Examining the test-retest reliability and construct validity of the Six-Spot Step Test in older adults with self-reported balance problems

Abstract

Objective: To evaluate the test-retest reliability and construct validity of the Six-Spot Step Test in older adults. Design: Cross-sectional study. Setting: Outpatient clinics. Subjects: Sixty-six adults aged ⩾65 years with self-reported balance problems. Main measure: The Six-Spot Step Test was completed twice on day 1. Three days later, the Six-Spot Step Test, the Timed “Up and Go” test, Dynamic Gait Index, the Mini-BESTest, and standing balance test were conducted. Reliability and validity were estimated using Bland-Altman statistics, minimal detectable change, Intraclass Correlation Coefficient (ICC), and Spearman’s rank correlation coefficient (ρ). Results: Participants had a mean (SD) age of 75.5 (6.6) years, a median (Q1–Q3) Montreal Cognitive Assessment test score of 26 (24–27), and a median (Q1–Q3) Falls Efficacy Scale International score of 22 (20–26). The study found within-day and between-day limits of agreements of ±15.7% and ±18.1%, respectively. The minimal detectable changes at the 95% level of confidence for within-day and between-day were 17.5% and 21.6%, and the ICCs were 0.96 (CI: 0.92–0.98) and 0.94 (CI: 0.86–0.97), respectively. A learning effect was observed between the first and second Six-Spot Step Test. Moderate to strong correlations were found between the Six-Spot Step Test and Mini-BESTest (ρ = −0.62), Dynamic Gait Index (ρ = −0.55), and the Timed “Up and Go” test (ρ = 0.79). Weak correlations were identified between the Six-Spot Step Test and static standing balance test (ρ < 0.38). Conclusion: The Six-Spot Step Test shows excellent relative consistency and moderate to high construct validity in older adults. A change of 21.6% is a true change in outcome.