McCartt, Anne T.; Farmer, Charles M.; Eichelberger, Angela H.
Proceedings of the 20th International Conference on Alcohol, Drugs, and Traffic Safety (CD-ROM)
The proportion of U.S. drivers in fatal crashes with blood alcohol concentrations (BACs) of 0.08% or higher declined substantially from 1982 through the mid-1990s but changed little since, whereas periodic national roadside surveys reported large declines in alcohol use.Aims:
To examine possible explanations for the apparent disconnect between trends in drinking and driving and trends in the proportion of drivers with BACs =0.08% in fatal crashes.Methods:
The study looked at data from national roadside surveys conducted on Friday and Saturday late nights (10 p.m.-3 a.m.) in 1986, 1996, and 2007 and data on fatal crashes for these times and years. Differences in the roadside survey samples and protocols were examined, as were trends in rates of BACs =0.08% among driver sub-groups in the roadside surveys and FARS.Results:
Even after adjustments to the roadside surveys for differences in sampling and weighting, there were large declines in drivers with BACs =0.08% (5.4% in 1986, 4.2% in 1996, 2.6% in 2007) not reflected in rates of drivers with BACs =0.08% in fatal crashes (55.3% in 1986, 52.1% in 1996, 49.7% in 2007). Drivers with BACs =0.08% in fatal crashes were more likely than other drivers to drive older vehicles, be speeding, be unrestrained, and have prior moving violations and crashes/suspensions. Except for belt use, the differences were fairly consistent over time. Belt use increased more slowly among drivers with BACs =0.08%.Conclusions:
Different methodologies of roadside surveys do not appear to explain why alcohol impairment in fatal crashes has declined much less than the rate of drinking drivers in general. Trends generally were similar for sub-groups of drivers in roadside surveys and in fatal crashes. The fact that drinking drivers have been slower to buckle up appears to account for some of the discrepancy in trends, but there also appear to be other unknown factors.