Wearable Respirometer for Pulmonary Disease Prediction

Breathing is a multifaceted indicator of health and performance, with pivotal parameters including tidal volume, inspiratory capacity, and forced expiratory volume crucial for healthcare assessment and disease detection. However, real-time, accurate, and continuous monitoring of these parameters is hindered by the limitations of cumbersome instruments. In response, this research introduces an innovative approach to estimate these critical parameters and predict associated diseases. Our method leverages descriptors extracted from breathing signals captured non-invasively via a wearable organic piezoelectric electronic patch, eliminating the need for invasive procedures. The wearable respirometer boasts impressive features, such as high sensitivity, with an exceptional signal-to-noise ratio, and a low limit of detection.<br/>To achieve precise and robust disease predictions, we conducted an extensive examination of various machine learning algorithms. Among these, gradient boosting regression emerged as the most suitable choice for predicting Chronic Obstructive Pulmonary Disease (COPD). Notably, our method achieved an accuracy of over 94% in predicting different COPD diseases. This research opens new avenues for continuous, real-time monitoring and early disease detection, addressing a critical need in healthcare and paving the way for improved patient outcomes.