Objective: To validate the sensitivity and accuracy of a video-based motor assessment platform (PRIMS) against a panel of expert MDS-UPDRS reviewers.
Background: The platform is a digital assessment device for remote assessment of Parkinson’s Disease (PD) utilizing a system of 3D cameras to capture motor symptoms, and rates disease severity with the MDS-UPDRS scale using machine learning algorithms. A valid tool for remote assessment of PD could address unmet needs of medically underserved patients, improving diagnosis speed and accuracy.
Method: The device and expert reviewers produced item-level scores on a 0–4 ordinal scale from the same videos across 20 items from MDS-UPDRS Part III (range 0-80). Analyses were restricted to complete video assessments. Missing items were removed for the device and all expert reviewers. The primary reference was determined by taking the median composite score across the three experts. Single‑measure intraclass correlation coefficient (ICC) assessed inter‑rater reliability between three expert reviewers using between‑subject and residual mean squares from a two‑way ANOVA (MS subjects−MS error)/(MS subjects+(k−1)*MS error).
Results: Individuals with physician-diagnosed PD (n=108) and healthy controls (n=16) were assessed by the device. Moderate agreement among the expert raters was calculated (ICC=0.67). The device had a positive strong correlation (r=.74) with the median expert score, whereas the correlation between the expert raters ranged from r=.60–.82). Systematic conservative bias amenable to calibration was observed when comparing the tool (17±5) to the median expert reviewer score (20±7) with a mean difference of -3 and mean percentage error of -8%. Linear regression revealed a significant model (f1,122=149, p<.001, R2=.55) predicting median expert score with y-intercept of 2.27 and βdevice=1.04.
Conclusion: Overall, results indicate the device demonstrated agreement with expert consensus comparable to inter-rater agreement among certified MDS-UPDRS reviewers. An accurate and portable tool capable of screening for PD and/or other conditions and assessing change over time would add substantial value in clinical settings. Further device development for assessments of other conditions may widen the scope and value of such a device.
To cite this abstract in AMA style:
N. Meier, O. Dimartino. Validation of an Objective Parkinson’s Disease Digital Assessment Tool in Individuals with PD and Healthy Controls [abstract]. Mov Disord. 2026; 41 (suppl 1). https://www.mdsabstracts.org/abstract/validation-of-an-objective-parkinsons-disease-digital-assessment-tool-in-individuals-with-pd-and-healthy-controls/. Accessed October 1, 2026.« Back to 2026 International Congress
MDS Abstracts - https://www.mdsabstracts.org/abstract/validation-of-an-objective-parkinsons-disease-digital-assessment-tool-in-individuals-with-pd-and-healthy-controls/