Improving occupational injury surveillance by using a severity threshold: development of a new occupational health indicator.

BACKGROUND Hospital discharge data are used for occupational injury surveillance, but observed hospitalisation trends are affected by trends in healthcare practices and workers' compensation coverage that may increasingly impair ascertainment of minor injuries relative to severe injuries. The objectives of this study were to (1) describe the development of a severe injury definition for surveillance purposes and (2) assess the impact of imposing a severity threshold on estimated occupational and non-occupational injury trends. METHODS Three independent methods were used to estimate injury severity for the severe injury definition. 10 population-based hospital discharge databases were used to estimate trends (1998-2009), including the National Hospital Discharge Survey (NHDS) and State Inpatient Databases (SID) from the Healthcare Cost and Utilization Project (HCUP), Agency for Healthcare Research and Quality. Negative binomial regression was used to model injury trends with and without severity restriction and to test trend divergence by severity. RESULTS Trend estimates for occupational injuries were biased downwards in the absence of severity restriction, more so than for non-occupational injuries. Imposing a severity threshold resulted in a markedly different historical picture. CONCLUSIONS Severity restriction can be used as an injury surveillance methodology to increase the accuracy of trend estimates, which can then be used by occupational health researchers, practitioners and policy-makers to identify prevention opportunities and to support state and national investments in occupational injury prevention efforts. The newly adopted state-based occupational health indicator, 'Work-Related Severe Traumatic Injury Hospitalizations', incorporates a severity threshold that will reduce temporal ascertainment threats to accurate trend estimates.