Jim Walker wrote:
> Most doctors know what signs that are "important" in most diseases but the
> problem as I see it is how to give each sign the proper weight.
I recall weekly seminars given by an old and wise neurologist who
said he approached each case with no pre-concptions, gathered
all the positive signs and symptoms without trying to force them
into a diagnostic category, and then attempted a synthesis,
which often ended up ignoring the odd and borderline. Maybe this
model is more applicable to specialities like nuerology than general
medicine surgery.
Another physician's dictum was "when presented with features that
do not fit patterns, think of a common disease presenting
atypically, rather than believing you have found a rarity."
> I have studied the diagnosis of appendicitis. For each variable I analysed the
> risk that the surgeon decided on an operation when the variable was present vs
> absent and compared this with the risk that the patient actually had
> appendicitis.
About 16 years ago as a student, I was shown a computer
programme: you fed in the clinical features, and it whirred and
buzzed and said something like "75% chance the patient has
acute appendicitis." Does anyone know if this sort of decision
support developed, or has been used?
Regards Alan O'Rourke
Alan O'Rourke
Information Officer
Wisdom Centre for Network Learning
http://www.wisdom.org.uk/
Institute of General Practice
Community Sciences Centre
Northern General Hospital Sheffield S5 7AU
Tel: 0114 271 5095 Fax: 0114 243 3762
E-mail: [log in to unmask]
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|