Realtime prediction of C-section risk for laboring mothers
summary
Cedars‑Sinai builds a real‑time C‑section risk pipeline that scores laboring mothers at admission and then updates real-time updates every 10 minutes using streaming EHR feeds, medication events, cervical exams and fetal heart tracings. An automated‑ML stack (random forest for admission, XGBoost for continuous) yields AUC 0.75 admission and improves to AUC 0.96 continuous after hours of labor. A Tableau dashboard shows per‑patient trajectories, feature attributions, and prospective validation and fairness monitoring (Fairlearn) are underway before clinical rollout.
AI-written summary · about
chapters (9)
- 0:00Introduction and speaker intro
- 1:05Project background and goals
- 9:53Target physician selection
- 12:00Model development roadmap
- 13:26Admission factors model results
- 16:33Intrapartum factors model results
- 20:43Productionalizing the model
- 25:40Ongoing work and validation
- 29:29Q&A session
related talks
23mEngineering AI-Powered Predictive Healthcare Systems
18mPipelines for Clinical AI: Real-Time Sepsis Detection
35mResponsible implementation of Machine Learning models for life-and-death decision making
25mLifesaving AI and JavaScript
24mDeveloping Innovative Systems Through Responsible AI In Healthcare
20mML applications in healthcare - Saudi Arabian hospital case study
54mCrossing the Last Mile with Predictable Hospital Readmissions
21mUsing R to develop production modeling workflows at Mayo Clinic