The effect of transient exposures on the risk of an acute illness with low hazard rate.

Abstract

There are many situations where intermittent short-term exposures of a certain kind are thought to temporarily enhance the risk of onset of an adverse health event (illness). When the hazard rate of the illness is small it is desirable to investigate this possible association using only data on cases occurring in a finite observation period. Here we extend a method for such an analysis by allowing the baseline hazard for the illness to depend on the increasing age over the observation period and using age at the times of exposure, a time dependent variable, as a covariate in the effect of the transient exposure. The method is illustrated with a study of the possible association of long-haul air travel and hospitalization for venous thromboembolism over an observation period of 19 years. It is demonstrated that allowing for aging over the observation period can avoid bias in the estimated effect size when the baseline hazard for the illness increases with age and exposures occur irregularly over time.