Re: Adjustment for Duration of Employment in Occupational Epidemiology

Details

• Personal Author:
  Bertke S ; Costello S ; Keil AP ; Kelly-Reif K ; Richardson DB
• Description:
  In regards to the recent manuscript by Ko et al. (2024), we are writing to express concerns with their conclusions regarding control of healthy worker survivor bias (HWSB), namely that HWSB should be addressed, when estimating associations between cumulative occupational exposure and mortality, either by adjustment for duration of employment as a confounder in Cox proportional hazards models or by inverse probability weighting (IPW) in Cox marginal structural models. We support this effort to identify simple approaches in this area. However, we assert that their results should not be used for general recommendations, in part because one of the main findings resulted from an analytic error. ... If exposure does not affect employment (as in the simulations of Ko et al. (2024)), then adjustment for employment as a time-dependent variable via adjustment or stratification is a valid approach to reducing bias. However, if exposure affects employment (e.g. by causing illness that leads to employment changes), an important aspect to HWSB, then g-methods (IPW, g-estimation, and g-computation) are needed. We have observed a number settings in which exposure is associated with the subsequent hazard of leaving work, so we would generally advocate in favor of g-methods. However, IPW, as implemented by Ko et al. (2024), is not an appropriate method for reducing healthy worker survivor bias in any setting that violates the positivity assumption. [Description provided by NIOSH]
• Subjects:
  Epidemiology Mortality Neoplasms Occupational Exposure Occupational Health Safety
• Keywords:
  Author Keywords: Ionizing Radiation Employee Health Epidemiologic Biases Mortality Studies Occupational Exposure Statistical Quality Control Survival Rate
• ISSN:
  1047-2797
• Document Type:
  Text
• Genre:
  Journal Article
• Place as Subject:
  California ; Chicago Region [OSHA] ; Maryland ; Ohio ; Philadelphia Region [OSHA] ; San Francisco Region [OSHA]
• CIO:
  National Institute for Occupational Safety and Health (NIOSH)
• Division:
  DFSE - Division of Field Studies and Engineering
• Topic:
  Workplace Safety & Health
• Location:
  North America
• Pages in Document:
  69-70
• Volume:
  98
• NIOSHTIC Number:
  nn:20070006
• Citation:
  Ann Epidemiol 2024 Oct; 98:69-70
• Contact Point Address:
  Alexander Keil, 9609 Medical Center Dr., Rockville, MD USA 20892
• Email:
  alex.keil@nih.gov
• Federal Fiscal Year:
  2025
• Peer Reviewed:
  False
• Source Full Name:
  Annals of Epidemiology
• Collection(s):
  National Institute for Occupational Safety and Health
• Main Document Checksum:
  urn:sha-512:a5a8e26a559abaf5d012ffce71d5f3a0ae1159217c2a67ba55a4930110cb38a5ad28be93dd5e7efa538a12b5512e00826482c3e81242c86cc9441d18ccb92791
• Download URL:
  https://stacks.cdc.gov/view/cdc/208315/cdc_208315_DS1.pdf
• File Type:
  [PDF - 315.19 KB ]