Alert system design based on experimental findings from long-term unobtrusive monitoring in COPD

Abstract An observational study using a cough count monitor was executed in the home of 30 patients with Chronic Obstructive Pulmonary Disease (COPD). The monitoring system was unobtrusive, allowing long-term monitoring. This paper illustrates the cough counts obtained from this monitoring using a personalized cough classifier. In particular, the data highlights a number of patient-specific parameters measured in baseline periods such as the average and mean cough count, and the pattern of the cough count distribution over the course of the day. Next to describing various specific patterns disclosed by the clinical study, the paper outlines the basic components in the design of an alert system compatible with the observed experimental findings. In particular, a novel cough count scale is introduced to facilitate the construction of an alert mechanism to detect worsening in respiratory condition.