This study examined the predictive validity of accelerometers (ACC) to estimate physical activity intensity (PAI) across age and differences in intensity predictions when expressed in relative and absolute PAI terms. Ninety adults categorized into 3 age groups (20–29, 40–49, and 60–69 yr) completed a treadmill calibration study with simultaneous ACC (7164 Actigraph) and oxygen-consumption assessment. Results revealed strong linear relations between ACC output and measured PAI (R2 = .62–.89) across age and similar ACC cut-point ranges across age delineating absolute PAI ranges compared with previous findings. Comparing measured metabolic equivalents (METs) with estimated METs derived from previously published regression equations revealed that age did not affect predictive validity of ACC estimates of absolute PAI. Comparing ACC output expressed in relative vs. absolute terms across age revealed substantial differences in PAI ACC count ranges. Further work is warranted to increase the applicability of ACC use relative to PAI differences associated with physiological changes with age.
The authors are with the Dept. of Human Movement Sciences, University of Wisconsin–Milwaukee.