FHIR Chat · tx.fhir.org/r4 support for NCIT · terminology

Stream: terminology

Topic: tx.fhir.org/r4 support for NCIT

May Terry (Apr 20 2021 at 13:59):

I am currently receiving an error validating my FHIR example: "None of the codings provided are in the value set http://hl7.org/fhir/us/shrae/ValueSet/adverse-event-seriousness-value-set". I have confirmed that my value set exists with the right term, and that the coding system is specified correctly as identified in terminology.hl7.org: http://ncithesaurus-stage.nci.nih.gov.
Question - is this URL recognized by the IG publisher terminology server tx.fhir.org/r4?
If so, any suggestions on what to check?
cc: @Michelle Casagni

Michael O'Keefe (Apr 20 2021 at 14:14):

@May Terry I don't see any ValueSet objects with that URL in tx.fhir.org/r4 when I go to http://tx.fhir.org/r4/ValueSet?url=http://hl7.org/fhir/us/shrae/ValueSet/adverse-event-seriousness-value-set

May Terry (Apr 20 2021 at 14:18):

hmm...okay. lemme look further into this.

Rob Hausam (Apr 20 2021 at 14:18):

Generally the IG Publisher should load the code systems and value sets that it needs to validate, so they don't necessarily already have to be on the server.

Rob Hausam (Apr 20 2021 at 14:20):

It also looks like 'share' in the url is misspelled (I don't see it with the proper spelling, either - but again that may not matter).

Michael O'Keefe (Apr 20 2021 at 14:22):

I think the shrae is deliberate, it's part of the SHR for Adverse Events IG I think

May Terry (Apr 20 2021 at 14:23):

thanks @Rob Hausam - yeh, that's not a misspell..it's the path where my fsh project for my IG resides. so basically, my example created in JSON looks like this:

    "coding" : [
      {
        "system" : "http://ncithesaurus-stage.nci.nih.gov",
        "code" : "C41336",
        "display" : "Non-serious Adverse Event"
      }
    ]
  }

Rob Hausam (Apr 20 2021 at 14:23):

Well, OK. I stand corrected. But it sure looks like a misspelling! :)

May Terry (Apr 20 2021 at 14:37):

so...basically, I'm wondering if tx.fhir.org/r4 recognizes the NCIT URL: http://ncithesaurus-stage.nci.nih.gov, or if the server has the NCIT terms to validate codes? For example, should I expect a response if I query the server as such:
http://tx.fhir.org/r4/CodeSystem/$lookup?system=http://ncithesaurus-stage.nci.nih.gov&code=41336

In this case, I get an error from the server that says

    {
      "severity" : "error",
      "code" : "invalid",
      "diagnostics" : "Unable to provide support for code system http://ncithesaurus-stage.nci.nih.gov"
    }

Rob Hausam (Apr 20 2021 at 14:44):

I don't believe that tx.fhir.org has any content loaded for NCIT (I don't recall any and I did a quick check). But if your IG needs and also provides that content, then the IG Publisher should still be able to validate it on tx.fhir.org.

Rob Hausam (Apr 20 2021 at 14:47):

I doubt that process can handle the full NCIT contents - but possibly a code system fragment with the content that you need could work. We can also look at the possibility of loading the NCIT content on tx.fhir.org. Whatever is done needs to be sure that the licensing is correct (but from NCI I expect that shouldn't be difficult).

Robert McClure (Apr 20 2021 at 14:51):

@May Terry Are you attempting to use the NCI Thesaurus for the code system? If so, "http://ncithesaurus-stage.nci.nih.gov" is not the canonical url for that code system. Always look at the HTA page for external code system information. NCIt is listed and the canonical url is " http://ncicb.nci.nih.gov/xml/owl/EVS/Thesaurus.owl" THO has not been updated yet. Not sure if tx.fhir.org has caught up either and who knows what codes are in the tx server.

