StethAid: A Digital Auscultation Platform for Pediatrics

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Medicine
• Pediatric Cardiology

Background:

• Mastery of auscultation is a challenge for healthcare providers.
• Artificial intelligence (AI)-powered digital tools are emerging to aid auscultation interpretation.
• Existing AI-augmented digital stethoscopes are not specifically designed for pediatric use.

Purpose of the Study:

• To develop a digital auscultation platform specifically for pediatric medicine.
• To create an AI-assisted tool for telehealth and auscultation in pediatrics.
• To build the first and largest pediatric cardiopulmonary datasets for AI model training.

Main Methods:

• Development of StethAid: a platform including a wireless digital stethoscope, mobile apps, patient-provider portals, and deep learning algorithms.
• Validation of the StethAid stethoscope's frequency response against commercial stethoscopes.
• Clinical application testing for Still's murmur identification and wheeze detection using trained deep learning models.

Main Results:

• StethAid stethoscope frequency response is comparable to leading commercial stethoscopes.
• High concordance between expert physician labels and bedside provider labels (79.3% lungs, 98.3% heart).
• Deep learning models demonstrated high sensitivity and specificity for Still's murmur (91.9%/92.6%) and wheeze detection (83.7%/84.4%).

Conclusions:

• A technically and clinically validated pediatric digital AI-enabled auscultation platform (StethAid) has been created.
• The platform has the potential to enhance the efficacy and efficiency of pediatric clinical care.
• Potential benefits include reduced parental anxiety and cost savings in pediatric healthcare.