Moving Forward with Backward Compatibility: Translating Wrist Accelerometer Data.

Area of Science:

• Pediatric physical activity monitoring
• Biomedical engineering
• Exercise science

Background:

• Accurate measurement of children's physical activity is crucial for public health.
• Wrist-worn accelerometers offer a convenient alternative to hip-worn devices.
• Standardizing data across different accelerometer types and placements is needed.

Purpose of the Study:

• To establish calibration methods for raw wrist accelerometer data.
• To align wrist-based measurements with established hip-worn ActiGraph moderate-to-vigorous physical activity (MVPA) estimates.
• To facilitate direct comparison of physical activity data across different measurement devices.

Main Methods:

• Secondary analysis of data from 238 children (aged 9-12 years) with concurrent wrist (GENEActiv) and hip (ActiGraph) accelerometer wear.
• Calculation of time above various normalized energy expenditure (ENMO) thresholds from wrist data.
• Comparison of wrist-derived ENMO thresholds with established hip-worn ActiGraph MVPA cut-points (Freedson/Trost, Pate, Evenson, Puyau) using different epoch lengths.

Main Results:

• Optimal ENMO thresholds were identified for aligning wrist data with hip-worn ActiGraph MVPA estimates across different cut-points and epoch lengths.
• For example, Freedson/Trost cut-points aligned with ENMO 150+ mg (ICC ≥ 0.65), while Puyau cut-points required higher thresholds (e.g., ENMO 400+ mg for 60-s epochs, ICC = 0.65).
• Cross-validation demonstrated robust agreement (ICC = 0.62-0.71) with minimal differences in MVPA estimates.

Conclusions:

• Incremental ENMO thresholds provide a simple and direct method for calibrating wrist accelerometer data.
• This calibration enables group-level comparison of children's physical activity data with historical estimates from hip-worn ActiGraphs.
• The findings support the use of wrist-worn accelerometers for consistent and comparable physical activity assessment in children.