Statalist The Stata Listserver

[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: c-statistic in Cox models

From   "Nick Cox" <>
To   <>
Subject   RE: st: c-statistic in Cox models
Date   Tue, 10 Oct 2006 16:07:30 +0100

The FAQ advises 

"The current version of Stata is 9. Please specify if you are 
using an earlier version; otherwise, the answer to your question 
is likely to refer to commands or features unavailable to you. 
Moreover, as bug fixes and new features are issued frequently by 
StataCorp, make sure that you update your Stata before posting a 
query, as your problem may already have been solved. It may help 
a lot if you give the results of typing -about- within your Stata, 
specifically the first two lines of the form 

Stata/SE 9.2 for Windows 
Born 06 Oct 2006" 

As you did not specify that you are using Stata 8, you 
got advice to use -estat- that does not apply to you, 
as -estat- was added in version 9. 

Roger Newson has documented the application of -somersd- about
as thoroughly as any user has documented their work. 

This leaves your attempt to use -stcstat-. Know that
"doesn't seem to be working" is not informative and cannot 
be answered. Again, advice to this effect is prominent 
in the FAQ. 

Finally, to spell out once more some advice given above: 
while we can do nothing about the institutional reasons 
why you are stuck with Stata 8.0, you, or someone 
responsible, should be able to -update- to Stata 8.2, 
which while it may not solve this problem is on other
grounds a good idea. 


Ashwin Ananthakrishnan
> Thanks for the suggestion.
> I installed the somersd package, but am not sure how
> to interpret the results or what represents the
> c-statistic. 
> I input the dependent variable and the independent
> variables, but it doesn't ask me for a specific time
> I'm looking at (for example failure at 12 months or at
> 24 months). It outputs coefficients for each of the
> variables I input, but doesn't give me a c-statistic
> or overall fit. 
> Am I not using it correctly?
> I'm also not able to use the 'estat concordance'
> command suggested by another user because it doesn't
> seem to run on Stata 8.0 which is the version I have
> at my workplace (which I cannot upgrade to 9.0 due to
> some policy reasons in the workroom).
Newson, Roger B

> > The -somersd- package, downloadable from SSC using
> > the -ssc- command,,
> > also calculates general Harrell c-indices, and also
> > Somers' D and
> > Kendall's tau-a. There is an example in the manual
> > -somersd.pdf-, which
> > is distributed with the package as an ancillary
> > file.

Jacki Buros
> > Try the stcox postestimation command -estat
> > concordance-, which
> > calculates Harrell's C. This may be what you want
> > for the survival
> > analysis, so long as you do not have time-varying
> > covariate effects.
> > 
> > I don't know whether there is a similar command for
> > svy regress.

Ashwin Ananthakrishnan

> > > Is there a command in stata where I can calculate
> > the
> > > c-statistic for predictive models in Cox
> > regression
> > > survival analysis?
> > >
> > > I tried the stcstat program that I found online
> > but
> > > that doesn't seem to be working. Are there other
> > > commands?
> > >
> > > Is there a similar c-statistic command for use in
> > > survey analysis also? (following a .svylogit or a
> > > .svyregress command) ?

*   For searches and help try:

© Copyright 1996–2021 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index