Hi,
Apologies in advance if this is the incorrect email list for this topic, but I thought it was the most relevant.
I’m a member of OpenMRS and there we are discussion a way to have users share the concepts (a limited form of an archetype) created in their systems. This means that for a single concept you could have many concepts from different implementations. This could be because of language or because different words refer to the same concept. For example, Gender in the US might be Sex in another country and Sexo in spanish.
I would like to see if OpenEHR has solved this problem so that perhaps OpenMRS could begin to use archetypes.
obvious a proper analysis would require a more detailed requirements expression, but in terms of languages, archetypes most likely do what you need. For an example, see the Apgar archetype at openEHR.orgClinical Knowledge Manager (CKM):
in the left hand side explorer, navigate to EHR Archetypes > Entry > Observation >Apgar Score
double click
in the toolbar in the main pane, choose any of the first 3 views (one page HTML, tabbed HTML, mind-map) and then choose the language you want on the right hand drop down
To see how this is represented in the raw form, use the 4th or 5th views (ADL or XML).
OpenEHR is a language and ontology neutral, standard. Archetype class is ready for translations via its ancestor: AUTHORED_RESOURCE that has unbounded instances of TRANSLATION_DETAILS. Several phrases (words) pertaining to a sole concept can be implemented via Terminology package that helps with assigning terminologies of any kind to map to Archetype parts.
The issue of openMRS implementations having different representations of the same thing is a common problem across clinical systems everywhere. Its this problem that is one of the things that we are trying to solve with archetypes. In general, what we find is that most clinical concept representations in clinical systems are subsets (based on a use case) of a fully specified concept. What we try to do in the archetypes is produce the fully specified concept and then constrain it for all the use cases. The different names that you see used for different concepts are not just language dependent, but are also region and usage dependent. This is also usually a matter of constraining the archetype or using a particular terminology subset to represent the information.
What openEHR offers openMRS is a single way of representing clinical information that becomes a logical record architecture. If openMRS adopted this approach, then any openMRS system could immediately share information with any other even if the other system hadn’t seen the information before. It also means that the burden of developing high quality, clinician validated information models is shifted away from the application developer to a global or regional space. This is going to become more and more critical, as we try to capture more complex clinical information and compute on it as well as share it.
Just to add to Hugh's excellent comments. One of the issues you will
find is that it is sometimes impossible to harmonise all the competing
perspectives, even for something as seemingly simple as demographics.
The archetype driven approach, with maximal dataset philosophy, allows
these competing views to co-exist, very visibly and acts as a spur for
future harmonisation. 'Minimal dataset' philosophies lose this
knowledge and any future attempts at rationalisation essentially have
to start from scratch. So, as well, as capturing shared, agreed
requirements, we are also capturing dissent, which we can try again to
resolve in the future.
You should have a look at the Demographics model archetypes on CKM at
www.openehr.org/knowledge
We have had a few discussions with openMRS people in the past, and I
am sure there is real value in collaboration. I think the archetype
methodology and review process would be of great value, even if
openEHR was not used formally at the back-end of openMRS systems.
Ian
Dr Ian McNicoll
office +44 (0)1536 414 994
fax +44 (0)1536 516317
mobile +44 (0)775 209 7859
skype ianmcnicoll
ian.mcnicoll@oceaninformatics.com
Clinical Modelling Consultant, Ocean Informatics, UK
openEHR Clinical Knowledge Editor www.openehr.org/knowledge
Honorary Senior Research Associate, CHIME, UCL
BCS Primary Health Care www.phcsg.org
Just to add to this. Another great aspect of openEHR is the separation of the technical and medical (content) aspect.
In the clinical knowlegde manager (which Thomas already referred to) clinicians can cooperate to create archetype without have to think about the technical aspects. As for the technical people: whatever the clinicians come up with, they don’t have to change any code, unless it would require a change in the reference model, which is extremely rare.
Thank you for all of the responses. I have forwarded them to the OpenMRS developer team. A couple of other questions arose that I was hoping to get cleared up.
1. Are there tools or a portal for archetypes where different users can share their archetypes?
2. Have archetypes for Tuberbulosis, HIV and Malaria been created that can be used in OpenMRS?
re Malaria and Tuberculosis - it depends exactly what you mean. There
is a problem-diagnosis archetype which might capture some of the
required information (making use of a terminology) but I suspact there
may be other disease-specific monitoring/ therapy information which
would be treated distinctly.
in openEHR we try to create small information components which can be
shared throughout and between systems, rather than necessarily
directly mirroring a specific dataset
e.g. for TB we might use
A Problem-diagnosis archetype for the original diagnosis details
A Medication archetype to describe therapy past/current
A Review archetype to handle follow-up
A Microbiology archetype to handle lab results
In use these would be brought together in a single template, which
better reflects the use-case for TB, but we can use all of the same
archetypes for malaria but using a different template.
and will try and find a little time to work out how we might approach
this with openEHR. The main thing to note with this approach is that
it takes longer to do the 'first module' but each subsequent module is
easier, and faster to model.
