LIVE
3 ALERTS
--:--:--
Blood Pressure
148/91
mmHg
↑ +8 vs avg Omron HEM-7155T
Heart Rate
72bpm
Regular sinus rhythm
→ Stable Fitbit Sense 2
SpO₂
97%
Oxygen saturation
→ Normal Apple Watch S9
Blood Glucose
6.4mmol/L
Post-prandial
↓ −0.8 Libre 3 CGM
Temperature
36.7°C
NormalWithings Thermo
Respiration
16br/min
NormalFitbit Sense 2
Steps Today
1,847steps
↑ +12%Fitbit Sense 2
Sleep Quality
5.2hrs
↓ −1.8hSleep Monitor
Vital Signs Trend — 6 Hours
BP
HR
SpO₂
Glucose
🔔 Active Alerts 3 Active
🩺
Hypertensive Episode
BP 148/91 — sustained >30 min above threshold (140/90)
09:14 · Escalated to RN
😴
Sleep Disruption
Total sleep 5.2h. Deep sleep only 18 min (low). 4 wake events
06:42 · Auto-logged
🚶
Reduced Mobility
Activity 40% below 7-day average. No motion 11:00–14:30
14:30 · Predictive flag
Clinical Risk Scores
Deterioration Risk
62%
Fall Risk
45%
Dehydration Risk
38%
Sleep Disruption
74%
Cardiac Event Risk
22%
Infection Risk
18%
Today's Timeline
09:14
Hypertensive alert triggered
BP 148/91, sustained >30 min. RN Sinéad O'Brien notified.
08:30
Morning vitals sync
All devices synced. CGM shows post-meal glucose 6.4 mmol/L.
06:42
Sleep session ended
5.2h total, fragmented. 4 wake events. Deep sleep deficit.
01:22
Night check — no motion alert
Position change detected. No fall detected. SpO₂ stable 96–98%.
22:10
Evening medication reminder
Amlodipine 5mg confirmed taken. BP pre-dose: 142/88.
ClinAI Insight
Powered by claude-sonnet-4-6
Click below to generate a real-time AI clinical summary for Mary O'Sullivan based on today's monitoring data.
Blood Pressure — 24h ELEVATED
Systolic avg 142 Diastolic avg 88 Peak 158/96 Readings 24
Heart Rate — 24h NORMAL
Avg 71 bpm Min 58 Max 94 HRV 28ms
Blood Glucose (CGM) — 24h IN RANGE
Time in range 78% Avg 6.8 mmol/L Highest 9.2 A1c est. 48 mmol/mol
SpO₂ & Respiration — 24h NORMAL
SpO₂ avg 97% Low 94% RR avg 16 br/min Events 0
Vital Signs Table — Last 12 Readings
TimeBP (mmHg)HRSpO₂GlucoseTempStatus
Vitals Analysis
AI pattern recognition
Run analysis to detect patterns and anomalies across all vital sign streams.
Reference Ranges
ParameterNormalTarget
Systolic BP<120 mmHg<130
Diastolic BP<80 mmHg<85
Heart Rate60–100 bpm60–90
SpO₂≥95%≥96%
Blood Glucose4–8 mmol/L5–7.5
Temperature36.1–37.2°C36.5–37.0
1
CRITICAL
2
WARNING
4
INFORMATIONAL
12
RESOLVED (24h)
Active Alerts
🩺
Hypertensive Episode — Mary O'Sullivan
CRITICAL
BP 148/91 mmHg sustained >30 min. Threshold: 140/90. Amlodipine dose 22:10 previous evening. Possible dose timing issue or stressor event.
09:14 today · Omron HEM-7155T · Auto-escalated to RN Sinéad O'Brien
😴
Sleep Disruption Pattern — Mary O'Sullivan
WARNING
5.2h total sleep (target ≥7h). Deep sleep 18 min (target >90 min). 4 wake events. Third consecutive night below threshold.
06:42 today · ResMed Sleep Mat · Pattern flag
🚶
Reduced Mobility — Mary O'Sullivan
WARNING
3.5h inactivity window 11:00–14:30. Steps 1,847 vs 7-day avg 3,200. Possible pain, fatigue, or mood-related. Motion sensor and wearable concur.
14:30 today · Motion sensor + Fitbit Sense 2 · Predictive flag
💧
Dehydration Risk — Patrick Brennan
INFO
Reduced activity and elevated temp (37.4°C) combined with low fluid intake log. Predictive model: 38% dehydration risk.
