Good morning 👋
Gauteng Health Network · 3 active alerts
Active Alerts
0
↑ 2 since yesterday
🚨
Lab Rerun Rate
0
↑ 6.2% vs baseline
🧪
Avg TAT (mins)
0
↑ 34 min above target
⏱️
QC Pass Rate
0
↓ 1.4% vs last week
✅
Trial Sites Active
0
1 flagged · Site B
📋
Protocol Deviations
0
↑ Critical · Site B
⚠️
Ward D Discharge Lag
0
↑ Correlated with TAT
🏥
Decision Cards Today
0
4 resolved · 3 open
📊
Active Decision CardsLoading…
All
🔴 Critical
🟡 Attention
Lab
Trial
Hospital
Loading decision cards…
System Health4 active
🧪LIMS
74%
📋CTMS
42%
🏥HIS/EMR
91%
📊QC Logs
68%
Rerun Rate · 7 DaysBaseline 12.2%
Current
0%
Delta
+6.2%
MonTueWedThuFriSatNow
Connected Sources
LIMS v4.2
Lab · Real-time
CTMS Pro
Trials · Sync
MedRecord HIS
Hospital · HL7
QC Logger
Manual · CSV
← Dashboard
← All Cards
Card Metadata
Detection Timeline
🎯 Confidence Radar
Actions
Decision Cards
All generated intelligence cards
All Decision Cards
Loading…
Intelligence Reports
Data science insights · 7-day rolling window · Gauteng Health Network
0
Cards Generated
↑ 43% vs last week
0
Resolution Rate
↑ 12% vs last week
0
Avg Time-to-Resolve
↓ 1.8h improvement
0
Detection Accuracy
↑ verified by audit
0
Standards Breached
ISO 15189 · ICH E6
0
Peak Rerun Rate
+6.2% above baseline
Lab Rerun Rate — 14 Day Trend
Daily QC rerun % · Baseline 12.2% · Anomaly threshold: 15%
Feb 4
Feb 5
Feb 6
Feb 7
Feb 8
Feb 9
Feb 10
Feb 11
Feb 12
Feb 13
Feb 14
Feb 15
Feb 16
Feb 18
■ Normal■ Attention■ Anomaly
Decision Card Distribution
Category breakdown · 7-day period · 23 total cards
Laboratory QC11 (48%)
Clinical Trial6 (26%)
Hospital Operations4 (17%)
Stable / Baseline2 (9%)
Statistical note: Lab QC cards disproportionately high (48% vs 32% baseline). Pearson r=0.91, p<0.01.
TAT vs. Discharge Lag — Correlation
Pearson r=0.87 · Ward D · n=18 · 3 shifts
Lab TAT above baseline (mins)
■ Normal■ Elevated■ Criticalr = 0.87
Descriptive Statistics — QC CV%
Batch 7 vs Baseline · Analyser 2 · Haematology
METRIC
BASELINE vs BATCH 7
Mean
2.20% → 5.43%
Std Dev (σ)
0.30 → 1.62
Z-Score (peak)
3.1σ ⚠
p-value (t-test)
p < 0.001 ***
CV Allowable Limit
ISO 15189: ≤3.0%
ISO 15189 §5.6.2 BREACH: QC CV exceeded allowable limit (3.0%) for 9 consecutive windows. Mandatory corrective action required.
Registered Users — Live from Supabase DatabasePostgreSQL · Real-time
Loading users…
Data Sources & Settings
Configure connected systems, thresholds, and notifications
🔌Data Sources
🔍Detection Engine
🔔Notifications
👥Users & Roles
🔐Security
Connected Data Sources
Read-only integrations. JLIGHT never writes back to source systems.
Active Connections
🧪● Connected
LIMS v4.2
Gauteng Lab · CSV + REST · Last sync: 2 min ago
📋● Connected
CTMS Pro
Phase II Trials · REST API · Last sync: 5 min ago
🏥● Connected
MedRecord HIS
HL7 v2.5 · Hospital System · Last sync: 1 min ago
📊● Connected
QC Logger
Daily CSV upload · Last sync: 08:00 today
🗄️○ Disconnected
FHIR R4 Endpoint
Not configured · Add server URL to enable
Sync Settings
Sync Interval
How often JLIGHT pulls from connected sources
Read-Only Mode
JLIGHT never writes back — always enforced
Detection Engine
Configure anomaly detection sensitivity and classification thresholds.
