The Relative Risk of Alcohol-Involved Crashes as a Function of Time of Day.

Abstract

Epidemiological research has repeatedly found that alcohol-impaired driving is associated with elevated risk of crash involvement in a dose-response fashion. Although experimental studies show that alcohol impairment of cognitive and psychomotor driving skills is exacerbated by sleep deprivation, there is less evidence that the combination of drowsiness and alcohol predicts actual motor vehicle crashes. We explored this by reanalyzing data from the National Highway Traffic Safety Administration Drug and Alcohol Crash Risk Study, constructing separate risk curves for daytime and nighttime hours. Controlling for driver demographics and driving exposure, and excluding other impairing drugs, we observed significantly greater risk of crash during the night versus the day at blood alcohol concentrations (BACs) between approximately .04 and .12 g/dl. Based on fitted point estimates, at .08 g/dl, the risk of crash at night was three times the risk of crash during the day. The association between BACs and crash risk was markedly different during daytime versus nighttime hours. Increased daytime risk was not observed until BACs exceeded the .08 g/dl per se legal limit. Results are interpreted as emphasizing the sedating role of alcohol. Implications are discussed.