The law and incomplete database information as confounders in epidemiologic research on occupational injuries and illnesses.

BACKGROUND Capture-recapture studies report undercounting of work injuries/illnesses with days away from work (DAFW) in the Bureau of Labor Statistics annual Survey of Occupational Injuries and Illnesses (BLS SOII) by 25-68% depending on the state and undercounting by various state workers' compensation (WC) systems of eligible claims by 5-35%. METHODS Statutory/regulatory criteria defining eligible cases are used to adjust counts in the 1998-2001 Minnesota's WC system and the BLS SOII to permit comparison and to evaluate the recent studies. Missing information in the employer database used in the capture-recapture studies is tabulated. An attempt is made to harmonize results with two additional databases counting work injuries. RESULTS Counts in the BLS SOII moderately undercount by 10-16% the number of WC cases. We believe that matching in capture-recapture studies is adversely affected by misperceptions regarding the application of statutory/regulatory eligibility criteria and by missing data. The result is that the reported undercounts in both the BLS SOII and several state WC databases are overstated in the capture-recapture studies. Although three of four databases can be approximately harmonized, the fourth cannot. CONCLUSIONS More precisely targeted information is needed before decisions regarding redesign of the BLS survey are made or before legislative or administrative changes in the WC are contemplated.