Proposed Breaking changes for FHIR R4

According to the FHIR maturity model that is part of the FHIR development process, we must consult the community if we are going to make breaking changes to artefacts that are labelled level 4 or 5 maturity (though there’s some lack of clarity as to what exactly is a breaking change, and what the right process should be).

At the recent meeting in San Diego, a number of breaking changes to maturity 4/5 resources were proposed, and we are seeking feedback on these:

  • Rename ValueSet.compose to ValueSet.definition (Vocab)
  • Question on the use of related types with Observation (OO)
  • Proposal to adopt GitHub Flavored Markdown (FHIR-I)

In addition, while we are consulting, we are seeking feedback on an additional change:

  • Proposal to remove ImagingManifest (II)

Each of the changes is described below, along with the proposed justification. If you have already implemented systems that will be impacted by these changes, and you’d like to have your say, please send an email to grahame@hl7.org by Friday 20th October that details your implementation, the impact the change will have, and your opinion about the change.

Note: if you don’t have an actual implementation, and you still have an opinion, we’d still like to hear it – you can let us know through one of our normal channels (see fhir.org).

Renaming ValueSet.compose

ValueSet.compose is a verb — “compose the value set this way”, but that’s not language that connects with the wider community as well as ValueSet.definition would – “this is the definition of the value set content” (which is the “Content Logical Definition”) in the Vocab work group’s wider value set work). Also, ValueSet.definition matches ValueSet.expansion grammatically. Finally, ValueSet.compose overlaps in name with the $compose operation.

For these reasons, the work group would like to rename .compose to .definition. There’s no other semantic changes – it’s just a straight rename, to reduce confusion.

Note that ValueSet is planned for normative in R4 – there won’t be any other chance to make this change.

Use of Observation.related and types

Are you using Observation.related? how? The Orders and Observations Work Group.is considering removing .related and replacing it with:

  • member : Observation 0..*   type: Reference(Observation)
  • derived: Observation 0..*  type: Reference(Observation)

The other things related does would be relegated to extensions, or to the Provenance resource. The task proposing this (GForge #10118) was discussed at the Sept WGM by the Orders and Observations Work Group and is the direct result of internal FHIR methodology review and discussion which resulted in a proposal to substantially change the .related structure. This would simplify the structure of Observation, but the OO work group has been hesitant to make a significant change without meaningful implementer input.

Because our past attempts to get implementer feedback by reaching out to the FHIR community did not result in any response, the work group has deferred this tracker several times. As the Observation resource approaches normative status, there is increasing concern regarding making substantive changes. Implementer feedback is sought regarding whether that perceived benefits of this proposal is outweighed by the costs it imposes on implementation.

Adopting GitHub Flavored Markdown

Currently, the markdown data type specifies that the content is Gruber’s original markdown syntax, with an STU note soliciting feedback about adopting CommonMark instead. Feedback on this has generally been in favor of adopting commonmark, except that it doesn’t support tables, and it allows raw HTMlL. The community has indicated a strong preference for GitHub Flavored Markdown (GFM), and we propose to adopt this as a replacement. Note that the table extension in GFM are likely to find their way into commonmark, which would mean that we’d be using CommonMark with a profile that says ‘no HTML’

In practice, this will mean updating some existing code – e.g. a different markdown processing engine, and different generation, though the markdown syntaxes are somewhat compatible, and markdown degrades gracefully anyway.

Note, again, that the markdown data type is slated to be normative next time, so there’ll be no more opportunity to make this change. (though we could settle for pure CommonMark if the GFM extensions are brought into CommonMark – that would be a non-substantiative change). For further discussion, see chat.fhir.org

Proposal to remove ImagingManifest

The II committee (joint between DICOM and HL7) would like to remove the ImagingManifest resource. It’s a little hard for most implementers to differentiate between ImagingStudy and ImagingManifest, and the DICOM use cases for which ImagingManifest was defined (XDS-I) have been replaced by MHD-I which uses WIA (Web Image Access), leaving the use case for ImagingManifest somewhat obscure.

Hence the proposal is to remove ImageManifest and ensure that ImagingStudy can meet the identified needs.

