Stream: netherlands
Topic: Code does not exist in valueset
Piotr Chabot Stadhouders (Oct 13 2021 at 08:00):
Hi,
I am validating a Practitioner resource with simplifier, scope nictiz.fhir.nl.stu3.zib2017 2.1.6 STU3
simplifier now generates an error for the telecom extension for mobile phone (see below)
Does this make sense?
CodeInvalid : Code 'MC' from system 'http://terminology.hl7.org/CodeSystem/v3-AddressUse' does not exist in valueset 'http://decor.nictiz.nl/fhir/ValueSet/2.16.840.1.113883.2.4.3.11.60.40.2.20.6.1--20171231000000'
Practitioner.telecom[0].extension[0].valueCodeableConcept[0]
Ardon Toonstra (Oct 13 2021 at 11:24):
At first sight it looks like a codesytem uri is used that is introduced since R4(considering the '..terminology.hl7.org...') Are you sure you have indeed the right system?
It should be http://hl7.org/fhir/v3/AddressUse for STU3
Piotr Chabot Stadhouders (Oct 14 2021 at 08:50):
Hi Ardon,
Since a couple of weeks I use NTS (Nationale Terminologie Server) to get the valuesets
When I do a GET on TelecomTypeCodelijst version 2017-12-31T00:00:00, the response has system 'http://terminology.hl7.org/CodeSystem/v3-AddressUse' for code 'MC' and 'PG'
Piotr Chabot Stadhouders (Oct 14 2021 at 09:17):
Another thing I don't get :
When I validate an medicationRequest, it returns a lot of warnings
For example, I generate an 'asNeededCodeableConcept' with value "1144" and system "urn:oid:2.16.840.1.113883.2.4.4.5"
The validator now returns "None of the codings provided are in the value set http://decor.nictiz.nl/fhir/ValueSet/2.16.840.1.113883.2.4.3.11.60.40.2.9.12.4--20171231000000"
However, "1144" is a valid value
Has this maybe something to do with the fact that the mentioned valueset is in XML format while I am generating JSON?
In NTS the Identifier of is ZonodigCriteriumCodelijst is "667592bd-6fd7-4ddc-965a-d0d654a111cd" by the way, and not "2.16.840.1.113883.2.4.3.11.60.40.2.9.12.4--20171231000000"
Pieter Edelman (Oct 14 2021 at 13:24):
AFAIK the Nationale Terminologieserver does only implement FHIR R4, while the zib2017 package is STU3 based. I don't think you can use the ValueSets from the NTS in combination with the zib2017 package.
Pieter Edelman (Oct 14 2021 at 13:29):
For example, I generate an 'asNeededCodeableConcept' with value "1144" and system "urn:oid:2.16.840.1.113883.2.4.4.5"
The validator now returns "None of the codings provided are in the value set http://decor.nictiz.nl/fhir/ValueSet/2.16.840.1.113883.2.4.3.11.60.40.2.9.12.4--20171231000000"
1144 is a valid code, but urn:oid:2.16.840.1.113883.2.4.4.5 is not the corresponding system -- this should be https://referentiemodel.nhg.org/tabellen/nhg-tabel-25-gebruiksvoorschrift#aanvullend-numeriek
Pieter Edelman (Oct 14 2021 at 13:31):
XML vs. JSON shouldn't be a problem, as is the id (not identifier), ValueSets are identified by the canonical URL.
Piotr Chabot Stadhouders (Oct 14 2021 at 13:35):
Hi Pieter,
When I retrieve version 2017-12-31T00:00:00 of the valueSet resource https://referentiemodel.nhg.org/tabellen/nhg-tabel-25-gebruiksvoorschrift#aanvullend-numeriek is indeed the system, so this looks good. Until now we used the OID's as system value, and I thought this should be no problem (in Zibs2017)
Piotr Chabot Stadhouders (Oct 14 2021 at 13:37):
@Pieter Edelman
And what about this one?
'http://terminology.hl7.org/CodeSystem/v3-AddressUse' for code 'MC' and 'PG'
Pieter Edelman (Oct 14 2021 at 14:15):
Until now we used the OID's as system value, and I thought this should be no problem (in Zibs2017)
Well, this is actually a bit of a nasty topic. FHIR requires a canonical URI, which is defined in formally for use in a FHIR context. An OID is not a URI, it is another identification mechanism altogether. You _can_ create OID based URI's, but this is (often) not the formally defined canonical URI for use within FHIR. So, taking a code system OID and prepending it with urn:oid: is usually not the way to go.
Pieter Edelman (Oct 14 2021 at 14:24):
And what about this one?
'http://terminology.hl7.org/CodeSystem/v3-AddressUse' for code 'MC' and 'PG'
This is a bit of a nasty topic as well. "http://terminology.hl7.org" is the URL base adopted in FHIR R4 for terminology, while until STU3 the base "http://hl7.org/fhir/" is used. If you use an R4 CodeSystem or ValueSet resource, it should use this new format, but any STU3-based resources should use the old format. I suspect that you're exporting an R4 resource from the NTS and base your instance on these resources. If you're validating against the STU3-based zib2017 package, it expects the old format: https://simplifier.net/packages/nictiz.fhir.nl.stu3.zib2017/2.1.6/files/406653
Piotr Chabot Stadhouders (Oct 15 2021 at 12:53):
@Pieter Edelman
I export the 2017-12-31T00:00:00 version of TelecomTypeCodelijst, with id 8a4b6a11-973d-4a77-8d84-5efe8f2609c6
This generates "http://terminology.hl7.org/CodeSystem/v3-AddressUse" for system for the values "MC" and "PG"
So, I think it is a "bug" in the valueSet resource
Pieter Edelman (Oct 15 2021 at 13:03):
Yes, the Nationale Terminologieserver only supports FHIR R4, you cannot use this in the context of the STU3 based zib2017 package. All ValueSets are included in the package (in STU3 format).
Last updated: Apr 12 2022 at 19:14 UTC
