Re: st: Survival analysis: finding best cut-off values

[Ron�n Conroy <rconroy@rcsi.ie>]

On 6 M�rta 2007, at 21:24, Diego Bellavia wrote:

> I am performing a survival analysis on a dataset with many  
> variables. Multivariate cox proportional-hazard models
> defined the best predictors (around 7 out of 270 variables). I  
> would like to give the readers some cut-off values
> they can use in the clinical practice,
My fear is that with a selection of 7 variables out of 270, your  
model shrinkage is likely to be a real problem. The associations you  
observe in your data may not be reproducible. Normally, model  
shrinkage is a function of the sample size and the number of  
predictors, but there's a Harrel paper in Statistical Methods in  
Medical Research (I'm sorry, I can't find it right now on the  
shelves) that argues that in variable selection models, the shrinkage  
is a function of the number of candidate variables, not the number in  
the eventual model.


=========
Ron�n Conroy
Royal College of Surgeons in Ireland
rconroy@rcsi.ie
+353 (0) 1 402 2431
+353 (0) 87 799 97 95
http://www.flickr.com/photos/ronanconroy




*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/