Machine Learning-Based Predictive Model for Fear of Childbirth in Late Pregnancy.

Abstract

This study aimed to develop and validate a machine learning-based predictive model for assessing the risk of fear of childbirth in pregnant women during late pregnancy. A cross-sectional observational study was conducted from November 2022 to July 2023, involving 406 pregnant women. Six machine learning algorithms, including Lasso-assisted logistic regression (LR), random forest (RF), eXtreme Gradient Boosting (XGB), support vector machine (SVM), Bayesian network (BN), and k-nearest neighbors (KNN), were used to construct the models with 10-fold cross-validation. The results showed that the XGB model achieved the best performance, with an area under the receiver operating characteristic curve (AUC) of 0.874, accuracy of 0.795, sensitivity of 0.764, and specificity of 0.878. The LR model also performed well (AUC = 0.873). Key predictors of fear of childbirth included pain catastrophizing, expectation for painless childbirth, childbirth delivery preferences, medication use during pregnancy, and use of birth-related apps. The LR model was used to create a nomogram for clinical use. These machine learning models can help healthcare professionals identify and intervene early in cases of fear of childbirth.