The prevalence of alcohol-involved crashes across high and low complexity road environments: Does knowing where drinking drivers crash help explain why they crash?

OBJECTIVE: Alcohol use has been linked to impairment of cognitive and psychomotor driving skills, yet the extent to which skill impairment contributes to actual crashes is unknown. A reasonable assumption is that some driving situations have higher skill demands than others. We contend that intersections, the presence of other vehicles or moving objects, and work zones are examples of common situations with higher skill demands. Accordingly, if skill deficits are largely responsible for alcohol-involved crashes, crashes involving a drinking driver (versus only sober drivers) should be overrepresented in these driving situations. METHOD: Publicly available FARS data from 2010 to 2017 were collected. Fatal crashes were coded as alcohol-involved (1+ driver with a blood alcohol concentration [BAC] ≥ .05 g/dl) or having no impaired driver (BACs = .000). Drug-positive crashes were excluded. Crashes were also coded as involving moving versus stationary objects, occurring at versus away from intersections, being multivehicle versus single vehicle, occurring at or away from work zones. RESULTS: Across multiple models, controlling for time of day and type of road, alcohol-involved crashes were significantly underrepresented in crashes at intersections, with moving objects, and other vehicles. Most strikingly, alcohol-involved crashes were 24 percentage points more likely to be with a stationary object than a moving object. CONCLUSIONS: No evidence supported the idea that skill reductions are a primary contributor to alcohol-involved crashes. Alternative explanations and limitations are discussed.