Note that this is not a change that affects a maturity >=4 resource, so community consultation is not needed. However the committee asked for wider feedback. For further discussion, see chat.fhir.org

FHIR Product Director’s Report from 2017 Sept (San Diego) Meeting

Last week, the HL7 and FHIR communities met for the annual Plenary WGM. This time, the meeting was held at the Hyatt Regency La Jolla, which was a great facility for the HL7 meeting. Thanks to Mary Ann Boyle, who’s taking over from Lillian in organizing the WGMs – they are a big event to organize.

HL7 Formal Strategy

The HL7 Board announced its core strategic goals during the Plenary session:

  • Enhance the public image and achieve recognition by stakeholders as the leading SDO for worldwide health data interoperability standards
  • Secure long-term sustainable revenue to realize the vision and improve customer experiences (internal and external)
  • Establish FHIR as the primary standard for global health data interoperability
  • Enhance and maintain quality of and accessibility to HL7 standards in current use

With regard to the FHIR specific goal, the board provided these strategic objectives:

  1. Increase understanding of FHIR usage and value of usage worldwide (Immediate)
  2. Achieve symbiotic link of brand and financial benefit between HL7 and FHIR.(Immediate)
  3. Demonstrate the value of FHIR in enabling interoperability (Midterm)
  4. Ensure resources are most effectively focused to enhance FHIR (Midterm)

As you can see, FHIR is front and center… I guess we’re all going to be busy in the coming year making this strategy a reality. At a later time I’ll blog about how we can realize these goals.

FHIR Foundation

The FHIR Foundation is now ready to take members. You can sign up here. The annual fee for individual members at this time is US$250. For that, you get:

  • The knowledge that you’re helping maintain the viability of the FHIR ecosystem
  • Access to the Product Director’s monthly report summarizing the FHIR community progress
  • Access to member’s forum (discuss future of the FHIR Foundation)
  • Access to the member’s market place where job/contract opportunities are
    publicized (and these are starting to flow)

We’ll be adding new additional benefits in the future, as discussed with members. We’re hoping that individual memberships will cover the basic year to year costs of keeping the FHIR Foundation viable (legal/accounting fees etc).

The FHIR Foundation isn’t yet ready to open for organizational members, but it’s our intent that we will be soon. We believe that any company or institution selling services related to FHIR (consulting, middleware, implementation tools etc) will want to be part of the FHIR Foundation, and we’re working on services related to  that now. Note that this is not the same as providing healthcare services using FHIR interfaces, though of course we’ll be building membership benefits for those kind of organizations too.

One again, the FHIR Foundation thanks Google for providing the cloud infrastructure on which the fhir.org services are hosted.

Connectathon

The meeting started with the biggest connectathon we’ve had to date – we had over 218 attendees. It’s become clear that we need to rework the way connectathons are managed – we need more organization and more preparation, as they continue to grow. With that in mind, we’ve asked whether anyone is interested in taking on a formal role as the event organizer (If you’re interested, contact me directly).

There were 22 proposed connectathon tracks for this connectathon. At least 2 of them didn’t get enough participants to get off the ground when it came to it, but most did. Each track provided a very brief summary presentation of outcomes, some of you can find in the links below.

As always, the connectathon is very important to us as a way to validate how the specification works. We will continue to add new streams as our community interests broaden. I plan to blog about some of the streams individually, and some streams have already blogged.

Reports and Blogs:

Clinicians on FHIR

The meeting closes with the Clinicians on FHIR event, where we engage Clinicians by working with ClinFHIR to record real clinical cases with the purpose of identifying gaps or deficiencies in the FHIR specification. This was our 10th event. The organizers of the CoF are developing processes to more clearly document clinical scenarios, testing, and feedback to the FHIR development process, and as part of that, we held a new breakout session for Beginners to help get them up to speed on the process.

The Clinicians on FHIR is another important way for us to validate how the specification works.

SMART App Launch Ballot

The Smart App Launch Specification was balloted in the lead up to this meeting. We received plenty of ballot comments, which is great. We’re confident that we’ll be able to publish the first STU for this specification later this year. Most of the comments related to clarification and clarity in the specification, though we’ll be noting some open issues to work on for future versions of the specification.

