Artificial intelligence in nutritional assessment and decision making

Purpose of the review – Artificial intelligence (AI) has become an non contourable tool in clinical nutrition practice. This review proposes to discuss the most recent advances that can support the physician in nutritional assessment and mainly physician decision support systems trying to predict and prevent nutritional related complications. Recent findings – Advanced data storage systems improve screening and assessment tools using large database analysis. Medical CT images analysis can determine patients suffering from sarcopenia and suggest outcome predictions accordingly. Decision making of the type of parenteral nutrition formula according to cluster obtained by machine of large databases has been shown to be superior to the prescription of neonatologists in preterm children. Machine learning can help to anticipate enteral feeding intolerance and predict enteral nutrition feeding success. Numerous digital technologies support analysis of the meal, allows for the passive monitoring of eating behaviors, including acoustic sensors for swallowing detection and motion sensors for tracking hand-to-mouth gestures. Summary – Each step in clinical nutrition, from screening to clinical decision-making, can be improved using AI. However, large data sources, the participation of data scientists, and advanced technologies are required. These improvements have the potential to transform clinical nutrition toward personalized nutrition, but AI integration should be carefully monitored to ensure patient benefit and safety.