plot.kmciLR {bpcp}

Plot and lines methods for kmci, kmciLR, kmciLRtidy, and kmciLRgroup objects.

Description

Plots survival curves and/or confidence intervals. kmciLR objects from the bpcp function will plot using base R plot. If a kmciLRtidy object is created using bpcpfit (plotstyle = "ggplot"), the plot will display using ggplot. If a kmciLRgroup object is created using bpcpfit (plotstyle = "standard"), the plot will display using base R.

Usage

## S3 method for class 'kmci'
plot(x, ...)

## S3 method for class 'kmciLR'
plot(x, XLAB = "time", YLAB = "Survival", YLIM = c(0, 1), 
    ciLTY = 2, ciCOL = gray(0.8), mark.time = NULL, 
    linetype = "both", ...)

## S3 method for class 'kmciLR'
lines(x, lty = c(2, 1), col = c(gray(0.8), gray(0)), 
    linetype = "both", mark.time = NULL, ...)

## S3 method for class 'kmci'
lines(x, ...)

## S3 method for class 'kmciLRtidy'
plot(x, ...)
                          
## S3 method for class 'kmciLRgroup'
plot(x,XLAB="Time",YLAB="Survival",
 YLIM=c(0,1),ciLTY=2,ciCOL=gray(.8), linetype="both", ...)

Arguments

x

kmci, kmciLR, kmciLRtidy, or kmciLRgroup object (created by functions described in kmtestALL)

XLAB

label for x axis

YLAB

label for y axis

YLIM

limits for y axis

ciLTY

lty (line type) for confidence intervals

ciCOL

col (color) for confidence intervals

col

vector of colors, first element used for ci second for survival curve

lty

vector of line types, first element used for ci second for survival curve

mark.time

put hash marks for censored objects (default puts marks of stype="km" but not if stype="mue")

linetype

character, which lines to draw: either 'both', 'surv' or 'ci'

...

Extra parameters to be passed. Any argument in plot.kmciLR can be passed from plot.kmci, similarly for line. Other parameters are usually graphical parameters passed to plot and segment calls within function.

Examples

data(leuk)
## kmciTG creates kmci object
fitTG<-kmciTG(leuk$time,leuk$status)
plot(fitTG)
## bpcp creates kmciLR object 
fitBP<-bpcp(leuk$time,leuk$status)
lines(fitBP,lwd=3,lty=1,col=gray(.5),linetype="ci")
legend(0,.2,legend=c("Kaplan-Meier","Thomas-Grunkemeier 95 pct C
I","Beta Product 95 pct CI"),
    lwd=c(1,1,3),lty=c(1,2,1),col=c(gray(0),gray(.8),gray(.5)))
data(leuk2)
## bpcpfit creates kmciLR object which is plotted with ggplot
fitBPsurv <- bpcpfit(Surv(time, status)~treatment, data=leuk2)
plot(fitBPsurv)
## this works even if no treatment variable is specified
plot(bpcpfit(Surv(time, status)~1, data=leuk2))

## if plotstyle "standard" is specified, a base R plot is produced
a <- bpcpfit(leuk2$time, leuk2$status, leuk2$treatment, plotstyle = "standard")
plot(a)

#can also create a tidy object to customize ggplot further
tidy <- tidykmciLR(fitBPsurv)
ggplot(tidy, aes(x = time, y = surv, ymin = lower, ymax = upper, col = group)) + 
  geom_line(show.legend=FALSE) + geom_ribbon(alpha = .2, aes(fill=group)) + 
  xlab("Time") + ylab("Survival") + 
  ggtitle("K-M curves with bpcp CIs")