Relationships Between 3 Classification Systems in Brachial Plexus Birth Palsy

Background: TheMallet scale, ActiveMovement Scale (AMS), and Toronto Test are validated for use in children with brachial plexus birth palsy (BPBP). However, the inability to compare these evaluation systems has led to difficulty gauging treatment efficacy and interpreting available literature in which multiple scoring systems are reported. Given the critical importance of physical examination, we compared 3 scoring systems to clarify statistical relationships between current validated evaluation methods. Methods: The medical records of children with BPBP treated at a single institution over a 14-year period were retrospectively reviewed. Modified Mallet, AMS, and Toronto scores were recorded throughout the entire period. Data were included if at least 2 complete scoring systems were documented during the same examination session. Spearman correlation coefficients were calculated for all composite and subscore combinations. A concordance table was constructed for select variables found to be highly correlated. Results: Total single-session score combinations were as follows: 157 Mallet and AMS, 325 AMS and Toronto, and 143 Mallet and Toronto. Composite AMS and Toronto scores were found to have a strong correlation (r= 0.928, P<0.001). A concordance table comparing these variables revealed that a Toronto score of 3.5 is concordant to an AMS score of 45. Modified Mallet scores had only a moderate correlation with composite AMS (r= 0.512, P<0.001) and Toronto (r= 0.458, P<0.001) scores. Specifically regarding the modified Mallet score, maneuvers requiring external rotation had stronger correlations with the composite modified Mallet score than maneuvers highlighting internal rotation. Conclusions: Modified Mallet scores do not correlate well with AMS or Toronto scores and should be utilized separately when managing children with BPBP. Similarly, AMS and Toronto scores are inadequate to guide clinical decisions for which the literature cites Mallet scores as outcome measures, and vice versa. Lastly, Mallet scores should incorporate an isolated internal rotation component to adequately assess midline function.