Ian
Dr Ian McNicoll
office +44 (0)1536 414 994
fax +44 (0)1536 516317
mobile +44 (0)775 209 7859
skype ianmcnicoll
ian.mcnicoll@oceaninformatics.com
Clinical Modelling Consultant, Ocean Informatics, UK
openEHR Clinical Knowledge Editor www.openehr.org/knowledge
Honorary Senior Research Associate, CHIME, UCL
BCS Primary Health Care www.phcsg.org
Have a look at www.openehr.org/knowledge. This is the clinical knowledge manager portal where you can share and download archetypes.
As for the archetypes that you mention, these are at the highest level a diagnosis which is covered by the diagnosis archetype with terminology to define the particular diagnosis. I suspect that you are thinking of all the other things you might want to record about these things. I would suggest that you would probably use a template which would allow you to include all the other things like medications etc by aggravating archetypes.
I remember that Cambodia group developed TB archetype at Kano labo in
Waseda university.
We can share information at our openehr.jp site. http://openehr.jp/news/9
Are the Cambodian TB archetypes visible or downloadable anywhere?
The tension in any archetype development is between modeling the
specific use-case e.g a TB - public health reporting application,
which is relatively easy and the more complex but ultimately
preferable approach of creating interoperable archetypes which make
use of clinical data collected as part of front-line activity and
requirements.
Ian
Dr Ian McNicoll
office +44 (0)1536 414 994
fax +44 (0)1536 516317
mobile +44 (0)775 209 7859
skype ianmcnicoll
ian.mcnicoll@oceaninformatics.com
Clinical Modelling Consultant, Ocean Informatics, UK
openEHR Clinical Knowledge Editor www.openehr.org/knowledge
Honorary Senior Research Associate, CHIME, UCL
BCS Primary Health Care www.phcsg.org
Very many thanks. I must have a nose around. You are right that they
are a number of archetype developed by 'the usual suspects' including
myself that are not yet on CKM. I have quite a number developed for a
paediatrics project which need to go up there. The problem, as ever,
is time. We do not want to put these archetypes , developed for a
particular project, up on to CKM until we have at least established
that they are appropriate for sharing to a wider audience.
We will get there honest.
Ian
Dr Ian McNicoll
office +44 (0)1536 414 994
fax +44 (0)1536 516317
mobile +44 (0)775 209 7859
skype ianmcnicoll
ian.mcnicoll@oceaninformatics.com
Clinical Modelling Consultant, Ocean Informatics, UK
openEHR Clinical Knowledge Editor www.openehr.org/knowledge
Honorary Senior Research Associate, CHIME, UCL
BCS Primary Health Care www.phcsg.org
The archetypes that Diego mentioned as being missing from CKM look to
me very much like an extensive set that were developed as a part of an
English NHS project some years ago. Many of the current CKM archetypes
come from the same source, but have been subsequently re-modelled.
Heather took an descision on which were 'good enough' to go into CKM
without too much extra work but there are others which can certainly
be used.
Ian
Dr Ian McNicoll
office +44 (0)1536 414 994
fax +44 (0)1536 516317
mobile +44 (0)775 209 7859
skype ianmcnicoll
ian.mcnicoll@oceaninformatics.com
Clinical Modelling Consultant, Ocean Informatics, UK
openEHR Clinical Knowledge Editor www.openehr.org/knowledge
Honorary Senior Research Associate, CHIME, UCL
BCS Primary Health Care www.phcsg.org
Clinical knowledge manager portal and its knowledge repository are yet at early stages. Most of efforts are aimed at creating basic building blocks of medical knowledge and the progress needs more contribution to speed up. As Leslie correctly mentioned, you may use Templates. Templates are like more localized, more constrained composition of existing archetypes and your local archetypes (may appear not agreeable by the community, later).
We have very detailed templates for HIV and Tuberculosis but they need rewriting using approved archetypes. If you are interested we may do it jointly.
Which archetypes to put on CKM and now has been an issue in my work as well...Some of you might have gotten sick of hearing this again but I'm dealing with very detailed (and specific) archetypes covering digestive endoscopy. Since they are now around 10 years old (!) and quite mature for our purpose unfortunately have little use for the rest of you. The main usage of CKM today is to share the common denominators for achieving a reasonable level of interoperability at a minimal cost. Don't forget that CKM work is not a funded project and people are doing it voluntarily (well perhaps has some fringe benefits but...).
So to cut it short we previously discussed the idea of having a 'production' space on CKM and also 'specialised' workspaces. Also a 'sandpit' workspace will be very very useful.