军事航空飞行人员听力损失风险预测模型的构建与验证  

Construction and Validation of a Nomogram for Predicting the Risk of Hearing Loss in Military Aviation Flight Pilots

作　　者：翟晓敏 张红蕾 董佳佳 梁雅竹 于笑 薛蓉 张昊 濮彧 霍小秀 郭睿 ZHAI Xiaomin;ZHANG Honglei;DONG Jiajia;LIANG Yazhu;YU Xiao;XUE Rong;ZHANG Hao;PU Yu;HUO Xiaoxiu;GUO Rui(Graduate School of Hebei North University,Zhangjiakou 075000,China;不详)

机构地区：[1]河北北方学院,张家口075000 [2]空军特色医学中心耳鼻咽喉头颈外科 [3]首都医科大学附属北京天坛医院耳鼻咽喉头颈外科

出　　处：《中华耳科学杂志》2025年第1期74-78,共5页Chinese Journal of Otology

基　　金：中国人民解放军空军特色医学中心院级临床研究项目(2021LC017)。

摘　　要：目的开发并验证一种预测军事航空飞行人员听力损失风险的列线图,并评估其临床应用价值。方法采用横断面研究,选取2019年1月至2020年1月空军特色医学中心进行体检的210例航空军事飞行人员。采用纯音测听法测量听力阈值。按照纳入与排除标准,将符合标准的195例航空军事飞行人员按照7∶3随机分为训练集(139例)和验证集(56例)。采用最小绝对收缩和选择算子回归(least absolute shrinkage and selection operator,LAASO)和二元logistic回归识别飞行员听力损失的危险因素。构建一个基于二元logistic回归模型和临床实践经验的列线图来预测飞行人员听力损失的风险因素。采用一致性指数、受试者工作特征曲线(receiver operating characteristic curve,ROC)下面积和校准曲线的方法对模型的区分度和校准度进行验证。结果训练集和验证集的一致性指数分别为0.72(95%CI:0.631~0.808),0.801(95%CI:0.676~0.926),训练集和验证集ROC曲线下面积分别为0.72和0.801。本研究结果显示,模型具有良好的区分能力。校准曲线也表明,该模型具有较好的辨别度。结论基于年龄、飞行时间和飞机类型的列线图有助于临床医师预测军事航空飞行人员发生听力损失疾病的风险。Objective To develop and validate nomogram for predicting the risk of hearing loss in military aviation pilots and evaluation of its clinical application.Methods A cross-sectional study,total of 210 aviation military flight personnel who underwent medical examinations at the Air Force Medical Centre from January 2019 to January 2020 were selected..Hearing thresholds were measured using pure-tone audiometry.Based on the strict inclusion and exclusion principle,the flight crew meeting the criteria were randomly divided into training and validation sets in a ratio of 7:3.LASSO regression and binary logistic regression were used to identify risk factors for hearing loss in pilots.Based on a binary logistic regression model and clinical practice experience,a monogram was constructed to predict the risk factors for hearing loss in flight crews.The model was differentiated and calibrated using validation methods such as consistency indices,the area under the ROC curve,and the calibration lot.The area under the ROC curve was 0.72 and 0.80 for the training and validation sets,respectively,indicating that the model has a good discriminatory ability.Results The consistency indices for the training and validation sets were 0.72(95%CI 0.63~0.80)and 0.80(95%CI 0.67~0.92),respectively.The areas under the ROC curves of the training set and the validation set are 0.72 and 0.80,respectively.The results of this study show that the model has good discriminatory power.The calibration curve also shows that the model has good discriminatory power.Conclusion A nomogram based on age,flight time,and aircraft type help clinicians predict the risk of hearing loss disease in aviation military fliers.

关 键 词：航空军事飞行员 听力损失 风险因素 列线图 

分 类 号：R856.76[医药卫生—航空、航天与航海医学]