Measurement properties of movement smoothness metrics for upper limb reaching movements in people with moderate to severe subacute stroke.

Only 31 participants were analysed, and six of the 37 motion-capture participants were excluded for incomplete or uninterpretable data. The sample had moderate to severe impairment in the subacute phase, so results may not apply to milder stroke, chronic stroke or other movement problems. Participants were pooled from two parent-trial rehabilitation groups, so the study cannot say which rehabilitation component caused the changes. Only single-variable correlations were used because data were non-normal and the sample was small, so confounding factors may bias relationships. No gold standard exists for smoothness, so responsiveness was judged against related measures rather than a definitive reference. Filtering was chosen from prior work, not optimised for each metric, and another filter could have changed the time-based metric performance. nSUB and NARJ had high within-person variability at baseline, and only three recorded movements were used, so more trials might change results. Motion capture requires instrumentation and expertise, so these metrics are not routine bedside tests.

Funded by the French ministry of health. The supplied text does not report other author conflicts.

Do not conclude that SPARC is a validated clinical test for every stroke patient or that it proves a treatment works. The study tested measurement properties in 31 people with moderate to severe subacute stroke using motion capture, and it pooled participants from two rehabilitation groups so it cannot show which rehabilitation caused change.