Comparison of base of support size during gait initiation using force-plate and motion-capture system: A Bland and Altman analysis

This study aimed to estimate the error made by investigators when force-plate data are used to approximate base of support size during gait initiation. Step length and step width obtained with a method based on motion capture system (Kinematics method, considered the “gold standard”) and with a method based on the centre of pressure traces obtained from a force-plate (Force-plate method) were purposely compared using descriptive statistics and the Bland and Altman (BA) method. Participants (N=19) performed series of gait initiation in Spontaneous and Maximal Velocity Conditions (SVC and MVC, respectively). BA analysis showed that 1) step length and width biases, corresponding to the difference between the two methods, were very small (<2.1%) in both velocity conditions and 2) the 95% limits of agreement of the BA plots ranged between 10% and 15% in absolute value. Repeated measures ANOVA showed that step length was significantly larger in MVC than in SVC, with no velocity X method interaction. There was no significant effect of the method on both step parameters. The present results suggest that the Force-plate method is sufficiently accurate to compare step parameters across conditions. However, researchers should be aware that non-negligible errors might occur when considering individual data.