OBJECTIVES: This research uses Cox regression to analyze relapse patterns of adults treated for substance use disorder (SUD). The objective is to evaluate the role psychosocial, treatment and environmental characteristics play in the relapse process. While relapse is a much-studied phenomenon, there is no research describing in detail how the risk of relapse changes over time (especially in the months immediately following primary treatment) given known protective factors. Since SUD is a chronic relapsing disorder, it is important to understand ways treatment gains can be maintained. METHODS: Subjects are 408 adults discharged between 2000-2009 from an ASAM-defined Level 1.A primary inpatient treatment program. Data were collected via a 215-item questionnaire as part of the treatment program’s annual outcomes evaluation. The sampling frame was people who successfully completed treatment and who gave consent. The response rate was 56 percent. The researchers obtained the treatment records of each person completing the questionnaire and matched the treatment outcomes from the questionnaire to treatment and sociodemographic variables present in the treatment records. A comprehensive data set was created from these two sources. Data were analyzed using Cox proportional hazard regression. RESULTS: Statistically significant covariates include: home environment variables (those with supportive people at home are 3.17 times less likely to relapse; those with supportive spouses are 4.91 times less likely to relapse) and support group variables (those attending self-help meetings are 1.61 times less likely to relapse; those with sponsors are 1.64 times less likely to relapse), cetertis paribus. Adjustor covariates include: Global Assessment of Functioning, addiction severity, DSM diagnosis and length of stay. CONCLUSIONS: Since this survival model relates the time that passes before relapse occurs to known risk factors for relapse, this research will help treatment providers identify not only the best types of interventions but also the correct timing of interventions.