Background: The healthy worker survivor bias is well-recognized in occupational epidemiology. Three component associations are necessary for this bias to occur: i) prior exposure and employment status; ii) employment status and subsequent exposure; and iii) employment status and mortality. Together, these associations result in time-varying confounding affected by prior exposure. We illustrate how these associations can be assessed using standard regression methods.
Methods: We use data from 2975 asbestos textile factory workers hired between January 1940 and December 1965 and followed for lung cancer mortality through December 2001.
Results: At entry, median age was 24 years, with 42% female and 19% non-Caucasian. Over follow-up, 21% and 17% of person-years were classified as at work and exposed to any asbestos, respectively. For a 100 fiber-year/mL increase in cumulative asbestos, the covariate-adjusted hazard of leaving work decreased by 52% (95% confidence interval [CI], 46-58). The association between employment status and subsequent asbestos exposure was strong due to nonpositivity: 88.3% of person-years at work (95% CI, 87.0-89.5) were classified as exposed to any asbestos; no person-years were classified as exposed to asbestos after leaving work. Finally, leaving active employment was associated with a 48% (95% CI, 9-71) decrease in the covariate-adjusted hazard of lung cancer mortality.
Conclusions: We found strong associations for the components of the healthy worker survivor bias in these data. Standard methods, which fail to properly account for time-varying confounding affected by prior exposure, may provide biased estimates of the effect of asbestos on lung cancer mortality under these conditions.
Copyright © 2013 Elsevier Inc. All rights reserved.