cindex: Concordance index calculation

In jixccf/QHScrnomo: Construct nomograms for competing risk models

Description

Calculate concordance index

Usage

cindex(
  prob,
  fstatus,
  ftime,
  type = "crr",
  failcode = 1,
  cencode = 0,
  tol = 1e-20
)

Arguments

prob	predicted risk of failure event, either probability or risk score
fstatus	failure(event) variable
ftime	follow-up time variable for survival or competing risks predictions
type	type of regression models corresponding to different type of outcomes. 'logis' is the default value for binary outcome, 'surv' for ordinary survival outcome and 'crr' for competing risks outcome.
failcode	coding for failure(event). 1 is the default value.
cencode	coding for censoring. 0 is the defaul
tol	error tolerance. the default value is 1e-20.

Details

to calculate the discrimination metric, concordance index for binary, time-to event and competing risks outcomes

Value

a vector of returned values.

N	the total number of observations in the input data
n	the nonmissing number of observations that was used f or calculation
usable	the total number of usable pairs.
oncordant	the number of concordant pairs
cindex	the concordance index that equal to the number of concordant pairs divided by the total number of usable pairs.

Author(s)

Changhong Yu, Michael Kattan, Brian Wells, Amy Nowacki.

Examples

seed <- 5555

# simulated data to test the function
n <- 20000
ftime <- rexp(n)
fstatus <- sample(0:2,n,replace=TRUE)
cov <- matrix(runif(3*n),nrow=n)
dimnames(cov)[[2]] <- c('x1','x2','x3')
dat <- data.frame(ftime,fstatus,cov)
require(Design)
print(z <- cph(Surv(ftime,fstatus == 1) ~ x1 + x2 + x3,data = dat))

# summary(z)
z.p <- predict(z)

#debug(cindex)
cindex(prob=z.p,fstatus=fstatus,type="logis")
cindex(prob=z.p,fstatus=fstatus,ftime=ftime,type="surv")
cindex(prob=z.p,fstatus=fstatus,ftime=ftime,type="crr")

jixccf/QHScrnomo documentation built on Dec. 21, 2021, 12:08 a.m.