Colloquium: Information Under the Curve: A Modern Approach to Leveraging Area Under Curves in Survival Analysis

Statistical Sciences Colloquium
Dr. Mitchell Paukner, Assistant Professor of Biostatistics Wake Forest University

Information Under the Curve: A Modern Approach to Leveraging Area Under Curves in Survival Analysis
24 Sept. — ZSR Auditorium (Room 404)

In this talk, I will discuss several different methods I have developed for the analysis of time-to-event data. With the often varied and heavily nuanced characteristics of different time-to- event settings, methods that rely on rigid assumptions lack utility in many areas of practice. The aim of my research has been, and is , to establish new methodologies for analyzing survival data that are not only flexible in application, but also maintain a clinically meaningful interpretation that i s accessible to more than just statisticians. In this talk, I will discuss the various methods I have developed to approach modern challenges in survival analysis by focusing on the statistical information that is“ under the curve”. I will introduce window mean survival time (WMST), an alternative to restricted mean survival time (RMST), as a method for evaluating survival data in the standard time- to- event setting. WMST, an estimate of mean survival in a specified window of time (characterized by the area under the Kaplan- Meier (KM) curve), offers the clinician and statistician great flexibility in choosing
window bounds without relying on distributional assumptions that nullify results or interpretation when not satisfied. I will discuss the basic methodology of WMST and its various extensions (formulation of estimates and tests, multiple testing, and trial design), and demonstrate its value through multiple simulation studies and real data examples.