OBJECTIVE: Alcohol-impaired driving contributes to more than 10 000 fatalities in the United States each year. This research estimated the unique effect of enforcement intensity on reductions in alcohol-impaired fatal crashes.
METHODS: We collected data from 30 states (including the District of Columbia) that experienced the greatest changes in alcohol-impaired fatal crashes from 1996 to 2006. Mixed-model regression was used to examine the extent to which year-over-year changes in the intensity of impaired driving enforcement predicted year-over-year reductions of drivers killed in alcohol-involved fatal crashes.
METHODS: Data from 30 states were obtained online.
METHODS: Aggregate state-level data from a total of 279 state-year combinations were analyzed.
METHODS: Our dependent measure was the ratio of drivers involved in fatal crashes with blood alcohol concentrations (BACs) ≥ 0.08 g/dl over drivers involved in fatal crashes with BACs = 0.00 g/dl. Per capita driving under the influence (DUI) arrests and traffic enforcement funding were the primary predictors. Covariates were estimated vehicle miles traveled (VMT); the proportional distributions of gender and racial/ethnic; geographic distribution; the proportion of drivers aged 21-34 years; median family income; and education level.
RESULTS: Analysis revealed that DUI arrests per capita uniquely and significantly predicted reductions in the ratio of fatal crashes (β = -0.753, t(238)  = 2.1, P < 0.05) after controlling the covariates. Exploratory analysis suggests the increase in arrest rates was associated with stronger reductions in urban versus rural settings.
CONCLUSIONS: Drunk driving enforcement intensity uniquely contributes to reductions in alcohol-impaired crash fatalities after controlling for other factors.

Driving under the influence of alcohol on curved roadway segments has a higher risk than driving on straight segments. To explore the effect of different breath alcohol concentration (BrAC) levels on driving performance in roadway curves, a driving simulation experiment was designed to collect 25 participants\' driving performance parameters (i.e., speed and lane position) under the influence of 4 BrAC levels (0.00%, 0.03%, 0.06% and 0.09%) on 6 types of roadway curves (3 radii×2 turning directions). Driving performance data for 22 participants were collected successfully. Then the average and standard deviation of the two parameters were analyzed, considering the entire curve and different sections of the curve, respectively. The results show that the speed throughout curves is higher when drinking and driving than during sober driving. The significant interaction between alcohol and radius exists in the middle and tangent segments after a curve exit, indicating that a small radius can reduce speed at high BrAC levels. The significant impairment of alcohol on the stability of speed occurs mainly in the curve section between the point of curve (PC) and point of tangent (PT), with no impairment noted in tangent sections. The stability of speed is significantly worsened at higher BrAC levels. Alcohol and radius have interactive effects on the standard deviation of speed in the entry segment of curves, indicating that the small radius amplifies the instability of speed at high BrAC levels. For lateral movement, drivers tend to travel on the right side of the lane when drinking and driving, mainly in the approach and middle segments of curves. Higher BrAC levels worsen the stability of lateral movement in every segment of the curve, regardless of its radius and turning direction. The results are expected to provide reference for detecting the drinking and driving state.