Key points are not available for this paper at this time.
Introduction Development of novel treatments for cough and its management are hampered by the lack of well validated objective cough frequency methodologies. Previous validations have been performed over limited time periods or in laboratory conditions not always representative of typical usage. We describe the rigorous validation of a semi-automated 24 hour ambulatory cough monitoring system (Vitalojak; Vitalograph; Buckinghamshire, England) operating in a manner which completely replicates routine practise. Methods In total 10 (4 female) patients (mean age 60.4 years SD ±14.1 including 6 chronic cough, 2 asthma, 1 COPD and 1 healthy control underwent full 24 Hour ambulatory monitoring (VitaloJAK™). These recordings were manually counted by trained cough counters who also recorded the time at which each cough occurred. These 24 hour recordings were then compressed using custom designed compression software and the sensitivity to cough and the reduced file times were determined. Importantly in each case we confirmed that cough sounds identified in the compressed files were the same sounds identified by the trained manual cough counters in the full 24 hour recording. We tested the software algorithm using three distinct compression levels (1, 2 and 3). Results All results are presented as median (IQR). Sensitivities to cough (%) for compression levels 1, 2 and 3 are 100(100, 100), 100(99.53, 100) and 99.92(99.33, 100) and for reduced file times (minutes) 65.89 (62.40, 83.07), 43.21 (35.94, 57.23) and 26.30 (25.07,46.81) respectively (Table 1). Conclusions The vitaloJAK™ is a reliable, robust and efficient tool for the objective measurement of cough frequency. Importantly it reduces 24 hour recordings by up to 98% whilst preserving close to 100% of recorded cough sounds. This development facilitates efficient and speedy manual cough counting and the level of compression achieved represents significant progress towards fully automated cough monitoring.
McGuinness et al. (Mon,) studied this question.