RESUMO
OBJECTIVE: To investigate: (i) the intra-rater, inter-rater and test-retest reliability of the 12-step stair test; (ii) its correlation with other stroke-specific impairments; and (iii) the cut-off scores that best discriminate patients with stroke from healthy elderly subjects. DESIGN: Cross-sectional study. SETTING: University-based rehabilitation centre. SUBJECTS: Thirty-five subjects with chronic stroke and 29 healthy elderly subjects. METHODS: The 12-step ascend and descend test was administered along with the Fugl-Meyer Motor Assessment for the lower extremities (FMA-LE), hand-held dynamometer measurements of hip abductor and knee extensor muscle strength, the Five Times Sit to Stand Test (FTSTST), assessment using the Berg Balance Scale (BBS), activities-specific balance confidence scale (ABC) assessment, the 10-metre walk test, and the Timed "Up and Go" (TUG) test. RESULTS: The 12-step ascend and descend test showed excellent intra-rater, inter-rater and test-retest reliability. The test was positively correlated with FTSTST times, gait velocity, and TUG times, and negatively correlated with FMA-LE scores and BBS scores. A test performance of 15.22 seconds or less was shown reliably to discriminate healthy elderly subjects from stroke survivors. CONCLUSION: The 12-step ascend and descend test is a reliable clinical test that is inexpensive and easy to implement, and is useful for assessing the stair-walking ability of patients with chronic stroke.