Stream: IPS
Topic: IPS & SHC
Grahame Grieve (Sep 13 2021 at 22:00):
If you'd like to convert an IPS to a SHC, you can do this:
-
Post your IPS as a transaction to http://test.fhir.org/r4 (web workaround: go to that address in your browser, paste the IPS into the text box at the bottom of the page, and choose 'upload')
-
observe the id of the patient created (look at the id of the second resource in the response)
-
fetch a healthcare card for the patient created (web workaround: http://test.fhir.org/r4/Patient/[id]/$health-cards-issue?credentialType=https://smarthealth.cards%23health-card)
bingo, you have yourself an SHC generated from the IPS.
Grahame Grieve (Sep 13 2021 at 22:01):
There's some issues with this, though. On the IPS side, you must ensure that the immunizations are validly coded for SHC - that means, in practice, they must have a CVX, SCT, ICD-11, ATC, or AIR code. (they can have other codes, but must have one of them)
Grahame Grieve (Sep 13 2021 at 22:03):
I generated this QR code from Peter Jordan's sample IPS.
Grahame Grieve (Sep 13 2021 at 22:03):
Jens Villadsen (Sep 13 2021 at 22:04):
I've always wondered how Peter looked on the inside ... now I know :grinning_face_with_smiling_eyes:
Josh Mandel (Sep 13 2021 at 22:05):
This is super cool, thanks Grahame! Is it driven by a structure map?
Grahame Grieve (Sep 13 2021 at 22:07):
no. the logic is a bit tricky for that. And also, I just used my generic 'generate a SHC card' code from the repository after loading the IPS into the repository
Grahame Grieve (Sep 13 2021 at 22:07):
the actual logic is here:
Grahame Grieve (Sep 13 2021 at 22:08):
https://github.com/grahamegrieve/fhirserver/blob/master/server/healthcard_generator.pas#L243
Grahame Grieve (Sep 13 2021 at 22:09):
to do for me is to populate the display for a performer if there's a reference not a display
Jens Villadsen (Sep 13 2021 at 22:15):
It should be fairly easy to also do this with the EU DCC
Grahame Grieve (Sep 13 2021 at 22:15):
if it's based on IPS... should be the same process?
Jens Villadsen (Sep 13 2021 at 22:16):
EU DCC is not based on IPS - it is based on ... well ... its own definitions
Jens Villadsen (Sep 13 2021 at 22:16):
if you're interested you can find the contents here: https://github.com/ehn-dcc-development/ehn-dcc-schema
Jens Villadsen (Sep 13 2021 at 22:17):
and all the community reference libraries here: https://github.com/orgs/eu-digital-green-certificates/repositories
Grahame Grieve (Sep 13 2021 at 22:18):
how is that related to https://build.fhir.org/ig/hl7-eu/dcc/index.html?
Grahame Grieve (Sep 13 2021 at 22:18):
because the process would be the same for anything conforming to that guide
Jens Villadsen (Sep 13 2021 at 22:19):
I did the implementation for DK using a library developed in Sweden. I don't think that IG was ever used. At least not to my knowledge
Jens Villadsen (Sep 13 2021 at 22:19):
The EU DCC is not related to FHIR
Jens Villadsen (Sep 13 2021 at 22:24):
@Giorgio Cangioli might know what the origins of that IG is - but as you can see, it never went beyond v0.0.1
Grahame Grieve (Sep 13 2021 at 22:29):
anyway, looking quickly at the DCC, it could be converted, but I'd actually have to do it to be sure
Jens Villadsen (Sep 14 2021 at 06:44):
we're probably a bit late to that party
Giorgio Cangioli (Sep 14 2021 at 07:24):
Jens Villadsen said:
Giorgio Cangioli might know what the origins of that IG is - but as you can see, it never went beyond v0.0.1
Yes the development was put on hold when the focus was given only on the QR code generation for getting a Yes/No answer.
The json schema specified https://github.com/ehn-dcc-development/ehn-dcc-schema defined for EU is reflected by this draft logical model (beside the root element) https://build.fhir.org/ig/hl7-eu/dcc/StructureDefinition-HC1.html. But it needs to be reviewed and used vocabularies will be added.
