BREATHWISE - An AI-Driven Early Warning System for COPD Exacerbations - PROJECT SUMMARY Chronic Obstructive Pulmonary Disease (COPD) is a major public health challenge, affecting over 16 million adults in the United States and contributing to more than 150,000 deaths annually. Exacerbations, which are defined as the acute worsening of respiratory symptoms requiring additional therapy, are the leading cause of hospitalization for COPD patients and account for a large portion of the disease’s estimated $50 billion annual economic burden. These events accelerate lung function decline, diminish quality of life, and increase the risk of premature death. Importantly, research shows that many exacerbations are preceded by subtle but detectable changes in physiological and behavioral signals, such as oxygen desaturation variability, elevated cough frequency, disturbed sleep, and decreased activity levels. However, current monitoring approaches, which typically rely on self-reports, adherence tracking, or intermittent clinical visits, remain narrow in scope and often miss the opportunity for timely intervention. Lynntech, Inc. proposes to develop BREATHWISE, an AI-driven early-warning system for COPD exacerbations that integrates multimodal physiological, behavioral, and environmental sensing with machine learning-based predictive analytics. BREATHWISE will continuously collect data from wearable devices and smartphones, apply advanced time-series feature extraction methods, and generate personalized risk predictions using interpretable machine learning models. By transforming raw data into clinically relevant decision-support alerts, the system will empower patients and clinicians to intervene earlier, reducing costly hospital readmissions and improving long-term outcomes. The goals of this Phase I SBIR project are to develop and demonstrate the feasibility of BREATHWISE’s core components. Our specific aims are to: (1) design and validate a passive sensing framework for monitoring COPD-related physiological and behavioral signals, (2) develop a personalized AI model capable of predicting exacerbation risk up to 72 hours in advance, and (3) implement an interpretable decision-support module that delivers actionable early-warning alerts to patients and clinicians. Successful completion of these aims will provide foundational evidence for BREATHWISE’s feasibility and clinical utility. This work will establish a scalable framework for personalized COPD management, laying the groundwork for Phase II development and future clinical trials. Ultimately, BREATHWISE has the potential to transform exacerbation prevention, reduce healthcare costs, and improve quality of life for millions of individuals living with COPD.
