Skip to content

Feature status

This page distinguishes what the repository implements today from the work still needed for broader production operation.

Area Current capability
Identity registry Source Patient ingestion, UUIDv7 enterprise identities, Person links and canonical Patient survivorship
FHIR R4 Patient create, update, read, search and $match; Person lookup; JSON/XML; ETags and OperationOutcome
HL7v2 MLLP ingestion for ADT A01, A04, A08, A28, A31, A40 and A47
Matching Configurable bounded blocking rounds, field weights, thresholds, identifier certainty and hard-conflict handling
Review Explainable probable-match cases, merge, reject, unlink/split, stale-case detection and dual approval
Tenancy Trusted tenant and source context on every identity, query, receipt, decision and audit event
Storage Development in-memory provider and durable GCP Healthcare API R4 provider behind one contract
Operations Blazor portal, source trust, policy editing, audit search, health checks and OpenTelemetry
Deployment Containers, Compose, Helm, Terraform foundations and a reproducible GCP demo
Capability Status
Online batch re-index after blocking-rule or key changes Not implemented
Scheduled population reconciliation Not implemented
Ground-truth recall and precision reporting Not implemented
Nickname dictionaries and additional comparator libraries Not implemented
Fellegi–Sunter probability calibration Not implemented
Adaptive or trainable ML classification Not implemented
Arbitrary non-patient entity resolution Out of scope
Broad webhook and integration catalogue Not implemented
Certified clinical product status Not claimed

The next safety-critical platform capability is controlled batch re-indexing. Until it exists, changing blocking inputs or authoritative identifier systems for a populated registry requires an offline migration plan.

Read the matching rules for exact behaviour and production readiness for the controls that remain the deploying organisation’s responsibility.