An acoustic method of automatically evaluating patient inhaler technique

Chronic respiratory diseases such as asthma and chronic obstructive pulmonary disease (COPD) affect millions of people worldwide. Inhalers are devices utilized to deliver medication in small doses directly to the airways in the treatment of asthma and COPD. Despite the proven effectiveness of inhaler medication in controlling symptoms, many patients suffer from technique errors leading to decreased levels of medication efficacy. This study employs a recording device attached to a commonly used dry powder inhaler (DPI) to obtain the acoustic signals of patients taking their inhaler medication. The audio files provide information on how a patient uses their inhaler over a period of one month. Manually listening to such a large quantity of audio files would be a time consuming and monotonous process and therefore an algorithm that could automatically carry out this task would be of great benefit. An algorithm was thus designed and developed to detect inhalation, exhalation and blister events in the audio signals, analyze the quantity of each event, the order in which the events took place and finally provide a score on the overall performance. The algorithm was tested on a dataset of 185 audio files obtained from five community dwelling asthmatic patients in real world environments. Evaluation of the algorithm on this dataset revealed that it had an accuracy of 92.8% in deciding the correct technique score compared to manual detection methods.