Background The recommendations in clinical guidelines for duration of therapy for alcohol use disorder (AUD) are based on consensus decisions. In reality, we do not know the optimal duration of an alcohol treatment course. Methods A systematic review and meta-regression of randomized controlled trials of psychosocial treatment in alcohol outpatient treatment centers. The population consisted of adults suffering from AUD, treated in an outpatient facility with at least two sessions of therapy. Meta-regression analysis was performed with treatment outcome as a function of duration of therapy across studies. Treatment outcome was defined as long-term alcohol use measured in percentage of days abstinent (PDA), percentage of heavy days drinking (PHD), and/or proportion of participants abstinent (ABS). Results 48 studies encompassing 8984 participants. Mean planned duration of therapy: 18 (8–82) weeks and 14 (2–36) sessions. Mean actual attended sessions: 9 (1–26). Mean follow-up time: 43 (8–104) weeks with a mean of 6 (2–18) research assessments. Neither planned weeks, duration of sessions, frequency of sessions per week, nor actual attended sessions were associated with long-term alcohol use outcomes. However, frequency of research assessments was positively associated with PDA and PHD. Conclusion No associations between long-term alcohol use outcomes and planned or actual attended duration of psychosocial treatment in outpatient care. Research assessments and, accordingly, the research project in itself may influence outcome in studies of psychosocial treatment for alcohol use disorder.