Test-retest reliability of trunk accelerometry during standing and walking

Abstract

Objective: To investigate repeatability of acceleration measured by a portable, triaxial accelerometer over the lumbar spine as a measure of balance during standing and walking. Design: Acceleration was measured along three axes and transformed to a horizontal-vertical coordinate system. Standing was tested on two feet, vision unobstructed and obstructed, and on one foot, vision unobstructed. Walking was tested in five different self-selected speeds on even and uneven ground. Retest was performed after 2 days. Setting: A flat floor with a 7m uneven mat. Subjects: Nineteen healthy students. Main Outcome Measures: Standing: Acceleration root mean square (RMS). Walking: Point estimate of acceleration RMS for a reference walking speed (1.2m/sec). All tests: Within-subject standard deviation (S W), intraclass correlation coefficients (ICCs). Results: The S w indicated high absolute test-retest repeatability for standing on two feet. Because of restricted ranges of value, relative reliability was low, however, with ICC(3,1) < .56. For standing on one foot, absolute and relative reliability were highest in the mediolateral direction, with ICC(3,1) = .84. For the walking tests, ICC(3,1) ranged from .79 to .94 for the three axes. Reliability was on the same level for even and uneven ground. Conclusion: The results indicate that accelerometry without need for stationary instrumentation is a reliable method that may be useful for studying standing balance and gait in the clinic.