This study’s objective was to investigate intertester and intratester reliability of the shoulder medial rotation test (MRT) and reliability differences depending on examiner expertise. Seventeen athletes with chronic shoulder pain took part in the study. Four independent observers with different experience levels simultaneously rated MRT performance as “correct” or “incorrect,” after a standardized assessment protocol, the same day (for intertester reliability) and in a 7-day interval (for intratester reliability). The intrarater reliability was admissible for 2 experts and one novice, with κ values ranging between 0.32 to 0.76 and poor for one novice (κ <0). Interrater agreement for all 4 assessors demonstrated slight agreement (κ = 0.06; 95% confidence interval: 0.06-0.47), increasing to fair agreement (κ = 0.33; 95% confidence interval: 0.21-0.69) when comparing the MRT findings between the 2 experienced assessors. Practice with the MRT in novices only slightly improved their degree of agreement.
Reliability of the MRT for detecting movement control of the shoulder girdle was fair at best for experienced examiners and poor overall. Dexterity and repetitive performance of the test is necessary for accurate interpretation of the MRT.