Marginal analysis of exposure data with repeated measures and non-detects.

Details

• Personal Author:
  Bertke SJ ; Chen I-C ; Estill, Cheryl F.
• Description:
  Exposure data with repeated measures from occupational studies are commonly right-skewed and in the presence of detection limits. The use of linear mixed effects models incorporating maximum likelihood method for repeated measures data with non-detects has been discussed to model log-normal exposure outcomes. However, this modeling has a disadvantage that assumes a correctly specified distribution for the random effect, which is practically unknown, and the estimation methods can result in bias and imprecision in finite-sample data even when distributional assumptions are met. In contrast with random effects modeling, which addresses subject-specific means by explicitly modeling subject-to-subject heterogeneity for the regression parameters, marginal modeling provides an alternative to analyzing data with repeated measurements, in which the parameter interpretations are with respect to marginal or population-averaged means. The interpretations are the same as for the ordinary linear regression, but in the presence of correlated data. In this study, we outline the theories of three marginal models, i.e., generalized estimating equations (GEE), quadratic inference functions (QIF), and generalized method of moments. With these approaches, we propose 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. In a simulation study and application examples, we demonstrate that the GEE method works well in terms of estimating the regression parameters, particularly in small sample sizes, while the QIF outperforms 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. [Description provided by NIOSH]
• Subjects:
  Environmental Exposure Occupational Health Safety
• Keywords:
  Author Keywords: Marginal Analysis; Left Censoring; Right Skewness; Limit Of Detection; Repeated Measures; Environmental Exposure Exposure Assessment; Limit Of Detection; LOD; Statistical Analysis; Statistical Quality Control; Mathematical Models; Environmental Exposure;
• ISSN:
  1556-5068
• Document Type:
  Text
• Genre:
  Journal Article
• Place as Subject:
  Ohio ; OSHA Region 5
• CIO:
  National Institute for Occupational Safety and Health (NIOSH)
• Division:
  DFSE - Division of Field Studies and Engineering
• Topic:
  Workplace Safety & Health
• Location:
  North America
• NIOSHTIC Number:
  nn:20065347
• Citation:
  SSRN 2022 May; :[Preprint]
• Contact Point Address:
  I-Chen Chen, Division of Field Studies and Engineering, National Institute for Occupational Safety and Health, Centers for Disease Control and Prevention, Cincinnati, Ohio, USA
• Email:
  okv0@cdc.gov
• Federal Fiscal Year:
  2022
• NORA Priority Area:
  Manufacturing
• Peer Reviewed:
  False
• Source Full Name:
  SSRN
• Collection(s):
  National Institute for Occupational Safety and Health
• Main Document Checksum:
  urn:sha-512:e18779b65dc8c5eac1ef3b4ea50c931ccd6ed912da68c152e547f48ed922a724846f555762b9b0cd2f1388da2fab74fdbeb41fa85cabfcab1e0762ef1916ff85
• Download URL:
  https://stacks.cdc.gov/view/cdc/249410/cdc_249410_DS1.pdf
• File Type:
  [PDF - 207.59 KB ]