Symptom trajectories in discontinuation trials.

Abstract

Many patients with clinical depression, particularly those with recurrent illness, are treated with maintenance antidepressants after an acute-phase response. But how effective is this approach? The effectiveness of maintenance treatment is most commonly tested in trials with a discontinuation design, 1 in which patients who have already responded to a treatment are randomly assigned to either ongoing treatment or placebo. Such studies provide a measure of the benefit of maintenance treatment over discontinuation. However, discontinuation studies have several limitations. The first is common to psychiatric treatment studies generally; how and when should we assess whether a patient's symptoms have changed after discontinuation? The second is more specific to discontinuation designs; the selection of patients based on a positive response to that treatment is likely to inflate the apparent beneficial effect of the treatment, because that subgroup of patients has the most to lose from stopping treatment.