Interobserver concordance analysis for the evaluation of internal rotation with elbow at the side during shoulder clinical examination.

Introduction: Internal rotation with the elbow against the body (IR1) is a crucial functional range of motion of the shoulder for daily activities. Its restoration following prosthetic surgery, particularly after reverse shoulder arthroplasty, remains unsatisfactory and unresolved. In clinical practice, its assessment can be challenging and subject to variability depending on the examiner. One method for evaluating IR range of motion is to determine the highest vertebral level reached by the patient. The objective of this study was to analyze interobserver agreement for the assessment of internal rotation using this commonly employed method, which is also utilized in the calculation of the Constant score.

Hypothesis: It was hypothesized that IR1 measurement based on the highest vertebral level reached would demonstrate a high degree of interobserver agreement.

Materials and methods: A total of 285 photographs of healthy volunteers performing a IR1 movement, as instructed prior to image acquisition, were independently and blindly assessed by four experienced shoulder surgeons. The range of motion was classified into six levels: thigh, buttocks, lumbosacral junction, waist (L3), T12 vertebra, and interscapular region (T7). Interobserver agreement was assessed using Fleiss' kappa, while pairwise agreement between raters was calculated using Cohen's kappa.

Results: Overall interobserver agreement was strong (K = 0.7, p-value < 0.04). In the paired analysis using Cohen's kappa, interobserver agreement was almost perfect between certain raters, with values reaching up to 0.89. Detailed kappa analysis showed almost perfect agreement for the IR1 levels corresponding to the thigh and the interscapular region (K = 0.91 and 0.85, respectively). Agreement was strong for the buttocks level (K = 0.66) and moderate for the intermediate levels (T12, L3, and the lumbosacral junction) with K values of 0.57, 0.47, and 0.53, respectively.

Discussion: The interobserver correlation for this method of assessing IR1 range of motion is strong, suggesting that evaluating IR1 based on the highest vertebral level reached is a reproducible method for clinical shoulder examination.

Level of evidence: III; prospective study.