NiceCumIncPlot produces a cumulative incidence curve plot in competing risk analysis

Cumulative incidence curve plot in competing risk analysis by split groups, including a table describing competing events and group comparison using the proportional subdistribution hazards regression model described in Fine and Gray (1999).

NiceCumIncPlot(TIMES, STATUS, SPLIT = NULL, CENSCODE = 0,
  TARGETCODE = 1, EVENTTABLE = TRUE, COMPARE = T, RATETABLE = T,
  TIMEPOINT = NULL, ORDER = T,
  MAIN = "Cumulative Incidence Function", XLAB = NULL,
  YLAB = "cumulative incidence", XLIM = NULL, YLIM = c(0, 1),
  SCALE = NULL, UNIT = "???", COL = NULL, EVENTNAME = "Event",
  COMPETENAME = "Competing", CEX = 0.8, ...)

Arguments

TIMES	time variable in competing risk setting
STATUS	status variable in competing risk setting default 0 = Censored 1= Events of interest, 2= Competing EVENT(s)
SPLIT	=NULL optional split variable
CENSCODE	=0, non-standard code for censored observations
TARGETCODE	=1, non-standard code for Events of interest
EVENTTABLE	= T, show event table
COMPARE	=T, show table with comparisons based on proportional subdistribution hazards regression model
RATETABLE	= T, show table with rate estimates at TIMEPOINT and their 95 percent CI
TIMEPOINT	=NULL, time point at which ordering of curves is read off if desired
ORDER	=T, order legends such that legends matched curves from top to bottom
MAIN	="Cumulative Incidence Function" plot title
XLAB	=NULL, x-axis label, default is "Time" with UNITS added in brackets if UNITS specified
YLAB	="cumulative incidence", y-axis label
XLIM	=NULL, range of x-axis - NOTE that for the comparison also Events beyond XLIM are used.
YLIM	=c(0,1), range of y-axis
SCALE	=NULL, unit for tick distance on time axis
UNIT	="???", specify time unit to be included in XLAB, e.g. "days", "Months", "years"
COL	=NULL, specify colours for curves to overwrite the default
EVENTNAME	="Event" colname for Event in Event table
COMPETENAME	="Competing" colname for competing events in Event table
CEX	= .8 letter size for legends
...	further plot parameter passed to graphics::plot

Value

produces a plot and invisibly returns the fitted proportional subdistribution hazard model

Details

NiceCumIncPlot

Examples

if (FALSE) {
 N<-200
 TimeToEvent<-rweibull(N,1)
 Events<-sample(0:2,N,replace=T,prob=c(.3,.5,.2))
 Arm<-sample(LETTERS[1:2],N,replace=T)
 xx<-NiceCumIncPlot(
  TIMES=TimeToEvent,
  STATUS=Events,
  SPLIT=Arm,
  CENSCODE=0,
  TARGETCODE=1,
  EVENTTABLE = F,
  COMPARE = T,
  RATETABLE = T,
  TIMEPOINT=2,
  ORDER=T,
  MAIN="Cumulative Incidence Function",
  XLAB=NULL,
  YLAB="cumulative incidence",
  XLIM=NULL,
  YLIM=c(0,1),
  SCALE=NULL,
  UNIT="years",
  COL=NULL,
  EVENTNAME ="Event",
  COMPETENAME ="Competing",
  CEX=.8)
}