Important issue since the cebm levels of evidence is the one most
commonly used by journals that rate the levels of evidence of their
published studies. In that vein changes should be thoughtful. I wonder
were the developers of the current version involved in the update since
that seems important? I have made a few preliminary comments.
1. I understand that SR have evolved since the levels first defined but
not sure that means than RCTs should become level 2; another trend has
been the mega-trial which might be more definitive than a SR of small
studies- so not sure that the revised system is an improvement over the
a,b,c level 1.
2. i understand the "grade up/grade down" footnote/rationale. So having
it in the table under level 2 treatment is redundant( " or exceptionally
dramatic effect") It should be clear how much you grade up or down- 1
level??- the option to do so makes sense- but without
definition/parameters. It will mean that reliability between raters
might be quite low.
3. IN treatment level 4 case series (i.e. single group retrospective)
basically disappears and since this is the most common study design in
some journals
Good luck with considerations
Joy MacDermid
Professor
McMaster university
-----Original Message-----
From: Evidence based health (EBH)
[mailto:[log in to unmask]] On Behalf Of Martin Dawes,
Dr.
Sent: May-28-10 2:43 PM
To: [log in to unmask]
Subject: Re: Beta OCEBM Levels! ready for special viewers!
Hi
this is a really clever way of approaching the problem
I think it works but.....
You need to link every term to a glossary with an example - people don't
know what a systematic review is - they just think they know.
what do you mean by current in current survey etc
it needs a manual
the manual should have examples
dare I suggest a podcast to introduce how you might use it and why
in the harms are 'individual case descriptions' a type that is missing -
esp now we have new journals for cases!
I dont think we ever 'do nothing' - just your presence in the room, the
lift of an eyebrow etc can be a major effect - it could be abbreviated
to 'what will happen?'
but those are minor comments
the commonest question a GP asks is 'what is the cause of symptom x' -
the second is 'what is the cause of sign x' the third is 'Is drug x (or
drug class x) indicated in situation y or for condition y?'
so causation is missing -
This is a major gap I would suggest is filled - or we try and fill
so my sandwich ends by again saying how much better this is and that
these comments are to be taken into account knowing that
Martin
On 2010-05-28, at 9:39 AM, Jeremy Howick wrote:
Dear All,
This is your chance to influence the quality of health care research!
I am pleased to announce that Oxford's Centre for EBM has launched a
beta version of its evidence 'hierarchy':
http://www.cebm.net/index.aspx?o=5513
We hope it is an improved tool for answering relevant clinical
questions.
We invite feedback (sent directly to jeremy.howick[at]dphpc.ox.ac.uk,
not this Listserve!) until September 1st. Soon after we will replace the
current levels, and also publish an extended explanatory document that
will take your comments into consideration.
An ESSENTIAL new feature of the levels is that you CANNOT access it
until you read the brief introduction. Our introduction both explains
some essential features of the table, and insists that the hierarchy be
interpreted as a heuristic tool rather than a universal guide to what
counts as good evidence.
I look forward to your comments!
Best wishes,
Jeremy
Jeremy Howick, PhD
Centre for Evidence-Based Medicine
Rosemary Rue Building
Old Road Campus
University of Oxford
Oxford OX3 7LF
United Kingdom
tel: +44 (0) 1865 289 363
fax: +44 (0) 1965 289 336
|