Accurate assessment of physical activity (PA) is necessary to identify the sedentary older individual who is in need of activity intervention. Activity monitors are quite popular, although it has been suggested that they are less accurate at slow gait speeds.
Purpose: To examine the accuracy of the three activity monitors in older individuals who walk at various gait speeds.
Methods: Participants were 34 community-dwelling older men and women (mean age 79.2) who were asked to simultaneously wear three activity monitors: the Yamax DigiWalker (DW) pedometer (hip), the Actigraph (AG) accelerometer (hip), and the StepWatch activity monitor (SAM) (ankle). Monitor accuracy was evaluated against observed steps taken during a 100-step walking test. Percent error of the monitors was calculated as [(monitor steps − observed steps)/observed steps] × 100. Participants were categorized into three groups (< 0.80, 0.80-1.0, > 1.0 m·s−1) according to gait speed, which was determined by a timed 4-m walk.
Results: Overall, the DW and AG failed to detect 16% and 7% of observed steps, respectively, and the SAM overestimated by 5.5%. When stratified by gait speed, all three monitors faired well at the gait speeds > 1.0 m·s−1. For gait speeds between 0.80 and 1.0 m·s−1, the SAM overestimated steps by 6.6%, and the AG and DW underestimated steps by 5.7% and 12.7%, respectively. However, at gait speeds < 0.80 m·s−1, the AG and DW performed poorly, underestimating steps by 19.1% and 31.2%, whereas the SAM performed better, having overestimated steps by 6.5%.
Conclusions: All three objective activity monitors performed well at moderate and higher walking speeds, but at decreased gait speeds, the SAM seemed to be the most accurate.
1Department of Epidemiology, University of Pittsburgh, Pittsburgh, PA; 2Department of Exercise and Wellness, Arizona State University, Mesa, AZ; 3Department of Physical Therapy, University of Pittsburgh, Pittsburgh, PA; and 4Department of Family Medicine, University of Michigan, and the Ann Arbor VA Medical Center, Ann Arbor, MI
Address for correspondence: Kristi L. Storti, Ph.D., MPH, University of Pittsburgh, Department of Epidemiology, 515 Parran Hall/GSPH, 130 DeSoto Street, Pittsburgh, PA; E-mail: email@example.com.
Submitted for publication March 2007.
Accepted for publication August 2007.