Early detection of COPD exacerbations based on non-obtrusive data coming from patient-friendly wearable biosensors

Early detection of COPD exacerbations allows early treatment. This is crucial to avoid the accelerated disease progression related to having exacerbations. Previous attempts to pick up signs and symptoms of exacerbations included active self-monitoring, putting burden to patients and resulting in low compliance. Multiple vital signs can now be measured using unobtrusive wearable biosensors, allowing a silent data collection and know to come with a good compliance.

I aim to test whether physical activity, measures of respiratory and cardiac autonomic function or their combination, all measured by 1 device worn at the wrist, can be early silent markers of having an exacerbation. To do so I will 1) validate the wrist-worn multi-sensor smartwatch in COPD; 2) investigate the impact of having exacerbations on the biosensor signals; and 3) Analyse the prognostic ability to detect an exacerbation based on these biosensor signals. For the latter aim machine-learning models from artificial intelligence will be used in collaboration with experts in this matter. The present project will result in an important clinically relevant deliverable by providing a prediction model to detect exacerbations early, based on unobtrusive measures.