Measurement and management of errors in quantitative gait data

Gait analysis is a valuable tool in the evaluation of children and adults with movement disorders. The data produced from gait analysis, however, is not necessarily free of errors. The purpose of this study was two-fold: (i) to estimate the errors associated with quantitative gait data; and (ii) to propose a method for incorporating the knowledge of these errors into the clinical interpretation process. An experimental protocol was designed that allowed within-subject, within-observer and between-observer errors to be computed at each point in the gait cycle. The estimates were then used in a practical scheme for detecting significant deviations in joint angles. The results of this study provide a means for managing error, while simultaneously improving the rigor and objectivity of clinical interpretations. (C) 2003 Elsevier B.V. All rights reserved.