Inter-rater reliability of the Rasch-modified medical research council scoring criteria for manual muscle testing in neuromuscular diseases.

It was argued that researchers and clinicians are not able to make judgments between most categories of the original Medical Research Council (MRC) scale and that a modified short version would reach higher agreement levels. We aimed to assess the inter-rater reliability for both the original and the Rasch-modified MRC scoring criteria of Manual Muscle Strength tests (MMSt) in patients with neuromuscular diseases. Two MRC scoring criteria were used to score muscle strength using MMSt in 40 muscle groups of the upper and lower limbs in patients with neuromuscular disorders. Three investigators performed the evaluations; the order of the MMSt and the use of the scales were performed according to the preferences of the investigators. The agreement coefficient (Gwet's AC 2 ) was used to compute the reliability. Sixty patients (mean age of 39.3 years ± 15.2) with neuromuscular diseases were included. The mean AC 2 for the muscle groups of the upper limbs ranged from 0.82 to 0.96 using the modified MRC scale and from 0.86 to 0.96 using the original MRC scale. The AC 2 for the lower limb muscle groups ranged from 0.80 to 0.91 (modified MRC scale) and from 0.87 to 0.93 (original MRC scale). These values might be interpreted as "almost perfect agreement" with no significant differences between the scales. The results indicate that both MRC scoring criteria have significant reliability among trained observers. Moreover, the Rasch-modified MRC scale is as reliable as the original MRC scale and can be used in future clinical studies.