Reproducibility and accuracy of angle measurements obtained under static conditions with the Motion Analysis video system.

The development of computerized and semi-automated motion analysis systems has made the study of human motion more widely available in research and clinical settings. Although many of these systems are currently used by physical therapists, the accuracy and reproducibility of some of these systems in estimating joint angles have not been reported. In this study, the accuracy and reproducibility of angle measurements obtained by use of the Motion Analysis video system were evaluated under static conditions using a standard goniometer. Reflective markers placed on a goniometer were recorded by two video cameras at 17 angles, from 20 to 180 degrees, in 10-degree increments. Recordings of the goniometer were made at three locations within the field of view of the cameras. The intraclass correlation coefficient for each location tested was .99. Average within-trial variability was less than 0.4 degree at all locations. A linear regression of the system-calculated angles and reference angles for all locations had slopes near unity (ie, 1) and intercepts that were not statistically different from zero. A preliminary evaluation of the system under dynamic conditions revealed that distances were slightly underestimated, regardless of where the movement occurred within the calibration cube.