Abstract
PURPOSE: The GEM scale is an objective assessment tool, specifically developed for older adults, to evaluate walking safety using standardized tasks. The purpose of this study was to estimate the interrater and test-retest agreement of the GEM scale. METHOD: Participants (n = 41; >/= 65 years) were recruited from geriatric units and assessed independently and simultaneously by three raters on two occasions using the GEM scale. Kappa coefficients and percentage agreement were calculated for each item of the scale. RESULTS: A majority of walking items (n = 22) showed fair to substantial interrater agreement (kappa >/= 0.25) and substantial to almost perfect test-retest agreement (kappa >/= 0.60). Mean percentage agreement was high for both interrater and test-retest agreement (79% +/- 15% and 83% +/- 16% respectively). Moreover, detailed analyses demonstrated that the relatively low agreement of some items resulted from changes in the performance of some participants and the low variability of scores. Although some walking items showed less agreement, the final decision regarding the participants' ability to walk safely resulted in moderate to substantial interrater and test-retest agreement. CONCLUSION: The GEM scale is a new assessment tool that can now be used with estimated interrater and test-retest properties to allow therapists to objectively evaluate walking safety among the elderly.