Stream: snomed
Topic: SRT
Grahame Grieve (Dec 13 2017 at 02:33):
Moving the SRT discussion to here.
Grahame Grieve (Dec 13 2017 at 02:34):
summary of previous discussion:
Grahame Grieve (Dec 13 2017 at 02:37):
- At present, http://build.fhir.org/snomedct.html says "SNOMED Legacy codes " are valid identifiers in the system "http://snomed.info/sct"
- these are widely used for legacy data, particularly around dicom, and the policy change that these are not be used going forward makes no difference here
- there's been a proposal that we should use http://snomed.info/srt instead of http://snomed.info/sct - this will make it easier to control use of the SRT codes rather than SCT codes
- there's some lack of clarity about what codes are valid, and it seems that DICOM are using codes that aren't? I have asked @David Clunie for comment on this
- the discussion really hasn't looked like settling on either status quo or changing to http://snomed.info/srt
Grahame Grieve (Dec 13 2017 at 23:34):
so having only partially successfully moved the discussion here.....
Grahame Grieve (Dec 13 2017 at 23:34):
does SI have an opinion on this subject? This is a kind of interesting space, in that FHIR publishes it, but it's as much under the authority of the SI technical committee as anywhere else
Dion McMurtrie (Dec 15 2017 at 01:30):
SI has a different governance model now, and there's no longer at technical committee with authority. Instead there are a series of advisory groups that are (as the name suggests) advisory and have no authority in themselves. They advise management team members from SI who do have authority.
Anyway, I think the most appropriate group to get a response out of from the SNOMED community is at https://confluence.ihtsdotools.org/display/FHIR/SNOMED+on+FHIR which is an interest group formed by the Members of SI to further collaboration, but is open to all. I think that group would form a recommended position for SI to consider.
A work item for this group has been created at https://confluence.ihtsdotools.org/display/FHIR/URIs+antecedent+SNOMED+versions - please feel free to comment!
As it happens this was discussed at the meeting this group had on Wednesday morning, and while a formal position was arrived at I think it is fair to say that there was support for defining the URI http://snomed.info/srt for this purpose. The main concern of the SI staff on the call seemed to be creation of new records using this code system which is deprecated and should not be used (as you note in your comment above). I don't think there was dissent on that being the right thing, but there are for example existing imaging machines that still do this despite DICOMs policy and changes, which is essentially impossible to control.
At any rate there was support for defining something (probably http://snomed.info/srt) because even if there weren't new records being created with these codes (which is bad) there is a body of existing legacy data which needs to be represented somehow.
Grahame Grieve (Dec 15 2017 at 01:35):
"The main concern of the SI staff on the call seemed to be creation of new records using this code system"... but no one is proposing that?
Dion McMurtrie (Dec 15 2017 at 02:10):
No, I think there is just concern that defining and owning URIs for the codes it would in some way encourage undesired behaviour. But that was just a concern raise, it is by no means a no or a blocker. I think we just need to make a concrete proposal and case, which is for supporting legacy data.
The concern was using "snomed.info" in the URI, but personally I think that's appropriate.
Grahame Grieve (Dec 15 2017 at 02:11):
what do we need to do to make the proposal more concrete then?
Dion McMurtrie (Dec 15 2017 at 03:34):
I think just specify it at https://confluence.ihtsdotools.org/display/FHIR/URIs+antecedent+SNOMED+versions stating the intended use, ownership and where it will be specified. Then get feedback/comments and ask SI to approve it
Dion McMurtrie (Dec 15 2017 at 03:34):
that's just my opinion though
Grahame Grieve (Dec 15 2017 at 03:35):
@Linda Bird - thoughts?
Michael Lawley (Feb 07 2018 at 23:49):
@Linda Bird / SNOMED International's response is here https://confluence.ihtsdotools.org/display/FHIR/URIs+antecedent+SNOMED+versions and, if I understand correctly, raises a number of issues.