Baseline Parameters
Baseline Window
Detection Sensitivity
CRITICAL Threshold
Cross-System Correlation
Trend Detection
Notifications
Email Alerts
CRITICAL Only
Daily Summary
Summary Time
Users & Roles
Live from Supabase database
Loading users…
Security & Compliance
POPIA Compliance
All processing adheres to SA POPIA requirements
Data Encryption
AES-256 at rest and in transit
Audit Logging
Data Residency
All data stored within South African borders
⚡ Live Intelligence Run
Watch JLIGHT process real laboratory data and generate a Decision Card in real time
📂 Input Data — LIMS Batch 7 · Analyser 2 · 14:20–14:32 SAST12 records
| Timestamp | Analyser | Batch | QC Mean | QC CV% | Reruns | Volume | Z-Score | Flag |
|---|
🔄 Intelligence Loop
📥IDLE
01 · Ingest
Waiting…
📐IDLE
02 · Baseline
Waiting…
🔍IDLE
03 · Detect
Waiting…
🔮IDLE
04 · Predict
Waiting…
🧠IDLE
05 · Reason
Waiting…
📋IDLE
05 · Deliver
Waiting…
🖥 System Log
[JLIGHT] System ready. Press ▶ Run Detection to begin.
◆ DECISION CARD GENERATED ◆
Detection Summary
JLIGHT — Built for Scale
JLIGHT is not just a prototype. It is a production-ready health intelligence platform designed to grow from a single laboratory to a national health network — without changing the core architecture.
⚖️ 10 International Standards
🔐 POPIA Compliant
🗄️ Supabase PostgreSQL
🌍 Built for Africa
📦 github.com/Fiadatumelo/jlight-health
5-Step Intelligence Loop
📥
01 · Ingest
LIMS · CTMS · HIS · QC
→
📐
02 · Baseline
7-day rolling adaptive
→
🔍
03 · Detect
Z-score · Correlation
→
🧠
04 · Reason
Root cause · Standards
→
📋
05 · Deliver
Decision Card · Database
🏥
Multi-Facility Architecture
Each facility operates in complete isolation using Supabase Row Level Security. A single JLIGHT deployment can serve hundreds of facilities simultaneously. Adding a new facility requires zero code changes — only a database record and data source configuration.
🔌
Data Source Plugins
New data sources — FHIR R4, DHIS2, OpenMRS, HL7 v2.5 — are added by registering the source in the database and writing one normalisation adapter function. The detection engine processes all sources identically regardless of origin.
📊
Detection Engine at Scale
The detection engine processes each facility independently using stateless workers. At scale, workers run in parallel via Supabase Edge Functions. Baseline calculations are cached per facility and updated only when new data arrives.
🌍
Built for the African Context
Data residency is configurable per country. Standards coverage includes SA, WHO, and ICH regulations. POPIA-compliant by design. Offline-capable for facilities with intermittent connectivity. Ready for DHIS2 public sector integration.
⚖️ International Standards Coverage10 standards · Automatically enforced
ISO 15189
Medical Laboratory Quality
QC CV limits, calibration intervals, non-conformance management
ICH E6(R2)
Good Clinical Practice
Informed consent timing, protocol deviation detection, monitoring obligations
ISO 14155
Clinical Investigation
Medical device trial consent compliance and documentation
ISO 22870
Point-of-Care Testing
TAT requirements aligned with clinical decision timelines
SA GCP / SAHPRA
SA Clinical Trial Regulations
15-day deviation notification, participant safety requirements
NCS Domain 6
National Core Standards
NDoH TAT compliance reporting and three-shift breach escalation
POPIA
Protection of Personal Information
SA data residency, audit logging, encryption enforcement
SANAS R47-01
SA Lab Accreditation
QC review after reagent lot change before clinical release
SANS 10212
Specimen Transport
Maximum time-to-analysis documentation requirements
HL7 v2.5
Health Data Interoperability
Observation request timing and data integrity standards
🚀 Production Roadmap
Phase 1 · Now
✅ Prototype Live
Real Supabase auth
Live database
Full detection loop
ISO standard flags
jlight-health.co.za
Phase 2 · 3–6 months
🔧 MVP Build
Next.js frontend
Real LIMS CSV import
Email notifications
Mobile-responsive
REST API layer
Phase 3 · 6–12 months
🏥 Production
HL7 v2.5 connector
FHIR R4 integration
Mobile app
Multi-facility dashboard
Enterprise SSO
Phase 4 · 12–24 months
🌍 National Scale
DHIS2 integration
OpenMRS connector
Predictive detection
SAHPRA auto-reporting
Pan-African deployment
🗄️ Database SchemaSupabase PostgreSQL · 7 tables
profiles
id · uuid (auth ref)email · textfull_name · textrole · textfacility_id · uuid
decision_cards
id · uuidcard_number · textseverity · crit/att/stabstatus · open/resolvedz_score · decimal
alerts
id · uuidcard_id · uuid (ref)message · textmetric_value · decimalbaseline_value · decimal
standards_violations
id · uuidcard_id · uuid (ref)standard_code · textdescription · textseverity · breach/risk
facilities
id · uuidname · texttype · textregion · textcountry · text (ZA)
audit_log
id · uuiduser_id · uuid (ref)action · textentity_type · textdetails · jsonb
View the Full Project on GitHub
Source code, database schema, architecture documentation, and README — all publicly available for judge review.