Sequential Parallel Comparison Design in Major Depressive Disorder Trials

• Activity Number: 394
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Yangchun Du*+ and Asli Memisoglu and Marc de Somer and Randall Marshall and Richard Leigh-Pemberton and Bernard Silverman and Elliot Ehrich and Fava Maurizio
• Companies: Alkermes and Alkermes and Alkermes and Alkermes and Alkermes and Alkermes and Alkermes and Massachusetts General Hospital
• Keywords: SPCD ; placebo response ; MDD

Abstract

The Sequential Parallel Comparison Design (SPCD) can improve signal detection in research populations with a high placebo response, typical of psychiatric clinical trials. SPCD stage 1 contrasts active drug to placebo in all subjects. Placebo non-responders in stage 1 are re-randomized in stage 2 to active drug or placebo. Inference is based on the weighted linear combination of stage 1 and stage 2 effect estimates. Prior statistical research has shown the validity and performance of SPCD.

The present simulation study expands the SPCD assumptions, informed by results of a completed phase 2 SPCD trial in major depressive disorder. Type I error rate was well-controlled using MMRM with an unstructured covariance matrix. Type I error control did not vary appreciably over a broad range of weights used to combine the stage estimates. The correlation between treatment effect estimates from each stage was negligible, supporting the validity of inference based on the combination statistics.

SPCD mitigates risk of clinical trial failure due to high placebo response and increases efficiency, while ensuring valid and reliable inference, including for confirmatory trials.