SLPG meeting cancelled this week

Thanks Linda

As something to consider prior to the July meeting, I was looking here earlier in the week at recent 'International Edition - SNOMED International Release notes' to see if there are any contemporary examples of 'retrieval-disrupting' model changes that I referred to in comments after an earlier meeting.

I've only done a time-limited review, but I think I've found a couple of examples (one clean, the other a long-running saga):

Traumatic injury variants:

Since approximately July 2020 'traumatic' injuries have been distinguished by the inclusion of a DUE TO = 773760007 |Traumatic event (event)| role in each definition. For the 15 years prior to this traumatic injury variants were identified by ASSOCIATED MORPHOLOGY = << 19130008 |Traumatic abnormality|.

Compositional expressions may have been created using the earlier convention (such as specialising 234506007 |Rupture of spleen (disorder)| to its traumatic variant by ASSOCIATED MORPHOLOGY value refinement using the now inactive 415747007 |Traumatic rupture (morphologic abnormality)| value), and these would no longer be returned by a 'modern' query that specified a DUE TO = 773760007 |Traumatic event (event)| role in its predicate.

Representing causal associations:

Having led the ECE group for some years I am sympathetic to the challenges here, but nonetheless the instability of any standard approach may well have relevance to compositional expression retrieval. I was reminded of it by the section on remodelling rheumatoid and rheumatic conditions here. This discussion is not the place to unpack all the details of disorder combination modelling, but suffice it to say that debates about disorder combination modelling (where there are strong or weak implications of causation and temporal overlap) unavoidably resurface in the context of post-coordinated efforts to achieve the same expressivity. There is a first order question of whether such practice should be encouraged/discouraged, but if it happens at all there are further considerations regarding the merits and risks of how and when conditions are combined (either by simple '+' conjunction or by role-based causal association) in compositional expressions, and how these (once fixed in records) can be retrieved when the reference data standard approach is still changing.

Ed