Shoulder outcome measures - A comparison of 6 functional tests

Background: Several shoulder function scores are used in research, with no universally adopted standard. This study compares 6 shoulder outcome scales. Hypothesis: Correlations exist between shoulder outcome scales, allowing conversion between scales. Shoulder scales are correlated with age. Study Design: Regression and correlation study. Methods: Seventy subjects with shoulder pain completed 6 shoulder outcome scales. Pearson correlations were calculated between the total scores of the 6 instruments, between the components of the scales, and with age. Regression equations were calculated between scales. Results: The range of r values for total scores was 0.495 less than or equal to r less than or equal to 0.770, P less than or equal to .01. In general, a scale's components were themselves highly correlated and added little new information to the scale (0.260 less than or equal to r less than or equal to 0.705, P less than or equal to .05). Most of the scale scores were highly correlated with age (0.291 less than or equal to r less than or equal to 0.582, P less than or equal to .05). Constant's reported corrections for age reduced (from r = -0.582 to r = -0.250, P < .05) but did not eliminate age as a confounding variable. Conclusions: Correlations exist between shoulder outcome scales, but existing shoulder scales are not equivalent in their assessments of function; they contain redundant information and, in some cases, may reflect a patient's age better than his/her shoulder function. The utility of conversion equations is minimized as a result of low to moderate correlations between scales.