Monitoring duration is important to the correct and accurate measurement of the phenomonon of cough, especially given its highly variable nature. Accordingly, Hyfe has carried out research looking into the optimal monitoring period to capture one’s true, underlying cough rate.
This document presents a 30-day observational study evaluating continuous, passive cough monitoring in individuals with persistent cough using Hyfe’s smartphone-based and wearable cough-detection technology. The objective was to characterize day-to-day variability in cough frequency, assess the limitations of traditional 24-hour monitoring windows, and evaluate the predictability of cough patterns over extended periods.
The study followed 57 users for 30 days and collected more than 20 daily monitoring hours per participant using two Hyfe applications (CoughPro and CoughTracker). Coughs were identified using automated acoustic detection and classification, with all audio processed in a privacy-preserving manner. Outcome measures included daily cough frequencies, daily cough variability within and between subjects, and predictive modeling of future cough load based on previous days’ data.
The results show extremely wide variability in cough frequency across individuals—ranging from approximately 6.5 to 182 coughs per hour. Within-subject cough frequency also fluctuated significantly from one day to the next. Despite this variability, predictability analysis showed that for participants with higher cough burdens (>6 coughs per hour), multi-day monitoring enabled reasonably accurate prediction of future cough frequency, particularly when looking at 7-day windows. The heat-map visualization (Figure 3) illustrates how predictability varies at the individual level across the 30-day period.
The study concludes that cough frequency is highly variable both between and within individuals, making single-day measurements insufficient for understanding meaningful trends. For people with persistent cough, at least 7 days of monitoring are needed to achieve reliable estimation of typical cough frequency and to detect clinically meaningful change. This reinforces the importance of longitudinal, passive, continuous cough monitoring for respiratory research and future clinical applications.