The two methods for testing protraction maximal isometric strength and the two methods for testing retraction maximal isometric strength indicated high intrasession and intersession reliability with a small standard of error of measurement. According to Portney and Watkins,36 values of ICC above 0.90 are indicative of good reliability for clinical studies.
The importance of these findings is twofold. First, these new procedures for measuring scapular muscle strength for protraction and retraction offer a possible clinical alternative to manual muscle testing. Second, two of the procedures offer an alternative to measuring protraction and retraction strength while avoiding the use of the humerus and rotator cuff.
The intraclass correlations (ICC's) found in our study were high and were similar to reports that used the Isobex® static dynamometer33
or measured protraction and retraction strength with other static dynamometers17, 19, 20, 25, 26
with a custom stationary apparatus16 or a modified isokinetic machine.22, 23
The ICC's observed in this study were higher than the 0.27 intrarater reliability for protraction (serratus anterior muscle) performed on baseball players using the traditional manual muscle testing method of applying a downward force on the subject's hand.21
This may have been due to the disparity of strength between the baseball players and testers, as well as the difficulty with the subject's coordination and stability when a force is applied at the distal end of the upper extremity kinetic-chain. One aim of this study was to determine if there was similar agreement between two alternative methods of measuring force output for protraction and retraction, with the difference between the two methods being the inclusion or exclusion of the GH joint during testing. Many studies have used Kendall et al's39
testing protocols which involve the GH joint when testing protraction and retraction. However, because some individuals may also present with GH joint dysfunction, it raises the issue if excluding GH joint during testing may affect force outcome measures. Although our ICC's agreement findings were considered strong, the Bland-Altman plots indicated that not involving the GH joint affected the readings.
For the two protraction methods there were only two outliers for the thirty force outputs at two standard deviations from the mean difference of −1.31. The small mean difference visually indicated that there was minimal disparity between the two methods of force measurement for protraction. However, as the force output increased there was a small bias towards the method not involving the GH joint. This bias may be due to the unaccustomed movement pattern and also to the perceived increased coordination necessary to involve the unsupported upper extremity.
For the two retraction methods there was only one outlier for the thirty force outputs at two standard deviations from the mean difference of 5.83. While both methods of measuring retraction were reliable, a mean difference of 5.83 indicated that the methods were not likely interchangeable. The bias towards retraction involving the GH joint may be due to the fact that pulling is a common movement pattern. In addition, since the humerus was elevated to ninety degrees when involving the GH joint, there may have been some affect from the deltoid and latissimus dorsi muscles. Furthermore, for retraction, there was a greater spread of force measurements within two standard deviations at all levels of force output.
There were limitations to this study. First, restricting the subject sample to young healthy females limits the generalizability of the findings. However, the intent was to establish reliability of the new measuring procedures in a confined homogenous population to minimize confounding. The reliability of these new testing methods and positions need to be further investigated with various population groups, validated, and taught to clinicians so they can assess scapular musculature without incorporating the GH joint. Second, the testing procedures appear to require stabilization from both sides of the trunk in the upright sitting position, both for countermovement stabilization and for prevention of synergistic trunk muscle activity in the direction being tested. Stabilization benches used for weight training purposes are readily available; however, one has to consider the width so it allows stabilization of the trunk while offering mobility of the shoulder girdle being tested. Alternative positioning methods using the supine and prone positions also deserve further study. The width of the trunk stabilization pad and the determination of the neutral position of the scapula should be considered. Third, some subjects had difficulty learning to quickly respond to the audio stimulus, which may have initially made it difficult to provide a maximal contraction effort for the full 5-second duration. Subjectively, it appeared that subjects that were athletic and familiar with resistance exercises learned the procedure more easily. Fourth, there was a trend of increasing mean force scores noted over the three testing sessions for all four methods. This suggested that there may have been some learning effect in spite of randomization and a strict testing procedure. However, further post hoc analysis using repeated-measures ANOVA revealed there was no real learning effect observed. Using three trials and using the mean measure helped to insure the truest score possible. Finally, while the Isobex® static dynamometer maximum measuring capability of 450 Newtons (approximately 46 kg) is appropriate for testing a normal patient population, it may not be sufficient for measuring individuals with strong, athletic physiques. With the appropriate stabilization it is unlikely that most athletes would exceed 450 Newtons. If extremely strong athletes are being tested there are various tension dynamometers that are available that could be used for the same protocol.
There were two key strengths of the study. First, several advantages of our new procedures using the Isobex® static tension dynamometer is that it is 1) a relatively inexpensive dynamometer, 2) is a more objective than manual testing grades,40
3) is relatively easy to learn to use, 4) is not dependent on the tester's strength, and 5) can be set-up in a typical patient management setting. Although the Isobex® dynamometer was chosen for this study, there are many other dynamometers that are commercially available. Second, it has been common to incorporate the GH joint when evaluating scapular dysfunction.39, 41–43
This study did not resolve the issue of whether to include or not include the GH joint when evaluating the scapular musculature in the healthy individual, and further study is warranted. However, in a clinical situation where the patient being evaluated for scapular muscle strength has problems such as limited GH range of motion, impingement with the GH joint at 90 degrees flexion, painful or dysfunctional rotator cuff musculature, elbow or wrist problems, or problems with their grip, then methods not requiring the upper extremity would seem prudent. The maximal strength force measurement is limited by the weakest link in the kinetic-chain.
The clinical application of this study is that it offers the clinician additional reliable, objective, and efficient quantitative methods for measuring force in the scapular stabilizing muscles, as well as being able to test subjects even if the subject has GH joint dysfunction or upper extremity problems. The clinician will be able to collect quantitative baseline information during the initial evaluation to determine muscle weakness or imbalances, as well as using the testing procedures for outcome measurements to determine the effect of interventions used for strengthening the scapular stabilizing muscles.