Supplementary Material for: Developing Smartphone-Based Objective Assessments of Physical Function in Rheumatoid Arthritis Patients: The PARADE Study
datasetposted on 05.05.2020, 14:41 by Hamy V., Garcia-Gancedo L., Pollard A., Myatt A., Liu J., Howland A., Beineke P., Quattrocchi E., Williams R., Crouthamel M.
Background: Digital biomarkers that measure physical activity and mobility are of great interest in the assessment of chronic diseases such as rheumatoid arthritis, as it provides insights on patients’ quality of life that can be reliably compared across a whole population. Objective: To investigate the feasibility of analyzing iPhone sensor data collected remotely by means of a mobile software application in order to derive meaningful information on functional ability in rheumatoid arthritis patients. Methods: Two objective, active tasks were made available to the study participants: a wrist joint motion test and a walk test, both performed remotely and without any medical supervision. During these tasks, gyroscope and accelerometer time-series data were captured. Processing schemes were developed using machine learning techniques such as logistic regression as well as explicitly programmed algorithms to assess data quality in both tasks. Motion-specific features including wrist joint range of motion (ROM) in flexion-extension (for the wrist motion test) and gait parameters (for the walk test) were extracted from high quality data and compared with subjective pain and mobility parameters, separately captured via the application. Results: Out of 646 wrist joint motion samples collected, 289 (45%) were high quality. Data collected for the walk test included 2,583 samples (through 867 executions of the test) from which 651 (25%) were high quality. Further analysis of high-quality data highlighted links between reduced mobility and increased symptom severity. ANOVA testing showed statistically significant differences in wrist joint ROM between groups with light-moderate (220 participants) versus severe (36 participants) wrist pain (p < 0.001) as well as in average step times between groups with slight versus moderate problems walking about (p < 0.03). Conclusion: These findings demonstrate the potential to capture and quantify meaningful objective clinical information remotely using iPhone sensors and represent an early step towards the development of patient-centric digital endpoints for clinical trials in rheumatoid arthritis.