Skip to content
View clinical-data-etl's full-sized avatar

Block or report clinical-data-etl

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
clinical-data-etl/README.md

clinical-data.org — Healthcare FHIR/HL7 Data Parsing & Clinical ETL Pipelines

clinical-data.org

Production engineering patterns for healthcare data pipelines.
FHIR R4 & HL7 v2 parsing · clinical ETL · terminology resolution · HIPAA de-identification & patient matching.

🌐 Visit the site → www.clinical-data.org


What this is

clinical-data.org is a deep, production-focused engineering reference for the teams who move clinical data for a living — health-tech engineers, clinical data scientists, ETL developers, and compliance teams. Every page is written for real systems: deterministic patterns, concrete failure modes, explicit type contracts, audit-ready logging, and runnable Python.

It answers the questions that break pipelines in production: how to tokenize a vendor-mangled HL7 v2 stream without dropping messages, how to resolve SNOMED CT to ICD-10 through a version-pinned decision table, how to keep a FHIR bulk export idempotent, how to de-identify a longitudinal cohort without fragmenting a patient, and how to match records across systems that share no common identifier.

The three knowledge areas

The standards contract end to end — HL7 v2 segment grammar and MLLP transport, FHIR R4 resource modelling, REST vs Bulk Data export, terminology-server integration, US Core conformance, SNOMED/LOINC mapping, and transaction-bundle referential integrity.

Turning raw clinical telemetry into trusted assets — Python-native parsing, type coercion, async batch processing, idempotent loads, data-quality validation frameworks, and orchestration with Airflow, Prefect, Spark, and dbt.

The identity boundary — Safe Harbor vs Expert Determination, removing the 18 identifiers, date-shifting, k-anonymity, deterministic and probabilistic (Fellegi–Sunter) patient matching, master patient index strategies, and tamper-evident PHI audit logging.

Why engineers use it

  • Runnable, not hand-wavy. Every implementation section ships correct, idiomatic Python you can lift into a service and test in isolation.
  • Standards-accurate. FHIR R4 (4.0.1), US Core, HL7 v2 encoding rules, UCUM/LOINC/SNOMED CT, and the HIPAA Privacy and Security Rules are treated as first-class facts, not afterthoughts.
  • Compliance built in. PHI handling, minimum-necessary, audit trails, and re-identification risk are engineered into every layer rather than bolted on.
  • Deeply interlinked. Hand-authored diagrams, spec tables, and a tight cross-reference graph make the whole domain navigable in one or two clicks.

Tech

Static site built with Eleventy, hand-authored inline SVG diagrams, structured data (JSON-LD TechArticle, BreadcrumbList, FAQPage), a PWA service worker, and full WCAG 2 AA accessibility. Deployed on Cloudflare.

Contributing & feedback

Spotted an error, or want a topic covered? Open an issue — corrections to the clinical, standards, or compliance details are especially welcome.


www.clinical-data.org

Popular repositories Loading

  1. clinical-data-etl clinical-data-etl Public

    Production engineering patterns for healthcare FHIR/HL7 data parsing, clinical ETL, and HIPAA de-identification & patient matching.

    JavaScript