Reflections from FHIR Camp: Trust, Interop, and the Power of Community
Wrapped up first day of FHIR Camp, it nice in Portugal (unless you land at 7:30 am in the morning and get stuck for an hour and a half in slow rush hour traffic heading to the hotel).
First I had the privilege of leading a Birds of a Feather session on conversational interoperability, and honestly, seeing firsthand the level of interest and energy from the community was incredible. The discussions ran deep trust, security, authentication, and some seriously forward-thinking ideas around smart scopes for agent authorization. We also explored use case validation, clinical guideline development for agent-based workflows, and agent confidence scoring. The takeaway? The community’s thinking about interoperability not just as APIs, but as conversation and collaboration.
Another highlight was the FHIR R6 preview session and it’s a big deal. All resources in this version will be normative, and any trial-use materials will move into implementation guides. That shift will stabilize the core spec and accelerate adoption. Some interesting tidbits:
USCDI and the US Core team are committed to full version migrations, not mixing versions so we will go all in on R6 or stay with R4.
Huge shoutout to Gino Canessa and the working group for the analysis and heavy lifting to understand R6’s downstream impact the level of detail and rigor here is just next-level.
Big kudos as well to Health Samurai and their crew for an incredible showcase from TypeScript code generation to Atomic EMR and more. Their approach continues to make FHIR practical, fast, and developer-friendly.
And maybe my favorite part connecting with people. So many passionate implementers from Finland, Sweden, Denmark, Australia, Chile, Portugal, and beyond. The global momentum is real. On the US side, there’s a ton of focus around CMS 0057 and Prior Auth with payers and providers finally converging on making real use of the investments they’ve already made.
There were also some deep dives on model context protocol alignment, and serialization standards, all in service of making mcp actually work at scale.
This community’s grit, curiosity, and sense of shared purpose is what keeps it all moving forward. I left both humbled and re-energized.





Discussion about this post