ML Model Predicts Third-Trimester Preeclampsia Risk
Weill Cornell Medicine investigators publish March 6 in JAMA Network Open a machine-learning model that continuously predicts short-term preeclampsia risk using electronic health record data collected late in pregnancy. Trained on 35,895 deliveries and evaluated across nearly 59,000 pregnancies at three NewYork-Presbyterian hospitals, the model best predicts risk around 34 weeks and may give clinicians lead time for monitoring and delivery planning.
Key Points
- 1Develops dynamic ML model using EHRs from ~59,000 pregnancies to predict short-term preeclampsia.
- 2Shows highest accuracy near 34 weeks, addressing gap in late-onset and term preeclampsia prediction.
- 3Enables continual risk updates for clinicians, informing enhanced monitoring, blood pressure management, and delivery planning.
Scoring Rationale
Strong peer-reviewed evidence and actionable clinical utility, tempered by limited external generalizability beyond the three NewYork-Presbyterian hospitals.
Sources
Public references used for this report.
Practice with real Health & Insurance data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Health & Insurance problems
