Interlaboratory Reliability of 3D Gait Analysis

Overview

The purpose of this study is to determine the interlaboratory reliability of 3D gait analysis performed on healthy subjects on an instrumented treadmill in three different laboratories. This will allow us to evaluate whether the pooling of control data is feasible in future studies. This will support our ambitions to enter into future collaborations to conduct studies with larger samples. Moreover, this way we will know if we can share patient data in the future so that patients can be specifically referred to a lab with expertise on a particular evaluation, e.g. the lab in Maastricht has expertise on perturbations, UZ Gent has expertise on full-body measurements and UHasselt has expertise on balance. During this study approximately 30 healthy subjects will be recruited (10 from each center) from the staff and students at each center. These subjects are own staff and students of the 3 centers. Each subject will undergo a 3D gait analysis at each center (at UZGent, UHasselt and UMaastricht). Therefore, each subject will be evaluated 3 times. The time between each evaluations will be a maximum of one week. The evaluations will be performed as per standard procedure in each lab. However, a standardized model (HBM2 lower limb), an agreed walking velocity (0.9 m/s, 1.0 m/s, 1.1 m/s) and an agreed duration per recording (3 min) will be used. The same software will be used to process the collected data. 3D kinematics and kinetic time-series data of the lower limbs will be extracted as outcome data. Absolute reliability indices will be calculated in the following way: First, the standard deviation (SD) of time-series data across different gait cycles will be calculated per participant per center. The SD values across different centers per participant will then be averaged. Finally, the SD values across different participants will form the inter-laboratory reliability indices. In addition, we will use statistical parametric mapping (Anova repeated measures) to compare the full time-series data of each outcome parameter.