A1522
Title: A suite of Spearman-like correlation measures for bivariate interval-censored data
Authors: Eric Kawaguchi - University of Southern California (United States) [presenting]
Abstract: Nonparametric tests of independence and their associated measures of association are important tools for studying bivariate failure time data. While pairwise concordance measures, such as Kendall's tau, have been studied extensively for bivariate interval-censored outcomes, comparatively less attention has been given to rank-based correlation measures. An extension of Spearman's rho for bivariate interval-censored data is proposed. The estimator relies only on marginal distributions and avoids direct estimation of the bivariate failure time surface. The method is further extended to allow for covariate adjustment, enhancing its practical relevance. Through simulation studies, it is demonstrated that the proposed estimator performs well across a range of scenarios. Finally, the utility of the approach is illustrated through applications to several real-world datasets.
