Based on Charlie's reply, maybe my own was not clear. openEHR is mostly
not about solving the problem of openEHR systems talkng to each other
(that is rare at this stage, as you might imagine; when it happens,
there is not much problem to solve, obviously - openEHR is a single data
standard), but in fact solving normal interoperability problems with
other systems and formats. Connecting to HL7v2 messages, HL7v3 messages
(these seem to be pretty rare so far, apart from the UK and Canada) and
CDA documents, other formats like text, PDF, proprietary data
structures, but also the IHE framework, Microsoft CHF services
framework.... all this is already happening in the openEHR world. The
main point is _how_ it is done. Instead of manually building a message
specification for say 'microbiology', the openEHR approach is as follows:
* build or source archetypes, like you can see at
http://www.openEHR.org/knowledge e.g. see Apgar at
http://www.openehr.org/knowledge/OKM.html#showArchetype_1013.1.172
* create templates from the archetypes for use-case specific use
* from templates we create various computable artefacts - such as
template-based message schemas (XSD) and template-based
programming objects.
* template-based message schemas are used to implement messages,
either as an integration solution, or as an end-to-end solution
* template-based programming objects are used to enable integration
with existing applications
* all querying done with AQL or other archetype-path based
languages, which enable portable, semantic querying for all data
represented in openEHR format.
This singe-source semantic modelling approach is visualised at
http://www.openehr.org/160-OE/version/default/part/ImageData/data/single_source_models.jpg
. I would go so far as to say that the age of hand-built messages will
be over soon - it won't make people in HL7 happy I guess, but the
technology is now beyond that approach, and firmly in the arena of
model-driven content and SOA infrastructure.
The final standardisation within openEHR of the templates and
template-related artefacts are happening at the moment, based on the
last couple of years' experience with this technology within some of the
commercial openEHR vendors. Our experience of archetype-based querying
has been extremely successful, in both EHR and population modes.
Querying is one of the weakest areas in the published standards, yet it
is probably the single most important factor in the use and re-use of
data in health. Having a comprehensive and semantically solid querying
approach is therefore crucial for useful interoperability (after all
there is not much point moving data somewhere else, or transforming it
if you still can't query it properly). We believe that the combined
power of archetype-based templates and archetype-based querying will
become a major advance for the field of health informatics, along with
the emerging openEHR service infrastructure.
- thomas beale*
*
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.openehr.org/mailman/private/openehr-technical_lists.openehr.org/attachments/20100201/1905fc1f/attachment.html>
