An Ontology and Rule-Based Clinical Decision Support System for Personalized Nutrition Recommendations in the Neonatal Intensive Care Unit

Abstract

Premature neonates hospitalized in the neonatal intensive care unit (NICU) have high nutritional needs to ensure optimal growth and development. Monitoring adherence to neonatal nutritional guidelines and providing personalized nutritional recommendations are essential for promoting their health. However, ensuring consistent adherence to these guidelines is challenging. To address this, we developed a clinical decision support system using an ontology and rule-based approach to offer personalized nutrition recommendations for preterm infants in the NICU. The Nutrition Recommendation Ontology (NRO) was developed, incorporating 121 classes, 366 axioms, and 157 semantic rules based on standard nutrition guidelines and retrospective data from 601 NICU patients, using a cumulative 8460 patient-days data collected from a single center between 2019-2021. While the ontology represented the conceptual knowledge, the rules encoded the procedural knowledge. The integrated NRO-based system serves as a reasoning engine, enabling the generation of patient-specific feeding recommendations and assisting in identifying deviations from established guidelines. To validate its efficacy, the NRO system was tested on 10 sample case studies and achieved 98% accuracy, as assessed by a panel of neonatologists. Our findings indicate that the NRO-based clinical decision support system can provide accurate personalized nutrition recommendations and assess guideline adherence in the NICU. With further real-world validation, we anticipate that this approach could significantly improve nutrition delivery, prevent malnutrition, and ultimately improve outcomes for preterm infants.

Keywords

• Premature neonates
• ontology
• clinical decision support
• semantic
• nutrition management