Related Experiment Video
Updated: Jun 2, 2026
07:24
A Method for Quantifying Upper Limb Performance in Daily Life Using Accelerometers
Published on: April 21, 2017
Identifying sedentary time using automated estimates of accelerometer wear time.
Elisabeth A H Winkler1, Paul A Gardiner, Bronwyn K Clark
1School of Population Health,The University of Queensland, Brisbane, Australia. e.winkler@uq.edu.au
British Journal of Sports Medicine
|April 21, 2011
Summary
Automated algorithms estimate accelerometer wear time accurately on average, but individual misclassification can be significant. Algorithm choice impacts sedentary time estimates, with limited movement allowance improving accuracy.
Area of Science:
- Physical Activity Measurement
- Biomedical Engineering
- Public Health Research
Background:
- Accurate assessment of physical activity and sedentary behavior is crucial for understanding health outcomes.
- Accelerometers are widely used to objectively measure physical activity, but accurate estimation of wear time is essential for valid results.
- Automated algorithms offer a potential solution for efficiently determining accelerometer wear time.
Purpose of the Study:
- To evaluate the accuracy of three automated accelerometer wear-time estimation algorithms.
- To assess the direct and indirect effects of these algorithms on sedentary time and moderate-to-vigorous physical activity (MVPA) estimates.
- To examine the implications of algorithm choice for population-level estimates of physical activity behaviors.
Main Methods:
- A subsample (n=148) from the Australian Diabetes, Obesity and Lifestyle Study wore accelerometers and completed activity logs.
- Three algorithms with varying allowances for movement during non-wear periods were compared.
- Population estimates were analyzed using US National Health and Nutrition Examination Survey data.
Main Results:
- While mean differences in wear time were small (≤11 min/day), 95% limits of agreement were wide (±≥2 h).
- Algorithms showed substantial misclassification of sedentary and non-wear time, with Algorithm 3 having the highest misclassification of non-wear as sedentary (51.3%).
- Algorithm 2 increased sedentary time estimates by 20 min/day compared to Algorithm 1, without affecting MVPA estimates.
Conclusions:
- Automated accelerometer wear-time estimation provides accurate average values but can lead to significant individual misclassification.
- The choice of algorithm significantly influences sedentary time estimates, highlighting the importance of algorithm selection.
- Allowing minimal movement during non-wear periods can enhance the accuracy of wear-time estimation and subsequent behavioral assessments.