11:20 today · Withings Thermo + motion sensor
Alert Escalation Framework
⚡ CRITICAL Alert
Immediate — threshold breach, fall detected, cardiac anomaly
↓ Auto-escalate <2 min
📞 RN Notification
SMS + app push + dashboard flag
↓ No response in 5 min
🏥 GP / On-Call Clinician
Phone + secure message
↓ If unresolved 15 min
🚨 Emergency Services
999 / NOK contact
Alert Thresholds
ParameterWarningCritical
Systolic BP≥140≥160
Diastolic BP≥90≥100
Heart Rate<55 / >95<45 / >110
SpO₂<95%<92%
Glucose<4 / >10<3 / >14
Temperature≥37.5°C≥38.5°C
Fall detectedStumbleImpact
🧠
Deterioration Detection
NEWS2-calibrated AI model
Continuous analysis of vital sign trajectories against personalised baselines. Early warning 4–12h before clinical deterioration.
Mary O'Sullivan ELEVATED · 62%
Patrick Brennan STABLE · 24%
Brigid Kelly STABLE · 17%
CONTRIBUTING SIGNALS — MARY
↑ BP trend +8 mmHg (2h)
↓ Activity −40% (morning)
↓ Sleep quality (3 nights)
→ Medication adherence OK
🚶
Mobility Intelligence
Gait, activity & motion analysis
Fuses smartwatch accelerometry, PIR motion sensors, and door sensors to build a 24h mobility profile.
Today: 1,847 steps Avg: 3,200 Active: 4.2h Sedentary: 8.1h
⚠ 3.5h inactivity window detected 11:00–14:30. Below personal baseline.
💧
Dehydration Risk Model
Multi-signal inference
No direct hydration sensor — AI infers dehydration risk from HR variability, activity, temp, CGM trend, and self-report patterns.
38%
Dehydration Risk Score
HR variability
+
Temp 36.7°C
Fluid log gap
+
Deep Clinical Analysis
Comprehensive AI-generated assessment across all monitoring streams
Click to generate a comprehensive clinical intelligence report integrating all device streams, risk models, and clinical context for Mary O'Sullivan.
AI Model Architecture
Device Data Ingestion
BLE · Cellular · WiFi · HL7 FHIR
Feature Engineering Pipeline
Time-series normalisation · Imputation
Anomaly Detection Models
LSTM · Isolation Forest · NEWS2-AI
Risk Scoring Engine
Ensemble model · Personalised baseline
Claude API Synthesis
Clinical narrative + recommendation
Total Sleep
5.2hrs
↓ −1.8h vs target
Deep Sleep
18min
↓ Low (target >90m)
REM Sleep
72min
→ Normal range
Wake Events
4×
↑ Above threshold
Sleep Architecture — Last Night
SLEEP STAGES
Awake
Light
Deep
REM
Awake
Light
REM
Awake
🟦 Light sleep · 42% ■ REM · 23% ■ Deep · 6% □ Awake · 29%
7-NIGHT SLEEP TREND
Night Vitals During Sleep
TimeSpO₂HREvent
22:1597%68Onset
00:4095%72Wake ×1
01:2296%70Reposition
02:5594%76Wake ×2
04:3096%65Wake ×3
05:5097%68Wake ×4
06:4297%71Wake up
Sleep Analysis
Pattern & intervention AI
Run analysis for AI-generated sleep quality assessment and intervention recommendations.
Falls (30 days)
0events
No falls detected
Fall Risk Score
45%
↑ STRATIFY: Medium
Gait Speed
0.82m/s
→ Normal range
TUG Test
13.2sec
↑ Borderline
Daily Activity Heatmap — Last 7 Days
Hours of significant movement per day. Grey = low activity, Teal = active, Amber = spike.
Gait Trend — 30 Days
Current: 0.82 m/s 30-day avg: 0.87 m/s Trend: ↓ slight
Fall Detection System
📱 Wearable Fall Detector
Fitbit Sense 2 accelerometer + gyroscope. Detects impact + post-fall inactivity. Response time <2 seconds. Last tested: yesterday.
ACTIVE Battery 78%
🏠 Motion Sensor Array
5× PIR sensors (bedroom, bathroom, hallway, kitchen, living room). Tracks room transitions and night-time toilet trips. Alert if no motion >4h daytime.
ALL ONLINE
🛌 Bed Exit Monitor
Pressure mat detects bed exit between 22:00–07:00. Alert if no bathroom return within 12 min.
ACTIVE 0 alerts last 7 nights
Falls Risk AI Assessment
STRATIFY + mobility AI + home environment
Generate an AI-powered falls risk assessment combining validated STRATIFY scores, gait data, mobility trends, and environmental factors.
