Factors associated with exposure in Occupational Safety and Health Administration data

This study investigated the possibility of making compliance data from the public and private sectors more amenable for multiple uses, by studying data from Occupational Safety and Health Administration (OSHA) inspections during 1979-1989. The potential association of five variables with mean and upper-end (in upper quartile) airborne exposures in similar exposure groups was investigated. The exposure groups reflected airborne exposures to lead in the battery manufacturing industry, to perchloroethylene among dry cleaners, and to iron oxide among welders in three metal fabrication industries. Variables examined were year, inspection type and scope, and size and union status of inspected establishments. Multiple linear regression and logistic regression models were used for the analyses. In small battery plants mean exposure levels were higher and the relative frequency of upper-end exposures (> 75th percentile) greater than in larger establishments. Evidence suggested a decline in mean lead exposures (5-9% per year). Neither type of inspection nor union status were associated with mean or upper-end levels of lead exposure, although the study's power to detect an association was sometimes modest. Some evidence showed that full scope inspections may be associated with higher mean exposures. Strong evidence showed a decline in mean perchloroethylene exposures among dry cleaners (7% per year), but no temporal trends for welder exposures to iron oxide. With few exceptions, the size, type, scope, and union variables were rarely associated with mean or upper-end exposure levels among dry cleaners or welders, although the power of the analyses to detect associations was at times modest. Results show that OSHA data is amenable to analysis that can provide valuable insights about workplace exposures. Several findings of the study are directly useful to the design of public policy.