Building a FHIR-Compliant EHR Integration Platform for a 14-Hospital NHS Trust
Challenge
Fourteen hospitals operated six different EHR systems with no interoperability. Clinicians treating transferred patients had no access to prior records, leading to patient safety risks, duplicated investigations, and delayed clinical decisions.
Solution
We designed and built an HL7 FHIR R4 integration platform aggregating records from all six systems into a unified patient timeline, accessible via a secure API and clinician-facing portal.
Results
All 14 hospitals connected within 8 months. Duplicate investigations reduced by 34%. Full patient history accessible in 4 seconds. Zero reportable data incidents in 18 months.
The Clinical Risk of Fragmented Records
Healthcare fragmentation is not an inconvenience. It is a patient safety issue.
At Northern Health NHS Trust:
- Patients transferred across sites regularly
- Average patient had attended 2.3 hospitals within the Trust
- No shared record access across systems
Clinician workflow reality:
- Call records department
- Wait hours (or unavailable out of hours)
- Make decisions with incomplete information
Consequences:
- Duplicate blood tests and imaging
- Medication errors (missing drug history)
- Delayed treatment decisions
The problem wasn’t data availability. It was data accessibility at the point of care.
System Design Approach
This was not a data warehouse project. It was a real-time clinical interoperability platform, designed to:
- surface complete patient history instantly
- unify heterogeneous EHR systems
- operate within NHS security and governance constraints
- integrate directly into clinical workflows
Technical Architecture
FHIR R4 as the Canonical Model
All source systems mapped into a single standard:
FHIR R4 resources used:
- Patient
- Encounter
- Observation
- MedicationRequest
- DiagnosticReport
- AllergyIntolerance
Source systems integrated:
- Cerner
- EMIS
- SystemOne
- Meditech
- Lorenzo
- In-house legacy system
Key outcome: Semantic interoperability across all clinical data
Identity Resolution Layer
Patient identity consistency is critical.
Challenges:
- Duplicate NHS numbers
- Name variations
- Incomplete records
Solution: probabilistic matching engine
- Jaro-Winkler string similarity (names)
- NHS number cross-referencing
- Date of birth validation
- Confidence scoring
Performance:
- 99.2% correct patient matching
- Low-confidence cases routed to manual review
Eliminates fragmented patient identities across systems
Real-Time Event-Driven Synchronisation
Batch sync is too slow for clinical environments.
Architecture:
- HL7 v2 ADT feeds from all systems
- Event-driven ingestion pipeline
- Streaming updates to unified record
Performance:
- Admissions, transfers, discharges → real-time
- Observations and results → synced within 60 seconds
Clinicians see near real-time patient state
Clinical Portal & API Layer
Two access modes:
1. Clinical Portal
- Chronological patient timeline
- Source system attribution
- Unified view across all hospitals
2. SMART-on-FHIR API
- Secure programmatic access
- Enables integration with clinical apps
- Extensible for future digital services
Supports both human and system-level consumption
Security & Compliance
Operating within NHS regulatory frameworks:
- NHS Data Security and Protection Toolkit compliant
- Encryption at rest and in transit
- Full audit logging of:
- Data residency: within NHS network boundaries
Zero external data exposure risk
Measured Outcomes
Hospital connectivity14 hospitals integrated within 8 months
Duplicate investigationsReduction: 34%Impact: cost savings + reduced patient burden
Access time to full patient historyBefore: hours (or unavailable)After: 4 seconds
Data security incidents0 reportable incidents over 18 months
Clinical Impact
Before:
- Fragmented records
- Delayed decisions
- Repeated investigations
- High clinical risk
After:
- Unified patient view
- Faster clinical decision-making
- Reduced duplication
- Improved patient safety
Why This Worked
- Used a universal standard (FHIR)Avoided vendor lock-in and ensured scalability
- Real-time architecture, not batchCritical for clinical relevance
- Solved identity firstNo unified record without accurate matching
- Designed for clinicians, not just systemsTimeline view aligned with clinical workflow
- Embedded compliance from day oneEnabled safe deployment at NHS scale
The Key Insight
Healthcare interoperability is not a data problem. It is a real-time systems problem constrained by safety, identity, and governance. Solving it requires:
- standardisation (FHIR)
- event-driven architecture
- rigorous identity resolution
Final Outcome
The platform transformed patient records from:
- fragmented → unified
- delayed → real-time
- inaccessible → clinically actionable
Result:
A Trust-wide patient record system that improves safety, efficiency, and clinical confidence.
Building Healthcare Systems That Work in the Real World?
Intagleo Systems designs interoperable, compliant healthcare platforms that integrate with existing infrastructure while improving clinical outcomes.