Development and Validation of a Predictive Model for Maternal Cardiovascular Morbidity Events in Patients With Hypertensive Disorders of Pregnancy
Authors' conclusion
Among individuals with HDP, our multilabel neural network model predicted CV events at delivery admission with good classification and events within 1 year of delivery with fair classification.
Study context
- Population
- Deliveries with an hypertensive disorders of pregnancy (HDP), discharge from October 1, 2015 to December 31, 2020deliveries with an HDP
- Participants
- 553,658 deliveries
- Comparison
- Index Model vs Readmission Model
- Outcome
- Prediction of all CV events and of heart failure, renal failure and cerebrovascular events after HDP
Methodological detailsmachine learning, multilabel neural network, training/validation/test split (60%/20%/20%), retrospective database (Premier Healthcare Database)
- Machine learning
- Multilabel neural network
- Training/validation/test split (60%/20%/20%)
- Retrospective database (Premier Healthcare Database)
Performance
As reported in the abstract
All CV events
| Test or model | Area under the receiver operating curve (AUROC) | Average precision (AP) | Specificity at ≥90% sensitivity |
|---|---|---|---|
| Index Model | 0.878 | 0.239 | 65.0% |
| Readmission Model | 0.717 | 0.022 |
Area under the receiver operating curve (AUROC)
Index Model0.878
Readmission Model0.717
Average precision (AP)
Index Model0.239
Readmission Model0.022
Specificity at ≥90% sensitivity
Index Model65.0%
Readmission Model— not stated in the abstract
Further findings
| Outcome | Index Model | Readmission Model |
|---|---|---|
| Cerebrovascular events | ||
| AUROC | 0.941 | 0.698 |
| Specificity at ≥90% sensitivity | 83.5% | — not stated in the abstract |
| Heart failure | ||
| AUROC | 0.898 | 0.734 |
| Specificity at ≥90% sensitivity | 68.6% | — not stated in the abstract |
| Renal failure | ||
| AUROC | 0.885 | 0.748 |
| Specificity at ≥90% sensitivity | 65.6% | — not stated in the abstract |
— not stated in the abstract
Cardiovascular (CV) event incidence
- CV event during delivery hospitalization
- 6501 (1.2%)
- CV events within 1 year of delivery
- 3018 (0.6%)
Feature importance
- Chronic renal disease, cardiac disease, pulmonary hypertension and preeclampsia with severe features had the greatest effect on prediction of CV events
Original publication
Development and Validation of a Predictive Model for Maternal Cardiovascular Morbidity Events in Patients With Hypertensive Disorders of Pregnancy. Anesth Analg. Published 6 November 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →