The prevention paradox, traffic safety, and driving-while-intoxicated treatment

In San Juan County, New Mexico, a 28-day jail/treatment program for first-time driving-while-intoxicated (DWI) offenders was established in 1994 to reduce both DWI recidivism and alcohol-related crashes. This paper assesses the impact of the program on both outcomes.

The data are composed of driving records of all people arrested for DWI in San Juan County from August 1994 through December 2001. Subsequent re-arrests and crashes were analyzed to compare people who had been sentenced to the jail/treatment program and those who had not. Kaplan-Meier survival curves and Cox proportional hazards regression analyses were used. Covariates included age, gender, blood alcohol content (BAC), number of prior arrests, and ethnicity (Native American, non-Hispanic white, and Hispanic).

Re-arrest rates were significantly lower for the treatment than the nontreatment group. Each of the three major ethnic groups showed similar effects. This was not observed for subsequent alcohol-related crashes, possibly as a result of insufficient numbers. BAC and number of previous arrests were, however, significant risk factors for subsequent crashes. Finally, although BAC and previous arrests were important risk factors for subsequent crashes, the vast majority of subsequent alcohol-related crashes occurred among people in the intermediate risk ranges.

The jail/treatment program is effective in reducing the probability of DWI re-arrests. The evidence with respect to crashes is equivocal. That most crashes occur to people in the intermediate risk range exemplifies the prevention paradox, and means that the courts, which deal most severely with high-risk individuals, cannot be expected to have a major impact on alcohol-related crashes.