Thanks to all who replied. And I just located an entry in the Encyclopedia of
Biostatistics that states that S, R, SAS (and as well STATA and SPSS) use the
Brookmeyer Crowley procedure for the median and its CI. I will need to locate
the manuals themselves for specific details of the calculations.
-----Original Message-----
From: Garrett, Robert [mailto:garrett@NRCan.gc.ca]
Sent: Tuesday, September 23, 2008 7:01 AM
To: Barker, Chris; Terry Therneau
Cc: s-news@lists.biostat.wustl.edu
Subject: RE: [S] Kaplan Meier - confidence interval for the median
Here is another alternative that was shared on S-News many years ago that I use
in a Tukey boxplot function.
The data are in temp and sssz is the data subset size:
if(notch) {
v <- sort(temp)
j <- qbinom(0.025, sssz, 0.5)
lci <- v[j]
uci <- v[sssz - j + 1]
x <- c(hih, low, low, mid, low, low, hih, hih, mid, hih, hih)
y <- c(bb[2], bb[2], lci, bb[3], uci, bb[4], bb[4], uci, bb[3],
lci, bb[2])
polygon(x, y, col = colr)
lines(x, y)
}
This same script is used in the R package 'rgr'
Bob Garrett
Robert G. (Bob) Garrett Robert G. (Bob) Garrett
Emeritus scientist scientique émérite
Geological Survey of Canada Commission géologique du Canada
Natural Resources Canada Ressources naturelles Canada
601 Booth St., Ottawa, Ontario 601 rue Booth, Ottawa (Ontario)
Canada K1A 0E8 Canada K1A 0E8
------------------------------------------------------------------------------------------------------
Tel.: 613+995-4517 FAX: 613+996-3726
Email: garrett@NRCan.gc.ca
website: http://gsc.nrcan.gc.ca/dir/index_e.php?id=4961
------------------------------------------------------------------------------------------------------
> ----------
> From:
> s-news-owner@lists.biostat.wustl.edu[SMTP:s-news-owner@lists.biostat.wustl.edu]
> on behalf of Terry Therneau[SMTP:therneau@mayo.edu]
> Reply To: Terry Therneau
> Sent: September 23, 2008 9:35 AM
> To: Chris.Barker@cvt.com
> Cc: s-news@lists.biostat.wustl.edu
> Subject: Re: [S] Kaplan Meier - confidence interval for the median
>
>
> >
> > I'd like to find out whether or not the method, used in SPLUS, for
> >computing the confidence interval for the median of a Kaplan Meier
> >estimate is the method of
> >Brookmeyer, Crowley, "A confidence Interval for the Median Survival
> >Time", Biometrics, 38, 29-41, 1982.
> >
> > Interestingly, for the simple dataset of , (10, 15, 23, 30, 35, 52,
> >100+) (and + indicates censoring)
> > SPLUS gives, the median and CI as 30 (15, NA), (R reports the upper
> >limit as inf)
> >
> > while SAS gives 30 (15,52)
> >
> > Chris Barker, Ph.D.
> >Consulting Statistician
> > CV Therapeutics
> >
>
> 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.
>
> Given a set of pointwise confidence bands for the survival curve S(t),
> draw a
> horizontal line on the graph at S = .5. The median survival is where that
> line
> intesects the curve S(t), the upper and lower confidence intervals are where
> it
> intersects the lower and upper pointwise bands, respectively. If it does not
> intersect, the corresponding end of the interval is unknown -- how to notate
> that to the user is a matter of personal preference.
>
> 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?
>
> Terry Therneau
>
>
> --------------------------------------------------------------------
> This message was distributed by s-news@lists.biostat.wustl.edu. To
> unsubscribe send e-mail to s-news-request@lists.biostat.wustl.edu with
> the BODY of the message: unsubscribe s-news
>
|