Incorporating simulated virtual patient outcomes as prior information in medical device clinical trials

Wednesday, March 7, 2018 - 11:30am - 12:30pm
Lind 305
Adam Himes (Medtronic)
Medical device manufacturers are increasingly using predictive computer models, also called virtual patient models, that simulate clinical outcomes. In some cases, these virtual patient models can be incorporated into a study in a way that is analogous to how some Bayesian clinical trials incorporate historical data as prior information. Benefits of this approach may include increased information from the clinical study, more confidence in the clinical outcome, and in some cases smaller or shorter duration studies.

The Medical Device Innovation Consortium (MDIC) working group on Clinical Trials Informed by Bench and Simulation includes members from industry and the U.S. Food and Drug Administration (FDA). Due to the novelty of incorporating virtual patients into a clinical study, and the mutual interest of both FDA and industry in advancing the use of modeling, a mock device submission was conducted as a means to evaluate and document the process. The sponsor team consisted of members representing medical device manufacturers, engineering software, MDIC, and FDA. A review team was formed at FDA, consisting of members representing engineering, statistical and medical functions, as if the submission were an actual product.

This talk will introduce the unique aspects of virtual patient models, describe the statistical methods required to incorporate virtual patient outcomes into a clinical trial, and highlight practical considerations that were raised during the collaborative work with FDA.