Analytics on FHIR® Recap

On October 30, 2024, Health Samurai, Antidote Solution & HL7 organized the “Analytics on FHIR®” online conference.

This event brought together leading industry experts from HL7, Microsoft, Google, Health Samurai, Mayo Clinic, and CSIRO to discuss SQL on FHIR specification & implementations and the future of Analytics on FHIR.

‍

FHIR Analytics & SQL on FHIR: An Introduction by Arjun Sanyal

Arjun Sanyal. Principal Antidote Solutions.

The slides

A high-level introduction to FHIR Analytics and the SQL on FHIR project touching on the motivations, the prior art, SQL on FHIR, use cases, and where we're headed next.

‍

A Technical Tour of the SQL on FHIR Spec with Ryan Brush

Ryan Brush. Senior Staff Software Engineer at Google

Ryan, a lead editor of SQL on FHIR specifications, will give a deep dive into the technical details of the specifications. This will include a close look at the heart of the spec, the ViewDefinition, and the decisions made during its design.

‍

SQL on FHIR in PostgreSQL. Nikolai Ryzhikov.

Nikolai Ryzhikov. Chief Technology Officer at Health Samurai

The slides

Nikolai will introduce two types of SQL on FHIR View Runners - ETL (in-memory) & ELT (in-database) and demo the ELT implementation for PostgreSQL using Aidbox, open-source ViewDefinition Builder, and Grafana.

‍

Using SQL on FHIR in the HL7 FHIR IG Publisher. Grahame Grieve

Grahame Grieve. Principal Health Intersections Pty Ltd.

The slides

The official HL7 FHIR IG publisher already has a built-in ViewDefinition runner, introduced by Grahame. It can be used to make it easier to generate content on spec pages with SQL queries. This runner is part of the open-source HAPI framework.

‍

Pathling: SQL on FHIR views for Apache Spark. John Grimes.

John Grimes. Principal Research Consultant ‍CSIRO

The slides

John presents Pathling, Apache Spark implementation of ViewDefinition runner.  John shows demo of real-time analytics using FHIR subscriptions, Pathling and Spark streaming, with live-dashboards on PostgreSQL & Superset.

‍

FlatQuack: FHIR resources to SQL tables with DuckDB. Dan Gottlieb

Dan Gottlieb. Principal, Data standards ‍Central Square Solutions

The slides

FlatQuack is a new, open source, View Definition ELT runner that targets DuckDB, a lightweight, high performance database engine. By compiling to SQL, FlatQuack can be easily integrated into existing data pipelines that use off-the-shelf orchestration tools like DBT and Apache Airflow.

‍

An Open Approach for Translating FHIR to OMOP. Carl Anderson.

Carl Anderson. Principal Integration Engineer. Mayo Clinic

Carl will share the history and implementation details of FHIR to OMOP  project and will discuss challenges and solutions.

‍

Transforming FHIR for efficient population queries using Open Health Stack. Bashir Sadjad.

Bashir Sadjad. Software Engineer at ‍Google

The slides

Bashir will discuss the evolution of the Google Open Health Stack  (OHS)  Project toward SQL on FHIR v2. OHS can continuously monitor data in an FHIR server and transform it into analytics-friendly formats for population-level queries.

‍

Safhire: Rust-y Ducklings in a Row. Kiran Ayyagari

Kiran Ayyagari. Freelance programmer, Sereen Systems

The slides

Safhire is an analytics engine based on SQL on FHIR specification written in Rust. Processed FHIR data is stored in an embedded DuckDB instance and can be queried using REST API.

‍

What’s next for SQL on FHIR? Roundtable.

Questions:

• What is the best way for folks who want to get started with SQL on FHIR?
• How to start working with SQL on FHIR?
• Demo of SQL on FHIR JavaScript implementation
• What is the SQL on FHIR group thinking about the next steps of development?
• What's next for SQL on FHIR in terms of becoming an official HL7 standard?