Skip to main content

Table 3 Effect of activity monitor removals during sleep on the prediction of physical activity and total energy expenditure.

From: Estimates of adherence and error analysis of physical activity data collected via accelerometry in a large study of free-living adults

 

Physical Activity

 

counts·min-1·day-1

log (counts·min-1·day-1)

Simulation

Mean (SD)

CV (%)

Mean (SD)

CV (%)

24 HR

228.4 (97.2)

-

5.35 (0.40)

-

A1

220.5 (97.6)*

3.1

5.30 (0.45)*

0.7

B1

332.9 (151.3)*

26.0

5.71 (0.43)*

4.8

C1

228.2 (98.8)

1.6

5.35 (0.42)

0.3

 

Total Energy Expenditure

 

MJ·day-1

log (MJ·day-1)

Simulation

Mean (SD)

CV (%)

Mean (SD)

CV (%)

24 HR

11.2 (3.3)

-

2.38 (0.29)

-

A1

11.2 (2.8)

9.0

2.37 (0.24)

4.1

B1

11.8 (4.2)

9.9

2.41 (0.26)

4.0

C1

11.2 (2.9)

9.0

2.38 (0.24)

4.0

  1. *p < 0.05 vs. 24 HR
  2. 24 HR: physical activity data from subset of highly adherent subjects (n = 35). Simulation A1: simulation of activity monitor removal during sleep by imputing zeroes. Simulation B1: simulation of compensating for activity monitor removal during sleep by measuring waking hours only (sleeping physical activity deleted). Simulation C1: simulation of compensating for activity monitor removal during sleep by imputing a constant (23.1 counts/min) CV(%)= coefficient of variation = ((standard deviation of 24 HR – Simulation A1/B1/C1)/(mean of 24 HR and Simulation A1/B1/C1)) × 100.