Long-term reliability of the timed up-and-go test among community-dwelling elders

Abstract

The purpose of this study was to describe the long-term reliability of the Timed Up and Go (TUG) test among community dwelling elders (mean age = 75 years). Sixteen women and 4 men were tested 3 times (baseline, 6 weeks, and 12 weeks) at a community senior center. The reliability across the 3 tests was good (intraclass correlation coefficient = .80). The TUG times decreased significantly (p=.02) over the first (mean = 10.1 sec), second (mean = 9.1 sec), and third (mean = 8.9 sec) tests, but between test comparisons (Bonferroni adjusted t-tests) demonstrated no significant differences between any pair of tests. The technical error of measurement was 3.2 sec between the first and second tests and 4.0 sec between the first and third tests; but it was only 0.1 sec between the second and third tests. Bland-Altman plots showed the following 95 percent limits of agreement: test 1 and test 2 (-2.9 to 4.9 sec), test 1 and test 3 (-3.1 to 5.7 sec), test 2 and test 3 (-3.4 to 3.9 sec). The results of this study corroborate those of previous studies supporting the reliability of the TUG test among cognitively intact individuals. The results, however, suggest that considerable differences may exist between tests conducted over a period of weeks. Consequently, differences of several seconds may be necessary to conclude that a real change in performance has occurred. This is particularly true between a first and second test. Copyright © 2006 by the Society of Physical Therapy Science.