FHIR Report for May 2017

The month of May was dominated by the madrid meeting, and planning for Release 4. This meeting reflected the growing vibrancy and depth in the FHIR community.

The most dominant event of the month was the Madrid WGM. From the FHIR team's point of view, attendance at this meeting was strong enough to allow us to continue as normal, which was a relief, though the connectathon was - for the first time - not larger than the last time. We should expect corresponding growth in the San Diego Connectathon. The WGM activities continue to grow in breadth and strength.

Some specific issues arose from the meeting - these will be discussed below. (Full Report).

After the Madrid meeting, Ewout Kramer and I met with the openEHR community in Oslo. This was a very positive meeting (see notes from this meeting). We identified several productive ways for the FHIR and openEHR communities to work together, which we will follow up with in due time.

At Madrid, the board agreed to hold a DevDays in USA in June, in association with Furore. Planning has started for this meeting, in terms of time, location, and program.

There have been a stream of other related meeting this month. The core team is all over the world doing presentations at various meetings.

The FHIR foundation is slowly coming into shape. We have a membership structure; it just remains for us to prepare the web site to do it. That will be one of my focus activities this coming month. However getting assistance with the FHIR.org web site is our most pressing resource dependency (not our most serious but definitely our most pressing)

Also the registry project is gradually moving forward. We have a summary document that represents a technical vision for how the registry will work, and we are awaiting further review before implementing the registry. We remain hopeful that we can have the registry in place prior to the September meeting.

Implementation Guides/Projects

It's no longer feasible to even pretend to track what's going on in the community. I routinely find out about FHIR Projects after they have gone live. There is a large and growing installation base (makes the FHIR.org website work urgent) with a lot of prototype work approaching implementation.

We are starting to see a profusion of implementation guides being published using the new IG publisher infrastructure, and seeing more and more pre-tooling chains, which is encouraging.

Planning

The formal product agenda was published as part of the Madrid report. All of the things on the agenda are already in train, and some will consume a great deal of time before the January meeting.

Quick Status Analysis:

• Normative : need to plan Sept Connectathon to be sure sufficient coverage before normative
• RDF : RDF sub-group is considering which ontology to choose for mapping. (Snomed? OBO?)
• Data Analytics - no action yet, but requirements have been agreed
• graph API - prototyped a graphQL interface, and implemented it on test.fhir.org. Other work to come
• Patterns - started first step, to get a single canonical code system for resource status codes, and to map actual status codes to it
• Services: nothing yet
• Deployment: focus is on getting Smart-On-FHIR ready for ballot

Connectathon planning: We have issued a call for proposals for streams. I am working with many groups to bring forward interesting and challenging new connectathon streams.

Collaborations

Snomed Intl: We proposed a collaboration, but this has gone nowhere due to lack of resources for HL7 volunteers to work on the project. This has caused some problems for SI, and unhappiness for their members. We are actively seeking some funding support for this work.

W3C: ongoing collaboration with the healthcare life sciences team + mayo is working well

CDISC: nothing is happening here for now

OpenMHealth: Nothing new here - will follow up

AMA: this is an ongoing prolbem that came up several times in Madrid: the licensing conditions around CPT-4 prevent us from working on it, and several approaches have gotten us nowhere. Implementers would like us to help them by agreeing the full details of how CPT-4 works in FHIR, but all we can say for now is 'licensing conditions prevent us from providing you with any assistance'. Any way to fix this?

Open Issues

• Internal infrastructure: we are working on both replacing gForge with Jira (Lloyd promises to get to this this month) and also to replace gForge subversion with Github, so we can better support a wider community. Both changes will be significant, and particularly the change to GitHub will make a lot of noise as we build out new processes around content management.
• Version Management - the community has asked us to do more work on inter-version management. Currently there are 2 streams of work:
  • FGB to describe our goals with regard to inter-version support
  • me to write up the current state of inter-version support in FHIR
  Then we will know where to go - this will be a subject for San Diego
• Smart on FHIR - looks like we will be good to ballot in September.
• The CQL specification doesn't appear to have a nominated product family, nor one of it's own. We propose to bring it into the FHIR family, and publish it under the same rules as FHIR.
• China - I continue to hear - and believe- that China is a great market for FHIR. We need to make sure that our vist there in August yields as good an outcome as we can. Likely approach points are around Chinese LOINC and the Chinese equivalent to RxNorm hosted on tx.fhir.org

Grahame Grieve, FHIR Product Director