[Modeling for clinical trial in psychiatry: rationale, benefits and limitations]

Drug development of new compounds implies to define the dosage as well as the conditions of their use (indication, treatment duration, drug interactions, warnings …). This information requires the identification of the time course response. The decisions made during the clinical phases are now based on mathematical models. These models are continuously described and improved during all phases of the drug development using data collected in healthy volunteers and patients. Their objectives are to describe the most precisely, the link between the compound characteristics (pharmacology), the patient demographics and the effects. Further, the natural history of the disease, the placebo effect and the probability of dropping out will be integrated into the model to optimize the evaluation of the compound. These technical improvements are not only statistical, in the sense that they allow a better understanding of the advantages and pitfalls of the new drug. This article presents these methods used in psychiatry and which will become the new standard of drug evaluation.