Can change in prolonged walking be inferred from a short test of gait speed among older adults who are initially well-functioning?

Background. The ability to walk for short and prolonged periods of time is often measured with separate walking tests. It is unclear whether decline in the 2-minute walk coincides with decline in a shorter 20-m walk among older adults. Objective. The aim of this study was to describe patterns of change in the 20-m walk and 2-minute walk over 8 years among a large cohort of older adults. Should change be similar between tests of walking ability, separate retesting of prolonged walking may need to be reconsidered. Design. A longitudinal, observational cohort study was conducted. Methods. Data were from 1,893 older adults who were well-functioning (÷70 years of age). The 20-m walk and 2-minute walk were repeatedly measured over 8 years to measure change during short and prolonged periods of walking, respectively. Change was examined using a dual group-based trajectory model (dual model), and agreement between walking trajectories was quantified with a weighted kappa statistic. Results. Three trajectory groups for the 20-m walk and 2-minute walk were identified. More than 86% of the participants were in similar trajectory groups for both tests from the dual model. There was high chance-corrected agreement (kappa.84; 95% confidence interval.82,.86) between the 20-m walk and 2-minute walk trajectory groups. Limitations. One-third of the original Health, Aging and Body Composition (Health ABC) study cohort was excluded from analysis due to missing clinic visits, followed by being excluded for health reasons for performing the 2-minute walk, limiting generalizability to healthy older adults. Conclusions. Patterns of change in the 2-minute walk are similar to those in the 20-m walk. Thus, separate retesting of the 2-minute walk may need to be reconsidered to gauge change in prolonged walking.