JMIR Publications

ABSTRACT

Background:

With the continuous evolution of technology, wearable accelerometers have become one of the most popular means of measuring daily physical activity (PA) levels. Despite the conventional use of the non-dominant wrist as a device placement in numerous PA studies, the impact of wrist-worn accelerometer placement on PA data outcomes remains uncertain.

Objective:

The objective of this study was to examine the degree of agreement between accelerometry data collected from ActiGraph CentrePoint Insight Watches (CPIW) worn on the dominant and non-dominant wrists of young adults in free-living conditions.

Methods:

Twenty-nine participants (Mage 20.2±1.6; 23 females) simultaneously wore an ActiGraph CPIW on both dominant and non-dominant wrists for 7 consecutive days during waking hours. A sampling frequency of 32Hz and Montoye 2020 cut-points were used to categorize the activity intensity based on vector magnitude counts per minutes. Data validity criteria included: (1) ≥600 min/day of monitor wear time, (2) a wear time difference between the dominant and non-dominant wrists of less than 1% of the average wear time, and (3) at least three valid days of monitor wear. Bland Altman plots and intraclass correlation coefficient (ICC) analyses were performed to compare the accelerometry data between the two device placements.

Results:

Average daily monitor wear time was 789.6±86.1min/day for dominant and 793.0±91.8 min/day for non-dominant wrists, respectively. All accelerometer variables, including sedentary time, light PA, moderate-to-vigorous PA (MVPA), steps, triaxial counts, and vector magnitude, showed good-to-excellent levels of reliability between the two measurements (ICC > 0.88 for all, p < .001). Bland Altman analysis calculated mean bias and standard deviation between the two device placements as follows: sedentary time (-18.8±27.6 min/day), light PA (2.7±15.9 min/day), MVPA (12.7±26.7 min/day), steps (218.1±476.6 counts/day), X axis (99.4±188.8 counts/min), Y axis (73.9±147.0 counts/min), Z axis (107.6±183.5 counts/min), and vector magnitude (161.2±273.4 counts/min). Bland Altman plots revealed that the upper and lower limits of agreement across most variables were considerably wide and that moderate proportional bias was present for steps (R2 = 0.091) and light PA (R2 = 0.105).

Conclusions:

Our findings align partially with previous research, demonstrating higher MVPA and step counts on the dominant wrist, while the non-dominant wrist produced a higher level of sedentary time compared to the dominant wrist. Despite the acceptable level of reliability between the two placements based on ICC analyses, the dominant wrist tended to produce greater outcomes as the intensity of PA increased compared to the non-dominant wrist, highlighting the need for careful consideration when determining the wear location of CPIW and interpreting data outcomes. Clinical Trial: N/A