M3 3127 and Zoom (email compmath@uwaterloo.ca for Zoom link)
Speaker
Michael Wallace | University of Waterloo
Title
Precision medicine with imprecise measurements: Measurement error and adaptive treatment strategies
Abstract
Adaptive treatment strategies are sequences of decision rules which take patient-level data as input and output treatment recommendations, and their is a large biostatistical literature devoted to their estimation. Much of this work assumes - usually implicitly - that observed data are measured without error, whereas in reality this is seldom the case. While measurement error has received considerable attention within the broader statistical literature, it is a topic that has received little attention in the context of precision medicine and adaptive treatment strategies. However, it presents some unique challenges within this context. These include non-adherence with personalized treatment plans, or making future decisions based on error-prone data.
In this talk, we will discuss the challenges at the interface of precision medicine and measurement error. We will establish why these problems arise through theoretical results, demonstrate their impact through interactive R Shiny applications, and summarize new methodology designed to account for measurement error within this framework.