INTERVENTIONS IN PLACE
✅ Non-slip mat in bathroom
✅ Grab rail — bath and toilet
✅ Bed raised to safe height
✅ Personal alarm pendant worn
⚠️ Hallway nightlight — check needed
❌ Stair handrail — awaiting installation
Open Interventions
Date/TimeTriggerActionClinicianStatus
17/06 09:14 BP 148/91 Medication review + GP call S. O'Brien RN OPEN
16/06 06:42 Sleep disruption Sleep hygiene review D. Nolan RN OPEN
15/06 14:10 Reduced mobility Physio referral submitted S. O'Brien RN PENDING
14/06 11:30 Glucose 9.8 Dietary adjustment note M. Walsh RN RESOLVED
12/06 08:00 SpO₂ 93% Chest assessment + GP D. Nolan RN RESOLVED
Outcomes Tracking
Interventions (30d): 8 Resolved: 6 Avg response: 4.2 min A&E prevented est.: 2
Intervention Recommendation
AI-generated care plan action
Generate an AI recommendation for the current hypertensive episode and reduced mobility alerts.
Clinical Governance — Intervention Protocol
HIQA Regulation 27 Compliance
All interventions documented in real-time, timestamped, and linked to triggering alert. Mandatory GP notification for BP >160/100.
Decision Support
AI provides recommendation only. All clinical decisions made by RN/GP. AI output is advisory — clinician documents acceptance or override.
Audit Trail
Full immutable log: alert → notification → response time → action → outcome. Available for HIQA inspection.
Mary O'Sullivan — Active Devices
Fitbit Sense 2
Smartwatch · HR, SpO₂, ECG, Temp, Steps, Sleep · Firmware 1.196.98
🔋 78%
CONNECTED
🩺
Omron HEM-7155T
Blood pressure · BLE 5.0 · Auto-sync 3×/day · Avg reading accuracy ±3 mmHg
🔋 92%
CONNECTED
📉
Libre 3 CGM
Continuous glucose · NFC + BLE · 14-day sensor · Real-time 1-min readings
Sensor: 6 days left
CONNECTED
🌡️
Withings ScanWatch 2
Temperature, SpO₂ overnight, AFib detection · Cellular backup
🔋 55%
SYNCING
🛏️
ResMed S+ Sleep Mat
Sleep stages, respiration, movement · WiFi · No contact required
Mains
CONNECTED
🏠
PIR Motion Sensors ×5
Bedroom, bathroom, hallway, kitchen, living room · Zigbee mesh · 2yr battery
All >80%
ALL ONLINE
📳
Tunstall i-vibra Fall Detector
Pendant fall sensor · GSM + WiFi · 3-axis accelerometer · 24h battery
🔋 88%
CONNECTED
Data Quality Dashboard
Completeness: 94% Latency avg: 2.1s Uptime: 99.2%
Connectivity Architecture
Devices
BLE 5.0 · Zigbee · NFC · WiFi
Edge Hub
Raspberry Pi 4 · Local processing
MQTT Broker
TLS 1.3 · QoS 2
SolidRock Cloud
FHIR R4 · HL7 · AWS HSM
AI Processing + Family Portal + Clinician Dashboard
Real-time WebSocket · HIQA audit log · GDPR-compliant data residency (IE)
Full System Data Flow
⌚ Smartwatch
Fitbit / Apple Watch
🩺 BP Monitor
Omron / Withings
📉 CGM
Libre 3
🏠 Motion PIR
Zigbee array
🛏️ Sleep Mat
ResMed S+
📳 Fall Detect
Tunstall
↓ BLE · WiFi · Zigbee · NFC · Cellular
🏠 Local Edge Gateway
Raspberry Pi 4 · Data aggregation · Offline resilience · Local alert execution
↓ MQTT over TLS 1.3 · Cellular 4G/5G backup
☁️ SolidRock RPM Cloud (AWS Dublin)
FHIR R4 datastore · Event-driven Lambda · Time-series DB (InfluxDB) · Kafka streams
🧠 AI Engine
ML models + Claude API
🔔 Alert Engine
PagerDuty + SNS
📊 Dashboards
Clinician + Family
Technology Stack
LayerTechnologyPurpose
Device comm.BLE 5.0, Zigbee 3.0Sensor connectivity
EdgeRPi 4, Python 3.12Local processing
TransportMQTT, HTTPS, WSSData streaming
Time-series DBInfluxDB 3.0Vital signs storage
FHIR StoreAzure Health DataClinical records
Stream proc.Apache KafkaReal-time events
AI/MLPyTorch, Claude APIIntelligence layer
BackendTanStack Start, NodeAPI + portal
AuthOAuth 2.0, PKCEIdentity
HostingAWS eu-west-1IE data residency
FHIR R4 Resource Mapping
Device DataFHIR Resource
Vital signsObservation
AlertDetectedIssue
InterventionClinicalImpression
DeviceDevice + DeviceUseStatement
