12-year-old EHR retired. 100% data migrated. $180K annual IT cost eliminated.
How a 22-branch home health agency safely escaped a legacy system holding back their entire AI roadmap — with a 6-month parallel run, zero data loss, and zero disruption to clinical operations at any point.
"We had been putting off this migration for years because we were scared of what could go wrong. Metricoid made it feel completely safe. The parallel run meant we could see everything working correctly before we committed. We have never looked back."
18 years of patient care data trapped in a system no one dared touch
The same 22-branch home health agency from CS-04. After the EHR integration project connected the legacy system to modern tools in 12 weeks, the organization had proven the concept — AI-assisted field documentation, automated OASIS submission, and payer portal connectivity were all working. The integration layer had delivered the clinical outcomes they needed.
But the legacy EHR itself remained a growing liability. The vendor had announced it was winding down the product line. Hardware was aging toward end-of-life. The $180,000 annual cost of maintaining on-premise infrastructure was hard to justify for a system that was being sunset. And every AI initiative the organization wanted to pursue — predictive analytics, remote patient monitoring, advanced OASIS AI — was blocked by a database format no modern system could read natively.
The integration layer from CS-04 had given them the confidence that modern systems worked. Now they needed to make the move permanent — and do it without risking 12 years of patient records that contained 340,000+ clinical episodes from hundreds of thousands of Medicare home health patients.
CS-04 covers the EHR integration project that preceded this migration — connecting the legacy system to mobile field documentation, automated OASIS submission, and payer portals in 12 weeks. This migration built on the confidence generated by that success.
Read CS-04 — EHR Integration →Three years of conversations. Zero action. The same fear.
Leadership had been discussing the migration for three years. Every conversation reached the same impasse: the organization had 340,000+ patient episodes in a proprietary binary database format with no documented export procedure. The legacy vendor confirmed they had no migration tool. Any extraction would require custom database engineering on undocumented schemas.
The stakes were high. If the migration failed — if records were lost, corrupted, or incompletely transferred — the consequences would be severe. Medicare audit exposure from incomplete clinical records. HIPAA liability from lost patient data. Potential inability to bill for completed care. And a clinical team with no access to historical patient information for ongoing patient management.
These fears were not irrational. They were exactly the risks that poorly executed migrations generate. The issue was not that migration was inherently dangerous — it was that the organization had never been shown a migration approach that systematically eliminated each risk before cutover.
Every year on the legacy system was another year of accumulating risk — aging hardware, unsupported software, growing security vulnerabilities, and $180K in IT costs that could fund actual improvements. The migration risk was finite and manageable. The inaction risk compounded annually.
12 years of data in a proprietary format
340,000+ patient episodes in a binary database format with no documented export procedure and no vendor-provided migration tool.
Active Medicare billing cannot be interrupted
The organization was billing Medicare for active home health episodes continuously. Any interruption in the billing system would delay payments and potentially put revenue at risk.
HIPAA liability for data loss or breach
Lost patient records would constitute a HIPAA breach. Incomplete clinical records could create Medicare audit exposure for historical claims.
Clinical team disruption during transition
330 clinical staff across 22 branches needed to transition to a new system without any gap in their ability to document and deliver patient care.
Six phases that eliminate every risk before the cutover happens.
Every fear leadership had was legitimate. Our response was not to dismiss those fears — it was to design a migration process that systematically addressed each one before any cutover commitment was made.
System audit & schema reverse-engineering
The legacy EHR database schema had no documentation. Our database engineers connected directly to the on-premise server, reverse-engineered the full schema across all tables, documented every field and relationship, and produced a complete field-by-field mapping from the legacy format to the modern target system. No migration work begins without this map.
New platform build & configuration
The modern cloud-native platform was built and fully configured in a staging environment before any patient data was touched. Clinical workflows, OASIS forms, billing configurations, and payer integrations were all validated against synthetic test data. The new system was certified ready before the first real patient record was imported.
Encrypted data extraction & migration
All 340,000+ patient episodes extracted from the legacy database using the reverse-engineered schema. Every field transformed to the modern format using the documented mapping rules. Extraction performed over an encrypted connection — no unencrypted PHI at any point. Full audit log of every record extracted, transformed, and loaded.
Dual validation — automated + manual audit
Every migrated record was validated twice: first by an automated checksum process comparing source and target record counts, field populations, and critical value consistency; then by a manual audit team spot-checking 5% of records (17,000+ episodes) across all record types, all branches, and all time periods. Migration did not close until both validation passes confirmed 100% accuracy.
6-month parallel run — both systems fully live
The most important phase. Both systems ran simultaneously for 6 months. Clinical staff used the new system for real patient documentation. The legacy system remained fully operational and writable as the safety net. Data entered in either system was reconciled nightly. Leadership could see the new system working in real clinical use — with the legacy system available as an immediate fallback — before any cutover commitment was made.
Cutover on the organization's timeline
When leadership decided they were confident — not when our schedule said — the cutover happened. Clean transition on a Monday morning. Legacy system retained in read-only mode for 90 days before decommissioning. On-premise hardware retired. $180K annual IT cost elimination realized. The organization chose the date; we never pushed.
340,000+ episodes. Every field mapped. Validated twice.
Data migration is where most healthcare modernizations fail. A missed field, a format mismatch, or an incomplete record can surface as a clinical or compliance problem months after cutover — when tracing it back to the migration is difficult and fixing it is expensive. Our protocol eliminates this risk through systematic mapping and dual validation.
Schema reverse-engineering
Full database schema documented — all 847 tables, all fields, all foreign key relationships, all legacy-specific data types and encoding formats. Four weeks of database engineering before a single record was touched.
Field-by-field transformation mapping
Every legacy field mapped to its modern equivalent with documented transformation rules — handling legacy date formats, proprietary code tables, custom fields, and historical records that predate current data standards.
Encrypted extraction & load
Full database extracted over AES-256 encrypted connection. Transformation applied. All 340,000+ episodes and 2.8M visit records loaded to the new system. Full audit log with timestamps for every record.
Automated checksum validation
Record counts, field population rates, and critical value consistency (patient IDs, episode dates, clinical codes) compared between source and target. Every discrepancy flagged and resolved before proceeding.
Manual spot-check audit
17,000+ records (5% sample) manually reviewed by auditors — comparing source screenshots against target records field by field across all record types, all branches, and all time periods back to 2009.
Migration closes only when both automated validation and manual spot-check audit confirm 100% accuracy.
Why 6 months of running both systems was the right call
The parallel run is the single most important risk mitigation in any healthcare system migration. It is also the phase that most migration vendors try to shorten — because it costs them time. We never shorten it. The organization needed 6 months to build confidence. We ran both systems for 6 months.
New system live. Legacy still operational.
Both systems fully writable. Clinical staff begin using new system at 4 pilot branches while all other branches continue on legacy. Data entered in either system reconciled nightly.
12 branches on new system. Confidence growing.
Pilot branches report high satisfaction with mobile documentation. OASIS acceptance at 94% from new system. Rollout extended to 12 of 22 branches. Legacy remains the fallback.
All 22 branches on new system. Legacy still live.
Full network operating on new system. Legacy retained as fallback. Leadership watching billing, OASIS acceptance, and clinical workflow metrics daily. All KPIs meeting or exceeding targets.
Leadership calls the cutover date.
CEO and COO review 5 months of parallel run data. Billing uninterrupted. OASIS acceptance at 94%. Staff satisfaction high. Rollback plan tested and documented. Cutover date chosen: the following Monday morning.
Clean transition. Legacy retired.
Monday morning cutover. Legacy system moved to read-only mode. On-premise infrastructure retired over the following 60 days. $180K annual IT cost elimination realized.
Our parallel run promise
These are the commitments we make to every organization before the parallel run begins — in writing.
We never pressure you to cut over before you are ready — even if the parallel run extends well beyond the original plan
The legacy system remains fully operational and writable throughout the entire parallel run — not read-only, not degraded
Data entered in either system during the parallel run is reconciled — no data loss from operating two live systems simultaneously
Rollback plan documented, tested, and on standby before parallel run begins — if something unexpected happens, you have a proven path back
We are on-call for immediate response throughout the parallel run — issues resolved within hours, not ticket queues
You choose the cutover date — we do not set a deadline that pressures a decision before confidence is established
"The parallel run was what made the difference. After 5 months of seeing the new system work perfectly in real clinical use, the cutover felt like the obvious and safe thing to do — not a risk we were taking."
— CEO, Home Health Agency, Mid-Atlantic USSeven deliverables. Modern platform. 12 years of history preserved.
The migration was not just a data transfer — it was a full platform build. The new system was purpose-built for home health operations, mobile-first, and AI-ready from day one.
Custom database extraction & transformation tool
FOUNDATIONBuilt specifically for this legacy EHR's proprietary schema — a custom extraction tool that reverse-engineered the undocumented database format and converted 340,000+ patient episodes to modern FHIR-compatible data structures. No off-the-shelf migration tool could handle this schema; it required original database engineering.
Cloud-native EHR on AWS US-East
INFRASTRUCTUREHIPAA-eligible AWS infrastructure with AES-256 encryption at rest, TLS 1.3 in transit, 99.9% uptime SLA, automated backups, and multi-AZ deployment. First time the organization's clinical data was hosted on infrastructure with a formal uptime guarantee and active security monitoring.
iOS and Android mobile documentation app
MOBILE-FIRSTPurpose-built for home health field workflows — visit notes, OASIS forms, vital signs, wound assessments, and medication reconciliation. Offline mode stores data locally and syncs when connectivity restores. GPS timestamping. Voice-to-text for faster note entry. Designed for use in patient homes, not offices.
FHIR R4 API layer
AI-READYThe new platform exposes full FHIR R4 APIs — Patient, Encounter, DocumentReference, Observation, Claim, and ServiceRequest resources — enabling every AI tool, analytics platform, and payer integration the organization wants to deploy in the future. The AI roadmap that was blocked by the legacy system is now completely open.
CMS iQIES & payer integrations
REVENUE CYCLEAll integrations from CS-04 — CMS iQIES for OASIS submission, Availity for eligibility, Waystar for claims — rebuilt on the new platform. PDGM grouping automation preserved and enhanced with direct clinical data linkage. No manual billing steps for standard encounters.
Dual validation migration protocol
DATA INTEGRITYThe process that guaranteed 100% data accuracy — automated checksum validation across all 340,000+ episodes, followed by 17,000+ record manual spot-check audit across all branches and all time periods. The protocol that gave leadership the confidence to proceed with cutover.
6-month parallel run infrastructure
RISK ELIMINATIONThe nightly reconciliation system that synchronized data between both systems during the parallel run — ensuring records entered in either system were reflected in both. The safety net that allowed 6 months of real-world testing without any data loss risk. Decommissioned cleanly on cutover day.
10 months from kickoff to legacy decommissioned.
System Audit
Reverse-engineered legacy EHR database schema — 847 tables, all fields, all relationships. Produced complete field-by-field mapping to modern target format. Identified all data transformation requirements including legacy-specific codes, proprietary date formats, and historical records. BAA executed. Rollback plan documented.
Platform Build
Cloud-native EHR provisioned on AWS US-East. Mobile app built for iOS and Android. OASIS validation engine configured against current CMS data specifications. CMS iQIES, Availity, and Waystar integrations rebuilt on new platform. All workflows tested against synthetic patient data. Platform certified ready before any real data migration.
Migration & Validation
All 340,000+ episodes and 2.8 million visit records extracted from legacy database using encrypted connection. Transformation applied using documented field mapping. Full audit log generated. Automated checksum validation: 100% pass. Manual spot-check audit of 17,000+ records: zero discrepancies found. Migration formally closed — confirmed complete.
