H M Arrighi1, I Hertz-Picciotto. 1. Department of Epidemiology, School of Public Health, University of North Carolina at Chapel Hill, USA.
Abstract
OBJECTIVE: This investigation sought to examine whether methods proposed to control the healthy worker survivor effect would influence the shape or magnitude of the dose-response curve for respiratory cancer induced by arsenic. METHODS: Results from an unadjusted analysis are compared with results obtained by applying four different methods for control of the healthy worker survivor effect to data on arsenic exposure and respiratory cancer. The four methods are: exposure lag, adjustment for work status, cohort restriction, and the G null test. RESULTS: Cohort restriction gave erratic results depending upon the minimum years of follow up used. Exposure lag substantially increased the rate ratios and a non-linear shape (decreasing slope) compared with an unlagged analysis. Adjusting for work status (currently employed upsilon retired or otherwise not employed) yielded slightly higher rate ratios than an unadjusted analysis, with an overall shape similar to the baseline analysis. Results from the G null test procedure of Robins (1986), although not directly comparable with the baseline analysis, did show an adverse effect of exposure that seemed to reach a maximum when exposure was lagged between 10 and 20 years. CONCLUSIONS: All results confirm an adverse effect of arsenic exposure on respiratory cancer. In these data, it seems that the healthy worker survivor effect was not strong enough to mask the strong effect of arsenic exposure on respiratory cancer. Nevertheless, several methods show a stronger association between arsenic exposure and respiratory cancer after adjustment for the healthy worker survivor effect, suggesting that for weaker causal associations, studies not controlling for this source of bias will have low power to detect results. Although the G methods are theoretically the most unbiased, further work elucidating the validity of the assumptions underlying lagging, adjustment for work status, and the G methods are needed before clear recommendations can be made.
OBJECTIVE: This investigation sought to examine whether methods proposed to control the healthy worker survivor effect would influence the shape or magnitude of the dose-response curve for respiratory cancer induced by arsenic. METHODS: Results from an unadjusted analysis are compared with results obtained by applying four different methods for control of the healthy worker survivor effect to data on arsenic exposure and respiratory cancer. The four methods are: exposure lag, adjustment for work status, cohort restriction, and the G null test. RESULTS: Cohort restriction gave erratic results depending upon the minimum years of follow up used. Exposure lag substantially increased the rate ratios and a non-linear shape (decreasing slope) compared with an unlagged analysis. Adjusting for work status (currently employed upsilon retired or otherwise not employed) yielded slightly higher rate ratios than an unadjusted analysis, with an overall shape similar to the baseline analysis. Results from the G null test procedure of Robins (1986), although not directly comparable with the baseline analysis, did show an adverse effect of exposure that seemed to reach a maximum when exposure was lagged between 10 and 20 years. CONCLUSIONS: All results confirm an adverse effect of arsenic exposure on respiratory cancer. In these data, it seems that the healthy worker survivor effect was not strong enough to mask the strong effect of arsenic exposure on respiratory cancer. Nevertheless, several methods show a stronger association between arsenic exposure and respiratory cancer after adjustment for the healthy worker survivor effect, suggesting that for weaker causal associations, studies not controlling for this source of bias will have low power to detect results. Although the G methods are theoretically the most unbiased, further work elucidating the validity of the assumptions underlying lagging, adjustment for work status, and the G methods are needed before clear recommendations can be made.
Authors: A Blair; P Stewart; M O'Berg; W Gaffey; J Walrath; J Ward; R Bales; S Kaplan; D Cubit Journal: J Natl Cancer Inst Date: 1986-06 Impact factor: 13.506
Authors: Lisa Fredman; Jennifer G Lyons; Jane A Cauley; Marc Hochberg; Katie M Applebaum Journal: J Gerontol A Biol Sci Med Sci Date: 2015-04-15 Impact factor: 6.053
Authors: Andreas M Neophytou; Sally Picciotto; Jaime E Hart; Eric Garshick; Ellen A Eisen; Francine Laden Journal: Occup Environ Med Date: 2014-04-12 Impact factor: 4.402
Authors: Eric Garshick; Francine Laden; Jaime E Hart; Mary E Davis; Ellen A Eisen; Thomas J Smith Journal: Environ Health Perspect Date: 2012-06-01 Impact factor: 9.031