. 2019 Apr 19;28(12):1659–1669. doi: 10.1007/s00787-019-01331-7

Table 2.

Probability models for being overweight in youth within the NeuroIMAGE cohort derived from generalized estimating equations

	OR (95% CI)
Default youth model
Intercept	0.50 (0.43; 0.59)^a
ADHD diagnosis	0.93 (0.79; 1.10)
Male gender	0.73 (0.61; 0.86)^a
Age (centered)	1.00 (1.00; 1.03)
Age (centered)²	1.00 (1.00; 1.01)
IQ (centered)	0.99 (0.98; 1.01)
N assessments		1749
Extended youth model
ADHD diagnosis × male gender	–
ADHD diagnosis × age (centered)	1.04 (1.01; 1.07)^a
Male gender × age (centered)	–
N assessments		1749
Parental youth model
Maternal age (centered)	–
Paternal age (centered)	–
Maternal education (centered)	1.05 (1.00; 1.10)^a
Paternal education (centered)	–
Maternal ADHD score above cut-off	–
Paternal ADHD score above cut-off	–
Maternal overweight	1.40 (1.14; 1.73)^a
Paternal overweight	1.83 (1.41; 2.36)^a
N assessments		828
Affected youth medication model
Psychostimulant use	0.98 (0.76; 1.26)
Antipsychotic use	0.81 (0.49; 1.34)
Melatonin use	0.74 (0.50; 1.07)
Psychostimulant use × male gender	–
Psychostimulant use × age (centered)	–
Antipsychotic use × male gender	–
Antipsychotic use × age (centered)	–
Melatonin use × male gender	–
Melatonin use × age (centered)	–
N assessments		641

^aSignificant after FDR correction