The Association Between Home and Crash Site Social Vulnerability on Injury and Mortality After Motor Vehicle Crashes: Implications for Traffic Policy

Introduction: There is a growing body of literature that shows geographic social vulnerability, which seeks to measure the resiliency of a community to withstand unforeseen disasters, may be associated with negative outcomes after traumatic injury. For motor vehicle collisions (MVCs) specifically, it is unknown how the resources of a patient's home environment may interact with resources of the environment where the crash occurred.

Methods: We merged publicly available crash data from the state of Michigan with the Michigan Trauma Quality Improvement dataset. A social vulnerability index (SVI) score was calculated for each ZIP code and was then cross-referenced between the location of the MVC (Crash-SVI) and the patient's home address (Home-SVI). SVI was divided into quintiles, with higher numbers indicating greater vulnerability. Adjusted logistic regression models using least absolute shrinkage and selection operator for feature selection and regularization were performed sequentially using patient, vehicular, and environmental variables to identify associations between Home-SVI and Crash-SVI, with mortality and injury severity score (ISS) greater than 15 (ISS15).

Results: Between January 2020 and December 2022, a total of 14,706 patients were identified. Most MVCs (75.3% of all patients) occurred in the second through fourth quintiles of SVI. In all cases, Crash-SVI occurred most frequently within the same quintile as the patient's Home-SVI. Average crash speed limits showed a significant negative association with increasing SVI. On adjusted logistic regression, there were significantly increased odds of mortality for the fifth quintile of Home-SVI in comparison to the first quintile when adjusted for patient factors; but this lost significance after the addition of vehicular or environmental variables. In contrast, there were decreased odds of ISS15 for the highest quintiles of Crash-SVI in all logistic regression models.

Conclusions: Geographic social vulnerability markers were associated with lower MVC-associated injury severity, perhaps in part because of the association with lower speed limit in these areas.