Category: Technology
Objective: To identify real-world mobility measures that are most reflective of PD severity and to assess their sensitivity to changes over time, as compared to the MDS-UPDRS.
Background: Real-world monitoring using wearable sensors has enormous potential for understanding and monitoring patient signs and symptoms. Many distinct features can be extracted that reflect multiple mobility domains (e.g., general physical activity, gait, nocturnal movement). However, it is not yet clear which features derived from wearables are related to PD severity, how they change over time, or how they are related to currently used clinical scales.
Method: Real-world data – collected continuously over one week – were collected using a wearable data logger (Axivity Ltd.) that was adhered to the lower back of 575 patients with PD (mean age: 68.1±8.4 yrs, 60% males, disease duration 6.3±5.8 yrs) and 69 healthy controls (58.4±8.7 years, 59.4% males). Machine-learning feature selection and regression algorithms were applied to discriminate between controls and early PD (<2 yrs from diagnosis), evaluate associations with disease severity, and explore sensitivity to change over one year.
Results: Recently diagnosed patients with PD were accurately (83.6%) discriminated from healthy controls using a combination of 13 extracted digital measures. The output of a regression model based on the digital measures was moderately correlated with MDS-UPDRS parts I (r=0.51) and II (r=0.60), primarily because of features reflecting nocturnal behavior and gait quality, and MDS-UPDRS-III (r=0.49), primarily related to activity quantity and its distribution across the day. The digital measures showed greater effect sizes (ES) (e.g., 0.37 activity distribution vs. 0.04 MDS-UPDRS part III) for change over 1 year; these effects became larger with disease progression (e.g., ES: gait speed: 0.64 vs. MDS-UPDRS part III: 0.32).
Conclusion: Real-world mobility measures show high discriminant abilities in early stage PD and moderate associations with clinical assessment, suggesting that the digital measures capture different aspects of capacity and function. Our findings also show that the wearable measures are more sensitive to change over one year than clinical assessment, demonstrating their utility in monitoring disease and disease progression.
To cite this abstract in AMA style:
A. Mirelman, J. Volkov, A. Solomon, E. Gazit, A. Nieuwboer, S. Del Din, L. Rochester, L. Avanzino, E. Pelosin, B. Bloem, U. Della Croce, A. Cereatti, A. Thaler, J. Shirvan, J. Cedarbaum, N. Giladi, J. Hausdorff. Association of real-world wearable mobility measures with Parkinson’s disease severity and disease progression [abstract]. Mov Disord. 2023; 38 (suppl 1). https://www.mdsabstracts.org/abstract/association-of-real-world-wearable-mobility-measures-with-parkinsons-disease-severity-and-disease-progression/. Accessed November 10, 2024.« Back to 2023 International Congress
MDS Abstracts - https://www.mdsabstracts.org/abstract/association-of-real-world-wearable-mobility-measures-with-parkinsons-disease-severity-and-disease-progression/