中老年人衰弱综合征危险因素分析及列线图预测模型的构建  

Analysis of risk factors and construction of line chart prediction model for middle-aged and elderly patients with frailty syndrome

作　　者：谢玉磊 沈炼伟 胡安丽 刘林茀 何雨洁 王寅旭[1] 吴青[1] XIE Yu-lei;SHEN Lian-wei;HU An-li;LIU Lin-fu;HE Yu-jie;WANG Yin-xu;WU Qing(Afiliated Hospital of North Sichuan Medical College,Nanchong,Sichuan 637000,China)

机构地区：[1]川北医学院附属医院康复科,四川南充637000 [2]首都医科大学康复医学院,中国康复研究中心,北京100068 [3]山东大学齐鲁医学院 [4]湖北经济学院

出　　处：《现代预防医学》2022年第24期4430-4436,共7页Modern Preventive Medicine

基　　金：四川省康复医学重点实验室开放课题(KFYXSZDSYS-04);四川南充市校合作科研专项基金(19SXHZ0103);川北医学院附属医院研发项目(2021ZD014)。

摘　　要：背景随着中国老龄化加剧及生活方式改变,中老年人疾病问题受到社会的广泛重视,衰弱综合征的患病率不断升高,分析筛选衰弱前期进展至衰弱综合征的危险因素及构建预测模型很有必要。目的探讨中国中老年人衰弱前期进展至衰弱综合征的危险因素及构建列线图预测模型,为中老年衰弱综合征的防控工作提供参考。方法研究数据来源于中国健康与养老追踪调查2015年随访数据,以是否患有衰弱综合征为因变量,纳入站立时间、体质量指数、腰围、身高、体重、亚洲人骨质疏松自我筛查工具(OSTA)指数、从椅子上起立时间、性别、年龄、吸烟、饮酒共11个变量探讨中老年人衰弱前期进展至衰弱综合征的相关因素,在SPSS上进行描述性分析,在Rstudio上将原始数据集分为训练集与验证集,训练集进行单因素、多因素回归分析、列线图模型构建及内部验证,验证集进行外部验证。结果共筛选中老年人1375人,其中检出有衰弱综合征505人,衰弱前期870人,构建原始数据集,以7:3将原始数据集随机分为训练集与验证集。模型的构建:根据训练集的二元logistic回归分析结果,选取出站立时间、从椅子上起立时间、OSTA指数等3个变量构建模型。模型的评价:模型ROC曲线下面积为0.757(95%CI:0.725~0.789),表明模型具有较高的区分度,校准曲线拟合良好说明模型具有较高的校准度。模型的内部验证:采用bootstrap法,生成的校准曲线拟合良好。模型的外部验证:由验证集检验,所得的ROC曲线下面积为0.787(95%CI:0.740~0.834),表明模型具有较高的区分度,校准曲线拟合良好说明模型具有较高的校准度,说明模型在验证队列也具有良好的效能。结论本研究构建的中老年人衰弱综合征预测模型具有较好的可信度,由此得出的列线图,可根据站立时间、从椅子上起立时间、OSTA指数等情况预测中老年人衰弱前期�Objective To explore the risk factors and construct a line chart prediction model of pre-frailty progression to frailty syndrome in middle-aged and elderly people in China,to provide reference for the prevention and control of frailty syndrome.Methods The research data were derived from the 2015 follow-up data of the China Health and Retirement Longitudinal Study.Taking whether suffering from frailty syndrome as dependent variables,11 variables including standing time,body mass index,waist circumference,height,weight,osteoporosis self-assessment tool for Asians(OSTA)index,standing time from chair,sex,age,smoking,and drinking were included to explore the related factors of pre-frailty progression to frailty syndrome in the middle-aged and elderly.Descriptive analysis was carried out using SPSS.The original data set was divided into training set and verification set in R studio.The training set was used for univariate,multivariate regression analysis,line chart model construction,and internal verification.The verification set was used for external validation.Results A total of 1375 middle-aged and elderly people were screened,of whom 505 had frailty syndrome and870 had pre-frailty.The original data set was constructed and then randomly divided into training set and verification set at the ratio of 7:3.According to the results of binary logistic regression analysis of the training set(modeling cohort),three variables including standing time,standing time from the chair,and OSTA index were selected to construct the model.The area under the ROC curve of the model was 0.757(95%CI:0.725-0.789),indicating that the model had a high degree of discrimination.The Bootstrop method was used to generate new samples after 1000 times of repeated sampling with replacement,and the generated calibration curves fit well.The area under ROC curve obtained by validation set(validation queue)was 0.787(95%CI:0.740-0.834),indicating that the model had a high degree of discrimination.Conclusion The prediction model of frailty syndrome in mi

关 键 词：中老年人 衰弱综合征 衰弱前期 列线图预测模型 

分 类 号：R493[医药卫生—康复医学]