Using injury severity to improve occupational injury trend estimates.

ABSTRACT

BACKGROUND: Hospitalization-based estimates of trends in injury incidence are also affected by trends in health care practices and payer coverage that may differentially impact minor injuries. This study assessed whether implementing a severity threshold would improve occupational injury surveillance. METHODS: Hospital discharge data from four states and a national survey were used to identify traumatic injuries (1998-2009). Negative binomial regression was used to model injury trends with/without severity restriction, and to test trend divergence by severity. RESULTS: Trend estimates were generally biased downward in the absence of severity restriction, more so for occupational than non-occupational injuries. Restriction to severe injuries provided a markedly different overall picture of trends. CONCLUSIONS: Severity restriction may improve occupational injury trend estimates by reducing temporal biases such as increasingly restrictive hospital admission practices, constricting workers' compensation coverage, and decreasing identification/reporting of minor work-related injuries. Injury severity measures should be developed for occupational injury surveillance systems.