Longitudinal changes of mental health problems in children and adolescents treated in a primary care-based health-coaching programme – results of the PrimA-QuO cohort study

Table 3 Longitudinal modelling on SDQ total scores

	Model 1: Unadjusted			Model 2: Adjusted for age and sex			Model 3: Fully adjusted
Variables	ß-Coefficient	Std Error	p-Value	ß-Coefficient	Std Error	p-Value	ß-Coefficient	Std Error	p-Value
Intercept	8.961	0.290	<0 .001	7.905	0.597	< 0.001	11.376	2.345	< 0.001
Intervention (Ref = control)	0.484	0.478	0.312	0.501	0.464	0.283	-0.237	0.501	0.636
Time (Ref = baseline)	-0.414	0.172	0.017	-0.414	0.175	0.017	-0.814	0.242	0.001
Interaction of intervention and time							0.802	0.344	0.020
Age of the child in years (Ref = 0–2)
3–5				-0.910	1.013	0.370	-1.060	1.011	0.295
6–8				-0.683	1.033	0.509	-0.893	1.036	0.389
9–11				0.861	1.085	0.428	0.388	1.078	0.719
12–14				1.183	1.222	0.334	0.252	1.222	0.837
15–17				2.781	2.359	0.239	1.533	2.344	0.513
Sex (Ref = female)				2.188	0.451	<0.001	2.000	0.450	< 0.001
Head/abdominal pain							-1.098	1.021	0.283
Speech and language							-0.465	0.936	0.619
Conduct disorder							1.778	0.940	0.060
Enuresis							0.254	1.060	0.811
High educational level (Ref = low)							-2.127	1.000	0.034
Intermediate educational level (Ref = low)							-1.792	1.020	0.079
Random intercept	24.162			22.544			21.650
Residual variance	5.254			5.255			5.213
AIC	7316.4			7269.7			7232.2

The β estimates the change in the dependent variable SDQ total score per unit of increase of continuous predictors or in the yes versus no group for binary predictors. SDQ-S and SDQ-P were combined
Negative β-coefficients represent a decrease in SDQ total scores per unit of increase of continuous predictors or in the yes versus no group for binary predictors
Random effects model adjusted for gender, age, educational level of the parents and diagnosis of the child with inverse probability of treatment weighting (IPTW) and random intercept (n = 599)

ISSN: 2731-4553