AI-Driven Solutions in Kidney Disease Management: A Systematic Literature Review

Abstract

Kidney disease is becoming more common worldwide. The primary reason for this rise is the very high rates of diabetes and hypertension, which are the main risk factors for chronic kidney disease (CKD) and are responsible for endstage kidney disease. During this final stage, patients are mostly treated with kidney transplantation or dialysis (hemodialysis or peritoneal dialysis), both of which are very costly treatments and have limited availability in different geographical areas. However, non-adherence to treatment may lead to more complications and, thus, hospital visits, while there is also a high risk of death. The combination of artificial intelligence and telemedicine in the nephrology area has really opened up a new perspective for kidney disease treatment. The authors of this article wanted to find out how exactly artificial intelligence helps in enhancing kidney disease management via a literature review that follows a systematic approach. The literature review we conducted incorporates papers published between 2014 and 2024, and the papers were sourced from various databases, including PubMed, Taylor & Francis, IEEE Xplore, Google Scholar, and others. According to the 2020 PRISMA statement, a total of 48 research publications were finally selected for the review, and these were organized into seven thematic categories of AI applications in kidney disease. A comparative analysis points out the following trends: predictive modeling, AI-assisted monitoring systems, generative AI, and digital tools for patient education and clinical support. Nevertheless, healthcare systems still face security, quality control, ethical, privacy, and clinical variability challenges, all of which impact the successful adoption of AI. Despite these limitations, current evidence suggests that artificial intelligence may have an important role in nephrology, with the need for future studies on larger, diverse, and real-time multicenter datasets with multimodal data integration.