AECOPD is defined as a “sustained worsening of the patient’s condition, from the stable state and beyond normal day-to-day variations, that is acute in onset and necessitates a change in regular medication in a patient with underlying COPD”. AECOPD are usually characterized by increased dyspnea, sputum production, and cough, shallow breathing, increased heart rate and body temperature, and possibly impaired mental state. Respiratory exacerbations are accompanied by abnormal breathing sounds such as wheezing.
In order to determine the best way to treat AECOPD and reduce its physical and financial burden upon patients, it is important to explore a variety of potential solutions. The pilot study explored in today’s blog posts tests a home-based respiratory sensor with computer analysis of respiratory sounds (CARS) and remote monitoring for prediction of AECOPD.
- Acute exacerbations of COPD (AECOPD) represent a major economic burden and are associated with increased mortality and morbidity in COPD patients.
- This pilot study tested a home-based respiratory sensor with computer analysis of respiratory sounds (CARS) and remote monitoring for prediction of AECOPD.
- During 6 months of remote monitoring, the system predicted 25 out of 33 AECOPD (75.8%) about 5 days before the event.
Remote monitoring is able to effectively predict AECOPD events with sufficient time for potential interventions.
COPD patients were equipped with a home based station and respiratory sensor to record their respiratory sounds for six months daily. Recordings were obtained while pressing the device against the suprasternal notch (the large visible dip in between the neck and the collarbone) and breathing normally. Recorded data were processed and filtered, then analyzed automatically by a trained classifier (a computer algorithm). An alarm was raised if the classifier identified positive outputs on two consecutive days.
During the study period, 33 AECOPD events occurred that qualified for analysis (i.e. they did not occur within the recovery phase of a previous AECOPD). The system was able to predict 25 out of these 33 AECOPD (75.8%) with a margin of 5 ± 1.9 days prior to the needed medical attention. Three false alarms were raised during the study period.
Remote monitoring of recorded respiratory sounds of COPD patients with subsequent computer analysis was able to correctly predict 75.8% of AECOPD. This demonstrates that remote monitoring of a non-invasive and user-friendly home-based device with minimum patient effort (once daily application) has the capacity to predict AECOPD with a sufficient margin (fiveÍ days before) for preventative interventions.