Table 10.
Characteristics of other included studies.
| Study | Country | Sample Size | Population | Model | Knowledge Source for AI | Type of Risk | Prediction Value of AI/AUC | Prediction Value of LR |
|---|---|---|---|---|---|---|---|---|
| Paydar et al. (2017) [75] | Iran | 149 | There was a retrospective analysis across pregnant women with systemic lupus erythematosus. For MLP (neuronal network-multi-layer perceptron), 70% were training cases, 15% were validation cases, and 15% were testing data. For RBF (radial basis function), 70% were training cases and 30% were testing data. |
CDSS (MLP, RBF) | Literature, specialist, patient’s history | High-risk pregnancy- (with SLE) | RBF 75.16%, MLP 90.6% |
NA |
| Li et al. (2017) [76] | China | 358 | There was a case-control study of pregnant women (119 fetuses with congenital heart disease and 239 controls). It was split into 85%, or n = 300 (101 cases and 199 controls), training cases and 15%, or n = 58 (18 cases and 40 controls), test cases. |
ANN − ANN + BPNN (backpropagation neural network) | Patient’s history, specialist | Congenital heart disease | Training 91%, testing 86% |
NA |
| Grossi et al. (2016) [77] | Italy | 137 | There was a retrospective analysis of pregnant women (45 mothers of autistic children and 68 mothers of typically developing children). A total of 24 siblings of 19 autistic children were an internal control group. All cases were used to test the model. |
Specialized ANNs (ANNs) | Interview of the mothers, literature, patient’s history | Autism | 80.2% | 46% |
| Valensise et al. (2006) [78] | Italy | 302 | There was a retrospective analysis of healthy post-term pregnancies (42 fetuses with labor distress and 260 without). All cases were used to test the model |
ANN | Patient’s history | Fetal distress | Accuracy 86%, positive predictive value 53%, negative predictive value 92% |
NA |
| Gao et al. (2019) [79] | United States | 45,858 | Electronic health record data of pregnant women were collected retrospectively. All cases were used to test the model. (NA if retrospectively or prospectively collected). |
Observational medical outcomes partnership | Patient’s history, internet | Severe maternal morbidity. | M0 AUC 0.790, M1 AUC 0.919, M3 AUC 0.937 |
NA |