Parallel Run Begins
Both systems live simultaneously. Pilot deployment to 4 branches. Clinical staff trained on mobile app. New system used for real patient documentation. Legacy system remains fully operational as fallback. Nightly reconciliation active. Leadership KPI dashboard live showing parallel performance comparison.
Parallel Expansion
Progressive rollout to all 22 branches over months 5–7. All branches fully operational on new system by month 7 while legacy remains live. Months 7–9: full parallel operation — both systems writable, all data reconciled. Leadership accumulates 5 months of real-world performance data before making cutover decision.
Cutover & Decommission
Monday morning cutover on organization-chosen date. Legacy EHR moved to read-only mode. All active billing, OASIS submissions, and clinical documentation routing through new system exclusively. On-premise hardware decommissioned over following 60 days. $180,000 annual IT cost elimination effective immediately.
Every feared outcome was prevented. Every metric improved.
The organization's specific fears going into this migration — data loss, billing disruption, clinical staff inability to work — were all prevented. Not by luck, but by a migration architecture that addressed each risk systematically before it could materialize.
The financial outcome — $180K annual IT cost elimination — was the most straightforward to calculate. The strategic outcome — a FHIR R4 platform with full AI-readiness — was harder to value but more important. The organization went from having an AI roadmap blocked by an incompatible legacy system to having an open platform that could support predictive analytics, remote patient monitoring, and advanced documentation AI. That transition is what CS-04 and CS-05 together were actually building toward.
All outcomes measured at 30 days and 90 days post-cutover. Data migration accuracy confirmed by dual validation before parallel run began. IT cost reduction calculated from actual infrastructure decommissioning and licensing termination. Zero data loss confirmed by ongoing audit of historical record accessibility in new system.
Patient episodes migrated — 340,000+ records, dual validated, zero discrepancies
Annual IT cost eliminated — hardware, licensing, and support from decommissioned infrastructure
Clinical disruption during entire 10-month migration — operations continued normally throughout
Modern AI-ready platform — every future AI initiative now unblocked and connectable
The AI roadmap that was blocked for years — now fully open
Once the legacy system was retired and the FHIR R4 platform was live, the organization could pursue the AI initiatives that had been impossible on the legacy EHR.
Advanced OASIS AI validation
OASIS validation AI trained on the organization's own CMS rejection history — going beyond the standard CMS rules to catch the specific patterns that generated rejections unique to this organization's clinical team.
Predictive readmission risk
ML model trained on the 12 years of historical patient data now accessible in FHIR format — predicting 30-day hospital readmission risk for current patients based on patterns from historical episodes.
PDGM optimization analytics
Analytics layer identifying PDGM revenue optimization opportunities — clinical documentation patterns that could support higher payment groupings without changing care delivery, just capturing it more accurately.
Built for Medicare home health's regulatory requirements — from scratch
HIPAA
Full Privacy & Security Rule
BAA Executed
Before any data moved
HITECH
Breach notification compliance
AES-256 + TLS 1.3
Encryption at rest & transit
FHIR R4
Modern interoperability
OASIS Compliant
CMS data specification rules
Full Audit Log
Every migration step logged
US-Hosted Only
AWS US-East, HIPAA-eligible
Rollback Plan
Documented & tested pre-cutover
SOC 2 Aligned
Security & availability controls
Explore the full capabilities
Legacy System Modernization
Zero data loss migration from on-premise legacy EHRs to modern cloud-native platforms — with 6-month parallel run and rollback plan standard.
EHR & System Integration
The integration project that preceded this migration — connecting the legacy EHR to mobile documentation, OASIS automation, and payer portals.
Home Health
See how Metricoid serves home health agencies — OASIS, PDGM, mobile documentation, and legacy EHR connectivity.
Other case studies
CS-04: Legacy EHR connected. OASIS 71% → 94%. Notes in the home.
Days in AR from 71 to 29 — $520K recovered annually
13.2% denial rate to 4.1% — 62% less documentation time
Running a legacy system you've been afraid to touch?
We will assess your system — data volume, schema complexity, migration risk profile — and show you exactly what a safe migration would look like for your organization. 15 minutes, no commitment.
