Prevalence and Trends in Alcohol Use by Drivers in the US

"The comparison of the BAC [Blood Alcohol Content] test results from the four NRS [National Roadside Survey] studies suggests that, during the most recent decade, there continues to be a downward trend in the proportion of drivers with positive BACs²¹ on U.S. roads on weekend nights, from 36.1 percent in 1973, 25.9 percent in 1986, 16.9 percent in 1996, to a low of 12.4 percent in 2007. Though the response rates we achieved in the 2007 NRS are somewhat lower than NRS studies conducted in previous decades, they are still well above those obtained with Random Digit-Dialing telephone surveys, which currently are typically lower than 50 percent (Battaglia, Frankel, & Link, 2008). We also obtained PAS [Passive Alcohol Sensor] readings from well over 90 percent of these drivers who did not provide actual breath tests. This allowed us to impute BAC values for nearly every driver eligible for an interview. Since the 1996 NRS, the proportion of drivers with BACs .08 g/dL or above on the road has declined substantially from 4.3 percent in 1996 to 2.2 percent in 2007.
"Across the four NRS surveys (1973, 1986, 1996, and 2007), reductions in .08 g/dL and above drivers in the NRS have been generally paralleled by reductions in fatal alcohol-related crashes involving drivers with a BAC of .08 or greater. The reduction in nighttime NRS drivers with BAC .08 g/dL or above from 1996 to 2007 appears to be greater than the reduction in FARS [Fatality Analysis Reporting System] from 1996 to 2007. Results from the FARS data analyses show that drivers with a .08 g/dL or higher in fatal crashes changed from 33.1 percent in 1996 to 32 percent in 2007, whereas the percentage of drivers at or above .08 in the 1996 NRS was 4.3 and fell to 2.2 in 2007. This is a departure from the trends observed from past NRS studies in that, from 1973 to 1986 and then from 1986 to 1996, the same pattern of reductions was observed both in fatal crashes and in the NRS."