Using 4+ to grade near-normal muscle strength does not improve agreement

Background: Manual assessment of muscle strength is often graded using the ordinal Medical Research Council (MRC) scale. The scale has a number of inherent weaknesses, including poorly defined limits between grades '4' and '5' and very large differences in the span of muscle strength encompassed by each of the six grades. It is not necessarily obvious how to convert a manual muscle test finding into an MRC grade. Several modifications which include intermediate grades have been suggested to improve the MRC scale and the current study examines whether agreement improves and variation in ratings decrease, with an intermediate grade between '4' and '5', in circumstances where such a grade would seem appropriate. The present study examined the hypothesis, that a modified MRC-scale which included the commonly used '4+' option, resulted in greater agreement between clinicians compared to the standard MRC-scale. Method: A questionnaire containing five simple clinical cases were distributed to a large convenience sample of chiropractors in Northern Europe, with instructions to grade the described muscle strength findings using the MRC scale. The scale was adapted (with/without an intermediate '4+' grade) depending on the preference of the individual respondent. The cases were designed in such a way as to suggest a muscle weakness in the grey area between '4' and '5', i.e. grade '4+' on the modified MRC scale. Results: A total of 225 questionnaires were returned (7% response rate). The average percentage agreement (across cases) in the standard MRC group was 64% [range 51%: 73%] (grade '4' in all cases). In the modified MRC group, the corresponding findings was 48% [38%: 74%] (grade '4' or '4+' in all cases). The mean average deviation analogue in the standard MRC group was 0.34 (range 0.34: 0.40), compared to 0.51 (range 0.39: 0.73) in the modified MRC group, indicating greater dispersion of scores in the modified MRC group. The Fleiss kappa was 0.02 (p < 0.001) and 0.13 (p < 0.001), respectively. Conclusions: Contrary to the original hypothesis, introduction of a '4+' grade did not clearly improve agreement or variability of ratings, despite eliminating the physical muscle testing by providing written descriptions of test findings and specifically designing these to suggest a weakness of grade '4+'.