An mHealth System for Monitoring Medication Adherence in Obstructive Respiratory Diseases Using Content Based Audio Classification

Abstract

Asthma and chronic obstructive pulmonary disease are obstructive respiratory diseases that affect negatively the quality of life for patients and their families worldwide. Despite the significance of these diseases, their management has been considered suboptimal around the world, whereas the improper inhaler use has been underlined as one of the main causes. Toward this direction, this paper presents an integrated mHealth system that provides real-time personalized feedback to patients for assessing the proper medication use and educating them and helping them avoid common mistakes. The identification of proper inhaler use is based on conventional and data-driven feature extraction and classification methods employed for the identification of four events (inhaler actuation, inhalation, exhalation, and background noise). The proposed scheme reaches 98% classification accuracy significantly outperforming recent and relevant state-of-the-art approaches. Finally, intuitive feedback interfaces were implemented in the form of a virtual guidance agent integrated with the mobile application, which can help patients follow their action plan and assess their inhaler technique in a more engaging manner. Extensive simulation studies, carried out using 12 subjects, demonstrated the efficiency of the proposed approaches in both indoor and outdoor environments.