Hello @jcase and @fkaag. Now that the dust has somewhat settled on our migration over to these forum pages, I thought we might be able to pick up our discussion on the potential for further alignment of the SNOMED CT and NUVA concept models.
As I see it, we have two lines of enquiry here (any others?):
- That NUVA specifies the disease that each vaccine is intended to protect against.
- That NUVA uses valence rather than ‘has ingredient’, and that this is the gap that prevents NUVA concepts from subsuming under SNOMED parents.
NUVA specifies the Disease
A blunt first step we could take here would be to add an attribute into the NUVA extension which links the vaccine to the disease, but that doesn’t enrich the International Edition at all, or improve subsumption. I think the problem on our side is that the disease we’re targeting cannot be considered a defining property of the vaccine. I wonder if it could be a candidate for expression as an “additional” relationship, that would not participate in classification.
We seem to be missing a link in SNOMED CT in that I can see the pathogen for a disorder, but there’s no link to the antigen (in terms of the model) that would let me arrive at the vaccine.
no ability to navigate to…
Valence vs Has Ingredient
Is it fair to say, Jim, that there is no short or medium term future in which the International Edition adopts the concept of Valence?
I wondered if - given that we know the target disease - there would be a way to augment the valence with a calculated active ingredient, and then define a role chain so that the active ingredient on the valence behaved as if it were directly on the vaccine. Might need to ask @ygao’s advice there.
Another approach that I’m thinking of here is : what if - in the NUVA extension only - we were to add a secondary axiom to International concepts which introduced the Valence, leading to the NUVA concept subsuming correctly. We could use the NUVA map to SNOMED vaccines in order to identify where these additional axioms should be added.