KMAN Publication Inc. (KMANPUB) Psychology of Woman Journal 3041-8615 In Press 2027 03 01 Machine Learning Prediction of Anxiety Symptoms in Pregnant Women Based on Pregnancy-Specific Stress, Social Support, and Health Literacy Machine Learning Prediction of Anxiety Symptoms in Pregnant Women Based on Pregnancy-Specific Stress, Social Support, and Health Literacy 1 14 10.61838/ EN Journal Article 2026 04 21 <p><strong>Objective:</strong> The present study aimed to develop and evaluate machine learning models for predicting anxiety symptoms among pregnant women based on pregnancy-specific stress, social support, and health literacy and to determine the relative importance of these predictors in identifying women at risk for elevated anxiety during pregnancy.</p> <p><strong>Methods and Materials: </strong>This applied cross-sectional predictive study was conducted among 384 pregnant women attending prenatal care centers and obstetric clinics in Tehran. Participants were selected through multistage convenience sampling and completed a demographic questionnaire, the Beck Anxiety Inventory, a Pregnancy-Specific Stress Questionnaire, the Multidimensional Scale of Perceived Social Support, and the Health Literacy for Iranian Adults Questionnaire. Following data preprocessing and descriptive analyses, machine learning techniques were employed to predict anxiety symptoms. The dataset was divided into training and testing subsets, and several supervised learning algorithms, including Logistic Regression, Support Vector Machine, K-Nearest Neighbors, Decision Tree, Random Forest, Gradient Boosting, and Extreme Gradient Boosting (XGBoost), were developed and evaluated using cross-validation procedures. Model performance was assessed using accuracy, precision, recall, F1-score, and area under the receiver operating characteristic curve (AUC). Feature importance analyses were conducted to identify the relative contribution of predictor variables.</p> <p><strong>Findings: </strong>Correlation analyses revealed a significant positive relationship between pregnancy-specific stress and anxiety symptoms and significant negative relationships between anxiety symptoms, social support, and health literacy. All machine learning algorithms demonstrated acceptable predictive performance; however, ensemble learning methods substantially outperformed traditional models. The XGBoost algorithm achieved the highest predictive accuracy (93%), precision (91%), recall (90%), F1-score (90%), and AUC (0.97), indicating excellent classification performance. Feature importance analysis showed that pregnancy-specific stress was the strongest predictor of anxiety symptoms, followed by social support and health literacy. Together, these findings demonstrated that psychosocial and informational factors could accurately predict anxiety risk among pregnant women and that advanced machine learning techniques effectively captured the complex relationships among these variables.</p> <p><strong>Conclusion: </strong>Pregnancy-specific stress, social support, and health literacy are significant predictors of anxiety symptoms during pregnancy, with pregnancy-specific stress representing the most influential factor. Machine learning models, particularly XGBoost, provide highly accurate prediction of maternal anxiety risk and may serve as valuable tools for early screening, prevention, and targeted intervention within prenatal healthcare settings. Strengthening social support systems and improving maternal health literacy may help reduce anxiety symptoms and enhance psychological well-being during pregnancy.</p> https://www.journals.kmanpub.com/index.php/psywoman/article/download/5747/10935