Stream: implementers
Topic: Best practice - ID(s) in FHIR messaging
Thea Mentz Soerensen (Jan 03 2022 at 10:14):
Hi,
We have created FHIR messaging Bundles and are unsure about what is best practice when it comes to cardinality for Bundle.id and MessageHeader.id.
In the specifications (https://hl7.org/fhir/R4/messaging.html#3.4.1.4) it is stated that there SHALL be a MessageHeader.id each time a new message is created – so, should this be translated to a cardinality of 1..1? And what about the cardinality for Bundle.id?
Thank you in advance for your input.
@Anders Jensen @Ole Vilstrup @Torben M. Hagensen
René Spronk (Jan 03 2022 at 10:53):
In the context of messaging bundles, MessageHeader.id is effectively 1..1
René Spronk (Jan 03 2022 at 10:59):
I'm not seeing any specific guidance on Bundle.id in the spec (it may be in some other place I've not been looking at) - its use is however essential in the detection of duplicates.
Jens Villadsen (Jan 03 2022 at 11:05):
Sounds like it should be expressed as fhirpath constraint since its conditional to the bundle type
Thea Mentz Soerensen (Jan 03 2022 at 11:43):
René Spronk said:
I'm not seeing any specific guidance on Bundle.id in the spec (it may be in some other place I've not been looking at) - its use is however essential in the detection of duplicates.
I agree - it is indeed essential. Would you advocate that the Bundle.id also should have the cardinality 1..1?
Thea Mentz Soerensen (Jan 03 2022 at 11:47):
Jens Villadsen said:
Sounds like it should be expressed as fhirpath constraint since its conditional to the bundle type
That's a good point.
René Spronk (Jan 03 2022 at 12:58):
In the context of messaging, if one uses a transport which is not reliable, then yes, Bundle.id should be 1..1. If one uses a reliable transport to support message exchanges, it may not be necessary. Therefore, it may well be a conscious decision not to create a constraint to mandate Bundle.id for message bundles.
Last updated: Apr 12 2022 at 19:14 UTC
