R help - Mar 2014 - How to use restricted cubic spline in survfit.cph function in "survival" package?

Hello,

I used PBC data set included in the "survival" package to fit a cox
model. The model included a restricted cubic spline transformation on
age. Then I tried using "survfit" function to predict a survival curve
using the first row of the original data. I got an error message. The
R code and error message is below.
> require(survival)
> fit<-coxph(Surv(time,status==2)~rcs(age,4)+sex, data=pbc)
> id1<-pbc[1,]
> surv.id1<-survfit(fit,newdata=id1)
Error in rcspline.eval(x, nk = nknots, inclx = TRUE, pc = pc) :
  fewer than 6 non-missing observations with knots omitted


Can anyone provide any suggestions?

Thanks,
Zhiyuan

Try cph in rms (where rcs is defined).

library(rms)
fit<-cph(Surv(time,status==2)~rcs(age,4)+sex, data=pbc, y=TRUE, x=TRUE)
id1<-pbc[1,]
surv.id1<-survfit(fit,newdata=id1)
plot(surv.id1)
summary(surv.id1)

-----Original Message-----
From: Zhiyuan Sun [mailto:sam.d.sun at gmail.com] 
Sent: Wednesday, March 05, 2014 10:52 PM
To: r-help at r-project.org
Cc: therneau.terry at mayo.edu
Subject: [R] How to use restricted cubic spline in survfit.cph function in
"survival" package?

Hello,

I used PBC data set included in the "survival" package to fit a cox
model. The model included a restricted cubic spline transformation on
age. Then I tried using "survfit" function to predict a survival curve
using the first row of the original data. I got an error message. The
R code and error message is below.
> require(survival)
> fit<-coxph(Surv(time,status==2)~rcs(age,4)+sex, data=pbc)
> id1<-pbc[1,]
> surv.id1<-survfit(fit,newdata=id1)
Error in rcspline.eval(x, nk = nknots, inclx = TRUE, pc = pc) :
  fewer than 6 non-missing observations with knots omitted


Can anyone provide any suggestions?

Thanks,
Zhiyuan


**********************************************************
Electronic Mail is not secure, may not be read every day, and should not be used
for urgent or sensitive issues