Estimating Energy Expenditure using Individualized, Power-Specific Gross Efficiencies

Abstract

Our purpose was to determine if using an individual's power-specific gross efficiency improves the accuracy of estimating energy expenditure from cycling power. 30 subjects performed a graded cycling test to develop 4 gross efficiencies: individual power-specific gross efficiencies, a group mean power-specific gross efficiency, individual fixed gross efficiencies, and a group mean fixed gross efficiency. Energy expenditure was estimated from power using these different gross efficiencies and compared to measured energy expenditure during moderate- and hard-intensity constant-power and 2 variable-power cycling bouts. Estimated energy expenditures using individual or group mean power-specific gross efficiencies were not different from measured energy expenditure across all cycling bouts (p>0.05). To examine the intra-individual variability of the estimates, absolute difference scores (absolute value of estimated minus measured energy expenditure) were compared, where values closer to zero represent more accurate individual estimates. The absolute difference score using individual power-specific gross efficiencies was significantly lower compared to the other gross efficiencies across all cycling bouts (p<0.01). Significant and strong correlations (r≥0.97, p<0.001) were found across all cycling bouts between estimated and measured energy expenditures using individual power-specific gross efficiencies. In conclusion, using an individual's power-specific gross efficiency significantly improves their energy expenditure estimate across different power outputs.

Determination of energy expenditure in professional cyclists using power data: Validation against doubly labeled water

“…Importantly, the developed equation can be used with a measured BMR and measured power output-EE relation to enhance the accuracy, but can also be used with an estimated BMR based on the proposed prediction equation (Equation 1 or 2), and using an estimated EE based on the average relation between power and EE (Equation 4). However, such a generic approach will reduce the accuracy (see e.g., 41 ). Indeed, while EE E was underestimated by only 3% at the group level when we assumed a gross efficiency of 20% to convert power output into EE (which is approximately similar to using our Equation 4), this error ranged from 17% underestimation to 8% overestimation at an individual level, thus highlighting the need for an individualized approach.…”