Hello and thanks to everybody involved in the
discussion around meta-analysis of ROC areas.
Thank you Tom for putting me on the right track and
explaining what is an SROC and you were of course
right, that was different from what I wanted to do.
Thanks also for the tips regarding using the same
cut-off point and comparing sensitivity and
specificity across studies (however this is not
possible in my material).
Thank you Thomas for offering to share the program you
have developed. I will contact you in a private
e-mail.
I managed to meta-analyse ROC curves in the end in an
easy way. I used the roctab command in order to obtain
the auc (from r(area)) and se for auc (from r(se)) for
each center and then I run meta command using auc and
se of auc. Is my approach correct?
Thank you also Roger for suggesting using the
-somersd-,-parmest- and -metaparm- packages. I suppose
that you meant an approach as the one recommended in
http://www.stata.com/statalist/archive/2002-12/msg00001.html
where you pointed as potential advantage the fact that
it is possible to calculate CI for the difference
between 2 ROC curves (compared to -roccomp-). More
comments or other advantages to have in mind?
With best regards,
Andrei
--- Roger Newson <[email protected]> wrote:
> Just in case anybody really does want to
> meta-analyse ROC areas, this can be
> done using the -somersd-, -parmest- and -metaparm-
> packages, all
> downloadable from SSC using the -ssc- command.
>
> I hope this helps.
>
> Roger
>
> --
> Roger Newson
> Lecturer in Medical Statistics
> Department of Public Health Sciences
> Division of Asthma, Allergy and Lung Biology
> King's College London
>
> 5th Floor, Capital House
> 42 Weston Street
> London SE1 3QD
> United Kingdom
>
> Tel: 020 7848 6648 International +44 20 7848 6648
> Fax: 020 7848 6620 International +44 20 7848 6620
> or 020 7848 6605 International +44 20 7848 6605
> Email: [email protected]
> Website: http://phs.kcl.ac.uk/rogernewson/
>
> Opinions expressed are those of the author, not the
> institution.
>
>
>
> Ben wrote:
> From: [email protected]
> [mailto:[email protected]]On
> Behalf Of Ben Dwamena
> Sent: 13 December 2005 19:46
> To: [email protected]
> Subject: RE: st: meta-analysis and ROC
>
>
> Tom,
> I am interested in using your code as a basis for
> developing a metaroc
> program incorporating other methods including the
> Littenberg-Moses
> algorithm.
> Thanks
> Ben
>
> Ben Adarkwa Dwamena, MD
>
> Assistant Professor of Radiology
> Division of Nuclear Medicine
> Department of Radiology
> University of Michigan Medical School
> 1500 E. Medical Center Drive
> B1 G505 University Hospital
> Ann Arbor, MI 48109-0028
>
> [email protected]
>
> http://sitemaker.umich.edu/metadiagnosis
>
>
>
> Staff Physician
> D748 Nuclear Medicine Service (115),
> VA Ann Arbor Health Care System
> 2215 Fuller Road
> Ann Arbor, MI 48105
> 734-761-7886 Phone
> 734-761-5229 Fax
>
>
>
> >>> [email protected] 12/13/05 1:54 PM >>>
> Two or three years ago, I started into an SROC Stata
> program following
> the method of Littenberg and Moses. My code was
> based, in
> part, on some Stata code written by Ben Littenberg
> but never published.
> The code performs a meta-analysis of AUC's as well
> as some
> diagnostic tests.
>
> Ben and I started an exchange on the functionality
> of the code but
> other tasks interfered and I had completed the
> specific work I
> needed to do, so I never returned to the SROC
> program. I seem to
> recall that Ben no longer believed that computations
> on Q* should
> be done and he preferred to use the AUC from 0 to
> the max observed
> value rather than from 0 to 1 (as my code does).
>
> Regardless, you a welcome to a copy of the code "as
> is". Drop me a
> note at [email protected] and I'll attach the ado,
> hlp and dlg to
> a reply message.
>
> Further, if anyone wants to develop this code
> further, I'd be happy to
> turn it over to you!
>
> Tom
>
> Thomas J. Steichen
> [email protected]
>
----------------------------------------------------------------------------
> Facts do not cease to exist because they are
> ignored. - Aldous
> Huxley
>
----------------------------------------------------------------------------
>
>
> > -----Original Message-----
> > From: [email protected]
> > [mailto:[email protected]] On
> Behalf Of
> > Tom Trikalinos
> > Sent: Tuesday, December 13, 2005 1:28 PM
> > To: [email protected]
> > Subject: Re: st: meta-analysis and ROC
> >
> >
> > Hi Andrei,
> >
> > Summary ROC (SROC) is not exactly a meta-analysis
> of "ROC"
> > curves. It's
> > somehow different.
> > SROC provides a summary measure of diagnostic
> performance
> > when you only
> > have sensitivity & specificity values per study.
> Each study is
> > represented by a pair of sensitivity-specificity
> values
> > rather than the
> > ROC curve. SROC was devised because often the
> cutoff points of the
> > diagnostic test variables differ across studies,
> and in fact SROC is
>
> > informative only when the cutoff points differ.
> >
> > If you really want to meta-analyze ROC curves, you
> can in principle
> > meta-analyze AUC values with an inverse-variance
> model (you need the
>
> > AUC variance for this).
> >
> > Otherwise you might be better off with a
> meta-analysis of sensitivity
>
> > and specificity across studies, after selecting
> the same cutoffs for
>
> > the patients of each center.
> >
> > Having all values for all patients, I might opt
> for different
> > analyses
> > rather than a SROC curve.
> >
> > For the record there's no module in Stata that
> does SROC analyses, to
>
> > my knowledge at least. You could find the SROC
> manually in
> > Stata, but
> > it seems you don't want this.
> >
> > tom
> >
> >
> >
> >
> > On Dec 12, 2005, at 10:07 PM, Andrei Malinovschi
> wrote:
> >
> > > Dear Statalisters,
> > >
> > > I am having a database where we are looking at
> > > different symptoms (categorical variables) and
> > > different diagnostic tests (continuous
> variables). The
> > > data comes from several centers and we are
> trying a meta-analysis
> > > approach because we can't pool the data (due to
> differences between
>
> > > methods used in different centers).
> > >
>
=== message truncated ===
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
*
* 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/
