Identifying sedentary time using automated estimates of accelerometer wear time

School of Population Health,The University of Queensland, Brisbane, Australia.
British Journal of Sports Medicine (Impact Factor: 5.03). 04/2011; 46(6):436-42. DOI: 10.1136/bjsm.2010.079699
Source: PubMed


The authors evaluated the accuracy of three automated accelerometer wear-time estimation algorithms against self-report. Direct effects on sedentary time (<100 cpm) and indirect effects on moderate-to-vigorous physical activity (MVPA, ≥1952 cpm) time were examined.
A subsample from the 2004/2005 Australian Diabetes, Obesity and Lifestyle Study (n=148) completed activity logs and wore accelerometers for a total of 987 days. A published algorithm that allows movement within non-wear periods (Algorithm 1) was compared with one that allows less movement (Algorithm 2) or no movement (Algorithm 3). Implications for population estimates were examined using 2003/2004 US National Health and Nutrition Examination Survey data.
Mean difference per day between the criterion and estimated wear time was negligible for all three algorithms (≤11 min), but 95% limits of agreement (LOA) were wide (±≥2 h). Respectively, the algorithms (1, 2 and 3) misclassified sedentary time as non-wear on 31.9%, 19.4% and 18% of days and misclassified non-wear time as sedentary on 42.8%, 43.7% and 51.3% of days. Use of Algorithm 2 (compared with Algorithm 1) affected population estimates of sedentary time (higher by 20 min/day) but not MVPA time. Agreement between Algorithms 1 and 2 was good for MVPA time (mean difference -0.08, LOA: -2.08, 1.91 min), but not for wear time or sedentary time.
Accelerometer wear time can be estimated accurately on average; however, misclassification can be substantial for individuals. Algorithm choice affects estimates of sedentary time. Allowing very limited movement within non-wear periods can improve accuracy.

