Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Multicenter Study
. 2025 Feb 28;25(1):225.
doi: 10.1186/s12884-025-07283-y.

Predicting outcomes of expectant and medical management in early pregnancy miscarriage using machine learning to develop and validate multivariable clinical prediction models

Sughashini Murugesu  1   2 , Kristofer Linton-Reid  3   4 , Emily Braun  3 , Jennifer Barcroft  3   5 , Nina Cooper  3   5 , Margaret Pikovsky  3   5 , Alex Novak  3   5 , Nina Parker  3   5 , Catriona Stalder  3 , Maya Al-Memar  3   5 , Srdjan Saso  3   5 , Eric O Aboagye  3   4 , Tom Bourne  3   4   6
Affiliations
Multicenter Study

Predicting outcomes of expectant and medical management in early pregnancy miscarriage using machine learning to develop and validate multivariable clinical prediction models

Sughashini Murugesu et al. BMC Pregnancy Childbirth. .

Abstract

Objective: To determine whether readily available patient, ultrasound and treatment outcome data can be used to develop, validate and externally test two machine learning (ML) models for predicting the success of expectant and medical management of miscarriage respectively.

Methods: A retrospective, multi-site study of patients opting for expectant or medical management of miscarriage was undertaken. A total of 1075 patients across two hospital early pregnancy units were eligible for inclusion. Data pre-processing derived 14 features for predictive modelling. A combination of eight linear, Bayesian, neural-net and tree-based machine learning algorithms were applied to ten different feature sets. The area under the receiver operating characteristic curve (AUC) scores were the metrics used to demonstrate the performance of the best performing model and feature selection combination for the training, validation and external data set for expectant and medical management separately.

Results: Parameters were in the majority well matched across training, validation and external test sets. The respective optimum training, validation and external test set AUC scores were as follows in the expectant management cohort: 0.72 (95% CI 0.67,0.77), 0.63 (95% CI 0.53,0.73) and 0.70 (95% CI 0.60,0.79) (Logistic Regression in combination with Least Absolute Shrinkage and Selection Operator (LASSO)). The AUC scores in the medical management cohort were 0.64 (95% CI 0.56,0.72), 0.62 (95% CI 0.45,0.77) and 0.71 (95% CI 0.58,0.83) (Logistic Regression in combination with Recursive Feature Elimination (RFE)).

Conclusions: Performance of our expectant and medical miscarriage management ML models demonstrate consistency across validation and external test sets. These ML methods, validated and externally tested, have the potential to offer personalised prediction outcome of expectant and medical management of miscarriage.

Keywords: Expectant management; Machine learning; Medical management; Miscarriage.

PubMed Disclaimer

Conflict of interest statement

Declarations. Ethics approval and consent to participate: This study was submitted to and approved by the Imperial College London Research Governance and Integrity Team (ICL RGIT), and the Health Research Authority and Health and Care Research Wales NHS, Approval Reference: 23/PR/0297. The need for consent to participate was waived by ICL RGIT due to the retrospective nature of the study. The research is in compliance with the Helsinki Declaration. Consent for publication: Not applicable. Competing interests: The authors declare no competing interests.

Figures

Fig. 1
Fig. 1
Predicting Outcome of Expectant Miscarriage Management. This figure includes a series of heatmaps, displaying the efficacy of various machine learning algorithms (displayed in columns) when combined with diverse feature reduction techniques (displayed in rows). Key for abbreviations: LASSO refers to Least Absolute Shrinkage and Selection Operator, E Net stands for Elastic-Net, RFE is Recursive Feature Elimination, Univariate LR corresponds to Univariate Logistic Regression, XGB represents Extreme Gradient Boosting Machine, NB is Naïve-Bayes, PSL corresponds to Partial Least Squares, L-SVM is Linear Support Vector Machine, NL-SVM is Non-linear (radial) SVM, RF refers to Random Forest, MDA denotes Mixture Discriminant Analysis, KNN signifies K-Nearest Neighbours, GLM is Generalised Linear Model, and NNET refers to Neural Network
Fig. 2
Fig. 2
ROC curves for the validation and test set for expectant management. Training Set AUC = 0.72 (95% CI 0.67,0.77). Validation Set AUC = 0.63 (95% CI 0.53,0.73). External Test Set AUC = 0.70 (95% CI 0.60,0.79)
Fig. 3
Fig. 3
Nomogram: Expectant Management of Miscarriage
Fig. 4
Fig. 4
Predicting Outcome of Medical Miscarriage Management. Heatmaps illustrating the performance of each machine learning algorithm (columns) with each feature reduction method (rows). Key for abbreviations: LASSO refers to Least Absolute Shrinkage and Selection Operator, E Net stands for Elastic-Net, RFE is Recursive Feature Elimination, Univariate LR corresponds to Univariate Logistic Regression, XGB represents Extreme Gradient Boosting Machine, NB is Naïve-Bayes, PSL corresponds to Partial Least Squares, L-SVM is Linear Support Vector Machine, NL-SVM is Non-linear (radial) SVM, RF refers to Random Forest, MDA denotes Mixture Discriminant Analysis, KNN signifies K-Nearest Neighbours, GLM is Generalised Linear Model, and NNET refers to Neural Network
Fig. 5
Fig. 5
ROC curves for the training, validation and test set for medical management. Training Set AUC = 0.64 (95% CI 0.56,0.72). Validation Set AUC = 0.62 (95% CI 0.45,0.77). External Test Set AUC = 0.71 (95% CI 0.58,0.83)
Fig. 6
Fig. 6
Nomogram: Medical Management of Miscarriage
Fig. 7
Fig. 7
Least Absolute Shrinkage and Selection Operator (LASSO) feature reduction feature reduction method used in combination with the logistic regression [21] model: Expectant Management of Miscarriage
Fig. 8
Fig. 8
Recursive Feature Elimination (RFE) feature reduction method used in combination with the logistic regression [21] model: Medical Management of Miscarriage
See this image and copyright information in PMC

References

    1. Quenby S, Gallos ID, Dhillon-Smith RK, et al. Miscarriage matters: the epidemiological, physical, psychological, and economic costs of early pregnancy loss. Lancet. 2021;397(10285):1658–67. - PubMed
    1. Farren J, Jalmbrant M, Falconieri N, et al. Posttraumatic stress, anxiety and depression following miscarriage and ectopic pregnancy: a multicenter, prospective, cohort study. Am J Obstet Gynecol. 2020;222(4):e3671–22. - PubMed
    1. Farren J, Jalmbrant M, Falconieri N, et al. Differences in post-traumatic stress, anxiety and depression following miscarriage or ectopic pregnancy between women and their partners: multicenter prospective cohort study. Ultrasound Obstet Gynecol. 2021;57(1):141–8. - PubMed
    1. Nielsen S, Hahlin M, Platz-Christensen J. Randomised trial comparing expectant with medical management for first trimester miscarriages. Br J Obstet Gynaecol. 1999;106(8):804–7. - PubMed
    1. Jurkovic D, Ross JA, Nicolaides KH. Expectant management of missed miscarriage. Br J Obstet Gynaecol. 1998;105(6):670–1. - PubMed

LinkOut - more resources