SNOMED requirements for clinical apps and SNOMED NZ roadmap for 2021

[event start=“2020-12-07 22:00” status=“standalone” end=“2020-12-07 23:00”]
[/event]
Thanks everyone for attending today’s session - here are slides for the topics I presented

SNOMED NZ Open Forum Dec 2020.pptx (2.6 MB)


Link: https://www.yammer.com/healthinformationstandards/messages/979977266618368

Thanks for sharing, Alastair. Apologies for not making today’s call, but I had a scheduling conflict.

Apologies i couldn’t make this. Are there more presentations lined up?

a couple of questions on slide 31:

Assume this is the core patient dataset?

  • Problems / Conditions are difficult to nail down: Manageable refset(s) would be great. SNOMED Disorders will not cover everything though (e.g. Symptoms prior to established diagnosis)
  • What about miscellaneous warnings such as those currently in national MWS? e.g. child protection, etc.
  • Medication Allergies/ADRs - we need a reliable set of drug allergy/ADR groups, mapped to NZULM concepts, to enable ADR alerting
  • Further issue with Problem disorder refsets is helping clinicians and applications make the right choices as to when they need to do coordinated expressions e.g. laterality (L/R) applies to some disorders (which eye, knee, etc), but not others

Author: Peter Jordan


Good points raised by Jens Andreas

  • There will always be this “Phrase Book v Dictionary” trade-off when using Reference Sets - particularly fixed (extensional) lists, as opposed to query-based (intensional) sets.
  • Miscellaneous warnings is an interesting topic - someone asked me recently about “access to firearms”; that’s certainly worthy of a warning, but does that fit into a medical ontology such as the SNOMED CT Concept Model?
  • Ideally, allergies and intolerances, would be linked to the SCT Substance Hierarchy, rather than NZULM concepts- mapping the NZULM Substance Table to SCT would be a good move in that direction.
  • Post-coordination is currently being discussed in some depth by the SNOMED International Languages Group. The underlying complexity is vast (MRCM, OWL, Role Grouping, Classifiers etc) - even for lateralizing body parts - and it’s very hard to see anyone getting this right without tooling.