AI-assisted Endotracheal Intubation

Intubation Performance Using AI Versus Conventional Videolaryngoscope in a Mannikin

intubation, mannikin, simulation, artificial intelligence, larynguide, neonates, infants, children

This study evaluates artificial intelligence (AI)-assisted videolaryngoscopy for endotracheal intubation in a simulated pediatric airway environment. Healthcare providers with varying levels of airway management experience will perform intubations on pediatric and neonatal mannequins using either AI-assisted videolaryngoscopy (larynGuide) or conventional videolaryngoscopy. Participants will be randomized to perform intubation tasks using one of the two techniques. The primary outcome is the time required for successful intubation. Secondary outcomes include first-attempt success rate, number of attempts, airway visualization (POGO score), usability of the AI system measured by the System Usability Scale (SUS), and gaze tracking metrics evaluating user interaction with visual guidance. This equivalence randomized controlled trial aims to determine whether AI-assisted videolaryngoscopy performs comparably to conventional videolaryngoscopy while potentially improving success rates and user experience.

Artificial intelligence is increasingly being applied in clinical medicine, including airway management. Machine vision algorithms have recently been developed to recognize airway anatomy and provide guidance during endotracheal intubation. LarynGuide is an AI-based system designed to guide endotracheal tube placement using real-time visual prompts during videolaryngoscopy. This study aims to evaluate whether AI-assisted intubation is equivalent to conventional videolaryngoscopy in terms of time required for intubation and whether AI guidance improves success rates or usability. This prospective randomized controlled simulation trial will recruit healthcare providers with varying levels of airway management experience. Participants will receive a short training session with the AI system and then will be randomized to perform intubations using either AI-assisted videolaryngoscopy or conventional videolaryngoscopy. Each participant will perform intubations on both pediatric and neonatal airway mannequins in a simulation setting. Outcomes including intubation time, success rate, number of attempts, POGO score, gaze tracking metrics, and user satisfaction will be collected. The study will be conducted at the Hospital for Sick Children Simulation Centre in Toronto, Canada.

No related papers found yet.

Canada