Assessment of physical activity patterns in patients with rheumatoid arthritis using the UK Biobank.

• Published in: PLoS ONE
• Main Authors: Valentin Hamy, Andrew Creagh, Luis Garcia-Gancedo
• Format: Article
• Language: English
• Published: Public Library of Science (PLoS) 2025-01-01
• Online Access: https://doi.org/10.1371/journal.pone.0319908

Measures of physical activity patterns that may characterize rheumatoid arthritis status were investigated, using actigraphy data from a large, prospective database study (UK Biobank). Population characterization identified 1080 individuals with rheumatoid arthritis who participated in accelerometer-measured physical activity data collection and met the eligibility criteria; these individuals were subsequently matched with 2160 non-rheumatoid arthritis controls. Raw actigraphy data were pre-processed to interpretable acceleration magnitude and general signal-based features were used to derive activity labels from a human activity recognition model. Qualitative assessment of average activity profiles indicated small differences between groups for activity in the first 5 hours of the day, engagement in moderate-to-vigorous activity, and evening sleep patterns. Of 145 metrics capturing different aspects of physical activity, 57 showed an ability to differentiate between participants with rheumatoid arthritis and non-rheumatoid arthritis controls, most notably activities related to moderate-to-vigorous activity, sleep and the ability to perform sustained activity, which remained different when adjusting for baseline imbalances. Objective measures derived from wrist-worn accelerometer data may be used to assess and quantify the impact of rheumatoid arthritis on daily activity and may reflect rheumatoid arthritis symptoms. This work represents an initial step towards the characterization of such impact. Importantly, this study offers a glimpse of the potential use of large-scale datasets to support the analysis of smaller clinical study datasets.