A recent meta-analysis showed that antidepressants are more effective than placebo. In placebo-controlled trials, the placebo effect is subtracted from the verum effect to find the specific effect of each antidepressant. In clinical practice, however, the situation is the other way around: non-specific ('placebo') effects are added to the specific effect of the antidepressant and together these effects determine the outcome. The non-specific aspects comprise strength of the therapeutic relation, conviction of both doctor and patient in the diagnosis and treatment, and hope that treatment will induce remission. In 2006, it was found that the psychiatrist who gave the treatment had a larger effect on the outcome than the effect of verum plus placebo. Some psychiatrists induced improvement, even with placebo, while others made patients worse, even with verum. This implies that while choosing the right antidepressant is important, choosing the right psychiatrist is even more important. When a doctor loses faith in antidepressants, so will his patients and non-specific aspects will have nocebo rather than placebo effects.