I-Chen Chen, Stephen J. Bertke, Cheryl Fairfield Estill
{"title":"Compare the marginal effects for environmental exposure and biomonitoring data with repeated measurements and values below the limit of detection","authors":"I-Chen Chen, Stephen J. Bertke, Cheryl Fairfield Estill","doi":"10.1038/s41370-024-00640-7","DOIUrl":null,"url":null,"abstract":"Environmental exposure and biomonitoring data with repeated measurements from environmental and occupational studies are commonly right-skewed and in the presence of limits of detection (LOD). However, existing model has not been discussed for small-sample properties and highly skewed data with non-detects and repeated measurements. Marginal modeling provides an alternative to analyzing longitudinal and cluster data, in which the parameter interpretations are with respect to marginal or population-averaged means. We outlined the theories of three marginal models, i.e., generalized estimating equations (GEE), quadratic inference functions (QIF), and generalized method of moments (GMM). With these approaches, we proposed to incorporate the fill-in methods, including single and multiple value imputation techniques, such that any measurements less than the limit of detection are assigned values. We demonstrated that the GEE method works well in terms of estimating the regression parameters in small sample sizes, while the QIF and GMM outperform in large-sample settings, as parameter estimates are consistent and have relatively smaller mean squared error. No specific fill-in method can be deemed superior as each has its own merits.","PeriodicalId":15684,"journal":{"name":"Journal of Exposure Science and Environmental Epidemiology","volume":"34 6","pages":"1018-1027"},"PeriodicalIF":4.1000,"publicationDate":"2024-01-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Exposure Science and Environmental Epidemiology","FirstCategoryId":"3","ListUrlMain":"https://www.nature.com/articles/s41370-024-00640-7","RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"ENVIRONMENTAL SCIENCES","Score":null,"Total":0}
引用次数: 0
Abstract
Environmental exposure and biomonitoring data with repeated measurements from environmental and occupational studies are commonly right-skewed and in the presence of limits of detection (LOD). However, existing model has not been discussed for small-sample properties and highly skewed data with non-detects and repeated measurements. Marginal modeling provides an alternative to analyzing longitudinal and cluster data, in which the parameter interpretations are with respect to marginal or population-averaged means. We outlined the theories of three marginal models, i.e., generalized estimating equations (GEE), quadratic inference functions (QIF), and generalized method of moments (GMM). With these approaches, we proposed to incorporate the fill-in methods, including single and multiple value imputation techniques, such that any measurements less than the limit of detection are assigned values. We demonstrated that the GEE method works well in terms of estimating the regression parameters in small sample sizes, while the QIF and GMM outperform in large-sample settings, as parameter estimates are consistent and have relatively smaller mean squared error. No specific fill-in method can be deemed superior as each has its own merits.
期刊介绍:
Journal of Exposure Science and Environmental Epidemiology (JESEE) aims to be the premier and authoritative source of information on advances in exposure science for professionals in a wide range of environmental and public health disciplines.
JESEE publishes original peer-reviewed research presenting significant advances in exposure science and exposure analysis, including development and application of the latest technologies for measuring exposures, and innovative computational approaches for translating novel data streams to characterize and predict exposures. The types of papers published in the research section of JESEE are original research articles, translation studies, and correspondence. Reported results should further understanding of the relationship between environmental exposure and human health, describe evaluated novel exposure science tools, or demonstrate potential of exposure science to enable decisions and actions that promote and protect human health.