Development of image processing-based smartphone App (Application Software) for self-management of inhaled medicines for patients with asthma or COPD

Background: Inhaled medications are the cornerstone of the treatment of asthma and COPD. Effective use of inhalers requires patients to follow the prescribed inhalation technique. Errors in device use can result in suboptimal drug delivery, reducing the effective medication dose and thus compromising treatment effectiveness. Many common errors are universal, such as failing to exhale before each inhalation and failing to hold the breath following inhalation. Improving the inhalation skills of asthmatics and COPD patients requires carefully trained inhalation. However, it is not really possible for a doctor to make sure that the patient is successfully inhaling medicine at home. Aim: To circumvent inhaler mishandling, we attempted to develop a smartphone App that can judge one9s own inhalation handling and learn inhalation maneuver. Method: This App was designed with a focus on image processing and deep learning techniques for analysing the movement of assisting respiratory muscles.Results: The major characteristics of this App are as follows: 1. The type of inhalation device is automatically classified. 2. The quality of inhalation skills are automatically judged. 3. Patients can learn the inhalation operation with streaming movies. 4. Doctors can track the usage status of patient9s inhaler. Results: With use of this system, we demonstrated that the type of inhaler can be classfied 100% correctly and correct inhalation can be judged with an accuracy of more than 75%. Conclusion: This smartphone App enables patients to learn inhalation skills through feedback by oneself and could provide internet-based remote medicine for doctors.