Inter-laboratory consistency of gait analysis measurements

The dissemination of gait analysis as a clinical assessment tool requires the results to be consistent, irrespective of the laboratory. In this work a baseline assessment of between site consistency of one healthy subject examined at 7 different laboratories is presented. Anthropometric and spatio-temporal parameters, pelvis and lower limb joint rotations, joint sagittal moments and powers, and ground reaction forces were compared. The consistency between laboratories for single parameters was assessed by the median absolute deviation and maximum difference, for curves by linear regression. Twenty-one lab-to-lab comparisons were performed and averaged. Large differences were found between the characteristics of the laboratories (i.e. motion capture systems and protocols). Different values for the anthropometric parameters were found, with the largest variability for a pelvis measurement. The spatio-temporal parameters were in general consistent. Segment and joint kinematics consistency was in general high (R2>0.90), except for hip and knee joint rotations. The main difference among curves was a vertical shift associated to the corresponding value in the static position. The consistency between joint sagittal moments ranged form R2=0.90 at the ankle to R2=0.66 at the hip, the latter was increasing when comparing separately laboratories using the same protocol. Pattern similarity was good for ankle power but not satisfactory for knee and hip power. The force was found the most consistent, as expected. The differences found were in general lower than the established minimum detectable changes for gait kinematics and kinetics for healthy adults.