长航官兵抑郁情绪发生风险预测模型的构建及验证

Construction and validation of predictive model for depression occurrence of navy serviceman during long voyage

目的构建基于睡眠、疲劳、体质等因素的长航官兵抑郁情绪发生风险预测模型并进行验证。方法采用随机整群抽样法,抽取2022年6—9月的1000名长航官兵,应用抑郁自评量表(self-rating depression scale,SDS)、匹兹堡睡眠质量指数(Pittsburgh sleep quality index,PSQI)量表、疲劳指数-14(fatigue scale-14,FS-14)量表、中医体质量表进行抑郁情绪及影响因素调查;采用logistics逐步回归分析方法筛选危险因素并构建列线图预测模型;采用曲线下面积、校准曲线、决策曲线对模型进行内部验证及效能分析。结果共回收有效量表949份,有效率为94.9%。在航行前、航行15天及航行30天后,舰艇官兵抑郁状态的检出率分别为14.33%、26.34%、19.18%,SDS评分随着航行时间的延长先增加后减少,组间差异有统计学意义(P<0.05),与国内常模比较差异均有统计学意义(均P<0.01)。筛选得到性别、年龄、工作年限、距上次长航任务时间、气虚质、痰湿质、血瘀质、气郁质、特禀质、睡眠障碍、疲劳状态11个变量可能是舰艇官兵长航15天后发生抑郁的独立危险因素,并构建出列线图预测模型,该模型曲线下面积为0.871(95%CI:0.840~0.901),校准曲线显示出良好的拟合性,决策曲线得到阈概率在5%~87%区间时,该模型均具有正的净效益。结论长航官兵抑郁情绪发生风险预测模型具有良好的预测精度和效能,为筛查及预防性干预长航官兵不良情绪提供了有效的评估工具。

Objective To construct and validate a risk predictive model for depression occurrence of navy serviceman during long voyage based on factors such as sleep,fatigue and constitution of traditional Chinese medicine(TCM).Methods A total of 1000 navy serviceman during long voyage from June 1 to September 30,2022 were selected as the subjects by random cluster sampling method.All of them were assessed by the self-rating depression scale(SDS),Pittsburgh sleep quality index(PSQI),fatigue scale-14(FS-14),TCM constitution scale to investigate the depressive emotions and influencing factors.The logistic stepwise regression analysis was used to screen risk factors and construct prediction nomogram model.The area under the curve,correction curve and decision curve were used for internal validation and predictive efficacy assessment.Results Totally 949 valid scales were collected,with an effective rate of 94.9%.Before the voyage,15 days and 30 days after voyage,the incidence of depression in navy serviceman was 14.33%,26.34%and 19.18%,respectively.The scores of SDS increased at first and then decreased with the increasing of the voyage time,with statistically significant differences among groups(P<0.05).There were statistically significant differences compared to the domestic norm(all P<0.01).There were 11 independent risk factors correlated with depression occurrence of navy serviceman 11 days after voyage,including sex,age,length of work,time from the last long voyage,Qi-deficiency,Phlegmdampness,Blood-stasis,Qi-stagnation,Special-constitution,sleep disorder and fatigue.A nomogram model was established according to the selected variables,whose area under the curve was 0.871(95%CI:0.840-0.901).The correction curve showed good fit.The decision curve showed that the model had a positive net benefit when the threshold probability was 5%-87%.Conclusion The predictive model for depression occurrence of navy serviceman during long voyage has good predictive accuracy and efficiency,which provides an effective assessment tool for screening high-risk cases of depression and implementing preventive interventions in navy serviceman during long voyage.