Determination of characteristics of fatally injured drivers.

The objective of this study was to identify driver characteristics that can be used to predict driver risk of fatal crashes. The study had 3 components: (1) Comparisons were made among drivers who were killed in single-vehicle crashes or were at fault or not at fault in multi-vehicle crashes using 1,115 drivers in the 1993 National Mortality Followback Survey linked with drivers in the Fatality Analysis Reporting System (NMFS-FARS cases); (2) NMFS-FARS drivers killed on Friday/Saturday nights from 10 PM to 3 AM were compared with drivers in the 1996 National Roadside Survey; and (3) NMFS-FARS drivers were compared with telephone respondents to the 1993 National Survey of Drinking and Driving Attitudes and Behaviour. The following results were obtained: (1) The risk of being killed in a single-vehicle crash is about 3 times as great for drivers who either drink and drive at least weekly or typically drink 5 or more drinks per occasion, and 4 times as great for those who used illicit drugs. (2) Based on comparisons with roadside surveys from 10 PM to 3 AM on Friday and Saturday nights, drivers with blood alcohol concentrations (BACs) of 0.10% or higher are 64 times (CI=33.6-122.8) as likely as drivers with BACs of 0 to 0.04% to be in a crash in which they are killed; drivers who do not use lap or shoulder belts are about 13 times as likely to be in a fatal crash as drivers who wear both lap and shoulder belts. (3) Based on the telephone survey: the likelihood of becoming a fatally injured driver is about 3 times as great for people who drink and drive at least weekly or have 5+ drinks per occasion; among drivers ages 30-64, the likelihood of driver fatality is 4.5 times as great for those with a CAGE score of 2 or higher (indicative of alcohol dependence) compared with those with a CAGE less than 2. Alcohol-related risk factors were more important than demographic factors in predicting driver involvement in fatal crashes. The conclusion of this study was that, when other factors are controlled for, alcohol and seat belt use are the strongest predictors of drivers' risk of involvement in crashes in which they are killed.