Variability and reliability study of overall physical activity and activity intensity levels using...

Area of Science:

• Biomedical Engineering
• Epidemiology
• Public Health

Background:

• Objective physical activity (PA) assessment using 24-hour accelerometry is crucial for observational studies.
• Key parameters like PA variability, required observation duration, and reliability remain unclear.

Purpose of the Study:

• To assess the variability of 24-hour physical activity (PA) using accelerometry.
• To determine the optimal observational time needed for reliable estimation of habitual PA.
• To evaluate the reliability of PA measurements across different days and days of the week.

Main Methods:

• Utilized triaxial accelerometers (ActiGraph GT3X+) to measure 24-hour PA in 50 participants over 2 weeks.
• Employed linear mixed-effects and random effects models to quantify PA variability.
• Calculated the necessary number of assessment days and PA reliability using intraclass correlation coefficients.

Main Results:

• Between- and within-person variance explained 34.4-45.5% and 54.5-65.6% of total PA variance, respectively.
• PA levels showed slight decreases in the initial 3 days and minor differences across days of the week (highest on Wednesday/Friday, lowest on Sunday/Monday).
• An average of 1 week of assessment was required for habitual PA estimation, with good week-to-week reliability (ICC range: 0.68-0.82).

Conclusions:

• Individual physical activity (PA) assessed via 24-hour accelerometry exhibits significant daily variability.
• The day of assessment or day of the week minimally impacts overall PA variance.
• A minimum of 1 week of accelerometry data is recommended for reliable estimation of habitual PA.