Neonatal & Maternity EHR on Medplum: NICU Monitoring, Growth Tracking, and Mother's Portal Built on Open-Source FHIR

Executive Summary

A women's hospital specializing in maternal-fetal medicine and neonatal intensive care (NICU) needed a specialty EHR that tracked the unique clinical workflows of pregnancy, labor and delivery, and newborn care. Generic EHRs treat neonatal patients as tiny adults — they don't understand gestational age, Apgar scoring, growth percentiles, or the mother-baby dyad relationship.

We built a neonatal and maternity EHR on top of Medplum — an open-source FHIR R4 platform — creating custom clinical interfaces for NICU monitoring, maternity care timelines, labor and delivery documentation, WHO growth chart tracking, feeding/intake-output management, and a mother's mobile portal. By building on Medplum rather than from scratch, we had a FHIR-compliant data platform with auth, subscriptions, and bot framework out of the box, letting us focus entirely on clinical UX.

Delivered in 4 months to a 60-bed NICU serving 1,200 deliveries per year.

The Problem: Generic EHRs Don't Understand Babies

Neonatal and maternity care has clinical requirements that general-purpose EHRs simply don't handle:

• Gestational age, not chronological age: a premature baby born at 28 weeks has different vitals norms, medication dosing, and developmental milestones than a full-term baby. Generic EHRs don't calculate corrected gestational age.
• Mother-baby linkage: the mother and baby are clinically linked — mother's medications affect breastfed baby, maternal conditions impact neonatal outcomes. Generic EHRs treat them as independent patients.
• Apgar scoring and birth documentation: delivery mode, Apgar scores at 1/5/10 minutes, cord blood gases, placenta status — these don't fit in standard encounter templates.
• Growth percentiles: neonatal growth tracking uses WHO growth curves with gestational-age-corrected percentiles. Generic EHRs have adult BMI calculators.
• Feeding tracking: NICU patients require precise intake/output monitoring — breast milk volumes, fortifier additions, IV fluid rates, diaper counts. This data drives daily clinical decisions.
• Vitals frequency: NICU patients have vitals checked every 1-4 hours, not once per visit. The EHR needs to handle high-frequency vital sign data with trending and alerting.

Why Medplum?

Medplum is an open-source, headless FHIR R4 platform that provides the backend infrastructure for healthcare applications — FHIR data store, OAuth2 authentication, subscriptions (webhooks for data changes), a bot framework for automation, and a built-in admin console.

By building on Medplum, we got for free:

• FHIR R4 data store: every clinical data point stored as a standard FHIR resource — Patient, Encounter, Observation, MedicationRequest, etc.
• Auth + RBAC: OAuth2 with role-based access control — different permissions for NICU nurse, attending physician, resident, mother
• Subscriptions: webhook triggers when data changes — new vitals → auto-alert if outside range, new lab result → auto-notify provider
• Bot framework: serverless functions that run on FHIR events — calculating corrected gestational age, updating growth percentiles, generating feeding recommendations

Our engineering effort went 100% into clinical UX — the unique screens, workflows, and algorithms that make neonatal care work.

NICU Patient Dashboard

The NICU dashboard is the nurse's and doctor's primary view for each baby:

• Header: baby name, gestational age at birth (34 weeks), current corrected age (36+2), birth weight (2.1 kg), current weight (2.3 kg), Apgar scores (7/9), NICU bed assignment
• Real-time vitals: heart rate, respiratory rate, SpO2, temperature — each with mini sparkline showing the last 6 hours. Color-coded: green (normal), yellow (borderline), red (critical). Vitals auto-ingested from bedside monitors via HL7 v2.
• Feeding tracker: current feeding plan (breast milk 45ml q3h), today's intake/output summary, weight change trend
• Active medications: caffeine citrate (for apnea of prematurity), vitamin D, iron — with dose-per-kg calculations auto-updated as weight changes
• Growth summary: current percentile status with link to full WHO growth curves
• Mother link: click to view mother's chart, medications (relevant for breastfeeding safety), and delivery summary

Maternity Timeline

The maternity timeline tracks the entire pregnancy journey as a visual, chronological narrative:

• Prenatal milestones: each visit, scan, and test plotted on a week-by-week timeline from first trimester through delivery
• Key findings at each point: ultrasound results, blood work, gestational diabetes screening, Rh status, group B strep
• Risk flags: high-risk conditions highlighted on the timeline — preeclampsia, gestational diabetes, placenta previa — with the clinical response at each flag
• Delivery event: comprehensive delivery summary including mode, duration, complications, newborn status
• Postpartum: recovery milestones, postpartum depression screening (Edinburgh score), breastfeeding support, follow-up scheduling

Architecture

Medplum Platform Layer

Medplum provides the foundation — we didn't build any of this:

• FHIR R4 server with full CRUD, search, and history
• OAuth2 + SMART on FHIR authentication
• Subscription triggers for data-driven automation
• Bot framework for clinical calculations
• Admin console for data management

Custom Clinical Layer (What We Built)

