中青年脑卒中后抑郁的影响因素及其预测模型构建与验证

Influencing Factors of Post-Stroke Depression and its Prediction Model Construction and Validation in Young and Middle-aged Patients

目的 探讨中青年脑卒中后抑郁(post-stroke depression, PSD)的影响因素,构建预测模型并验证,以期为医护人员预警中青年PSD风险提供参考。方法 回顾性选取2017年1月至2020年12月苏州市某三级医院收治的329例中青年脑卒中患者为模型开发,以是否发生抑郁分为PSD组、非PSD组,分析中青年PSD影响因素并建立预测模型。采用便利抽样法选取2021年1月至2022年4月苏州市某三级医院收治的103例中青年脑卒中患者进行模型验证。结果 家庭人均月收入≤5000元、多灶病变、美国国立卫生研究院卒中量表评分高、卒中病耻感评分高和希望评分低是中青年PSD的影响因素(均P<0.05)。构建的预测模型的受试者工作特征曲线下面积为0.901,经验证模型预测的总准确率84.47%,拟合度较好(χ~2=2.734、P=0.542)。结论 医护人员应关注中青年脑卒中患者的家庭收入、病灶数、神经功能缺损程度、卒中病耻感及希望水平;构建的预测模型可预测PSD风险,值得推广。

Objective To explore the influencing factors of post-stroke depression(PSD)in young and middle-aged patients,to construct and verify a prediction model to provide medical staff reference for a early warning of PSD risk in young and middle-aged patient.Methods A total of 329 young and middle-aged stroke patients admitted to a tertiary A hospital in Suzhou from January 2017 to December 2020 were retrospectively selected as the model development,and were divided into PSD group and non-PSD group according to whether depression occurred.The influencing factors of PSD in young and middle-aged patients were analyzed and a prediction model was established.A total of 103 young and middle-aged stroke patients admitted to a tertiary A hospital in Suzhou from January 2021 to April 2022 were selected by convenience sampling method for model verification.Results Family per capita monthly income≤5000 yuan,multifocal disease,high stroke stigma score and low hope score were the influencing factors of PSD in young and middle-aged adults(all P<0.05).The area under the subject operating characteristic curve of the constructed prediction model was 0.901,and the total accuracy of the verified model was 84.47%,with good fit(χ2=2.734,P=0.542).Conclusions Medical staff should pay attention to the family income,number of lesions,degree of neurological impairment,stroke stigma and hope level of young and middle-aged stroke patients.The constructed prediction model can predict the risk of PSD and is worth promotion.