Community Consultation

Part of the FHIR Maturity Process is that once an artifact is at level 4/5, breaking changes require formal community consultation. In the next week, I’ll be issuing requests for comment on (at least) the following breaking changes:

  • ValueSet.compose, rename to ValueSet.definition (Vocab)
  • Question on the use of related types with Observation (OO)
  • (II) Proposal to remove ImagingManifest
  • (FHIR-I) Proposal to adopt GitHub Flavored Markdown

Notifications on these will be provided in the following places:

Normative Plans

Overall, our plans are unchanged from last time, though we’ve clarified the
time lines. High points:

  • Draft for comment, Dec/Jan: lining up for ballot
  • Mixed normative/STU, Apr/May: Main normative ballot
  • 2nd Normative + STU, Aug/Sept: 2nd chance ballot
  • Publication Mid-Dec

Those are pretty firm deadlines – we’ll slip normative content back to STU rather than hold up the timeline, since various jurisdictions and consortium initiatives are synchronizing their timelines to this (though HL7 can never provide any firm promises in this regard – we have to follow our processes)

A few additional resources are being considered candidates for normative. Last call for comments about which resources should be candidates will be in November (watch this space). After the draft for comment goes out, resources can be dropped from the normative list, but not added.

Bulk Data

The US ONC has asked to work on adding new capabilities to the FHIR specification to increase support for API-based access and push of data for large number of patients in support of provider-based exchange, analytics and other value-based services. This is a high priority work item for them this year. See write up of plans for further information regarding our proposal – this should lead to a connectathon track at HL7’s January meeting in New Orleans, at which all are welcome.

Tooling EcoSystem

The FHIR tooling ecosystem is starting to fill out quite nicely. In particular, the FHIR registry is now entering pilot mode. Please feel free to use, and (for now) provide feedback through chat.fhir.org.

We’ve been working on documenting the tools we have and need – see the HL7 wiki for further information. This is still draft work, and we’re looking for further feedback and community participation.

Certification

The FHIR Proficiency Certification process was trialed at the San Diego meeting. So far, preliminary results show that we’re broadly on the right track. We still plan to have the certification fully available for the January meeting in New Orleans. Congratulations to

Yunwei Wang, IMO

who was the first person to pass the certification. Intending candidates should note that the test aims to ensure that you are familiar with the scope and shape of the full FHIR specification. Knowing the RESTful API and the resources is not enough for proficiency certification (people who know this stuff well from the connectathons/other implementations will still need to read up on the rest of the specification – terminology, intent, licensing, maintenance procedures, etc).

New Areas for the specification

The HL7 technical committees are taking up new areas of functionality and
adding them to the FHIR Specification as draft/STU for the forthcoming
release 4:

  • Public Health Case Reporting and Reportability Responses (PHER)
  • Occupational Data for Health (PHER)
  • Laboratory Test Catalog (OO)
  • BiologicallyDerivedProduct ( blood transfusion, and hematopoietic cell transplant material.) (OO)
  • Medical Device Nomenclature/Vocabulary Service (Dev)
  • Insurance Plans (FM)

If you’re interested in any of these, please get in contact with the relevant committees.

Forthcoming Events

 

FHIR Release 3 Posted

HL7 is pleased to announce that Release 3 of FHIR has just been published.

The FHIR community invested a huge amount of work in this release – hundreds of people have contributed to the specification, and there have been thousands of Change Proposals processed (>2400). Most of these change proposals arose from 3 different places:

  • Implementation Experience (Trial use is working)
  • Alignment with other standards
  • Internal Quality Review processes

Some of the key changes:

  • Added support for Clinical Decision Support and Clinical Quality Measures
  • Broadened functionality to cover key clinical workflows
  • Further development of Terminology Services, and support for Financial Management
  • Defined an RDF format, and how FHIR relates to Linked Data
  • Incremental improvements and increased maturity of the RESTful API and conformance framework

The FHIR specification is very much the living record of the community of users who share the experience of trying to solve problems with it. It’s getting ever more difficult to provide meaning recognition to all the people and organizations who contribute. Still, we’ve tried – see the FHIR credits page.

