Time and causation in studies of long-term exposure to environmental factors.

The various shortcomings of the prevalence study as an instrument for causal research are evaluated by comparing the prevalence design with the classical experimental design. On the basis of these shortcomings it is suggested that epidemiologists refine their research strategies in two respects, both related to time. (1) it is suggested that they include formerly exposed as well as presently exposed in their sampling frames and (2) that they avoid using present status with respect to causal factors as indicators of life-experience of these factors in analysis. On the basis of some studies of the relationship between long-term exposure to presumed noxae and chronic lung disease, numerical examples of the undesirable consequences of ignoring time are given.