oship-dev team mailing list archive

["Sergio M. Freire" <sergio@xxxxxxxxxxxxxxx>]

Hi Luciana,

If you take your line of reasoning about the body weight archetype further, you may think that even the body weight measurement itself and how the patient was dressed at the time of measurement should not be part of the archetype, because they are not inherent to the concept of body weight. And what is left for the body weight archetype?

Well, I think that archetypes, IMHO, are not designed to represent domain concepts. This may look like a heresy, but for representing domain concepts and their relationships we have ontologies. 

The correct definition of archetype is "reference model constraints". The meaning of these constraints in a specifica domain should be looked at more carefully.

All EHR models are software constructs designed to show how to represent information that should be recorded in the EHR. openEHR approach is no exception, only instead of a single model we have two, the reference model and the archetype model. If we need to record a body weigth measurement, for example, than we need to record who took the measurement, who was weighed, when and where was the measurement taken, which device was used to take the measurement, the state of the patient at the time of measurement, and of course the measurement itself. Have I forgotten something? 

In openEHR, this measurement would be recorded in a composition. The reference model offers several classes and attributes to record part of the context of the measurement: who took the measurement, the patient, the time of measurement, and where the measurement was taken, but it does not specify how to record the measurement itself, the state of the patient and the device used. Then comes the body weight archetype and fills the missing part in. 

So, archetypes are artefacts that help to specify how to record information in an EHR. They are not aimed at representing domain concepts. If you agree with this, maybe you will find a role for archetype slots after all.

Cheers,

Sergio



----- Mensagem original -----
De: "Luciana Tricai Cavalini" <lutricav@xxxxxxxxx>
Cc: "OSHIP-Dev" <oship-dev@xxxxxxxxxxxxxxxxxxx>
Enviadas: Sexta-feira, 7 de Maio de 2010 20:28:15
Assunto: Re: [Oship-dev] Fwd: IHTSDO meeting - term binding presentation available

Hi Sergio and OSHIP-Dev,

Sergio M. Freire escreveu:

Hi,

I feel rather confortable with the archetype specification as it is. I
don't think we should change anything on the openEHR specifications
(especially because we can't). This is something for us to discuss about
the MLHIM specifications.


I won´t go into the philosophical discussions of what a concept is. But
that's what I would love to do, because I think we should consider
including more epistemological/scientific rigueur to the specifications.
Of course they are great the way they are right now, but they can be
improved, and I would like to give my contribution to that. Being inside
the academic world and having time and other conditions to develop those
reflections as a scientific project, in my opinion, is something that
gives us the priviledge of discussing in depht the immense philosophical
background of the multilevel modeling approach.
I think it is possible to discuss those theoretical subjects along with
the practical implementation of the specifications, especially because
OSHIP 1.x must be 100% openEHR compliant, what's important for our proof
of concept.


When Rigoleta and me designed the demographic archetypes we used and
abused of archetype slots because it is a construct that expressed well
what we meant. This is something for me to think about, because your
demographic archetypes are great, especially because they are openEHR
*and* ISO compliant at the same time. That *might* demonstrate the
conceptual differences between the demographic and the clinical models.
But then I go to the CKM and I download that body weight archetype with
that slot for the scale, something there just doesn't sound good to me.
Body weight is a well defined clinical concept, so it deserves an
archetype. Scale is another well defined clinical concept, so it
deserves another archetype. The body weight archetype and the scale
archetype can be used in combination to other clinical concepts. Scales
are not only used to weight bodies, otherwise it should be part of the
body weight archetype, because it would be part of the body weight
concept. So, why not combining those two archetypes in a template and
just forget about the internal slot on the body weight archetype? I
think this approach makes the knowledge modeling much easier for
healthcare professionals.


What are the practical reasons for discarding archetype slots? For OSHIP
1.x, I can't see any. For the MLHIM specifications, I think we should
keep discussing and thinking about moving forward.
Because I really think that isolating the healthcare concepts inside the
archetypes and only allowing their combination inside the templates is a
much "cleaner" approach that, I repeat, I think it could make more sense
for healthcare professionals.



