12 responses to my question produced an interesting response:
1. Several organisations have developed national, regional or local question
answering services (some of whom admit to having curtailed searches to meet
deadlines)
2. Library services are available and we should not forget them in
practising EBM
3. Several discussed the learning value of the process of EBM rather than
the final evidence itself
4. Dont rely purely on the abstracts
BUT
No-one can tell me what the likely harm is in doing quick and dirty - so I
have been developing my idea of a question and the process.
If one took 10 (or 100) questions occuring in practice this morning in
family medicine
Wrote down one's answer based on current knowledge - ie drug a reduces
endpoint effectively - nnt about 12, nnh for common non serious side effect
about 20, not sure about cost effectiveness. Main question for me now is the
NNT as this drug is expensive.
& did a search using any or all of the following
one electronic bibliographic source of appraised evidence - ATTRACT,
BESTBETS, Clinical Evidence, Best Evidence etc
one on-line search for Medline
one systematic review source sych as Cochrane
Perhaps limiting the search to 60 minutes??
1. How long would it take
2. How many articles would one find
3. How many articles would one like to have full text access to
4. How many articles does one have full text access to
5. What is the answer according to appraisal of accessible full text
articles
6. Does that answer change according to appraisal of full text articles one
does not have access to.
7. Does a more thorough librarian search a) reveal more articles b) that
produce a different answer
I still guess the answer would be that maybe 5 out of 100 questions might
change with access to ALL full text that one wanted.
Martin
|