Using Injury Severity to Improve Occupational Traumatic Injury Trend Estimates

BACKGROUND: Acute work-related trauma is a leading cause of death and disability among U.S. workers. Work-related injury rates are reportedly trending downward, but substantial evidence of increasing barriers to complete surveillance casts some doubt. Healthcare-based surveillance avoids some reporting filters, e.g., whether a workers’ compensation (WC) claim was filed/accepted. However, changing hospitalization practices may increasingly reduce capture of minor injuries but have little effect on severe injuries, which could create a false impression of declining trends in injury burden. The potential importance of severity restriction to occupational injury surveillance has not been systematically investigated. Washington State Trauma Registry-based exploratory research found significant upward trends in severe work-related injury rates, in contrast to flat trends when minor injuries were also included. The current study extends that research to assess the impact of severity restriction on occupational injury trend estimates using several population-based databases.

METHODS: Work-related traumatic injuries among those ages 16+ were identified in the National Hospital Discharge Survey (NHDS) and New Jersey’s hospital discharge records (1998-2009), based on first-listed diagnosis and primary payer=WC. Replication/validation using hospital discharge records for three other states is in progress. Severity was estimated from the first-listed ICD-9-CM code (severe=Abbreviated Injury Scale>2). Negative binomial regression was used to model injury rates and trends, with and without severity restriction. Temporal trend divergence by severity was tested using an interaction term.

 RESULTS: The trend for minor injuries diverged significantly downward compared with the trend for severe injuries in each data set/sample (P≤.004). Hospital discharge records for New Jersey residents contained 21,415 traumatic occupational injuries, with 4,174 classified as severe. For all injuries, the average annual decrease in age-adjusted injury rates was 5.1% [95% CI: -3.9%, -6.3%]. When restricted to severe injuries, the annual decrease was halved to 2.1% [95% CI: -1.3%, -2.8%]. The NHDS contained 4,237 survey records for traumatic occupational injuries, representing 555,052 nationally, with 132,783 classified as severe. For all injuries, the average annual decrease was 3.7% [95% CI: -1.6%, -5.8%], but the trend was essentially flat when restricted to severe injuries [0.4% increase; 95% CI: -2.7%, 3.5%]. 

CONCLUSIONS:  Restriction to severe injuries provides a different picture of occupational injury trends than does including all injury-related hospitalizations. Inadequate severity characterization coupled with increasingly limited hospitalization, WC coverage, and/or reporting of minor injuries may be contributing to unwarranted optimism about overall occupational injury trends. A new occupational health surveillance indicator for severe traumatic injuries will be developed.