15581-m-nanninga

PROBLEMS AT ENTRY, CARE RECEIVED, OUTCOMES 113 negative attributes of children with regard to emotional problems, behavioural problems, hyperactivity, peer problems, and prosocial behaviour. We measured internalizing problems as the sum of ten items related to emotional and peer problems (Cronbach’s α parents =.78 (T1, T2, T3), adolescents = .75 (T1), .76 (T2), .73 (T3)), and externalizing problems as the sum of ten items related to behavioural problems and hyperactivity (Cronbach’s α parents = .83 (T1, T2, T3), adolescents =.74 (T1), .76 (T2), .73 (T3)) [22]. Scores were (a) dichotomized as ‘problems’ if either the parent score or the adolescent score was increased, versus ‘no problems’, and (b) counted as change in mean severity (ranges 0-20). Parenting problems were measured using total parent scores on the nine-item version of the Alabama Parenting Questionnaire (APQ) [23]. The APQ includes a five- point Likert scale on the domains poor supervision, inconsistent disciplining and positive parenting (maximum of three missing items, Cronbach’s α=.66 (T1, T2), .69 (T3)). The total APQ scale was dichotomized into ‘problems’ (score 2.25-5), based on the 20% highest scores in the total reference cohort on T1 and otherwise into ‘no problems’, i.e. reflecting the SDQ cut-off points. Scores were also summed per measurement, leading to changes in severity between measurements (range 1-5). Family problems were measured using the General Functioning Scale (GF) of the McMaster Family Assessment Device (FAD) [24,25]. Using 12 items, the scale included the dimensions problem solving, communication, roles, affective responsiveness and involvement, and behavioural control. Parents rated their agreement on a four-point scale from ‘totally disagree’ to ‘totally agree’ (maximum of two missing items, Cronbach’s α=.87 (T1, T2), .83 (T3)). The GF was (a) dichotomized into ‘problems’, i.e. unhealthy family functioning, or ‘no problems’, i.e. healthy family functioning, and (b) counted as change in mean severity (range 1-4) [26]. Number of problems was measured by combining the dichotomized scores on internalizing, externalizing, parenting and family problems, resulting in five categories ranging from problems in no domain to problems in four domains. For example, a child with both internalizing and externalizing problems, and a child with both internalizing problems and parents who had parenting problems fell in the second category. Types of psychosocial care referred to the psychosocial care service by which children entered this study, categorized as ‘PCH’, ‘CASC’ or ‘CAMH’. Children in the reference cohort were categorized as ‘Not in care’. Duration of care as a process-outcome was included and defined as (0) ‘0-3 months’, (1) ‘>3 months-1 year’, and (2) ‘>1 year’, based on information from adolescents, and from the parents if no child information was available.