Abstract
The reliability of several gravity correction procedures was examined using a KinCom trunk dynamometer in thirty subjects with no history of low back pain. The procedures, including the KinCom software procedure, an isometric mode and a passive force curve mode were evaluated for variability. Mean differences and standard deviation of difference scores were used to indicate systematic and random variability as well as coefficient of variation of method error to express variability relative to score magnitude. Overall, the isometric and eccentric passive force curve modes were most reliable while the KinCom software method showed greatest variability. In all modes, considerable variability was present between the angles of 5 to 15 degrees for flexion and between 5 to 20 degrees for extension. These findings indicate that alternative methods of gravity correction which may be more reliable than those provided by the supplier, should be used to ensure greater reproducibility of results.