Critique du modèle additif de l'essai clinique randomisé

Randomized, double-blind, placebo-controlled clinical trials are currently the best way to demonstrate the clinical effectiveness of drugs. Its methodology relies on the method of difference (John Stuart Mill), through which the observed difference between two groups (drug vs placebo) can be attributed to the pharmacological effect of the drug being tested. However, this additive model can be questioned in the event of statistical interactions between the pharmacological and the placebo effects. Evidence in different domains has shown that the placebo effect can influence the effect of the active principle. This article evaluates the methodological, clinical and epistemological consequences of this phenomenon. Topics treated include extrapolating results, accounting for heterogeneous results, demonstrating the existence of several factors in the placebo effect, the necessity to take these factors into account for given symptoms or pathologies, as well as the problem of the “specific” effect.