噪声时域特征对职业噪声性听力损失作用研究  

The effecs of time-domain features of noise on occupational noise-induced hearing loss

作　　者：曾安可 李金哲 辛佳芮 邱伟 孙新[1] 邹华[4] 张美辨 ZENG Anke;LI Jinzhe;XIN Jiarui;QIU Wei;SUN Xin;ZOU Hua;ZHANG Meibian(National Institute for Occupational Health and Poison Control,Chinese Center for Disease Control and Prevention,Beijing 100050,China)

机构地区：[1]中国疾病预防控制中心职业卫生与中毒控制所,北京100050 [2]北京市疾病预防控制中心 [3]浙江清华长三角研究院 [4]浙江省疾病预防控制中心

出　　处：《工业卫生与职业病》2025年第1期4-11,共8页Industrial Health and Occupational Diseases

基　　金：“十四五”国家重点研发项目(2022YFC2503200,2022YFC2503203);美国国立卫生院(NIH)资助项目(1R01DC015990-01)。

摘　　要：目的探索有效的时域指标以评估复杂噪声对职业噪声性听力损失的影响。方法基于横断面设计,分析966名男性制造业劳动者接触的全班次职业噪声的时域要素(噪声事件数量、噪声相对值以及占空比)与间接时域指标(峰度、波形因子、脉冲因子、峰值因子以及裕度因子)。分析噪声时域要素与间接时域指标间的关联;采用LASSO回归、多元线性回归分析有效间接时域指标对1、2、3、4 kHz频率下噪声导致的永久性听阈位移平均值(NIPTS_(1234))的作用;基于Boruta算法选择特征。结果5个噪声间接时域指标与各时域要素均显著相关(P<0.01);LASSO回归筛除了3个间接时域指标(脉冲因子、裕度因子与波形因子,r=0);LASSO回归与多元线性回归结果均表明峰值因子与NIPTS_(1234)负相关,峰度与NIPTS_(1234)正相关,均被Boruta算法确认为NIPTS_(1234)的重要预测变量。结论峰度是工作场所噪声评估与噪声性听力损失预测的潜在有效指标。Objective To explore effective time-domain indicators for assessing the impact of non-Gaussian noise on occupational noise-induced hearing loss(ONIHL).Methods A cross-sectional study was conducted among966 male manufacturing workers to examine both time-domain elements(number of noise events,peak relative value,and duty factor)and indirect time-domain indicators(kurtosis,shape factor,impulse factor,crest factor,and margin factor)of whole-shift noise recordings.Specifically,we examined the associations between time-domain elements and indirect indicators;the LASSO regression and multiple linear regression were applied to assess the effects of significant indirect indicators on noise-induced permanent threshold shifts at 1,2,3,4 kHz(NIPTS_(1234));the Boruta algorithm was used for feature selection.Results Significant correlations were found between five indirect indicators and all time-domain elements(P<0.01).LASSO regression filtered three indirect indicators(i.e.impulse factor,margin factor,and shape factor,r=0).Both LASSO and multiple linear regression results indicated that crest factor was negatively correlated with NIPTS_(1234),while kurtosis was positively correlated with NIPTS_(1234).The Boruta algorithm further validated both the crest factor and kurtosis as significant predictors for NIPTS_(1234).Conclusion Kurtosis was a potential effective indicator for workplace noise exposure assessment and ONIHL prediction.

关 键 词：工作场所 噪声 噪声事件 时域特征 噪声性听力损失 

分 类 号：R136.2[医药卫生—劳动卫生]