My first FHIR project should not have been a FHIR project.
It stored healthcare data but once ingested that data was never shared with any other systems or apps.
There was no interoperability requirement.
The data and project had an 18 month life span before everything would be purged.
There was no regulatory requirement in this case to use FHIR.
Users were already authenticated by existing single-sign-on methods before reaching the app and had full access to all data.
There was no Smart-on-FHIR requirement or use case.
A couple of months in to the project I was on a call with a team of devs (no business or product people) and I asked the question:
Why are we using FHIR?
It took me that long to get my head around FHIR and to be confident enough that my concerns were valid.
The devs on the call had no answer that made sense.
- Other projects in the business were using FHIR
- Everyone’s talking about FHIR now
- We use Azure for everything else
- “FHIR is the future!”
All answers to a degree, but none of them made business or technical sense.
The project and its data requirements did not need or benefit from FHIR. Using it was slowing them down. A simple SQL database with a schema that matched the data would have been better.
It was a terrible introduction to FHIR, but on the flip side a beneficial one for me. I saw the bad before the good in FHIR.
A year ago I was on a call with a startup in Ireland who were at the early stages of building a local solution in the elderly care space.
The first question their tech lead had for me was:
Should we build a FHIR façade or use a managed FHIR server on AWS?
He’d been doing a lot of Googling and a lot of reading.
After digging a little I learned that none of the organisations they were sharing data with understood FHIR.
FHIR did not exist in Ireland at that time outside dev teams working on US projects. (Still doesn’t, but that’s a disappointing post for another day.)
By the end of our call they decided that the project would not benefit from FHIR at all.
They were right.
FHIR should never be the default answer for storing and sharing healthcare data. Sometimes it’s the right answer. Sometimes it’s not.
Be careful of people selling you FHIR as the solution to all problems.
---