The Potential Impact of Physiological Computer Models in Medicine: Regulation and Considerations for Ensuring Patient Safety
Richard A. Gray, PhD, Senior Research Biomedical Engineer, Food and Drug Administration, Silver Spring, MD
Computational modeling has the potential to accelerate the ongoing revolution in healthcare underway attributable to personalized medicine and information technology. Some potentially clinically useful computer models are based on the underlying mechanisms of human physiology which have unique benefits, but also present unique challenges. The benefits are numerous and span the entire medical product life-cycle from design to clinical outcomes as well as regulatory evaluation. The challenges are also numerous and include establishing trust, obtaining useful data, incorporating patient variability (if necessary) as well as ensuring models are validated and their use doesn’t endanger patients. This talk will begin with a brief overview of ongoing modeling efforts at the FDA, and then focus on a discussion of the factors and methodologies regarding the appropriate evaluation metrics for physiological computer models to be used clinically. Examples of verification, validation, and uncertainty quantification (VVUQ) for complex cardiac electrophysiological models will be presented as well as discussing methods for assessing credibility for specific contexts of use.