OptimalSurrogate (version 1.0)

pte_cont: PTE estimation with a continuous surrogate marker

Description

PTE estimation with continuous surrogate marker

Usage

pte_cont(sob, yob, aob, var = TRUE, conf.int = TRUE, rep = 500)

Value

Estimates

Estimates of the treatment effect on the priamry outcome, delta, the treatment effect on the transformation of the surrogate, delta.gs, two versions of the proportion of treatment effect explained by the surrogate, pte1 and pte2; if var = TRUE, standard error estimates are also provided (se); if conf.int = TRUE, 95% confidence intervals are also provided

Transformed.S

the transformed surrogate, g(s), for each value of the surrogate, s; if var = TRUE, standard error estimates are also provided (se); if conf.int = TRUE, 95% confidence intervals are also provided

Arguments

sob

CONTINUOUS surrogate marker

yob

outcome of interest (continuous or binary)

aob

treatment assignment (1: treatment; 0: control)

var

whether variance should be calculated (TRUE/FALSE)

conf.int

whether 95% confidence intervals should be calculated (TRUE/FALSE)

rep

number of resampling replications (default is 500)

Examples

data(marker_cont)
out <- pte_cont(
  sob = marker_cont$sob, 
  yob = marker_cont$yob, 
  aob = marker_cont$aob, var = FALSE, conf.int = FALSE)
out

x <- as.numeric(names(out$Transformed.S))
plot(x, out$Transformed.S, ylim = range(out$Transformed.S), type = "l", 
     las = 1, xlab = "Surrogate Marker", ylab = "Optimal Transformation")

# \donttest{
out <- pte_cont(
  sob = marker_cont$sob, 
  yob = marker_cont$yob, 
  aob = marker_cont$aob, 
  var = TRUE, conf.int = TRUE, rep = 1000)
out$Estimates

x <- as.numeric(rownames(out$Transformed.S))
plot(x, out$Transformed.S[, "est"], ylim = range(out$Transformed.S[, -2]), type = "l", 
     las = 1, xlab = "Surrogate Marker", ylab = "Optimal Transformation")
  lines(x, out$Transformed.S[, "lower"], lty = 2)
  lines(x, out$Transformed.S[, "upper"], lty = 2)
# }