Optimizing CAPD Patient Monitoring Through Automated Vs Rule-Based Artificial Intelligence: A Systematic Comparative...

Area of Science:

• Nephrology
• Medical Informatics
• Artificial Intelligence

Background:

• Continuous Ambulatory Peritoneal Dialysis (CAPD) is a vital renal replacement therapy, particularly in developing nations.
• CAPD faces challenges like peritonitis and fluid overload, necessitating improved monitoring and decision support.
• Artificial intelligence (AI) presents potential solutions for enhancing CAPD management.

Purpose of the Study:

• To systematically review and compare rule-based AI systems and automatic machine learning approaches for CAPD monitoring.
• To evaluate AI's effectiveness in clinical outcomes, patient adherence, operational efficiency, cost, and usability.
• To explore automated AI for dialysate image classification in CAPD.

Main Methods:

• A systematic review of literature published between January 1, 2020, and May 20, 2025.
• Inclusion of studies from PubMed, Scopus, Google Scholar, and IEE Xplore, with 24 studies synthesized.
• Assessment of 14 eligible studies focusing on AI implementation for CAPD monitoring and management.

Main Results:

• Automated AI systems demonstrate superior precision and earlier detection capabilities.
• Rule-based AI models offer practical advantages in resource-limited healthcare environments.
• AI, including machine learning for dialysate image classification, shows promise in CAPD management.

Conclusions:

• A hybrid AI integration model combining rule-based and automated approaches is proposed for CAPD monitoring.
• This hybrid model aims to maximize clinical accuracy, cost-effectiveness, and accessibility.
• Tailoring AI solutions to national guidelines and insurance schemes is crucial for successful implementation.