Using the Actigraph to measure physical activity of people with disabilities: an investigation into measurement issues.

This study investigates measurement issues when using the Actigraph motion sensor to measure the physical activity of people with disabilities. Four volunteers with rheumatoid arthritis participated in the study. Activity levels were measured using the Actigraph over four consecutive weekdays. Activity counts were determined using Actigraph Standard Software. The daily percent of waking time spent at each intensity level was calculated and a weighted average calculated to determine a single daily measure of activity. The variability of measurement, over 1, 2, 3 and 4 days, was examined using these weighted and rolling averages. Activity counts were very low, with over 50% of waking time spent in the 'very low' intensity category, demonstrating a marked floor effect. Little variability (+/-0.11 counts/min) in the level of physical activity was found across 1-4 days. The limited variability (consistency across days) in combination with the floor effect indicates that a 1-day data collection period is adequate. The limitations of the Actigraph are thus related to the existing software options, which have been designed for people without rheumatoid arthritis. The Actigraph does not appear to be a sensitive measure in people performing very low physical activity.