Objective: Alcohol-involved drivers or those with blood alcohol concentrations greater than 0.00 percent have more frequent and more severe crashes than other drivers. Alcohol use, because it delays perception and response and impairs coordination, increases the risk of a crash. However, those using alcohol may take additional driving risks, which may also lead to crashes. This study was done to learn whether risks besides alcohol involvement contributed to crash initiation and whether crash severity increased with alcohol involvement or with those other risky behaviors.
Methods: Data that represented nearly 1.4 million motor vehicle crashes were accessed from an NHTSA database. Analyses evaluated whether alcohol-involved driving was associated with other driving risks and whether driver alcohol involvement, alone or together with other risks, increased the likelihood of initiating a 2-vehicle crash or in the event of a crash or increased crash severity.
Results: Alcohol-involved drivers were less likely to use seat belts, drove faster, and were more likely to be distracted than others. Those who initiated 2-vehicle crashes were more likely to be alcohol involved or to have taken other driving risks than others from the same crashes. Crash severity was significantly greater for alcohol-involved drivers than for other drivers, but severity increased further if additional risks were taken. Crashes involving only drivers who had not used alcohol were also sometimes severe, and that severity was associated with risky driving behaviors. When crashes involved 2 drivers, the behaviors of both affected crash severity.
Conclusions: Risky driving behaviors, including alcohol involvement, increased the risk of a crash. Crash severity tended to increase with any risky behavior and to increase further with multiple risky behaviors. Other risky behaviors were associated with both alcohol involvement and crashes. Therefore, if effects from those other risky behaviors were not accommodated for, those effects would confound apparent associations between alcohol involvement and crashes. Therefore, this study's use of multivariate models that accommodated for effects from those other behaviors provided a truer picture of alcohol's association with crashes than simpler models would have. Supplemental materials are available for this article. Go to the publisher's online edition of Traffic Injury Prevention to view the supplemental file.