Title: Double-robust inference for differences in restricted mean lifetimes using pseudo-observations
Authors: Sangbum Choi - Korea University (Korea, South) [presenting]
Abstract: When comparing survival times between two treatment groups, restricted mean lifetime, defined as the expectation of the survival function restricted to a prespecified time point, is often of direct interest, as it is easily understood by clinician investigators and does not require restrictive assumptions, such as proportionality. If the treatments are not randomized as in observational studies, covariate adjustment is needed to account for treatment imbalances in confounding factors. We propose a simple pseudo-value approach to estimate the difference of the restricted mean lifetime between two groups while accounting for confounders, which can be used as a metric for average causal effect (ACE). The proposed method combines two general approaches, (1) group-specific regression models (Q-model) for the time-to-event and covariate information, and (2) inverse probability of treatment assignment weights (A-model), where the restricted mean lifetimes are replaced by the corresponding pseudo-observations for survival outcomes. We show that the proposed estimator is double robust in the sense that it is consistent if at least one of the two working models remains correct. Simulation studies are conducted to assess its finite-sample performance and the method is applied to kidney transplant data.