Skip to contents

DTEAssurance 1.2.0

Breaking changes

  • calc_dte_assurance_adaptive() no longer returns a Final_Decision column. That column was recomputed independently from a Cox Wald statistic at a flat threshold, ignoring analysis_model$method and the design’s boundaries, and could disagree with Decision. Use the new logical Success column, which is derived only from Decision. A Converged column (MCMC convergence at the BPP interim; NA for other futility types) is also returned.

New features

Bug fixes

  • survival_test(): "MW" now returns Z with the same sign convention as "LRT"/"WLRT" (positive Z favours treatment); "WLRT" uses the signed statistic from nph::logrank.test() directly.
  • survival_test() no longer fails with “could not find function Surv” when the survival package is not attached.
  • apply_GSD_to_trial() computes interim and final statistics with analysis_model$method via survival_test() instead of a hard-coded Cox Wald statistic, and passes update_priors_sims/n_BPP_sims through.
  • calc_dte_assurance_adaptive() and make_rpact_design_from_GSD_model() recognise the "MatchedZ" futility type.

DTEAssurance 1.1.1

DTEAssurance 1.0.1

CRAN release: 2025-10-24

DTEAssurance 1.0.0

CRAN release: 2025-10-14

  • Initial CRAN release.
  • Added support for delayed treatment effects using elicited prior distributions.
  • Introduced calc_dte_assurance() for fixed designs.
  • Introduced calc_dte_assurance_interim() for group sequential designs.
  • Added interactive Shiny applications for both design types.
  • Included vignettes and a pkgdown site for documentation.