This article is an initial report from a review of alcohol and drug treatment studies with follow-ups of 2 years or more. The goals of the review are to examine the stability of substance use outcomes and the factors that moderate or mediate these outcomes. Results from 12 studies that generated multiple research reports are presented, and methodological problems encountered in the review are discussed. Substance use outcomes at the group level were generally stable, although moderate within-subject variation in substance use status over time was observed. Of factors assessed at baseline, psychiatric severity was a significant predictor of outcome in the highest percentage of reports, although the nature of the relationship varied. Stronger motivation and coping at baseline also consistently predicted better drinking outcomes. Better progress while in treatment, and the performance of pro-recovery behaviors and low problem severity in associated areas following treatment, consistently predicted better substance use outcomes.