R/confintCalc.R

In nickreich/clusterPower: Power Calculations for Cluster-Randomized and Cluster-Randomized Crossover Trials

#' Power simulations for cluster-randomized trials: get exact confidence intervals.
#'
#' This function uses binom.test() to get exact
#' confidence intervals for the power in Monte Carlo power estimation.
#' It is called by some of the clusterPower functions.
#'
#' @author Alexandria C. Sakrejda (\email{acbro0@@umass.edu} and Ken Kleinman (\email{ken.kleinman@@gmail.com})
#'
#' @param alpha A numeric; the user-selected alpha cutoff.
#' @param p.val A vector of p-values.
#'
#' @return A dataframe
#' \describe{
#'   \item{power.parms}{Exact confidence intervals produced using \code{binom.test()}}
#' }
#'
#' @noRd
confintCalc <- function(alpha = alpha,
                        p.val = p.val,
                        nsim = nsim,
                        multi = FALSE) {
  sig.val <-  ifelse(p.val < alpha, 1, 0)
  if (isTRUE(is.data.frame(sig.val)) ||
      isTRUE(is.matrix(sig.val))) {
    pval.power <- apply(sig.val, 2, FUN = sum)
  }
  if (isTRUE(is.vector(sig.val))) {
    pval.power <- sum(sig.val)
  }
  power.parms <- list()
  
  for (q in 1:length(unlist(pval.power))) {
    power.parms[[q]] <-
      binom.test(
        p = alpha,
        n = length(unlist(p.val)),
        x = pval.power[q],
        alternative = "two.sided"
      )
  }
  Power <- vector(length = length(pval.power))
  Lower.95.CI <- vector(length = length(pval.power))
  Upper.95.CI <- vector(length = length(pval.power))
  Alpha <- vector(length = length(pval.power))
  Beta <- vector(length = length(pval.power))
  Names <- vector(length = length(pval.power))
  
  for (o in 1:length(pval.power)) {
    Power[o] = power.parms[[o]]$estimate
    Lower.95.CI[o] = power.parms[[o]]$conf.int[1]
    Upper.95.CI[o] = power.parms[[o]]$conf.int[2]
    Alpha[o] = alpha
    Names[o] = paste0("Arm.", o + 1)
  }
  Beta <- 1 - Power
  power.parms <-
    data.frame(Power, Lower.95.CI, Upper.95.CI, Alpha, Beta)
  rownames(power.parms) <- Names
  if (multi == FALSE){
  num.returned <- data.frame("Converged" = length(p.val), 
                             "Requested" = nsim)
  power.parms <- cbind(power.parms, num.returned)
  }
  return(power.parms)
}