The Standardized Three-metre Walking Test for elderly people (WALK3m): repeatability and real change

Abstract

Objectives: (i) To examine the relative and absolute retest within-session reliability of the Standardized Three-metre Walking Test for elderly people (WALK3m), using walking time and number of steps as measures. (ii) To give clinicians details of how they can use the minimal detectable change in order to decide whether a patient's WALK3m performance has really changed. (iii) To give rules to decide whether one trial rather than three is sufficient.

Design and analysis: Subjects made three trials of WALK3m in the same session. Variability and level of performance were related so the data were log (ln) transformed. Relative reliability was measured by the intraclass correlation coefficient (ICC) then absolute reliability by the repeatability factor (RF) and minimal detectable change (MDC) (one-tailed and two-tailed) was derived from the within-subjects standard deviation.

Setting: Rehabilitation areas in the acute geriatric medical units at three south-west London hospitals.

Subjects: The standardizing sample consisted of 81 inpatients approaching discharge to their own homes.

Results: The ICC for walking time was 0.97 and for step counts 0.96. The MDC for walking is a 36% decrease between first and second measurements. The corresponding value for step counts is 21%.

Conclusion: WALK3m, whether measured by time or by step counts has excellent repeatability. Knowing the minimal detectable change can be of great value to clinicians who can calculate whether their patient is really making progress.