Comparison of a Simple Bayesian Decision-Theoretic Design for Dose-Finding Trials with Traditional 3-Plus-3 Method

• Activity Number: 350
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Lynn Eudey*+ and Shenghua Kelly Fan and Tianyao Lu
• Companies: California State University, East Bay and California State University, East Bay and California State University, East Bay
• Keywords: Baysian adaptive-design ; dose-finding trial ; decision theory

Abstract

Phase I dose-finding trials using a new drug are conducted to study the safety, explore the pharmacokinetics and pharmacodynamics and to identify the maximum tolerated dose (MTD) of the new drug. MTD is defined as a specific percentile of the dose-toxicity curve. Due to ethical considerations the trials use adaptive designs, increasing the dose in patients based on patient results at all doses. The traditional 3 + 3 design is often used but is inefficient. Fan, Lu and Wang (2012, Statistics in Medicine) propose a simple and flexible Bayesian decision-theoretic design. Their method uses a working model with conjugate priors which produce analytic posterior distributions for monotonic dose-toxicity curves. Moreover, a utility function can be selected to reflect the interest of the trial. In this poster we present simulation results comparing their method with the traditional 3 + 3 method. The possible problem due to prior misspecification is also investigated.