Predicting Motor Vehicle Crash (MVC) Injury Severity
Overview
Motor vehicle crashes (MVC) occur frequently and can be deadly. Given the possibility of occupant injury, emergency response personnel are typically sent to the crash scene. Fortunately, the majority of crashes do not cau
se significant injury. But this means that the limited number of emergency vehicles may be deployed unnecessarily. If it were possible to predict the absence of injury, emergency response resources could be conserved and made more available when needed.
NWPERRC researchers are developing models capable of doing this kind of prediction. They are exploring how Advanced Automatic collision Notification (AACN) systems, such as OnStar, can be used to identify crash occupants who are uninjured or have minor injuries. Findings are intended to inform methods for improving the timeliness of crash notification and response.
This one-year project was funded in 2009 by the Centers for Disease Control and Prevention. It is being led by researchers at the Harborview Injury and Prevention Research Center.
Methods
Researchers are using nationally available MVC data to examine the feasibility of constructing an algorithm that accurately predicts MVCs with a low likelihood of occupant injury. The algorithms they are developing and testing use data that is currently or will soon be available in AACN systems. Researchers are assessing if the algorithms can generate immediate (prior to emergency response) feedback to emergency dispatch services. This feedback would identify MVCs with little risk of occupant injury and therefore no need of intensive emergency response. Researchers for this project will also identify the costs and benefits of utilizing the algorithms across a wide variety of MVC types.
Researchers
Mary A. Kernic, PhD, MPH, Project Director/Principal Investigator
Beth Ebel, MD, MSc, MPH, Co-Investigator
Christopher D. Mack, MS, Research Consultant
Robert Kaufman, BS, Project Coordinator
For questions about this project, please contact Mary Kernic at mkernic@u.washington.edu
