SmartPD: An AI-Enhanced System for Optimizing Home-Based Continuous Ambulatory Peritoneal Dialysis

Authors:

Abstract
In this paper, SmartPD, which is an artificial intelligence of things-based monitoring system designed for continuous ambulatory peritoneal dialysis (CAPD) treatment at home, is proposed. SmartPD comprises a heating and thermal preservation module, an infusion detection and cutoff module, a turbidity identification module, and a display module. It efficiently raises and maintains the temperature of fresh dialysate at 37°C. Additionally, the system monitors infusion volume by detecting changes in the weight of the dialysate bag, automatically shutting off infusion upon reaching a preset threshold to prevent over- or under-infusion. SmartPD also utilizes deep learning technology to identify the turbidity of the drainage fluid from the abdominal cavity, assisting patients in determining whether peritoneal infection symptoms are present. These monitoring data are transmitted and recorded in the proposed cloud-based information management (CIM) platform. The experimental results show that SmartPD can quickly heat 2,000 c.c. of fresh dialysate to 37°C within 30 min, with overall turbidity recognition recall, precision, accuracy and area under the curve (AUC) values exceeding 0.95. Consequently, SmartPD improves the efficiency of CAPD treatment, helps patients assess the severity of associated peritoneal symptoms, and prevents the occurrence of peritonitis. The construction of the CIM platform meets the needs of hospitals for remote patient monitoring, thereby enhancing the quality of CAPD home treatment and reducing costs.
Funding Information
  • Ditmanson Medical Foundation Chia-Yi Christian Hospital, Chiayi, Taiwan, (I110-303, I112-206)

This publication has 24 references indexed in Scilit: