Background: Research on group therapy indicates that various dimensions of the helpful relationship qualities (cohesion, climate, empathy, alliance) are associated with outcome. However, the use of a wide variety of empirical scales makes comparisons between studies as well as generalizations somewhat difficult. Although a generic, trans-theoretical measure such as the Group Climate Questionnaire-Short Form (GCQ-S; MacKenzie, 1983) is available and applicable to most treatment conditions, it has never been tested with cognitive-behavioural group therapy. Aims: To investigate perceived dimensions of group climate (engagement, avoidance and conflict) as predictors of long-term (1 year) follow-up in a manualized, structured time-limited cognitive-behavioural group therapy (CBGT) for out-patients with comorbid psychiatric disorders. Methods: Data from 27 patients were analysed using hierarchical multiple regression analyses. Outcome measures used were general symptomatic complaints (SCL-90-R), interpersonal problems (IIP-64), specific mood- and anxiety symptoms (BDI; BAI) and early maladaptive schemas (YSQ). After controlling for scores on the relevant dependent variables at both intake and treatment termination, dimensions of group climate measured close to termination were entered as predictors in separate analyses. Results: Higher ratings of engagement were associated with reduced scores on all outcome measures at follow-up, except for anxiety symptoms (BAI). Higher ratings of avoidance were associated with lower anxiety symptoms at follow up, whereas ratings of conflict were unrelated to all follow-up scores. Conclusions: The results provide partial support for the use of the GCQ-S as a predictor of long-term follow-up in CBGT, and highlights perceived engagement as the most important dimension. Clinical implications are discussed.