FHIR Connectathon of the IHE-MHD Profile

At FHIR Connectathon, three brave shoes are ready to experience the IHE-MHD profile.

Diego Kaminker (Almadok Project / Argentina)
Fernando Campos (Project Almadok / Argentina)
Simon Keane (NHS Digital / United Kingdom)
They made every effort. I just answered the questions like this …

They have a specific goal: they want to test the IHE MHD profile on the general FHIR servers. In particular, FHIR servers do not advertise on the IHE MHD profile but support the essential elements required by IHE-MHD. This work is important because it demonstrates my success in writing an IHD MHD profile. How close to FHIR (STU3) and how to notify the public about deviations.

I’m trying to write IHE-MHD in a remote place where the general FHIR server works. However, it is not intended to be covered by IHE-MHD. The desired range or IHE MHD is an API for an XDS, XDR, or XCA environment. When I write IHE-MHD, I had to restrict it in a way that would not support Universal FHIR server.

However, there are some wonderful surprises.

I summarize this report with examples of real world and XML guides.

Take the test
1. Document Source: New Document (1 binary attachment in package)
2. Document source: No binary package attachments: The binary file referenced by the external server
3. Buyer’s Document: find a filter combination

1. ID Theme
2. Types of Type / Document Type
3. Document creation set date.
4. Consumer Document: Get a direct binary document

Notes: We did not specify the security / authentication requirements for this connection, such as: Token / RBAC / etc.

Test: Server: with HAPI, Vonk, Grahame Server

Clients: Typical REST with Chrome Restart Client (Almadoc), Postman / Eclipse (LDP)

results
1 – By ITI-65 document availability package that carries both binary documents and references binary documents expected MHD that the server DocumentReference.content.attachment.url “repaired”. If the server continues to run binary if you specify a binary URI ID (GUID) for the URL server, the new URL is stored in DocumentReference.content.attachment.url with the same URI (GUID) how there are packages. If the Fhir data type is used and working with earlier versions of Fhir (new from the reference of a complex data type, not just a URL) is a good example of this behavior Fhir solution specification

I made CP to fix the problem, see GF # 13823
It is recommended that you do the URL correction for the transaction as you would for the reference.
Some chat recommends that zulip uses DocumentReference.content.attachment.data, but has many problems. Most of all, it works with IHE XD * to use cases.
If it does not melt, it may be that the next revision MHD ITI-65 is referred to as the Fhir process, rather than using the transaction. I do not think it should do it until FHIR STU4. But it makes it easier for someone who wants to define it as an option. However, I would like to remind everyone that IHE MHD does not attempt to use Fhir server for general purposes, but it is an API for XD *.
Cold is two transactions. To create a binary file and create a DocumentReference in the client’s static URL. This can be done with FHIR’s general servers, but is against IP-MHD as an API for XD *.
I will write CP to MHD to clarify it as required by MHD MHD server.
2 – Known Issues # MHD_048 – This is not a new problem because we learned about it. In MHD text, the problem is not clear, so there are problems. The default behavior of the Universal FHIR Server is that “incoming” resources can not be searched. At IHE-MHD we leave it as an open question and do not try to solve it completely.
It turns out that one of the questions is for resources.
Bad news can not tell you if it works with server compatibility provider.
I wrote CP on this feature, which I did not notice.
3 – Search for patient identification: a problem that I just did not show, as much as possible. This search is sequential and works only if the querying server also hosts the patient resources.

Leave a Reply

Your email address will not be published. Required fields are marked *