Storage & Analytics in San-Antonio

We finished #3 Storage & Analytics track during FHIR connectathon in San-Antonio!

The team was about 6 active participants with visitors from other tracks: Nikolai Ryzhikov (Health Samurai), Toby Hu (Google), Brian Wright (Mayo Clinic), Alex Zautke(Firely), Shantha (Accenture) and Stefan Lang (QHUL).

Brian Wright did a great presentation about Analytics on FHIR project in Mayo Clinic. There is a hope that we will publish his post about this project here — stay alert!

One of the hot topics was getting HL7v2 feed into FHIR-aware database for analytics. HL7v2 has a deep adoption and can serve as a robust source of data!

All participants agree that schema/storage representation of FHIR resources should be tuned for database and we need back and forth transformations from standard representation (in JSON). I’m personally interested in bidirectional transformation language and started my investigation and experiments (if you interested about — ping me).

Many people interested in real-time analytics and we definitely have to collaborate with FHIR Subscriptions group.

Toby Hu reported that Google guys are going to take SQL on FHIR spec seriously and we will do the same from the fhirbase side. We updated GitHub issues for SQL on FHIR after track discussions and you are welcome to contribute!

To chase the ultimate goal of portable SQL on FHIR over implementations we decided work on a common dataset and reference SQL queries, which can be used to test implementation for spec conformance.

See you in zulip or on next connectathon in Quebec!