Stream: implementers
Topic: Meta.source as extension in R2?
Dan Gottlieb (Apr 01 2019 at 21:09):
I'd like to include the Meta.source metadata element from R4 in R2 resources. Is there an official extension url I should use to backport this?
Lloyd McKenzie (Apr 01 2019 at 21:13):
http://build.fhir.org/versions.html#extensions (though we haven't published them to my knowledge - @Grahame Grieve ?)
Dan Gottlieb (Apr 01 2019 at 21:19):
Thanks @Lloyd McKenzie - exactly what I was looking for!
Dan Gottlieb (Apr 01 2019 at 21:23):
So for meta, should I do "http://hl7.org/fhir/4.0/StructureDefinition/extension-Meta.source" on the type or does it have to be resource specific like "http://hl7.org/fhir/4.0/StructureDefinition/extension-Patient.meta.source"?
Lloyd McKenzie (Apr 01 2019 at 21:28):
extension-Meta.source is fine
Lloyd McKenzie (Apr 01 2019 at 21:29):
It should be the "path" where the element is defined. (If you don't think that's clear in the spec, feel free to submit a change request for us to clarify further.)
Josh Mandel (Apr 05 2019 at 17:10):
Just to check on expected values for Meta.source... whether for R2 or R4, I created GF#20688 to try to understand whether Meta.source is intended to be the URL of a system from which the resource came, or a full URL to a source resource? Is there a clear expectation on this?
Igor Sirkovich (Apr 05 2019 at 19:04):
Interestingly, Australia had defined their own extension for Metadata Source and a few other R4 pre-adoption extensions at http://build.fhir.org/ig/hl7au/au-fhir-base/extensions.html (I know because we had considered leveraging their Metadata Source Extensions at some point). Maybe someone would be able to communicate the "extension-Meta.source" approach to the Australian Base Profiles folks?
Lloyd McKenzie (Apr 05 2019 at 19:34):
Can't hurt :)
Grahame Grieve (Apr 05 2019 at 20:42):
@Brett Esler
Grahame Grieve (Apr 05 2019 at 20:44):
I commented in the task - can be either.
Brett Esler (Apr 07 2019 at 23:16):
Australian pre-adopting of R4 elements was done on a case by case basis - pre-adopt is based on what the definitions of the element are in the referenced FHIR version and since this was the CI build of R4 had the potential to change until R4 was done - would the http://build.fhir.org/versions.html#extensions approach be the official way to include cross version elements? Having local national versions for the same things seems a bit problematic for me.
Brett Esler (Apr 07 2019 at 23:17):
i'll get this looked at with AU working groups...
John Moehrke (Apr 08 2019 at 15:09):
when I saw this show up as a URL; I figured our intention was very broad possibilities. If you have very specific Provenance, then use Provenance resource... I expected those that would use meta.source would have some local IG that defines the meaning of the URL. Some might want it to always be to a formal FHIR Resource, some might want it to be a binary URL, other might just use it to hold a organization identification (public web site).
Brett Esler (Apr 10 2019 at 02:05):
ok a query on pre-adoption in general - the scheme as per http://build.fhir.org/versions.html#extensions works great if you have a published major version to reference - how to handle case where you want to reference elements before the final publication e,g the CI build or ballot snapshot
this will come up now for R5; can't really refer to elements in R5 when there is no final R5 - any suggestions? Just live with it and deprecate the 'custom' interim extensions in favor of the systematic standard approach when versions are final?
Grahame Grieve (Apr 10 2019 at 02:17):
not sure. This is a practical use case now?
Brett Esler (Apr 15 2019 at 06:25):
it was for STU3 to R4 before R4 was finalised...
Brett Esler (Apr 15 2019 at 06:29):
@Brian Postlethwaite is suggesting I should just use the standard scheme and if it turns out wrong change it
Last updated: Apr 12 2022 at 19:14 UTC
