The objective of this study was to systematically review the reliability of physical examination procedures used in the clinical examination of patients with shoulder pain. MEDLINE, PEDro, AMED, PsychInfo, Cochrane Library (2009) and CINAHL were searched from the earliest record on the database to June 2009. Reliability studies that included any patients with shoulder pain were analysed for their quality and reliability results. Pre-established criteria were used to judge the quality of the studies (high quality >60% methods score) and satisfactory levels of reliability (kappa or intraclass correlation coefficient >/=0.85, sensitivity analysis 0.70). A qualitative synthesis was performed based on levels of evidence.Thirty-six studies were included with a mean methods score of 57%. Seventeen studies were deemed to be of high quality; high-quality studies were less likely to meet the pre-agreed level of reliability. The majority of studies indicated poor reliability for all procedures investigated. Overall, the evidence regarding reliability was contradictory.
There is no consistent evidence that any examination procedure used in shoulder assessments has acceptable levels of reliability. Alternate methods of classification which are reliable should be used to classify patients with shoulder problems.