Exposure-response modeling of anti-depressant treatments: the confounding role of placebo effect.

Emerging evidence indicates that the selection of patients and the selection of recruitment centers in placebo-controlled, randomized clinical trials (RCTs) in depression have a substantial impact on the probability of observing a treatment effect of a novel medication. The objective of this work was to evaluate the role of placebo in characterizing the exposure-response relationship and proposes a new methodology based on band-pass filtering for assessing the relationship between drug exposure, level of clinical response conditioned to the level of placebo response, and the potency of the antidepressant drug evaluated. Clinical trial simulation (CTS) was used to demonstrate that the conventional analyses of data generated in multi-centre RCTs including centres exhibiting heterogeneous placebo response can lead to contradictory and inappropriate assessment of the exposure-response relationship. To address this issue, a novel modelling approach has been developed for establishing the exposure-response relationship by using a pharmacodynamic model accounting for the placebo effect. The proposed model demonstrated that the expected drug related treatment effect in a placebo-controlled RCT can be predicted as a function of the dose, the drug potency and the level of placebo response when data from informative recruitment centres are considered. With this novel approach, a more accurate estimate of the dose/exposure response can be derived and more informed go/no-go decisions can be made in developing drugs for psychiatric disorders.