Intra-session test–retest reliability of the Timed “Up & Go” Test when performed by patients with hip fractures

Abstract

Abstract The aim of this study was to assess the reliability of the Timed “Up & Go” Test (TUG) when performed three times in one session by elderly hip fracture patients. Ninety-two patients in the orthogeriatric unit aged ≥ 65 years old were asked to perform the TUG three times at hospital discharge. Relative and absolute reliability was examined using intra-class correlation coefficients (ICCs) and standard error of measurement (SEM), respectively. The smallest detectable change on an individual level (SDCind) was calculated. Thirty-seven patients completed three trials and were included in the reliability analysis. For each trial, mean performance time became shorter. For TUG trial 2–3 both ICC 1.1 and 3.1 were 0.96 [95% confidence interval (CI) ICC 1.1: 0.78–0.94 and ICC 3.1: 0.92–0.98], indicating no learning effect between these trials. For TUG trial 1–2, ICC 3.1 (0.90, 95% CI: 0.82–0.95) was larger than ICC 1.1 (0.88, 95% CI: 0.78–0.94), indicating a learning effect. The 1.96*SEM was ± 10.22 seconds from TUG trial 1–2 and ± 5.32 seconds from trial 2–3. SDCind was 14.5 seconds and 7.5 seconds for TUG trials 1–2 and 2–3, respectively. In conclusion, in this small study, TUG scores had high intra-session test–retest reliability when three trials were conducted. Only 40.2% of the patients managed to complete all trials. Performing the test twice may be sufficient to provide a reliable TUG measurement in our patient group.