RelatedPersons test data

Has anyone identified a patient who has RelatedPersons that are ‘carer’s’ or ‘next of kin’ in the INT environment. Randomly clicking test patients in the all test patient list doesn’t seem to produce any.

If you need specific test data, you can always ask for it

I have them all and none of them have related persons mentioned as in this one has such and such related persons. Short of trying each one to look I was hoping it was obvious.

Hi @iminshall

Please contact the test data team directly and they will create the data you need.

Test data email address:

Kind regards,

Thanks I have requested it.

Patient/9449305889/RelatedPerson returns a RelatedPerson resource, with RelatedPerson.relationship of Carer and Guardian.

However, RelatedPerson.patient.identifier.value is 9651181834 which is a different patient! Looks like another error to workaround in the client :face_exhaling:

Thanks. Yes they do have a Carer… at least its something I can work with.

1 Like

Is that not correct?

I was looking at the examples in the OAS spec and Jane Smith appears as a Patient and as a RelatedPerson, the related person is Jane Smith

Can’t tell for certain as the RelatedPerson doesn’t have an identifer.

deleted this message

Let’s assume we ask for “Persons related to ALICE”, and get a response with a single RelatedPerson, BOB. I would expect BOB.patient to be ALICE.

If not, where in the data model is BOB linked to ALICE so that the request can be answered?

The FHIR spec says…

I think that is pretty clear…

That patient should be the Patient of the RelatedPerson. The NHS number we see in the patient.identifier should be in the RelatedPerson identifier array so you can get to the RelatedPersons demographics if not in the resource itself

Assuming by NHS Number you mean the RelatedPerson’s NHS Number

But NHSE seem to have their own rules…

Removed original reply

I am referring to ab example like this.

Swap the NHS Numbers around and then change the relationship

I think the schema for RelatedPerson has redefined the FHIR definition?

As I read the linkage PDS is using, this field can’t be optional and @dunmail is correct. It shouldn’t be a reference to another NHS Number.

The NHS Number in RelatedPerson.patient.identifier.value, the one in the path Patient/{id}/RelatedPerson and the one Patient.identifier.value should all match.

The other NHS Number should be in RelatedPerson.identifier.value and is optional.

1 Like

Happy to disagree here Kevin. The example you quote is incorrect as to the spirit of the specifications, you cannot have the same identifier in each identifier array. The patient should point at the patient (not the RelatedPerson) the identifier of the RelatedPerson can contain the identifier for the Person…

According to the latest R4 Core HL7 FHIR® UK Core R4 | UK Core RelatedPerson - SIMPLIFIER.NET

RelatedPerson.patient.identifier = The Patient identifier

RelatedPerson.identifer = The identifer of the RelatedPerson

Seems we are at an impass and the PDS lookup is incorrect in my opinion.

I don’t think we are disagreeing? You’ve said the same as me and PDS specification I quoted looks to be wrong.

If a RelatedPerson and Patient are the same person, then the RelatedPerson.identifier (not present in PDS) and Patient.identifier will have the same NHS Number

If they are different then these will be different. The linkage which shows the relationship is RelatedPerson.patient - the Patient this references or identifies should be different to the RelatedPerson that contains this relationship.

I have raised it with the PDS team. Let see what they say

1 Like

your not disagreeing with what I say? (as I’m also about to raise)

My point will be this is deviating from FHIR specification.

This is describing use of RelatedPerson.identifier, not RelatedPerson.patient.identifier

But need clarification on what join does this id represents


Which identifier does it mean?
a. RelatedPerson.idenfifier
b. RelatedPerson.patient.identifier ← I think it’s this