Skip to main content
. 2017 Apr 3;5(2):e6. doi: 10.2196/games.6421

Table 3.

Multivariable association between baseline characteristics and program usage and physical activity (n=51): separate models using linear mixed models.

Characteristics of program usage: models Moderate to vigorous physical activitya
Sources of variance, SDb (95% CI) Β (95% CI)
Model: description Independent
variables
(additional)
Categories Between eHealth teams Between participants
Model 1: raw

N/Ac 0.98 (0.80 to 1.19) N/A
Model 2: baseline characteristics

N/A 0.99 (0.81 to 1.22)

Age


0.01 (−0.03 to 0.04)

Sex Women

Reference


Men

0.43 (−0.33 to 1.19)

BMId at baseline


−0.02 (−0.08 to 0.04)
Model 3: model 2
+ eHealth team


0.88 (0.51 to 1.55) 0.54 (0.43 to 0.67)
Model 4: model 3
+ compliance


0.86 (0.49 to 1.52) 0.55 (0.44 to 0.69)

Compliancee <85%

Reference


85%-95%

0.15 (−0.34 to 0.65)


≥95%

0.17 (−0.31 to 0.64)
Model 5: model 3
+ engagement


0.83 (0.46 to 1.50) 0.54 (0.43 to 0.68)

Engagementf ≤3

Reference


4 or 5

0.19 (−0.25 to 0.64)
Model 6: model 3
+ compliance
+ engagement


0.82 (0.45 to 1.48) 0.56 (0.44 to 0.70)

Compliance <85%

Reference


85%-95%

0.17 (−0.34 to 0.67)


≥95%

0.06 (−0.50 to 0.62)

Engagement ≤3

Reference


4 or 5

0.17 (−0.32 to 0.66)

aLog-transformed average moderate to vigorous physical activity (metabolic equivalent task or MET>3), in MET hours.

bSD is the standard deviation of the random effect between teams or between participants.

cN/A: not applicable.

dBMI: body mass index.

eCompliance is expressed as percentage of days with >10 hours of physical activity registration (accelerometer wear).

fEngagement is expressed as the number of times at least 100% of the target was reached (1-5).