In this study, Wang et al. tests two clinical prediction frameworks to develop an early alert system for intradialytic hypertension (IDH). Models were developed using the following machine learning algorithms; Light Gradient Boosting Machine (LGBM), Support Vector machine (SVM), and TabNet. The first framework (IDH-1) estimates hypertension risk by analyzing pre-dialysis vital signs and longitudinal treatment patterns, while the second framework (IDH-2) predicts future risk by analyzing real-time dialysis parameters with historical biomarkers. Model performance was validated using AUC-ROC, sensitivity, accuracy, and F1 score. After using 185,125 hemodialysis (HD) sessions as training data and 71,427 sessions as testing data, the LGBM model for both frameworks achieved the highest performance. For IDH-1, the LGBM model had superior discriminative capacity (AUC: 0.87; recall: 0.73; F1 score: 0.36) and had significant parameters of pre-dialysis diastolic pressures, historical mean arterial pressure, and historical average IDH episodes. For the IDH-2 model, an AUC of 0.74 was achieved with the most important parameters being historical average IDH episodes and post-dialysis systolic pressures. Notably, this study does include some limitations for its retrospective multicenter approach such as, reliance on data from Asian populations in the dataset used.