What are the alternatives now? Are they better? Right now, as I said
before, no alternatives, let's go 100% compliant to the openEHR
specifications for OSHIP 1.x, and sorry if I am being repetitive about
it, but I am thinking about the newbies and students, so I am just
trying to be didactic. But I think we could take the opportunity of
being developing a research project and allow ourselves to perform those
experiments, because we have this freedom.
If they are better or not, there's only one way for us to know: by
performing the experiments and testing our hypothesis. How could we do
that? No other way: let's follow the openEHR Foundation principle #1:
implementation, implementation, implementation. Our new approach for
archetype governance, with expert panels and free and open debate and
polling on HKCR, might bring more light to this subject. Only the
development of the project will give us those answers, in my opinion.

Cheers, Luciana.



Sergio


----- Mensagem original -----
De: "Luciana Tricai Cavalini" <lutricav@xxxxxxxxx> Para: "OSHIP-Dev"
<oship-dev@xxxxxxxxxxxxxxxxxxx> Enviadas: Sexta-feira, 7 de Maio de 2010
11:17:29
Assunto: Re: [Oship-dev] Fwd: IHTSDO meeting - term binding presentation
available

Hi OSHIP-Dev,

I'm going to express my opinion on what I feel comfortable about:
MLHIM specifications' concepts should be clearer than the openEHR ones,
that are still open to a few amount of misinterpretations.
For me, an Clinical Constraint Definition (= "archetype") is like an
atom, so it should be complete and self-contained. It should not allow
slots. Templates are molecules, and the connections between archetypes
needed to create those "molecules" are the slots in my opinion. So,
slots should be limited to templates.

Sorry for the chemical analogy...

Debate! Debate! ;-)

Cheers, Luciana.


Tim Cook escreveu:


Hi All,

Below is a thread from the openEHR lists regarding terminology binding.
It should be read by all addressed here.

I have considered these issues in the past and since I am presenting
MLHIM to a US company in two weeks with a focus on just this problem. I
wanted to let you all know what I am 'planning' to put in the MLHIM
draft specs.

The problem that is most vexing is one of post-coordinated binding. This
problem is closely related to what I am talked with some of you about
before. The complexity and counter-intuitiveness of slots in concept
constraint definitions (CCDs aka. archetypes).

Given: an archetype is defined as being a "complete representation of a
single clinical concept".

Then: Why archetype slots? Everything for that concept SHOULD be a part
of that CCD/archetype.

Therefore: A CCD should be a complete clinical concept. If/when the
concept is too broad to utilize pre-coordinated terminology, then it
should be marked as ABSTRACT (in the object-oriented sense of the word).
This allows a CCD to be created and then subclassed from (aka. archetype
specialization) but it prevents that specific abstract CCD from being
instantiated.

Of course in the specs this will be expanded on so it makes more sense
to new readers. But I wanted to pass this by knowledgeable people to get
feedback. The concept of abstract schemas is support in XML schemas.

Regards,
Tim


---------- Forwarded message ----------
From: Seabury Tom (NHS Connecting for Health) < tom.seabury@xxxxxxx >
Date: Thu, May 6, 2010 at 11:32 AM
Subject: RE: IHTSDO meeting - term binding presentation available
To: For openEHR clinical discussions < openehr-clinical@xxxxxxxxxxxxxxx
Hi Mikael, Thomas



There are certainly instances of sets being generated by authorities
other than national release centres, ones that conform to the SNOMED CT
guidance, which requires their identifiers include a recognised
namespace.

Two aspects of this topic are 1) that you need access to tools which
allocate unique identifiers, in accord with the standard and 2) that the
allocation of namespace identifiers cannot be assumed to be a
free-for-all although at last count there were in the region of 110,
some of which were allocated to individuals. To use a Namespace you are
obliged to conform to the SNOMED CT licence conditions and similar good
behaviour.



Those with an interest in this topic may wish to join, for example, the
SNOMED CT Implementation Special Interest Group via this link:-
http://www.ihtsdo.org/join-us/help-improve-snomed-ct/ To join an
established Special Interest Group mail to: info(at)
ihtsdo.org stating which group you wish to join,

‘Special Interest Groups (SIGs) are open and ongoing bodies that examine
issues related to specified topics that are relevant to the IHTSDO and
its members.’



