Virtual exam for Parkinson's disease enables frequent and reliable remote measurements of motor function

Abstract

Sensor-based remote monitoring could help better track Parkinson's disease (PD) progression, and measure patients' response to putative disease-modifying therapeutic interventions. To be useful, the remotely-collected measurements should be valid, reliable, and sensitive to change, and people with PD must engage with the technology. We developed a smartwatch-based active assessment that enables unsupervised measurement of motor signs of PD. Participants with early-stage PD (N = 388, 64% men, average age 63) wore a smartwatch for a median of 390 days. Participants performed unsupervised motor tasks both in-clinic (once) and remotely (twice weekly for one year). Dropout rate was 5.4%. Median wear-time was 21.1 h/day, and 59% of per-protocol remote assessments were completed. Analytical validation was established for in-clinic measurements, which showed moderate-to-strong correlations with consensus MDS-UPDRS Part III ratings for rest tremor (⍴ = 0.70), bradykinesia (⍴ = -0.62), and gait (⍴ = -0.46). Test-retest reliability of remote measurements, aggregated monthly, was good-to-excellent (ICC = 0.75-0.96). Remote measurements were sensitive to the known effects of dopaminergic medication (on vs off Cohen's d = 0.19-0.54). Of note, in-clinic assessments often did not reflect the patients' typical status at home. This demonstrates the feasibility of smartwatch-based unsupervised active tests, and establishes the analytical validity of associated digital measurements. Weekly measurements provide a real-life distribution of disease severity, as it fluctuates longitudinally. Sensitivity to medication-induced change and improved reliability imply that these methods could help reduce sample sizes needed to demonstrate a response to therapeutic interventions or disease progression.

Fig. 1. Study design and participant-facing interface.

a Study design for the Personalized Parkinson’s Project. b Study schematic defining Sets 1 and 2 for the PD VME substudy. Set 1 consists of participants for whom consensus ratings of the MDS-UPDRS were available. Set 2 consists of participants who opted into the PD-VME substudy and performed at least one PD-VME. c User interface to prompt medication logging and PD-VME tasks. Seated rest, arm raise and arm twist (20 s duration), up-and-go (60 s duration).

Fig. 2. Rest tremor acceleration.

a Lateral tremor acceleration (log scale), measured during the in-clinic examination, by rest tremor (MDS-UPDRS 3.17) consensus score. Center lines: median, boxes: first and third quartiles, whiskers: 1.5x inter-quartile range. b Illustrative examples of raw lateral acceleration signals for each score on the MDS-UPDRS 3.17. Measurement values, as computed by the PD-VME, are also indicated. c Difference between the remote measurements in on and off states, aggregated over PD-VMEs obtained during the first two months from each participant. Mean and 95% confidence intervals across participants are represented. d Intra-class correlation (ICC) between at-home measurements, for various durations of aggregation. Whiskers represent 95% confidence intervals. The dotted blue line represents the published test-retest ICC of 0.79 for the whole rest tremor UPDRS Part III subcomponent (all four extremities + lip & jaw). e Distribution of PD-VME tremor measurements (off state) obtained during the in-clinic PD-VME (orange dot, representing a single measurement) and remote PD-VMEs (blue bar, representing the 25th to 75th percentile of PD-VMEs within 90 days of the in-clinic PD-VME), sorted on the remote PD-VMEs.

Fig. 3. Upper extremity bradykinesia.

a Arm twist amplitude measured during the in-clinic examination, by pronation-supination (MDS-UPDRS 3.6) consensus scores. Center lines: median, boxes: first and third quartiles, whiskers: 1.5x inter-quartile range. b Illustrative examples of raw gyroscope signals, along the x axis, for each score on the MDS-UPDRS 3.6. Measurement values, as computed by the PD-VME, are also indicated. c Difference between the remote measurements in on and off states, aggregated over PD-VMEs obtained during the first two months from each participant. Mean and 95% confidence intervals across participants are represented. d Intra-class correlation (ICC) between at-home measurements, for various durations of aggregation. Whiskers represent 95% confidence intervals. The dotted blue line represents the published test-retest ICC of 0.89 for the whole bradykinesia subcomponent of the UPDRS Part III. e Distribution of PD-VME arm-twist measurements (off state) obtained during the in-clinic PD-VME (orange dot, representing a single measurement) and remote PD-VMEs (blue bar, representing the 25th to 75th percentile of PD-VMEs within 90 days of the in-clinic PD-VME), sorted on the remote PD-VMEs.

Fig. 4. Gait.

a Arm swing acceleration measured during the in-clinic examination, separated by gait (MDS-UPDRS 3.10) consensus scores. Center lines: median, boxes: first and third quartiles, whiskers: 1.5x inter-quartile range. b Illustrative examples of raw accelerometer signals, for each score on the MDS-UPDRS 3.10. Measurement values, as computed by the PD-VME, are also indicated. c Difference between the remote measurements in on and off states, aggregated over PD-VMEs obtained during the first two months from each participant. Mean and 95% confidence intervals across participants are represented. d Intra-class correlation (ICC) between at-home measurements, for various durations of aggregation. Whiskers represent 95% confidence intervals. e Distribution of PD-VME gait measurements (off state) obtained during the in-clinic PD-VME (orange dot, representing a single measurement) and remote PD-VMEs (blue bar, representing the 25th to 75th percentile of PD-VMEs within 90 days of the in-clinic PD-VME), sorted on the at-home PD-VMEs.