R. Gomeni,1 N. Goyal,2 F. Bressolle,1 M. Fava3; 1Pharmacometrica, La Fouillade, France, 2GlaxoSmithKline, King of Prussia, PA, 3Massachusetts General Hospital, Boston, MA

BACKGROUND: The main reason for the inefficiency of multicenter randomized clinical trials (RCTs) in depression is the excessively high placebo response. The aim of this work was to propose a novel methodology to analyze RCT using the center-specific level of placebo response as a weighting factor, assuming that centers with high placebo response are less informative than the others for estimating the ‘true’ treatment effect (TE = difference between active and placebo).
METHODS: A linear mixed-effect modeling approach for repeated measures (MMRM) was used as a reference approach. The new method for estimating TE was based on a non-linear longitudinal modeling of the clinical scores (NLMMRM). NLMMRM estimates TE by associating to the data collected in each center a weighting factor defined by the level of placebo response in that center. The weight was defined by the posterior probability of detecting a clinically relevant separation of active treatment from placebo in that center [Clin Pharmacol Ther. 2008; 84(3):378-84].
RESULTS: Data from a RCT in depression (38 centers, placebo, low and high dose of a drug) was used to compare MMRM with NLMMRM. In this study, the level of placebo response was significantly high: ~55% of the subjects were classified as remitters. In absence of weighting factor MMRM and NLMMRM provided identical estimate of TE. However, when the weighting factor was included in the NLMMRM analysis a substantial improvement in the TE value was observed. The improvement was higher in the low dose arm (~40%) than in the higher dose arm (~6%).
CONCLUSION: The NLMMRM approach provides a tool for controlling the confounding effect of high placebo response and a more reliable estimate of the ‘true’ TE by controlling false negative results associated with excessively high placebo response.