Shuai Chen, PhD, Assistant Professor, UC Davis
It is widely recognized that treatments often have substantially different effects across a population. Many statistical methods have recently been developed for identifying subgroups of patients who may benefit from different available treatments. In many clinical and observational studies to evaluate the treatment benefits (eg, survival time) and harms (eg, toxicity and medical costs), censored data pose challenges to the analysis. Due to the induced dependent censoring problem, standard survival analysis techniques are often invalid for censored costs and toxicity duration. We propose a method for estimating individualized treatment benefits and harms with censored date in both randomized clinical trials and observational studies, which would provide a tool for physicians and patients to make decision based on personal benefit-harm tradeoff. Our method bypasses the modelling of main effect, and hence involves minimum modeling for the relationship between the outcome and covariates pertinent to measuring individual treatment benefit-harm tradeoff. The proposed method also allows variable selection via regularization. We then conducted numerical studies to evaluate the performance of proposed method.