In article <[log in to unmask]>, [log in to unmask]
writes
>> The human could have done the calculation, but not in a time that
was at all relevant to the patient with belly-ache, who would either
have died or got better, as the case may be before the human came up
with a result. <<
Too true, but I find that driving a computer in a normal consultation is
sooo slow in comparison to thinking. However, when I'm stuck then I
would like the decision support it could give me as a further
investigation - where it could act as a logical exploration of
available facts.
>>A computer is NOT just an abacus; it is a stored program device; <<
I always thought the abacus had its program stored in its structure -
form = function and all that.
>>Interesting bit on this evening's news about the analysis of NMR
scans, where it would appear that the use of a genetic algorithm in the
diagnosis of brain tumours is capable of yielding results which are at
least comparable with those of a trained radiologist.<<
It's getting there, where you put the patient into a machine at one end
and a diagnosis comes out the other. From what I read this was similar
to Dombul's work where data was compared to a synopsis of many doctor's
experience which was applied consistently and without bias using digital
NMR scans rather than analogue signs and symptoms. I assume, probably
innacurately, the only difference is the type of data and a few more 'if
- then - else' statements in the 'genetic' algorithm. However, I hope
it's more sophisticated than that.
The computer is good if you feed it very specific, very accurate data
and decision nodes, I've done that stuff myself, but I long for the day
when it applies to the daily grind of half-remembered, confused,
innacurate, unrelated morass of a normal surgery consultation (and
that's just the doctor!).
Regards
George Myszka
Peterborough GP
---------------
Z J M
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|