Compute Bayesian Predictive Probability of Success

Function to compute the Bayesian predictive probability of success for a clinical trial using Bayesian analysis. The function simulates future patient outcomes based on posterior distributions derived from observed interim data and evaluates the probability of trial success at the end.

Usage

ComputeBayesianPredictiveProbabilityWithBayesianAnalysis(
  dataS,
  dataE,
  priorAlphaS,
  priorBetaS,
  priorAlphaE,
  priorBetaE,
  nQtyOfPatsS,
  nQtyOfPatsE,
  nSimulations,
  finalBoundary,
  lAnalysisParams
)

Arguments

dataS

A vector of binary outcomes (0 or 1) for the control treatment observed at the interim analysis.

dataE

A vector of binary outcomes (0 or 1) for the experimental treatment observed at the interim analysis.

priorAlphaS

The alpha parameter of the Beta prior for the control treatment.

priorBetaS

The beta parameter of the Beta prior for the control treatment.

priorAlphaE

The alpha parameter of the Beta prior for the experimental treatment.

priorBetaE

The beta parameter of the Beta prior for the experimental treatment.

nQtyOfPatsS

The total number of patients for the control treatment expected by the end of the trial.

nQtyOfPatsE

The total number of patients for the experimental treatment expected by the end of the trial.

nSimulations

The number of virtual trials to simulate for predictive probability computation.

finalBoundary

The cutoff threshold for posterior probability to determine trial success.

lAnalysisParams

A list of analysis parameters for posterior computation, including priors for the control and experimental treatments.

Value

A list containing:

predictiveProbabilityS

The Bayesian predictive probability of trial success.

Details

This function computes the Bayesian predictive probability of success for a clinical trial. It uses observed interim data to update the Beta priors into posterior distributions for success probabilities of both control and experimental treatments. Future patient outcomes are simulated based on these posteriors, and the trial success is evaluated based on the probability that the experimental treatment has a higher success rate than the control treatment. The predictive probability is calculated as the proportion of simulated trials meeting the success criteria.