May Terry (Apr 20 2021 at 15:04):

Thanks for the additional clarification, RobM and RobH. So... maybe this brings up another source of confusion.
Someone had pointed me to https://terminology.hl7.org as the source for coding system URLs. And https://terminology.hl7.org/CodeSystem-v3-nciThesaurus.html specifies http://ncithesaurus-stage.nci.nih.gov as the Defining URL. So I guess this is not where we should be looking, and instead assume that HTA is the authoritative source?
On a related note, there is also https://www.hl7.org/fhir/terminologies-systems.html, so that is an additional source of confusion on where to look for the right coding system to specify in my examples so it passes validation.

Rob Hausam (Apr 20 2021 at 15:08):

We're working to have everything caught up and in sync, but obviously it is a process and takes some time, and we are in the midst of it so some things will not be completely up to date. If and when there is a difference, HTA is the authoritative source (but tx.fhir.org or any server may not necessarily be up to date with it at a particular point in time). Sorry about that! :(

May Terry (Apr 20 2021 at 15:28):

no worries, RobH. Totally understandable. :grinning: I just like to get to a root cause. My AdverseEvent IG is not being balloted anytime soon, so it's more about figuring out whether I can just accept the error for now or keep troubleshooting. Thanks again!

Davera Gabriel (Apr 20 2021 at 19:51):

Hello @May Terry We have had a series of discussion with the NCI regarding adoption of a canonical URL for the NCiT. Where that landed was the NCI selection of the URL as indicated in the HTA record for NCiT here --> https://confluence.hl7.org/display/TA/National+Cancer+Institute+-+Thesaurus being explicitly: http://ncicb.nci.nih.gov/xml/owl/EVS/Thesaurus.owl The HTA membership raised concerns regarding the use of the ".owl" url so what the NCI adopted was a MIME-specific redirect that points inbound HTML transactions to this page --> https://ncit.nci.nih.gov/ncitbrowser/ The reasoning they provided was that albeit they would like to be good partners with HL7 and cooperate with the requirements for a canonical URL supporting FHIR, there are decades-old legacy NCI technical requirements for persisting the ".owl" URL. I hope that information helps...

Grahame Grieve (Apr 20 2021 at 20:50):

@May Terry tx.fhir.org does not support the NIC terminologies. The validator still might be able to validate, but it depends on how the value set is specified. Can I reproduce this?

May Terry (Apr 21 2021 at 13:23):

Thanks @Davera Gabriel , @Grahame Grieve , and Rob x2.
Looks like changing my URL from https://terminology.hl7.org/CodeSystem-v3-nciThesaurus.html (which is referenced on the terminology.hl7/org publication) to http://ncicb.nci.nih.gov/xml/owl/EVS/Thesaurus.owl has now changed the validation of it from severity error to a warning (which is expected since as Grahame mentioned, tx.fhir.org does not support the NCIT). I suppose that while it's not supported, at least the coding system is recognized.
So moving forward, if the coding system is not on that HTA confluence page, am I to suspect that I will always get this error?
On a related note, should we not rely on the coding system URLs in terminology.hl7.org given that there are discrepancies like this - https://terminology.hl7.org/CodeSystem-v3-nciThesaurus.html with the HTA page? Should I open a JIRA ticket when I find discrepancies like this?
Thanks again all! This is very helpful for improving my understanding of this process.

Robert McClure (Apr 21 2021 at 17:35):

@May Terry Yes, the warning is proper given it's recognized but has no concepts. Yes, always start with the HTA page and if you find that THO has something different, please log a JIRA ticket so we can work through the misalignments that we hopefully already know about. Even better if you searched to find if it's already known but please just log a new ticket if you can't find anything. Yes, if the an external code system is not on HTA, you should ask but in fact there are external code systems in tx. fhir.org and in THO that do not have pages at HTA. How would you find that you might ask, we are slowly trying to make this all work better but not completed and far from implemented. If you need a code system you do not see at HTA, please contact them and let them help determine next steps.

Patrick Werner (Aug 19 2021 at 09:18):

Thanks for bringin this issue to attention @May Terry

Patrick Werner (Aug 19 2021 at 09:18):

Will use http://ncicb.nci.nih.gov/xml/owl/EVS/Thesaurus.owl in german onco profiles as well

Last updated: Apr 12 2022 at 19:14 UTC