基于模型预测的老年糖尿病视网膜病变患者视力损害因素分析  

Construction and verification of visual disability risk prediction model in patients with elderly diabetic retinopathy

作　　者：张雅萍 钱莉 郑宏强 申玉英[1,4] ZHANG Yaping;QIAN Li;ZHENG Hongqiang;SHEN Yuying l(Nanjing Medical University,Jiangsu Nanjing 211166,China;Department of Jiangning,Baijiahu Community Health Service Center,Jiangsu Nanjing 211101,China;Sir Run Run Hospital Nanjing Medical University,Jiangsu Nanjing 211101,China;Medical Department of Nanjing Hospital Affiliated to Nanjing Medical University(Nanjing First Hospital),Jiangsu Nanjing 211100,China)

机构地区：[1]南京医科大学,江苏南京211166 [2]南京市江宁区百家湖社区卫生服务中心,江苏南京211101 [3]南京医科大学附属逸夫医院,江苏南京211101 [4]南京医科大学附属南京医院(南京市第一医院),江苏南京211100

出　　处：《中国实验诊断学》2024年第12期1392-1397,共6页Chinese Journal of Laboratory Diagnosis

基　　金：南京市卫生科技发展专项资金资助项目(YKK22202)。

摘　　要：目的探究老年糖尿病视网膜病变患者(DR)发生视力残疾的相关因素。方法选取2020年5月至2023年12月就诊的糖尿病视网膜病变患者,按时间顺序前198名作为建模组,后184名作为验证组。结果建模组DR患者中视力残疾患病率为43.0%。视力残疾与性别、空腹血糖、糖尿病病程、共患疾病、尿蛋白和眼压等因素有关(P<0.05)。多因素Logistic回归表明上述因素是发生视力残疾的独立危险因素。Hosmer-Lemeshow拟合优度检验表明建模组与验证组中模型拟合较好(χ^(2)=9.33,P=0.31;χ^(2)=11.21,P=0.19)。ROC曲线下面积建模组为0.919,最大约登指数0.674,灵敏度0.782,特异度0.892;验证组为0.841,最大约登指数0.564,灵敏度为0.827,特异度为0.737。结论该研究建立的模型能有效预测糖尿病视网膜病变患者视力残疾的发生,为临床早期诊断及干预提供价值。Objective To investigate the factors associated with the occurrence of visual disability in elderly patients with diabetic retinopathy(DR).Methods Patients with diabetic retinopathy who attended the clinic from May 2020 to December 2023 were enrolled,and the first 198 patients were included in the modelling group and the rest 184 patients were included in the validation group according to the chronological order of the clinic visit.Results The prevalence of visual disability among DR patients in the modelling group was 43.0%.Visual disability was associated with gender,fasting glucose,duration of diabetes mellitus,co-morbidities,urinary protein and intraocular pressure(P<0.05).Multifactorial logistic regression indicated that the above factors were independent risk factors for the development of visual disability.Hosmer-Lemeshow goodness-of-fit test showed a better model fit in the modelling group versus the validation group(χ^(2)=9.33,P=0.31;χ^(2)=11.21,P=0.19).The area under the ROC curve was 0.919 in the modelling group,with a maximal Yoden index of 0.674,sensitivity 0.782,and specificity 0.892;the validation group was 0.841,maximum Jordon's index 0.564,sensitivity 0.827,and specificity 0.737.Conclusion The model established in this study can effectively predict the occurrence of visual disability in patients with diabetic retinopathy,and provide value for early clinical diagnosis and intervention.

关 键 词：老年糖尿病 视网膜病变 预测模型 列线图 

分 类 号：R587.1[医药卫生—内分泌] R774.1[医药卫生—内科学]