Wearable inertial sensors tracking 3,424 older adults reveal a dramatic 0.41 m/s difference between clinical corridor gait speed and real-world home walking—a 42% reduction. Crucially, this "ecological gap" is not evenly split: step length accounts for 67.3% of the deficit while cadence contributes only 33.7%, meaning people shorten their strides roughly twice as much as they slow their stepping rhythm. Double-support time (both feet on ground simultaneously) nearly doubled from 0.18 to 0.43 seconds, while single-limb swing time remained preserved—suggesting a deliberate stability-seeking strategy rather than simple fatigue.
This finding reframes how clinicians should interpret gait speed assessments. The well-established 0.8 m/s slow-gait threshold—linked in prior research to fall risk, sarcopenia, and mortality—misclassifies 88.6% of home walkers as impaired when applied directly to home data. The authors propose an equipercentile-equated home cutoff of approximately 0.5 m/s, which has significant implications for fall-risk screening and rehabilitation goal-setting. Step-length reduction specifically aligns with known markers of cautious gait and fear of falling, distinct from the cadence changes more typical of neurological decline. The large, well-characterized cohort (n=3,424) strengthens statistical confidence, though the cross-sectional, observational design limits causal inference, and separate clinical and home cohorts introduce selection differences. As a preprint not yet peer-reviewed, these translated cutoffs and mechanistic interpretations require independent validation before clinical adoption. Still, this work appears paradigm-shifting for mobility assessment: measurement context is inseparable from the measurement itself.
