Introduction

The physical examination is one of the building blocks of the diagnostic process in patients with acute shoulder injuries. The use of tests with good diagnostic properties is very important, and reliability is an important aspect in rating the diagnostic value of a test. Therefore, this study examined the reliability of 13 common shoulder tests used for detecting acute rotator cuff tears in patients with acute soft tissue shoulder injuries.

Methods

As a part of a diagnostic accuracy study, this study used a subset of patients to examine the reliability of 13 common shoulder tests (see figure below). Patients were recruited from a combined primary and secondary care emergency department. When patients following an acute shoulder injury presented with pain intensity of 4 or more at rest or during activity, with reduced active abduction range of motion or with decreased active external rotation range of motion, they were referred for follow-up measurements.

Results

So which of the 13 common shoulder tests appeared to be of use? Forty-eight participants were included in the study after a mean of 12 days after the shoulder injury. Substantial to excellent interrater agreement was found for the abduction strength test while the external rotation strength test showed moderate to excellent agreement and poor to substantial when performed conventionally and with the small finger test, respectively.

The inability to abduct beyond 90° showed substantial to excellent agreement, the reduction of external range of motion by 20° or more showed moderate to excellent agreement. Abduction range of motion assessed continuously showed good to excellent reliability while external range of motion had moderate to excellent reliability.

Questions and thoughts

Almost perfect agreement when evaluating the inability to abduct beyond 90° and substantial agreement when registering a loss of external rotation ≥20° compared to the uninjured side were observed. However, as these findings made part of the requirements for inclusion in the study, it seems logical that these observations were rated equally by both examiners.

Non-shoulder-specialized examiners were taking these measurements, which is very interesting since often these reliability examinations are held in highly specialized clinics where so called “shoulder experts” are conducting these tests. Therefore, we can assume that the tests, as described above, can be of value to be used in clinical practice by therapists with varying levels of experience. Furthermore, the range of motion and strength measurements in this study were assessed clinically without the use of specialized equipment, which may be interesting since not every physiotherapy practice has access to goniometers or dynamometers.

The study concludes“the results of the present study indicate which physical examination tests may be effective in detecting acute rotator cuff tears in patients during the acute phase of shoulder injury in the first line setting.” However, given the fact that only a minority of these patients (27%) had a rotator cuff tear AND given that this study did not examine the diagnostic properties of these shoulder tests, this statement is not correct. As this study only investigated the interrater reliability, we only know that these tests are reliable and therefore repeatable, producing consistent results when performed by different reviewers. We do not know from these results to what extent these shoulder tests are effective in predicting/diagnosing an acute rotator cuff tear. For this, a diagnostic accuracy study is needed.

Talk nerdy to me

A limitation of this study is the fact that no sample size calculation was performed. Rather the authors based their required sample size on other reliability studies of shoulders with different pathologies.

Generally speaking, tests examining pain showed lower reliability than tests assessing range of motion and strength. Therefore, we can advise relying on assessments of motion and strength, rather than on pain measurements. Therefore, the Hawkins test, for example, can be of lower value as it has a lower reliability than for example the active abduction range of motion test.

The confidence intervals were sometimes wide, therefore we have interpreted both the lower and upper limits to carefully depict the agreement. A possible reason for the wide intervals include that a training effect may have impacted the findings, for example: when a patient experienced pain during the first test, they may have been more hesitant during the second, and on the other side, patients who did not experience much pain may have pushed their limit further during the second test.

When looking at the Bland-Altman plots, we can see that the scores for the continuously measured active abduction range of motion and active external range of motion show no systemic bias. Using the standard error of measurement as displayed in Table 3, we can calculate the smallest detectable change of 42° and 25° for active abduction and external rotation range of motion respectively, which is relatively high considering the normal active range of motion.

Take home messages

The inability to actively abduct the shoulder beyond 90° and a loss of 20° or more in external rotation are reliable measurements. Evaluating the inability to abduct beyond 90° showed higher reliability than a loss of external rotation of 20° or more compared to the uninjured side. Equally, abduction strength was more reliable than external rotation strength. The total active abduction range of motion showed higher reliability than the total active external rotation range. This study however did not examine the diagnostic accuracy of these tests to detect acute rotator cuff tears. Therefore, based on these results alone, we cannot conclude its usefulness for diagnostic purposes. 

“The main result of the present study is that clinical assessment of active range of abduction and external rotation (expressed by the inability to abduct >90° and external rotation reduced by ≥20° compared to uninjured side) and abduction and external rotation strength expressed best reliability among the included tests in patients in the acute phase of shoulder injury.”

Reference

Schmidt, M., Enger, M., Pripp, A. H., Nordsletten, L., Moosmayer, S., Melhuus, K., & Brox, J. I. (2021). Interrater reliability of physical examination tests in the acute phase of shoulder injuries. BMC musculoskeletal disorders, 22, 1-9.