Longitudinal Cough Frequency Monitoring in Subjects with Persistent Cough: Daily Variability and Predictability Over 30 Days

<bold>Background:</bold> Chronic cough is defined by a cough lasting over 8 weeks. Although the pattern of coughs over a monitoring period of one day or week has been investigated, its pattern over longer periods is not known. <bold>Research Question:</bold> What is the daily variability of coughing over 30 days in subjects with persistent cough? <bold>Study Design and Methods:</bold> The Hyfe Cough Tracker is a smartphone app that uses AI algorithms to passively and continuously recognize and timestamp coughs, while preserving privacy. From January to August 2023, 25,128 individuals downloaded the app to their smartphones. Ninety-seven of these users monitored for over 30 days and had median cough counts of at least 5 coughs/hour. The statistical properties of coughing in these subjects were analyzed. <bold>Results:</bold> The daily median cough rates per hour varied from 6.2 to 56.0 coughs per hour over 24 hours (interquartile range 1.30 - 41.5 coughs/hour). There was a correlation between mean daily cough frequency and the standard deviation. The predictability of cough rates for each day was determined by analyzing the percent of daily cough counts that fell within the 95% confidence intervals. In ⅔ of subjects, the variability in their daily cough rates was so great that 24h measurements did not accurately reflect longer term trends. <bold>Interpretation:</bold> Cough rates among subjects with persistent cough recorded continuously over 30 days are stochastic. The day to day variability in cough frequency is such that one day of monitoring is often insufficient to describe a person’s cough accurately. Continuous monitoring for longer periods of time is required to understand trends and changes with interventions.