In addition to FHIR Release 3, today we’ve also published the first release of the US Core Implementation Guide. This generalizes the lessons of learnt through the Argonaut process, and publishes the agreements as a base profile for all use of FHIR in the US context. ONC and other (e.g. HSPC) implementation guides will build on this, and the Core Implementation Guide will provide a general base for consistency across all these contexts.

We expect that the FHIR specification will continue to evolve in the future as we responds to the interoperability needs of the robust FHIR implementation community. Our priority is to advance the well tested platform parts of the FHIR standard to a full ANSI-approved normative standard. I will post more about this later.

 

FHIR Product Roadmap January 2017

R3 plans

The FHIR project is presently finalising “STU3” (Standard for Trial Use, release 3). This 3rd major milestone is currently close to completion. We’ve been meeting in San Antonio this week to finalise ballot reconciliation, perform testing and quality activities, and we are now focusing on preparing the final publication package. Following our publication plan we expect to be publishing release 3 on or about Mar 20.

R4 plans

Once R3 is published, we will start working on release 4. The various committees that manage the different parts of Release 4 have been discussing their scope of work for R4, and planning their engagement and implementation activities to support that this week.

Some of the major things under consideration for Release 4:

  • Improvements across all domains
  • Cds-hooks integrated in FHIR Specification
  • Query language framework
  • Support for integrating research and clinical practice

The most significant change is that R4 is expected to be the first ‘normative version’. It’s important to understand what that means. We will continue to follow our overall maturity model, where content gradually matures through testing and implementation activities that demonstrate success in the real world. The end of the process is “normative” (“FMM level 6”), where the standard becomes stable, and breaking changes are no longer considered.

Only some portions of the specification are candidates for being normative. We are currently considering balloting the following parts of the specification as normative:

  • Infrastructure (API, data types, XML/JSON formats, conformance layer resources like StructureDefinition and ValueSet)
  • Administration (amongst others Patient, Organization, Practitioner)

We will continue to seek and receive comments about this. Some clinical resources may be considered, depending how implementation experience unfolds this year.

Overall planned R4 timeline:

  • Dec 2017: publish first draft of R4 for comment (finalise plans for normative sections)
  • Jan 2018: first R4 based connectathon(s)
  • April 2018: ballot R4
  • May – Sept 2018: ballot reconciliation
  • Oct 2018: publish FHIR R4

We will conduct a round of market consultations in Aug/Sept 2017 to seek comment on this timeline from the FHIR community.

Note that this timelines anticipates that we publish R4 in October 2017 irrespective of the outcome of the normative ballot. Anything that has not passed normative ballot will continue to published as STU. We are still working on preparation, quality and balloting processes to support the normative FHIR ballot.

Longer term, we anticipate following R4 with a roughly 18 month publication cycle, with increasing normative content.

Implementation Progress

FHIR is a specification for a common API for exchanging healthcare data between information systems. Any information system supporting the healthcare system can choose to implement the common API, and exchange data following the rules. FHIR enables a ‘healthcare web’ to exist, but doesn’t actually create it.

HL7 is pleased to work on the FHIR specification with many hundreds of partners, who are all implementing the specification to exchange data in service of the healthcare needs to their enterprises, their customers, and, ultimately, patients. HL7 does not ‘implement’ the specification (other than various prototype/test services) – our partners and other healthcare services do.

Argonaut Project

One particularly important sub-group of the FHIR community is the Argonaut project, which is a joint project of major US EHR vendors to advance industry adoption of FHIR and we’ve had many questions about the Argonaut implementation timeline for EHR access. With regard to the Argonaut community:

  • The Argonaut STU2 specification for Common Clinical Data Set elements is close to being finalized and will be announced shortly.  The Argonaut STU3 specification for Provider Directory will be published after final balloting of STU3
  • Most Argonaut members who are certifying an API with ONC are using the Argonaut specification; most certifications are expected in Q1/Q2 2017
  • Software roll-outs have commenced — progress will vary depending on the vendor
  • It is presently unknown what the adoption rate by provider institutions will be — MU and MACRA in the US provide incentives to make a patient-facing API available by the end of 2017
  • Some of the Argonaut interfaces provide additional functionality not yet described in the Argonaut specification, but there is considerable appetite for additional services beyond what is currently available. The Argonaut project will be making announcements about its future plans shortly which will be announced in a press release, through collaboration channels, and at www.argonautproject.org

