Improving the criterion validity of the activPAL in determining physical activity intensity during laboratory and free-living conditions.
Level 3 - non-randomized controlled study
Non-randomized cross-sectional validation study comparing an assessment method against a reference standard.
PubMed 33203323 · doi:10.1080/02640414.2020.1847503
What was done
Forty healthy adults (mean age 25 ± 6 years, mean BMI 23.3 ± 4.1 kg/m²) completed a 7-stage progressive treadmill walking protocol (1.5 to 4.5 mph) wearing an activPAL monitor, with indirect calorimetry serving as the criterion measure. A sub-sample of 32 participants wore the device during free-living conditions for 7 days, using a PiezoRxD monitor as the comparison criterion. The study evaluated the criterion validity of the standard built-in linear cadence–metabolic equivalent (MET) equation against an individualized, height-adjusted curvilinear cadence–MET equation.
What was found
In laboratory testing, the built-in linear equation overestimated METs at slow speeds (1.5–3.0 mph) and underestimated METs at faster speeds (3.5–4.5 mph) (all p < 0.001). In free-living conditions, the linear equation overestimated time spent in light-intensity activity, underestimated moderate-intensity activity (both p < 0.001), and detected no vigorous-intensity activity. The height-adjusted curvilinear equation yielded estimates statistically equivalent to indirect calorimetry across treadmill stages 1 through 6 (1.5–4.0 mph) and equivalent to PiezoRxD classifications for light- and moderate-intensity activity in free-living conditions. Absolute numerical values and error margins were not reported in the abstract.
Why it matters
Standard activPAL processing systematically misclassifies physical activity intensity. Applying a height-adjusted curvilinear equation enables researchers to accurately estimate light- and moderate-intensity activity from existing activPAL cadence data without supplementary hardware.
Limits
The sample was small (n = 40 total, n = 32 free-living) and limited to young, lean adults, restricting generalizability to older, pediatric, or clinical cohorts. The free-living reference standard was another wearable accelerometer (PiezoRxD) rather than direct observation or physiological energy expenditure measurement. Accuracy was not demonstrated for the fastest treadmill speed (4.5 mph) or for vigorous-intensity free-living activity.
Cited by
- supports Standard metabolic equivalent of task (MET) categories classify light activity as 0 to 3 METs, moderate activity as 3 to 6 METs, and vigorous activity as greater than 6 METs.