[R] stepwise selection cox model

[David Winsemius]

On May 25, 2011, at 5:28 AM, linda Porz wrote:

> Sorry, I have wrote a wrong subject in the first email!
>
> Regards,
> Linda
>
> ---------- Forwarded message ----------
> From: linda Porz <linda.porz at gmail.com>
> Date: 2011/5/25
> Subject: combined odds ratio
> To: r-help at r-project.org
> Cc: r-help-request at stat.math.ethz.ch
>
>
> Dear all,
>
> I am looking for an R function which does stepwise selection cox  
> model in r
> (delta chisq likelihood ratio test) similar to the stepwise, pe  
> (0.05) lr:
> stcox in STATA.
>


Does the Stata method apply appropriate penalization to its stepwise  
procedures?

I suspect you will find that the experts in survival analysis around  
these parts take a very dim view of stepwise procedures and I would  
not be surprised if they purposely put a barrier in front of naive  
users to protect them from falling into the well-described but perhaps  
not widely understood pitfalls of such methods.

I do know that Harrell provides for some support for penalized methods  
in his cph related functions. See the function pentrace.  He also has  
a fastbw function in rms which is provided mainly so one can  
investigate and demonstrate those aforementioned pitfalls.

Note: You should not be adding the address r-help-request at stat.math.ethz.ch 
  to your postings. You may get cryptic replies in your Inbox. It is  
the address for interacting with the mail-server to manage your  
subscription options.


David Winsemius, MD
West Hartford, CT