Purpose. To explore (1) test-retest reliability, (2) measurement error at one point in time (standard error of measurement, SEM), and (3) measurement error between two measurements (smallest real difference, SRD) of 6 mobility tests in the community-dwelling elderly. The SRD can be used to determine whether a real change has occurred.
Methods. A physiotherapist administered the 6 mobility tests (functional reach [FR]; usual and fastest gait speed [UGS & FGS]; timed chair stands [TCS]; timed up & go [TUG]; and 6-minute walk distance [6MW]) for 77 community-dwelling elderly in 2 sessions, 1 week apart.
Results. Test-retest reliability was excellent (intra-class correlation coefficient2,1=0.80-0.95) for all measurements. The Pearson’s correlation coefficient (r) among the 6 mobility tests varied from 0.2 to 0.9; UGS & FGS, 6MW, and TUG were highly correlated (r?0.8). The standard error of the measurement and smallest real difference of all measurements
were within 10% and 26%, respectively. The FGS showed the highest test-retest reliability (ICC2,1=0.95) and was most responsive for detecting a real change (SRD%=10.3%). The FGS was highly correlated with 6MW (r=0.87-0.88) and UGS (r=0.79-0.80).
Conclusion. The 6 mobility tests provided reliable measurements of
mobility functions for the community-dwelling elderly.
