Composable Healthcare
summary
A Clojure + Datomic + React system reads FHIR and UMLS concepts, runs a rule engine that materializes suggestions (diagnoses, weight‑based amoxicillin dosing, orders, billing codes) into encounter pages in real time; selecting a suggested diagnosis cascades prescriptions and checklist items. Backend is ~4,000 LOC, frontend ~20,000 LOC, and rules are content‑addressed so identical rules deduplicate. Demonstration works live with optimistic commits and subscriptions. Key pieces: Composable rules authored in-app, Storefront client with Datomic-like API.
AI-written summary · about
chapters (15)
- 0:00Start
- 0:19Agenda
- 0:41Health Information Problem
- 3:27The Medical Record
- 3:49Encounter Notes
- 7:26Medical Scribes
- 9:46Augmenting the Human Intellect
- 11:03Composable Healthcare
- 12:51Essential Complexity
- 14:41Central State
- 27:58Strep Test
- 28:34The Unified Medical Language System
- 29:38Prototypical Rules
- 30:14Feedback Loop
- 34:09Unified Medical Language System
related talks
52mFull stack FP in Datomic, Clojure, and ClojureScript
42mA Practical Functional Relational Architecture
43mScaling Full-Stack Applications
26mClojure vs Sepsis: path to real time enterprise data science
38m"Growing Data Center networking mgmt UI using ClojureScript, Reagent and re-frame"
44mFrom REST to CQRS with Clojure, Kafka, & Datomic
35mMaterial UI on FHIR
35mIt's Just Data