Stream: smart/health-cards
Topic: Comments on Specification
Grahame Grieve (Sep 02 2021 at 01:43):
I'm just diving into Smart Health Cards (finally) and I've got some comments on the specification
- the spec should really explain the idea of 'fall back' profiles better. It's partially explained, but I had to go hunting for that.
- It's a bit mystifying that Patient.name.period is allowed, but only one Patient.name is allowed.
- I would have thought that Patient.name.text could be excluded for US use. But given that it can't be and the comments on it. it's a bit weird to not mark it as must-support
- there's no need to download the package for validation - just refer to it by package id (and version)
- is there any usefulness in the FHIR Validator being able to validate the JWS directly?
- what is the expectation for patient profiles for other countries? Same IG, or different?
- I believe that the DM profile for patient will be a problem in Australia, since the clinical standard here is three facts for identifying people, while the card only has two - name, dob. Why not include gender as at least allowed? (I may be wrong in my surmise; this may never qualify as clinical - I've given up predicting how these things play out)
- The identity proofing levels for vaccination don't match what's happening here in Australia at GPs: the GPs don't explicitly check your identity when you get the injection, but you're 50% likely to get a person who knows you personally when you turn up, and the GPs talk to you about it before and afterwards. So: no ID was used, but there's actually a pretty good verification system in place (enough? I'll leave that to others to debate)
- why is immunization status 'entered=in-error' allowed in this context? under what circumstances would that ever appear in a card? I would have thought that if it's allowed, it's use should be discussed. I think of entered-in-error as a primary record source thing
- where does dose number go in the immunization resource? The example descriptions talk about dose number but I don't see it in the example
- The lab binding for Observation.code is bound to a US Specific value set?
- surely the lab profile could fix to a dateTime not a period for effective[x}?
- There's no use case for a card that contains both Immunization and lab result records?
Peter Jordan (Sep 02 2021 at 04:51):
Not sure if this an 'entered-in-error' use case, but there have been some cases in NZ where patients mistakenly received a dose of saline solution, instead of a COVID-19 vaccination. Pre-COVID-19 vaccination Identity checks (plus general health and 'any previous adverse reaction' ?s) are taking place here, notably where administration is at pharmacies or large-scale events; but I get also get asked those questions before my annual flu jab at the local general practice where I'm well-known.
Grahame Grieve (Sep 02 2021 at 05:08):
indeed there have been such cases, and in those cases, the primary source should absolutely have an 'entered-in-error' record. But why would you get a card for that?
Peter Jordan (Sep 02 2021 at 05:20):
Guess it depends on what the card is supposed to prove, or disprove. Many have a physical card given when vaccinated, but the electronic record may be modified or augmented after that so the SMART Heath Card might differ - although this is probably an edge case with regard to COVID-19 vaccinations anyway.
Josh Mandel (Sep 02 2021 at 15:11):
Thanks for this feedback @Grahame Grieve -- will you also submit through ballot, or shall we track here? I'll create sub-topics here for a few where I have some thoughts, at any rate.
Grahame Grieve (Sep 02 2021 at 18:51):
I didn't sign up for ballot - was out in the bush when sign up was open
Josh Mandel (Sep 02 2021 at 18:51):
I am signed up so if you want me to put ballot weight behind any comments for you on jira I am happy to do so :-)
Last updated: Apr 12 2022 at 19:14 UTC
