|
Activity Number:
|
166
- Non-Clinical Statistics, Personalized Medicine, and Other Topics
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 10, 2021 : 10:00 AM to 11:50 AM
|
|
Sponsor:
|
Biopharmaceutical Section
|
|
Abstract #318797
|
|
|
Title:
|
Addressing Covariate Unbalance in Subgroup Identification for Censored Survival Data via Weighted Gradient Tree Boosting
|
|
Author(s):
|
JINCHUN ZHANG* and PINGYE ZHANG and Yue Shentu and JUNSHUI MA
|
|
Companies:
|
Merck & Co. and MERCK & Co. and Merck Sharp & Dohme and Merck & Co.
|
|
Keywords:
|
subgroup identification;
individual treatment rule;
survival outcome
|
|
Abstract:
|
Identifying subgroups of participants that have heterogeneous treatment effects serves as an important step towards precision medicine which has attracted great attention recently. However, covariate unbalance between assigned treatment arms often presents in the observational studies, and posit challenges in subgroup identification. To address this challenge, we propose a Weighted Value Guided Gradient Tree Boosting method to identify subgroups for censored survival data when covariate unbalances present. The original Gradient Tree Boosting algorithm is guided by a value function that focuses on maximizing the between-group treatment effect heterogeneity. And we extend this method by weight the restricted mean survival time using propensity score in each split that accounts for the observational nature of data. We evaluate the performance of proposed methods using simulations.
|
Authors who are presenting talks have a * after their name.
