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 at the end of May, and its replacement, statalist.org is already up and running.


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

Re: st: R: compare median survival times? (flag: Stata 9.2/SE)


From   Martin X <martinx30@yahoo.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: R: compare median survival times? (flag: Stata 9.2/SE)
Date   Fri, 11 Jun 2010 07:48:22 -0700 (PDT)

Thanks for your input, Carlo and Ronan. I agree with Ronan that the code below will only compare survival functions with defined median survival times. The hypothesis itself is motivated by my desire to be able to say that half of group A survived 30 months and half of group B survived 20 months, and the 10 month difference between A and B is statistically significant. I've already used cox reg to compare the survival functions (and logistic reg to compare survival at a given time endpoint). Now, I just want to be able to provide some "common language" results (i.e. months of survival for half the group).    

Any other thoughts on this? I'm not familiar with interval quantile regression, but will look into it...

Thanks,
Marty



----- Original Message ----
From: Ronan Conroy <rconroy@rcsi.ie>
To: "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
Sent: Fri, June 11, 2010 7:23:28 AM
Subject: Re: st: R: compare median survival times? (flag: Stata 9.2/SE)

On 11 Meith 2010, at 08:44, Carlo Lazzaro wrote:

> Marty may want to customize what follows:
> --------- code begins-----------
> sysuse cancer.dta, clear
> stset studytime, failure(died==1)
> sts test drug if r(p50), logrank
> -------- code ends---------------

But doesn't this just compare the survival function in those in whom the median survival time is definable?

The hypothesis is a very peculiar one indeed: specifically, that the 50th percentile of the survival function differs between two groups. I don't know why anyone would formulate this hypothesis (why the 50th centile and not some other quantile? why ignore differences at other quantiles?) but it's a tricky one to try to test. It seems to me that you would need interval quantile regression, something whose complexities cause one to have to sit down and drink a glass of water.

Or am I missing something?


Ronan Conroy
=================================

rconroy@rcsi.ie
Royal College of Surgeons in Ireland
Epidemiology Department,
Beaux Lane House, Dublin 2, Ireland
+353 (0)1 402 2431
+353 (0)87 799 97 95
+353 (0)1 402 2764 (Fax - remember them?)
http://rcsi.academia.edu/RonanConroy

P    Before printing, think about the environment




*
*  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