Reliability of upper limb kinematic variables: adults with unilateral cerebral palsy performing a standardized drinking task.

Only 25 adults were tested, and the authors note that there is no established consensus on the required number of participants for these reliability estimates. Most participants had high function, with 60% at GMFCS I and 40% at MACS I, so the results may not apply to more severely affected adults. The reliability estimates apply only when the same marker placement, cup setup, seating, and testing conditions are used. The model did not measure finger, thumb, forearm rotation, wrist deviation, or head and trunk rotation, which can affect grasp in cerebral palsy. Participants did not use their regular hand or finger orthosis, which could change how they grasp the cup. Only the most affected limb was tested, not the least affected limb. Optical motion capture requires a laboratory and technical expertise, so it is not intended for routine clinic use. Three participants had outlier values, and outliers can raise ICC values, although all variables remained above 0.90 after outlier removal.

Do not read the high reliability as proof that the drinking task improves function. This study measured whether motion-capture variables repeated across two days; it did not test a treatment effect.