This proposal examines the 3- and 5-year course of problems in an "intent to treat" sample of 400 adolescent intakes in four managed care substance abuse treatment sites, and assesses the relationship of substance abuse problems and treatment to medical utilization and cost. Baseline, 6- and 12-month follow-up data have been collected. The proposed study builds on the longitudinal and treatment outcome literature on adolescent substance abuse, and identifies the determinants of 5-year time paths of substance use patterns and problems. Time paths, or "trajectories" of substance use are of both theoretical and applied relevance to the adolescent drug treatment field, where substance abuse problems are increasingly viewed as chronic and relapsing, and at the same time, many individuals "mature out" of problems. Analyses draw on latent curve analysis techniques to compare the shape of trajectories of substance use and their determinants. We examine the influences that demographic characteristics, problem severity (i.e., substance use type and severity, medical and psychiataric comorbidities, social problems, and life stressors), treatment (i.e., starting treatment and length of stay of index treatment, and readmissions), and "extra-treatment" (e.g., peer influences, family functioning, twelve-step participation) factors have on the 5-year course of problems. Another important contribution of the proposed study to the longitudinal literature on adolescent substance abuse is the examination of co-morbid medical conditions, and their relationship to the course of adolescent substance abuse. We have also interviewed a parent at each wave of the study, and we examine collateral reports and validity of self-report through drug testing. The HMO has on-going computerized data on the dates and length of stay of treatment episodes and all psychiatric and medical services received, as well as their costs. We collect self-report data from teens and parents on treatment and medical service utilization not covered by the health plan. We follow adolescents for 5 years during the crucial "becoming young adult" period when many individuals "mature out" of problems, which allows us to examine the iindividual, environmental, and service factors that underlie this process, for those who succeed and those who do not.