Is slicing on $this correct?
Shouldn't https://github.com/hl7ch/ch-core/blob/944e7bf370ef629d8d72f2bef1dc167da0b7d3b6/input/fsh/profiles-resources/CHCoreOrganization.fsh#L9C15-L9C41 have been system instead of $this?
The idea behind using $this is to allow for example ONLY type to serve as a discriminator for an additional identifier slice, alongside system, as we have it in the CH Core Patient profile for identifier.
Proposal to leave the slicing as it is to be discussed at an FHIR telco.
then you are technically allowing multiple identifieres with the same system which would be a bit weird, IMHO.
then you are technically allowing multiple identifieres with the same
systemwhich would be a bit weird, IMHO.
yes, but you would need to differentiate (e.g. int the type) that it could be used as slices.
Telco: we don't see added value of switching just to the system discriminator, @jkiddo do you have an issue with that if we stay the way it is?
Do you have a case where the system doesnt implicitly points to a type?
Do you have a case where the system doesnt implicitly points to a type?
maybe a local organization identifier (like the local PID for pat):
{
"identifier": [
{
"type": {
"coding": [
{
"system": "http://terminology.hl7.org/CodeSystem/v2-0203",
"code": "XX"
}
]
},
"system": "http://hospital-group-abc.ch/local-identifier",
"value": "LOCAL-ORG-001"
}
]
}
Using $this as the discriminator in the core profile provides flexibility for derived profiles compared to system-only discrimination. While system-based slicing locks derived profiles into the same discrimination pattern, $this enables derived profiles to slice by any combination of identifier elements (system, type, use, value patterns) as needed for their specific use cases.
I'm yet to see a case where it doesn't make sense to use system for identifiers. If you prefer to stick with $this, then do so, and you can close this issue.
Since we currently have no specific use case for identifiers with slicing via type, we will switch to the system as requested by @jkiddo.
fixed with #389