长航官兵疲劳发生风险预测模型的构建及验证  

Construction and validation of predictive model for fatigue in soldiers during long voyage

作　　者：戚秀中 于磊 王剑 吴汇鑫 周丰宝 Qi Xiuzhong;Yu Lei;Wang Jian;Wu Huixin;Zhou Fengbao(Department of Traditional Chinese Medicine,Qingdao Special Service Sanatorium of Navy,Qingdao 266071,China)

机构地区：[1]海军青岛特勤疗养中心中医科,山东青岛266071

出　　处：《海军医学杂志》2025年第2期138-143,共6页Journal of Navy Medicine

基　　金：军队中医药服务能力培育与提升专项计划(2021ZY016)

摘　　要：目的构建基于睡眠、情绪、体质等因素的长航官兵疲劳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1000名长航官兵,应用疲劳指数(FS-14)量表、匹兹堡睡眠质量指数(PSQI)量表、焦虑自评量表(SAS)、抑郁自评量表(SDS)、中医体质量表等进行疲劳状态及影响因素调查。采用logistic逐步回归分析方法筛选危险因素并构建列线图预测模型。采用受试者操作特征曲线下面积(AUC)、校准曲线、Hosmer-Lemeshow拟合优度检验、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935份,有效率93.5%。排除航行前已存在疲劳的306名长航官兵,629名官兵在航行15、30 d后新增疲劳状态的检出率分别为26.39%、42.13%,FS-14量表总分及躯体、脑力疲劳得分均随着航行时间的延长而增加,组间差异具有统计学意义(P<0.01)。中医体质量表、SAS、SDS、PSQI量表筛选得到气虚质、阳虚质、阴虚质、气郁质、特禀质、焦虑状态、抑郁状态、睡眠障碍是长航官兵航行30 d后发生疲劳的独立危险因素,并构建出列线图预测模型,该模型AUC为0.828(95%CI:0.797~0.858),校准曲线及Hosmer-Lemeshow检验显示模型拟合优度较好(χ^(2)=15.384,P=0.052),决策曲线得到阈概率在10%~95%时该模型具有正的净效益。结论长航官兵疲劳发生风险预测模型具有良好的预测精度和效能,为开展长航官兵疲劳发生风险筛查及预防性干预提供了有效的评估工具。Objective To construct and validate a risk predictive model for fatigue in soldiers during long voyage based on sleep,mood and constitution.Methods One thousand soldiers during long voyage were selected as research objects by random cluster sampling.The fatigue scale-14(FS-14),Pittsburgh sleep quality index(PSQI),self-rating anxiety scale(SAS),self-rating depression scale(SDS),and TCM constitution scale were used to assess the occurrence of fatigue and its influence factors.Logistic stepwise regression analysis was used to screen variables,and the prediction nomogram model was constructed based on the risk factors of fatigue.The area under the receiver operator characteristic curve(AUC),calibration curve,Hosmer-Lemeshow test and decision curve were used for internal validation and predictive efficacy assessment.Results A total of 935 effective scales were collected,with an effective rate of 93.5%.Excluding 306 soldiers who were already fatigued before voyage,the fatigue incidence rate of 629 soldiers after 15 days and 30 days of voyage were 26.39% and 42.13%,respectively.The total score of FS-14 and the scores of physical fatigue and mental fatigue were significantly proportional to the voyage time(P<0.01).Eight variables including Qi-deficiency,Yang-deficiency,Yin-deficiency,Qi-stagnation,Special-constitution,anxiety,depression,and sleep disorder might be the independent risk factors for the occurrence of fatigue 30 days after voyage.A nomogram model was established according to the selected variables,and the AUC of the model was 0.828(95%CI:0.797-0.858).Calibration curve and Hosmer-Lemeshow test showed that the goodness of fit of the model was good(χ^(2)=15.384,P=0.052).Decision curve showed that the model had a positive net benefit when the threshold probability was 10%-95%.Conclusion The predictive model for fatigue in soldiers during long voyage has good predictive accuracy and efficiency.It is an effective assessment tool for screening high-risk cases of fatigue and implementing preventive interventions in

关 键 词：长航官兵 疲劳 影响因素 预测模型 列线图 

分 类 号：R181.2[医药卫生—流行病学]