DRsurvCRT: Doubly-Robust Estimation for Survival Outcomes in
Cluster-Randomized Trials
Cluster-randomized trials (CRTs) assign treatment to groups rather than individuals, so valid analyses must distinguish cluster-level and individual-level effects and define estimands within a potential-outcomes framework. This package supports right-censored survival outcomes for both single-state (binary) and multi-state settings. For single-state outcomes, it provides estimands based on stage-specific survival contrasts (SPCE) and restricted mean survival time (RMST). For multi-state outcomes, it provides SPCE as well as a generalized win-based restricted mean time-in-favor estimand (RMT-IF). The package implements doubly robust estimators that accommodate covariate-dependent censoring and remain consistent if either the outcome model or the censoring model is correctly specified. Users can choose marginal Cox or gamma-frailty Cox working models for nuisance estimation, and inference is supported via leave-one-cluster-out jackknife variance and confidence interval estimation. Methods are described in Fang et al. (2025) "Estimands and doubly robust estimation for cluster-randomized trials with survival outcomes" <doi:10.48550/arXiv.2510.08438>.
Documentation:
Downloads:
Linking:
Please use the canonical form
https://CRAN.R-project.org/package=DRsurvCRT
to link to this page.