Research In Action

Motor vehicle crashes are a leading cause of injury and death for both children and adults, and complete and accurate crash injury data are critical for effective prevention and intervention efforts. Many researchers and public health professionals rely on either crash report data filed by responding officers or on hospital discharge data, but each data source alone has limitations and may produce a biased undercount of the crash-injured population.

For example, hospital discharge data do not include factors from the scene of the crash that may be associated with injury (such as seat belt use) and may miss minor injuries for which no medical care was sought. Conversely, crash report data typically classify injuries in broad categories and may miss injuries that are only able to be diagnosed in a clinical setting (such as concussion).

Linking Data for a More Complete Picture

Linking hospital discharge data with crash reports give us a more complete picture of the “true” number of crash-related injuries, as well as an understanding of which groups are underrepresented in each source. A long-term goal of our team at CIRP is to develop novel sources of data that overcome these limitations and biases and effectively inform prevention strategies.

For these reasons, we leveraged an integrated data source, the New Jersey Safety and Health Outcomes (NJ-SHO) Data Warehouse, to examine the share of the entire crash-injured population captured by each of these two data source: overall, and by age, race/ethnicity, sex, injury severity, and road user type. This study, co-authored with my CIRP colleagues Melissa Pfeiffer, MPH; Kristina Metzger, PhD, MPH; Rachel Myers, PhD, MS; and Allison Curry, PhD, MPH, was recently published in Traffic Injury Prevention.

Highlights of the findings include:

• Over 260,000 individuals were injured in a motor vehicle crash during the 2016-2017 study period.
• Hospital discharge data and crash report data each captured nearly two-thirds of injured individuals, meaning that each data source alone missed over one-third of crash injuries.
• Compared with crash report data, hospital discharge data captured a larger share of individuals who were ages 65+, Black or Hispanic, had higher severity injuries, and were bicyclists or motorcyclists

These findings make clear the continued need for data integration efforts to classify the true burden of crash-related injuries and to better inform mitigation efforts—particularly related to populations most at risk for crash-related injuries.