白内障患者发生抑郁症的影响因素及风险预警平台的构建

Influencing factors for the occurrence of depression in patients with cataract and the establishment of risk warning platform

目的分析白内障患者发生抑郁症的影响因素,并构建白内障患者抑郁症发生风险预警平台。方法选取476例白内障患者,根据汉密尔顿抑郁量表评分将患者分为心理健康组和抑郁症组。使用基线信息采集表收集患者的基线信息,将基线信息指标纳入随机森林模型和Logistic回归模型筛选白内障患者发生抑郁症的影响因素。根据获得的影响因素构建白内障患者抑郁症发生风险模型,并使用列线图对模型进行可视化,采用一致性指数、受试者工作特征(ROC)曲线及校正曲线来评估模型的诊断效能和稳定性。使用R软件的Shiny包搭建白内障患者抑郁症发生风险预警平台。结果476例白内障患者中,共有49例患者被诊断为抑郁症。经单因素Logistic回归分析和随机森林模型分析筛选得到6个重要特征变量,包括年龄、烟酒史、入院时的视力水平、白内障并发症、就诊陪护、性别。多因素Logistic回归分析结果显示,年龄、性别、入院时的视力水平、就诊陪护、白内障并发症是影响白内障患者抑郁症发生风险的独立危险因素(均P<0.05)。根据上述5个独立危险因素及其相关系数构建白内障患者抑郁症发生风险预警模型,模型的一致性指数为0.783,ROC曲线下面积为0.712,校正曲线图显示预测曲线与理想曲线几乎一致,提示模型具有良好的预测效能和稳定性。基于此模型搭建的预警平台(http://ckr123.synology.me:3838/DRAS/)可快速提示白内障患者的抑郁症风险概率并提供相应的护理干预建议。结论年龄、性别、入院时的视力水平、就诊陪护及白内障并发症是影响白内障患者发生抑郁症的独立危险因素。根据上述变量构建的白内障患者抑郁症发生风险预警模型和预警平台具有较高的准确性和稳定性。

Objective To analyze the influencing factors for the occurrence of depression in patients with cataract,and to establish the risk warning platform for the occurrence of depression in patients with cataract.Methods A total of 476 patients with cataract were selected,and they were assigned to mental health group or depression group according to Hamilton Depression Scale score.The baseline information of patients was collected by using baseline information acquisition table.The baseline information indices were enrolled in random forest and Logistic regression models for screening influencing factors for the occurrence of depression in patients with cataract.The risk model of cataract patients suffering from depression was established according to the influencing factors acquired,and visualization was performed on the model by employing nomogram.The diagnostic efficiency and stability of the model were evaluated by employing the consistency index,receiver operating characteristic(ROC)curve,and calibration curve.The risk warning platform for the occurrence of depression in patients with cataract was established by using R software Shiny package.Results A total of 49 patients were diagnosed as depression among 476 cataract patients.Six important characteristic variables were screened by univariate Logistic regression analysis and random forest analysis,including age,history of smoking and alcohol drinking,visual acuity level on admission,cataract complications,medical attendant,and gender.The results of multivariate Logistic regression analysis revealed that age,gender,visual acuity level on admission,medical attendant,and cataract complications were independent risk factors for the occurrence of depression in patients with cataract(all P<0.05).According to the five aforementioned independent risk factors and their correlation coefficients,the risk warning model for the occurrence of depression in patients with cataract was established,and the consistency index of the model was 0.783,the area under the ROC curve was 0.712,as well as the calibration curve depicted that the prediction curve was almost the same as the ideal curve,which indicated that the model exerted favorable predictive efficiency and stability.The warning platform established based on the model as above(http://ckr123.synology.me:3838/DRAS/)could quickly prompt the risk probability of depression in cataract patients and provide corresponding nursing intervention suggestions.Conclusion Age,gender,visual acuity level on admission,medical attendant,and cataract complications are the independent risk factors in affecting the occurrence of depression in patients with cataract.The risk warning model and platform for the occurrence of depression in patients with cataract established based on variables as above exert relatively high accuracy and stability.