PatientPatient
Care planCarePlan
⚖️HIQA ComplianceACTIVE
✅ Regulation 5 — Individual assessment
✅ Regulation 6 — Healthcare plan
✅ Regulation 7 — Positive behaviour support
✅ Regulation 25 — Temporary absence
✅ Regulation 27 — Infection prevention
✅ Regulation 31 — Notification of incidents
✅ Regulation 34 — Complaints procedure
⚠️ Regulation 17 — Premises (ongoing audit)
🔒GDPR & Data ProtectionCOMPLIANT
✅ DPIA completed (RPM platform)
✅ Data residency: AWS eu-west-1 (Dublin)
✅ AES-256 encryption at rest
✅ TLS 1.3 encryption in transit
✅ Role-based access control
✅ Data minimisation — device-specific
✅ Right to erasure workflow
✅ DPO appointed: M. Burke
✅ Consent management system
✅ 7-year retention policy
🏥Clinical SafetyACTIVE
✅ AI is advisory — clinician decides
✅ All alerts logged + timestamped
✅ RN response SLA: <5 min critical
✅ GP escalation protocol documented
✅ Device calibration schedule
✅ False positive rate monitored
✅ Clinical lead sign-off: Dr. M. Flanagan
✅ Indemnity insurance: VHI + Marsh
⚠️ Annual safety review: July 2026
Audit Log — Recent Entries
TimestampEventUserAction
17/06 09:14Critical alert firedSYSTEMAUTO-ESCALATE
17/06 09:16Alert acknowledgedS.OBrien RNACK
17/06 09:20GP contactedS.OBrien RNCALL LOGGED
17/06 08:30Device syncSYSTEMSUCCESS
16/06 22:10Medication reminder sentSYSTEMSENT
16/06 18:00Care plan reviewedD.Nolan RNSIGNED
Consent Management
Mary O'Sullivan
✅ RPM monitoring consent signed — 12/03/2026
✅ Data sharing (GP): consented
✅ Family portal access (daughter Siobhán): consented
✅ AI-assisted analysis: consented
⚠️ Research data sharing: not consented
Consent Renewal
Next review due: 12/09/2026 (6-month cycle)
Reminder set: 30 days prior
Core Endpoints
GET /api/v2/patients/{id}/vitals
Real-time and historical vital signs. Query params: from, to, interval, metrics[]
GET /api/v2/patients/{id}/alerts
Active and historical alerts. Filter: status, severity, device, dateRange
POST /api/v2/patients/{id}/interventions
Log a clinical intervention linked to an alert. Body: alertId, action, clinicianId, notes
POST /api/v2/ai/analyse
Trigger AI analysis for a patient. Returns structured clinical narrative via Claude API.
GET /api/v2/devices/{patientId}
List connected devices, battery, last-seen, firmware version, data quality score
DEL /api/v2/alerts/{alertId}/acknowledge
Acknowledge and close an alert. Requires clinician auth token.
WebSocket — Real-time Stream
// Connect to live vital stream
const ws = new WebSocket(
  'wss://rpm.solidrockhomecare.ie/live'
);

ws.onmessage = (e) => {
  const data = JSON.parse(e.data);
  // data.type: 'vital' | 'alert' | 'device'
  // data.patientId, data.metric, data.value
  // data.timestamp (ISO 8601)
};
Webhook Events
EventTriggerPayload
vital.alert.criticalThreshold breachpatient, metric, value, device
vital.alert.warningWarning levelpatient, metric, value, trend
fall.detectedImpact + inactivitypatient, location, severity
device.offlineNo ping >15 mindeviceId, lastSeen, type
sleep.reportMorning, dailystages, duration, quality
activity.lowInactivity flagpatient, windowHrs, steps
glucose.highCGM >10 mmol/Lvalue, trend, timestamp
ai.insight.readyAnalysis completereportId, summaryText
Authentication
OAuth 2.0 + PKCE — All API calls require a Bearer token from the SolidRock Identity Service. Tokens expire after 1 hour. Refresh tokens valid 30 days.
Authorization: Bearer <JWT>
X-SolidRock-OrgId: sr-chfo7
X-SolidRock-Version: 2.1
Role scopes: rpm:read · rpm:write · alerts:ack · ai:analyse · admin:all
Rate Limits
Endpoint TypeLimitWindow
Vital reads1,000 reqper minute
Alert operations200 reqper minute
AI analysis10 reqper minute
Device status500 reqper minute
WebSocket msgsUnlimitedsubscribed