Quantifying Uncertainties in Medical Decisions

Series: Program in Applied Mathematics Brown Bag Seminar
Location: Math 402
Presenter: Nicholas Henscheid, Program in Applied Mathematics, University of Arizona

Increasingly, medical decisions will be made using a combination of multiscale mathematical biology models and massive amounts of patient-specific data, including imaging data.  These models tend to rely on infinite dimensional parameters that are only partially or imprecisely known for a given patient, so predictive uncertainties exist and must be quantified in order to assess risks and make good decisions. While imaging data acts to reduce uncertainties, it is noisy, indirect, and incomplete, so uncertainty still exists. In this talk, I'll describe how we can quantify the impact of imaging using probability and information theory.