Research-Grade Accelerometers to Assess Intervention Effectiveness and Commercial Wearables as a Self-Monitoring Tool: Are the Two Aligned on Moderate to Vigorous Physical Activity?

Abstract

There is limited evidence about how comparable commercial monitors and moderate to vigorous physical activity (MVPA) cut-points frequently used with research-grade accelerometers are in estimating time spent in MVPA when both are worn on the wrist. We used data from 14 adults (57.3 years) who concurrently wore Fitbit (Versa 2) and Axivity (AX3) monitors to assess comparability between the Fitbit and Axivity data, aggregated to 5-s and 60-s epochs, for time spent in daily MVPA and MVPA in bouts of ≥1 – <2 min, ≥2 – <5 min, and ≥5 min. The Fitbit provided higher estimates of total daily MVPA (93.8 ± 66.3 min/day) compared with Axivity-assessed MVPA derived from either 5-s (59.2 ± 37.3 min/day) or 60-s (47.4 ± 36.2 min/day) epochs. The number and duration of MVPA bouts were also higher from Fitbit compared with Axivity (5-s and 60-s epochs). Intraclass correlation coefficients showed moderate to good reliability between MVPA assessed by the Fitbit and Axivity devices (5 s and 60 s). Bland–Altman analysis suggested proportional bias between the Fitbit and Axivity outcomes with wide limits of agreement. No MVPA outcomes were within the 10% equivalency zone. Researchers utilizing the Fitbit as a self-monitoring tool within an intervention should be aware of the potential misalignment with the research-grade accelerometer used to evaluate intervention effectiveness. Feedback from a Fitbit may inform participants they have achieved the required MVPA, while outcomes from the research-grade device may classify them as not meeting their target.