Automatic prediction of chronic obstructive pulmonary disease exacerbations through home telemonitoring of symptoms

Abstract

Chronic Obstructive Pulmonary Disease (COPD) is a progressive disease of the lung with a great prevalence and a remarkable socio-economic impact on patients and health systems. Early detection of exacerbations could diminish the adverse effects on patients' health and cut down costs burdened on patients with COPD. A group of 16 patients were telemonitored at home using a novel electronic daily symptoms questionnaire during a 6-months field trial. Recorded data were used to train and validate a Probabilistic Neural Network (PNN) classifier in order to enable the automatic prediction of exacerbations. The proposed system was able to predict COPD exacerbations early with a margin of 4.8 ± 1.8 days (average ± SD). Detection accuracy was 80.5% (33 out of 41 exacerbations were early detected); 78.8% (26 out of 33) of theses detected events were reported exacerbation and 87.5% (7 out of 8) were unreported episodes. The proposed questionnaire and the designed automatic classifier could support the early detection of COPD exacerbations of benefit to both physicians and patients.