Proposal for a Two-Stage Adaptive Dose-Response Trial Design Using Repeated Measures Data

• Activity Number: 516
• Type: Contributed
• Date/Time: Wednesday, August 3, 2011 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Anjela Tzontcheva*+ and Susan Huyck
• Companies: Merck Research Laboratories and Merck Research Laboratories
• Keywords: Dose-response characterization ; repeated measures ; simulation ; two-stage design ; adaptive design ; constrained longitudinal data analysis

Abstract

The chances of achieving successful late-stage drug development are greatly improved when the dose response has been definitively characterized allowing for the selection of the optimal dose for inclusion in confirmatory studies. Adequate characterization can be defined as identifying 4 dose levels where the lowest dose is not different than placebo, the 2 top doses plateau at the peak drug effect and the remaining dose is better than placebo, but lower than the plateau. Phase II dose-finding studies often represent the first large-scale evaluation of a new drug product in patients and adaptively designed studies provide opportunities for design adjustments that can occur because the true dose response range is unknown. This talk will present an adaptive dose response approach that improves the likelihood of adequate dose response characterization at the end of the study. Models are based on repeated measures data analyzed using a constrained longitudinal approach. Hypotheses first evaluated the presence of a dose response and then examine pairwise differences of drug vs. placebo. Simulations will be shown to demonstrate the operating characteristics.