Development of Software for Simulation and Implementation of Optimal Response-Adaptive Randomization Designs in Time-to-Event Trials

• Activity Number: 522
• Type: Contributed
• Date/Time: Wednesday, August 1, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Oleksandr Sverdlov*+ and Yevgen Ryeznik and Weng-Kee Wong
• Companies: Bristol-Myers Squibb and Kharkov National University of Economics and University of California at Los Angeles Fielding School of Public Health
• Keywords: Adaptive design ; Optimal allocation ; Multiple-objective clinical trial ; Time-to-event outcome ; Randomization design

Abstract

Response-adaptive randomization (RAR) designs are becoming increasingly popular in clinical trial practice due to their ability to handle multiple experimental objectives simultaneously. In this talk, we describe user interface software developed in MATLAB to facilitate implementation of optimal RAR designs in multi-arm clinical trials with multiple objectives and censored time-to-event primary outcomes. Our software has the following important utilities: 1). It can compute different types of optimal treatment allocation ratios for a user-defined set of experimental parameters. 2). It can simulate RAR procedures targeting selected optimal allocations. The simulated design characteristics include summary statistics and graphical displays of the distributions of treatment allocation proportions, estimators of the model parameters, design efficiencies, and ethical criteria. 3). It allows interactive recalculation of treatment randomization probabilities given history of treatment assignments and responses from patients in the trial. We demonstrate the utility of our software by redesigning a time-to-event clinical trial from the literature and make specific comparisons.