Digital Health Technologies Superior to Traditional Approaches for Measuring Disease Progression in Early Parkinson Disease

Newly published in Nature, findings from an analysis on the multicenter, observational WATCH-PD study (NCT03681015) demonstrated that digital composite measures for upper and lower extremity bradykinesia were more sensitive to 1-year longitudinal disease progression for early Parkinson disease (PD) compared with traditional measurement approaches.¹ These results suggest the potential of digital health technologies for improving sensitivity to disease progression and may be used as a basis for treatment development tools in clinical research.

In a comparison of digital features between baseline and 12 months in both healthy and participants with PD, few individual digital features displayed strong sensitivity to longitudinal change. Notably, both pronation-supination (Wilcoxon p = 0.018, V = 429, CI: −2.76 to −0.27, effect size = 0.45, proportion = 0.62) and toe-tapping (Wilcoxon p = 0.011, V = 485, CI: −1.83 to −0.23, effect size = 0.30, proportion = 0.66) composite scores demonstrated significant differences after 12 months.

“From a clinical trial perspective, better measures of progression are needed to optimize trial designs and decrease the duration needed to detect signals of efficacy in proof-of-concept trials of novel therapies,” lead author Matthew D. Czech, PhD, associate director of digital science at AbbVie, and colleagues wrote.¹ “The current results represent a step toward better long-term characterization of disease course and response to treatment in PD patients that may ultimately lead to enhanced therapeutic trials and care.”

READ MORE: AC Immune SA Shares Phase 2 VacSYn Trial Update for ACI-7104.056 in Parkinson Disease

In the current analysis, investigators developed a wearable sensor-based digital algorithm for deriving features of upper and lower-body bradykinesia and assessed the sensitivity of digital measures to longitudinal progression using data from WATCH-PD.² Investigators recruited 82 participants and 50 age-matched controls, who took part in a variety of motor tasks over a 12-month period while wearing body-worn inertial sensors. Authors established clinical validity of sensor-based digital measures by assessing convergent validity with appropriate clinical constructs, known groups validity by differentiating patients with PD from healthy participants, and test-retest reliability by comparing measurements between visits.

When analyzing the connection between digital composites and corresponding MDS-UPDRS items and subscores, the digital pronation-supination composite score revealed a significant relationship with MDS-UPDRS pronation-supination item score (item 3.6; Kruskal–Wallis p = 0.028; Spearman’s rho = 0.31, P = 0.007). The digital pronation-supination composite score significantly distinguished between scores of 0 and 1 (Wilcoxon p = 0.02) did not between scores of 1 and 2, or 2 and 3. Also, the digital toe-tapping composite score showed a significant relationship with MDS-UPDRS toe-tapping item score (item 3.7; Kruskal–Wallis p = 0.001; Spearman’s rho = 0.38, P <.001). The digital toe-tapping composite score significantly distinguished between scores of 0 and 1 (Wilcoxon P <.001) but also not between scores of 1 and 2, or 2 and 3.

In a further analysis on the ability of these measures to differentiate between PD and healthy comparison participants, both pronation-supination (Wilcoxon p < 0.001, W = 757, CI: −4.86–−1.96) and toe-tapping (Wilcoxon p = 0.029, W = 1143, CI: −2.23–−0.13) digital composites showed significant differences between patients with PD and healthy participants. Additionally, both MDS-UPDRS pronation-supination (Wilcoxon p < 0.001, W = 278, CI: −1.00 to−1.00) and toe-tapping (Wilcoxon p < 0.001, W = 510, CI: −1.00 to −1.00) subscores significantly distinguished between patients with PD and healthy participants.

“Our results also suggest that composite scoring may be an effective method to enhance sensitivity to the progression of bradykinesia or other factors, likely because of heterogeneity in the way bradykinesia may manifest across individuals during structured tasks. The demonstrated composite approach attempts to better align sensor-based measures with how bradykinesia is clinically assessed, namely by examining multiple features, including amplitude, speed, and decrement over time of repetitive movements,” Czech et al noted.¹ “Thus, our results provide evidence that composite scoring may enhance sensitivity to progression by combining various individual features that do not individually change significantly over time but likely trend in a common direction.”

In the test-retest analysis assessing reliability of digital features, digital composite scores for both pronation-supination (Pearson R = 0.42, p = 0.016; ICC = 0.40) and toe-tapping (Pearson R = 0.51, p = 0.004; ICC = 0.5) showed moderate test-retest reliability between baseline and 1-month visits for healthy participants. In participants with PD, digital composite scores for pronation-supination (Pearson R = 0.54, p < 0.001; ICC = 0.52) and toe-tapping (Pearson R = 0.32, p = 0.013; ICC = 0.31) displayed moderate and poor test-retest reliability, respectively, between baseline and 1-month visits.

The digital measures used in the study were limited by the moderate to poor test-retest reliability observed with the digital bradykinesia composite scores between baseline and 1-month visits. In another limitation, researchers considered the possibility that participants might have felt more comfortable with the assessments toward the end of the study, which may reduce the variability between assessments. Since the recordings were made contemporaneously with clinician ratings, the number of movements per assessment used for the study varied significantly. Additionally, the digital composites had the inability to distinguish between MDS-UPDRS scores of 1, 2, and 3.

“Digital health technologies have great potential for improving the clinical assessment of motor features in early PD, however there is a lack of data demonstrating that measures derived from DHTs are more sensitive to decline in function over time,” Czech et al noted.¹ “Future research should extend our findings by evaluating the sensitivity of digital measures to disease progression and treatment effects relative to current clinical assessments across indications and disease stages.”

REFERENCES
1. Czech MD, Badley D, Yang L, et al. Improved measurement of disease progression in people living with early Parkinson's disease using digital health technologies. Commun Med (Lond). 2024;4(1):49. Published 2024 Mar 15. doi:10.1038/s43856-024-00481-3
2, Adams JL, Kangarloo T, Tracey B, et al. Using a smartwatch and smartphone to assess early Parkinson's disease in the WATCH-PD study. NPJ Parkinsons Dis. 2023;9(1):64. Published 2023 Apr 17. doi:10.1038/s41531-023-00497-x