A precision-nutrition decision platform for one of medicine's most unforgiving environments.
AI that turns the Medical Client’s proprietary neonatal dataset into protocol-backed feeding guidance at the bedside, in real time.
- EpicEHR integrated via HL7 / FHIR
- 0→1Built end-to-end · no legacy base
At a glance
- ClientA precision-nutrition company sitting on a proprietary clinical dataset.
- ChallengeDecision intelligence in a setting where the margin for error is millimetres.
- ApproachThe dataset is the intelligence core. The application's job is to make it queryable at the point of care.
- ResultsFrom disconnected data to actionable intelligence at the bedside.
Client's data was scientific. Their software was nonexistent.
A US company focused on preterm infants, holding one of the most comprehensive neonatal nutrition datasets: feeding protocols, microbiome profiles, longitudinal growth metrics.
The data sat in disconnected systems, unusable in real time; feeding was decided from memory and judgement.
Preterm care is unforgiving. The cognitive load is enormous.
Feeding decisions shape gut development, growth velocity and long-term neurological outcomes — made several times a day by nurses managing several critically ill patients at once.
No existing codebase
No prototype, no legacy system: everything built from zero, data model to clinical UI.
Multi-stream data complexity
Every recommendation combines intake history, growth curves, microbiome indicators and protocol parameters, all varying per patient.
Clinical-trust requirements
Augment clinical judgement, don’t replace it: every AI suggestion traceable, overridable, with a logged reason.
EHR integration
Bidirectional Epic sync — patient records in, compliance data out — without interrupting hospital workflows.
A four-layer build, scoped to the clinical workflow.
Inputs
- Feeding protocol
- Microbiome data
- Growth metrics
Decision engine
Protocol-matched feeding decision tree
Outputs
- Nurse dashboard
- Rounding report
- Epic EHR sync
- 01Data layer
Unified patient data model
One PostgreSQL schema for feeding history, growth velocity, microbiome profiles and protocol parameters — logged on admission, updated throughout care.
- 02AI decision engine
Protocol-matched feeding decision tree
Weighs prior outcomes, growth speed, tolerance and contraindications against the Client’s protocol library; nurses see why and can override.
- 03Clinical interface
Nurse dashboard & rounding reports
An Angular front-end: recommendations, a rounding-report tab, growth trends per infant and cohort, logged protocol changes.
- 04EHR & analytics
Epic integration and performance analytics
Two-way Epic sync, no re-entry; adherence and outcome correlations tracked over time, exportable for research and reporting.
Nine capabilities, designed as one clinical decision-support platform.
- AI feeding recommendations
- Growth-monitoring charts
- Rounding report tab
- Protocol-compliance tracking
- Epic EHR integration
- Central identity server
- Organisation analytics
- Data and report export
- Microbiome data integration
Compliance went from untracked to measured. Decisions got faster. Nurses got their cognitive capacity back.
Protocol adherence, finally measurable
Consistent adherence across participating NICUs; before the software, deviations were common and untracked.
Cognitive load, back where it belongs
Less time on feeding calculations, more on direct patient care; nurses report confidence acting on suggestions.
Real-time record currency
Epic sync removed manual entry between systems — fewer transcription errors, patient records kept current.
End-to-end greenfield delivery
Data model, AI engine, clinical interface, EHR integration, analytics — all built by our team.
From the team
Client's proprietary dataset integrates feeding protocols, microbiome profiles and clinical information. We synchronised it with the application so it generates actionable suggestions in the form of a decision tree — giving nurses protocol-backed recommendations at the exact moment they need them.
Stack
Chosen for reliability, clinical performance and a track record in regulated healthcare — not novelty.
Building something for clinical teams?
Clinical decision support, neonatal software, regulated AI — write to a principal engineer. We reply within one business day. No discovery call.
More case studies
