Validation of the StepWatch device for measurement of free-living ambulatory activity in patients with chronic obstructive pulmonary disease.

This study examined the validity of the StepWatch activity monitor and the optimal number of monitoring days in patients with chronic obstructive pulmonary disease (COPD). StepWatch activity data were obtained from databases of two studies: one of patients with COPD who participated in a trial of two exercise persistence interventions (n = 17) and a second, cross-sectional study of activity in chronically healthy older adults (n = 60). Independent t tests, Pearson's correlations, and Bland-Altman plots were used to examine the activity data (total steps, activity intensity, and distribution of activity). Results indicated that patients with COPD registered 50% fewer total steps and had lower levels of walking intensity compared with chronically healthy older adults. Activity parameters had moderate-to-high correlations with disease severity, peak cycle workload, and physical functioning. More monitoring days were associated with greater measurement precision. This study provides preliminary evidence to support the validity of the StepWatch device for measuring free-living ambulatory physical activity in COPD. Keywords: physical activity; COPD; measurement; validation Epidemiological studies based on self-reported physical activity show that higher levels of activity are associated with lower risk of incident chronic obstructive pulmonary disease (COPD) in smokers and in patients with COPD, decreased risk of hospital admissions, exacerbations, and mortality (Garcia-Aymerich et al., 2003; Garcia- Aymerich, Lange, Benet, Schnohr, & Anto, 2006, 2007). Given the centrality of physical activity in health and illness and the increased availability of affordable activity monitoring devices that can precisely measure what patients with COPD actually do in their daily life instead of what they report or what they are capable of with laboratory testing, the importance of validating these various devices for research and clinical purposes assumes high priority (Walker, Burnett, Flavahan, & Calverley, 2008). Several accelerometer-based devices including the SenseWear Pro armband (BodyMedia, Inc., Pittsburgh, PA; Watz et al., 2008; Watz, Waschki, Meyer, & Magnussen, 2009), Actiwatch (Cambridge Neurotechnology Ltd., UK; Walker et al., 2008), DynaPort activity monitor (McRoberts B. V., The Hague, Netherlands; Pitta et al., 2006; Walker et al., 2008), and RT3 (Stayhealthy, Monrovia, CA; Steele et al., 2008) have recently been used to measure various dimensions of physical activity in patients with COPD. The selection of any device is driven by several factors including costs, evidence of reliability and validity, patient comfort, and study goals. The StepWatch device (Orthocare Innovations, Mountlake Terrace, WA) is another accelerometer-based activity monitor, originally designed to measure free-living ambulatory activity in individuals wearing prosthetics (Coleman, Smith, Boone, Joseph, & del Aguila, 1999). Although several published studies have confirmed the high accuracy of the StepWatch in capturing step counts (98%-99%) in other populations (Cavanaugh, Coleman, Gaines, Laing, & Morey, 2007; Gardner, Montgomery, Scott, Afaq, & Blevins, 2007; Resnick, Nahm, Orwig, Zimmerman, & Magaziner, 2001), there are no published data in patients with COPD. Because several ongoing COPD studies (B. Steele, personal communications, September 1, 2009; M. Moy, personal communications, March 1, 2009) are using the StepWatch to measure physical activity as a primary outcome and use of the device will only increase because of its purported accuracy in capturing ambulatory activity, there is a need to test the validity of the StepWatch for use in this clinical population. Aside from examining device validity, there is also a practical, unanswered question of the optimal monitoring duration. Although previous studies reported only modest dayto- day variability in physical activity levels, these studies had relatively short monitoring durations of between 2 and 7 days (Pitta, Troosters, et al. …