BOINTable: BOIN Decision Table for Dose-Finding Trials

In lrstat: Power and Sample Size Calculation for Non-Proportional Hazards and Beyond

BOIN Decision Table for Dose-Finding Trials

Description

Generates the decision table for the Bayesian Optimal Interval (BOIN) design, a widely used approach for dose-escalation trials that guides dose-finding decisions based on observed toxicity rates.

Usage

BOINTable(
  nMax = NA_integer_,
  pT = 0.3,
  phi1 = 0.6 * pT,
  phi2 = 1.4 * pT,
  a = 1,
  b = 1,
  pExcessTox = 0.95
)

Arguments

nMax	The maximum number of subjects allowed in a dose cohort.
pT	The target toxicity probability. Defaults to 0.3.
phi1	The lower equivalence limit for the target toxicity probability.
phi2	The upper equivalence limit for the target toxicity probability.
a	The prior toxicity shape parameter for the Beta prior.
b	The prior non-toxicity shape parameter for the Beta prior.
pExcessTox	The threshold for excessive toxicity. If the posterior probability that the true toxicity rate exceeds pT is greater than pExcessTox, the current and all higher doses will be excluded from further use to protect future participants. Defaults to 0.95.

Value

An S3 class BOINTable object with the following components:

• settings: The input settings data frame with the following variables:

  • nMax: The maximum number of subjects in a dose cohort.

  • pT: The target toxicity probability.

  • phi1: The lower equivalence limit for target toxicity probability.

  • phi2: The upper equivalence limit for target toxicity probability.

  • lambda1: The lower decision boundary for observed toxicity probability.

  • lambda2: The upper decision boundary for observed toxicity probability.

  • a: The prior toxicity parameter for the beta prior.

  • b: The prior non-toxicity parameter for the beta prior.

  • pExcessTox: The threshold for excessive toxicity.

• decisionDataFrame: A data frame listing dose-finding decisions for each combination of sample size (n) and number of observed toxicities (y):

  • n: Cohort size.

  • y: Number of observed toxicities.

  • decision: Recommended action: escalate, de-escalate, or stay at the current dose.

• decisionMatrix: A matrix version of the decision table showing the recommended action based on the number of toxicities for each possible cohort size.

Author(s)

Kaifeng Lu, [e-mail]

References

Liu, S., & Yuan, Y. (2015). Bayesian optimal interval designs for phase I clinical trials. Journal of the Royal Statistical Society: Series C (Applied Statistics), 64(3), 507-523.

Examples

BOINTable(nMax = 18, pT = 0.3, phi = 0.6*0.3, phi2 = 1.4*0.3)

lrstat documentation built on Aug. 8, 2025, 7:36 p.m.