I have come across an issue when implementing the orchestration for the save encounter workflow (see pseudo code here - https://wiki.ohie.org/display/documents/Save+encounter+workflow). We have always, as a community, said that both a provider (health worker) ID and a facility ID must be present for clinical data to be stored. This is so that we can validate the locations that send the clinical content and the person who is responsible for the data.
Now, in XDS.b is seems that one one of these is required or both can be present. Here are the words from the profile text: “At least an authorPerson, authorTelecommunication or authorInstitution sub-attribute shall be present when the author attribute is included in the metadata.” (ITI vol 3 line 939). So this means we won’t always have both a author person and a author institution. So here is my question:
Should we require that PoC systems sending documents to OpenHIM always send both an authorPerson and a authorInsitution attribute (this is more specific than the profile) or do we relax our validation so that only one is required?
Let me know what you think or if I have misunderstood the text.
Cheers,
Ryan
···
–
Ryan Crichton
Lead Developer, Jembi Health Systems | SOUTH AFRICA
I think that we don’t want to enforce sending both unless it is appropriate for the use case. The SHR supports either (I believe). I think the more important requirement to place on Point of Care systems is that if they are to send author information, that they use the full XCN/XON encoding including an identifier (this is more restrictive than XDS but I think it is necessary for resolving of the EPID/ELID).
Also, as a note; the HIM should also be updating the CDA header information as it resolves this data as well, I see the IL doing the following resolutions:
Putting the ECID in the patientId externalIdentifier element
(not modifying the sourcePatient slots)
Putting the EPID in the authorPerson and formatting it according
to the appropriate HL7v2 datatype (I think it is XCN, basically a name with
identifier is the datatype)
Putting the ELID in the authorInstitution and formatting it
accordingly (same as authorInstitution)
Adding the ECID to the CDA header recordTarget
Adding the EPID to the CDA header author elements
Adding the ELID to any CDA organizations that appear
Validating any codes in the CDA body
A better strategy (for CDA content that is) may be to do the resolution on the CDA document data and then promote that to the XDS meta-data by the IL. For example, it would be much easier to do a simple x-path on the CDA body for authors (//v3:author/v3:assignedAuthor[/v3:assignedPerson] to get all authorPersons) , do the resolution on those identifiers and then promote the E[C|P|L]ID to the XDS meta-data using the PCC bindings. Of course, this would only work with CDA content
Cheers
-Justin
···
On Wednesday, October 29, 2014 9:47:17 AM UTC-4, Ryan Crichton wrote:
Hi all,
I have come across an issue when implementing the orchestration for the save encounter workflow (see pseudo code here - https://wiki.ohie.org/display/documents/Save+encounter+workflow). We have always, as a community, said that both a provider (health worker) ID and a facility ID must be present for clinical data to be stored. This is so that we can validate the locations that send the clinical content and the person who is responsible for the data.
Now, in XDS.b is seems that one one of these is required or both can be present. Here are the words from the profile text: “At least an authorPerson, authorTelecommunication or authorInstitution sub-attribute shall be present when the author attribute is included in the metadata.” (ITI vol 3 line 939). So this means we won’t always have both a author person and a author institution. So here is my question:
Should we require that PoC systems sending documents to OpenHIM always send both an authorPerson and a authorInsitution attribute (this is more specific than the profile) or do we relax our validation so that only one is required?
Let me know what you think or if I have misunderstood the text.
Cheers,
Ryan
–
Ryan Crichton
Lead Developer, Jembi Health Systems | SOUTH AFRICA
Thanks Justin. I agree the more important restriction is that the document metadata must contain author information. At a minimum this must either be an authorPerson or an authorInsitution or both. These must be encoded full encoding so that we have a identifier to validate. I’l add some text to our workflow about this.
Thanks for the resolution steps, those are helpful. I’l update the pseudo code with the missing steps.
Extracting the identifiers from the CDA would be ideal but as you say this would only work for CDA content so at the moment I’m extracting these from the metadata rather. I’ve already gone through the trouble of extracting these so a may as well use them
I think that we don’t want to enforce sending both unless it is appropriate for the use case. The SHR supports either (I believe). I think the more important requirement to place on Point of Care systems is that if they are to send author information, that they use the full XCN/XON encoding including an identifier (this is more restrictive than XDS but I think it is necessary for resolving of the EPID/ELID).
Also, as a note; the HIM should also be updating the CDA header information as it resolves this data as well, I see the IL doing the following resolutions:
Putting the ECID in the patientId externalIdentifier element
(not modifying the sourcePatient slots)
Putting the EPID in the authorPerson and formatting it according
to the appropriate HL7v2 datatype (I think it is XCN, basically a name with
identifier is the datatype)
Putting the ELID in the authorInstitution and formatting it
accordingly (same as authorInstitution)
Adding the ECID to the CDA header recordTarget
Adding the EPID to the CDA header author elements
Adding the ELID to any CDA organizations that appear
Validating any codes in the CDA body
A better strategy (for CDA content that is) may be to do the resolution on the CDA document data and then promote that to the XDS meta-data by the IL. For example, it would be much easier to do a simple x-path on the CDA body for authors (//v3:author/v3:assignedAuthor[/v3:assignedPerson] to get all authorPersons) , do the resolution on those identifiers and then promote the E[C|P|L]ID to the XDS meta-data using the PCC bindings. Of course, this would only work with CDA content
Cheers
-Justin
On Wednesday, October 29, 2014 9:47:17 AM UTC-4, Ryan Crichton wrote:
Hi all,
I have come across an issue when implementing the orchestration for the save encounter workflow (see pseudo code here - https://wiki.ohie.org/display/documents/Save+encounter+workflow). We have always, as a community, said that both a provider (health worker) ID and a facility ID must be present for clinical data to be stored. This is so that we can validate the locations that send the clinical content and the person who is responsible for the data.
Now, in XDS.b is seems that one one of these is required or both can be present. Here are the words from the profile text: “At least an authorPerson, authorTelecommunication or authorInstitution sub-attribute shall be present when the author attribute is included in the metadata.” (ITI vol 3 line 939). So this means we won’t always have both a author person and a author institution. So here is my question:
Should we require that PoC systems sending documents to OpenHIM always send both an authorPerson and a authorInsitution attribute (this is more specific than the profile) or do we relax our validation so that only one is required?
Let me know what you think or if I have misunderstood the text.
Cheers,
Ryan
–
Ryan Crichton
Lead Developer, Jembi Health Systems | SOUTH AFRICA
