Namespace for identifier

[bdc-ehealth]

@MatonAnthony
The namespace for the UHMEP identifier is: https://www.ehealth.fgov.be/standards/fhir/referral/NamingSystem/be-ns-uhmep (see IG).

Other namespaces can be used by clients (but they have to mention the UHMEP namespace, if there is one available, and if it is available that is the one that will be used by UHMEP. If no UHMEP identifier is available (at creation time) it will be created by UHMEP). Do all parties involved agree with this? (-> refer to the discussion in the architecture workgroup).

[MatonAnthony]

Hi, following our conversation yesterday on this very matter.

• Other namespace cannot be provided by clients on a free text basis.
• While I understand the advice made in the architecture workgroup and we take them into consideration, at this stage this analysis is not over for us and has to be treated internally.

I will take the point internally with the relevant people and we can schedule this point when we are ready.

TLDR; For me no decision at this stage, it's on our radar

[bdc-ehealth]

WG: we will add the argumentation of the SMALS security team here. We still need a solution for the problem situation (network problem, ...) to indicate that this information was already sent.

[bdc-ehealth]

WG: the proposal after discussion is to create a transaction identifier in the header of the HTTP message to allow for a possible duplicate of the information , e.g. because of network failure. Bart will check whether this is accepted by FHIR, but as this is a basic HTTP feature, chances are very high.

[AlexisVZ-MDS]

Just to be sure; so if we sent a request but receive no answer, we have to send the request again, with the same transactionid in HTTP Header, until we receive a response with the UUID generated by UHMEP server, which we then have to store. The transactionID will not be resend to us ?

[bdc-ehealth]

@AlexisVZ-MDS
That is what I understood.

[MatonAnthony]

Yes, but this should happen very rarely. Then whether or not you store the prescriptionId somewhere is not my concerns. We also offer all the facilities to research and find the prescription without the prescription identifier.

[bdc-ehealth]

I have found no indication up to now that this approach is forbidden by FHIR. I propose to use this header: https://en.wikipedia.org/wiki/List_of_HTTP_header_fields#cite_note-44:~:text=Csrf%2DToken%3A%20i8XNjC4b8KVok4uw5RftR38Wgp2BFwql-,X%2DRequest%2DID,-%2C%5Bstackoverflow2%201

[bdc-ehealth]

WG: Bart Reekmans will check this before next week. Hans and Geert and Robin agree.

[BartReekmans]

We also agree

[bdc-ehealth]

WG: the workgroup agrees to use this solution for referral prescription.