Healthcare & Operations · production
Enterprise data migration for NHS PAS & EPR upgrades.
OpenLI Data Migration is a contract-driven ETL platform built for the most complex enterprise data migrations in healthcare. Migrate hundreds of source tables and thousands of fields from legacy NHS Patient Administration Systems (PAS) and Electronic Patient Record (EPR) platforms to modern targets — with semantic mapping, quality gates, release gates and an immutable audit trail.
Built on the OpenLI Codex foundation. Same governance, RBAC, and dual-runner architecture as the rest of the OpenLI portfolio. Pairs naturally with OpenLI TIE Migration for full system upgrade programmes.
v0.2.5 production
Contract-driven ETL
Quality gates · release gates
Whitelabel-ready
dm.openli.ai
Why contract-driven?
Most data migration tools are connector-driven: you pick a source, you pick a target, you write transforms. That works for simple migrations. It does not work for NHS PAS and EPR migrations where 5,000+ fields need to land correctly, where every mapping decision needs to be reviewable, and where regulators expect a paper trail of every row that moved.
OpenLI Data Migration is contract-driven. The contract between source and target is a first-class artefact: versioned, reviewable, signed off by the integration lead, and enforced by the runtime. Quality gates fail the migration if the contract is violated. Release gates ensure no row reaches production without passing the contract in pre-production and staging first.
Contract-first ETL
Source schema, target schema, semantic mapping and validation rules are all part of one versioned contract. The runtime enforces the contract on every row that moves.
Quality gates
Configurable quality KPIs (completeness, accuracy, referential integrity, code translation coverage). The migration fails if the gate threshold is not met — no surprise data corruption in production.
Release gates
Three-tier release gating (pre_production, staging, production). Each tier has its own approval requirements. Promotion to the next tier is a deliberate human action, not an automatic event.
Capabilities (v0.2.5, in production today)
Schema extraction & mapping
Auto-extract source schema from CSV, IRIS Emulator, PostgreSQL Emulator or JSON test sources. Map 417 source tables and 5,387 fields to 38 target LOAD tables and 880 fields with semantic mapping.
Semantic mapping workbench
Versioned mapping definitions with reviewable diffs. Each mapping rule is annotated with the clinical or operational rationale. Mapping changes go through approval workflows.
Crosswalk management
Code translation dictionaries managed as first-class artefacts. ICD-10 to SNOMED. NHS Number formats. Local clinical codes to national vocabularies. All versioned, all auditable.
Mock data generation
Generate realistic synthetic test data (20 to 20K rows, seeded for reproducibility). Test the migration end-to-end before touching real patient data.
Audit explorer
Immutable event trail of every row that moved, every mapping rule that fired, every quality gate decision. Exportable for regulatory submission.
Multi-tenant org/workspace/project
Org → workspace → project hierarchy with full data isolation. One Data Migration deployment can serve a managed-service operator running multiple Trust migrations in parallel.
Enterprise RBAC
Role-permission matrix with user lifecycle controls (lock, unlock, session reset). Audit-logged on every privileged action.
Deployment verification
Healthchecks and smoke tests baked into the deployment pipeline. Migrations promote between gates only when verification passes.
Built on OpenLI Codex
Inherits the dual-runner architecture, audit framework and policy hooks from the OpenLI Codex foundation. Same security review covers OpenLI Data Migration plus every other OpenLI product.
How a migration runs
1. Define the contract
Source schema, target schema, semantic mapping rules, code translation dictionaries and quality KPIs are captured as a versioned contract. Reviewable as a diff.
2. Test on synthetic data
Generate seeded synthetic data, run the contract end-to-end, inspect the audit trail. Iterate the contract until quality gates pass.
3. Promote through release gates
Pre-production → staging → production. Each gate has its own approvers. Each promotion is audit-logged. No row reaches production without passing all gates.
4. Run on real data
Execute against real source data. Quality gates fail the migration if the contract is violated — no silent corruption. Audit trail captures every row that moved.
5. Forensics & export
Audit explorer surfaces every event. Export the audit trail for the regulator, for the clinical safety officer, or for the post-migration health-check.
Cross-product
Pair with OpenLI TIE Migration for the integration engine layer of the upgrade, and with OpenLI HIE for the post-migration interface estate.
Try the live evaluation portal
The OpenLI Data Migration evaluation portal is hosted at dm.openli.ai.
Request access to evaluate the platform with the included synthetic dataset.
FOUNDATION
Built on OpenLI Codex
Dual-runner architecture, governance framework, audit trail and prompt template library all come from the OpenLI Codex foundation. One security review covers Data Migration plus every other OpenLI product your organisation adopts.
NHS PAS & EPR DATA MIGRATION
Ready to run an enterprise data migration with full audit trail?
Open the live portal, view the source on GitHub, or request a guided evaluation with the included synthetic dataset.