Healthcare integration insights
Standards breakdowns and architecture notes on FHIR, TEFCA, USCDI, HEDIS, clinical data platforms, and medical document pipelines. Written from systems we run in production, not from other articles.
Everything here comes out of the healthcare integration practice: the standards questions clients ask before a build, and the architecture decisions that turned out to matter after one. Each piece links to the production case it was learned on.
TEFCA vs FHIR: what each one actually does
TEFCA is a trust framework for network exchange, FHIR is the API standard the data rides on. How they differ, where they meet, and what you build for each.
USCDI and FHIR: data classes vs the API that moves them
USCDI defines which data elements must be exchangeable, US Core maps them onto FHIR R4 resources. The versions, the mapping, and the gaps engineers hit.
FHIR AuditEvent: an audit trail that survives review
How the FHIR AuditEvent resource works, what an access trail must prove under review, and the patterns that keep PHI out of logs while keeping evidence in.
Clinical data aggregation: architecture that scales
What a clinical data aggregation platform does, the hub pattern that stops every consumer re-integrating every source, and the failure modes to design against.
Clinical data warehouse: design notes from production
What separates a clinical data warehouse from a generic one: history-preserving normalization, idempotent ETL, forward-only migrations, and read replicas.
Medical document OCR: an intake-to-record pipeline
The stages of a medical document OCR pipeline: capture, enhancement, classification, extraction, and routing, with the failure modes each stage exists to stop.
OCR fallback and barcodes in document pipelines
What an OCR fallback path is, when barcodes beat text recognition entirely, and how confidence scoring decides which route a document takes through a pipeline.
HEDIS automation: a pipeline auditors can reproduce
How a HEDIS measurement pipeline automates intake, classification, and measure calculation so numerators and denominators reproduce on demand for NCQA audit.
Healthcare schema markup that earns rich results
The schema.org types that matter on clinical sites: MedicalClinic, Physician, MedicalProcedure, FAQPage, and the implementation details search engines check.
Up against one of these standards right now?
Describe the system and where it hurts. The engineer who answers is the engineer who ships it.