Comparing ActiGraph equations for estimating energy expenditure in older adults

Area of Science:

• Gerontology
• Biomedical Engineering
• Physical Activity Measurement

Background:

• Accurate estimation of energy expenditure (EE) in older adults using accelerometers presents a significant challenge.
• Existing ActiGraph (AG) equations require validation for this specific demographic.

Purpose of the Study:

• To validate five different ActiGraph (AG) equations for predicting EE in older adults.
• To compare the accuracy and precision of these equations across various physical activity intensities.

Main Methods:

• Forty older adults (mean age 77.4 years) performed household and gardening activities.
• Energy expenditure was measured using a portable calorimeter (MetaMax 3B) as the criterion measure.
• ActiGraph GT3X+ accelerometers were worn at the hip to predict EE using Freedson, refined Crouter, Sasaki, and Santos-Lozano equations.

Main Results:

• The Sasaki equation demonstrated the lowest root-mean-square error (RMSE) across all activities and physical activity intensities (PAIs).
• Freedson and Santos-Lozano equations overestimated EE during sedentary activities and underestimated it during moderate-to-vigorous physical activity (MVPA).
• Refined Crouter and Sasaki equations exhibited minimal systematic bias, though they showed slight over/underestimation across PAIs.

Conclusions:

• No single ActiGraph equation provided complete accuracy for predicting EE across all PAIs in older adults.
• The refined Crouter and Sasaki equations offered superior overall accuracy and precision compared to the other evaluated methods.