Reliability of gait variability assessment in older individuals during a six-minute walk test

Gait variability is an important indicator of gait performance. However, the reliability of the parameters used for its quantification, obtained from trunk linear accelerations, has still not been thoroughly investigated. The aim of this study is to assess the reliability of gait variability assessment in healthy older individuals based on lower trunk accelerations during a six-minute walk test and to examine the reliability of the data acquired in shorter periods. Twenty-nine subjects (84±5 years) performed the test while wearing one inertial sensor on the lower trunk. Gait variability parameters (standard deviation and coefficient of variation of the stride duration, and correlation coefficients of accelerations between neighbouring strides) were calculated from the accelerations over 35 rectilinear strides observed during six series of one-minute intervals extracted from the original signal. The reliability of these parameters was assessed using intraclass correlation coefficients (ICC). Results showed no significant changes across the six series for any of the parameters, with very high ICC values (0.93-0.95), indicating a strong reliability of the observed quantities. Therefore, gait variability analysis based on lower trunk acceleration data is a reliable and informative quantity in gait performance assessment in older individuals, and 1min interval is sufficient to ensure reliable results.