Summarise calibration-grid output over a grid of BPP futility thresholds

Applies each candidate BPP futility threshold \(\lambda\) to the raw output of run_calibration_grid: a trial stops for futility if its interim BPP is below \(\lambda\), otherwise it takes its simulated continuation outcome.

Usage

summarize_grid_by_lambda(raw, lambda_grid, conf_level = 0.9)

Arguments

raw

The raw element returned by run_calibration_grid.

lambda_grid

Numeric vector of candidate BPP thresholds.

conf_level

Confidence level for the one-sided lower confidence bound on power (Clopper-Pearson; default 0.90).

Value

A data frame with one row per \(\lambda\) and columns lambda, P_early_fut, power_or_typeI, power_LCB, ESS (expected sample size) and duration (expected trial duration).

See also

run_calibration_grid, select_lambda_star