Reliability and agreement in gait measurements among patients with brain injury

Abstract

The aim of this study was to evaluate the agreement and test–retest reliability in two walking tests. A consecutive sample of 20 voluntary subjects with brain injury was recruited. Velocity, cadence, right- and left-step length were measured using the GAITrite electronic walkway system and the Ten Meter Walking (TMWT) paper and pencil test. The gait evaluation methods (GAITrite and TMWT) were placed on top of each other. The subjects performed three comfortably paced walking trials approximately 30 min apart. Cronbach's alpha, typical error, change in the mean and Bland–Altman plots, 95% limits of agreement (LOA) were used as statistics of reliability for the agreement between the tests. The intra-class correlation coefficient (ICC), LOA and coefficient of variation (CV) were used as statistics for test–retest reliability. In addition, the objective was to ascertain inter- and intra-rater reliability of these tests. Overall, Cronbach's alpha (0.99) values showed excellent results in both tests. Test–retest and inter-rater ICC values were good (0.95–0.99) in both tests and measured parameters. The absolute reliability results were good in all tests. This study showed acceptable agreement, test–retest and inter-rater reliability of gait measured in both GAITrite and TMW tests in brain-injured patients.

• Agreement
• brain
• gait
• rehabilitation
• reproducibility of results
• stroke
• trauma
• walking