Stream: IPA
Topic: ballot without connectathon?
Isaac Vetter (Oct 27 2021 at 16:12):
Dear FMG co-chairs, (@David Hay , @Lloyd McKenzie ),
The IPA project team has not held a connectathon track. We intend to in Jan, 2022. We'd really like to ballot at STU in Jan. Content is largely complete.
Isaac Vetter (Oct 27 2021 at 16:12):
IPA is the International Patient access specification (different from IPS!). It's a scaled down, minimal version of national FHIR profiles, such as US Core. It's intended to provide a proven foundation on which both new countries can use as a starting place for the regulation of patient access via FHIR APIs, and also as a base level of functionality for multi-national apps and FHIR servers.
Isaac Vetter (Oct 27 2021 at 16:12):
The project team believes that the intent of the connectathon requirement has been met in this specific case through other connectathon tracks, such as the Argonaut US Core track, SMART tracks as well as significant production deployment of RESTful FHIR APIs for common clinical data set for patient access.
Isaac Vetter (Oct 27 2021 at 16:12):
What do you guys think about this? If we submitted a NiB for Jan, would FMG reject it out of hand?
Lloyd McKenzie (Oct 27 2021 at 16:48):
What we care about is that the content of the specification has been reduced to code and do you have a community that will use the spec. Essentially there are three questions:
- Has the technology represented in the spec been implemented and do we have some sense it works
- Is the specification sufficiently clear that independent implementers can pick it up, write code, and successfully interoperate?
- Is there a set of folks who care about this IG enough that they're willing to write code to test it? (because if not, why are we bringing it to ballot?)
You're probably fine wrt the first in terms of past experience with other implementations. The real question is the second. If the IG is largely copy & paste from US core with a few minor changes, you're probably fine. If there's any significant variation that might bring into question how implementable/well-written the spec is, then the FMG would have grounds for seeking a connectathon first or seeking justification on why there was urgency to go to ballot before there was implementation experience with this particular spec.
I'm assuming that for #3, you have sufficient evidence of intention to implement that the lack of a connectathon before start of ballot wouldn't be a barrier.
In short: certainly wouldn't reject it out-of-hand, but might have some questions and need some justification.
Isaac Vetter (Oct 27 2021 at 17:49):
Lloyd, thank you, sir! This explanation is thorough, comprehensive and logical. I'll work with the WG to see if we can submit a NiB. Towards the goal of facilitating FMG discussion & decision, I've added an agenda item on FMG's call next week. Happy to push back to whenever is convenient for FMG.
Isaac Vetter (Oct 27 2021 at 21:57):
fyi IPA'ers. The Patient Care WG voted today to submit a January NiB for IPA; however, it looks like our PSS may not have moved from a status of WG Approval to TSC review and therefore may not be a candidate for a Jan ballot. I'm following up.
Rob Hausam (Oct 27 2021 at 21:59):
That's unfortunate. But I'm expecting that you may be able to make a case for getting an exception for that, too - which presumably is what you are following up on.
Isaac Vetter (Oct 28 2021 at 14:05):
NiB created: http://www.hl7.org/special/committees/tsc/ballotmanagement/DisplayNIB.cfm?ballot_document_sdo_id=1199 :fireworks:
Mikael Rinnetmäki (Nov 03 2021 at 14:20):
Regarding @Lloyd McKenzie's point 3 (about who cares enough):
In Finland, we have a new legislation and patient access to health data will be implemented in 2022.
A big risk is that without a specification, we will implement something proprietary, or something that’s close to IPA, but not quite. And that will lead into problems later on.
Having a specification with a higher status would be an important signal for the implementers here that this is something that is worth adopting, and help them commit to following the international specification.
I'm sure we can find people to write code to test the specification, from the local community. But the governmental organization actually responsible for implementing the patient access API for the national healthcare IT infrastructure have indicated they'd need more than just a draft of a spec in order to commit to adopting it.
Mikael Rinnetmäki (Nov 03 2021 at 14:22):
See https://chat.fhir.org/#narrow/stream/194447-nordics/topic/Use.20case.20for.20IPA.20in.20Finland/near/239859922 for some more details.
Last updated: Apr 12 2022 at 19:14 UTC