Download full-text


Available from: Genevieve N Healy, Jul 21, 2014
  • Source
    • "For example, consistent with the anatomical location of the anterior iliac spine, with the accelerometer is placed in a vertical position and the accelerometer worn on the belt or waistband of the clothing. Alternately, subjects were provided with a belt to secure the accelerometer to the proper location at the waist in the event that it could not be attached properly to their clothing,20 which seemed to be the standard wearing site. But, a triaxial accelerometer can be worn on various sites, such as wrist, ankle, upper arm, and waist using a band or other accessories. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Purpose We performed a study to determine the best appropriate wearing site of a triaxial accelerometer at different exercise speeds. Materials and Methods We conducted an observational study with 66 healthy Korean adults (26 men and 40 women). Resting metabolic rate (RMR) before exercise, physical activity-related energy expenditure (PAEE) by cardiorespiratory gas analyzer and Signal Vector Magnitude (SVM) were measured while wearing four triaxial accelerometers on four different sites (wrist, waist, upper arm, and ankle) at exercise speeds from 2-10 km/h. Results The mean RMR was 4.03 mL/kg/min and Actual METs (oxygen consumption at different exercise speeds divided by individual RMR) compared with the calculated METs (oxygen consumption divided by 3.5 mL/kg/min) showed relatively low value. The overall correlation between PAEE and SVM was highest when the accelerometer was worn on the wrist at low exercise speed (r=0.751, p<0.001), waist at a moderate speed (r=0.821, p<0.001), and ankle at a high speed (r=0.559, p<0.001). Using regression analysis, it was shown that the ankle at a low speed (R2=0.564, p<0.001), high speed (R2=0.559, p<0.001), and the waist at a moderate speed (R2=0.821, p<0.001) were the best appropriate sites. Conclusion When measuring the PAEE and SVM at different exercise speeds, the ankle in low and high exercise speed, and waist in moderate speed are the most appropriate sites for an accelerometer.
    Yonsei Medical Journal 07/2014; 55(4):1145-51. DOI:10.3349/ymj.2014.55.4.1145 · 1.29 Impact Factor
  • Source
    • "Another issue that can affect the accuracy of reported sedentary time is the ability to differentiate between non-wear time and sedentary time [94]. This is of particular concern for older adults’ accelerometry data because the large amount of time they spend in sedentary behaviors can potentially lead to the misclassification of sedentary time as non-wear time [94]. Of the included publications, 50 reported some assumptions for their data cleaning procedure to identify spurious data or non-wear time (Table 1). "
    [Show abstract] [Hide abstract]
    ABSTRACT: Accelerometers objectively monitor physical activity and sedentary patterns and are increasingly used in the research setting. It is important to maintain consistency in data analysis and reporting, therefore, we: (1) systematically identified studies using accelerometry (ActiGraph, Pensacola, FL, USA) to measure moderate-to-vigorous physical activity (MVPA) and sedentary time in older adults, and (2) based on the review findings, we used different cut-points obtained to analyze accelerometry data from a sample of community-dwelling older women. We identified 59 articles with cut-points ranging between 574 and 3,250 counts/min for MVPA and 50 and 500 counts/min for sedentary time. Using these cut-points and data from women (mean age, 70 years), the median MVPA minutes per day ranged between 4 and 80 min while percentage of sedentary time per day ranged between 62 % and 86 %. These data highlight (1) the importance of reporting detailed information on the analysis assumptions and (2) that results can differ greatly depending on analysis parameters.
    European Review of Aging and Physical Activity 04/2014; 11(1):35-49. DOI:10.1007/s11556-013-0132-x · 0.68 Impact Factor
  • Source
    • "Analyses will be limited to days with ≥10 hours of wear and no excessive counts ≥20, 000 cpm (valid days). Bouts of ≥60 minutes of 0 cpm (allowing for < 3 minutes of counts 1–49 cpm) will be excluded as non-wear time [55]; self-reported sleep time will also be excluded for those wearing the monitor to bed. Outcomes used from the accelerometer data will include: average minutes per valid day of moderate-vigorous activity; and, average number of bouts per valid day of moderate-vigorous activity accumulated in ten minutes bouts (allowing for two one-minute epochs below this threshold). "
    [Show abstract] [Hide abstract]
    ABSTRACT: Behavioural lifestyle interventions can be effective at promoting initial weight loss and supporting physical activity and dietary behaviour change, however maintaining improvements in these outcomes is often more difficult to achieve. Extending intervention contact to reinforce learnt behavioural skills has been shown to improve maintenance of behaviour change and weight loss. This trial aims to evaluate the feasibility, acceptability and efficacy of a text message-delivered extended contact intervention to enhance or maintain change in physical activity, dietary behaviour and weight loss among participants who have completed a six month Government-funded, population-based telephone coaching lifestyle program: the Get Healthy Information and Coaching Service (GHS). GHS completers will be randomised to the 6-month extended contact intervention (Get Healthy, Stay Healthy, GHSH) or a no contact control group (standard practice following GHS completion). GHSH participants determine the timing and frequency of the text messages (3-13 per fortnight) and content is tailored to their behavioural and weight goals and support preferences. Two telephone tailoring calls are made (baseline, 12-weeks) to facilitate message tailoring. Primary outcomes, anthropometric (body weight and waist circumference via self-report) and behavioural (moderate-vigorous physical activity via self-report and accelerometer, fruit and vegetable intake via self-report), will be assessed at baseline (at GHS completion), 6-months (end of extended contact intervention) and 12-months (6-months post intervention contact). Secondary aims include evaluation of: the feasibility of program delivery; the acceptability for participants; theoretically-guided, potential mediators and moderators of behaviour change; dose-responsiveness; and, costs of program delivery. Findings from this trial will inform the delivery of the GHS in relation to the maintenance of behaviour change and weight loss, and will contribute to the broader science of text message lifestyle interventions delivered in population health settings.Trial registration: ACTRN12613000949785.
    BMC Public Health 02/2014; 14(1):112. DOI:10.1186/1471-2458-14-112 · 2.26 Impact Factor
Show more