Reliability and validity of a dual-task force platform assessment of balance performance: Effect of age, balance impairment, and cognitive task

OBJECTIVES: The primary aim of this study was to identify whether performing a balance task on a force platform while performing a cognitive task (backwards counting by threes) could discriminate between healthy older people and older people at mildly increased risk of failing better than performing the balance task alone. Additional aims were to identify the effect of age on balance performance and to investigate the retest reliability and concurrent validity of the balance measures. DESIGN: Cross-sectional-cohort, three-group comparison design. SETTING: Gait and balance testing laboratory. PARTICIPANTS: Twenty healthy young adults (HY, mean age standard deviation = 26.4 +/- 6.1 years), 20 healthy older adults (HO, 73.8 +/- 6.0 ycars), and 20 older adults with mild increase in falls risk (MIFR, 74.8 +/- 7.3 years). MEASUREMENTS: All subjects were tested on the balance task alone and while performing a concurrent cognitive task oil the Chartecx Balance System (CBS) under three platform conditions: stable platform (stable), dynamic platform with forward-backward tilting (F-B), and dynamic platform with side-to-side tilting (S-S) and on the step test, timed up-and-go, and gait measures. RESULTS: There was a significant difference between HY and HO on the S-S and F-B conditions, both with and without the addition of the cognitive task (P <.05), but not under stable platform conditions. There were also significant differences between the 110 and MIFR groups under stable and F-B conditions (P <.05). A three-way analysis of variance identified significant main effects for group, platform condition and task (P <.05), group-by-platform interaction, and platform-by-task interaction. The F-B condition with the cognitive task most effectively discriminated between groups. Post hoc analyses revealed that the F-B condition with cognitive task was associated with the highest accuracy in classifying subjects from the two older groups, with a sensitivity of 0.8 and a specificity of 0.8. Retest reliability for most CBS measures was moderate to high, and correlations between dynamic CBS measures and clinical measures were high. CONCLUSION: Measurement of dynamic balance on the CBS (F-B) with a cognitive task is a reliable measure that discriminates well between healthy older people and those with a mild increase in risk of falling.