Measuring underreporting by comparing workers' compensation and bureau of labor statistics injury data

Background and Objectives. The Bureau of Labor Statistics' annual Survey of Occupational Injuries and Illnesses (SOII) and state workers' compensation (WC) data are two systems commonly used to gauge the frequency and severity of work-related injuries. However, researchers have raised concerns about underreporting to each of these systems. We describe the protocols and results of several recent studies that link the SOII and WC data sources at the level of the injury in order to directly evaluate the degree of underreporting.

Methods. The studies link individual injury cases from administrative and survey records using detailed information on the worker, employer and the injury. This identifies each case as being either common to both sources, or as being from one source only. Subsequent to linkage harmonized inclusion criteria are imposed so that analysis only includes cases potentially reportable to both systems. The degree of SOII underreporting is estimated from the propensity for cases to be found in the WC data only. To examine factors affecting reporting, the research subsets the data or models reporting propensities based on observable characteristics.

Results. The linkage studies we describe all indicate a substantial underreporting of injuries to the SOII. The studies identify large variation in underreporting by state and identify certain situations where underreporting is more likely.

Conclusions. Injury and illness statistics reported in the SOII are likely to substantially underestimate true risk. Also, the degree of underreporting appears to vary considerably by state. Researchers should use caution in comparing SOII injury rates across states.