ERS 2024: Differentiating Cough Origins in Non-Contact AI Monitoring: Analyzing 'Non-User Coughs' with the Hyfe CoughMonitor Smartwatch
Summary
This pilot work clarifies a methodology for quantifying this phenomenon and provides preliminary evidence that non-user coughs have a negligible effect on the cough estimates of the Hyfe CoughMonitor Watch in real-world settings.
The Hyfe CoughMonitor Suite allows for continuous, non-intrusive acoustic cough monitoring that can last for months. This allows the interpretation of longitudinal signals providing important insights for clinicians, researchers, drug developers and regulators. There were however concerns that coughs from non-user bystanders may contaminate the acoustic environment and skew the estimated cough rates.
We analysed 223 hours (almost 10 days) of monitoring submitted by non-coughing volunteers across ten different everyday environments. Acoustic contamination by non-participant coughs ranged from 0 coughs per hour in a restaurant/bar outdoors to 1.04 coughs per hour in a dense, indoor space (symposium) and the same rate for periods of grocery shopping.
Scenario 1 - office worker with a family: spends 8 hours sleeping at home with their partner, 8 hours in the office, 1 hour grocery shopping, 1 hour on public transportation, 1 hour in a restaurant, and 5 hours at home with family. They pick up 2.6 coughs from others in 24 hours.
Scenario 2 - business traveler: spends 7 hours sleeping at home with their partner, 5 hours on public transportation, 1 hour in a restaurant indoors, 1 hour in a bar outdoors, 5 hours at a symposium, 2 hours at home with family, and 3 hours alone in a hotel. They pick up 4.61 coughs from others in 24 hours.
This pilot study establishes a methodology for quantifying this phenomenon and offers preliminary evidence that non-user coughs have minimal impact on the cough estimates provided by the Hyfe CoughMonitor Watch in real-world settings.
