Mayo Clinic Study Uses Wearables, Machine Learning to Predict COPD Rehab Participation

The study addresses a key challenge in COPD care, low adherence to 12-week remote rehabilitation programs. 

Researchers at Mayo Clinic have developed a machine learning model that uses wearable data to predict patient participation in remote pulmonary rehabilitation programs for Chronic Obstructive Pulmonary Disease (COPD), according to a study published in Mayo Clinic Proceedings: Digital Health.

The study addresses a key challenge in COPD care, low adherence to 12-week remote rehabilitation programs. Patients with COPD often experience disrupted sleep due to breathing difficulties, leading to reduced energy levels and increased likelihood of dropping out of structured rehabilitation regimens.

To assess patient readiness and likelihood of participation, researchers collected wearable data using wrist-based activity monitors for one week prior to the start of home-based rehabilitation. This data was used to generate a “Composite Sleep Health Score,” designed to capture sleep quality and patterns.

The score was then combined with traditional clinical data and analyzed using a machine learning model. According to the study, this integrated approach significantly improved the model’s ability to predict patient engagement over a three-month rehabilitation period.

“Our goal was to explore how wearable data could improve dropout rates in remote pulmonary rehabilitation programs,” said Dr. Stephanie Zawada, research associate at Mayo Clinic and the study’s first author.

COPD is characterized by inflamed and narrowed airways along with mucus buildup, which can impair breathing and disrupt sleep cycles. These symptoms can directly affect a patient’s ability to adhere to exercise and education components of rehabilitation programs, particularly in remote settings.

The findings suggest that early identification of patients at risk of low participation could enable more targeted interventions, potentially improving adherence rates in remote care models.

Stay tuned for more such updates on Digital Health News