s-news
[Top] [All Lists]

Re: Kaplan Meier - confidence interval for the median

To: Terry Therneau <therneau@mayo.edu>
Subject: Re: Kaplan Meier - confidence interval for the median
From: Thomas Lumley <tlumley@u.washington.edu>
Date: Tue, 23 Sep 2008 12:54:09 -0700 (PDT)
Cc: Chris.Barker@cvt.com, s-news@lists.biostat.wustl.edu
In-reply-to: <200809231335.m8NDZaa19096@hsrnfs-101.mayo.edu>
References: <200809231335.m8NDZaa19096@hsrnfs-101.mayo.edu>
On Tue, 23 Sep 2008, Terry Therneau wrote:

  I don't have a copy of that issue of Biometrics so can't answer the question
directly, but I can tell you what survfit does do.  I'm moderately sure that it
is the method found in said paper.


I was asked this (about the R version) a couple of weeks ago.

It is the same in spirit but not in letter. Brookmeyer & Crowley invert a confidence interval for the median, but they create the confidence interval on the probability scale, not a transformed scale.


  Beware!  The quality of this procedure depends on the quality of the original
pointwise bands.  There are several methods to generate the bands, and lots of
papers comparing them.
   empirical likelihood (Thomas and Grunkmeier): excellent
   bootstrap methods: okay to excellent, depending on details of how its done
   transformed CI, log, logit, log-log, etc: very good.  Different papers rank
        different ones of these first in the list.
   plain,  S(t) +- 1.96* std: poor

There is no excuse for a modern package to use the last of these, given the
literature.  Guess which one is the default in SAS?


Probably the one used by Brookmeyer & Crowley.

        -thomas

Thomas Lumley                   Assoc. Professor, Biostatistics
tlumley@u.washington.edu        University of Washington, Seattle

<Prev in Thread] Current Thread [Next in Thread>