Comparative and joint prediction of DUI recidivism from alcohol ignition interlock and driver records.

Abstract

This work was conducted to find practical predictors that anticipate which driving under the influence (DUI) offenders will continue to drink and drive after a period of alcohol ignition interlock-controlled driving ends. The interlock prevents impaired driving by requiring a low blood alcohol concentration (BAC) breath sample before allowing an engine to start. Each breath test is recorded. The study evaluated the interlock record as a predictor of future DUI offenses relative to driver records and self-report items. Subjects were 2,273 DUI offenders in Alberta, Canada, who used an interlock to gain full reinstatement of driving privileges; for 2,134, the installed periods ranged from 5 to 30 months. A median of 8.1 breath tests was logged for each installed day; 9.9 tests were taken on each day of vehicle use (4.3 starts plus 5.6 running retests). Predictors of postinterlock repeat DUI were compared by sensitivity and survival analyses. Although 69% of all interlock users had at least one BAC test > or = .04% (a "fail" test) within the first 5 months, only 9% were reconvicted up to 4 years after interlock removal. Failed interlock tests proportional to all BAC tests taken was the best predictor of driver recidivism risk during the years following interlock removal. The interlock record provides new information, particularly about drivers with no prior DUI offenses. Prior moving violations and driving while suspended convictions, although better predictors than questionnaire data, were poorer than interlock records and prior DUI offenses. The alcohol interlock, already recognized as a useful control device, warrants attention for DUI prediction as well.