Utilisation of AI in CKD Management

Chronic kidney disease (CKD) represents a major global health burden, affecting approximately 10.0% of Taiwanese adults (1.91 million individuals). The rising prevalence of CKD and end-stage kidney disease (ESKD) imposes substantial clinical and economic burdens while also contributing to a significant environmental footprint due to the resource-intensive nature of long-term dialysis. In this context, artificial intelligence (AI)-driven strategies that enhance early detection and clinical decision-making have the potential to improve patient outcomes while promoting more sustainable healthcare delivery.

At our institution, several AI-based systems have been implemented across the continuum of CKD care. To address acute kidney injury (AKI), a major contributor to CKD progression, we integrated an AI-powered AKI warning system into the inpatient computerized physician order entry system. The platform continuously monitors laboratory data and generates alerts based on KDIGO-defined AKI criteria. Implementation of this system reduced nursing workload by approximately 75 hours per month and achieved a renal recovery rate of 58% among identified patients.

We further developed AI-assisted prediction models for sepsis-associated AKI (SA-AKI), acute kidney disease (SA-AKD), and sepsis-associated CKD (SA-CKD). These models incorporate serum creatinine, blood urea nitrogen (BUN), and peripheral immune cell phenotype data to identify high-risk patients and facilitate early targeted intervention.

For patients receiving maintenance hemodialysis, we implemented a contactless AI physiological monitoring platform within the dialysis unit. Trained using clinical data from more than 1,600 patients, the system continuously monitors respiratory rate, heart rate, and blood pressure throughout dialysis sessions. It generates automated alerts up to 30 minutes before the onset of complications, such as intradialytic hypotension and muscle cramping, enabling timely preventive intervention. In addition, AI-powered robots have been introduced to support patient education and assist with transportation within the dialysis unit.

By integrating AI technologies across both inpatient and dialysis settings, our institution has established a comprehensive framework for CKD management that strengthens clinical surveillance, enables timely intervention, improves operational efficiency, and supports the sustainable use of healthcare resources.