Artificial Neural Network Modeling to Predict Neonatal Metabolic Bone Disease in the Prenatal and Postnatal Periods
Authors' conclusion
In this diagnostic study, ANN appeared to be a simple and efficient tool for identifying neonates at risk for MBD. Combining prenatal and postnatal factors or using postnatal exposures alone provided the most precise prediction. Extremely low birth weight was the most significant predictive factor, whereas magnesium sulfate use during pregnancy could be an important bellwether for MBD before delivery.
Study context
- Population
- Pregnant women with singleton pregnancy, followed up until 1 month after parturition, and their infants, Shanghai, China, 2012–2021; mean age 29.7 years; 138 (1.3%) infants with metabolic bone disease (MBD)
- Participants
- 10,801 mother-infant pairs
- Comparison
- 5 artificial neural network (ANN) models with different exposures from prenatal to postnatal periods
- Outcome
- Model performance for predicting neonatal metabolic bone disease (MBD), by AUC
Methodological detailsartificial neural network prediction models
- Artificial neural network prediction models
Performance
As reported in the abstract
Prediction of MBD (AUC)
| Test or model | AUC |
|---|---|
| Model 1 | 0.981 (95% CI, 0.970-0.992) |
| Model 2 | 0.647 (95% CI, 0.571-0.723) |
| Model 3 | 0.808 (95% CI, 0.726-0.891) |
| Model 4 | 0.850 (95% CI, 0.785-0.915) |
| Model 5 | 0.977 (95% CI, 0.966-0.988) |
AUC
Model 10.981 (95% CI, 0.970-0.992)
Model 20.647 (95% CI, 0.571-0.723)
Model 30.808 (95% CI, 0.726-0.891)
Model 40.850 (95% CI, 0.785-0.915)
Model 50.977 (95% CI, 0.966-0.988)
Further findings
Feature importance
- Extremely low birth weight importance (model 1; most powerful factor)
- 50.5%
- Magnesium sulfate use during pregnancy importance (model 4; most significant prenatal predictor)
- 21.2%
- Most important model 1 characteristics: birth weight, maternal age at pregnancy, neonatal disorders (anemia, respiratory distress syndrome, septicemia)
Original publication
Artificial Neural Network Modeling to Predict Neonatal Metabolic Bone Disease in the Prenatal and Postnatal Periods. JAMA Netw Open. Published 3 January 2023. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →