At 12:21 pm +1000 11/5/06, Horst Herb wrote:
On Thursday 11 May 2006 12:08, Andrew Patterson wrote:
> HL7 V3 goes a long way to eliminating variability, but this is no
> small task and real implementations are sparse.
See this is where I think the disconnect between proponents
of HL7 and I are..
I have spent the last 2 weeks "outputting" HL7v3 from a
clinical system and from my point of view the specs
go nowhere near eliminating variability - I was left
with about 50 "choices" as to how the HL7 would be structured
(is a 'progress note' an observation and how should it be coded -
if I have a set of 'allergy' observations, where do I put them in a
patient record and how do I distinguish them from other observations).
So here I am, making choices as a technical person that I
don't feel at all comfortable making. I don't want to make those
The problem is generalisation.
HL7 is aiming at being as general as possible.
In extremis, generalisation would mean "simply" re-implementing a general
purpose language like English. HL7 adds a little structure to the ambiguity,
but it doesn't solve the problem.
In our domain, we don't need overly generalized messaging. We have a very
limited scope that really matters. Messaging for limited scopes would be very
easy and quick to implement - alas, we are too focussed on overly
generalizing.
Messaging for limited scopes is useless for broader interoperability
- you can have a single point to point interface in minutes - it will
work for a single purpose between two systems. But try to connect to
more than two system. That method works only for M$ and others with
monopoly power.
OpenEHR has the better approach IMHO - but my personal preference would be
even more domain specific (and thus simpler)
Horst
But if you want interoperability, you need "generalisation" to the
degree that "(nearly) everybody has to agree". And that is the
fundamental problem.
If 5 people decide the right answer and try to make the other 400
agree, it may work or it may not, depending on the politics and trust.
You can write a specification for messaging. So could I, but who is
going to listen?? Solo efforts in standardisation are rather futile.
Ian.
--
Dr Ian R Cheong, BMedSc, FRACGP, GradDipCompSc, MBA(Exec)
Health Informatics Consultant, Brisbane, Australia
Internet: [EMAIL PROTECTED]
(for urgent matters, please send a copy to my practice email as well:
[EMAIL PROTECTED])
PRIVACY NOTE
I am happy for others to forward on email sent by me to public email lists.
Please ask my permission first if you wish to forward private email
to other parties.
_______________________________________________
Gpcg_talk mailing list
[email protected]
http://ozdocit.org/cgi-bin/mailman/listinfo/gpcg_talk
