What's Happening?
Researchers at the Icahn School of Medicine at Mount Sinai have developed an artificial intelligence (AI) tool named NutriSightT, designed to predict which critically ill patients on ventilators are at risk
of underfeeding. The study, published in Nature Communications, highlights the importance of providing adequate nutrition during the first week of ventilation, a period when patients' nutritional needs can change rapidly. The AI tool analyzes routine ICU data, including vital signs, lab results, medications, and feeding information, to predict underfeeding risks hours in advance. The model updates predictions every four hours, allowing clinicians to intervene early and adjust care to ensure patients receive the necessary nutritional support. The study found that underfeeding is common, with 41 to 53 percent of patients underfed by day three and 25 to 35 percent by day seven. The tool aims to serve as an early-warning system to guide timely nutrition interventions without replacing clinicians.
Why It's Important?
The development of NutriSightT represents a significant advancement in personalized healthcare, particularly in the intensive care unit (ICU) setting. By enabling early identification of patients at risk of underfeeding, the tool can help improve recovery outcomes and reduce complications associated with malnutrition in critically ill patients. This approach aligns with the broader trend towards personalized medicine, where treatments and interventions are tailored to individual patient needs. The potential to integrate such AI tools into electronic health records could streamline clinical workflows and enhance decision-making processes. The research underscores the growing role of AI in healthcare, offering a model for how technology can support clinicians in delivering more effective and efficient care.
What's Next?
The research team plans to conduct prospective multi-site trials to evaluate whether acting on the AI tool's predictions can improve patient outcomes. They also aim to integrate NutriSightT into electronic health records and expand its application to broader individualized nutrition targets. These steps are crucial for validating the tool's effectiveness in real-world settings and ensuring its seamless adoption in clinical practice. The outcomes of these trials could influence future guidelines and standards for nutritional care in ICUs, potentially leading to widespread implementation of AI-driven solutions in healthcare.
