Terminology bindings in openEHR

Technical (archive)
1

Hi,

Can I get some insight into how terminology bindings work in openEHR, or how they are supposed to work?

It appears openEHR supports multiple terminologies, in addition to its own internal coding structure (atxxxx). I think there’s a whole terminology service dedicated to this. But my confusion is how to implement the feature.

In the Ocean Template designer as well as the Archetype editor, there are features that allow us specify what terminology we want to use and to ‘manually’ input the codes we want to represent. Some data elements already have fixed ‘Valuesets’ (e.g. Regular, Irregular as Coded_text values for the ‘Regular?’ element in the Pulse archetype).

When I use free online tools to create a FHIR resource for upload to a FHIR repository, where ‘CodeableConcepts’ exist the form field has a code/database lookup service to LOINC, SNOMED CT, internal FHIR codes etc., so that as I enter the first three characters of the concept I want the system presents a list of options which I can override if I don’t find what I want. When I select the concept I want, the system inputs the correct code as well as the URL of the source terminology server.

I want to imagine openEHR supports this sort of ‘lookup’ to existing terminologies, but how do I accomplish this? Otherwise how does it handle terminology binding?

Thanks.

2

Hello Eneimi

The Data Value entities hold information on which terminology a specific encoded term may be coming from to enforce further data validation.

What you describe in your question, involving FHIR, is what happens when you interact with a terminology server (https://en.wikipedia.org/wiki/Terminology_server).

openEHR itself does not specify the terminology server part (yet), which is what you would use if you wanted to lookup the codes that will then makeup your values.

The binding can be handled in a number of ways. You could create your own simple lookup service based on the raw datasets of ICD, READ, SNOMED, LOINC and others as it was done in OSHIP (http://bazaar.launchpad.net/~higorpinto/oship/cr-ehr/view/head:/oship/src/oship/rxterms/createrxterms.py), use LexBIG (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2732233/) as it was done in opereffa (http://www.openehr.org/pt/news_events/releases.php?id=33) or CTS (http://clinfowiki.org/wiki/index.php/Common_Terminology_Services) as in the LinkEHR editor (http://linkehr.com/).

Hope this helps.

All the best
Athanasios Anastasiou
Lecturer in Health Data Science | Darlithydd mewn Gwyddoniaeth Data Iechyd

3

Some background information from the ADL2 spec may help. This description explains the same semantics as for ADL 1.4. archetypes, except that in ADL 1.4, ‘at’ and ‘ac’ codes are used, whereas ADL2 has ‘id’, ‘at’ and ‘ac’ codes.

With respect to tools, there is not yet a standard openEHR terminology look-up tool, and the direction is most likely that openEHR modelling tools will integrate with other available tools such as the ones you mention.

  • thomas
4

Thanks, that clears things up a bit.