Skip to content


Bayes or Not Bayes, Is This the Question?

Chair:  Robert Barrier (Parexel)

Speaker: Lauren Kanapka, PhD (UNC)
Title: Bayesian Versus Frequentist Approaches to Basket Trials
A basket trial aims to expedite the drug development process by evaluating a new therapy in multiple populations within the same clinical trial. Each population, referred to as a “basket”, can be defined by disease type, biomarkers, or other patient characteristics. The objective of a basket trial is to identify the subset of baskets for which the new therapy shows promise. The conventional approach would be to analyze each of the baskets independently. Alternatively, several Bayesian dynamic borrowing methods have been proposed that share data across baskets when responses appear similar. These methods can achieve higher power than independent testing in exchange for a risk of some inflation in the type 1 error rate. We propose a frequentist approach to dynamic borrowing for basket trials using adaptive lasso. The proposed approach performs similar to Bayesian methods but is faster and is easier to implement. 

Speaker: Matt Psioda, PhD (GSK)
Title: A Hybrid Bayesian /Frequentist Approach for Augmenting Control Arms in Oncology Clinical Trials
Abstract: Use of Bayesian methods (e.g., flexible priors) to dynamically borrow information from existing data sources in the design and analysis of new clinical trials has increased dramatically in recent years. More recently, consistent with increased penetration of causal inference methodology into clinical trials practice, there have been a number of methodological contributions that integrate classical causal inference concepts (e.g., propensity score weighting or matching) with Bayesian methods (e.g, informative power priors), resulting in methods that exhibit desirable frequentist properties whilst still retaining a Bayesian interpretation (e.g., based on posterior probabilities). In this talk we describe the use of inverse probability weighted robust mixture priors and illustrate their application through the lens of an ongoing trial that utilizes a hybrid control arm comprised of internal and dynamically borrowed external controls. We discuss both statistical and pragmatic considerations for using such methods in practice.

Speaker: Anastasia Ivanovo, PhD (UNC)
Title: Dose-finding Studies: Bayes or Frequentist, Which is Better?
Abstract: When designing a dose-finding study in oncology a frequentist or a Bayesian method can be used. A well-known continual reassessment method comes in both a frequentist and a Bayesian version. We examine frequentist and Bayesian dose-finding designs to find which approach, frequentist or Bayesian, is better for dose-finding.