Manual strength testing in 14 upper limb muscles
A study of inter-rater reliability

Abstract

Background Manual muscle testing has been termed a “lost art” and is often considered to be of minor value. The aim of this investigation was to study the inter-rater reliability of manual examination of the maximal voluntary strength in a sample of upper limb muscles. Patients and methods The material consisted of a series of 41 consecutive patients (82 limbs) who had been referred to a clinic of occupational medicine for various reasons. Two examiners who were blinded as to patient-related information classified 14 muscles in terms of normal or reduced strength. In order to optimize the evaluation, the individual strength was assessed simultaneously on the right and left sides with the limbs in standardized positions that were specific for each muscle. Information on upper limb complaints (pain, weakness and/or numbness/tingling) collected by two other examiners resulted in 38 limbs being classified as symptomatic and 44 as asymptomatic. For each muscle the inter-rater reliability of the assessment of strength into normal or reduced was estimated by κ-statistics. In addition, the odds ratio for the relation to symptoms of the definition in agreement of strength was calculated. Results The median κ-value for strength in the muscles examined was 0.54 (0.25–0.72). With a median odds ratio of 4.0 (2.5–7.7), reduced strength was significantly associated with the presence of symptoms. Interpretation This study suggests that manual muscle testing in upper limb disorders has diagnostic potential.