I still believe that that DCC guide may have sense when the interest will not be limited to a Yes/No answer; but on getting vaccination information cross-border
Jens Villadsen (Sep 14 2021 at 10:10):
The DCC is not limited to a Yes/No - the DCC 'simply' states the data on person. It is up to any given authority to examine whether that data allows for 'passing the gates'
Giorgio Cangioli (Sep 14 2021 at 10:13):
Yes but the scope is to generate a pass/no pass answer; not to share data about vaccinations
Carl Leitner (Sep 14 2021 at 14:26):
Grahame Grieve said:
how is that related to https://build.fhir.org/ig/hl7-eu/dcc/index.html?
the IG is considered a non-normative representation (as I understand) of the specs @Jens Villadsen referenced
Jens Villadsen (Sep 14 2021 at 14:49):
Giorgio Cangioli said:
Yes but the scope is to generate a pass/no pass answer; not to share data about vaccinations
Actually not, AFAIK. The scope is to be able to expose data relevant for the DCC data model. It is correct that you wont have a certificate if you dont have any data that meets the data model, but it is definetly not a yes/no model.
Christof Gessner (Sep 14 2021 at 18:19):
Actually, the implementers decided that the FHIR spec would be an unnecessary layer of abstraction (for the purpose). The group preferred to work from a JSON schema as a reference. All the mechanics of valueSets/codeSystems were agreed “ad hoc”, likewise some details on data and string formats that were not expressable in the JSON schema. Based on the initial FHIR work, some underlying structures of the JSON may still reflect kind of FHIR-based ideas
Christof Gessner (Sep 14 2021 at 18:48):
A central part of the DCC implementations appears to be the decision support that is implemented gathering qualified “business rules” expressed in JSONlogic. Applied to the data, those rules yield the desired automated yes/no answer, specific for the ruleset of the target country or region.
Christof Gessner (Sep 14 2021 at 19:27):
https://ec.europa.eu/health/sites/default/files/ehealth/docs/eu-dcc_validation-rules_en.pdf
https://github.com/ehn-dcc-development/dgc-business-rules
Grahame Grieve (Sep 14 2021 at 20:03):
can't imagine what could go wrong with ad hocing all that stuff...
Neelima Karipineni (Sep 14 2021 at 20:11):
Thanks to all who joined the IPS & SHC breakout session. Major points to follow-up (@Max Masnick @Paul Denning @Reece Adamson feel free to add) -
-
CVX support - consider adding slice for CDC CVX to Immunization-uv-ips, which currently slices SNOMED GPS and WHO ATC. CDC CVX is one of the few currently openly licensed code sets that supports vaccine codes to the granularity needed to identify specific medicinal product and is committed to representing international (non-US approved vaccine) codes. We are considering making CVX must-support in the SHC vaccination and testing IG.
-
credential types for IPS - if there is a use case for transporting 'raw' IPS resources via health card, can consider defining credential types (see https://spec.smarthealth.cards/vocabulary/) for the IPS domain to allow querying and identification of these cards. While a QR code has data size limitations, they can transport a fair amount of data as in https://spec.smarthealth.cards/examples/example-02-a-fhirBundle.json, which can be transported in 3 QRs.
Paul Denning (Sep 15 2021 at 12:17):
See about getting COVID vaccinations into SNOMED GPS
Rob Hausam (Sep 15 2021 at 12:49):
We can further discuss @Neelima Karipineni's points (particularly #1) on our IPS Touchpoint and also the regular weekly HL7 IPS call (11:00 AM ET) today. And, as @Paul Denning suggests we should gather the data to make a request to SNOMED International for requesting additional COVID vaccination codes in the SNOMED CT GPS.
Jens Villadsen (Sep 17 2021 at 20:04):
Grahame Grieve said:
can't imagine what could go wrong with ad hocing all that stuff...
it actually works, AFAIK
Grahame Grieve (Sep 20 2021 at 18:44):
typically, things like that do but are fragile against the future. We have enough problems with the future...
Jens Villadsen (Sep 21 2021 at 17:40):
I dont believe there is a future where SHC/DCC is widely used
Josh Mandel (Sep 21 2021 at 17:43):
For SMART Health Cards, we anticipate many common workflows including presenting childrens' vaccination records to schools. (In many parts of USA, this is a manual, frustrating, and error prone process -- has been for decades and won't decline in importance.)
John Moehrke (Sep 21 2021 at 17:55):
long standing need for immunizations cards for many things. Often we all forget when these events happen.
John Moehrke (Sep 21 2021 at 17:56):
I had to dig up my immunization cards for Visa applications to many countries.
John Moehrke (Sep 21 2021 at 17:57):
Today, this is often just a statement by the individual. But if we had an interoperable technical solution, it would be more used.
John Moehrke (Sep 21 2021 at 17:58):
but, I am with Jens that this future is unclear. COVID has driven the use-case due to urgency.
Last updated: Apr 12 2022 at 19:14 UTC