Simulations and Asymptotic Theory Illustrate Improvement on a Response-Adaptive Biased Coin Design

• Activity Number: 699
• Type: Contributed
• Date/Time: Thursday, August 13, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Aleksandra Stein*
• Keywords: phase III ; adaptive randomization ; biased coin ; ethical allocation

Abstract

Modern medical experiments accrue patients--and hence response data--throughout the duration of a trial. Designs which prospectively plan to modify patient allocation by leveraging accumulating outcome data are response-adaptive. This paper improves on an adaptive design that randomizes patients relative to current treatment estimates. This paper proposes an intuitive replacement for an arbitrary scaling constant that allows the design to allocate more patients to the superior treatment while mitigating the loss of power and increased bias inherent in many adaptive designs. This paper also expands the applicability of the original design by relaxing the assumptions of known, equal variances; of equal treatment-covariate interaction; and of immediate responses. The utility and benefits of the modified design are illustrated by a real-world application.