Paul
It would be interesting how many schools teach the representation of
rates and proportions as "natural frequencies", with "tree graphics" to
represent the relevant populations and sub-populations, instead of (or,
as well as) the standard 2x2 table. These are tools for avoiding
statistical errors such as confusing sensitivity with positive
predictive value.
The issue of representation of data is highly topical: the UK General
Medical Council has found Sir Roy Meadow guilty of giving erroneous and
misleading evidence. People who have read "Reckoning with Risk: Learning
to Live with Uncertainty" by Gerd Gigerenzer will know that Roy Meadow
committed "the prosecutor's fallacy" in his statistical reasoning. They
will also know that most medical (and legal professionals) would also
commit this fallacy, and that an effective tool for avoiding this sort
of error is to represent the relevant populations and sub-populations as
tree graphs, and to represent the statistics as "natural frequencies"
(explicitly giving the denominators and numerators, and explicitly
describing the relevant populations and sub-populations).
The GMC are now considering if Roy Meadow is guilty of serious
professional misconduct. If he is found guilty for making an error that
is "known" to be very common, would this be grounds for finding teachers
of statistical reasoning to be guilty of professional misconduct if they
do not provide their students with the best tools for avoiding common
mistakes?
Michael
Michael Power BSc Hons, MB BCh, MD, DCH
Clinical knowledge author, guideline developer, and informatician
Sowerby Centre for Health Informatics at Newcastle Ltd
University of Newcastle
-----Original Message-----
From: Evidence based health (EBH)
[mailto:[log in to unmask]] On Behalf Of
EVIDENCE-BASED-HEALTH automatic digest system
Sent: 13 July 2005 00:01
To: [log in to unmask]
Subject: EVIDENCE-BASED-HEALTH Digest - 11 Jul 2005 to 12 Jul 2005
(#2005-130)
There is 1 message totalling 38 lines in this issue.
Topics of the day:
1. survey of EBM curriculum
----------------------------------------------------------------------
Date: Tue, 12 Jul 2005 07:07:11 +0100
From: Paul Glasziou <[log in to unmask]>
Subject: survey of EBM curriculum
Dear All,
Does anyone have experience in trying to assess what EBM is taught in
various schools?
We are considering a survey and would like to know what has been done
previously.
The sorts of areas we are considering are:
1. Who is involved in teaching EBM (librarians, statisticians,
clinicians, ...)
2. What elements are explicitly taught in the curriculum
a. question recognition and formulation
b. searching skills
c. critical appraisal of
treatment studies
diagnostic studies
prognostic studies
systematic reviews
d. application to individual patients
3. What is the total face-to-face contact time spent on the above
elements,
either in lectures or tutorials?
4. Do students present a "critically appraised topic*" during the
course?
If so, how many times?
All suggestions welcome!
Thanks
Paul Glasziou
Department of Primary Health Care &
Director, Centre for Evidence-Based Practice, Oxford
ph: 44-1865-227055 www.cebm.net
------------------------------
End of EVIDENCE-BASED-HEALTH Digest - 11 Jul 2005 to 12 Jul 2005
(#2005-130)
************************************************************************
****
|