Estimating heterogeneous treatment effects with right-censored data via causal survival forests
This addresses the challenge of personalized treatment effect estimation in survival analysis for researchers and practitioners, though it is incremental as it builds on existing forest-based methods.
The paper tackles the problem of estimating heterogeneous treatment effects in survival and observational settings with right-censored data, introducing causal survival forests that perform well relative to baselines.
Forest-based methods have recently gained in popularity for non-parametric treatment effect estimation. Building on this line of work, we introduce causal survival forests, which can be used to estimate heterogeneous treatment effects in a survival and observational setting where outcomes may be right-censored. Our approach relies on orthogonal estimating equations to robustly adjust for both censoring and selection effects under unconfoundedness. In our experiments, we find our approach to perform well relative to a number of baselines.