Skip to content


  • Poster presentation
  • Open Access

Adaptive designs for phase II cross-over dose-finding trials using Bayesian model averaging

  • 1,
  • 1,
  • 1 and
  • 2
Trials201516 (Suppl 2) :P148

  • Published:


  • Placebo
  • Prior Distribution
  • Model Uncertainty
  • Interim Analysis
  • Candidate Model

Finding the right dose of a novel treatment is one of the most important tasks in early drug development. However, there is often uncertainty about the form of the relationship between dose and patient response at the time that the dose-finding trials are designed. In this presentation we develop Bayesian adaptive designs for Phase II cross-over trials conducted to estimate the minimum effective dose (MED), that is, the dose corresponding to the minimum clinically relevant effect over placebo.

We propose designs which stipulate that patients enter the trial in cohorts, and each cohort receives placebo and three active doses from a total of K active doses available for use. After each cohort of patients has been treated using a Williams square design an interim analysis is performed to identify which doses the next cohort of patients should receive. Bayesian model averaging is used to account for model uncertainty. Prior distributions will be based on pseudo data collected from experts by eliciting quantiles of the response distribution on placebo and a number of active doses to be used in the trial. A number of plausible models for the dose-response relationship are identified before the trial begins and accumulating evidence on the shape of this relationship is summarised by posterior model probabilities. Doses for the next cohort of patients are chosen adaptively to minimise the weighted average variance of the log(MED), averaging across all candidate models and weighting according to posterior model probabilities.

Authors’ Affiliations

Lancaster University, Lancaster, UK
Novartis, Basel, Switzerland


© Simpson et al. 2015

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated.


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate. Please note that comments may be removed without notice if they are flagged by another user or do not comply with our community guidelines.