Does a bio-impedance-derived estimate better agree with measured resting energy expenditure than the harris-benedict equation in older adults?