Identification of Potential Sources of Measurement Errors in an Isokinetic Dynamometer : Reliability Analysis of Shoulder Abduction and Flexion Data

Abstract: The evaluation of shoulder abduction and flexion strength is important in the rehabilitation after rotator cuff tear. The purpose of this work is to assess the intra and inter-rater reliability of measurement data from an isokinetic dynamometer used to evaluate shoulder abduction and flexion strength, with the aim to identify sources of measurement errors and suggest improvements. The measurement data was collected by a research group at Karolinska Institute and contained load and torque data from thirteen healthy subjects in the ages of 25 to 87 years. The measurements were carried out on two occasions, one week apart. Systematic differences between occasions are analyzed using the Shapiro Wilk test, the paired t-test, and Wilcoxon signed rank test. The agreement of the measurements is analyzed quantitatively using the coefficient of variation and the Bland Altman plot, and quantitively, using the intraclass correlation coefficient. A significant systematic difference in shoulder abduction and flexion load measurements was found, and the recommendation to prevent this is that components should be calibrated in a standardized way. The measurements showed varying reliability within and between measurement occasions and that after familiarization with the isokinetic dynamometer, repeatability improved. The findings indicate a need of a standardized protocol for patient education and placement. Measurements from the position sensor contributed to more random torque values. To improve the repeatability in measurements from the position sensor, axis of rotation should be kept aligned. An increasing variability in measurements with increasing load and torque was found. The recommendation is to use a preload for patients using more force in the movement, to make sure a preset speed is not exceeded, which contributes to more reliable measurements.