Enhancing nutritional management in peritoneal dialysis patients through a generative pre-trained transformers-based...

Area of Science:

• Nephrology
• Nutritional Science
• Artificial Intelligence

Background:

• Peritoneal dialysis (PD) patients frequently experience nutritional deficiencies.
• Conventional nutritional management strategies often fall short of personalized patient needs.
• Micro-inflammation and nutrient loss contribute to malnutrition in PD patients.

Purpose of the Study:

• To develop and evaluate an AI-powered recipe generation application using Generative Pre-trained Transformers (GPT).
• To assess the efficacy of the AI tool in improving the nutritional status of PD patients.
• To determine the safety and practical value of AI in personalized nutritional management for chronic disease.

Main Methods:

• A self-controlled prospective study involving 35 PD patients.
• Two phases: conventional dietary education followed by intervention with a GPT-based dietary guidance tool.
• Nutritional status assessed by comparing serum prealbumin, albumin, and phosphate levels before and after intervention.

Main Results:

• Significant improvement in mean serum prealbumin levels (p=0.001).
• A non-significant increase in serum albumin levels.
• Serum phosphate levels remained stable and within safe limits (p=0.241).

Conclusions:

• The AI recipe generation application effectively improved prealbumin levels in PD patients.
• The tool demonstrated safety by maintaining stable phosphate levels.
• AI technology holds significant potential for personalized nutritional management in chronic disease patients.