Kozey · Medicine and science in sports and exercise 2010 · cross-sectional calibration study · n=277

Accelerometer output and MET values of common physical activities.

Cited 177 times in the scientific literature.

Level 4 - case-series / case-control

Cross-sectional physiological measurement and device calibration study

PubMed 20142781 · doi:10.1249/MSS.0b013e3181d479f2 · record verified 2026-08-29

What was done

A total of 277 participants performed 11 physical activities for 7 minutes each from a menu of 23 activities while wearing an accelerometer and a portable metabolic system measuring oxygen consumption (VO2). Energy expenditure was determined as measured METs (VO2 / measured resting metabolic rate) and standard METs (VO2 / 3.5 mL·kg⁻¹·min⁻¹). MET values were compared with the Compendium of Physical Activities across the total cohort and by subgroups (age < 40 vs. ≥ 40 years, sex, and BMI < 25 vs. ≥ 25 kg·m⁻²). Linear association between accelerometer counts per minute and METs was evaluated.

What was found

Compendium MET values differed significantly from measured METs for 17 of 21 evaluated activities (81%), with similar discrepancies across subgroups and standard MET definitions. Mean activity counts ranged from 11 counts per minute (dishes) to 7,490 counts per minute (treadmill at 2.23 m·s⁻¹, 3% grade). The linear correlation (r) between accelerometer counts per minute and METs was 0.65.

Why it matters

Standard Compendium MET values frequently misestimate actual energy expenditure during common activities, and linear models based on accelerometer counts are insufficient for accurately predicting METs.

Limits

Activities were evaluated in brief 7-minute structured bouts rather than free-living conditions, and each participant completed only a subset (11 of 23) of activities. The abstract does not provide exact MET numerical values or confidence interval bounds for individual activities.

Cited by