search

HL7 / fhir-shc-vaccination-ig

FHIR Implementation Guide describing clinical and patient data contained within a SMART Health Card
https://vci.org/ig/vaccination-and-testing
35 stars 18 forks source link

Canadian and UK SNOMED CT Vaccine Codes #163

Closed igorsirkovich closed 3 years ago

igorsirkovich commented 3 years ago

Please consider adding ValueSets with the COVID-19 Vaccine Codes from the Canadian and UK SNOMED CT Editions:

https://browser.ihtsdotools.org/?perspective=full&conceptId1=28531000087107&edition=MAIN/SNOMEDCT-CA/2021-03-31&release=&languages=en,fr

https://termbrowser.nhs.uk/?perspective=full&conceptId1=39330711000001103&edition=uk-edition&release=v20210512&server=https://termbrowser.nhs.uk/sct-browser-api/snomed&langRefset=999001261000000100,999000691000001104

masnick commented 3 years ago

It looks like both the Canada- and UK-specific SNOMED codes you linked to above are descendants of 787859002, which is what the value set is currently based off of.

I believe that using the locale-specific SNOMED codes you link to would therefore conform, but you may get a false-positive validation error if you use a terminology server that isn't loading from the locale-specific SNOMED edition.

@igorsirkovich is it sufficient to provide some written guidance related to this, or is there a specific validation issue you're trying to resolve?

igorsirkovich commented 3 years ago

@masnick, I would just like to ensure that using these locale-specific SNOMED codes is conformant with the Smart Health Cards IG and passes validation. Unfortunately, at this stage these codes fail validation. If I search for all the descendants of 787859002 in the international edition of SNOMED CT, these codes don't get returned. I'm wondering whether it would be possible to define a new vaccineCode.coding:snomed-canada slice in the Immunization profile which would be bound to descendants of 787859002 in the Canadian edition of SNOMED CT (version http://snomed.info/sct/20611000087101/version/20210331).

masnick commented 3 years ago

@igorsirkovich I'm actually not sure what the best practice is here. I'm asking in the terminology stream on Zulip: https://chat.fhir.org/#narrow/stream/179202-terminology/topic/Locale-specific.20SNOMEDs.20and.20validation

masnick commented 3 years ago

Following the Zulip thread posted above, it appears this is fundamentally an issue with the terminology server used by the FHIR validator (by default tx.fhir.org).

In theory, I believe that if one had a terminology server that responded with the CA edition of SNOMED when the validator requested expansion of the value set (asking for descendants of 787859002), then a resource using 28531000087107 would validate.

It doesn't seem like there's anything we can do inside the IG to force this to happen, so unfortunately I don't have a satisfying answer of how to successfully validate. Pragmatically you can test by filling in a SNOMED code that's part of the int'l edition that tx.fhir.org uses, and if that resource validates then you can safely assume that replacing that code with a national edition code will yield a conforming resource.

igorsirkovich commented 3 years ago

@masnick, thank you for the Zulip link. It's a great discussion indeed. Hopefully, update of tx.fhir.org with the codes from the Canadian SNOMED CT edition would do the trick.

masnick commented 3 years ago

@igorsirkovich We were able to get the Canadian edition of SNOMED explicitly included in the STU1 ballot version of the IG: http://hl7.org/fhir/uv/shc-vaccination/2021Sep/ValueSet-vaccine-snomed.html

We are limited to the SNOMED editions that tx.fhir.org supports, which does not currently include UK SNOMED. But now that we have a working method for including specific SNOMED editions, we will continue to update the IG whenever possible with the full suite of SNOMED editions.