[R] nomogram function (design library)

Gondrie, M. M.Gondrie at umcutrecht.nl
Mon Sep 29 16:05:29 CEST 2008

Dear colleagues,

I hope someone can help me with my problem.

I have fitted a cox model with the following syntax:
# cox01def <-cph(Surv(TEVENT,EVENT) ~  ifelse(AGE>50, (AGE-50)^2,0) +
as.factor(ALBUMIN)+STENOSIS+IMT,data # = XC,  x=T, y=T, surv=T) *1

Furthermore  I have estimated my beta's also with a Lasso method -
Coxpath ( from Glmpath pckage) , namely: 

# LASSOdata <- list(x=cox01$x, time=XC$TEVENT, status=XC$EVENT)
# summary(LASSOdata)
# LASSOpath <- coxpath(data=SMARTdata)

My problem is the following:

I want to make a nomogram with the beta's from the lasso. Is that
possible? I know of the nomogram-function from the "design"-library. I
thought that when I would replace the coefficients of the cph-fit*1 by
the lasso-coefs, the nomogram would change; However it doesn't

# Lasso.coefs <- LASSOpath$b.predictor
# Cox01def$coef <- Lasso.coefs

If it isn't possible to solve this problem with the nomogram-function
from the design library, Then I would like to know if there is an
alternative function available? 
I hope to get a reply soon, 

Martijn Gondrie ( Utrecht, The Netherlands)  

More information about the R-help mailing list