Adaptive and Bayesian Clinical Trial Design

Ben Saville, PhD, is President and Lead Statistical Scientist of Adaptix Trials, LLC, where he specializes in the design of innovative Bayesian and adaptive clinical trials.  He works primarily with medical device and pharmaceutical companies to solve challenging problems via Bayesian designs. Dr. Saville has designed Bayesian clinical trials in a variety of therapeutic areas, including cardiovascular diseases, rare and progressive diseases, infectious diseases, stroke and traumatic brain injury, respiratory and digestive disorders, and neurological diseases. He has extensive regulatory experience representing clients in FDA meetings, submissions, and advisory panels.  With a strong background in teaching, Dr. Saville has a unique ability to communicate complex statistical methods to clinical collaborators.  Dr. Saville has authored over 100 peer-reviewed publications in the statistical and medical literature, with expertise in Bayesian hierarchical models and Bayesian adaptive platform trial designs.

Clinical trial designs rely on assumptions of key trial parameters, e.g. endpoint variability and potential treatment effect, to make important decisions about sample size and other trial features.  Unfortunately, we often learn at the end of the trial that our assumptions were inaccurate, leading to a failed or inconclusive trial result.  To address the uncertainty in these assumptions, an adaptive trial design uses a pre-specified algorithm to learn about key trial design parameters from the accruing trial data.  This allows adaptations based on the observed data such as adaptive sample size, population enrichment via subgroups, and arm dropping.  Bayesian methods enable natural and flexible decision criteria for these adaptations, as well as the ability to leverage external data via Bayesian borrowing.  We use virtual clinical trial simulations to customize an adaptive clinical trial design for your specific objectives.   This allows us to fully evaluate our adaptive algorithm and provide compelling justification to FDA regulators.