Possibly, though that could miss other correlations that occur with race. I would want a study that also includes location, year of most recent road renovation, local visibility at time of fatality, distance traveled per day by car/bike/foot, and visibility of clothing, just to name a few. That said, putting together a more in-depth dataset may be impossible given the level of information collected. Looking for correlations with race, which is typically recorded in police reports following collisions, is a reasonable first step that can be used to motivate and drive further research.