Study shows that digital devices can provide objective and real-world measures of Parkinson disease, capturing key motor and nonmotor features of for early diagnosis of the neurodegenerative disorder.
Results from the 12-month, multicenter, observational study, WATCH-PD (NCT03681015), recently published in Nature, revealed significant associations between digital biomarkers with a wearable device (Clinical ink) and conventional clinical scoring methods used in Parkinson disease (PD).1 The findings provided evidence on different digital measures derived from smart devices for the detection of motor and nonmotor features in patients with early PD.
In the gait analysis, smartphone and smartwatch data from participants with PD (n = 72) and controls (n =41) showed that patients with PD had smaller arm swing magnitude and exhibited various gait abnormalities detected by the smartphone (27.8 [17.0] degrees vs 49.2 [21.8] degrees; P <.001). Notably, gait parameters measured by the smartwatch and smartphone showed as strongly correlating with those obtained from research-grade wearable sensors (0.36 < r <.79).
"Digital health technologies have great potential to improve our understanding of Parkinson's disease," lead author Jamie L. Adams, MD, associate professor, Department of Neurology, University of Rochester Medical Center, and colleagues wrote.1 "Our findings suggest that these technologies can provide objective measures of PD symptoms that are not captured by traditional rating scales. However, further research is needed to address some of the limitations associated with these devices."
Investigators explored whether a smartwatch and smartphone application could measure features in 82 patients with early, untreated PD and 50 age-matched controls. The participants wore research-grade sensors, a smartwatch, and a smartphone while performing standardized assessments in the clinic. The wearable and sensory devices were loaded with a movement disorders application, consisting of digital assessments on cognition, speech, and motor performance (arm swing, the proportion of time with tremor, and finger tapping). For 7 days, participants wore the smartwatch after each clinic visit at home and completed motor, speech, and cognitive tasks on the smartphone every other week.
“Parkinson’s disease is a neurodegenerative disease with increasing prevalence. However, diagnosis of early Parkinson’s disease remains difficult due to complexity of symptoms,” David Anderson, PhD, principal scientist, Clinical ink, said in a statement.2 “This publication demonstrates the potential use for consumer wearables in the detection and staging of early-stage Parkinson’s disease.”
In the psychomotor function analysis , PD participants (n = 78) exhibited slower finger-tapping speeds (dominant, 104.5 [SD, 40.5] taps per 30 s vs 130.2 [SD, 40.9] taps per 30 s; P <.001; nondominant 106.4 [SD, 39.9] taps vs 122.2 [SD, 34.6] taps; P = .02) and longer inter-tap intervals (dominant, 169.9 [SD, 68.2] ms vs 137.3 (SD, 38.4) ms; P = .008; nondominant hand, 173.0 [SD, 67.7] ms vs 141.9 [SD, 32.1] ms; P = .02) compared with controls (n = 45). In a fine motor test, individuals with PD completed fewer tasks compared with controls in both their dominant (3.4 [SD, 1.7] vs 4.6 [SD, 1.9]; P <.001) and nondominant hand (3.4 [SD, 1.7] vs 4.0 [SD, 1.9]; P <.05).
The monitoring period showed that participants with PD (n = 44) had a significantly higher proportion of time with tremors compared with controls (n = 22) (15.9% vs. 0.6%; P <.001). Additionally, among participants with PD, the measured tremor fraction correlated moderately with self-reported tremor severity (MDS-Unified Parkinson Disease Rating Scale [UPDRS] part II, item 10, r = 0.43, P = 0.003) and strongly with clinician-reported upper extremity rest tremor amplitude MDS-UPDRS part III, item 17, r = 0.86, P < 0.001) and rest tremor constancy (MDS-UPDRS part III, item 18, r = 0.79, P < 0.001).
In smartphone cognitive test data, participants with PD (n = 82) demonstrated poorer performance on the Trail Making Test Part A (54.5 [SD, 23.8] vs 48.0 [SD, 36.0] seconds; P <.05) and had fewer correct responses on the Symbol Digit Modalities Test (18.3 [SD, 8.2] vs 20.4 [SD, 8.9]; P = .05) compared with controls (n = 49). Higher scores on the Montreal Cognitive Assessment weakly correlated with faster completion time on Trails A (r = −0.20, P = 0.14) and Trails B, Trails B (r = −0.38, P <.01), more correct matches on the Symbol Digit Modalities Test (r = 0.25, P <.05), and a higher proportion of correct answers on the Visuospatial Working Memory Test (r = 0.18, P = .16).
Baseline reading task data (participants with PD, n = 79; controls, n = 46) and with phonation task data (participants with PD, n = 53; controls, n = 41) were also analyzed. Participants with PD showed a reduced average pitch range in the reading task compared to controls (4.6 [SD, 1.2] vs 5.6 [SD, 1.2]; P =.00004). Notably, individuals with PD rated as having "normal" speech on the MDS-UPDRS showed a decreased pitch range (4.9 [SD, 1.2] vs 5.6 [SD, 1.2]; P = 0.015).
“We are thrilled to co-author this study in npj Parkinson’s Disease,” added Jonathan Goldman, MD, chief executive officer of Clinical ink in statement.2 “I am delighted that Clinical ink is taking a leadership role in clinical research applications of these novel technologies.I hope that wearables and associated analytic tools can improve the lives of patients living with Parkinson’s disease and other movement disorders.”
The study had several limitations, including missing data, variability in wearing the smartwatch, unfamiliarity with tasks, a homogeneous study population, and concerns about the significance of the measurements. The main limitation was the loss of data because of device permission issues. Also, participants with PD and controls wore the smartwatch on one wrist on different sides. Furthermore, the some of tasks on the app were new to participants, and the study excluded those with cognitive impairments, which may limit the feasibility of digital tasks in more advanced populations.