>I agree - SNOWMED is the way forward, and I don't have a problem with
>that.
>
>However, my understanding is that whereas Read is hierarchical - and
>SNOWMED isn't.
>
>I don't think you can regard them as the same thing: the *Terms* will be
>the same - but the *structure* - and the way the coding structure is
>handled - won't.
>
>Any views on the technical side?
>
>MaryH
Mary,
1) You're correct that SNOMED is not hierarchical. This avoids the
tuberculous meningitis paradox where one condition correctly has two
codes. All codes MUST now be unique.
2) Individual issues such as side, eg. L/R, site, eg tibia, medical event
eg. fracture each attract one code "atom", and the code for a fracture of
the R tibia is then a "molecule" composed of these atoms.
3) According to a presentation at EMIS NUG (by a coder) the back end SNOMED
might even be invisible in some GP computer systems, and hidden behind a
Read 5 byte front end.
I think that would be a pity, but don't doubt there are complex issues to
resolve in any GP implementation of such a rich coding system and huge
issues in any retrospective data transfer.
I'll leave others to answer the truly technical questions.
Julian
|