Component	Technology	Purpose
NICU Dashboard	React + TypeScript	Bedside clinical view for nurses and physicians
Maternity Portal	React	Prenatal timeline, labor tracking, postpartum
Growth Charts	React + D3.js	WHO growth curves with corrected gestational age
Feeding Tracker	React	Intake/output logging, feeding plan management
Mother App	React Native	Baby updates, feeding log, appointments, education
Vitals Interface	Node.js + HL7 parser	Ingest real-time vitals from bedside monitors
Clinical Bots	Medplum Bot Framework	Gestational age calc, growth percentiles, alerts

WHO Growth Charts

Growth tracking is critical for NICU babies. Our implementation:

• WHO growth standards: weight-for-age, length-for-age, head circumference-for-age with percentile curves (3rd through 97th)
• Corrected gestational age: a baby born at 32 weeks who is now 8 weeks old is plotted at corrected age of 40 weeks (full term), not 8 weeks chronological. This dramatically changes percentile interpretation.
• Trend tracking: baby's measurements plotted over time, showing whether they're tracking along a percentile line (healthy) or crossing percentiles (concern)
• Printable reports: one-click PDF generation for parent handouts and referring pediatrician records

Feeding & Intake/Output Tracker

NICU feeding management is a clinical workflow unto itself:

• Feed logging: every feed recorded — type (breast milk, formula, fortified), volume, duration (for direct breastfeeding), route (oral, NG tube, gavage)
• Running totals: real-time daily intake vs. target (calculated from weight-based ml/kg/day goals). Alert if intake falls below threshold.
• Output tracking: wet diapers, stool frequency and character — key indicators of hydration and feeding tolerance
• Weight delta: daily weight change displayed prominently. NICU babies should gain 15-20g/day once feeding is established. Below this → feeding plan review.
• Feeding schedule: visual 24-hour timeline showing feeding blocks — helping nurses maintain consistent intervals

Mother's Portal

NICU parents are anxious and separated from their baby. The mother's portal provides:

• Baby updates: real-time status — weight, feeding progress, care milestones
• Growth tracking: simplified growth chart view with "on track" status indicators
• Feeding contribution: mothers providing expressed breast milk can log pumping sessions and see their contribution to baby's nutrition
• Discharge checklist: progressive checklist of milestones needed before going home — feeding independently, maintaining temperature, car seat test passed
• Educational content: curated articles and videos on breastfeeding, safe sleep, developmental milestones, when to call the pediatrician

Labor & Delivery Documentation

The L&D module captures the most clinically intense period:

• Partograph: cervical dilation plotted over time against expected progress curves (alert and action lines). The visual partograph is the gold standard for labor monitoring.
• Maternal vitals strip: continuous display of maternal heart rate, blood pressure, and fetal heart rate with contraction monitoring
• Medication log: epidural timing, pitocin drip rates, antibiotics — all time-stamped for legal documentation
• Delivery summary: structured capture of delivery mode, time, Apgar scores (1/5/10 min), birth weight, cord blood gases, placenta delivery, estimated blood loss
• Newborn transition: immediate post-delivery documentation — resuscitation if needed, skin-to-skin timing, first feeding

Results

Metric	Before	After	Impact
Build time	N/A	4 months	Medplum accelerated 60% vs from-scratch
Vitals documentation	Manual charting from monitors	Auto-ingested via HL7	Zero manual transcription
Growth percentile accuracy	Hand-plotted on paper charts	Auto-calculated with corrected GA	Eliminated plotting errors
Feeding documentation time	8 min per feed event	45 seconds (tap logging)	91% faster
Parent engagement	Daily phone calls from nurses	Real-time app updates	24/7 visibility for parents
Discharge readiness tracking	Informal, undocumented	Structured checklist	14% shorter NICU stays
Medication dosing errors	Manual weight-based calculation	Auto-calculated per current weight	Zero dosing calculation errors

Timeline

Month	Deliverables
Month 1	Medplum deployment, FHIR resource design (neonatal profiles), vitals HL7 interface, NICU dashboard MVP
Month 2	Growth charts (WHO curves + corrected GA), feeding tracker, mother-baby linkage, medication management
Month 3	Maternity timeline, labor & delivery module, mother's portal (React Native), discharge checklist
Month 4	Clinical bots (auto-alerts, percentile calculations), staff training, pilot deployment, go-live

Team: 3 engineers + 1 clinical informatics specialist. Medplum's platform layer saved an estimated 2-3 months of infrastructure work.

Lessons Learned

• Medplum as a platform accelerates dramatically. We estimated 6-7 months to build from scratch. Medplum gave us auth, FHIR storage, subscriptions, and bots out of the box — saving 2-3 months. The ROI on choosing an open-source FHIR platform over building custom was clear.
• Corrected gestational age changes everything. Every clinical calculation in neonatal care — drug dosing, growth percentiles, developmental milestones — must use corrected gestational age. Building this into the core data model (not as an afterthought) was critical.
• Parents need a portal, not a patient portal. NICU parents don't want to log into a generic patient portal and navigate to their baby's chart. They want a dedicated app that shows baby updates, growth progress, and discharge readiness — designed with empathy for an anxious parent, not a generic healthcare consumer.
• HL7 v2 vitals integration is the first thing to get right. NICU runs on vitals. Auto-ingesting from bedside monitors eliminated the #1 source of documentation burden and transcription errors. This single integration saved more nursing time than all other features combined.