estSeqSim: estSeqSim

Description

Simulate Comparisons For Use in Sequential Clinical Trial Simulations

Usage

estSeqSim(parameter, looks, gendat, fitter, nsim = 1, progress = FALSE)

Arguments

parameter

vector of true parameter (effects; group differences) values

looks

integer vector of observation numbers at which posterior probabilities are computed

gendat

a function of three arguments: true parameter value (scalar), sample size for first group, sample size for second group

fitter

a function of two arguments: 0/1 group indicator vector and the dependent variable vector

nsim

number of simulations (default is 1)

progress

set to TRUE to send current iteration number to the console

Value

a data frame with number of rows equal to the product of nsim, the length of looks, and the length of parameter.

Details

Simulates sequential clinical trials. Looks are done sequentially at observation numbers given in the vector looks with the earliest possible look being at observation 2. For each true effect parameter value, simulation, and at each look, runs a function to compute the estimate of the parameter of interest along with its variance. For each simulation, data are first simulated for the last look, and these data are sequentially revealed for earlier looks. The user provides a function gendat that given a true effect of parameter and the two sample sizes (for treatment groups 1 and 2) returns a list with vectors y1 and y2 containing simulated data. The user also provides a function fitter with arguments x (group indicator 0/1) and y (response variable) that returns a 2-vector containing the effect estimate and its variance. parameter is usually on the scale of a regression coefficient, e.g., a log odds ratio.

Examples

Run this code

# NOT RUN {
# Run 100 simulations, 5 looks, 2 true parameter values
# Total simulation time: 2s
lfit <- function(x, y) {
f <- rms::lrm.fit(x, y)
  k <- length(coef(f))
  c(coef(f)[k], vcov(f)[k, k])
}
gdat <- function(beta, n1, n2) {
  # Cell probabilities for a 7-category ordinal outcome for the control group
  p <- c(2, 1, 2, 7, 8, 38, 42) / 100

  # Compute cell probabilities for the treated group
  p2 <- pomodm(p=p, odds.ratio=exp(beta))
  y1 <- sample(1 : 7, n1, p,  replace=TRUE)
  y2 <- sample(1 : 7, n2, p2, replace=TRUE)
  list(y1=y1, y2=y2)
}

set.seed(1)
est <- estSeqSim(c(0, log(0.7)), looks=c(50, 75, 95, 100, 200),
                  gendat=gdat,
                  fitter=lfit, nsim=100)
head(est)
# }