粉尘作业工人肺功能异常风险预测模型的建立与验证  

Establishment and verification of risk prediction model of pulmonary function abnormality in workers exposed to dust

作　　者：张奇 杨志辉 杨甲飞 刘志宏[1] ZHANG Qi;YANG Zhi-hui;YANG Jia-fei;LIU Zhi-hong(School of Public Health,Ningxia Medical University,Yinchuan,Ningxia 750004,China;不详)

机构地区：[1]宁夏医科大学公共卫生学院,宁夏银川750004 [2]宁夏颐阳老年病医院健康体检科

出　　处：《现代预防医学》2023年第21期3878-3882,3894,共6页Modern Preventive Medicine

基　　金：宁夏回族自治区重点研发项目(2021BEG02030)。

摘　　要：目的探讨银川市粉尘作业工人肺功能异常的影响因素,建立肺功能异常风险预测模型并验证。方法选取银川市2022年3月—7月进行职业健康体检的粉尘作业工人进行问卷调查、体格检查和实验室检查,共纳入923名调查对象。采用logistic回归分析粉尘作业工人肺功能异常影响因素并构建肺功能异常风险预测列线图模型,采用受试者工作特征(receiver operator characteristic,ROC)曲线、校准曲线和决策分析曲线对模型进行评价。结果多因素logistic回归结果显示,工作时无个人防护(OR=7.806,95%CI:5.803~11.989)、工龄[与0~5年相比,6~10年OR=3.025,95%CI:1.764~5.188;>10年OR=2.546,95%CI:1.545~4.193]、高血压(OR=3.909,95%CI:2.223~6.873)、BMI(≥24.0 kg/m2)(OR=1.553,95%CI:1.004~2.401)均是肺功能异常的独立危险因素,文化程度为大专(OR=0.503,95%CI:0.299~0.847)是肺功能异常的保护因素。风险预测列线图ROC曲线下面积为0.806,校准曲线表现出良好的一致性。临床决策曲线结果显示,当列线图模型阈值概率为10%~88%时进行肺功能异常干预具有更大的收益。结论通过logistic回归构建的列线图模型直观的显示了多个影响因素的作用,在对粉尘作业工人肺功能异常风险进行预测时具有良好的准确性。Objective To investigate the influencing factors of abnormal pulmonary function in workers exposed to dust in Yinchuan,and to establish and verify a risk prediction model of abnormal pulmonary function.Methods A total of 923 dust workers who underwent occupational health examination in Yinchuan from March to July 2022 were investigated by questionnaire,physical examination,and laboratory examination.The influencing factors of abnormal pulmonary function in dust exposed workers were analyzed by Logistic regression,and the risk prediction model of abnormal pulmonary function was established.The Receiver Operator Characteristic(ROC)curve,calibration curve,and decision analysis curve were used to evaluate the model.Results The results of multivariate Logistic regression showed that no personal protection(OR=7.806,95%CI:5.803-11.989),length of service at work(compared with 0 to 5 years,6 to 10 years:OR=3.025,95%CI:1.764-5.188;>10 years:OR=2.546,95%CI:1.545-4.193),hypertension(OR=3.909,95%CI:2.223-6.873),and BMI(≥24.0 kg/m2)(OR=1.553,95%CI:1.004-2.401)were independent risk factors of pulmonary dysfunction,while education(college)(OR=0.503,95%CI:0.299-0.847)was the protective factor of pulmonary dysfunction.The area under the ROC curve of the risk prediction line chart was 0.806,and the calibration curve showed good consistency.The results of clinical decision curve showed that the intervention of abnormal lung function had greater benefits when the threshold probability of the line chart model was 10%to 88%.Conclusion The nomogram constructed by Logistic regression visually demonstrates the role of many influencing factors and has good accuracy in predicting the risk of abnormal pulmonary function in workers exposed to dust.

关 键 词：肺功能 影响因素 预测模型 列线图 

分 类 号：R135.2[医药卫生—劳动卫生]