The FHIR Implementation Guide https://simplifier.net/guide/fhir-genomics-implementation-guide/Home/FHIRAssets/Profiles/All-Profiles/UKCore-ServiceRequest
Contains a definition for reasonCode (Procedure) and reasonReference (Condition). In FHIR R5 these are combined into a CodeableReference called reason.
Should Genomics follow the same convention now and so both these definitions should point to a Condition (maybe reasonCode should be bound to https://simplifier.net/hl7fhirukcorer4/valueset-ukcore-conditioncode)
If we follow HL7 FHIR ServiceRequest mappings to HL7 v2 (ORC-16) https://www.hl7.org/fhir/servicerequest-mappings.html#v2, the element we should populate here is reasonCode which is a coded entry and so it is unlikely we will be able to populate reasonReference.
1 Like
Hi Kevin,
Thanks for raising the issue. I believe the modelling for the Genomics ‘Reason for Testing’ is incorrect and should be moved to ServiceRequest.category. This should reserve the reason[X ] field for the condition being tested for. I will discuss the change internally, and if agreed, the new mapping will be available in the next IG release.
However, the reason field will still likely need to be a reference to a Condition resource rather than a single code, due to the requirement for WGS testing to record additional information against the condition, such as the onset date and verification status.
Could you possibly work around this by creating a Condition resource with only the code present (from the HL7v2 mapping) and further populating the resource through subsequent messages?
Regarding the binding to the UK Core valueset, at the moment, this is being inherited from UK Core, with the valueset being slightly wider than just clinical findings (as the reason for testing could also be indicated through an observation etc.). The binding to UK Core ConditionCode is preferred within the Condition resource however. Please note that the code may not always be a SNOMED concept, particularly for rare disease where the condition being tested for is not yet in SNOMED and only recorded in the Genomics Test Directory.
Best,
Omar
1 Like
Can I come back to you on that? I need to refer a view on that to a wider audience (within the region).
I may have an issue around the lab order practically being an ‘archetype’ (defined as a FHIR Questionnaire) rather than a composition (/Clinical Headings).
The former says Condition code (v2 CWE data type), whilst the latter says is saying Condition.
2 Likes