Stream: argonaut
Topic: How much churn is there in argonaut specs?
Adam Flinton (Jul 29 2019 at 07:56):
Putting together a medical system with lots of incoming messages in lots of message formats (edifact, proprietary, hl7 v2, v3, etc.etc. ) i.e. the std range of medical messages built up over the years. Looking at if we should transform to a common standard internally (still carried via messaging over aws sns/sqs etc) . 2 basic options, our own json created by the initial/main endpoint for the message (e.g. clinical record for a lab report) or something like fhir4/argonaut.
Internal Json - main advantage is it's entirely under our control
fhir/argonaut - main advantage we may (probably will) want to give other vendors access to system and thus it is more of a standard for external parties.
So the question is, is the argonaut / fhir4 spec going to be a reasonably static target (i.e. we have no interest in uplifting internally frequently) and if not yet, then any ideas as to when?
Micky Tripathi (Jul 29 2019 at 10:45):
The ONC final rule will be published sometime in late fall 2019 and will very likely point to a specific US Core/Argonaut R4 implementation guide version that will be balloted in early fall. Because that version will be enshrined in regulation, it should be both stable and widely adopted in the US.
Lloyd McKenzie (Jul 29 2019 at 14:04):
Many of the resources that Argonaut/US Core constrains are not yet locked down as 'normative', however they are quite mature - so we shouldn't see a lot of breaking changes before the 'final' version - which for many will likely be late 2020. However, breaking changes are still possible.
Adam Flinton (Jul 30 2019 at 07:02):
Thanks both. Late 2020 might be a bit late but fall 2019 wouldn't be. Will discuss internally.
Grahame Grieve (Jul 30 2019 at 07:04):
you might not like the fact that an external standard is subject to change, but any internal thing you do will also be subject to change....
Adam Flinton (Jul 30 2019 at 10:10):
yup change is a constant :-)
Josh Mandel (Jul 30 2019 at 14:46):
(My typical recommendation here is to consider FHIR as a starting point for your internal models -- use r4 internally and run with it; worst case you can always fork it for your own needs over time, but more likely you'll appreciate the community process around developing/improving FHIR.)
Kevin Mayfield (Jul 31 2019 at 06:48):
I've used fhir internally in a couple of orgs, not profiled though but influenced by them. It was far more tolerant of change than our previous models, we did make our endpoints more tolerant to change. This refactoring took the most time but it was needed.
Brett Marquard (Jul 31 2019 at 14:14):
High probability of a new US Core version with Argonaut lessons late US Fall 2019....It still won't be 'normative' but will capture all the latest lessons
Last updated: Apr 12 2022 at 19:14 UTC
