Fifty-three child and adolescent psychiatric patients with depressive disorders were randomly allocated to brief cognitive-behaviour therapy (CBT) or to a control treatment, relaxation training. Forty-eight patients completed the treatment phase of the trial, which comprised 5-8 treatment sessions. Post-treatment assessments showed a clear advantage of CBT over relaxation on measures of both depression and overall outcome. However, there were no significant differences between the treatments on comorbid anxiety and conduct symptoms. At follow-up, the differences between the groups were reduced, partly because of a high relapse rate in the DTP group and partly because subjects in the relaxation group continued to recover.