Association of daily lung condition in COPD patients with wearable speech and physiological data

Abstract

Abstract Chronic obstructive pulmonary disease (COPD) is a prevalent condition that imposes significant burden on patients and healthcare systems, with COPD exacerbations being a leading cause of hospitalizations and mortality worldwide. Continuous monitoring of lung function is crucial for effective management, yet traditional methods are often cumbersome and require clinic visits. Wearable technology offers a non-invasive alternative, enabling the monitoring of physiological parameters in real-world settings. In this study, we analyzed free-living speech data collected via smartwatches from 18 COPD patients over an average period of 198.9 ± 122.6 days. Utilizing linear mixed-effects models, we discovered that phonation features are negatively associated with adverse COPD outcomes, while prosodic speech features show a positive correlation with increased exacerbation risk. Further analysis revealed that these associations are significantly moderated by physiological covariates such as heart rate variability and physical activity levels. These findings highlight the complex interplay between respiratory function, autonomic regulation, and vocal production, suggesting that the integration of speech analysis with physiological monitoring through wearables can lead to the development of composite digital biomarkers of impaired lung function.