Hi Ven
I assume this is related to your previous question about special
characters because you have examples of sites stuffing a textual
identity domain disambiguator in with the patient ID value. Using
a prefix or suffix to qualify Patient ID or Accession Number is a
not uncommon ad hoc way of avoiding collisions when institutions
merge or interact (though the 16 character SH VR of Accession
Number makes this harder, compared to the 64 character LO limit
for Patient ID).
A "better" way to address this is using the mechanisms described in
the Multiple Image Manager/Archive (MIMA) profile:
http://www.ihe.net/technical_frameworks/#radiology
In short, this specifies the use (including mapping and defaults) for
these attributes (amongst others):
- Issuer of Patient ID (0010,0021)
- Issuer of Accession Number Sequence (0008,0051)
- Other Patient IDs Sequence (0010,1002)
- Institution Name (0008,0080)
- Institution Code Sequence (0008,0082)
By the way, it is important to consider two potentially distinct ways
in which this information can be used:
- to communicate the identity domain in which the identifier is defined
- to communicate in what domain the instance was acquired/created
One cannot assume that because an identifier in a particular domain is
present in an object, that it was acquired in that domain, because it
may have been coerced (e.g., into the local domain) during ingestion
(etc.). Hence the use of Institution Code Sequence in MIMA.
The MIMA profile also describes the concept of an AE title based mapping
for when the sending system lacks support for these attributes, i.e., the
receiving system adds the issuer information (etc.) based on (configured)
knowledge about where an object came from.
David
PS. Related to this is IHE CP-RAD-222, work in progress, which addresses
the Acquisition Modality populating these de novo (with SWF.b, unrelated
to MIMA), by adding an Enterprise Identity Option, requiring the presence
(and query support for) Issuer of Patient ID and Issuer of Accession Number
Sequence.