Artificial Intelligence and Machine Learning in Predicting Intradialytic Hypotension in Hemodialysis Patients: A Systematic Review

Intradialytic hypotension (IDH) is a common and potentially life-threatening complication in hemodialysis patients. Traditional preventive measures have shown limited effectiveness in reducing IDH incidence. This systematic review evaluates the existing literature on the use of artificial intelligence (AI) and machine learning (ML) models for predicting IDH in hemodialysis patients. A comprehensive literature search identified five eligible studies employing diverse AI/ML algorithms, including artificial neural networks, decision trees, support vector machines, XGBoost, random forests, and LightGBM. These models utilized various features such as patient demographics, clinical data, laboratory findings, and dialysis-related parameters. The studies reported promising results, with several models achieving high prediction accuracies, sensitivities, specificities, and area under the receiver operating characteristic curve values for predicting IDH. However, limitations include variations in study populations, retrospective designs, and the need for prospective validation. Future research should focus on multicenter prospective studies, assessing clinical utility, and integrating interpretable AI/ML models into clinical decision support systems.