Step‐counting accuracy of activity monitors in persons with Down syndrome. Issue 1 (21st September 2018)

• Record Type: Journal Article
• Authors: Bertapelli, F.; Curtis, J. S.; Carlson, B.; Johnson, M.; Abadie, B.; Agiovlasitis, S.
• Is Part Of: Journal of intellectual disability research. Volume 63:Issue 1(2019)
• Page Start: 21
• Page End: 30
• Publication Date: 2018-09-21
• Subjects: activity monitor -- disability -- measurement -- physical activity -- walking; Mental retardation -- Research -- Periodicals; 362.2
• DOI: 10.1111/jir.12550
• Languages: English
• ISSNs: 0964-2633

Abstract

Abstract: Background: Accelerometers and pedometers have been used to monitor the number of steps. However, the evidence on the step‐counting accuracy of these devices – especially accelerometers – is limited in persons with Down syndrome (DS). This study therefore examined the accuracy of accelerometers placed on the hip or wrist and of a pedometer with a uni‐axial accelerometer mechanism in measuring steps in persons with DS and whether device error is associated with walking speed, height, weight, body mass index, waist circumference, leg length, age or sex. Method: Seventeen persons with DS (eight women and nine men; age 33 ± 15 years) walked over‐ground for 6 min at their preferred speed. The steps were measured with a hip‐worn and a wrist‐worn ActiGraph accelerometer using the manufacturer's default (DF) and low‐frequency extension (LFE) filters, and with the NL‐1000 New Lifestyles pedometer on the hip. Steps were also measured with hand tally which served as the criterion. Results: Absolute percent error was considerable and differed statistically between devices ( P = 0.001); however, error improved for accelerometers when LFE was applied (Hip‐DF: 31.6 ± 18.8%; Hip‐LFE: 9.7 ± 12.8%; Wrist‐DF: 32.7 ± 14.2%; Wrist‐LFE: 13.6 ± 10.2%; Pedometer: 23.2 ± 22.8%). Bland–Altman plots indicated underestimation of steps for accelerometers and the pedometer. Application of LFE, however, improved the prediction of the accelerometers. The number of steps measured by the hip accelerometer with LFE and by the pedometer did not differ statistically from actual steps. Steps by the remaining methods were significantly lower than hand tally ( P ≤ 0.001). Correlations between percent error for each device and walking speed, anthropometry, age or sex ranged between −0.28 and +0.48, and were non‐significant, except for age. Conclusions: The results demonstrated that the pedometer and ActiGraph accelerometers have considerable error in measuring steps of persons with DS. Application of LFE, however, significantly improved the step‐counting performance of the Actigraph accelerometers.