Quantile regression for longitudinal data with values below the limit of detection and time-dependent covariates-application to modeling carbon nanotube and nanofiber exposures.
IF 1.8 4区 医学Q3 PUBLIC, ENVIRONMENTAL & OCCUPATIONAL HEALTH
{"title":"Quantile regression for longitudinal data with values below the limit of detection and time-dependent covariates-application to modeling carbon nanotube and nanofiber exposures.","authors":"I-Chen Chen, Stephen J Bertke, Matthew M Dahm","doi":"10.1093/annweh/wxae068","DOIUrl":null,"url":null,"abstract":"<p><strong>Background: </strong>In studies of occupational health, longitudinal environmental exposure, and biomonitoring data are often subject to right skewing and left censoring, in which measurements fall below the limit of detection (LOD). To address right-skewed data, it is common practice to log-transform the data and model the geometric mean, assuming a log-normal distribution. However, if the transformed data do not follow a known distribution, modeling the mean of exposure may result in bias and reduce efficiency. In addition, when examining longitudinal data, it is possible that certain covariates may vary over time.</p><p><strong>Objective: </strong>To develop predictive quantile regression models to resolve the issues of left censoring and time-dependent covariates and to quantitatively evaluate if previous and current covariates can predict current and/or future exposure levels.</p><p><strong>Methods: </strong>To address these gaps, we suggested incorporating different substitution approaches into quantile regression and utilizing a method for selecting a working type of time dependency for covariates.</p><p><strong>Results: </strong>In a simulation study, we demonstrated that, under different types of time-dependent covariates, the approach of multiple random value imputation outperformed the other approaches. We also applied our methods to a carbon nanotube and nanofiber exposure study. The dependent variables are the left-censored mass of elemental carbon at both the respirable and inhalable aerosol size fractions. In this study, we identified some potential time-dependent covariates with respect to worker-level determinants and job tasks.</p><p><strong>Conclusion: </strong>Time dependency for covariates is rarely accounted for when analyzing longitudinal environmental exposure and biomonitoring data with values less than the LOD through predictive modeling. Mistreating the time-dependency as time-independency will lead to an efficiency loss of regression parameter estimation. Therefore, we addressed time-varying covariates in longitudinal exposure and biomonitoring data with left-censored measurements and illustrated an entire conditional distribution through different quantiles.</p>","PeriodicalId":8362,"journal":{"name":"Annals Of Work Exposures and Health","volume":" ","pages":"846-858"},"PeriodicalIF":1.8000,"publicationDate":"2024-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Annals Of Work Exposures and Health","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1093/annweh/wxae068","RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"PUBLIC, ENVIRONMENTAL & OCCUPATIONAL HEALTH","Score":null,"Total":0}
引用次数: 0
Abstract
Background: In studies of occupational health, longitudinal environmental exposure, and biomonitoring data are often subject to right skewing and left censoring, in which measurements fall below the limit of detection (LOD). To address right-skewed data, it is common practice to log-transform the data and model the geometric mean, assuming a log-normal distribution. However, if the transformed data do not follow a known distribution, modeling the mean of exposure may result in bias and reduce efficiency. In addition, when examining longitudinal data, it is possible that certain covariates may vary over time.
Objective: To develop predictive quantile regression models to resolve the issues of left censoring and time-dependent covariates and to quantitatively evaluate if previous and current covariates can predict current and/or future exposure levels.
Methods: To address these gaps, we suggested incorporating different substitution approaches into quantile regression and utilizing a method for selecting a working type of time dependency for covariates.
Results: In a simulation study, we demonstrated that, under different types of time-dependent covariates, the approach of multiple random value imputation outperformed the other approaches. We also applied our methods to a carbon nanotube and nanofiber exposure study. The dependent variables are the left-censored mass of elemental carbon at both the respirable and inhalable aerosol size fractions. In this study, we identified some potential time-dependent covariates with respect to worker-level determinants and job tasks.
Conclusion: Time dependency for covariates is rarely accounted for when analyzing longitudinal environmental exposure and biomonitoring data with values less than the LOD through predictive modeling. Mistreating the time-dependency as time-independency will lead to an efficiency loss of regression parameter estimation. Therefore, we addressed time-varying covariates in longitudinal exposure and biomonitoring data with left-censored measurements and illustrated an entire conditional distribution through different quantiles.
期刊介绍：
About the Journal
Annals of Work Exposures and Health is dedicated to presenting advances in exposure science supporting the recognition, quantification, and control of exposures at work, and epidemiological studies on their effects on human health and well-being. A key question we apply to submission is, "Is this paper going to help readers better understand, quantify, and control conditions at work that adversely or positively affect health and well-being?"
We are interested in high quality scientific research addressing:
the quantification of work exposures, including chemical, biological, physical, biomechanical, and psychosocial, and the elements of work organization giving rise to such exposures;
the relationship between these exposures and the acute and chronic health consequences for those exposed and their families and communities;
populations at special risk of work-related exposures including women, under-represented minorities, immigrants, and other vulnerable groups such as temporary, contingent and informal sector workers;
the effectiveness of interventions addressing exposure and risk including production technologies, work process engineering, and personal protective systems;
policies and management approaches to reduce risk and improve health and well-being among workers, their families or communities;
methodologies and mechanisms that underlie the quantification and/or control of exposure and risk.
There is heavy pressure on space in the journal, and the above interests mean that we do not usually publish papers that simply report local conditions without generalizable results. We are also unlikely to publish reports on human health and well-being without information on the work exposure characteristics giving rise to the effects. We particularly welcome contributions from scientists based in, or addressing conditions in, developing economies that fall within the above scope.
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号
