Ali Amirsavadkouhi

Abstract

Background: Nutritional management in critically ill patients remains one of the most complex and impactful aspects of ICU care. Traditional approaches often fall short in adapting to real-time patient variability, leading to underfeeding, overfeeding, or suboptimal metabolic outcomes. With the advent of artificial intelligence (AI), there is now an opportunity to revolutionize ICU nutrition strategies. Objective: This presentation aims to explore how AI-based systems can personalize and optimize nutritional interventions for ICU patients by integrating dynamic clinical data, metabolic biomarkers, and predictive modeling to improve outcomes. Methods: Drawing from recent advancements in AI and digital health, we review case-based examples and pilot data from critical care settings where AI models have been utilized for real-time caloric assessment, macro/micronutrient adjustment, and enteral/parenteral feeding decisions. Integration with EHR and ICU monitoring platforms enables continuous learning and decision support. Results: AI-driven nutrition algorithms have shown promising improvements in glycemic control, reduced ICU length of stay, and decreased incidence of feeding-related complications. Additionally, personalized nutrition protocols based on AI predictions demonstrate potential in reducing mortality and improving long-term recovery. Conclusion: The incorporation of AI in critical care nutrition marks a paradigm shift from reactive to proactive and personalized ICU care. This talk will highlight practical frameworks for AI integration, challenges in implementation, and future pathways toward a data-driven ICU nutrition ecosystem.