https://github.com/cran/clinfun
Tip revision: 9ef289b3d571cf9efd9d6e67a67b72255e6ae181 authored by Venkatraman E. Seshan on 04 November 2010, 00:00:00 UTC
version 0.9.1
version 0.9.1
Tip revision: 9ef289b
coxphQuantile.Rd
\name{coxphQuantile}
\title{Survival time quantile as a function of covariate}
\alias{coxphQuantile}
\keyword{survival}
\description{
Draws a quantile curve of survival distribution as a function of covariate.
}
\usage{
coxphQuantile(phfit, xrange, p=0.5, whichx=1, otherx=NULL, ...)
}
\arguments{
\item{phfit}{output from a proportional hazards fit.}
\item{xrange}{the range of covariate values for which the quantiles of
survival times are computed.}
\item{p}{the probability level for the quantile (default is median).}
\item{whichx}{if there are more than one covariates in the Cox model,
the one chosen for the quantile plot.}
\item{otherx}{the values for other covariates in the Cox model. If
missing uses their average values.}
\item{...}{additional parameters to be passed on to the lines command.}
}
\details{
This function is used to draw quantile curves. It requires a plot of
the data (time & covariate of interest) to be present. See example.
It invisibly returns the observed failure times and the covariate
values at which the estimated survival probability is (exactly) p.
}
\examples{
\dontrun{ library(survival)
data(pbc)
pbcfit <- coxph(Surv(time, status==2) ~ trt + log(copper), pbc,
subset=(trt>0 & copper>0))
plot(log(pbc$copper[pbc$trt>0 & pbc$copper>0]), pbc$time[pbc$trt>0 &
pbc$copper>0], pch=c("o","x")[1+pbc$status[pbc$trt>0 & pbc$copper>0]],
xlab="log Copper", ylab="Survival time")
coxphQuantile(pbcfit, c(2.5,6), whichx=2, otherx=1)
coxphQuantile(pbcfit, c(2.5,6), p=0.75, whichx=2, otherx=2, col=2)}
}
\references{
Heller G. and Simonoff J.S. (1992) Prediction in censored survival data:
A comparison of the proportional hazards and linear regression models.
\emph{Biometrics} 48, 101-115.
}