ML Models Overestimate Preeclampsia Prediction Transferability
This systematic review and meta-analysis (searches through February 2025) evaluated 31 machine-learning models for predicting preeclampsia across 26 studies, finding a pooled AUC of 0.91 (95% CI 0.87–0.92) but extreme heterogeneity (I2>99%). Prediction intervals for sensitivity ranged widely (0.32–0.96), and external-validation studies (n=6) showed lower pooled sensitivity (0.68; PI 0.25–0.94). The authors call for multicenter prospective external validation and recalibration to improve transferability.
Scoring Rationale
Comprehensive, peer-reviewed meta-analysis with robust PI estimates, but limited external validation and high heterogeneity reducing generalizability.
Practice interview problems based on real data
1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
