Continuous Cough Monitoring Using Ambient Sound Recording During Convalescence from a COPD Exacerbation

Michael G. Crooks, Albertus den Brinker, Yvette Hayman, James D. Williamson, Andrew Innes, Caroline E. Wright, Peter Hill, Alyn H. Morice

Cough is common in chronic obstructive pulmonary disease (COPD) and is associated with frequent exacerbations and increased mortality. Cough increases during acute exacerbations (AE-COPD), representing a possible metric of clinical deterioration. Conventional cough monitors accurately report cough counts over short time periods. We describe a novel monitoring system which we used to record cough continuously for up to 45 days during AE-COPD convalescence. This is a longitudinal, observational study of cough monitoring in AE-COPD patients discharged from a single teaching hospital. Ambient sound was recorded from two sites in the domestic environment and analysed using novel cough classifier software. For comparison, the validated hybrid HACC/LCM cough monitoring system was used on days 1, 5, 20 and 45. Patients were asked to record symptoms daily using diaries. Cough monitoring data were available for 16 subjects with a total of 568 monitored days. Daily cough count fell significantly from mean ± SEM 272.7 ± 54.5 on day 1 to 110.9 ± 26.3 on day 9 (p < 0.01) before plateauing. The absolute cough count detected by the continuous monitoring system was significantly lower than detected by the hybrid HACC/LCM system but normalised counts strongly correlated (r = 0.88, p < 0.01) demonstrating an ability to detect trends. Objective cough count and subjective cough scores modestly correlated (r = 0.46). Cough frequency declines significantly following AE-COPD and the reducing trend can be detected using continuous ambient sound recording and novel cough classifier software. Objective measurement of cough frequency has the potential to enhance our ability to monitor the clinical state in patients with COPD.