This week, I’m sharing a special conversation I had with Aaron Neiderhiser of Tuva Health and Phil Ballentine from Atropos Health (together they host theHigh Dimensional Health Data Podcast) . We’ve been trying to schedule this for months, and I’m glad we finally hit record because we got right into the weeds of why healthcare analytics is still so painful and how we are fixing it.
If you’ve followed my work, you know I’m obsessed with the intersection of FHIR and Analytics. But there is often a disconnect between the “high priests” of standards and the data engineers trying to run SQL queries in relational databases.
In this episode, we bridge that gap.
We discuss:
* The SQL on FHIR Conference: Why the industry is standardizing “View Definitions” to make FHIR portable across different data warehouses.
* CQL vs. SQL: Why Clinical Quality Language is great for requirements, but SQL (aided by AI) is another way to execute at scale.
* The “Dump Truck” Theory: Why APIs aren’t always the answer for analytics teams who just want bulk data access.
* Automating Interop: How I’m using AI Agents to do the grunt work of mapping and flattening data, so you don’t have to wait on “Larry” (the fictional, overworked hospital IT admin) to run a report for you.
It was a blast recording with the High Dimensional team they bring a refreshing, “not just drinking the FHIR Kool-Aid ” reality to the conversation that is often missing from standards discussions.
Enjoy the listen!
Gene
Spicy Takes
* The “Windows ME” of Standards: We discussed how the industry seems to be collectively skipping FHIR R5 (like Windows 95/98 to 2000) and waiting for FHIR R6 to become the normative standard, avoiding the “trough of disillusionment” associated with intermediate versions.
* The “Larry” Bottleneck: We personified the interoperability problem as “Larry” the overworked IT guy at the health system who smokes two packs a day, takes three lunches, and ignores your ticket for a claims feed. The goal of Bulk FHIR is to bypass the “Larry problem,” but we aren’t quite there yet.
* Phil’s Hot Take: In a moment of pure chaos, Phil claimed Windows ME was the best operating system ever made. (We agreed to disagree).
Fun Facts & Why You Should Listen
* The Portuguese Connection: I shared a story about meeting Grahame Grieve (the “Father of FHIR”) at a FHIR camp in Portugal, highlighting how the “open” nature of the standard was intentional, even if it creates headaches for data modelers today.
* Real-Time Solutions: We discuss actual tools (like the Tuva Project and Health Samurai) that are solving these problems now, not just theoretically.
* Why Listen: If you have ever stared at a nested JSON object and wondered how to get it into a dashboard without crying, this episode explains exactly how the industry is solving that problem.
