Influence of Data Elements in OSHA Air Sampling Database on Occupational Exposure Levels

Abstract

The influence of five variables in the Occupational Safety and Health Administration (OSHA) sampling data collected from 1979-1997 in three exposure groups were examined for their association with the reported levels of occupational exposure to airborne chemicals. The exposure groups evaluated were: 1) asbestos abatement workers exposed to asbestos fibers; 2) automotive body repair shop workers exposed to toluene during spray painting; and 3) embalmers exposed to formaldehyde. The variables (data elements) examined were type, year and scope of inspection, size of the employer, and union status. Multiple linear regression was used to investigate, using three statistical models, the influence of these variables on occupational exposure levels. In the asbestos and toluene exposure groups, a consistent time-related trend in mean exposures was observed with decreasing exposure levels being seen under each statistical model. By contrast, no consistent relationship between year of inspection and exposure levels was seen for the formaldehyde exposure group. Higher exposures were detected during comprehensive inspections than during partial scope inspections in each exposure group under all three regression models, suggesting that scope may be a source of bias in this data. Higher mean exposures were measured during complaint inspections than planned inspections across all regression models in the toluene and formaldehyde exposure groups. However, this trend was reversed for the asbestos case study, in which higher mean exposures were measured during planned inspections under each model. Type of inspection and union status were not consistently found to be associated with mean exposure levels in any of the exposure groups. The strengths and limitations of the OSHA sampling data in providing insights regarding occupational exposures, as well as their utility beyond assessing regulatory compliance are discussed. The results of this study are compared with those obtained in a previous study of OSHA compliance data that examined the potential association of the same variables with exposure levels in three different exposure groups.