Talk:Master File/Registry Control Act Wrapper

There is an inconsistency in the MPI example and, I think, a problem of interpretation:

"The receiver of MPI messages should therefore consider the entire list comprised of (PatientRole.id 1..1, OtherIDs.id 0..*) to be identifiers for the scoping person, PatientRole.id should not be ignored 8even though mostly it is an MPI internal ID). "

In fact the identifiers are for the "player", rather than scoper.

The problem: if the external id's supplied to the MPI originated from different organizations then they reflect the fact that the entity is a patient in two different contexts. To put it another way: the person in question is a patient twice-over. The identifiers are therefore not synonymous. [An interesting wrinkle on this that the role instances could in fact have the *same* numeric identifier but issued by different scopers (with different OID's).] The id created by the MPI reflects neither of these scopers, but is some other kind of artifact entirely. Arguably, it is not even a 'patient' role id.

Robert grant 01:12, 2 Jun 2006 (CDT)