Test-retest reliability and variability of self-reported normal treadmill walking speeds in adults aged 20–80 years old

Background

Overground walking speeds show good to excellent reliability. However, intra-session test-retest reliability and variability of self-reported speeds during treadmill walking have not been quantified. Additionally, literature surrounding the effects of age on variability of overground walking speeds has presented mixed findings.

Objective

This study evaluated intra-session test-retest reliability and variability of self-reported normal speeds during a treadmill walking speed protocol, and changes in these measures across adults aged 20–80 years old.

Methods

Self-reported normal treadmill walking speeds were collected in two bouts (six and three speeds, respectively) during one session from 59 participants (20–80 years old). Intraclass correlation coefficients (ICC), standard errors of measurement (SEM), and Bland-Altman plots were derived for four sets of self-reported speeds, for the full sample and each age group (grouped by decade).

Results

Most ICCs presented good to excellent reliability (ICCs ≥ 0.80), with SEMs ≤ 0.45 km/h and ≤ 12 %. The earliest and latest self-reported speeds in each set of analyzed speeds contributed the most variation. The oldest groups demonstrated decreased reliability and increased variability compared to adults in the middle age groups, with similarly low reliability among the youngest group.

Conclusion

Results provide insight into the magnitude, patterns, and sources of reliability and variability in self-reported normal speeds during treadmill walking. Furthermore, reliability and variability of self-reported speeds during treadmill walking may be affected by age and/or walking speed. These findings may have utility for research and clinical settings, for applications in which reliably identifying normal walking speeds while minimizing time burden is an important consideration.