Last Updated on
Last week, I attended the HL7 FHIR DevDays Conferences hosted at Microsoft’s headquarters in Redmond, Washington. Overall, this was a great conference that was well worth attending. I walked away feeling inspired and hopeful for the future of health IT interoperability. There is a community of amazing talent, working hard to make a real difference in how patient records are shared between trusted applications; both for patient and clinician access.
This was a global event – I met people from all over the US, Netherlands, Russia, New Zealand, Canada, and Germany – just to name a few. Several vendors did live demos and hands-on coding sessions, which were fabulous to see.
I attended several sessions focused on successful uses of FHIR. Most notably, I enjoyed listening to both Steve Posnack and Grahame Grieve. Steve Posnack is a dynamic presence – an articulate speaker and very knowledgeable about IT. His position at the ONC is going to be pivotal with moving legislation forward, which offers a clear path for removing legacy barriers to innovation. Grahame Grieve is equally polished and a rockstar of sorts in health IT. He is doing God’s work and has the respect of the entire health IT community (which is a feat in itself).
This Server is on FHIR
It seems that everyone has built a FHIR server. Not a terrible thing, but it was surprising to see most companies decided to build their own server rather than using the open source versions that are available today. All servers have similar characteristics, namely, a database used to persist records which are then queryable through FHIR resources.
I did not see any companies describe their databases in a way that traverses through a FHIR resource to store discrete data elements in separate fields. Rather, the database descriptions were more along the lines of taking JSON structures and storing the JSON. Those JSON resources are available through traditional relation database scripting so information is quickly obtained through indexes created as a result of the resource elements themselves. An approach I’ve discussed with our Chief Architect, Chong Yu, in which we came to the same conclusion that the design pattern is a good thing. We applaud the work done thus far.
Many people ask me or my team why we didn’t implement a FHIR model from Sansoro Health’s beginning. Back in early 2014, we evaluated the state of FHIR resources, their maturity against what we set out to accomplish, and we concluded there were many uses cases not covered by FHIR. We built our own API data models to cover use cases that we knew were not yet supported. I often said, “We will control the data that we access, when we want to access it. We let consumers and app developers dictate the data they want, when they need the data and for what use case.” Our theory, at the time, was once FHIR is more mature and clearly the broadly accepted standard for interoperability, we would take our great work and do a simple transform from our model to FHIR.
That day is now upon us.
All FHIR-ed Up
For Sansoro Health, this will result in a FHIR platform that normalizes and standardizes our current capabilities into FHIR resources that are consistent, regardless of the source system. No need to deal with all the variants of FHIR implementations based on how different vendors implemented or the profiles they decided (or neglected) to support. We will still control the outward facing specification, the ability to get the data we want and provide a consistent, rigid implementation, which will remain our unfair advantage. As we identify data elements, we support outside a normative FHIR resource specification, we will take advantage of extensions to make that data element available. In other words, there are no data gaps.
The Future is So Bright, I Have to Wear Shades
As I stated at the beginning, I am very excited about the future of health IT. I met with, heard from, and have seen enough to realize the dark days of inept interoperability are crumbling. The health IT community is full of bright, motivated engineers who will make interoperability work. There is no other option.
It is time for health care to move forward. Embrace innovation and start doing the cool things in health IT that should have happened years ago. Those that stand in the way will either evolve or lose market share and fade away. That simple. We are very close to a new age in Health IT. I won’t be content until I see it (and neither should you) …