What Peter Lewis has touched on but not drawn the distinction is that there
are two (at least - can anyone think of more?) levels of appraisal:
1. We need a response to a problem in our busy lives as clinicians, usually
presented by patients or developments or service proposals. A *quick and
dirty* answer (no offence to editors intended) is needed, which may at times
(if lucky) come from level 2 appraisals - eg a look at existing digests of
evidence such as Clinical Evidence.
2. There is a need for topics to be critically appraised for wider
application - they become CATS, clinical practice guidelines, policy such as
whether or not riluzole should be available on NHS prescription. Systematic
reviews with all the work that they entail are essential to level 2
appraisals.
--
Kev Hopayian, Seahills, Leiston Rd, Aldeburgh, Suffolk IP15 5PL, England
----------
>From: Peter Lewis <[log in to unmask]>
>To: Evidence-Based-Health <[log in to unmask]>
>Subject: algorithm
>Date: Tue, Feb 1, 2000, 11:51 am
>
> The list seems a bit quiet so I thought I would run this past you....
>
> The standard algorithm for appraising evidence is:
>
> 1. Pose question
> 2. Establish method for appraising evidence
> 3. Appraise evidence
> 4. Publish results
> 5. Hear from people who use different method
> 6. Recognise need for method for deciding which method is best
> 7. Pose question
> 8. Establish which method should be used for deciding which method is best
> 9. Appraise methods for deciding which method is best
> 10. Publish results
> 11. Hear from people who use different method for deciding which method is
> best
> 12. Recognise need for method for deciding which method is best for deciding
> which method is best
> 13. (or later) spot the pattern & wonder.....
>
> Sooner or later you realise that somewhere in all of this you have to put a
> value judgement to break the cycle otherwise you spend all your time
> inventing methods and no time taking even half sensible decisions. We
> cannot expect any manager (clinician or otherwise) to take perfect decisions
> because we just dont have perfect data, so why strive for the hypothetical
> perfect method of appraising? What EBM tries to do (IMHO) is get a decent
> summary of what you should do in given circumstances. It doesnt matter
> which of n methods you use for evaluating evidence as long as we all use the
> same one and it is reasonably sensible.
> The alternative is that you finish up by having your clinical management
> overtly determined by where you live, and I thought we were trying to avoid
> that.
>
> Peter
>
> Peter Lewis [log in to unmask]
> Senior Lecturer in Bio-statistics
> Public Health Group
> School of Postgraduate Medicine
> University of Bath, Bath, BA2 7AY
> desk 01225-323368 sec 01225-826400
> fax 01225-323833
>
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|