Using Occupational Safety and Health Administration accident investigations to study patterns in work fatalities.

Investigations of fatalities by the Occupational Safety and Health Administration (OSHA) provide the most detailed available information about traumatic workplace deaths that are potentially related to violations of existing safety standards. Comparison of the number of such deaths investigated by OSHA from 1977 to 1986 with the comparable category of deaths reported to the Bureau of Labor Statistics Survey of Occupational Injuries and Illnesses indicates that the overall magnitudes have been roughly similar. The OSHA data contain more information than other sources and are especially valuable for analyses of fatalities at smaller workplaces. The OSHA data show that death rates decline sharply with establishment size; the inverted "U" pattern for lost workday injury rates is absent. Because accident investigations are conducted as part of an administrative system, the OSHA data can be influenced by changes in administrative policies. Changes over time in the percent of fatalities in which violations of OSHA standards were cited have clearly been influenced by changes in OSHA citation policy and thus do not provide a valid measure of the rate of violation-caused deaths. Realization of the epidemiological value of this data source depends upon a commitment from OSHA to maintain consistency in investigating accidents and to improve its data collection methods.