Artificial Intelligence in The Care of Kidney Failure Patients Undergoing Dialysis

Abstract

Artificial intelligence (AI) is increasingly transforming the care of patients with kidney failure undergoing dialysis, driven by advances in machine learning (ML) and deep learning (DL), alongside the availability of large, structured, and longitudinal clinical datasets and advances in computational resources. Dialysis practice is uniquely suited to AI applications due to its data-dense environment, characterised by repetitive treatments, continuous physiological monitoring, and integration with electronic health records and global registries. Early rule-based systems from the 1980s have evolved into sophisticated data-driven models capable of capturing complex, non-linear clinical relationships.

This review examines contemporary and emerging AI applications across the dialysis care continuum, including dialysis initiation, vascular access planning and surveillance, dialysis adequacy assessment, peritoneal dialysis peritonitis management, nutritional optimisation, intradialytic haemodynamic stability, anaemia management, and prediction of cardiovascular events and mortality. Across these domains, AI models—ranging from artificial neural networks to ensemble and reinforcement learning approaches—have consistently demonstrated superior predictive performance compared with traditional statistical methods, with potential to enable real-time, individualised clinical decision-making capable of continuous performance improvement.

AI-driven tools are increasingly transitioning from retrospective risk prediction to prospective, actionable decision support, including integration into dialysis machines and clinical dashboards. These developments support a shift toward proactive, personalised, and adaptive dialysis care. However, significant challenges remain, including data heterogeneity, limited external validation, model interpretability, workflow integration barriers, and concerns regarding data governance, privacy, and medico-legal accountability.

Future progress will depend on multicentre validation, harmonisation of registry data, and development of explainable, clinician-in-the-loop systems. With these advances, AI has the potential to redefine dialysis care through continuously learning, patient-centred clinical ecosystems.