FHIR-29563Getting issue details...
- Vocab to develop a response to this JIRA - understanding that this ticket represents an issue that is not specific to US Core, however it has been brought up in the context of US Core.
- A concept in a value set scopes meaning. When you provide a VS in a binding, you are intending to send codes that map to this concept's meaning. You may send a code from the VS, and another code.
- How does the binding strength impact this?
- VS are allowed to have concepts in them that are scoped by other concepts (procedure, surgical procedure, )
- It is implied that the most specific/accurate concept should be "sent" (rather than a broad, general concept) - the most specific concept code is conformant
- Rob M and Ted: we are conflating what is legal and what people should do as best practice
- Reuben: LM's issue goes beyond legal and stupid. LM's point is that for an extensible binding as referenced in the ticket (e.g. scoped as a procedure), it would be illegal to put a concept from any other code system in the element
- From the ticket: "The problem is that, value or not, with an extensible binding, they MUST send a SNOMED code if one applies - and one will always apply. "
- Reuben: disagrees
- RM: disagrees with RD
- Ted: policy from the Conformance WG is needed to declare that the most specific concept should be selected/sent. If the specific concept is not in the value set expansion, then the closest, more general code should be sent.
- Reuben: The definition of extensible is problematic (from the spec)
- To be conformant, the concept in this element SHALL be from the specified value set if any of the codes within the value set can apply to the concept being communicated. If the value set does not cover the concept (based on human review), alternate codings (or, data type allowing, text) may be included instead.
- Problematic phrases
- When there is a VS where the CLD results in an expansion with broad concepts, (e.g. all procedures), there is always a code that applies.
- Ted: there is an example where for example, there are procedures that will never be in SNOMED CT, but implementations are using them. Their implementation is sending the general SNOMED CT code for Procedure, and then the specific code from the other code system as another coding. This qualifies as a translation (kind of).
https://www.hl7.org/fhir/terminologies.html#extensible (Look here for examples)
Definition of conformant?
- The value in a coded element is machine testable - verifiable?
- Human conformance testing is valid
- RM feels LM would contend that conformance always defaults to the human based environment - cannot always rely on machine testable conformance testing
Some feel the high level concepts in the VS definitions, should be removed (e.g. Procedure, Surgical Procedure) to avoid this crazy situation.
Others do not.
Rob H: the FHIR terminology course teaches the LM view. For the folks who have attended the class, they may have this understanding.
Reuben: Many implementations are not aligned with the LM view. Representing a procedure correctly is important.
Clearly, the definition of extensible is not sufficient. It should be revisited and clarified.
Reminder: in FHIR, when multiple codes are provided, they are not ordered. There is no primary code. How would a receiver know which is the more general code, and which is the specific?
An IG author may
- Define the VS to exclude/not include the broad concepts and use extensible binding strength
An implementer may decide
- If the concept is not represented in the value set, send the closest more general code and the code / code system that accurately reflects the concept to be communicated and use extensible binding strength
Scribe lost track of where the required binding strength options comes in.
RD: ask the community if a task force is desired (Conformance, FHIR-I, Vocab, M&M) to improve and clarify the definition of extensible.
Some in the meeting feel the existing definition is clear, others don't. At the next co-chair call, we will craft a more precise definition of extensible.
To be continued at next co-chair call.