Background: Addiction‐focused mutual‐help group participation is associated with better substance use disorder (SUD) treatment outcomes. However, little has been documented regarding which types of mutual‐help organizations patients attend, what levels of participation may be beneficial, and which patients, in particular, are more or less likely to participate. Furthermore, much of the evidence supporting the use of these organizations comes from studies examining participation and outcomes concurrently, raising doubts about cause–effect connections, and little is known about influences that may moderate the degree of any general benefit.

Method: Alcohol‐dependent outpatients (N=227; 27% female; M age=42) enrolled in a randomized‐controlled telephone case monitoring trial were assessed at treatment intake and at 1, 2, and 3 years postdischarge. Lagged‐panel, hierarchical linear models tested whether mutual‐help group participation in the first and second year following treatment predicted subsequent outcomes and whether these effects were moderated by gender, concurrent axis I diagnosis, religious preference, and prior mutual‐help experience. Robust regression curve analysis was used to examine dose–response relationships between mutual‐help and outcomes.

Results: Mutual‐help participation was associated with both greater abstinence and fewer drinks per drinking day and this relationship was not found to be influenced by gender, Axis I diagnosis, religious preference, or prior mutual‐help participation. Mutual‐help participants attended predominantly Alcoholics Anonymous and tended to be Caucasian, be more educated, have prior mutual‐help experience, and have more severe alcohol involvement. Dose–response curve analyses suggested that even small amounts of participation may be helpful in increasing abstinence, whereas higher doses may be needed to reduce relapse intensity.

Conclusions: Use of mutual‐help groups following intensive outpatient SUD treatment appears to be beneficial for many different types of patients and even modest levels of participation may be helpful. Future emphasis should be placed on ways to engage individuals with these cost‐effective resources over time and to gather and disseminate evidence regarding additional mutual‐help organizations.