; Voas, R.B.
; and Tippetts, A.S.
In Supplement to Alcoholism. Clinical and Experimental Research. 2000 Scientific Meeting of the Research Society on Alcoholism. June 24-29, 2000—Denver, CO
pgs. 114A , Santa Barbara, CA: Lippincott Williams & Wilkins (2000)Prior work has shown the alcohol interlock recorder (AIR) to be useful for studying the compliance of DUI offenders ordered to drive with an interlock as a condition of license reinstatement. The AIR is an event recorder, which logs driver BAC data, time of day, engine starts and stops. The interlock effectively suppresses new DUI convictions while installed but the rate of repeat offenses climbs sharply after the interlock is removed. The question under study here is how well the pattern of failed BAC tests (>40 mg/dl) logged by the AIR during the interlock period can be used to predict repeat offenses during the years following interlock removal. In a sample of 2159 interlock users in this Alberta, Canada study, interlocks were used for periods ranging 3-24 month. Our published work has shown that the overall highest proportion of failed BAC tests occur most often in the earliest few months of interlock use. Cox survival analysis has found the proportion of fails relative to start attempts in the first five months of interlock use is a potent predictor of repeat offenses at 12 months (R=.20, P<.0001) and 24 months (R=.22, P<.0001) after relicensing. Those who log the most failed tests on the AIR have familiar characteristics: more reported drinks per drinking day, more drinking days, more prior offenses. Half of the interlock sample never failed a BAC test. By 12 months after interlock removal only 1.5% of those with no AIR fails had a repeat DUI offense, but for the 15% of the sample with the highest rate of AIR fails, 6.5% had been reconvicted of DUI. This is more than 4 times as many repeats. Serial 3 month intervals of AIR BAC fails and warn data were examined but the earliest 5 months of failed BAC tests had the best predictive power for repeat offenses. This information may be of value to licensing authorities who make decisions about which interlock users should have their licenses reinstated after completing their sanction. Sensitivity analysis is underway to characterize the false negatives and positives of such a decision matrix.