Longitudinal wearable-sensor data identified multiple walking, non-walking, and composite digital measures that sensitively track progression in early- to mid-stage Parkinson's disease—some outperforming conventional clinical scales and detecting change within ~10 months.

These objective, remote digital biomarkers could increase sensitivity and temporal resolution in PD trials and monitoring, enabling smaller/shorter studies and earlier detection of treatment effects to accelerate therapeutic development.

Digital measures of walking and sedentariness may objectively quantify Parkinson's disease (PD) progression. We analyzed longitudinal wearable-sensor data to evaluate the sensitivity and specificity of digital walking and non-walking measures in ambulatory people with PD. We selected 26 individual and 6 composite measures with sufficient sensitivity and test-retest reliability in a development dataset (N = 171). Twenty measures showed significant within-participant changes, and 7 showed larger 2-year effect-size than gold-standard clinical measures in people with early-stage PD (N = 101, mean number of years since diagnosis [YSD], 2.2). Significant changes in non-walking and composite measures were detectable as early as 10 months. Twelve measures showed greater change in people with more advanced PD (N = 67; mean YSD 6.5) compared to matched non-PD individuals (N = 171). Sensitivity and specificity results indicate that measures capturing walking and especially non-walking behaviors hold promise as PD progression markers in early to mid-stage PD.