Key result
Support vector machine model predicts the need for delivery within 7 days with ~77% sensitivity.
Why the study?
The timing of hospitalization, corticosteroids, and delivery in early onset preeclampsia remains a clinical challenge requiring improved prediction tools.
Do advanced machine-learning models improve the prediction of delivery within 7 days and HELLP/abruptio placentae compared to basal demographic models in patients with early onset preeclampsia?
Population
215 singleton pregnancies with early onset preeclampsia and attempted expectant management
Comparison
Advanced machine-learning models with diagnostic data vs basal models with demographic characteristics only
Design
Retrospective cohort study using machine-learning models with evolutionary feature selection
Follow-up
Median 8 days
Authors
Loading...
May aid delivery timing decisions in early-onset preeclampsia; hypothesis-generating until validated against simpler models in prospective studies.
Cohort (n=215)
Do advanced machine-learning models improve the prediction of delivery within 7 days and HELLP/abruptio placentae compared to basal demographic models in patients with early onset preeclampsia?
Machine learning models using routinely obtained maternal characteristics and markers at diagnosis can accurately predict the need for delivery within 7 days and the development of HELLP/abruptio placentae in early-onset preeclampsia.
Villalaín et al. (2022) conducted a cohort in Early onset preeclampsia (n=215). Advanced machine-learning models (Support Vector Machine) at diagnosis vs. Basal models (demographic characteristics only) was evaluated on Need for delivery within 7 days of diagnosis and risk of HELLP syndrome or abruptio placentae. A support vector machine model using maternal characteristics and markers at diagnosis predicted the need for delivery within 7 days with 77.3% sensitivity and 80.1% specificity.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: