Stream: argonaut
Topic: ICD-9 and ICD-10 system URIs
Michele Mottini (Mar 31 2020 at 16:20):
The system URIs to use for ICD-9 and ICD-10 codes are http://hl7.org/fhir/sid/icd-9-cm and http://hl7.org/fhir/sid/icd-10-cm for both procedures and conditions, correct?
Michele Mottini (Mar 31 2020 at 16:21):
(having the same URIs for both turns out to complicate things a lot in our code....)
Rob Hausam (Apr 01 2020 at 03:13):
@Michele Mottini It turns out that's not quite the case. It is true that both the diagnoses and procedures are included in ICD-9-CM, but ICD-10-CM does not include procedures - they are coded separately in ICD-10-PCS. From a general (not FHIR) standpoint it's explained pretty well here, I think. From the FHIR perspective, I see two Zulip topic threads that we've had on this here and here.
It seems that we haven't actually finished settling the ICD-10-PCS code system url issue. In the second Zulip thread there seemed to be support for using http://hl7.org/fhir/sid/icd-10-pcs. There is a value set for it in US Core R4 here, but it still lists the code system url as http://www.icd10data.com/icd10pcs. As I mentioned in the Zulip thread, I don't think that's a good (correct) choice, as that is a url for a particular commercial service which is definitely not the code system owner or an officially authoritative source. So do we need to get together (US Core, Vocab WG, HTA) and take another look at this (and hopefully finish it)? @Eric Haas @Brett Marquard
Josh Mandel (Apr 01 2020 at 13:14):
Linking to the ICD-10 data commercial site just seems like an error (I'm guessing that company invests more in search engine optimization than WHO does). ARe there people who think we should not correct this error?
Michele Mottini (Apr 01 2020 at 13:39):
I see, so the current situation is like this:?
Michele Mottini (Apr 01 2020 at 13:41):
(deleted)
Michele Mottini (Apr 01 2020 at 13:41):
Brett Marquard (Apr 01 2020 at 15:10):
The value set @Rob Hausam referenced is not used in any formal profile -- it's just listed on the general US Core terminology page. This still should be fixed. @Michele Mottini do you mind logging a tracker?
Brett Marquard (Apr 01 2020 at 15:10):
@Rob Hausam We will update to the recommendation of Vocab/HTA...whenever there is another release.
Michele Mottini (Apr 01 2020 at 15:15):
Sorry, a tracker to do what? To change https://www.hl7.org/fhir/us/core/ValueSet-us-core-procedure-icd10pcs.html to use a different system URI ? . . .which one . . . ?
Michele Mottini (Apr 01 2020 at 16:33):
@Brett Marquard ^
Brett Marquard (Apr 01 2020 at 16:34):
Tracker can be to update the value set to the HTA/Vocab approved URI or remove the value set. thank you
Michele Mottini (Apr 01 2020 at 17:54):
Done, https://jira.hl7.org/browse/FHIR-26679
Michele Mottini (Apr 01 2020 at 18:00):
Reading those threads @Rob Hausam linked it seems that the 'correct' system URI would be http://hl7.org/fhir/sid/icd-10-pcs, but it is not up yet, correct?
Michele Mottini (Apr 01 2020 at 18:02):
(we are going in production with R4, and ICD codes are kind of important, I'd rather not have to change system URIs.....)
Rob Hausam (Apr 01 2020 at 18:26):
This is a US code system and will be used in a US realm IG. The proposed URL of http://hl7.org/fhir/sid/icd-10-pcs is a FHIR url (which would nominally be universal). An alternative (although still generally "universal") might be http://terminology.hl7.org/icd-10-pcs (or something like that, as per UTG), as I don't see a particular reason that the code system url must be in the "fhir" space. Since it's an external code system probably HTA should make the decision on the url (or at least should provide their input). @Robert McClure? (I would also copy to Ted Klein and Carol Macumber or another current HTA member, but they're not subscribed to this stream)
Robert McClure (Apr 02 2020 at 02:42):
It seems odd that we use the hl7.org base for ICD code system canonical urls but I see that a number of realm variants have allowed that. I honestly don't understand why we don't use the WHO site or something for the international version. As for the US variant, it should be the CMS site I would think.
But this is not for us to decide, it really should be the HTA, as Rob noted. @Carol Macumber and @Julie James and I just subscribed them!
Grahame Grieve (Apr 02 2020 at 03:29):
for any authority that wouldn't engage with us and define a canonical url or a permlink, we defined our own for them
Carol Macumber (Apr 02 2020 at 13:38):
Hi All - I recall the previous effort and the fact that we were unable to get a definitive answer from either organization. I have found the following information for each source from the NLM:
ICD-10-PCS - note the URL listed: https://www.cms.gov/Medicare/Coding/ICD10/index.html
image.png
ICD-10-CM - note the URL listed: https://www.cdc.gov/nchs/icd/icd10cm.htm
image.png
HTA will be reaching out to the contacts, per HTA policy with regards to obtaining codesystem identifiers for use in FHIR.
@Julie James @Robert McClure @Ted Klein
In the meantime, I feel the use of the above URLs would be more authoratative than the FHIR defined ones as they resolve to (what appears to be) stable pages from the source. However, i recognize that many people have used the existing URIs...
