摘要
目的 构建前庭性眩晕患者伴发抑郁危险预测模型并验证。方法 选取341例前庭性眩晕患者,将患者按照7∶3比例随机分为建模组(239例)与验证组(102例)。根据患者是否出现抑郁将建模组分为抑郁组(66例)和非抑郁组(173例)。前庭性眩晕患者发生抑郁的影响因素采用多因素Logistic回归分析;前庭性眩晕患者发生抑郁的列线图模型采用R软件构建。绘制ROC曲线评估前庭性眩晕患者发生抑郁列线图模型的区分度;模型一致性绘制校准曲线评估;使用临床决策曲线(DCA)对该模型的临床应用价值进行评估。结果 239例患者中有66例发生抑郁,发生率为27.62%。抑郁组和非抑郁组在年龄、病程、学历、DHI评分、是否独居方面比较差异均有统计学意义(P<0.05)。多因素Logistic回归分析结果显示,年龄≥60岁、病程≥4周、小学及以下、DHI评分≥60分、独居是前庭性眩晕患者发生抑郁的危险因素(P<0.05)。建模组ROC曲线下面积(AUC)为0.968,校准曲线斜率与1接近,H-L检验为χ^(2)=6.703,P=0.710,一致性良好。外部验证AUC为0.938,校准曲线斜率与1接近,H-L检验为χ^(2)=6.537,P=0.692,一致性良好。DCA曲线提示高风险阈值概率在0.04~0.91时,该列线图模型预测前庭性眩晕患者发生抑郁的临床使用价值较高。结论 年龄、病程、学历、DHI评分、是否独居是前庭性眩晕患者发生抑郁的危险因素,以此构建的列线图模型区分度与一致性良好,可预测前庭性眩晕患者发生抑郁的风险。
Objective To construct a predictive model for the risk of depression in patients with vestibular vertigo and validate it. Methods A total of 341 patients with vestibular vertigo were selected and randomly divided into modeling group(239 cases) and validation group(102 cases) in a 7∶3 ratio. The modeling group was divided into depression group(66 cases) and non-depression group(173 cases) based on whether the patient experienced depression. The influencing factors of depression in patients with vestibular vertigo were analyzed using multivariate logistic regression. The nomogram model of depression in patients with vestibular vertigo was constructed using R software. ROC curve was plotted to evaluate the discriminability of the nomogram model, and calibration curve was used to evaluated the consistency of the model. Decision Curve Analysis(DCA) was applied to evaluate the clinical application value of the model. Results Out of 239 patients in this study, 66 developed depression, with an incidence rate of 27.62%. There were significant differences between depression group and non-depression group in terms of age, disease course, education level, DHI score, and whether living alone(P<0.05). Multivariate logistic regression revealed age≥60 years old, disease duration≥4 weeks, primary school and below, DHI score≥60 points, and living alone were risk factors for depression in patients with vestibular vertigo(P<0.05). The area under the ROC curve(AUC) of modeling group was 0.968, and the slope of the calibration curve was close to 1. H-L test showed χ~2=6.703, P=0.710, and the consistency was good. The external validation AUC was 0.938, and the calibration curve slope was close to 1. H-L test showed χ~2=6.537, P=0.692, and the consistency was good. DCA curve indicated that when the probability of high-risk threshold was between 0.04 and 0.91, the nomogram model had high clinical value in predicting depression in patients with vestibular vertigo. Conclusion Age, disease course, education level, DHI score, and w
作者
高红玉
贡冉
毋少华
王胜利
苏琼琼
GAO Hongyu;GONG Ran;WU Shaohua(Henan University of Science and Technology Affiliated Yellow River Hospital,The Huanghe Sanmenxia Hospital,Sanmenxia 472000,China)
出处
《精神医学杂志》
2023年第6期605-609,共5页
Journal of Psychiatry
基金
河南省医学科技攻关计划项目(编号:LHGJ20200779)。
