[FOSS_health] Re: interoperability

[Tim C]

On 04/06/07, Adrian Midgley <amidgley2 at defoam.net> wrote:
>
> Joseph Dal Molin wrote:
> > In both open and proprietary systems users of the same application
> > across multiple organizations usually have the power to add new
> > terminology etc. without some centralized "terminology"
> > standardization process.
> >
>
> If they can be, by dint of strenuous efforts at education,
> proselytising, engineering etc, persuaded and enabled to do so by means
> of a proper inheritance and sub-classing approach, then some of the
> effort and vagueness might be reduced.


Yes. openEHR, and so some extent the HL7 v3.x people, have proposed (but not
completely implemented) technical infrastructure to deal with these
problems. There are also a range of mature but rather technically-oriented
knowledge management/ontology development tools available for developing
shared ontologies and semantic representations, including tools used to
build on or subset SNOMED CT.

However, I can't help feeling that the technical infrastructure is the
trivial part, and that it is the sociological aspects which are the real
challenge in achieving long-term, sustainable semantic interoperability
between health apps.

Many nations have, or are beginning to develop, national health informatics
standards which may (or often may not) address some limited aspects of
knowedge representation (eg they may specify a specific code set for a
particular data item) - see for example teh Australian standards at
http://www.saiglobal.com/shop/Script/PortalInformatics.asp but specific code
sets are scarce within them - and some countries have national health data
dictionaries or knowledgebases - see for example
http://meteor.aihw.gov.au/- but these are far from comprehensive with
respect to clinical data,
although the Australian Meteor facility has good coverage of administratve
health data "metadata" (it is not strictly metadata which it contains, it is
meta-metadata, but we'll ignore that).

However, all these things are at the national level...

It seems to me, that with the advent of the SNOMED CT SDO (see
http://www.snomed.org/news/documents/snomed_sdo.pdf ) and the prospect of
affordable national licenses for developing and transitional countries, that
it might be possible to develop at least regional, multi-lingual semantic
representation (data encoding) standards for many health sub-domains, based
on subsets of SNOMED CT, thus leapfrogging the national-only scope of
earlier efforts in some developed countries.

Tim Churches
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.oshca.org/pipermail/foss_health/attachments/20070604/9dace377/attachment.htm