认知障碍患者抑郁风险预测模型的构建及验证  

Construction and validation of a depression risk prediction model for patients with cognitive impairment

作　　者：廖黎 李雪芬[1] 石静萍[1] 李晓芳 谈莉莉 叶晨 康艳[1] Liao Li;Li Xuefen;Shi Jingping;Li Xiaofang;Tan Lili;Ye Chen;Kang Yan(Neurology Department,the Affiliated Brain Hospital of Nanjing Medical University,Nanjing 210029,China)

机构地区：[1]南京医科大学附属脑科医院神经内科,南京210029

出　　处：《中华现代护理杂志》2023年第20期2701-2707,共7页Chinese Journal of Modern Nursing

基　　金：南京医科大学科技发展基金项目(NMUB2020244)。

摘　　要：目的探讨认知障碍患者发生抑郁的危险因素并构建预测模型,初步验证该模型的预测效果,旨在为医护人员提供高危人群的筛选工具。方法于2020年1月—2021年12月,采用便利抽样法选取南京医科大学附属南京脑科医院收治的1130例认知障碍患者为研究对象,以7︰3比例将研究对象分为建模组(n=791)与验证组(n=339)。采用二项Logistic回归分析探讨认知障碍患者发生抑郁的影响因素并建立风险预测模型,采用受试者工作特征(ROC)曲线检验预测模型的预测效能。结果1130例认知障碍患者的抑郁发生率为51.3%(580/1130)。二项Logistic回归分析结果显示,患者发生抑郁的影响因素包括年龄、日常生活自理能力量表评分、汉密尔顿焦虑量表评分、匹兹堡睡眠质量指数评分、路易体复合风险评分(P<0.05)。建模组的ROC曲线下面积为0.921,Youden指数为0.716,敏感度为0.834,特异度为0.882,预测正确率为0.858;验证组的ROC曲线下面积为0.896,Youden指数为0.651,敏感度为0.824,特异度为0.827,预测正确率为0.825。结论本研究构建的认知障碍患者抑郁风险预测模型能较好地预测认知障碍患者的抑郁发生风险,可为医护人员提供高危人群的筛选工具。Objective To explore the risk factors for depression in patients with cognitive impairment and construct a prediction model to preliminarily validate the predictive performance of the model,aiming to provide medical and nursing staff with a screening tool for high-risk groups.MethodsFrom January 2020 to December 2021,convenience sampling was used to select 1130 patients with cognitive impairment admitted to the Affiliated Brain Hospital of Nanjing Medical University as the research subject.The research subjects were divided into a modeling group(n=791)and a validation group(n=339)at a ratio of 7∶3.The influencing factors of depression in patients with cognitive impairment were determined using binomial Logistic regression and a risk prediction model was established.The predictive performance of the prediction model was tested using the receiver operating characteristic(ROC)curve.ResultsThe incidence of depression in 1130 patients with cognitive impairment was 51.3%(580/1130).Binomial Logistic regression analysis showed that the influencing factors for depression in patients included age,Activities of Daily Living Scale score,Hamilton Anxiety Scale score,Pittsburgh Sleep Quality Index score,and Lewy Body Composite Risk score(P<0.05).In the modeling group,the area under the ROC curve was 0.921,the Youden index was 0.716,the sensitivity was 0.834,the specificity was 0.882,and the prediction accuracy was 0.858.In the validation group,the area under the ROC curve was 0.896,the Youden index was 0.651,the sensitivity was 0.824,the specificity was 0.827,and the prediction accuracy was 0.825.ConclusionsThe depression risk prediction model can effectively predict the risk of depression in patients with cognitive impairment,and can provide a screening tool for high-risk groups for medical and nursing staff.

关 键 词：抑郁 认知障碍 危险因素 预测模型 

分 类 号：R749.1[医药卫生—神经病学与精神病学] R749.4[医药卫生—临床医学]