Development of a Hand Motion-based Assessment System for Endotracheal Intubation Training.

Area of Science:

• Medical Simulation
• Biomedical Engineering
• Artificial Intelligence in Healthcare

Background:

• Endotracheal intubation (ETI) is a critical airway management skill requiring continuous training.
• Current ETI skill assessment relies on human supervisors, leading to potential inconsistencies.
• Developing objective and automated assessment tools is crucial for effective medical training.

Purpose of the Study:

• To develop an automated system for assessing endotracheal intubation proficiency.
• To investigate the feasibility of using only hand motion features for skill classification.
• To differentiate between novice and experienced trainees using objective metrics.

Main Methods:

• Extracted 18 hand motion features (time/frequency domains) and 12 force features.
• Applied feature selection algorithms to identify optimal feature sets.
• Developed artificial neural network (ANN) classifiers using selected features.

Main Results:

• An ANN classifier using five selected hand motion features achieved 91.17% accuracy.
• An ANN classifier using five force features achieved 80.06% accuracy.
• Hand motion features proved more effective for classifying ETI proficiency.

Conclusions:

• A simplified assessment system based on hand motion features is effective for ETI training.
• Automated assessment using hand motion offers a viable and affordable solution.
• This technology can assist in standardizing and improving ETI skill evaluation.