Test–retest reliability of step counts with the ActivPAL™ device in common daily activities

The ActivPAL device is a well-established physical activity monitor for assessment of physical activity. Aim: To investigate test–retest reliability of step counts and establish minimal detectable changes (MDC) in step count to account for intra device error over time in various physical activities. Methods: Healthy participants (n = 24, age range, 19–28 years) performed activities on two occasions, 1 week apart, in a laboratory setting; self-paced floor walking, treadmill walking at three different speeds (3.2 km/h, 4.5 km/h and 4.5 km/h with incline), treadmill jogging (8.0 km/h), stair walking and cycling on an exercise bike at three speeds (45 rpm, 60 rpm and 75 rpm). Relative reliability was calculated using intraclass correlation coefficient (ICC) and Spearman correlation. Absolute reliability was assessed using standard error of measurement (SEM) and coefficient of repeatability (CR). Results: The ActivPAL showed high to very high relative reliability for treadmill walking at all speeds and stair walking, while self-paced normal floor walking showed moderate reliability. The absolute reliability was the best for treadmill walking activities, slightly increased for self-paced walking, followed by stair walking and jogging. The use of activity monitors during cycling has been questioned and our results confirm a low absolute and relative reliability. MDC values varied according to the type of activity e.g. treadmill walking 4.5 km/h (10 steps), walking on the floor (45 steps). Data loss in this study (10–13%) was higher than previously reported. Conclusions: The ActivPAL is reliable for treadmill walking, jogging and self-paced walking. MCD varies according to the activity and should be considered when establishing true change over time.