Alcohol involvement is a major contributing factor in the occurrence of traffic crashes. Alcohol has been found to be more prevalent in fatal crashes than in personal injury and property-damage-only crashes. The Fatality Analysis Reporting System (FARS) collects information on all fatal motor vehicle crashes that occur on public roads, if the fatality occurs within 30 days of the data of the crash. The most direct measure of a driver's or a nonoccupant's (pedestrian's or pedalcyclist's) alcohol involvement is a known BAC test result, either based on breath tests administered by police, or blood tests. BAC test results for many drivers and nonoccupants involved in fatal crashes are not known. The significant number of missing BAC values greatly inhibits the ability to describe the extent and trends of alcohol involvement in fatal crashes, to identify high-risk groups and times for targeting countermeasures, and to evaluate the effectiveness of anti-drunk driving programs. This Research Note describes the implementation of a new methodology that generates multiple imputations of missing Blood Alcohol Concentration values in FARS. (A)
Abstract