Blaze: A High-Performance Open Source FHIR Server with Embedded CQL Evaluation Engine
Abstract
Blaze is an open-source, high-performance FHIR server with an embedded Clinical Quality Language (CQL) engine, initially developed to support large-scale federated feasibility queries in the German Cancer Consortium (DKTK) and later extended in the German Medical Informatics Initiative (MII) and other research consortia. Built in Clojure, Blaze uses a functional, immutable database architecture based on RocksDB to enable lock-free concurrency, full resource history, and efficient query execution. Our benchmarks on datasets exceeding one billion FHIR resources show that it supports cohort queries over millions of patients in under one minute. Deployed in production at over 60 national and international research sites, Blaze demonstrates that a functional, embedded database design can achieve scalable, standards-compliant performance for clinical research data networks.
Fields of Study & Subjects
Machine Learning in Healthcare
Electronic Health Records Systems
Data Quality and Management
Publication details
- Journal / source
- Studies in health technology and informatics
- Publisher
- IOS Press
- Volume
- 336
- Pages
- 1481–1485
- Publication date
- 21 May 2026
- Language
- EN
- Document type
- Book chapter