Stream: fmg
Topic: QA - GF#12144 - Approved
Lloyd McKenzie (Mar 12 2017 at 15:39):
GF#12144 - This involves defining a profile on Provenance as part of the main build, which is technically a substantive change. The only impact on existing content is updating the usage notes for those resources with a relevantHistory element to point to the new profile
Lloyd McKenzie (Mar 12 2017 at 15:39):
+1
Paul Knapp (Mar 13 2017 at 10:30):
I'm concerned about this proposal. It seems to be central to the pharmacy handling of information yet it was handled by FHIR-I not pharmacy with no indication of support from Pharmacy or Security, and would represent an un-reviewed core profile being attached to an FMM 1 or higher resource.
John Moehrke (Mar 13 2017 at 12:48):
I am not clear on how this is not a core use-case for Provenance. Is this not simply narrative documentation? I am not objecting to the overall concept, just the urgency to force this through in the 11th hour with minimal oversite. I could support a narrative update now, with a more normative fix post STU3.
John Moehrke (Mar 13 2017 at 12:48):
Is FHIR-I ready to take ownership of Provenance? It seems to be taking on more of a multi-purpose infrastructure flavor. I think Security WG has taken it as far as they can given their experience. They would continue to want to keep it aligned with security concepts, but there seems to be many proposed uses that are infrastructure, not security.
Lloyd McKenzie (Mar 13 2017 at 15:47):
FHIR-I isn't taking ownership of Provenance. We do have ownership of Workflow. This is to create an FMM 0 profile that would be pointed to by several Request (and Event) resources that follow the Workflow pattern of pointing to Provenance for history to give guidance on how to use the resource for that purpose. Pharmacy actually asked for the profile to be created. Security will certainly be involved in reviewing the profile before it can progress to FMM 1.
Grahame Grieve (Mar 13 2017 at 20:17):
so is the profile defined?
Lloyd McKenzie (Mar 13 2017 at 20:42):
Yes, though not committed
Grahame Grieve (Mar 13 2017 at 20:46):
what does it say?
Lloyd McKenzie (Mar 13 2017 at 20:50):
Must support target, period, reason, activity, agent. Activity must be present and is constrained to codes that actually deal with record changes (create, update, suspend, resume, etc.) Must specify agent.role as 1..1 and propose initial value set covering author, data enterer, witness and verifier (significantly cut down from base value set and focusing on participation types)
Grahame Grieve (Mar 13 2017 at 20:54):
considering this, and reading back through the issues, I am in favor of the change. This is just documentation, and it's just a FMM 0 profile.
Grahame Grieve (Mar 13 2017 at 20:54):
and pharmacy is asking for the change
Lloyd McKenzie (Mar 13 2017 at 21:03):
Can you record the +1? (Makes it easier for me to count :))
Grahame Grieve (Mar 13 2017 at 21:06):
+1
John Moehrke (Mar 13 2017 at 22:13):
Lloyd, I was not asking if you were embarking on a hostile takeover. I was suggesting that it is time for FHIR-I to own Provenance.
Grahame Grieve (Mar 13 2017 at 22:13):
I read it the other way too. I think this is a good discussion to have in Madrid
John Moehrke (Mar 13 2017 at 22:15):
Creating the Profile does not seem too worrisome, although there is already similar in the EHR Lifecycle; right? The concern I had was to then propagate it as a mandatory profile for all uses of relevantHistory. It is this last change that seems rather fragile to make so late in the process.
Lloyd McKenzie (Mar 13 2017 at 22:33):
It won't be propagated as mandatory. Just adding a usage note that points to the profile as "providing helpful guidance".
Lloyd McKenzie (Mar 13 2017 at 22:34):
I'm open to FHIR-I taking on Provenance if we think that's necessary, though I think we'd need a stronger argument than "workflow has profiled it"
John Moehrke (Mar 13 2017 at 22:40):
like we have discussed on other Resources (Imaging), the owning workgroup should be aware of Profiling that is going on. Security was not aware of this. I don't think there is a problem. I am more worried about the efforts that I don't know about, and the unintended impact of a late-breaking change with very little oversight.
John Moehrke (Mar 13 2017 at 22:42):
So, how is your profile on Provenance different from EHR? http://build.fhir.org/ehrsrle/ehrsrle-provenance.html
David Hay (Mar 14 2017 at 00:22):
+1
Brian Postlethwaite (Mar 14 2017 at 05:26):
Just confirming that this is the inclusion of a new profile on Provenance, and some explanative text on some pharmacy resources on their event history property that says see [link], and no other changes correct?
Lloyd McKenzie (Mar 14 2017 at 05:33):
Explanatory text will go on whatever resources have adopted the "relevantHistory" element from the pattern to point to Reference(Provenance). But other than that, yes. Text will be in usage notes and say something like "For guidance on using Provenance to capture event history information, see [link]" where the link would point to the draft profile.
Brian Postlethwaite (Mar 14 2017 at 05:46):
+1
Lloyd McKenzie (Mar 14 2017 at 06:02):
In answer to @John Moehrke, the EHRSRLE profile doesn't focus on making things easy for implementers to understand, doesn't tighten down the vocabulary to what's relevant. This profile is much simpler and focuses on "how do you use Provenance to capture the clinically relevant information for event history".
Paul Knapp (Mar 14 2017 at 10:55):
+1
John Moehrke (Mar 14 2017 at 14:52):
+1
Lloyd McKenzie (Mar 14 2017 at 16:42):
Approved
Last updated: Apr 12 2022 at 19:14 UTC
