Properties of the 'timed up and go' test: more than meets the eye.

BACKGROUND The 'timed up and go' test (TUG) is a simple, quick and widely used clinical performance-based measure of lower extremity function, mobility and fall risk. We speculated that its properties may be different from other performance-based tests and assessed whether cognitive function may contribute to the differences among these tests in a cohort of healthy older adults. OBJECTIVE To evaluate psychometric properties of the TUG in healthy older adults in comparison to the Berg balance test (BBT) and the Dynamic Gait Index (DGI). METHODS The TUG, DGI and BBT were assessed in 265 healthy older adults (76.4 ± 4.3 years; 58.3% women) who participated in a 3-year prospective study. The Mini-Mental State Examination, digit span and verbal fluency measured cognitive function. The one-sample Kolmogorov-Smirnov test evaluated deviations from a normal distribution and Pearson's correlation coefficients quantified associations. RESULTS The mean scores of the BBT, DGI and TUG were: 54.0 ± 2.4, 22.8 ± 1.5, 9.5 ± 1.7 s, respectively. The BBT and the DGI were not normally distributed (p < 0.001), but the TUG was (p = 0.713). The TUG times were mildly associated (p < 0.01) with digit span and verbal fluency and were related to future falls, while the BBT and the DGI were not. CONCLUSIONS The TUG appears to be an appropriate tool for clinical assessment of functional mobility even in healthy older adults. It does not suffer from ceiling effect limitations, is normally distributed and is apparently related to executive function. The BBT and the DGI do not share these beneficial properties. Perhaps the transferring and turning components of the TUG help to convert this relatively simple motor task into a more complex measure that also depends on cognitive resources.