Recently we discussed terminology bindings. We probably still have not got them right, but we don’t have a model of what we think they should be. I posted a quick idea of a possible more structured version:
I noted that the right hand side of a binding can be a few different things, each of which would be accompanied by various meta-data, including:
a single concept code
a single code or other id referring to an external value set in an external terminology (in SNOMED it is a SNOMED code; for e.g. ICD10, there is no standard that I know of)
a composition expression that refers to a more refined concept
possible a constraint expression that locally determines a value set intensionally, to be resolved by application to the Terminology service.
I’d rather avoid the last, because of the brittleness of intensional ref-set query syntax expressions. In any case, we need a better idea of what meta-data are needed. E.g.:
something to do with (min) version of terminology required for the reference to be valid
something to do with purpose?
other notes - a tagged list of basic types?
I would like to get a better idea of the requirements.
When you add the descriptions in SNOMED, language of the SNOMED-database would be important, version is already there, I would say "version" instead of min_version, it makes it more generic usable.
I think min_version can be problematic as certain terms can be deprecated in future versions and then this naming could be misleading. That said for SNOMED it’ll still be present in future releases just marked as inactive. For other terminologies this cannot be guaranteed. BTW SNOMED uses term Effective Time
Now that I have more experience with SNOMED expressions, I like the idea of doing the binding with an expression, also I think an expression includes the single code binding, if that is correct there is no need of defining a different notation for single code binding, just use a simple expression formed by one specific concept code. Also the expression being something processable and very versatile, we can express complex concepts with a few codes, which will help on adding knowledge to the archetype and serve to a better and simpler CDS.
About the metadata, there should be expressed against which SNOMED release this expression was created. We can’t be sure only with min version. I should be responsibility of the user to check if the expression works on a different version/release of SNOMED. Another metadata is if the version is a local extension, some countries have their own extensions.
I don’t know if we need to support other terminologies (technically) and if doing that is useful (strategically). Terminology services can do SNOMED to ICD, and ICD is not clinical relevant. LOINC is useful, but there is a SNOMED-LOINC collaboration, so we might expect an official mapping in the future (https://loinc.org/collaboration/snomed-international/). IMO we should focus on SNOMED.
Yes, it is correct that expressions include single code binding. Those kinds of bindings are just the simplest variants of expressions.
I think that in a few years’ time nearly all implementations of SNOMED CT not only implement the international version, but also one are a few international, national or local extensions, so this use case is probably the normal use case and not the exceptional use case.
Regards
Mikael
(Among other things SNOMED CT Implementation Advisor)
I do that too. It seems like more and more people are moving away from the position that SNOMED CT is complex and expensive to a position that SNOMED CT is manageable and an affordable way of getting rid of local terminologies and add value.
CIMI made the decision to use LOINC for the ‘question’ part of the statement.
And SNOMED for the ‘answer’ part.
Leading to: Question = Answer, or something coded in LOINC is something coded in SNOMED.
Nodes in an archetype coded in LOINC and data coded in SNOMED.
are you able to provide more information on the reasoning that led to this decision? Maybe links to documents or any other insights? This would be quite interesting for our acitivities in Germany.
It depends on what you mean by scope … There is an agreement between SNOMED International and Regenstrief Institute to not batch include concepts that directly match LOINC codes in SNOMED CT. However, in general SNOMED CT can have concepts that represent the same or similar ideas as LOINC codes.
Without knowing the internal decision process, I would say that has something to do with the fact that HL7 guys in CIMI already used Loinc for describing document and section codes in CDA (not being uncommon that they just define new Loinc codes to new kinds of documents and sections).
Same can be achieved with Snomed BTW, Spanish Snomed Extension includes codes for the Compositions, Sections and Entries defined in the national archetypes
FHIR also supports the expression language in the URL with, for example, http://snomed.info/sct?fhir_vs=ecl/<<123464:474748=<<84848484
But note that these URIs (the above and your isa/ one below) are defined by HL7 FHIR, not SNOMED International. Technically they identify FHIR ValueSets that expand to the set of codes you want.
You could do a lot worse than adopting the FHIR ValueSet mechanism for binding. There are some excellent terminology servers out there (full disclosure, one of them, Ontoserver, is mine).
There is currently some kind of interesting momentum against Snomed.
It can come from governments that refuse to pay for it (current mood in
France), of from practitioners who, after having been asked by their gov
to "sort out their Snomed subset" came to the conclusion that it doesn't
exist.
<Troll>Some also predict that the most certain result of keeping up
trying to build systems using such shitty fully endemic components is to
have medical doctors disappear from missing the "information society"
turn.</Troll>
Have some of you been aware of the Meriterm (European) project?
Maybe SNOMED International’s document ”Guidance on use of SNOMED CT and LOINC together”, http://snomed.org/snomedloinc , could be of interest to some of you?
Have anybody ever heard about a health-it-project that hadn't a smaller or larger group of sceptic people that try to build momentum against the project?
For SNOMED CT the trend at least seems to be that fewer and fewer people belongs to the sceptic group, and about half of the European Union member countries seems to be member in SNOMED International (https://www.snomed.org/members).
Will France as usual be the last country that adopt something that originate from Great Britain?
LOINC defines a way to asks clinical questions which coded answers may be represented by SNOMED-CT. LOINC has the worldwide integration and SNOMED-CT has the detailed semantics, and is the leading global clinical terminology. So the both are well matched partners, and as a consequence the owners of both coding systems, IHTSDO and the Regenstrief Institute agreed to a plan to integrate both coding systems to one coding system. This plan defines a cooperative work, which started in 2014, and according to this plan, all LOINC-concepts will have SNOMED-CT concepts somewhere in 2018.
