Task- and time-dependent weighting factors in a retrospective exposure assessment of chemical laboratory workers.

A chemical exposure assessment was conducted for a cohort mortality study of 6157 chemical laboratory workers employed between 1943 and 1998 at four Department of Energy sites in Oak Ridge, Tennessee, and Aiken, South Carolina. Previous studies of chemical laboratory workers have included members within professional societies where exposure assessment was either limited or not feasible, or chemical processing employees where laboratory and production workers were combined. Because sufficient industrial hygiene records were unavailable for all four sites, weighted duration of employment was used as a surrogate for the magnitude of exposure. Potential exposure indices were calculated for each worker using number of days employed and weighting factors for frequency of contact and year of employment. A total of 591 unique laboratory job titles indicative of a chemical laboratory worker were collapsed into 18 general job title categories. Through discussions with current and retired workers, along with examination of historical organizational charts and job descriptions, the percentage of time with activities involving the direct handling of chemicals in the laboratory was estimated for each job title category. Scaled weighting factors of 1, 0.6, 0.3, and 0.05 were assigned to the job title categories representing 100%, 60%, 30%, and 5% of daily activities handling chemicals, respectively. Based on limited industrial hygiene monitoring data, personal radiation monitoring records, and professional judgment, weighting factors that declined 4% annually were applied to each year to account for improvements in laboratory technique, advancements in instrumentation, improvement in engineering controls, and increased safety awareness through time. The study cohort was separated into three categories of chemical exposures based on department level information: (1) inorganic, (2) mixed inorganic and organic, and (3) unknown. Potential exposure indices ranged from 0.15 to 6824.5 with a median value of 377.5 and a mean equal to 884.2. This exposure assessment method is useful for epidemiologic analyses when quantitative exposure data are absent or insufficient.