Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: How to estimate adjusted survival curves after fitting Cox model


From   Sanam P <sanamp25@yahoo.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: How to estimate adjusted survival curves after fitting Cox model
Date   Wed, 2 Jun 2010 20:09:31 -0700 (PDT)

Thank you Maarten for your response
 
I have been able to estimate the adjusted survivals at individual level using below codes using Professor Yi Li notes at Department of Biostatistics, Harvard SPH:
 
http://biowww.dfci.harvard.edu/~yili/lect5notes.pdf
 
 
**baseline survivals:
 xi:stcox i.var1  i.var2 i.var3  var4, basesurv(prsurv)  
 
predict betaz, xb
 
gen newterm=exp(betaz)
 
gen predsurv=prsurv^newterm
 
However I dont know what should do next to get the adjusted survivals for combinations of the covariates and not at individual level.
 
Regards,
Sadaf
 
 


----- Original Message ----
From: Maarten buis <maartenbuis@yahoo.co.uk>
To: stata list <statalist@hsphsun2.harvard.edu>
Sent: Wed, June 2, 2010 5:47:03 PM
Subject: Fw: st: How to estimate adjusted survival curves after fitting Cox model

--- On Wed, 2/6/10, Sanam P wrote:
> I was wondering what is the best way for
> calculating adjusted survival curves after fitting a cox
> regression model in stata. 
> 
> I think in the Kaplan miere method using "sts graph" and
> "adjust for " the calculated adjusted curves are for when
> all the coefficients are equal to zero which is not be the
> best method.

I am guessing that you are looking for something similar to
an average marginal effect: i.e. a survival curve that averages
over the distribution of the explanatory variables. However,
what is the the distribution of the explanatory variable in a
survival analysis? The distribution at t=0, or at each individual
time point. In some sense the latter seems more attractive, but
notice that changes in the survival curve then also represent
changes in the distribution of the explanatory variables in the
at risk population. However, the whole point why we add controll 
variables is that we want to keep them constant...

The trouble is that you are dealing with non-linear models, so 
looking for a single "best way" is usually not fruitful. You are 
much better of by considering the different summary statistics 
possible and find out what it is they say and understand and report
why they give different results.

Hope this helps,
Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://www.maartenbuis.nl
--------------------------


      

*
*  For searches and help try:
*  http://www.stata.com/help.cgi?searchhttp://www.stata.com/support/statalist/faqhttp://www.ats.ucla.edu/stat/stata/



      


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index