Hello,
When looking at the published archetypes in the international CKM I see a lot of archetype members that do not have a terminology binding, but which could possibly have one. The "Body Surface Area" for example has a member "Body Surface Area" which has no bound code but could have the Snomed Code 128178001. Or "Blood Pressure" has terminology bindings for its members "Systolic" and "Diastolic", but not for the "Mean arterial pressure". Is there a reason for those non existing bindings or is it just because the bindings have not yet been needed and so nobody added them yet ?
Greetings
Georg
Hi Georg,
It is partly a matter of resources and partly a question of it actually being very hard to do, as you are generally faced with a number of choices and omissions in SNOMED CT - that’s not a criticism, just a reality.
It is also a question of ‘where’s the value’? Querying for element names within an openEHR CDR does not generally rely on external reference terminologies like SNOMED so there has not been much drive to do this work from vendors, particularly as actual SNOMED CT adoption remains relatively low.
Having said that, any input would be very welcome, and CKM is well set up to support the collaborative review of suggested bindings. Thee may be a bit of a push from the UK, which is intending to use SNOMED CT to drive FHIR Observation coding / profiling rather than LOINC, but the process does need folks with some experience and authority in terms of dtermining the correct bindings.
Ian
Hello Georg,
Defining bindings has been a topic of discussion for many years. I would say that term bindings (bindings to the labels of the nodes in the archetype) are more difficult than the constraint bindings (bindings to a subset of valid terms). This is mostly because when creating “label” bindings most of the time the meaning of a given item in snomed (e.g. blood pressure) is not exactly the same as “measure of a blood pressure”. Some countries such as Spain have gone the other way around by defining Snomed codes in the extension with the meaning provided by the archetype element itself. Other standards bodies use Loinc with this same principle.
Also, using wrong codes is quite easy if you are not aware of the Snomed hierarchy (specimen vs procedure, etc.)
Regards
Thanks Diego,
I think you expressed what I was trying to say much more clearly!!
I would also add that opinions about what is the ‘right’ code can differ substantially between the terminology experts!!
Ian