 

Summary Report from SNOMED Meeting

Last week, IHTSDO held their main annual business meeting and SNOMED CT Expo in Wellington, New Zealand. I attended, along with a number of other key players from the FHIR terminology community, and we took advantage of the opportunity to hold a joint meeting between the FHIR community and the IHTSDO community.

From the FHIR community’s point of view, this was an important meeting because many of the FHIR stakeholders make extensive use of SNOMED CT and we still have a long way to go before we’ve truly mastered all the issues associated with using SNOMED CT. From IHTSDO’s point of view, FHIR is a very important implementation mechanism by which SNOMED CT content is used in production, and their stake holders have identified working well with FHIR as a priority.

The joint meeting was split into 2 parts – a general exploration of the known issues, and a technical exploration of some of them. During those 2 meetings, we discussed the following issues:

  • Progress since Montreal. In the HL7 working meeting in Montreal (May 2016), the HL7 Vocabulary WG (on behalf of IHTSDO) asked the FHIR project to make several changes around the way the FHIR Publication tools handle SNOMED CT editions, versions, copyright statements, and mappings. These were almost all completed prior to the the Wellington meeting, but we’ve been unable to transition fully to the International Edition for the base FHIR specification because of some specific value sets using US specific content. We agreed that we will change to the International Edition once this content has been promoted to the International Edition.
  • Terminology Service Implementation Guide for SCT. The Terminology service API published as part of FHIR has a very active community, and we’ve been testing the API with SNOMED CT content since we first started using it (it’s already in production in Australia). However our work has primarily focused on the API, and there’s an opportunity for us to improve the service for consumers by being more consistent in how SNOMED CT structures and semantics are exposed through the API. To that end, it was recommended that HL7 and IHTSDO should collaborate on a set of detailed rules about how the SNOMED CT concepts are exposed on the terminology servers. Most of these are things that the server providers have already discussed and agreed on informally, but it would be useful to get formally documented agreement around this. IHTSDO will also look at the feasibility of providing a reference FHIR terminology server.
  • Versioning – IHTSDO is planning to move towards a more continuous release cycle. While the details of this aren’t worked out yet, it may present some issues to the way we use SNOMED CT in FHIR – though FHIR is not special in this regard: all users will have the same challenges and opportunities. We agree that we will have to consider the impact this would have on the FHIR eco-system
  • Cross-Edition support – Presently, the SNOMED CT implementation space is divided by the national release center system – many affiliates have their own edition, with additional concepts, descriptions, relationships, and reference sets. As long as all records come from within the same jurisdiction, this is ok, because you just choose an edition. But if your records cross jurisdictions – which is something that will happen for many systems implementing patient focused record support, then there is currently no specific guidance to support this. IHTSDO agreed that this is something that implementers will need.
  • Value set review – IHTSDO has performed a review of a number of SNOMED CT value sets in the FHIR specification and provided feedback to HL7 on the results. IHTSDO has also developed a reproducible review process that may be applied to other SNOMED CT value sets. It was agreed that the review of SNOMED CT value sets in the FHIR specification should be completed, and that additional SNOMED CT value sets may be identified. IHTSDO has also previously offered to provide HL7 international with a set of SNOMED CT concepts that may be used for free in its international products. Part of the value set review will include looking at opportunities where this might be useful for the FHIR specification.
  • Binding/Mapping progress / Implementation Advice – Linda Bird from IHTSDO has been working on bindings and mappings (both words are used slightly differently across IHTSDO and HL7) between SNOMED CT and FHIR resources, both a simple attribute mapping, and a more complete template binding. So far, we’ve looked at Condition, AllergyIntolerance, Observation, Procedure, Goal and Family Member History. This has identified a number of issues in terms of gaps between the relevant SNOMED CT concepts and the design of the FHIR resources. Some of these may be flaws in the FHIR resources, or in the SNOMED CT content, or they may just be inevitable results of different perspectives – that’s yet to be resolved. We need to work on this jointly, and that might lead to changes to either FHIR or SNOMED CT, or implementation guidance for FHIR implementers that use SNOMED CT, or a formal FHIR Implementation Guide. We’re not sure where that will end up yet.
  • Simplification – we also discussed an idea for providing a framework for sharing additional views on SNOMED CT as a way for helping implementers. We agreed to pursue ideas around making SNOMED CT easier to use for beginning implementers informally for now. The draft SNOMED CT Machine Readable Concept Model (MRCM) currently out for community review (http://snomed.org/mrcm) may allow these simplified views to be created automatically.

