Feature Selection and Classification Models for Endotracheal Intubation Training Devices

Introduction: Endotracheal intubation (ETI) is a medical procedure by which a tube is inserted into the trachea to open the airway and support breathing. Since unsuccessful attempts and prolonged time to complete can lead serious complications, continuous ETI training is required, even for senior doctors, to acquire and refine the skills for ETI to be performed in a short period of time. However, current practice of ETI training relies heavily on the use of bulky and costly mannequin with supervisor’s observation, which complicates continuous training. To overcome the challenge, this project aims at developing wearable gloves embedded with sensors for self-training. To this end, this study presents identification of features that can discriminate between novices and experienced providers by analyzing data collected from accelerometers. Based on the selected features, classification models are developed and tested to evaluate the potential of this approach in developing efficient devices for ETI training.