Police-reported pedestrian crash matching and injury severity misclassification by body region in New Mexico, USA.

Between 2009 and 2019, pedestrian fatalities in the U.S. increased 51.0% while all other traffic fatalities increased 0.4%. To mitigate pedestrian safety issues, practitioners increasingly use police-reported data to identify and treat locations that experience either serious or fatal injuries. We investigated how many and which types of pedestrian injuries were misclassified by police-reported data in New Mexico between 2014 and 2018 by matching pedestrian-vehicle crash victims reported in New Mexico Department of Transportation (NMDOT) crash data to patients treated at University of New Mexico Health-Science Center, an American College of Surgeons-certified level 1 trauma center (n = 3097 pedestrians in NMDOT data; n = 512 matched pedestrians). Findings suggest that injuries involving older pedestrians, males, alcohol, more serious injuries, and those that occur at night are more likely to match to the hospital data. Of the non-fatally injured pedestrians who police estimated as seriously-injured (n = 207), 21.7% were no more than minorly-injured (n = 45) (KABCO A and ISS < 9). Of pedestrians who police estimated as minorly-injured (n = 239), 55.6% were seriously-injured (n = 133) (KABCO B,C,O and ISS ≥ 9). Of pedestrians with true serious injuries (n = 295) (ISS ≥ 9), 45.1% were under-estimated by police (n = 133) (KABCO B,C,O and ISS ≥ 9) whereas 29.8% of pedestrians with true minor injuries (n = 151) (ISS < 9) were over-estimated by police (n = 45) (KABCO A and ISS < 9). Minorly-injured pedestrians who were over-estimated by police (KABCO A and ISS < 9) were more likely to have lower extremity injuries (62.2% vs 42.5%, p-value = 0.013) compared to minorly-injured pedestrians whose injury severities were estimated correctly (KABCO B,C,O and ISS < 9). Seriously-injured pedestrians who were under-estimated (KABCO B,C,O and ISS ≥ 9) were less likely to have injuries to the head (39.8% vs. 55.6%, p-value = 0.003), spine (30.1% vs. 50.0%, p-value < 0.001), thorax (53.4% vs. 66.7%, p-value = 0.0139), or abdomen (18.8% vs. 32.1%, p-value = 0.005) compared to seriously-injured pedestrians whose injury severities were estimated correctly (KABCO A and ISS ≥ 9). This research illustrates the importance of linking police and health outcome databases to provide a more complete understanding of traffic safety.