Measurement error and minimum detectable change in 4-meter gait speed in older adults

Abstract

Gait speed is a commonly-used assessment and outcomes measure in geriatric clinical and research settings. Although relative reliability of usual gait speed has been well studied in community- dwelling older adults, less emphasis has been placed on a measure of absolute reliability (the standard error of measurement [SEM]), and on an associated clinically relevant index of real change in gait speed, minimum detectable change (MDC). The purpose of this study was to quantify measurement error and MDC for usual gait speed over 4 meters in community-dwelling older adults ambulating at intermediate and fast speeds. Community-dwelling older adults ambulating at intermediate gait speed (IGS), (n=15, mean age 74.2 yrs) and fast gait speed (FGS), (n=15, mean age 72.1) were included in this study. Participants performed two trials of gait speed over a distance of 4 meters. SEM and MDC at the 95% confidence level (MDC95) were computed for the IGS and FGS groups. Mean gait speed was 85.4 cm/s (IGS) and 129.9 cm/s (FGS). Measurement error (<5% of mean gait speed) and minimum detectable change (<13% of mean gait speed) were low in both groups. MDC95 was computed as 10.8 cm/s and 14.4 cm/s for the IGS and FGS groups, respectively. To be considered real change beyond the bounds of measurement error, change in 4-meter gait speed should exceed 10.8 cm/s (for intermediate speed ambulators) or 14.4 cm/s (for fast speed ambulators). Low measurement error in assessing 4-meter gait speed in community-dwelling older adults suggests that gait speed assessed over short distances has excellent reproducibility across trials. Low minimum change values suggest that 4-meter gait speed may be responsive and sensitive to change.