Procedures used to develop estimates of exposure to acrylonitrile for a cohort study (>25 000 workers in 8 monomer, fiber, and resin companies from 1952 to 1983) are presented. Visits to the companies were made, interviews of workers were conducted, historical records were made, and measurements were taken. On the basis of similar tasks, locations, other exposures, and a similar distribution of exposures to acrylonitrile, 3600 exposure groups were formed. Special procedures were used to reduce the misclassification of workers performing tasks that varied in time but that were inadequately reflected in the job title. A software program organized and retained all exposure information on each exposure group. Quantitative estimates of acrylonitrile exposure were developed using a hierarchical approach in a software program that documented the derivation of each estimate and facilitated data review. Two of the estimation methods were evaluated in a comparison with measurement data.

Key terms cancer; occupation