[R] Nomogram with stratified cph in rms package, how to get failure probability
Frank Harrell
f.harrell at vanderbilt.edu
Thu Dec 1 17:14:58 CET 2011
Hazard() is not implemented except for parametric survival models.
You are not calling nomogram() correctly; in rms the plotting step is
separated from the nomogram computations.
To plot cumulative event rates do something like:
mort10 <- function(lp) 1 - surv(10,lp)
and tell nomogram about mort10 using fun=.
Frank
min wrote
>
> Hello,
>
> I am using Dr. Harrell's rms package to make a nomogram. I was able to
> make a beautiful one. However, I want to change 5-year survival
> probability to 5-year failure probability.
>
> I couldn’t get hazard rate from Hazard(f1) because I used cph for the
> model.
>
> Here is my code:
>
> library(rms)
> f1 <- cph(Surv(retime,dfs) ~
> age+her2+t_stage+n_stage+er+grade+cytcyt+Cyt_PCDK2 , data=data11,
> surv=T, x=T, y=T, time.inc=5)
>
> surv<- Survival(f1)
> haz<- Hazard(f1)
> Here is the Error in UseMethod("Hazard") :
> no applicable method for 'Hazard' applied to an object of class
> "c('cph', 'Design', 'coxph')"
>
>
> surv10 <- function(lp) surv(10,lp)
> surv5 <- function(lp) surv(5,lp)
> quant <- Quantile(f1)
>
> at.surv <- c(0.1, 0.3, 0.5, 0.7, 0.9)
>
> at.med1<-c(2,3,4, 5,6,7,8, 10,15,20,25, 30)
>
>
> par(cex=0.8)
> nom<- nomogram(f1, conf.int=F,
> fun=list(surv5, surv10), funlabel=c('5-Year Survival Probability',
> '10-Year Survival Probability' ), lp=F,
> fun.at=c(at.surv, at.surv),label.every=1, force.label=FALSE, cex.axis=0.8,
> verbose=TRUE, cex.var=0.8)
>
> How can I show failure probability instead of survival probability?
>
> I would very much appreciate any assistance in this matter.
> Thank you Very much.
>
-----
Frank Harrell
Department of Biostatistics, Vanderbilt University
--
View this message in context: http://r.789695.n4.nabble.com/Nomogram-with-stratified-cph-in-rms-package-how-to-get-failure-probability-tp4123249p4129173.html
Sent from the R help mailing list archive at Nabble.com.
More information about the R-help
mailing list