ML Model May Predict Preeclampsia Risk
Dynamic EHR-based approach may enable earlier intervention in late pregnancy.
-
By
-
Olivia Anderson
-
March 18, 2026
-
Clinical Scorecard: ML Model May Predict Preeclampsia Risk
At a Glance
| Category | Detail |
|---|
| Condition | Preeclampsia |
| Key Mechanisms | Machine learning model using electronic health record data to predict risk. |
| Target Population | Pregnant individuals in late pregnancy. |
| Care Setting | Multisite hospitals within a single health system. |
Key Highlights
- Preeclampsia affects 2% to 8% of pregnancies globally.
- Machine learning models demonstrated good discrimination for predicting preeclampsia onset.
- Blood pressure is the most influential predictor across all time points.
- Model performance improved through the third trimester, peaking at 34 weeks.
- High negative predictive values indicate strong ability to rule out near-term risk.
Guideline-Based Recommendations
Diagnosis
- Utilize machine learning models to assess risk dynamically as new data becomes available.
Management
- Consider earlier intervention based on updated risk assessments.
Monitoring & Follow-up
- Implement closer monitoring for patients identified at higher risk.
Risks
- Retrospective design limits findings; requires prospective validation.
Patient & Prescribing Data
Pregnant individuals monitored for preeclampsia risk.
Potential for tailored interventions based on site-specific data.
Clinical Best Practices
- Incorporate routinely available clinical variables in risk assessment.
- Adapt machine learning models to local clinical settings for improved accuracy.
- Continuously update risk assessments as new clinical data is available.
References
