Research Review By Dr. Rob Rodine©

Date Posted:

January 2010

Study Title:

Reliability and diagnostic accuracy of 5 physical examination tests and combination of tests for subacromial impingement


Michener LA, Walsworth MK, Doukas WC & Murphy KP

Author's Affiliations:

Department of Physical Therapy, Virginia Commonwealth University, Medical College of Virginia Campus

Publication Information:

Archives of Physical Medicine & Rehabilitation 2009; 90: 1898-1903.

Background Information:

Shoulder pain is a common problem. As Michener et al. point out it accounts for one-third of physician office visits in the U.S. However, the shoulder is an elusive joint structure when attempting to reach a diagnosis. Jia et al. (1) discuss several diagnostic considerations in their review of shoulder pain and examination. Here they explore the contribution of rotator cuff disease (including tendinosis, bursitis, partial- and full-thickness tears), acromio-clavicular joint disease, gleno-humeral instability, labral lesions and biceps tendon pathology.

Jia et al. state that impingement syndrome should be considered during the early stages of rotator cuff disease, when inflammation and swelling causes compression and aberrant mechanics within the subacromial joint space. Provocation of pain through compression in this space is the mechanism behind orthopeadic tests such as the Hawkins-Kennedy and Neer tests, which are tailored to ‘impingement syndrome.’

Most interestingly however is the controversy noted by Michener et al. regarding ‘impingement syndrome,’ whereby no consensus exists for its diagnosis, despite being considered the most frequent cause of shoulder pain.

The primary aim of this study was threefold. Firstly, the intent was to determine the inter-rater reliability of 5 orthopeadic tests for subacromial shoulder impingement. A secondary intent was to determine the diagnostic accuracy of orthopeadic tests when compared to a surgical reference. Uniquely however, this study not only sought to examine the diagnostic accuracy of tests in isolation, but also specific combinations of tests and aiming to determine a threshold of positive tests for the diagnosis of subacromial impingement.

Pertinent Results:

Of the 55 subjects who underwent arthroscopy, 16 were diagnosed with subacromial impingement syndrome. All but 1 of these 16 patients had a concomitant shoulder pathology such as a rotator-cuff tear, labral tear, instability or degenerative joint disease of the acromio-clavicular joint. This has significant impact on the study’s results as concomitant pathology may affect patient responses to specific orthopeadic tests. For example, the positive test result may be cause by the mechanical impact of a rotator-cuff tear as compared to an enlarged or fibrotic bursa.

Inter-tester reliability (agreement between different clinicians) was found to be moderate for all five orthopeadic tests. The highest percentage agreement was found with the external rotation resistance test (87%), followed by the empty-can test (76%), the painful-arc (73%), the Neer test (71%) and the Hawkins-Kennedy test (69%).

The sensitivity, specificity, positive likelihood and negative likelihood ratios for each test is reported below, respectively:
  • Hawkins-Kennedy: 63%, 62%, 1.63 and 0.61
  • Neer impingement: 81%, 54%, 1.76 and 0.35
  • Painful-arc: 75%, 67%, 2.25 and 0.38
  • Empty-can: 50%, 87%, 3.90 and 0.57
  • External rotation resistance: 56%, 87%, 4.39 and 0.50
As the Hawkins-Kennedy and Neer test were found to have poor clinical relevance within this study, their data were not used when evaluating combinations of tests. Regardless, no combination of clinical testing was found to correlate to the surgical reference when entered into a logistic regression model.

The receiver operator characteristic analysis determined that regardless of the combination of positive clinical tests, a threshold of 3/5 positive’s correlated with the surgical reference with 75% sensitivity and 74% specificity. The positive likelihood ratio of 3/5 positive tests was found to be 2.93.

Clinical Application & Conclusions:

While this study is not without limitations (see below), it does offer insight into the use of clinical examination procedures related to evaluation of subacromial impingement syndrome.

While providing details of isolated test value, this study guides the clinician to using a collection of tests and accepting a threshold as an indication of the presence of disease. This scenario is more applicable to the average patient, rather than just isolated tests which are applicable to the ‘text book’ patient.

More and more research is beginning to look at clinical testing procedures in this manner. Research support of such approaches helps the clinician through the diagnostic process and allows greater confidence in a diagnostic label. Unfortunately, as evidenced in this study, the answer is not always clear.

Study Methods:

This prospective single-blinded cohort design invited consecutive patients who presented to an orthopeadic surgeon’s office with a primary complaint of shoulder pain of at least one week’s duration to participate in this study. A total of 65 subjects were invited, with 55 subjects accepting and following through with subsequent surgery.

Subject demographics included a sample primarily composed of men (47:8), an average age of 40.6 yrs (SD 15.1; 18-83) and symptom duration of 33.8 months (SD 48.9; range 2-230).

All subjects underwent two separate, and blinded, physical examinations, one performed by a board-certified orthopeadic surgeon with 17 years experience and one performed by a board-certified physical therapist with 8 years of experience. Each examination included a full and independent historical and physical examination, without the knowledge of any previous imaging findings.

The five orthopeadic examinations performed on the shoulder were adequately described and performed in a consistent manner with other peer-reviewed publications as well as orthpeadic text-books. The Neer and Hawkins-Kennedy tests indicated a positive result when superior shoulder pain was reproduced. The painful-arc test was considered positive when pain was noted between 60 and 120 degrees of abduction.

The empty-can test and external rotation resistance tests indicated a positive result when weakness was detected in the involved shoulder. Following the clinical examination, the involved subjects underwent an arthroscopic procedure on the involved shoulder, providing a surgical reference standard.

Surgical examinations were performed an average of 2.6 months (SD 2.7; 1 day – 8 months) after the clinical examination. The involved surgeon was blinded to the previous clinical examination findings of each subject. A surgical diagnosis of subacromial shoulder impingement syndrome was made if arthroscopy found a visually enlarged or fibrotic bursa, a degenerated supraspinatus tendon. Subjects with full-thickness rotator-cuff tears, labral tears or instability would have been excluded from the study.

Inter-rater reliability was calculated via Kappa reliability coefficients. Diagnostic accuracy was calculated via sensitivity, specificity, and positive and negative likelihood ratios. Logistic regression was used to determine the accuracy of test combinations and a receiver operating characteristic curve analysis was used to determine the diagnostic accuracy of a threshold for the number of positive tests needed to identify subacromial impingement syndrome.

Study Strengths / Weaknesses:

Studies like this are often not perfect. The main weaknesses of this study are common to many studies investigating orthopaedic tests, and in this case include:
  • not including clinical history details
  • not utilizing all available tests related to impingement syndrome
  • looking at a primarily middle-aged male population seeking surgical consultation after experiencing predominantly chronic pain

Additional References:

  1. Jia X et al. Examination of the shoulder: the past, the present, and the future. JBJS 2009; 91(Suppl 6): 10-18.