Scholarlink

<p>Aim: To develop a cost-effective, predictive model for hypertensive disorders of pregnancy (HDP) in advanced-aged pregnant women based on demographic and lifestyle factors. Methods: A large prospective, population-based, multicenter cohort study was conducted among advanced maternal-age pregnancies in China. Demographic and blood pressure data were collected from questionnaires of the first prenatal visits. The least absolute shrinkage and selection operator (Lasso) regression was applied for feature selection of risk factors, followed by XGBoost model construction and SHAP (SHapley Additive exPlanations) visualization. Results: Lasso regression identified 9 risk factors, including systolic blood pressure in the first trimester (SBP1), diastolic blood pressure in the first trimester (DBP1), body mass index (BMI), family history of hypertension, multiparous parity, age, alcohol assumption, assisted reproductive technology (ART), and screen use. The XGBoost model was set with an optimized tune grid. The AUC of the model was 0.82, AUPRC of 0.41, with an accuracy of 0.88, sensitivity of 0.46, and specificity of 0.92. The SHAP demonstrated a novel predictive performance and clinical applicability. Conclusion: The XGBoost-derived model offers a practical and simplified tool for individualized risk assessment in advanced maternal age pregnancies, facilitating early intervention and enhanced prenatal care.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Prediction of hypertensive disorders of pregnancy in advanced-age pregnant women using SHAP value and XGBoost

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月18日

查看全文
  • Jue Wang,
  • Hao Zhu,
  • Weirong Gu

摘要

Aim: To develop a cost-effective, predictive model for hypertensive disorders of pregnancy (HDP) in advanced-aged pregnant women based on demographic and lifestyle factors. Methods: A large prospective, population-based, multicenter cohort study was conducted among advanced maternal-age pregnancies in China. Demographic and blood pressure data were collected from questionnaires of the first prenatal visits. The least absolute shrinkage and selection operator (Lasso) regression was applied for feature selection of risk factors, followed by XGBoost model construction and SHAP (SHapley Additive exPlanations) visualization. Results: Lasso regression identified 9 risk factors, including systolic blood pressure in the first trimester (SBP1), diastolic blood pressure in the first trimester (DBP1), body mass index (BMI), family history of hypertension, multiparous parity, age, alcohol assumption, assisted reproductive technology (ART), and screen use. The XGBoost model was set with an optimized tune grid. The AUC of the model was 0.82, AUPRC of 0.41, with an accuracy of 0.88, sensitivity of 0.46, and specificity of 0.92. The SHAP demonstrated a novel predictive performance and clinical applicability. Conclusion: The XGBoost-derived model offers a practical and simplified tool for individualized risk assessment in advanced maternal age pregnancies, facilitating early intervention and enhanced prenatal care.