Stream: terminology
Topic: post coordination transformations
Michael Lawley (Oct 06 2016 at 05:22):
[Moving to new thread]
This example suggests that you're not applying the close to use form transformations
http://fhir3.healthintersections.com.au/open/CodeSystem/$lookup?system=http://snomed.info/sct&code=125605004%3A272741003%3D7771000
If you were, then the laterality refinement would be shifted into the normalised expression and onto bone structure
Grahame Grieve (Oct 07 2016 at 09:27):
should I be doing that?
Michael Lawley (Oct 07 2016 at 11:37):
From the Technical Implementation Guide: "If an attribute representing a value for 272741003 | laterality | is present in the refinement and is applied to a focus concept that is not subsumed by 123037004 | body structure |, the laterality attribute should be applied to any and every lateralisable | body structure | specified in the resulting refinement."
- see http://www.snomed.org/tig?t=nfg_transform_normalizeExpression_mergeRefine_laterality
Futher example here: http://www.snomed.org/tig?t=nfg_normalForm_expression_laterality
However, I will note that this section of the guide is categorised as "draft" and "guidance" so it is not "current" nor "standard", but it's been around since I started looking at SNOMED (~10 years now) and it is not uncommon in practice to find expressions with laterality applied to non body structures.
Grahame Grieve (Oct 07 2016 at 11:44):
what's a lateralisable body structure? And other than that, this is just a special rule?
Michael Lawley (Oct 07 2016 at 11:56):
"what's a lateralisable body structure" - very very good question. It's not documented anywhere. At one point I took it to mean any body structure (and its descendants) that appeared with a laterality relationship in the distribution, which is at least minimally correct. Now I just allow and body structure because otherwise all you can do is reject the expression as malformed and that's not usually useful. There are some cases where a concept is modelled with two sites where only one is lateralisable and thus to apply to to both is technically wrong, but again, without knowledge from the terminology of what is lateralisable there's not a lot of options.
Yes, it's just a special rule. There are others, also, earlier in that first section about merging role groups. They have the same "draft guidance" rider and are probably safer to skip. However, it's still worthwhile knowing about them.\
It's an area I've tried to get clarified but there are so few people with knowledge of this detail and (until now) tools that are impacted, that I've gotten very little traction.
Last updated: Apr 12 2022 at 19:14 UTC
