Background: There is a paucity of reliability data for walking speed tests in complex conditions to assess functioning in healthy older individuals.
Aims: To evaluate the absolute intra- and intertest reliability of walking speed performed in basic and complex conditions in healthy older individuals.
Methods: Fifty-two men and women of mean age 69.7 ± 3.2 years were tested for habitual and maximal walking speed. Maximal speed was also assessed under different conditions, including walking on a path of reduced width; picking up objects; stepping over hurdles; stepping over hurdles wearing sunglasses and finally, carrying a box. Two testing sessions (separated by 4 weeks) of two trials each were administered. Reliability was analysed by intra-class correlation coefficient (ICC), minimal detectable change (MDC) and Bland-Altman plots with limits of agreement (LOA).
Results: Intrasession ICCs ranged from good to excellent (0.89-0.95) except for picking up objects (0.44). Intersession ICCs were moderate to good (0.60-0.78) and %MDCs were acceptable (14-24%). Bland-Altman plots suggested a good agreement between the two testing sessions at group level (mean differences from - 0.02 to - 0.11 m/s), and limited agreement between testing sessions at individual level (upper LOA from 0.13 to 0.37 m/s and lower LOA from - 0.29 to - 0.49 m/s).
Conclusions: Complex walking speed tests are generally reliable measures displaying good and moderate intra- and inter-session reliability. Such tests seem a more suitable functional assessment tool for heathy older subjects compared with simple walking. Some learning effect may be present and further reliability studies are needed.
Keywords: Aging; Bland–Altman plot; Functional assessment; Gait speed; Minimal detectable change; Reproducibility.