Thase et al use a sophisticated model to assess the ‘true’ effect of active antidepressant therapy v. placebo. ^{Reference Thase, Larsen and Kennedy1}

Health authorities generally evaluate the efficacy of new medications from randomised controlled trials (RCTs) v. placebo which are well documented and rely on such a simple statistical paradigm that they can resist the major financial conflicts of interest inherent in the evaluation of pharmaceuticals. Concerning antidepressants, these studies generally identify small, average drug–placebo differences. ^{Reference Kirsch, Deacon, Huedo-Medina, Scoboria, Moore and Johnson2}

Using statistical modelling, other authors have addressed the question of outcome measurement ^{Reference Moncrieff and Kirsch3} and found that efficacy is better understood as a large effect in a subgroup of patients. This is consistent with the common clinical viewpoint.

However, Thase et al's model leads to a curious phenomenon: everything happens as if some patients were considered as non-benefiters, whereas their final score is markedly less than the score for patients considered as benefiters. As they state, ‘Essentially, all models are wrong, but some are useful’ a ‘true’ effect of active antidepressant v. placebo be built on such a ‘wrong’ model?

Surely not for a health authority. Nevertheless, it could be useful for researchers and clinicians as it generates hypotheses on the manner in which antidepressants are different from placebo. In this view, it is necessary to go further and compare the characteristics of benefiters with non-benefiters with two additional perspectives:

1. 1 to perform RCTs in populations of benefiters in order to maximise the signal and to minimise the noise – this could help to limit the number of ‘negative studies’;

2. 2 to use antidepressants only in this subpopulation of treatment benefiters and to propose alternatives to other patients (e.g. psychotherapy, repetitive transcranial magnetic stimulation, electroconvulsive therapy).

Finally, Thase et al's model is based on RCTs which if applied to major depressive disorder raises fundamental questions regarding internal ^{Reference Ioannidis4} and external validity. ^{Reference Naudet, Maria and Falissard5} Even if a ‘true’ effect of active antidepressants exists, I'm not sure that it could be derived from RCTs.