[R] [newbie] Cox Baseline Hazard

gb gb at stat.umu.se
Fri Feb 23 09:17:03 CET 2001

On Fri, 23 Feb 2001, Meles MELES wrote:

> I must have a vocabulary problem.
> I've tryed the method you gave me and I get a Cumulative Hazard for my
> Data.
> But, what I'm looking for is the Baseline Hazard (I don't know if the
> term is appropriate).
> If I consider a cox model for the intantaneous risk of death:
> h(t,Z)=h0(t)exp(b'Z), what I'm looking for is H0(t).
> What I get with the surfit() methot is H(t).
> I might be completly wrong, but when I compute my data in SPSS, I get a
> Cumulative Baseline Hazard of 0.062, in R, with the surfit way, I get
> 0.33, which correspond in the SPSS result to the Cumulative Hazard.
> I hope that what i'm writing is not too fuzzy.
> I hope my questions are not too boring :-)

Certainly not!
I think the problem here is that what Thomas and Peter call the
"baseline hazard" actually is calculated at the mean values of
the covariates. You can get what you want by using the "newdata"
parameter in  survfit. Try

  fit <- coxph( Surv(futime,fustat)~resid.ds+rx+ecog.ps,data=ovarian)
  base <- data.frame(age=0, resid.ds=0, rx = 0, ecog.ps = 0)
  ss <- survfit(fit)
  ss1 <- survfit(fit, newdata = base)
  par(mfrow = c(2, 1))
  plot(ss, fun = "cumhaz")
  plot(ss1, fun = "cumhaz")

and you see the difference.

 Göran Broström                      tel: +46 90 786 5223
 professor                           fax: +46 90 786 6614
 Department of Statistics            http://www.stat.umu.se/egna/gb/
 Umeå University
 SE-90187 Umeå, Sweden             e-mail: gb at stat.umu.se

r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch

More information about the R-help mailing list