Gerard Flaherty writes:
>Welcome Rob Hampton of NHS CCC to GP-UK
G'day
>I am unable to understand some of your ideas fully. Especial ennumeration
>and compostitional.
Enumerative coding schemes are lists of different individual concepts. Each
varaition needs a new code. This, in my experience, is the natural way that
doctors usually build up schemes for their own use. It is good for small coded
sets because it's easy to use. However, it soon loses out as a method to cover
more detailed terminology. For example 'Psoriasis of L elbow' would need a
different code to 'Psoriasis of R elbow' and another for 'Guttate Psoriasis of L
elbow' etc. An explosion of codes and terms results and becomes impossible to
navigate. Imagine typing psoriasis and getting a hit on all individual
variations of it's type, site and side.
Compositional schemes code individual concepts and combine them to build up
complex statements. It is the scheme that people come up with when locked in a
room and asked to develop a model for clinicla terminology. There is a seperate
code for 'skin of elbow' 'Right', 'Left' and 'Psoriasis'. By combining the codes
for Psoriasis|Skin of elbow|Right you can build up the concept. The cluster of
codes to describe the same thing on the Left would simply replace one code of
the three. The beauty of this is that the number of codes in the scheme is kept
reasonably small and it is easily maintained. However it is more complex to
actually build into a working system, hence much of the GP-UK debate on V3 use
and usability to date. Also it raises the spectre of ensuring rules are enforced
to prevent inappropriate combinations such as Left|diabetes|eyebrow. More rules
introduce a more difficult implementation.
Read V0, V1 and V2 are pure enumerative schemes. People rightly complain that
there are certain things they can't say with these versions. Hence the effort
(and yes - Govt money) into the creation of Read version 3. This has an
enumerative framework but allows composition through linked 'qualifiers' the
name given for codes describing Left/Right, anatomy etc.
>Formal evaluation by an American based organisation seems odd.
There is an Institute funded by American IT industry to look at improving the
use of computers in healthcare. They have looked at Read V3 as compared to
SNOMED and the US National Library of Medicine clinical coded thesaurus for
their own evaluation of a recommended scheme for the future. I'll inform you of
the outcome when I can. The NHS CCC would welcome any similar evaluation in
the UK, Europe or abroad. We can't do it. BIAS would always be the accusation
if we took on the evaluation.
More later. I hope this answers a couple of your queries
Rob Hampton NHS CCC
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|