Methodological problems of time-related variables in occupational cohort studies.

In studies of long-term health effects of occupational exposures it is important that the time patterns of the study exposure and of the relevant confounders should be taken into account in the analysis. The time-related confounders that have been most frequently considered include age at risk, calendar period, duration of employment, length of follow-up, and employment status (active or retired). These factors are related to the healthy worker effect which is more pronounced among active workers, and declines with length of follow-up. Thus, these time-related factors are potential confounders in occupational cohort studies (and in nested case-control studies). It is therefore advisable to routinely control for length of follow-up and employment status, in addition to the usual confounding factors of age at risk and calendar period. However, it should be stressed that this approach will minimise, but not eliminate, confounding due to the healthy worker effect. In particular, direct control for employment status is inadvisable if this constitutes an intermediate variable in the pathway between exposure and disease, and more complex analytical procedures are necessary in such situations. It is also important that the principle exposure under study should also be analyzed in a time-related manner, taking account of the likely induction and latency periods, and the relative etiological importance of exposure intensity, exposure duration and cumulative exposure. The simplest approach is to analyse the cumulative exposure in a time-related manner, and this may suffice when the aim is merely to consider whether or not there is an effect of exposure. However, once it has been provisionally assumed that an effect exists, attention then shifts to understanding the nature of the effect. In this context, the temporal pattern of exposure and outcome can be considered by examining the effects of exposures in specific time windows while controlling for time-related confounders, and for the effects of exposures in other time windows. A more sophisticated approach is direct fitting of a theoretical model of carcinogenesis, such as the Armitage-Doll or Moolgavkar models. However, it should be emphasized that occupational cohort studies only rarely have sufficient numbers, and data of sufficient quality, to permit meaningful conclusions to be drawn from more sophisticated analyses of this type.