Search saves lives: solving healthcare problems with search

on YouTube ↗37m

" summary

NHS testing moves from straight‑line distance to travel‑time matching using a driving model + graph traversal that returns driving times for up to 100,000 locations per request, serves peaks of 100,000 concurrent users, and delivers results under 50 milliseconds by switching JSON to protobuf, running in‑memory on dedicated redundant servers (six instances) with no PII written to disk. It operates globally (92 countries), ingests GTFS from ~15,000 agencies (3.5M stops, 200k routes), and powers 111/999, job and workforce matching; it works in production.

AI-written summary · about

" chapters (28)

1. 0:00Introduction and speaker background
2. 1:25Agenda and talk overview
3. 2:08COVID-19 testing challenges in 2020
4. 3:57Problems with straight-line distance search
5. 6:27Limitations of bounding boxes and polygons
6. 8:38Searching using transport networks
7. 12:32Requirements for healthcare search solution
8. 16:28Building the driving time model
9. 17:55Handling large location datasets
10. 18:58Achieving sub-50ms response times
11. 19:39Scaling for high concurrent demand
12. 21:01Ensuring user data privacy
13. 21:42Customizing the solution for healthcare
14. 22:44Live implementation and results
15. 23:48Expanding beyond COVID-19 testing
16. 25:53Searching by time vs distance for transport
17. 27:20Challenges of public transport data
18. 28:23Global public transport coverage
19. 29:25Additional transport modes: cycling
20. 30:09Multimodal transport search capabilities
21. 30:31Workforce planning use cases
22. 31:13Emergency response and mobile doctors
23. 31:55Healthcare planning at scale
24. 32:41Why search relevancy matters in healthcare
25. 33:01Future of time-based healthcare search
26. 33:43Q&A: Offline pre-calculations
27. 34:04Q&A: User feedback and model improvements
28. 34:48Q&A: Handling road closures and events