Stream-enabled chronic disease management at population scale
summary
NUHS runs CHAMP on the ENDEAVOUR AI platform using AWS Lex, Lambda and WhatsApp to stream vitals, send automated WhatsApp chatbot reminders and surface clinical alerts; models run on a streaming pipeline that nudges patients, ranks FAQ matches and supports branching dialogs up to 7–8 layers. System averages 9,000+ conversations/month with 10–14 turns and a fallback rate of 1.7% (industry ~3%), has reduced inappropriate pediatric A&E visits, and alerts clinicians when inputs (eg. BP 190/100) require escalation.
AI-written summary · about
chapters (8)
- 0:00Introduction and setup
- 1:06Healthcare's three Cs
- 4:59AWS healthcare cloud journey
- 12:53NUHS chat bot overview
- 23:01Chat bot branching logic
- 32:06Conversational chat bot design
- 36:45CHAMP program launch
- 43:52Summary and next steps
related talks
12mRevolutionizing Healthcare Population Management
46mReimagining clinical engagement using Amazon Connect
09mEnhance Healthcare Chatbot Transparency
26mAccelerating innovation in the healthcare and life sciences industry
31mBuilding SwasthaX: Engineering a Production-Scale AI Platform for Healthcare
45mCreando soluciones inteligentes en el sector salud
24mBuilding Cloud-based Solutions in Healthcare
23mEngineering AI-Powered Predictive Healthcare Systems