Stream: argonaut
Topic: Epic server
Michele Mottini (Sep 13 2019 at 14:16):
@Danielle Friend This is what I got testing:
Warning importing Patient/eroCd17NbCffM3E3WXi7eOQ3: The identifier with namespace 'urn:oid:2.16.840.1.113883.4.1' has no value
Error importing AllergyIntolerance?patient=eroCd17NbCffM3E3WXi7eOQ3: Search AllergyIntolerance failed: Operation was unsuccessful because of a client error (BadRequest). OperationOutcome: Overall result: FAILURE (1 errors and 0 warnings)
[FATAL] Content invalid against the specification or a profile.(further diagnostics: Required clinical-status parameter is not provided.) (at /f:clinical-status).
Error importing Condition?subject=eroCd17NbCffM3E3WXi7eOQ3: Search Condition failed: Operation was unsuccessful because of a client error (BadRequest). OperationOutcome: Overall result: FAILURE (1 errors and 0 warnings)
[FATAL] A required element is missing. (at /f:category).
Error importing Observation?patient=eroCd17NbCffM3E3WXi7eOQ3: Search Observation failed: Operation was unsuccessful because of a client error (BadRequest). OperationOutcome: Overall result: FAILURE (1 errors and 0 warnings)
[FATAL] A required element is missing.(further diagnostics: Must have either code or category.) (at /f:code/category).
Cooper Thompson (Sep 15 2019 at 18:22):
The issue with urn:oid:2.16.840.1.113883.4.1 is due to our use of the rendered-value extension. In order to be able to support masked SSNs, I think we might actually want to look at updating the US Core Patient profile to allow either value or the rendered-value extension.
Cooper Thompson (Sep 15 2019 at 18:22):
See https://chat.fhir.org/#narrow/stream/179166-implementers/topic/SSN.20(.2FNational.20ID).20masking
Michele Mottini (Sep 15 2019 at 18:25):
OK - on our side is just a warning, not a real issue - we are able to import the resource
Michele Mottini (Sep 15 2019 at 18:27):
The other errors are your server validating its own data against the profile - and returning 400 if it fails? I am reading it wrong?
Cooper Thompson (Sep 15 2019 at 18:47):
I submitted GF#24443 to consider the SSN masking case.
Danielle Friend (Sep 15 2019 at 19:28):
The other errors you mention is our server returning operation outcomes for additional required search parameters.
Danielle Friend (Sep 15 2019 at 19:28):
For Observation, the spec does not require a patient only search - we expect a patient and code or patient and category
Michele Mottini (Sep 15 2019 at 19:56):
AH got it - I'll fix that on our side and re-test
Eric Haas (Sep 16 2019 at 13:37):
@Cooper Thompson see my follow up on GF#24443 There is no issue when using this extension on the value element since the element is present. Will you withdraw the comment - if yes I will mark as withdrawn
Michele Mottini (Sep 16 2019 at 13:50):
(@Danielle Friend The error message could maybe be something like 'Missing required search parameter xxxx' - 'Requiring element is missing' seems to refer to the resource itself)
Cooper Thompson (Sep 16 2019 at 14:01):
@Eric Haas Gaah! We implemented that extension wrong :(. We (and by we I mean me, since I did the dev) put it as an extension on Identifier, not on Identifier.value. Yeah, I'll withdraw and submit a bug report on our implementation. :/
Robert Scanlon (Sep 16 2019 at 17:49):
@Danielle Friend -- you had discussed potentially pulling up a patch to this server regarding case sensitivity on the epic-user-id header. Is that something you are going to have a chance to do? I never closed the loop on that w/you.
Michael Donnelly (Sep 17 2019 at 07:20):
I'm ~50% sure I did a development log to take care of that.
Michael Donnelly (Sep 17 2019 at 07:20):
@Danielle Friend or @Cooper Thompson, do you know if that made it to the sandbox (if it happened)?
Michael Donnelly (Sep 17 2019 at 07:21):
Because you're right, @Robert Scanlon, that definitely should not be case sensitive.
Danielle Friend (Sep 17 2019 at 11:39):
Correct - We've fixed it but the connectathon environment just needs to be updated. @Michael Donnelly I'll send ya the reference to it, because it wasn't as simple to move as I expected and then I got distracted - sorry @Robert Scanlon! We can let you know when that update is in.
Robert Scanlon (Sep 17 2019 at 13:29):
Thanks @Michael Donnelly , @Danielle Friend for the update. I thought it would be easy to change on our end too but to get it to behave this way may involve monkey patching Ruby's http module ;-) . FYI @Reece Adamson
Danielle Friend (Sep 19 2019 at 19:33):
@Robert Scanlon mind giving it another try? We put the update in :)
Last updated: Apr 12 2022 at 19:14 UTC
