Time-Dependent Pseudo R-Squared for Assessing Predictive Performance in Competing Risks Data

Music City Center 

Evaluating and validating the performance of prediction models is a fundamental task in statistics, machine learning, and their diverse applications. However, developing robust performance metrics for competing risks time-to-event data poses unique challenges. We first highlight how certain conventional predictive performance metrics for competing risks time-to-event data, such as the C-index, Brier Score, and time-dependent AUC, can yield unexpected results when comparing predictive performance between different prediction models. To address this research gap, we introduce a novel time-dependent pseudo R-squared measure to evaluate the predictive performance of a predictive cumulative incidence function over a restricted time domain under right-censored competing risks time-to-event data. Specifically, we first propose a population-level time-dependent pseudo R-squared measures for the competing risk event of interest and then define their corresponding sample versions based on right-censored competing risks time-to-event data. We investigate the asymptotic properties of the proposed measure and demonstrate its advantages over conventional metrics through comprehensive simulation studies and three real-data applications.


Authors: Zian Zhaung, Wen Su, Eric Kawaguchi, and Gang Li*

*Presenter: Gang Li, Ph.D.
Professor of Biostatistics and Computational Medicine, University of California at Los Angeles