青少年抑郁症患者继发性失眠的Nomogram模型构建与验证

Construction and validation of Nomogram model for secondary insomnia after adolescent depression

目的构建青少年抑郁症患者继发性失眠的Nomogram预测模型,为临床预防青少年抑郁症患者继发性失眠提供参考和依据。方法便利抽样法选取2021年5月至2022年3月在绍兴市第七人民医院儿童精神科接受治疗的158例青少年抑郁症患者作为研究对象,根据匹兹堡睡眠质量指数问卷(PSQI)得分分为睡眠正常组(PSQI≤7分)和继发性失眠组(PSQI>7分)。采用汉密尔顿抑郁量表、汉密尔顿焦虑量表、社会支持评定量表对其进行横断面调研。使用多因素Logistic回归分析继发性失眠发生的影响因素并构建青少年抑郁症患者继发性失眠的Nomogram预测模型,模型准确度以ROC曲线判断。另取2022年5月至2023年3月在绍兴市第七人民医院接受治疗的70例青少年抑郁症患者进行成功率预测,与实际观测到的结果进行比较,并通过Hosmer-Lemeshow检验评价该模型的拟合度。结果睡眠正常组72例,其中男40例,女32例,年龄(15.34±2.62)岁,继发性失眠组86例,其中男51例,女35例,年龄(14.89±2.13)岁。多因素分析结果显示,抑郁(OR=6.381,95%CI 1.548~26.295,P<0.05)、焦虑(OR=6.248,95%CI 1.445~27.005,P<0.05)及社会支持(OR=0.586,95%CI 0.346~0.994,P<0.05)为继发性失眠是否发生的独立影响因素,在此基础上建立风险预测模型,受试者工作特征曲线下面积为0.999,敏感度为98.80%,特异度为98.60%,约登指数为0.974。外部验证模型预测总准确率为99.80%。结论青少年抑郁症患者继发性失眠的Nomogram预测模型具有极好的预测效能,抑郁程度、焦虑程度及社会支持均是患者产生继发性失眠的主要影响因素,医护人员应为患者提供更有效的心理支持。

Objective The Nomogram prediction model of secondary insomnia in adolescent depression patients was constructed to provide reference and basis for clinical prevention of secondary insomnia in adolescent depression patients.Methods A total of 158 adolescent depression patients who were treated in the Department of Pediatric Psychiatry of Shaoxing Seventh People′s Hospital from May 2021 to March 2022 were selected as the research objects by convenient sampling method.According to the Pittsburgh Sleep Quality Index(PSQI)score,the group was divided into a normal sleep group(PSQI≤7 points)and a secondary insomnia group(PSQI>7 points).A cross-sectional survey was conducted using the Hamilton Depression Scale,Hamilton Anxiety Scaleand Social Support Rating Scale.Multivariate binary Logistic regression was used to analyze the influencing factors of secondary insomnia and construct a Nomogram prediction model for secondary insomnia in adolescents with depression.The accuracy of the model was judged by receiver operating characteristic curve.In addition,70 adolescent depression patients who were treated in Shaoxing Seventh People′s Hospital from May 2022 to March 2023 were selected for success rate prediction.The results were compared with the actual observed results,and the fitting degree of the model was evaluated by Hosmer-Lemeshow test.Results There were 72 cases in the normal sleep group,including 40 males and 32 females,aged(15.34±2.62)years old,and 86 cases in the secondary insomnia group,including 51 males and 35 females,aged(14.89±2.13)years old.The results of multivariate analysis showed that depression(OR=6.381,95%CI 1.548-26.295,P<0.05),anxiety(OR=6.248,95%CI 1.445-27.005,P<0.05)and social support(OR=0.586,95%CI 0.346-0.994,P<0.05)were independent influencing factors for the occurrence of secondary insomnia in the two groups.On this basis,a risk prediction model was established.The AUC was 0.999,the sensitivity was 98.80%,the specificity was 98.60%,and the Youden index was 0.974.The total prediction accuracy of the external validation model was 99.80%.Conclusions The Nomogram prediction model of secondary insomnia in adolescent depression patients has excellent predictive efficacy.The degree of depression,anxiety and social support are the main influencing factors of secondary insomnia in patients.Medical staff should provide more effective psychological support for patients.