In the early years of FHIR, evangelism was the order of the day. Sharing patient data, improving patient outcomes, working together to get healthcare data right at last.
To some extent that evangelism still exists, but what I’m seeing lately is a more “clinical” and business approach to FHIR. Especially when it comes to private companies.
Yes, we’re all for improved patient outcomes. But what’s the business case?
On many of my introductory calls with tech and product leaders I hear the same things.
“Regulation and AI.”
Regulations: “We have to.”
More and more European companies cite regulations as the nudge that tipped them over the edge. They were familiar with FHIR for many years but had made no major progress in adopting it. “Upcoming” regulations were not always enough and there was no extra pressure to make big changes.
EHDS was on the horizon, as well as more country specific regulations that mandated FHIR use. There was no escaping FHIR but companies were still deferring the work.
What changed?
AI: “We need to.”
While regulations may have been the spark, AI is truly igniting the fire.
Most private businesses who consume or create healthcare data are sitting on top of large legacy systems filled with valuable data. Often it’s hard to get to, siloed away in one database or data warehouse after another.
But its value has grown on the back of its potential use by AI.
Big business has been talking about Big Data for decades. It rarely moved beyond talk.
Not any more.
All that legacy data is now fuel to the AI fire.
The push to AI is being felt from board level right down to engineering. If there’s an AI use case, the money will be found to work on it.
While the initial impetus for investing in FHIR often begins with regulation, the potential for introducing value through AI is how mid-sized FHIR projects turn into massive, company wide data transformation projects.
I’ve sat in on meetings where “AI first, regulations second” was the message.
A far cry from the community evangelism of 10 years ago.
What do these projects look like?
They start small, targeted with meeting the needs of upcoming regulations. This usually means FHIR in one form or another, alongside investment in existing data cataloging.
What do we have, where is it, how do we get it out?
The end result of this first stage is often a single FHIR API or FHIR server, populated with data from existing systems as required by specific regulations.
The next stage involves pulling in more data, all standardized to FHIR and using recognized terminologies.
Once the legacy data has been transformed it can be pushed out of FHIR in various formats – either “as is” or anonymized – and into different data stores.
From there it’s accessible to AI in formats that it can understand and work with.
The message here is that FHIR is no longer optional.
Businesses who see FHIR as just a tool to enable compliance are missing out. FHIR is the perfect linchpin to enable larger AI projects.
Don’t look on FHIR compliance as the end goal. Treat it as a foundation on which to build better data and better access to your data.
---