Statalist The Stata Listserver


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

Re: st: survival analysis: given median survival time 95% CI, how do es one calculate SE, t, or exact p?


From   "Erick Turner" <turnere@ohsu.edu>
To   rogerharbord@bigfoot.com, statalist@hsphsun2.harvard.edu, "Erick Turner" <turnere@ohsu.edu>
Subject   Re: st: survival analysis: given median survival time 95% CI, how do es one calculate SE, t, or exact p?
Date   Sun, 22 Oct 2006 09:32:25 -0700

Unfortunately, that paper assumes you're given the hazard ratio.
The Parmar paper is cited by a paper by Michiels with a promising title involving median survival times, but it also assumes you've moved beyond median survival times and have the hazard ratios. I guess I need something more at the level of the median survival times. 
Thanks.

>>> "Roger Harbord" <rogerharbord@bigfoot.com> 10/22/06 1:44 AM >>>
Sounds like you might find this useful:
Extracting summary statistics to perform meta-analyses of the published 
literature for survival endpoints
Mahesh K. B. Parmar, Valter Torri , Lesley Stewart
Statistics in Medicine 17 (24):2815-2834
<http://www3.interscience.wiley.com/cgi-bin/abstract/25000467/ABSTRACT>

Roger.

Erick Turner wrote:
> Dear Stata users,
>
> I am working with a report of data analyzed using survival analysis.
> Specifically, the survival probabilities were calculated using the 
> Kaplan-Meier method, and the estimated survival distributions were 
> compared between groups using the log-rank test.
>
> For each group, they report the N, median survival time, and the 95% 
> CI around the median. (The CIs were not calculated using a simple 1.96 
> x SE, because the upper and lower CI widths are different from one 
> another.)
>
> They also report exact P values, but only for where they get a value 
> under .05.
> Where the comparison between groups is nonsignificant, they just 
> report "NS".
>
> My goal is to find the exact P value (which I can get to if I can 
> first get SE or t).
>
> Would someone know how I can achieve this using Stata or some 
> not-too-difficult formula?
>
> Thank you.
>
> Erick





*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   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   |   What's new   |   Site index