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: p-values for K-M point survival estimates using sts list


From   Steve Samuels <sjsamuels@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: p-values for K-M point survival estimates using sts list
Date   Wed, 25 Aug 2010 20:09:01 -0400

--

Cynthia, if S(t) is the KM estimate, the CIs are based on
log(-log(S(t)), which removes range restrictions. The SE itself is a
function of S(t) and so does not make a good denominator.

Let S = S(t)

vl = var(log S)= (1/S^2)x var(S)

 w = log(-log(S))
var(w) = vl/(logS)^2

Base the test of difference on w.

Steve

Steven Samuels
sjsamuels@gmail.com
18 Cantine's Island
Saugerties NY 12477
USA
Voice: 845-246-0774
Fax:    206-202-4783

On Wed, Aug 25, 2010 at 11:41 AM, Bell, Cynthia
<Cynthia.Bell@uth.tmc.edu> wrote:
> The "sts list" command provides point estimates, std errors, and confidence intervals at exact times during my observation period. I prefer confidence intervals to compare, say, 10-year survival estimates between groups but a journal requested p-values comparing group point estimates at specific time points. Is this possible in Stata? If not, it shouldn't be complicated to compute a p-value given a distribution for the K-M survival point estimate and the standard error, right?
>
> Thanks very much,
>
> Cynthia Bell
> Division of Pediatric Nephrology and Hypertension
> University of Texas-Houston, Medical School
>
>

*
*   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