The program used a multimodel treatment strategy that included daily educational presentations, individual and group counseling sessions, family counseling services, participation in the 12 steps of Alcoholics Anonymous and Narcotics Anonymous, and 6 months of aftercare. The program had a failure rate typical of such programs. The study participants were all released from the program during its first 4.25 years of operation between May 1988 and August 1992. The research extracted six factors from the independent variables and used the factor scores in a logistic regression analysis to determine the characteristics associated with client rearrest. Results indicated that socioeconomic status, criminality, and time at risk predicted the individual's success following treatment. In contrast, staff prognosis, maturity, child abuse, family history of alcohol abuse, and the number of prior drunk driving offenses did not consistently predict success. Findings indicate that the program would be well served if it screened out offenders with criminal backgrounds beyond drunk driving offenses. Tables and 38 references (Author abstract modified)