Clinic gait measures do not map cleanly onto home walking after stroke
A common assumption in rehabilitation is that a patient’s gait in a clinic corridor reflects how they walk in the real world. This study put that assumption under direct pressure. Using smartphone or tablet video recorded in both environments, investigators compared gait kinematics in adults with and without stroke and found that the relationship between clinic and home performance was uneven, with the strongest alignment seen for walking speed and weaker agreement for asymmetry measures.
The work matters because clinicians often make mobility judgments from brief observations under controlled conditions. Those observations may capture capacity, but not always everyday behavior. After stroke, a person can appear stable in a supervised hallway and then move differently in a familiar home environment where turning, spacing, attention, and pace all shift. The study tested that gap using video-based pose estimation and MATLAB analysis rather than relying on subjective observation alone.
Video analysis captured step asymmetry and speed across two environments
Data from 43 participants were analyzed, including 18 adults with a history of stroke and 25 without stroke. Each participant walked in both a clinic setting and at home, and those recordings were processed to extract three gait variables: step length asymmetry (SLA), step time asymmetry (STA), and gait speed (GS). The question was straightforward: do measurements obtained in the clinic resemble the same person’s gait in daily life?
The answer was mixed. Spearman correlation analyses showed weak-to-strong associations between environments, with the most clinically meaningful relationships appearing in the stroke group. For step length asymmetry, the correlation was ρ = 0.455 with p = 0.058. For step time asymmetry, the correlation was ρ = 0.519 with p = 0.027. For gait speed, the correlation reached ρ = 0.752 with p = 0.002. Those values suggest that the same individual’s walking speed translated relatively well from clinic to home, while the asymmetry measures were more variable.
Walking speed tracked across settings better than limb symmetry did.
Stroke-related gait impairments may be more stable than expected, but not fully interchangeable
Among participants with stroke, the stronger associations imply that some aspects of impaired walking are preserved across environments, especially when the gait metric is gross and behaviorally robust. Gait speed may be less sensitive to environmental nuance because it reflects a broad output: how quickly a person can move from one point to another under a given set of constraints. In contrast, asymmetry measures depend on limb coordination and temporal-spatial control, which can shift with attention, surface, path length, and comfort. A person may walk briskly in a clinic because the route is straight and supervised, then alter step timing at home where obstacles and household tasks interrupt rhythm.
For rehabilitation, that distinction is not academic. A clinic-based gait evaluation can still be useful, but it should be read as an estimate of capacity rather than a complete portrait of lived mobility. For patients after stroke, this becomes especially important when clinicians are deciding whether a change in therapy has truly improved everyday walking or merely improved performance in a controlled setting.
The data also challenge the idea that one environment can substitute for the other without loss. Home walking contains the variables that clinics suppress: distraction, fatigue accumulated over the day, self-selected pacing, and the need to negotiate real household space. Those pressures can expose deficits that a short hallway walk leaves hidden. At the same time, the moderate-to-strong relationship for gait speed suggests that not every clinical metric is equally fragile. Some measures appear portable across contexts; others are highly context dependent.
Why the association for asymmetry may matter more than the raw speed value
Asymmetry metrics are often treated as secondary details, but they can be central to understanding post-stroke gait quality. Step length asymmetry and step time asymmetry reflect how evenly the body distributes work between limbs, and chronic asymmetry can be linked to inefficient movement and compensatory strategies. Even when speed improves, a person may still rely on the same asymmetrical pattern. That matters for balance, joint loading, and long-term mobility because faster walking is not necessarily more symmetrical walking.
In practical terms, the study suggests that clinic testing may be better at identifying a person’s maximum or near-maximum walking speed than at reproducing the subtler mechanics of day-to-day ambulation. Home-based assessment, even when captured through something as accessible as a smartphone camera, may reveal the motor pattern a patient actually lives with. That opens a realistic path for tele-rehabilitation, remote monitoring, and more ecologically valid outcome assessment without requiring sophisticated gait labs.
A small cohort limits certainty, but the measurement approach is clinically relevant
The sample was modest, with 43 total participants and 18 in the stroke group, so the estimates should be interpreted cautiously. Correlation values can also be sensitive to sample size and to the range of walking ability represented in a cohort. Still, the study provides a concrete method for comparing gait across settings using common digital tools rather than expensive instrumentation. That is especially relevant in stroke care, where access to specialized gait laboratories is limited and where home function often matters more than isolated clinic performance.
Because the analysis used video recordings processed with pose estimation, the study also points toward a future in which gait can be evaluated outside the laboratory without sacrificing too much analytic rigor. The larger clinical question is not whether the clinic should disappear; it is which gait metrics are trustworthy across contexts and which need direct home measurement to be interpreted correctly. Here, speed looked more transferable than asymmetry, and that pattern should shape how clinicians think about routine gait assessment after stroke.
The central takeaway is narrow but important: clinic gait evaluation is informative, yet it does not fully replace home-based observation, especially for asymmetry. In stroke rehabilitation, the environment is part of the measurement.