Skip to main content
. 2021 Sep 12;32(2):343–352. doi: 10.1007/s00787-021-01853-z

Table 2.

Association between time and both health-related quality of life and psychosocial problems tested in generalized estimating equations (GEE) models

Health-related quality of lifea Psychosocial problemsa,b
Coefficient 95% CI Coefficient 95% CI
Lower limit Upper limit Lower limit Upper limit
Timec
 1 Ref Ref Ref Ref Ref Ref
 2 1.68 0.04 3.31 − 0.05 − 1.21 1.12
 3 2.05 0.66 3.44 − 0.15 − 1.18 0.87
 4 3.04 1.24 4.84 − 1.33 − 2.50 − 0.17
 5 − 0.10 − 1.83 1.63 −  −  − 
Sex
 Boys Ref Ref Ref Ref Ref Ref
 Girls 0.11 − 1.54 1.76 − 0.75 − 1.89 0.39
Age − 0.28 − 2.08 1.53 − 0.06 − 1.17 1.06
SES 0.40 − 0.12 0.92 − 0.61 − 0.99 − 0.23
Ethnic background
 Dutch Ref Ref Ref Ref Ref Ref
 Western − 4.90 − 9.47 − 0.32 − 0.30 − 3.52 2.93
 Non− Western − 0.03 − 2.44 2.38 − 1.88 − 3.38 − 0.38
Grade
 5 Ref Ref Ref −  −  − 
 6 2.33 − 0.02 4.69
 7 4.88 0.84 8.93 Ref Ref Ref
 8 4.14 − 1.48 9.76 − 0.64 − 2.09 0.82

In these models, school was additionally included as a categorical covariate, but the school-specific estimates are not shown

SES = socio-economic status; CI = confidence interval

aHigher scores indicate higher quality of life and more psychosocial problems

bPsychosocial problems were only measured in grade 7 and 8 (age 9–13 years) and it was not possible to include T5 in the GEE models (comparing T5 with T1 separately: B = 1.18, 95% CI [− 0.20, 2.56])

cCorresponding to T1–T5, with T1 containing the measurement occasions at both 0 and 4 months