Michele Mottini (Apr 02 2020 at 14:27):
Cerner is already using the 'bad' http://www.icd10data.com/icd10pcs, so if there is no quick decision about an alternative that will become the de-facto standard
Rob Hausam (Apr 02 2020 at 14:43):
I really do think we need to get rid of that "bad" one. How soon is "quick" in this context?
Michele Mottini (Apr 02 2020 at 15:00):
This month? Maybe this quarter? No later than that I'd say
Vassil Peytchev (Apr 02 2020 at 15:03):
Note that the page https://www.cms.gov/Medicare/Coding/ICD10/index.html lists diffferent versions of the US ICD codes, e.g.
https://www.cms.gov/Medicare/Coding/ICD10/2019-ICD-10-PCS
https://www.cms.gov/Medicare/Coding/ICD10/2020-ICD-10-PCS
Carol Macumber (Apr 04 2020 at 14:39):
Yes, it's a URI for the ICD-10-PCS code system, not a particular version. Similar to http://snomed.info/sct as a URI for SNOMED as a codesystem and https://loinc.org/ for LOiNC.
Carol Macumber (Apr 04 2020 at 14:42):
@Robert McClure @Rob Hausam @Ted Klein @Reuben Daniels @Carmela Couderc @Heather Grain
We need to discuss this on the next Vocab call. Hopefully HTA will have heard back from the contacts. However, if they haven't, the process has been defined to then have HTA (in collaboration with Vocab) to set the URIs.
Rob Hausam (Apr 04 2020 at 18:09):
Sounds reasonable, @Michele Mottini. @Carol Macumber, I'm not sure exactly what your comment on "not a particular version" was in response to? The links @Vassil Peytchev posted are to the FY2019 and FY2020 versions (but I expect aren't the urls/uris that we'll use). Agree about discussing this on the next Vocab call (and can join HTA calls, if needed).
Carol Macumber (Apr 04 2020 at 20:10):
Yup, that's what I was replying to...the different version URLs would be inappropriate for the Codesystem URI.
Rob Hausam (Apr 05 2020 at 12:16):
Yes, agree - those urls couldn't be the code system uris.
Carol Macumber (Apr 16 2020 at 16:09):
Update - Pending a few questions posed to @Ted Klein, to be clarified with @Grahame Grieve on the validity of some proposed URIs, HTA will be resolving with the owners of ICD-10-CM and ICD-10-PCS
Michele Mottini (Apr 16 2020 at 19:52):
Thanks for the update @Carol Macumber
Carol Macumber (Apr 29 2020 at 19:33):
Hi All - progress! slow and steady, but progress none the less. We have worked with CMS with regards to ICD10PCS. The codesystem URL will be http://www.cms.gov/Medicare/Coding/ICD10, this meets the requirements noted by Vocab WG for URL syntax (located here https://confluence.hl7.org/display/VOC/Proposed+Vocabulary+WG+Policy+on+Creation+of+Canonical+URLs+for+UTG+Code+Systems) AND has been blessed by the authority/CMS. @Reuben Daniels will be working on creating the content for a Using ICD10PCS with FHIR page based on the information we recieved from CMS (including code representation and versioning).
I am still working with CDC/NCHS on the official URL for ICD-10-CM. Stay tuned there.
Michele Mottini (Apr 29 2020 at 20:32):
Great, thanks @Carol Macumber , so for procedures we can go ahead and use http://www.cms.gov/Medicare/Coding/ICD10, correct?
Carol Macumber (Apr 29 2020 at 20:36):
Yes, absolutely…with the caveat that CMS is doing some validation on their end to update the content on the endpoint (e.g., some of the information is outdated, validating all the links are active etc).
Robert McClure (Apr 30 2020 at 04:16):
@Carol Macumber Seriously? ICD-10-PCS will have the canonical url that ends in ..."ICD10"?
Brett Marquard (May 01 2020 at 01:20):
@Robert McClure are you helping chase the solution?
Robert McClure (May 01 2020 at 17:28):
Yes, but Carol is on this. I'm sensing final decisions are not in our hands.
Carol Macumber (May 05 2020 at 14:03):
Robert McClure said:
Yes, but Carol is on this. I'm sensing final decisions are not in our hands.
Correct. The URL is from the source (CMS), is valid (per Vocab guidelines) and as far as HTA is concerned, final.
Rob Hausam (May 06 2020 at 03:32):
@Carol Macumber The page that uri (as url) resolves to includes links for both ICD-10-CM and ICD-10-PCS (plus also GEMS). So if this is the CMS uri for ICD-10-PCS, what is their uri for ICD-10-CM?
Carol Macumber (May 06 2020 at 03:41):
@Rob Hausam
ICD10CM is maintained by NCHS (CDC).
CMS maintains ICD10PCS.
CMS only republished ICD10CM and designates billable codes for reimbursement.
Rob Hausam (May 06 2020 at 03:43):
Yes, that is true. So is the uri for ICD-10-CM still being determined?
Grahame Grieve (May 06 2020 at 04:27):
no - we published that long ago: http://hl7.org/fhir/sid/icd-10-cm
Rob Hausam (May 06 2020 at 15:02):
Yes, I know that we did.
Last updated: Apr 12 2022 at 19:14 UTC
