To evaluate the validity of in-clinic standardized mobility assessments in people with multiple sclerosis (pwMS) using a smart insole-based system (Insole).

Objectively assessing mobility in pwMS is essential for clinical evaluation and timely rehabilitation. However, conventional instrumented gait analysis demands costly, space-intensive equipment and specialized staff, limiting routine use in MS centres. Supervised, Insole–based methods may provide a low-cost, clinic-ready alternative.

Fifty-two pwMS completed the Timed 25-Foot Walk (T25FW) and the 2-Minute Walk Test (2MWT); forty-six pwMS also performed four jump assessments on force plates (FP; 10-Second Hop Test, Countermovement Jump, Single-leg Countermovement Jump right/left). An independent cohort of 18 healthy participants performed jump tasks. All activities were supervised and were conducted under a standardized clinical workflow aligned with the Dresden Protocol for Multidimensional Walking Assessment, which operationalizes routine, clinic-feasible gait assessments in MS.

For every task, participants wore Insoles that streamed raw plantar pressure, accelerometer, and gyroscope data to a smartphone application (Celestra Health Systems, Canada). Custom algorithms recognized ambulatory activities, computed clinical endpoints (2MWT distance; T25FW time), and derived spatiotemporal features (68 walking- and 74 turning-related metrics). Clinical endpoints were compared against ground truth (wheel odometer for 2MWT, photoelectric timing gate for T25FW). For all jumping tasks, 17 temporal, kinematic, and kinetic parameters were compared against FP-based measures.

In the 2MWT, average absolute distance from the odometer was 4.4 m (2.7%). In the T25FW, average absolute time difference was 0.2 s (4.0%). FP agreement (ICC(3,k)) was ≥0.75 for 15 of 17 metrics across all jumping activities. Walking and turning comparisons will be presented at the conference.

Insole-derived measures were comparable to references and fell within minimum detectable change thresholds, while providing richer spatiotemporal and turning insights. Given the small deviations (<5%), this supervised, insole-based system appears suitable for objective mobility assessments in neurorehabilitation and MS outpatient care.