Influence of Sex and Body Size on the Validity of the Microsoft Kinect for Frontal Plane Knee Kinematics During Landings

Abstract

Three-dimensional (3D) motion capture is inaccessible, and the Microsoft Kinect is an alternative to measure surrogates of knee valgus that may contribute to anterior cruciate ligament (ACL) injury risk. We evaluated the influence of sex and body size on the agreement between methods. A total of 40 (10 per sex and BMI group) participants were included. The Kinect and motion capture measured knee ankle separation ratio (KASR) and knee abduction angles (KAAs). Intraclass correlation coefficients (ICCs) evaluated agreement between methods. 2 (sex) by 2 (BMI) by 2 (method) ANOVA compared kinematics between groups. Agreement between methods was moderate-to-good for KASR (initial contact ICCs 0.667–0.86; peak flexion ICCs 0.766–0.882). Agreement for KAA was low-to-moderate (initial contact ICCs 0.128–0.575; peak flexion ICCs 0.315–0.760). There was a BMI-by-method interaction for KASR at initial contact (p < 0.01) and a main effect of method (p < 0.01). There were BMI-by-method interactions for KAA (initial contact p > 0.01; peak knee flexion p < 0.01). The high BMI group had greater KAAs than the low BMI group, but only using motion capture. The Kinect is an alternative for measuring KASR, but not KAA. The high BMI group had greater KAAs than the low BMI group, but only when measured with motion capture.