The program was located in a medium-security Federal correctional institution. Eighty-two participants remained in prison over the entire 2-year followup period, while 36 were released and lived in the community an average of 8 months prior to the end of the period. Results revealed a marginally significant predictive effect. Participants with elevated DLSI scores of 12 or higher displayed more subsequent alcohol and drug misuse (21.7 percent) than participants who had lower scores (9.7 percent). The LCSF had a somewhat stronger predictive relationship with future outcome in that 27.3 percent of the participants with high scores (10 or higher) subsequently abused alcohol or other drugs, compared to only 4.4 percent of the low-scoring participants. Regression analysis also revealed the superiority of the LCSF in predicting future outcome. Findings suggest that criminal background should be considered when assessing the relapse potential of drug- involved offenders once they leave treatment. Tables and 8 references (Author abstract modified)