A small group of IHTSDO and HL7 participants has been selected to form a joint project group to turn all these ideas into a working project plan. We’d welcome further comment from people who are interested in this area who weren’t at the joint meeting. On the HL7 side, contact me. For the IHTSDO perspective, contact Jane Millar. We’ll be working on this over the next few months.

FHIR Meeting Report – Montreal, May 2016

Preparation for STU 3

The main focus of the meeting was preparing for the September ballot of release 3, which will be “STU  3” (note that we’re now using the title “STU” – Standard for Trial Use –  rather than “DSTU”  – Draft Standard for Trial Use – to reflect that important parts of the spec are well past the draft stage)

Planned key dates:

  • ballot sign up starts: Jul 27
  • ballot opens: Aug 12
  • ballot closes: Sept 12
  • Baltimore Connectathon / HL7 meeting: Sept 16-23
  • tentative target release date for STU 3: Dec 31

STU 3 is currently planned to include:

  • formats: no change to XML & JSON formats, but we will generate JSON schema. Introduction of RDF, tied to an ontological base
  • RESTful API: no change to existing API, bar some clarifications around transactions. Maybe add Patch?
  • Conformance – split out CodeSystem from Value set and minor changes to other resources, including the use of FluentPath instead of XPath
  • Core Clinical, Administrative & Financial resources – ongoing minor changes in response to trial use and improved quality
  • Continuing improvements to the Clinical Decision Support / Quality Measure framework
  • A new framework for workflow / task management
  • Draft mapping framework and CCDA/FHIR mapping guides

Alongside STU3, we expect to be providing a full tool chain to support implementation guides, covering editing, publishing, validation, and a public registry

Growing Maturity

One clear conclusion from this is that growing importance of the FMM (maturity model) – some areas of the specification are quite stable now, and are being managed accordingly. STU3 will probably be the last version of FHIR that doesn’t include any normative content.

With regard to FMM levels, In the lead up to the Montreal meeting, we surveyed the FHIR use base. From the results of that survey, we were able to make the following list of resources that are a priority for implementers:

Patient Observation Medication Condition
Practitioner DiagnosticReport MedicationOrder AllergyIntolerance
Organization DiagnosticOrder MedicationStatement Immunization
Encounter Bundle MedicationAdministration CarePlan
Person Conformance MedicationDispense Procedure

Specifically, these are resources that the community would most like to see move up through the maturity levels, so the HL7 committees will prioritize these resources when preparing for ballot. Note, though, that it’s mostly up to the community to trial these resources now.

Specification Road map

The FHIR specification is now starting to have real breadth and depth, and we’ve had some comment about growing complexity of the specification. In particular, these areas have attracted comments:

  • RDF / Base Ontology work
  • Fluent Path and the new mapping language
  • Clinical decision support resources
  • CIMI / logical model development

The concern around this work indicates that we need to provide better a better road map to the specification on the main page, and the documentation page, so that people can better understand how these parts of the specification relate to each other, and which parts are relevant for them to implement.

We will be working on this over the next few months.

Acknowledgements

The FHIR community continues to grow in both size and importance. 1000s of implementers have taken part in connectathons, and the number of active contributers – committers, editors, work group co-chairs, facilitators, evangelists – just keeps growing. And not just the number, but the volume and depth of their input.

The FHIR community is our biggest asset – and it’s getting bigger every day.