Title:Can a wearable device detect airway obstruction?

Abstract:Objective: Lung health monitoring may enable early detection and treatment of exacerbations in respiratory diseases such as chronic obstructive pulmonary disease (COPD). Our objective is to explore the feasibility of using a wearable device to continuously monitor lung health. Towards that goal, this work studies the relationship between wearable device measurements of tidal breathing and spirometer lung function parameters. Methods: Data was collected from 25 single visit adult volunteers with a confirmed or suspected diagnosis of COPD. A respiratory chest belt was used to measure the fractional inspiratory time, respiratory rate and tidal volume of subjects during quiet breathing. The subjects also performed standard spirometry in a pulmonary function testing laboratory under the supervision of trained clinical staff to produce forced expiratory volume in one second (FEV1), forced vital capacity (FVC) and the ratio FEV1/FVC. Two classification models were built and trained: one to detect the presence of lung airway obstruction, and another to stratify the severity of airway obstruction. Results: The classifier detected airway obstruction with a sensitivity of 95% and a specificity of 80%. Severity of airway obstruction was classified as either mild/moderate or severe/very severe with a sensitivity of 91% and a specificity of 78%. Conclusion: Tidal breathing parameters that are measured with a wearable device can be used to detect airway obstruction with a level of accuracy that is comparable to that of conventional screening tests used in primary care settings. In addition, these parameters can reliably distinguish between severe and mild airway obstruction.