AHTauDesign: Truncation Time Selection for Average Hazard Analyses
Implements a two-stage, design-informed framework for selecting the
truncation time in time-to-event trials analyzed with the average hazard
estimand, motivated by pediatric oncology settings with small samples,
slow accrual, and limited follow-up. Stage 1 validates a clinically
proposed truncation time against the planned design using simulation-based
diagnostics for risk-set support, follow-up coverage, estimability, and
estimator stability, classifying it as Pass, Borderline, or Fail. Stage 2
performs constrained optimization over a grid of candidate truncation
times within a clinical-distance window, maximizing a utility subject to
feasibility constraints, with an independent evaluation run to assess
the selected time. Supports proportional-hazards, early-, and delayed-effect
patterns, uniform accrual with administrative censoring, and
calibrated exponential random censoring.
Documentation:
Downloads:
Linking:
Please use the canonical form
https://CRAN.R-project.org/package=AHTauDesign
to link to this page.