Simulate Continuous Patient Outcomes with Probability of Zero from a Beta Distribution

Simulates patient outcomes from a normal distribution, with the probability of a zero outcome being random and sampled from a Beta distribution. The probability of a zero outcome is determined as follows:

• For the control treatment, it is sampled from a \(Beta(UserParam$dCtrlBetaParam1, UserParam$dCtrlBetaParam2)\) distribution.

• For the experimental treatment, it is sampled from a \(Beta(UserParam$dExpBetaParam1, UserParam$dExpBetaParam2)\) distribution. This approach incorporates variability in the unknown probability of no response.

Usage

SimulateContinuousPatientOutcomePercentAtZeroBetaDist(
  NumSub,
  TreatmentID,
  Mean,
  StdDev,
  UserParam = NULL
)

Arguments

NumSub

The number of subjects to simulate. Must be an integer value.

TreatmentID

A vector of treatment IDs, where:

• 0 represents Treatment 1.

• 1 represents Treatment 2. The length of TreatmentID must equal NumSub.

Mean

A numeric vector of length 2 specifying the mean values for the two treatments.

StdDev

A numeric vector of length 2 specifying the standard deviations for each treatment.

UserParam

A list of user-defined parameters. Must contain the following named elements:

UserParam$dCtrlBetaParam1

Numeric; first parameter in the Beta distribution for the control (Treatment 1).

UserParam$dCtrlBetaParam2

Numeric; second parameter in the Beta distribution for the control (Treatment 1).

UserParam$dExpBetaParam1

Numeric; first parameter in the Beta distribution for the experimental (Treatment 2).

UserParam$dExpBetaParam2

Numeric; second parameter in the Beta distribution for the experimental (Treatment 2).

Value

A list containing the following elements:

Response

A numeric vector representing the simulated outcomes for each patient.

ErrorCode

Optional integer value:

0

No error.

> 0

Non-fatal error; current simulation is aborted but subsequent simulations continue.

< 0

Fatal error; no further simulations are attempted.