Linking occupational injury and illness data bases.
Rosenman-KD; Gardiner-JC; Reilly-MJ; Kalush-A; Reeves-M; Kalinowski-D
NIOSH 2004 Dec; :1-21
The national surveillance system for occupational injuries and illnesses, which is administered by the U. S. Department of Labor Bureau of Labor Statistics (BLS) is based on reports from employers. The national estimates are derived from a sampling strategy rather than a census of all work-related injuries and illnesses. In response to a National Academy of Sciences report in 1987 (NRC, 1987) which showed that the BLS national estimates missed 50% of acute work-related deaths, BLS began the Census of Fatal Occupational Injuries (CFOI). CFOI is a complete census that uses multiple data sources, covers all workers and is not dependent on an employer either being aware of the condition or responding to a survey. No such change was implemented to improve the national estimates for non-fatal work-related injuries and illnesses. There have been an increased number of studies documenting that the BLS national estimates for work-related injuries and illnesses undercount both chronic conditions and acute injuries (Boyle et al, 2000; Islam et al, 2003; Landrigan and Baker, 1991; Leigh et al, 1997; Leigh et al, 2004; Nelson et al, 1992; Park et al, 1992; Roscoe et al, 2002; Rosenman et al, 2003; Stanburyet al, 2003; Windau et al, 1991). These previous studies, which estimated undercounting by the BLS system have been based on comparison of counts of work-related injuries or illnesses in non-employer based data sources with the BLS estimates and not of actual matching of individuals identified in the different systems. In this study, adhering to the strict confidentiality rules of the BLS, we present the results of matching both individual workers and companies from non-employer based data sources with the actual individuals reported by the companies who participated in the BLS annual survey in Michigan in 1999, 2000 and 2001. We present the percentage of work-related injuries and illnesses missed in Michigan by the BLS survey estimates and use capture-recapture analysis to estimate the number of injuries and illnesses missed in Michigan by all data sources. We present this data for overall injuries and illnesses and also by specific conditions and industrial sectors.
Surveillance-programs; Information-retrieval-systems; Injuries; Diseases; Occupational-accidents; Occupational-diseases; Statistical-quality-control; Data-processing; Statistical-analysis
Kenneth D. Rosenman, MD, Michigan State University, 117 West Fee, East Lansing, MI 48824
Final Grant Report
NTIS Accession No.
National Institute for Occupational Safety and Health
Michigan State University, East Lansing, Michigan