction model for frailty after ischemic stroke in elderly patients

作　　者：张玲慧 王亚喜 陈晨[1] 杜钦霞[2] 庞旭峰[2] ZHANG Linghui;WANG Yaxi;CHEN Chen;DU Qinxia;PANG Xufeng(School of Nursing,Qingdao University,Shandong 266021 China)

机构地区：[1]青岛大学护理学院,266021 [2]青岛大学附属医院,266000

出　　处：《全科护理》2024年第8期1398-1404,共7页Chinese General Practice Nursing

摘　　要：目的:调查老年缺血性脑卒中病人发生衰弱的影响因素,构建衰弱风险决策树预测模型。方法:选取2022年7月-2023年4月青岛市某三级甲等医院神经内科收治的485例老年缺血性脑卒中病人,其中441例病人完成随访,根据出院后3个月内是否发生衰弱分为衰弱组(134例)和非衰弱组(307例)。采用单因素分析、Logistic回归分析老年缺血性脑卒中病人发生衰弱的影响因素。采用Python软件构建预测老年缺血性脑卒中病人发生衰弱的决策树模型。结果:脑卒中次数(OR=5.899)、美国国立卫生研究院卒中量表(NIHSS)评分(OR=2.150)、老年抑郁(OR=1.673)、握力(OR=0.921)、Barthel指数(OR=0.954)、一般自我效能感(OR=0.797)、社会支持(OR=0.860)是老年缺血性脑卒中病人发生衰弱的重要影响因素。分类回归树算法(CART)决策树模型结果显示,NIHSS评分、老年抑郁、握力、一般自我效能感和社会支持是老年缺血性脑卒中病人发生衰弱的预测因子,其中对模型贡献性高的变量前3名依次为NIHSS评分(45.342%)、老年抑郁(26.124%)、握力(15.297%)。决策树模型性能优秀,在训练集和测试集上的ROC曲线下面积分别达到0.94,0.92。结论:通过CART决策树模型可方便、准确地筛选出发生衰弱的高危病人,可为制定针对性的干预措施提供依据,避免老年缺血性脑卒中病人衰弱发生。Objective:To investigate the influencing factors of frailty after ischemic stroke in elderly patients and construct frailty risk decision tree prediction model.Methods:A total of 485 elderly patients with ischemic stroke admitted to the department of neurology in a third grade A hospital in Qingdao from July 2022 to April 2023.Among them,441 patients completed follow-up and were divided into a frailty group(134 cases)and a non-frailty group(307 cases)according to whether frailty occurred within 3 months after discharge.Univariate analysis and Logistic regression were used to analyze the influencing factors of frailty in elderly patients with ischemic stroke.Python software was used to construct a decision tree model for predicting frailty in elderly patients with ischemic stroke.Results:Stroke frequency(OR=5.899),National Institutes of Health Stroke Scale(NIHSS)score(OR=2.150),geriatric depression(OR=1.673),grip strength(OR=0.921),Barthel index(OR=0.954),general self-efficacy(OR=0.797),social support(OR=0.860)were the important factors affecting the debilitation of elderly patients with ischemic stroke.The results of CART decision tree model showed that NIHSS score,geriatric depression,grip strength,general self-efficacy and social support were predictive factors of occurrence of frailty in elderly patients with ischemic stroke.The top three variables with high contribution to the model were NIHSS score(45.342%),geriatric depression(26.124%)and grip strength(15.297%).The performance of the decision tree model is excellent,and the area under the ROC curve on the training set and the test set reach 0.94 and 0.92 respectively.Conclusions:CART decision tree model can easily and accurately screen out high-risk patients with frailty,which can provide basis for formulating targeted intervention measures to avoid frailty in elderly patients with ischemic stroke.

关 键 词：缺血性脑卒中 衰弱 预测模型 决策树 

分 类 号：R473.74[医药卫生—护理学]