Unleashing the power of FHIR data with knowledge graphs
summary
A project loads synthetic FHIR R4 JSON into Neo4j + APOC, creating about 1.8 million nodes and 10 million relationships to represent resources, codes and explicit references. The graph makes multi-resource queries (claims→encounters→codes, patient timelines around first opioid prescriptions) far more direct than many SQL joins, supports visual exploration with Neo4j Bloom, and proves conceptually promising though no full-scale performance benchmarking is completed.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 0:55Agenda
- 4:10FHIR Resources
- 6:00Knowledge Graphs
- 10:25Searching in a graph
- 11:30The opioid crisis
- 12:30Building relationships
- 13:15Demo
- 16:00Questions
