FHIR Chat · FHIR Validator and Synthea and Deprecated Codes

Stream: tooling

Topic: FHIR Validator and Synthea and Deprecated Codes

Jason Walonoski (Jun 18 2021 at 13:47):

So, a GitHub issue was posted on Synthea today, saying that the FHIR validator errors out on our synthetic data, because we're using deprecated codes.

https://github.com/synthetichealth/synthea/issues/902

Error: HTTP 422 Unprocessable Entity: Unknown code {http://snomed.info/sct}698360004 for 'http://snomed.info/sct#698360004'

698360004 is deprecated the correct code is  http://snomed.info/sct#735985000

This has probably been discussed a lot previously, but is this the behavior we should expect from the validator? Part of me says "yes, the codes are no longer valid" and another part of me says "we can't expect older data to be updated, possibly incorrectly, just because a code was deprecated. At one point it was valid."

I suppose I could write a script to occasionally scan Synthea for deprecated codes and we can replace them, but that just sort of sweeps the issue under the table.

Would the correct thing to do be to change my use of the system from http://snomed.info/sct to http://snomed.info/sct/[sctid] or http://snomed.info/sct/[sctid]/version/[YYYYMMDD] as according to http://hl7.org/fhir/R4/snomedct.html#4.3.1.0.3?

Which in practice would look like changing this:

{
        "coding": [ {
          "system": "http://snomed.info/sct",
          "code": "698360004",
          "display": "Diabetes self management plan"
        } ],
        "text": "Diabetes self management plan"
      }

to this:

{
        "coding": [ {
          "system": "http://snomed.info/sct/900000000000207008/version/20210131",
          "code": "735985000",
          "display": "Diabetes self management plan"
        } ],
        "text": "Diabetes self management plan"
      }

Would the FHIR Validator handle that correctly?

Grahame Grieve (Jun 18 2021 at 21:57):

you'd leave the system alone and introduce a version. But where exactly is this error coming from? I don't recognise it

Rob Hausam (Jun 19 2021 at 03:18):

Nor do I.

Jason Walonoski (Jun 21 2021 at 12:44):

I'm not sure where the error is coming from, the person who posted the issue just reported the FHIR validator -- I'll try to get more information.

Jason Walonoski (Jun 21 2021 at 14:10):

FWIW, the person who reported the bug said it was the HAPI FHIR Server (https://hapifhir.io/), their own installation with MariaDB as backend database – With FHIR validator activated.

Last updated: Apr 12 2022 at 19:14 UTC

Main menu