ATNA essentially supercedes the Basic Security profile that was added
in earlier versions of the IHE Radiology technical framework, which
was implemented for various different IHE connectathons.

The updated radiology-specific messages are now being defined in the
IHE Audit Trail Option supplement. The public comment text of this is
available at:

http://www.ihe.net/tf/IHE_TF_Suppl_R...2005-02-21.pdf

A draft for trial use should be out shortly.

David

jonny31415@yahoo.com wrote:

> I am looking at implementing an Auditing System and was hoping that
> someone in this group could verify some of the things that I have
> discovered in my research and perhaps answer some questions.
>
> 1) It seems that there are 3 separate documents that describe Security
> Auditing: Dicom Supplement 95, RFC 3881, and IHE ATNA. The IHE is built
> on top of the lexicon defined in DICOM and DICOM on top of that defined
> in RFC 3881.
>
> 2) IHE had a provisional schema, but new implementations should use the
> RFC 3881 schema and perhaps provide a XSLT transform to map from the
> provisional schema to the one in RFC 3881.
>
> 3) The IHE provisional schema looks like a nicer schema to use because
> it is more structured and presents a better domain interface. The one
> in RFC 3881 is more flexible.
>
> 4) It seems that the concept of Organization is not well defined in the
> schema. RFC 3881 section 5.5.6 "Participant Object ID" states that the
> id field can be a composite data field. Does this mean if you want to
> scope, for instance a patient, by the organization then it would be
> something like org1\12345 (Organization\PatientID)? If so, then it
> seems the contents of this field would be implementation specific and
> not generally consumable. It seems there should be an "Issuer"
> attribute for scoping Participant Object ID's in a
> multi-organization/multi-RIS environment.
>
> 5) If an implementation is both the Audit consumer and supplier and
> wants to pass around GUIDs for patients, procedures, etc., would the
> right thing to do be to add a ParticipantObjectDetail type/value pair
> that identifies the object in question?
>
> 6) Why is there so little discussion of this standard in any groups?
> Are people waiting for it to be finalized before they implement it?
>

Thanks very much for the info David.