FHIR Chat · Provenance.agent.role · implementers

Stream: implementers

Topic: Provenance.agent.role


view this post on Zulip Michele Mottini (May 22 2019 at 00:39):

What's the correct or suggested value for Provenance.agent.role in the case of data received from an outside system (via HL7v2, FHIR or some other interop system)? Author?
(We set Provenance.agent.who to an Organization resource corresponding to the outside system)

view this post on Zulip John Moehrke (May 22 2019 at 14:02):

There is some guidance in the spec on this usecase http://build.fhir.org/provenance.html#6.3.5.5

view this post on Zulip John Moehrke (May 22 2019 at 14:04):

You might find the agent.type vocabulary valueset that AuditEvent uses more useful. You can use these vocabulary as Provenance.agent.type http://build.fhir.org/valueset-participation-role-type.html

view this post on Zulip John Moehrke (May 22 2019 at 14:05):

The Provenance.agent.role would likely be left empty, unless you knew the security roles used. This is unlikely, but possible, so we have the element to record it.

view this post on Zulip Michele Mottini (May 22 2019 at 14:18):

In DSTU2 and STU3 there is not Type , only Role - and for DSTU2 the best (?) value seems 'Author' - so should I still use Type and not Role in R4?

view this post on Zulip Michele Mottini (May 22 2019 at 14:19):

Note that the 'who' is not the import process (eg HL7v2) but the originator (eg Sisters of Mercy Hospital)

view this post on Zulip Michele Mottini (May 22 2019 at 14:20):

(I have to generate Provenance for all three versions)

view this post on Zulip John Moehrke (May 22 2019 at 14:22):

You can use the new valuesets with the DSTU2 structures

view this post on Zulip Michele Mottini (May 22 2019 at 14:24):

Ah I see

view this post on Zulip Michele Mottini (May 22 2019 at 14:25):

But put it in type for R4 and role for the others? Seems strange

view this post on Zulip John Moehrke (May 22 2019 at 14:35):

Trial Use has educated us on the need to make R4 better than DSTU2.

view this post on Zulip Michele Mottini (May 22 2019 at 14:36):

?

view this post on Zulip John Moehrke (May 22 2019 at 14:40):

DSTU2 was published 4 years ago... We have learned that there is a need for an agent.type as well as agent.role. The meaning of .type and .role have also been harmonized among many Resources in FHIR to be more consistent.

view this post on Zulip John Moehrke (May 22 2019 at 14:40):

DSTU -- Draft Standard for Trial Use.... aka, not a standard

view this post on Zulip Lloyd McKenzie (May 22 2019 at 14:41):

Or more specifically, not a locked-down standard :)

view this post on Zulip Michele Mottini (May 22 2019 at 14:42):

So you mean 'yes use type for R4 and role for the DSTU2 and STU3' ?

view this post on Zulip John Moehrke (May 22 2019 at 14:43):

use the vocabulary defined in current specification for Provenance.agent.type in your DSTU2 or STU3 implementation. Vocabulary transcends versions of FHIR specification.

view this post on Zulip Michele Mottini (May 22 2019 at 14:44):

I understand that - my question is where to put the value for R4 - because in R4 there are two separate elements - type and role - should I use type? even if I am putting the same exact value in role in the other two?

view this post on Zulip John Moehrke (May 22 2019 at 14:45):

yes

view this post on Zulip Michele Mottini (May 22 2019 at 14:45):

Thanks


Last updated: Apr 12 2022 at 19:14 UTC