Related Experiment Videos
Comparison of MTI accelerometer cut-points for predicting time spent in physical activity
S J Strath1, D R Bassett, A M Swartz
1Department of Health and Exercise Science, The University of Tennessee, Knoxville, Tennessee 37996-2700, USA.
International Journal of Sports Medicine
|June 5, 2003
Summary
Different accelerometer cut-points significantly impact physical activity intensity estimates. The Swartz method showed accuracy on a group level but large individual errors, highlighting limitations in current accelerometer use for activity pattern analysis.
Area of Science:
- Exercise Physiology
- Biomedical Engineering
- Physical Activity Measurement
Background:
- Accelerometers are widely used to estimate physical activity intensity during free-living conditions.
- Existing regression equations and cut-points vary, potentially leading to inconsistent data interpretation.
- Accurate assessment of physical activity is crucial for public health and exercise science research.
Purpose of the Study:
- To evaluate the accuracy of five published accelerometer regression equations for predicting time spent in different physical activity intensities.
- To compare the performance of various cut-points (e.g., Freedson, Hendelman, Swartz, Nichols) in classifying activity levels.
- To identify limitations of hip-mounted accelerometers in reflecting accurate physical activity patterns.
Main Methods:
- Ten participants engaged in free-living activities for 5-6 hours.
- Oxygen uptake and accelerometer data were collected concurrently.
- Time spent in resting/light, moderate, and hard activities was calculated using 3 and 6 MET cut-points from five regression equations.
Main Results:
- The Swartz cut-points showed no significant difference from criterion measures for overall activity intensity.
- Freedson, Hendelman, and Nichols cut-points demonstrated significant over- or under-estimation of time spent in various activity intensities.
- Substantial individual-level errors were observed across all tested regression formulas, despite group-level accuracy.
Conclusions:
- Different accelerometer cut-points yield substantially different estimates of time spent in physical activity intensities.
- Researchers must be aware of the limitations and potential inaccuracies when using hip-mounted accelerometers for physical activity pattern analysis.
- The choice of regression equation and cut-points significantly influences the interpretation of accelerometer-derived physical activity data.