Table 1 Summary of the reviewed literature
From: Presenting a prediction model for HELLP syndrome through data mining
Study/Year | Country | Methods Used | Datasets | Performance Metrics | Outcomes |
|---|---|---|---|---|---|
Moreira/2021[1] | Brazil | ANNs and fuzzy logic | This study considered 205 parturient women diagnosed with HDP. Among these, seven pregnant women presented HELLP syndrome. | Precision, Recall, F1 score, AUC | Performance of the neuro-fuzzy model for HELLP syndrome: Precision: 0.692 Recall: 0.600 F1-score: 0.643 AUC: 0.807 |
Melinte-Popescu/2023 [20] | Romania | DT, NB, K-NN, and RF | The 161 pregnant patients affected by HELLP syndrome were subsequently divided into the following subgroups according to the Mississippi classification: Subgroup 1 (Class 1, n = 21), subgroup 2 (Class 2, n = 35), and subgroup 3 (Class 3, n = 25) | Precision, AUC, F1 score | F1 score for all HELLP types: DT: 80% RF: 89% NB: 87% KNN: 87% |
Melinte-Popescu/2023 [21] | Romania | DT, NB, SVM, RF | A total of 233 patients were included in the study. | AUC, Precision, Recall, F1 Score | F1 Score for all types of preeclampsia: DT: 0.93 NB: 0.98 SVM: 0.88 RF: 0.93 |
Villalaín/2022 [22] | Spain | SVM, KNN, GNB, DT | A total of 215 patients were included, among them, 103 required delivery within seven days of diagnosis. | Sensitivity, Specificity, AUC | K-NN in D1: precision of 0.68 ± 0.09, with 63.6% sensitivity, 71.4% specificity. Models at diagnosis SVM D2: precision improved to 0.79 ± 0.05 with 77.3% sensitivity and 80.1% specificity. |
Zheng/2022 [23] | Italy | LR, KNN, RF, MLP, SVM, DT, Linear discriminant analysis (LDA) | A dataset containing 733 women diagnosed with preeclampsia. Participants were grouped by four diferent pregnancy outcomes. | Sensitivity, Specificity, AUC | AUC for Adverse maternal outcome: KNN: 0.911 DT: 0.908 RF: 0.963 SVM: 0.976 MLP: 0.973 LDA: 0.961 LR: 0.958 |
Chen/2022 [24] | China | RF, C5.0, bagged CART, boosted trees, KNNs, neural network, flexible discriminant analysis, boosted LR, naïve Bayesian, single C5.0 tree, boosted generalized linear model, elastic net, partial least squares, nearest shrunken centroids, bagged MARS, and tree models from genetic algorithms | This study included pregnant women who delivered at the Tongzhou Maternal and Child Health Care Hospital of Beijing | Sensitivity, Specificity, AUC | The maternal model using the RF algorithm produced an AUC of 0.984 |
Huang/2022 [25] | China | ANN | A total of 270 pregnant patients with preeclampsia complicated by Fetal growth restriction who delivered in the Obstetrics Department of the Fujian Provincial Maternity and Children’s Hospital between January 2010 and December 2018 were selected as the study group | Accuracy, Sensitivity, Specificity, F1 Score | Preeclampsia complicated by fetal growth: Accuracy: 84.3% sensitivity = 97.7% specificity = 78% F1 score = 81% |
Ejiwale/2021[26] | USA | LR, RF, DT, SVM, SC, and Keras Classifier | This study analyzed a sub-sample (n = 4624, n_features = 38) of a labelled maternal perinatal dataset collected by the PeriData.Net® database from a participating community hospital in Southeast Wisconsin | Precision, Recall, AUC, F1 score | For F1, its highest mean score was from DTree Model Set 2 (4.93%) |
Marik/2020 [27] | USA | Elastic net, gradient boosting algorithm | A total of 16,370 records were used in this study | AUC, sensitivity | The obtained prediction model for preeclampsia yielded an AUC of 0.79 and sensitivity of 45.2%. |
Moreira/2019 [28] | Brazil | NB, TAN, average one dependence estimators | Not reported | Precision, F1 Score, AUC, MCC | F1 Score for prediction of the delivery outcome for the pregnant woman: NB: 0.923 TAN: 0.950 average one dependence estimators: 0.944 |