Reliability of isometric muscle strength tests for the trunk, hands and shoulders

The objective of this study was to determine the reliability of isometric muscle strength tests for the trunk, hand and shoulders. At the same time the inadequacy of correlation coefficients as useful manifestation of reliability was highlighted. Nineteen healthy subjects performed a test and retest with exactly 1 week interval. The measurements consisted of recordings of maximal voluntary contractions (MVC) during isometric back flexion, back extension, shoulder elevation, shoulder abduction and handgrip. No habituation to the equipment was made before the initial test since this is often for practical reasons not possible in occupational settings. All Pearson and Intraclass correlation coefficients showed r-values above 0.9. However, a significant increase from test to retest was present in the muscle strength measures for back flexion, back extension and handgrip. In conclusion all tests showed reasonable reliability on group level judged by 95% confidence intervals, however on the individual level the variation was wide according to the calculated limits of agreement. If this variation is not taken into consideration the tests can be of limited use on individual level. Finally, this study clearly showed the inadequacy of correlation coefficients as a single number estimate of reliability. (C) 2001 Elsevier Science B.V. All rights reserved.