A0944
Title: Learning individualized minimal clinically important difference (iMCID) from high-dimensional data
Authors: Jiwei Zhao - University of Wisconsin-Madison (United States) [presenting]
Abstract: Statistical significance has been widely used to infer the treatment effect in assessing the efficacy of a treatment or intervention; however, there has been a growing recognition that statistical significance has limitations. On the contrary, clinical significance is usually desirable in practice as it provides a better assessment of clinically meaningful improvement. A critical concept in evaluating clinical significance is a minimal clinically important difference (MCID), the smallest change in the outcome that an individual patient would identify as important. A statistical learning framework for estimating the individualized MCID (iMCID) from high-dimensional data will be presented. In particular, a path-following iterative algorithm and some novel nonregular theoretical results will be presented. Additionally, simulation studies that reinforce the theoretical findings and an application to the study of chondral lesions in knee surgery to demonstrate the usefulness of the proposed approach will also be discussed.