Overview
System Purpose
Real-time AI detection of unusual disease clusters across UAE health facilities, enabling outbreak response within 24 hours rather than 7-14 days.
Data Sources
- Hospital admission records (near-real-time API feeds)
- Pharmacy dispensing data (daily batch)
- Emergency department chief complaint logs
- Laboratory test orders (not results, for privacy)
Daily Operations
Morning Check (7:30 AM)
- Open surveillance dashboard
- Review overnight alerts (threshold: 2x normal rate for any syndrome)
- For each alert: mark as reviewed + classify (noise / signal / escalate)
- Forward escalated signals to Public Health Officer
Syndrome Categories Monitored
| Syndrome | Trigger (7-day rolling) |
|---|---|
| Respiratory | 15% above seasonal baseline |
| Gastrointestinal | 20% above baseline |
| Fever + rash | Any cluster of 5+ within 48h |
| Neurological | Any cluster of 3+ within 48h |
Escalation Protocol
Level 1 (Monitor): 1.5x baseline → Continue daily review Level 2 (Alert): 2x baseline → Notify Regional Health Director within 4h Level 3 (Emergency): 3x baseline or novel syndrome → Notify MOHAP Incident Command within 1h
Weekly Tasks
- Review false positive rate (target: < 20% of alerts are noise)
- Update seasonal baselines
- Generate weekly epidemiological summary
Monthly Tasks
- Sensitivity analysis: did we catch any real outbreaks?
- Refine detection algorithms based on false positive/negative review
- Update contact lists for notification chain