In an attempt to build on earlier work that outlined relapse trajectories in adolescents discharged from residential treatment, this study empirically derived relapse trajectory groups among a population of adolescents discharged from outpatient treatment across different substances. The study used cluster analysis over five follow-up waves that extended out to 30 months after intake, in order to identify homogeneous groups based on multiple alcohol and other drug use variables, along with days in a controlled environment. A sample of 563 adolescents, age 12 to 18, who were assigned randomly to 1 of 5 outpatient interventions in 4 United States communities were examined with follow-up at 3, 6, 9, 12, and 30 months post intake. Data were drawn from adolescent interviews using the Global Appraisal of Individual Needs (GAIN) with the use of a cluster analysis to derive empirically a typology of five different post-treatment relapse trajectories. It was found that adolescents could be grouped empirically into five relapse trajectories over 30 months following outpatient treatment. Heterogeneity was found in the relapse trajectories, varying by the initial response to treatment, the stability of the response, and an interaction with time in a controlled environment. Cannabis and alcohol were found to be the two main substances involved. Tables, references