Do we underestimate the benefits of antidepressants?

Do we underestimate the benefi ts of antidepressants? In the past 5 years, doubts have been raised about the therapeutic eff ectiveness of antidepressants in patients with depressive disorders, because of the small diff erences in symptom improvement between antidepressants and placebo recorded in randomised controlled trials (RCTs). With the recent debates about lowering of disease thresholds in the Diagnostic and Statistical Manual of Mental Disorders, fi fth edition, and the medicalisation of normal bereavement, this scepticism has increased. For the large group of patients with mild depression, the diff erences between antidepressants and placebo are not thought to be large enough to be clinically signifi cant—ie, at least three points on the Hamilton Depression Rating Scale, HAMD-17. 1 Therefore, several guidelines 2,3 no longer recommend antidepressants as fi rst-line treatment for patients with mild and moderate depression, and instead generally favour psychotherapy. We are concerned that scepticism about the benefi ts of antidepressants goes too far, and risks depriving many patients with depression of eff ective treatment. The crucial question for both patients and doctors is how much symptom improvement can a patient expect if he or she tolerates and reliably takes antidepressants, compared with watchful waiting, psychotherapy, or other alternatives? The present approach to estimation of the clinical signifi cance of antidepressants is misleading, for several reasons. 4 Induction of hope of treatment benefi t is an important factor in antidepressant treatment. 5 In double-masked RCTs, less hope is induced in patients receiving an active drug than in routine care, because patients are left uncertain about which treatment group they belong to; however, more hope is induced in those receiving placebo compared with those who are undergoing watchful waiting, because the pill could be an antidepressant. Both factors together reduce the antidepressant–placebo diff erence in RCTs compared with the unmasked real-life situation. Clinical signifi cance of therapeutic eff ects of anti-depressants is estimated on the basis of antidepressant– placebo diff erences 6 in improvement of depressive symptoms. In this context, the method of intention-to-treat analysis with last observation carried forward is usually applied. This conservative method is clearly useful to prove effi cacy. However, the benefi t that a compliant patient can expect from antidepressants in routine care cannot be estimated from study samples which include early drop outs for various study-related reasons. A better approach to estimate real-life benefi t would be to include only those who …