Clinical Decision Support for Total Parenteral Nutrition in Neonatal Intensive Care Unit Patients: A Pilot Study

What this Trial Is About

Researchers are evaluating an artificial intelligence (AI) tool designed to assist doctors in ordering total parenteral nutrition (TPN) for newborns and infants in the neonatal intensive care unit (NICU). Premature babies who cannot eat by mouth often need nutrition delivered through an IV, but ordering TPN is complex and prone to errors. This pilot study aims to assess whether the AI tool can improve the TPN ordering process by suggesting formulas based on each baby's lab values and health data, while doctors retain full control over final decisions. The study involves two periods: in the first, doctors use the current standard practice to order TPN without AI assistance; in the second, they use the AI-driven TPN decision support tool integrated within the electronic health record system. This tool recommends TPN compositions using patient lab results, age, weight, gestational age, and physician inputs. Doctors can accept, modify, or reject the AI suggestions. The study plans to enroll about 260 neonates who require TPN, typically for around 14 days. During participation, babies will have daily lab tests and continuous monitoring of health outcomes such as weight changes, days on TPN, and common complications in premature infants. Researchers will measure how often doctors accept AI recommendations, the time taken to complete TPN orders, and clinical outcomes including lung disease, brain bleeding, infections, and other conditions. Provider satisfaction and lab value abnormalities will also be tracked. The AI tool aims to support doctors without changing standard care practices, and each baby remains in the study while receiving TPN.

AI-driven Total Parenteral Nutrition Platform