Design Consideration in Phase 2 Dose Response Trial in the Presence of Possible Non-Monotonicity -- Comparison of Bayesian Emax and NDLM Model

• Activity Number: 32 - Statistical Methods in Dose-Finding Studies
• Type: Contributed
• Date/Time: Sunday, July 28, 2019 : 2:00 PM to 3:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Feng Liu* and Stephen Walters and Steven Julious
• Companies: AstraZeneca and University of Sheffield and University of Sheffield
• Keywords: Emax; NDLM; dose response

Abstract

Phase 2 clinical trials are critical in understanding the dose response relationship, so the optimal doses can then be selected for subsequent late phase trials. Emax dose response model was widely used in the phase 2 clinical trials of new drug development. Bayesian Emax model was found to be suboptimal to model the potential non-monotonical dose response and alternatives approaches of a Normal dynamic linear model (NDLM) is proposed. This paper compares the statistical properties of the Bayesian Emax model and NDLM model and both models are evaluated using simulation in the context of adaptive Phase 2 design under a variety of assumed dose response curves: linear, Emax like, U-shaped like, and flat response. It is shown that the Bayesian NDLM method is flexible and can handle a wide variety of dose-responses, including monotonic and non-monotonic relationships. The type I error, probability of incorrectly concluding a drug may work when it does not, is inflated with the Bayesian NDLM model in comparison to Emax model which would represent a development risk to pharmaceutical company.