Issues regarding archetypes:

AH > >Are these archetypes required for new GEHR/CEN compliance?

TB > No, they wouldn't be. They would be basic archetypes designed to a)
show
> software developers and users how the system works, and b) save a lot of
> work for users who can use the example archetypes (e.g. a basic set of
> GP archetypes would go a long way).

I think that this still needs to be discussed. We have found from our
initial research in Australia that it is possible to create very generic
archetypes which cater for a large range of information. It is my hope that
this will translate internationally. As we have based this work on the HL7
and LOINC experience it is likely to.

As sharing of base archetypes is the basis for interoperability (the
information model provides information sharing) it is clear that we need to
maximise the sharing of this hopefully small number of archetypes (perhaps
as low as 300).

AH >How will deviations from the "reference/default archetypes" be handled?
> >
TB > Well, archetypes are designed to have any amount of "deviation" you
> want, by either a) define a specialised (more restricted) archetype
> based on an existing one, or b) define a new version of an exisitng
> archetype (to fix an error), or c) just define a new archetype.

New versions of base archetypes are likely to loosen the constraint on the
prior version (this could be seen as an error but may relate to new
requirements). It is our plan at the moment to have a set of base archetypes
called 'any' which can be constrained locally to meet needs not met with the
other archetypes. It will be clear that these cannot be processed locally
unless they are known to the local system. Due to the multilevel
architecture it is not envisaged that these will lead to the same problems
we have seen with Z segments but this is yet to be proven.

TB > Note that higher-level archetypes tend to have pieces in them which
> allow almost anything, and lower level ones specialise this by further
> restricting these parts to be particular structures. So specialising
> does not mean you can't add features; in fact you can. See the
> biochemistry example in the archetypes paper.

TB > There will be a copy of the archetype editor we use on the web in the
> next couple of weeks, along with 60 or so archetypes. It is undergoing
> some last changes before release. It is based on the existing GEHR
> models, so is in a sense already legacy. But it will show what building
> archetypes is like, and generates real schema-compliant XML. If you
> write archetypes with this tool, we should be able to convert them to
> the new openEHR form when it is finalised.

Cheers, Sam
_______________________________________
Sam Heard
Director
General Practice Education and Research Unit
NT Clinical School, Flinders University
PO Box 41326, Casuarina NT 0811

Ph: +61 (0)8 8922 7937
Fx: +61 (0)8 8922 7928
W: www.gperu.org

The Good Electronic Health Record
www.gehr.org
_______________________________________


Reply via email to