心力衰竭患者抑郁风险预测模型的构建与验证  

Construction and validation of a predictive model for depression risk in heart failure patients

作　　者：李艳[1] 刘大军 房信响 史云[1] LI Yan;LIU Dajun;FANG Xinxiang;SHI Yun(Department of Cardiology,Tianjin Fourth Central Hospital,Tianjin 300142,China;Department of Respiratory and Critical Care Medicine,Tianjin Fourth Central Hospital,Tianjin 300142,China;Department of Clinical Medicine,First Clinical College of Changsha Medical College,Changsha 300240,China)

机构地区：[1]天津市第四中心医院心内科,天津300142 [2]天津市第四中心医院呼吸与危重症医学科,天津300142 [3]长沙医学院第一临床学院临床医学系,长沙300240

出　　处：《天津医科大学学报》2024年第5期434-439,共6页Journal of Tianjin Medical University

基　　金：天津市卫生健康科技项目(TJWJ2022MS022);天津市卫生健康科技项目(TJVJ2021MS024)。

摘　　要：目的:分析心力衰竭(HF)与抑郁的相关因素,基于临床特征建立预测抑郁情绪模型并进行验证。方法:随机抽取2020年1月至2022年12月在天津市第四中心医院就诊的399例HF患者作为研究对象。入院时收集患者的基线信息。使用医院焦虑和抑郁量表(HADS)评估并定义抑郁症状。研究队列按照随机数字表法随机分为训练集和验证集,比例为7:3。利用Logistic回归模型探索与抑郁相关的因素。构建一个列线图来发展预测模型。应用校准曲线评估列线图模型的准确性。通过受试者工作特征(ROC)曲线和决策曲线分析评估预测性能。结果:共140例HF患者(35.09%)出现抑郁症状。多变量Logistic回归分析发现,年龄(OR=0.937,95%CI:0.911~0.964)、婚姻状况(OR=2.466,95%CI:1.099~5.536)、水肿(OR=1.866,95%CI:1.043~3.337)、体力活动水平(OR=1.942,95%CI:1.200~3.143)、生活质量(OR=1.026,95%CI:1.013~1.040)是HF患者抑郁的预测因子。利用以上因素构建的列线图模型显示出良好的一致性和准确性。预测模型和内部验证组的曲线下面积(AUC)分别为0.814(95%CI:0.763~0.864)和0.915(95%CI:0.866~0.964)。Hosmer-Lemeshow检验的P值分别为0.212和0.486(均大于0.05)。结论:年龄、婚姻状况、水肿、体力活动水平、生活质量是HF患者抑郁的预测因子。本抑郁风险预测模型在HF患者中显示出良好的预测性能。Objective:To analyze the factors associated with depression in patients hospitalized with heart failure(HF),and to estab-lish and validate a model for predicting depressive mood based on clinical characteristics.Methods:A total of 399 patients diagnosed with HF were randomly selected from Tianjin Fourth Central Hospital from January 2020 to December 2022.Baseline information was collected at admission.Depressive symptoms were assessed and defined using the Hospital Anxiety and Depression Scale(HADS).The study cohort was randomized into training set and validation set using a random number table method,with a ratio of 7:3.Logistic regression models were used to explore factors associated with depression.A column-line diagram was constructed to develop the pre-dictive model.Calibration curves were applied to assess the accuracy of the column-line diagram model.Predictive performance was as-sessed by receiver operating characteristics(ROC)curve and decision curve analysis.Results:A total of 140 patients with HF(35.09%)showed depressive symptoms.Multivariate Logistic regression analysis identified age(OR=0.937,95%CI:0.911-0.964),marital status(OR=2.466,95%CI:1.099-5.536),edema(OR=1.866,95%CI:1.043-3.337),physical activity level(OR=1.942,95%CI:1.200-3.143),and quality of life(OR=1.026,95%CI:1.013-1.040)as predictive factors of depression in patients with HF.These factors were used to construct a column-line graphical model that showed good agreement and accuracy.The area under the curve(AUC)for the predictive model and internal validation group were 0.814(95%CI:0.763-0.864)and 0.915(95%CI:0.866-0.964),respectively.The P-values for the Hosmer-Lemeshow test were 0.212 and 0.486(both greater than 0.05).Conclusion:Age,marital status,edema,physical activity level,and quality of life are predictors of depression in patients with HF.This predictive model for depression risk showes good predictive performance in patients with HF.

关 键 词：心力衰竭 抑郁 列线图 风险预测模型 

分 类 号：R541.6[医药卫生—心血管疾病]