The first HL7 FHIR Developer Days conference held in the Americas just wrapped up. The Boston conference gave us three packed days of presentations, demos, comradery, and stunning floor-to-ceiling views of Boston Harbor in the beautiful State Street venue.
Today’s guest blog is by J2 subject matter expert Ayesha Lefebvre. She attended FHIR Dev Days in Boston last week and wrote about her favorite takeaways:
Setting the Scene
Most of the 384 participants were from the US but Argentina and Canada were represented as well as a strong presence from the Netherlands courtesy of the sponsoring Fire.ly team. My experience at the conference leaves me hopeful for innovation in our industry, and I am eager to see what the next year brings us.
One of the keynotes was given by Eyal Oren of the artificial intelligence-focused Google Brain team. It was standing-room-only and focused on the quality of healthcare data and the application of machine learning algorithms to that data to make clinical predictions. There are definite use cases, such as automating some of the routine administrative tasks involved with patient care delivery including pharmacy refills and transportation that can be improved by applying Google AI. It is important to remember though that these algorithms were applied to retrospective data and that actually demonstrating how these models can improve care delivery is quite a complex process.
For more about this keynote address, as well as my thoughts on Dr. John Halamaka’s keynote, stay tuned for Part II of this blog post.
Get ready for FHIR R4 and its new features, to be published sometime in October 2018! Some of the major things under consideration for Release 4:
- Normative – content will be backward compatible to previous versions
- Versioning – support for multiple versioning
R4 will be the first ‘normative version’, which means as the FHIR standard becomes stable, with backwards compatibility with previous published versions. FHIR guru Grahame Grieve pointed out that only some portions of the specification are candidates for being normative, such as patient, practitioner, administration and organization as well as infrastructure (API, xml/json format) and conformance layer resources like StructureDefinition and ValueSet.
FHIR versioning is still not easy but one doesn’t have to maintain multiple endpoints anymore. Unless you want to. Going forward, it will be possible to support multiple versions on the same end-point, by specifying which version to support using a mime type parameter:
Accept: application/fhir+json; fhirVersion=3.0
Other interesting changes to expect are CDS-Hooks integration as well as a framework for querying. Also, there is an open community call for help with mapping transformations between versions. So if you’d like to contribute to the FHIR standard, get in touch with Grahame Grieve (@grahamegrieve)!
In addition to R4 news, Dev Days Boston showed the FHIR developer community has a number of robust API data sources on which to build apps:
- VA API Management Platform: This is the new name for the gateway to access the mountain of data possessed by Veterans Affairs. The site is still in beta but this is an exciting chance for the FHIR and interoperability community to expand healthcare for our veterans through standardized APIs. There is a common clinical data set available, as well as on-going work on FHIR resources to address scheduling, patient encounters, patient feedback, and clinical documentation.
Check it out.
- Blue Button 2.0 – This API was announced at HIMSS18 and we got a chance to see how CMS will use FHIR to allow app developers to deliver Medicare claims to its 53 million beneficiaries. The tutorial session demonstrated how to connect an application (through an OAuth 2.0 interface) to the CMS Blue Button Sandbox environment and retrieve a patient record as well as coverage bundle and ExplanationOfBenefit resources.
Sign up for the developer sandbox at: https://sandbox.bluebutton.cms.gov.
- Synthea: If you don’t know about Synthea already, it is an open source synthetic patient generator built by Mitre. This session demonstrated how to install and set up the software and then generate the realistic health data, based on publicly available demographic parameters. Access to EHR data has long been a challenge for healthcare IT developers. Mitre’s software models more than 500 clinical concepts using the HAPI FHIR library and generates a FHIR bundle for each patient. When you can’t access real EHR data, Synthea is about as close as you can get.
- All of Us / Sync for Science (S4S)S4S was launched in 2016 and is a project coordinated by Harvard Medical School’s Department of Biomedical Informatics, the National Institutes of Health and the Office of the National Coordinator for HIT (ONC) with support from many big EHR vendors. It is a fairly new model of patient-enabled research using open APIs to donate health data. S4S promotes access to health data by both patients and researchers. It also aligns with the objectives of the PMI (Precision Medicine Initiative), a $215 million government research venture with the goal of advancing individualized medical treatments. In this track, we discussed the 2018 Pilot, the All of Us Research Program. Track leader Josh Mandel walked us through the All of Us Participant Portal and showed how research participants can share health data with a just a few clicks, simply by registering as a participant in the portal and giving consent to share their data. Pilot launches have already taken place at Cerner, Allscripts and Epic, this year. The neat thing about All of Us/S4S is that it enables:
a) researchers to request real EHR data donations in a simple, transparent and consistent manner,
b) conforms to HIPAA to enable individuals access to their own health data as well as deciding where and how to share that data, and
c) and is based on open APIs and the Meaningful Use Common Clinical Data Set.
There are also marketplace implications because a vendor implementing S4S will be able to leverage the growing ecosystem of consumer apps that collect and manage consumer health data.
Read more here.
These new resources set the stage for dramatic innovation and collaboration. The team atmosphere in the room made me optimistic that regardless of which organization or role someone represented, everyone involved in FHIR development has a single goal: improving patient care.
This is just Part I of a multi-post recap of 2018’s FHIR Dev Days. In my next installment, I will dive into the keynote addresses by Dr. John Halamka and Google’s Eyal Oren and the vendor FHIR sandboxes. Stay tuned!
PS: Quick shout-out to the conference’s sponsors, including our partner InterSystems, and to my team at J2 who covered my day job so I could have the opportunity to spend an incredibly fun and informative three days with colleagues and renowned FHIR pioneers!
Continued in Part II: The Hype and the Reality.