We connected a regional health provider's separate EHR systems with a shared record format, live data sharing and patient matching across sites. Records became 100% consistent, and response times fell 85%.
Industry: HealthcareService: SaaS & SoftwareUpdated
Overview
The challenge
Each clinical site ran its own older EHR, from Epic and Cerner to custom SQL databases. Clinicians could not see a patient's full medical history in real time, and staff matched patients across sites by hand, which left records split and incomplete.
When a patient moved between sites, their data could take 4-24 hours to follow. Every new connection between systems needed custom code.
Allergy and medication lists did not match across sites, which raised the chance of preventable drug reactions.
Clinicians repeated diagnostic tests because they could not see earlier results, at an estimated cost of $2.4M a year.
The older systems had no shared logging, so 21 CFR Part 11 and HIPAA audits were slow, manual and prone to error.
The solution
A translation layer converts older HL7 v2 and SQL data into the FHIR R4 health-record standard as soon as it arrives.
A live event stream shares admissions, lab results and medication orders with all sites in under a second.
A Master Patient Index (MPI) uses consistent matching rules to confirm each patient's identity across systems before any records are merged.
Before and after
Before
Mismatched HL7 feeds
Non-standard data structures meant custom code for every new system connection.
After
One FHIR record
Standard FHIR R4 resources work with any compatible application from day one.
Before
Site-by-site controls
Privacy controls lived inside each separate EHR, with no single view of who accessed patient data.
After
Shared controls
One set of HIPAA-aligned access controls, with a traceable record of every data request.
Before
Batch transfers
Patient data often arrived 4-24 hours late when a patient moved between sites.
After
Live streams
Clinical events reach every site in < 500ms.
Key features
Every change is stored as a traceable event, so the team can replay a record's history to fix issues or answer an audit.
Business impact
Zero inconsistent records
Patient health information is filtered out of data streams based on each user's role, which also lets researchers use the data safely.
Business impact
100% regulatory alignment
Each EHR adapter runs in a container that scales with message volume, cutting cloud spend during quiet night shifts.
Business impact
30% infrastructure cost savings
Results
100% Record Consistency
All 14 data sources now feed one schema, and not a single record was rejected along the way.
Before: 0% StdAfter: 100% Std
6x Faster Onboarding
Connecting a new clinical site now takes 2 weeks instead of 12.
Before: 12 WeeksAfter: 2 Weeks
85% Data Access Response Time
Live auditing and redaction give authorised clinicians patient data in under 50ms, where requests used to take hours.
Before: HoursAfter: < 50ms
Faster delivery
These components saved about 12 weeks of setup work. The team used the time for site-specific mappings and patient matching.
Provided tested sign-in and role-based permissions for clinicians and researchers using the shared record.
Supplied ready-made HL7 v2 to FHIR mappings, which we adapted to each site's EHR version.
Showed message flow and errors for every site connection, so failed feeds were caught quickly.
Kept patient data inside the approved AWS regions from the first release.
Trust and control
Multi-region AWS residency keeps all patient health information within the required regulatory boundaries.
HIPAAData is encrypted with AES-256 at rest and TLS 1.3 in transit across the whole streaming system.
EncryptedSelf-healing cloud clusters keep critical lab results available 99.99% of the time.
High availabilityOn completion, the client owns 100% of the health-record adapters and translation logic.
100% ownedEach of its 14 clinical sites ran its own older EHR, including Epic, Cerner and custom SQL databases. Records moved between sites in batches that took 4-24 hours, and patients were matched by hand. Allergy and medication lists did not line up, and duplicate tests cost an estimated $2.4M a year.
A translation layer converts HL7 v2 and SQL data into the FHIR R4 standard as it arrives. A live event stream then shares admissions, lab results and medication orders across all sites in under 500ms. A Master Patient Index matches each patient across systems with 99.9% accuracy before records merge.
Access is controlled centrally, and every data request leaves a traceable audit trail. Patient health information is redacted from data streams according to each user's role, which also allows safe use for research. Data is encrypted with AES-256 at rest and TLS 1.3 in transit, and stays in the required AWS regions.
If your sites run different EHRs that do not talk to each other, we can help you connect them into one shared record without replacing any system.