A1104
Title: Rank estimation of monotone individualized treatment regimes for survival outcomes
Authors: Taehwa Choi - Sungshin Women\'s University (Korea, South) [presenting]
Seohyeon Park - Korea University (Korea, South)
Hyeonseok Oh - Korea University (Korea, South)
Zhezhen Jin - Columbia University (United States)
Sangbum Choi - Korea University (Korea, South)
Abstract: An optimal individualized treatment regimen (ITR) recommends tailored treatment decisions based on patients' genetics and demographic information, thereby providing greater clinical benefits compared to traditional clinical trials that randomly assign binary treatments to patients without considering their clinical status. While many studies have been conducted over several decades to identify the optimal ITR, the majority of them have covered simpler models that cannot be generalized to more complex models, such as the single-index model. The aim is to propose an inferential procedure for the surrogate maximum rank correlation (SMRC) to determine the optimal ITR under the single-index transformation model. Unlike the existing methods, the proposed approach provides greater flexibility, avoids strict model assumptions, and eliminates the need for complex computations while maintaining statistical accountability. Furthermore, an efficient variance estimation procedure is developed, using induced smoothing. Large sample properties are investigated, and various numerical examples demonstrate the usefulness of the method.
