pob2ef

pob2ef implements a partially optimized Bayesian single-arm phase II design for a binary efficacy endpoint with efficacy and futility stopping. The design searches candidate response-count boundaries and, subject to prespecified error constraints, selects the design that minimizes the grey-zone at the final look. The package creates decision tables and evaluates operating characteristics by exact trial-path enumeration or Monte Carlo simulation.

Installation

After CRAN publication, install the released package with:

install.packages("pob2ef")

To install the development version from GitHub:

remotes::install_github("clinrfnl/pob2ef")

Statistical framework

Let \(\theta\) denote the true response rate, with unacceptable benchmark \(p_0\) and clinically desirable target \(p_1 > p_0\). At an analysis after \(n\) evaluable participants, the design declares:

The boundary family is

\[ C_F(n) = \lambda_1(n/N)^\gamma, \qquad C_E(n) = 1 - \lambda_2(n/N)^\eta. \]

The tuning parameters satisfy

\[ 0 < \lambda_1 \leq 0.5, \qquad 0 < \lambda_2 < 0.5, \qquad \gamma, \eta \geq 0. \]

The design is calibrated to constrain superiority probability at \(p_0\) by \(\alpha\) and futility probability at \(p_1\) by \(\beta\). Among feasible boundary settings, it minimizes the final grey-zone probability at \(p_1\), then expected sample size at \(p_0\).

pob2ef is a distinct method. Related published methods include BOP2 and BOP2-FE:

Zhou, H., Lee, J. J., and Yuan, Y. (2017). BOP2: Bayesian optimal design for phase II clinical trials with simple and complex endpoints. Statistics in Medicine, 36(21), 3302-3314. https://doi.org/10.1002/sim.7338

Xu, X., Hashimoto, A., Yimer, B. B., and Takeda, K. (2025). BOP2-FE: Bayesian optimal phase II design with futility and efficacy-stopping boundaries. Journal of Biopharmaceutical Statistics. https://doi.org/10.1080/10543406.2025.2558142

Example

library(pob2ef)

design <- select_design(
  n_max = 20,
  looks = c(10, 20),
  p0 = 0.20,
  p1 = 0.40,
  alpha = 0.10,
  beta = 0.20,
  lambda1_grid = c(0.10, 0.30, 0.50),
  lambda2_grid = c(0.05, 0.10, 0.20),
  gamma_grid = 0:2,
  eta_grid = 0:2
)

design$selected
decision_table(design)
plot_decision_boundaries(design)

simulation <- simulate_design(
  design,
  true_rates = c(0.20, 0.30, 0.40),
  n_sim = 1000,
  seed = 2026
)

simulation$results
plot_operating_characteristics(
  simulation,
  metric = "superiority_probability"
)