Henry Glyde, University of Bristol

 

Henry Glyde, University of Bristol

Topic: Predicting COPD Exacerbations Using Real-World Digital Health Data: Lessons from Wearables and Symptoms

Date: November 11th, 2025 14:00-15:00 GMT
Location: In Person Wills Memorial Building Room 3.23 (University of Bristol only) & Online

Abstract:

Chronic obstructive pulmonary disease (COPD) exacerbations remain a leading cause of hospitalisation and health service burden. Predicting when an individual is about to deteriorate is a major clinical challenge, but advances in digital health and real-world data offer new opportunities.

Using daily data from the myCOPD app, Dr Henry Glyde and colleagues modelled short-term exacerbation risk based on self-reported symptoms, COPD Assessment Test (CAT) scores, and step-count data from connected wrist-worn wearables. Through a rigorous nested cross-validation pipeline, the study compared symptom-based and multimodal models, demonstrating stable predictive performance (ROC-AUC ~0.75) and showing that activity features provide complementary information to symptom trends.

The talk will explore the methodological lessons learned—data quality, baseline modelling, and evaluation design—as well as their clinical implications. This work underpins the development of FORESEE-COPD, an open research platform integrating wearable, symptom, and environmental monitoring to enable personalised, proactive COPD care.

Bio: Dr Henry Glyde is a Research Associate in the Academic Respiratory Unit, University of Bristol, and an NIHR Bristol BRC Postdoctoral Fellow. His research integrates AI, digital health, and respiratory medicine to develop predictive models for COPD using wearable, symptom, and environmental data.