Background: Healthy worker survivor bias (HWSB) occurs when healthier individuals remain employed for longer durations than those who leave employment potentially due to occupational exposures, resulting in underestimation of the true effect of an exposure-outcome association. To our knowledge, g-estimation has only been used once to calculate counterfactual unexposed survival times to account for the impact of HWSB on survival time related to lung cancer mortality among styrene-exposed workers (Bertke, 2021).

Objectives: We evaluated the influence of HWSB on survival time (time from cohort entry to death) related to lung cancer mortality among a historical cohort of styrene-exposed workers employed across 30 U.S. reinforced plastics facilities between 1948 and 1977.

Methods: The impact of HWSB on survival time was assessed among a cohort of 15,826 styrene-exposed workers followed from 1948 through 2019. We used g-estimation of a structural nested model adjusting for past exposures, sex, age, and calendar period, conditioned on employment history. Using g-estimation, an exposure-response parameter that compared survival times among exposed versus counterfactually unexposed workers was calculated.

Results: We identified an exposure-response parameter of -0.56, indicating styrene exposure was not associated with a shorter time to lung cancer death after accounting for HWSB. Specifically, this result implies that 1 year of exposure greater than 25 ppm does not accelerate time to lung cancer death.

Conclusion: As the second study that has assessed HWSB among a cohort of styrene-exposed workers, this study suggests HWSB did not have an impact on survival times related to lung cancer mortality.