Occupational mortality studies have identified a large proportion of known human carcinogens. However , such studies are sometimes difficult to interpret due to selective forces collectively known as the healthy worker effect. One aspect of the healthy worker effect is the healthy worker survivor effect (HWSE), in which those who remain employed are healthier than those who terminate employment. The effect of the HWSE is to attenuate any effect of an occupational exposure towards the appearance of no effect. For this reason, the development and validation of appropriate statistical methods to control for the HWSE would be expected to improve the sensitivity of occupational cancer studies. The primary objective of this study is to validate and compare several methods of adjusting for the HWSE. Using existing databases on three occupationally exposed cohorts, we will reassess previously reported analyses that showed: (i) a strong association and a dose-response relationship between arsenic exposure and lung cancer mortality in a cohort of copper smelter workers, (ii) a weak association between low-level external penetrating ionizing radiation exposure and deaths from cancer among employees at Oak Ridge National Laboratories, (iii) no association between formaldehyde and cancer mortality among workers exposed in several industries. A second objective is to explore two other issues related to analysis of occupational studies with quantified exposure data. Specifically, we will (i) assess the ability of ratio, as opposed to difference, measures of effect to detect dose-response relationships for cancer mortality, and (i i) examine the comparability of standardized mortality ratios at different levels of cumulative exposure and determine the degree of confounding due to differences in age and calendar-year distributions, and active/terminated employment status. This investigation will result in a deeper understanding of both the degree to which the HWSE may attenuate estimates of carcinogenic effects from occupational exposures, and the appropriate methods to control this bias. Further, since occupational studies are often used as a basis for risk assessments to establish environmental or occupational standards for exposure levels, an understanding of this bias and the methods for its control would have policy ramifications.
| Arrighi, H M; Hertz-Picciotto, I (1996) Controlling the healthy worker survivor effect: an example of arsenic exposure and respiratory cancer. Occup Environ Med 53:455-62 |
| Arrighi, H M; Hertz-Picciotto, I (1995) Controlling for time-since-hire in occupational studies using internal comparisons and cumulative exposure. Epidemiology 6:415-8 |
| Arrighi, H M; Hertz-Picciotto, I (1994) The evolving concept of the healthy worker survivor effect. Epidemiology 5:189-96 |
