Retest reliability of the temporal and distance characteristics of hemiplegic gait using a footswitch system

Using a footswitch system, the retest reliability of the temporal and distance parameters of gait was investigated within a session for 22 stroke patients in the early phase of rehabilitation. High to very high reliability was found for the temporal and distance parameters of gait, and the temporal symmetry index based on the difference in single-limb support duration between each leg (r = 0.85 to 0.98; intraclass correlation coefficients (ICC)(2,1) = 0.82 to 0.98). Significant differences were found between the two trials for velocity, stride length, and total double support (p < .05). Despite the high reliability coefficients, 95% confidence intervals, which take into account the random and systematic error, were wide for all parameters. These wide confidence intervals indicate that the use of two consecutive measurements for interpreting an individual patient's change would not be a sensitive method for monitoring progress or deterioration during rehabilitation. Strategies that may improve the clinical usefulness of temporal and distance gait measures in stroke rehabilitation are discussed. These include further reducing error sources, increasing data collection per measurement, using serial measurements on each patient, or using less rigorous confidence intervals.