Stream: patient administration WG
Topic: mother's maiden name extension
Craig Newman (Oct 21 2021 at 16:04):
Curious why the mother's maiden name extension (http://hl7.org/implement/standards/fhir/extension-patient-mothersmaidenname.html) uses the string data type rather than HumanName. Typically, it is the mother's maiden surname that is captured and shared, but why not offer the ability to send discrete name parts?
René Spronk (Oct 22 2021 at 07:55):
As I understand the US context mothers maiden name is a 'shared secret', akin to the name of your first dog, the city of birth of your father, etc. As such the 'name-structure' aspect doesn't really matter. In Europe (different law system), if we were to have mothersName as an extension, it would be using HumanName as a data type. But in the European context, that's not a shared secret, but part of a patient's demographics.
Brian Postlethwaite (Oct 22 2021 at 09:56):
I understood it to simply be the pre-marital surname, not any other parts.
I guess you coukd also just uss another old name in the repeatin name on patient if you wanted more details.
Craig Newman (Oct 22 2021 at 14:25):
In many use cases (even in the US), it's treated as demographic information to aid in patient matching. I agree that it's sometimes used as a security verification question, but it's more that just that.
Craig Newman (Oct 22 2021 at 14:26):
I wouldn't think that it would be a repetition of the patient's name as they've (likely) never used that surname (assuming the mother changed her name at marriage) or they are currently using it (if she did not and the child uses her surname)
Brian Postlethwaite (Oct 23 2021 at 02:31):
Sorry, missed the "mothers" part in there... I should have read more closely.
Hans Buitendijk (Dec 06 2021 at 18:51):
@Brian Postlethwaite : So since v2 comes in with a full name data type, should we merge the components into a string OR create an origignalHumanName extension to preserve the details OR do you plan to update the core extension and make it a HumanName data type?
Brian Postlethwaite (Dec 06 2021 at 23:16):
@Cooper Thompson any thoughts on this one?
René Spronk (Dec 07 2021 at 06:44):
Note that some countries already use an extension for 'mothersName' (with a HumanName data type). 'maiden' is a name type. That would fit this use case as well.
John Moehrke (Dec 07 2021 at 12:57):
just because it was done that way in v2, should not make us invent use-cases to justify something that our current use-cases don't call for. Meaning, the use-cases I know of explicitly call for just the maiden name. Providing more than that is considered unnecessary and a risk.
The alternative you do have is to create a RelatedPerson where you carry the full name and other information.
Cooper Thompson (Dec 07 2021 at 14:16):
If you have the full mother's name, why not just put that in Patient.contact.name instead of the extension? That way you can represent the mother's name and the father's name the same way. It is probably worth adding some clarification around when to use Patient.contact and when to use the extension. But honestly I'm not sure when you'd use that extension instead of the Patient.contact.
René Spronk (Dec 07 2021 at 14:40):
Well, one needs to be able to contact a contact person. Without contact details, or in the case of a deceased person, it'd be difficult to have them as contacts. We're talking about demographic details of the patient, not about their contact parties. If it's purely about demographics then RelatedPerson also doesn't fit the bill. You're stuck with some sort of extension.
Cooper Thompson (Dec 07 2021 at 15:17):
The invariant on the Patient.contact structure doesn't require actual contact info. Having a name is sufficient. The other option is RelatedPerson. I'm not sure I like the idea of having a third way of identifying someone's mother's name.
Last updated: Apr 12 2022 at 19:14 UTC
