Assessing the accuracy of 3-D kinematic data: Implications for laboratory setup

Abstract

The reproducibility of quantitative gait analysis measurements is an important consideration when analyzing data of both normal subjects and patients. Waveform similarity statistics were used to assess the reproducibility of lower extremity kinematic and kinetic data collected on 5 normal adult subjects. For each subject, gait evaluations were done on 3 separate test days with 3 trials collected each day. Pelvis angles were fairly repeatable, however, with notably poor reproducibility in sagittal plane tilt. Re-application errors of the sacral wand coupled with a small range of motion are believed to be the principle contributors to variability in this pelvis angle. The sagittal plane angles for the hip, knee, and ankle demonstrated excellent repeatability within test days and between test days. Frontal and transverse plane angles were fairly repeatable within test days but between-day repeatability was considerably lower. The `downstream' errors accompanying Euler angle calculations coupled with the smaller ranges of motion in the non-sagittal plane angles are suggested to contribute to the ambient variability already in these data and, thus, affect the within-day repeatability. Errors in re-application of midthigh and midcalf wands further contribute to the variability of these data between test days. Net resultant joint forces and moments were repeatable overall with slightly lower between-day statistical values. The exception to this observation was the adduction moment of the ankle which was notably variable between test days. Several explanations for this variability are presented. Results of this study have led to modifications in some marker placement procedures and reinforced the need for others already being used.