Innovating Clinical Research with AI and Real World Data
summary
Describes using double machine learning on an observational cohort of 2–3k patients treated with secukinumab to predict individual quality‑of‑life gains and produce a simple decision tree that guides when to up‑dose; reports the model deployed to clinicians with positive uptake. Describes building digital twins that simulate a patient’s outcome on standard‑of‑care to create virtual control arms, validated retrospectively then prospectively with plans to shift randomization from 50/50 toward 80/20. Notes barriers: siloed data, 97% unused, and Germany EHR adoption ~65%.
AI-written summary · about
chapters (8)
- 0:00AI and health data in clinical research
- 1:07Clinical trials failure and AI role
- 3:16Health data abundance challenge
- 5:47Secukinumab precision medicine case
- 10:22Digital twins for trial design
- 15:00AI adoption barriers in healthcare
- 16:25Call to action for innovation
- 17:07Q&A
related talks
26mFrom Pixel to Patient: Making Medical AI Work in the Real World
59mKeynote Panel AI in Healthcare
36mReimagining the Future of Clinical Trials
28mAI in healthcare - Deep learning for precision psychiatry
19mThe future of AI in powering clinical trial innovation
24mDeveloping Innovative Systems Through Responsible AI In Healthcare
50mA Sustainable Future for Learning AI in Healthcare
16mMultimodal AI: Transforming Patient-Centered Healthcare