R/alpha.R

In esm-ispm-unibe-ch/sequentialnma2: Sequential network meta-analysis

###########################################################################################################
###### A function that calculates the adjusted nominal level and the associated sequential bounaries ######
###########################################################################################################
#method: the method to construct the boundaries, it can take the values "BF" (O'Brien Flemming), 
#"POC" (Pocock), "LIN" (Linear), "PFUN" (power function) 
#see Jihao Zhou, Glen Andrews. Alpha spending function. In: Encyclopedia of biopharmaceutical statistics. 3rd ed. Taylor &
#Francis, pp. 38–44. for an overview of the available methods to be approximated in constructed alpha spending functions
#t: the fraction of the accumulated information

alpha <- function(method, t, typeIerror=typeIerror) {
  za <- qnorm(1 - typeIerror/2)
  #calculation of adjusted nominal level
  if (method == "BF") at <- 2 * (1 - pnorm(za/sqrt(t)))
  if (method == "POC") at <- a * log(1 + (exp(1) - 1) * t)
  if (method == "LIN") at <- a * t
  if (method == "PFUN") at <- a * (t^2)
  #calculation of associated boundary
  E = qnorm(1 - at/2)
  #store fraction of accumulated information, nominal level and boundary
  output <- data.frame(t = t, at = at, E = E)
  #replace boundaries with abs(1.96) if t>1 
  output$E[output$t>1 & output$E>0]<-za
  output$E[output$t>1 & output$E<0]<--za
  return(output)
  result <- array(0, 1, length(output))
  for (j in 1:length(output)) {
    result[j] <- output[j]
  }
  return(result) 
}