Specifically:
1. What should be the response to GET [base]/CodeSystem/$lookup?system=http://snomed.info/sct&code=DD-1302A
2. What should be the response to GET [base]/CodeSystem/$validate-code?url=http://snomed.info/sct&code=DD-1302A
3. What should be the response to GET [base]/ValueSet/$validate-code?system=http://snomed.info/sct&code=DD-1302A&url=http://snomed.info/sct?fhir_vs
4. Should DD-1302A be in the expansion of GET [base]/ValueSet/$expand?system=http://snomed.info/sct?fhir_vs (i.e. "All SNOMED concepts")
5. What should be the response to GET [base]/CodeSystem/$subsumes?system=http://snomed.info/sct&codeA=DD-1302A&codeB=447395005 where 447395005 |Closed fracture of fibula| is the corresponding code to DD-1302A as per the last Identifier Reference Set der2_sRefset_SimpleMapFull_INT_20170731.txt
Rob Hausam (Feb 08 2018 at 18:07):
I added a response in the Confluence thread. We can continue discussion here, if needed.
Michael Lawley (Feb 09 2018 at 04:59):
Thanks Rob, I've followed up there, but wanted to post here to try to connect up the different communities
Rob Hausam (Feb 09 2018 at 21:41):
Yes, that's why we have this stream - thanks.
David Flade (Feb 07 2022 at 08:09):
Hello all,
I would like to follow up on this chat and wanted to ask if - in the meantime - there is any new light on the topic of which System URL to use for (alpha-numeric) SNOMED RT codes in FHIR? I guess they don't fall in the category of SNOMED RT codes that can be treated as SNOMED CT codes (something like: SNOMED RT code for "Carotid Artery": "T-45010", corresponding SNOMED CT code: "69105007").
Would using a proprietary system URL for the SNOMED RT codes make sense which are no valid SCTIDs? Or are you aware of any non-SNOMED managed proprietary system URL for non-valid-SCTID-SNOMED RT codes (such as "T-45010") that one can use?
Thanks a lot and best wishes,
David
Rob Hausam (Feb 07 2022 at 15:53):
@David Flade If you haven't looked at it recently, I think it's probably worth reading or re-reading the discussion in SNOMED CT Confluence on this. One point that Linda Bird makes in her response, is that the legacy alphanumeric SNOMED identifiers are not "SNOMED RT" codes/identifiers, but are actually "SNOMED 3.x" identifiers. I think we need to try to keep that straight and use those labels as consistently as possible, in order to avoid confusion. I'm not thinking of any other discussion that has occurred on this since 2018. I assume that probably something may be needed to handle this in FHIR, for the mapping of legacy SNOMED 3.x coded data. But I am curious what anyone who actually has a need for that has been doing with it over the past 4 years, when we haven't really yet landed on a specific agreed solution.
David Flade (Feb 07 2022 at 16:16):
@Rob Hausam Thanks for clarification! I read the article and I interpreted the term "SNOMED 3.x ID" as "old generation" alpha-numeric SNOMED RT codes, but I won't call it as such anymore.
Then what is listed as "SNOMED RT ID" for instance here in DICOM Part 16 (Table CID 1002. Anatomical Reference Basis - Head):
https://dicom.nema.org/medical/dicom/current/output/chtml/part16/sect_CID_1002.html
is in fact a "SNOMED 3.x ID".
Rob Hausam (Feb 07 2022 at 16:32):
Yes, that's what I would say.
David Flade (Feb 07 2022 at 17:13):
@Rob Hausam Thanks for the clarification.
Elliot Silver (Feb 07 2022 at 18:35):
(@David Clunie do we need to clean up the phrasing used in DICOM?)
David Clunie (Feb 07 2022 at 19:15):
No, I don't think we need to clean up the phrasing used in DICOM wrt. SRT. The phrase "SNOMED RT identifier" is pervasive. Search for it (quoted as a single phrase) on the Internet.
Grahame Grieve (Feb 07 2022 at 19:44):
@David Clunie the FHIR value sets use the SCT identifier - which is correct. However they don't include the SNOMED 3.x ID anywhere. This could be a different designation, or it could be provided in a concept map, but it seems to me that it would be useful for implementers since many of them use the old style identifiers in their implementations
David Clunie (Feb 07 2022 at 20:32):
https://dicom.nema.org/medical/dicom/current/output/chtml/part16/chapter_O.html
Grahame Grieve (Feb 07 2022 at 20:35):
Thanks @David Clunie , that's a lovely table and exactly what I had in mind but it's not published as a FHIR concept map. Perhaps you could consider doing that? I'm happy to help if necessary
Last updated: Apr 12 2022 at 19:14 UTC
