In clinical gait analysis it is standard practice to compare gait data from one patient with a reference data set of healthy subjects. Deviations from these reference curves are interpreted as gait impairments. There are however two problems with this approach. First, the variability in the reference curves, usually presented as a grey band, only takes into account the variability between subjects. It does not include variation between strides, as the curves per subject are already averaged across strides. Secondly, the timing of gait events may differ between subjects. This causes a temporal misalignment of the peaks and flattens the curve. This paper illustrates the impact of these problems for clinical gait analysis of children. When strides from typically developing children are compared with a reference data set, a large number of these strides are incorrectly classified as abnormal (on average 28%, 50%, and 51% for joint kinematics, moment, and powers, respectively). One solution when you are interested in measurement of peak values, is to compare them to the actual peak values of the group instead of looking at the curves. Moreover, the authors propose using an instrumented treadmill for gait analysis, allowing collection of multiple strides and using averaged curves instead of single stride data.
Check out the full article here: Oudenhove L.M. et al (2019) Gait & Posture