Tom Seabury

IHTSDO



From: openehr-clinical-bounces@xxxxxxxxxxxxxxx [mailto:
openehr-clinical-bounces@xxxxxxxxxxxxxxx ] On Behalf Of Mikael Nyström
Sent: 06 May 2010 16:51
To: 'For openEHR clinical discussions'
Subject: RE: IHTSDO meeting - term binding presentation available




Hi Thomas,



As another member of IHTSDO Technical Committee I would like to ask from
where you have got your impression that RefSets should mainly be
maintained by the national release centers, because that is not at all
my impression. Some of the national release centers, like NHS in UK,
maintain Subsets/RefSets, but I have never heard anything about that
only the national release centers should create and maintain RefSets. My
impression is that whoever that find it useful to create and maintain an
own RefSet should be able to do it.



The only requirement to create a RefSet is the technical requirement to
have an own namespace to create the RefSet into. However, all
organizations and persons with a valid SNOMED CT license can get an own
namespace.



(I have my own namespace, so if I would I could immediately produce and
release as many RefSets that I want.)



Greetings,

Mikael







From: openehr-clinical-bounces@xxxxxxxxxxxxxxx [mailto:
openehr-clinical-bounces@xxxxxxxxxxxxxxx ] On Behalf Of Thomas Beale
Sent: den 6 maj 2010 12:15
To: openehr-clinical@xxxxxxxxxxx Subject: Re: IHTSDO meeting - term
binding presentation available



On 06/05/2010 08:48, Sebastian Garde wrote:



Hi Thomas,

do you know if there is a formal way of how RefSets (=the resulting
Snomed CT codes etc.) and the RefSet query (=the query on Snomed CT to
get to the RefSet) are expressed and shared?


the ref set results are defined by Snomed RF2 ref set specs, now in
draft. I don't know whether they are allowed to be shared outside
IHTSDO, but will find out. The query language is still an open question.



Similar to what is described here but based on RefSets:
http://www.openehr.org/wiki/display/term/Ocean+Terminology+Query+Language+%28TQL%29
I agree that RefSets are a good way forward, but they need to be
available, reusable and sharable, etc.


they do. The current IHTSDO thinking is that ref sets will mainly be
defined by national release centres and maintained in the national
extensions, with some major refsets being in the international release.
I think this is far too restrictive a view, and that in future they will
be defined and maintained in health care facilities, regional
information centres and anywhere else that needs them. The IHTSDO
representation and extension mechanism should be usable at all these
levels. How it could be used for openEHR I have yet to find out, but am
making enquiries as a member of the IHTSDO Technical Committee).

- thomas
********************************************************************************************************************

This message may contain confidential information. If you are not the
intended recipient please inform the
sender that you have received the message in error before deleting it.
Please do not disclose, copy or distribute information in this e-mail or
take any action in reliance on its contents:
to do so is strictly prohibited and may be unlawful.

Thank you for your co-operation.

NHSmail is the secure email and directory service available for all NHS
staff in England and Scotland
NHSmail is approved for exchanging patient data and other sensitive
information with NHSmail and GSI recipients
NHSmail provides an email address for your career in the NHS and can be
accessed anywhere
For more information and to find out how you can switch, visit
www.connectingforhealth.nhs.uk/nhsmail
********************************************************************************************************************

_______________________________________________ openEHR-clinical mailing
list openEHR-clinical@xxxxxxxxxxx
http://lists.chime.ucl.ac.uk/mailman/listinfo/openehr-clinical -- Esta
mensagem foi verificada pelo sistema de antiv�rus e
acredita-se estar livre de perigo.
_______________________________________________ Mailing list:
https://launchpad.net/~oship-dev Post to : oship-dev@xxxxxxxxxxxxxxxxxxx
Unsubscribe : https://launchpad.net/~oship-dev More help :
https://help.launchpad.net/ListHelp -- Esta mensagem foi verificada pelo
sistema de antiv�rus e
acredita-se estar livre de perigo.
_______________________________________________ Mailing list:
https://launchpad.net/~oship-dev Post to : oship-dev@xxxxxxxxxxxxxxxxxxx
Unsubscribe : https://launchpad.net/~oship-dev More help :
https://help.launchpad.net/ListHelp
-- Esta mensagem foi verificada pelo sistema de antiv�rus e
acredita-se estar livre de perigo.
_______________________________________________ Mailing list:
https://launchpad.net/~oship-dev Post to : oship-dev@xxxxxxxxxxxxxxxxxxx
Unsubscribe : https://launchpad.net/~oship-dev
More help : https://help.launchpad.net/ListHelp