The question of the clinical effectiveness of ‘evidence based’ treatments is addressed by investigating the long-term outcome of a controlled study in which various treatment for panic disorder with agoraphobia were compared. Two years after completion of the trial patients were interviewed about their complaints and the amount of additional treatment they had received after the trial was assessed. Initially, short-term outcome of the trial had revealed that all treatments had a significant positive impact on the complaints. Medication (fluvoxamine) plus exposure in vivo had been more effective than placebo medication plus exposure, breathing retraining plus exposure, or exposure alone. The long-term results revealed no longer any difference between the treatment groups: the gains of medication plus exposure were maintained and patients of the other three groups had experienced further improvement. Further improvement was due to the additional treatment that was provided to the patients of the latter three groups. Furthermore, the mean number of additional sessions was well in excess of the 12 sessions of the original treatment. These results suggest that the clinical effectiveness of evidence based treatments is limited, even when their efficacy has been demonstrated in a controlled trial. Few patients recover from a short course of treatment and the vast majority of patients needs prolonged additional treatment. Finally, the issue is addressed whether medication is a useful adjunct to exposure therapy. The long-term results of the study suggest that medication accelerates the improvement, resulting in less need for additional treatment. However, continued use of medication is required for a substantial number of patients. Exposure treatment does not have this drawback and, on the long run, this treatment is equally effective to medication combined with exposure. Exposure does, however, require a prolonged treatment investment.