An examination of bias in shoulder scoring instruments among healthy collegiate and recreational athletes.

This investigation examined whether gender, activity level, hand dominance, or age displayed bias for 5 common shoulder scores among 120 healthy collegiate or recreational athletes. Data were collected for 5 instruments: Constant-Murley, UCLA rating, pre-1994 American Shoulder and Elbow Surgeons, Shoulder Pain and Disability Index, and Oxford. Normalized total and subscale scores and effect sizes were analyzed to determine how each demographic variable affected the scores. The normalized scores for the 5 instruments were not equivalent. The normalized total scores were significantly lower for the Constant-Murley score (P <.0001) compared with those for the other instruments. Men had significantly higher Constant-Murley scores, primarily because of the strength subscale. The Constant-Murley score also displayed gender and age biases. The other instruments had mostly negligible variance attributable to gender, activity level, hand dominance, or age. In the absence of a universal validated method, shoulder scoring instruments should be carefully chosen to match the population and the purpose of the study.