The role of parental attributions in predicting parenting intervention outcomes in the treatment of child conduct problems

Highlights

• Parents differ in the attributions they make about their child and his/her behaviours.

• Little is known about whether parental attributions predict parenting intervention outcomes.

• Negative and change resistant parental attributions uniquely predicted poor treatment outcomes in the current study.

• Parental attributions may be possibly inlcuded in treatment to maximise treatment gains for children with conduct problems.

Abstract

Background

Parent attributions about the causes of their children's behaviour problems are a known predictor of problematic parenting and ongoing child problems. However, their importance in parenting interventions remains unknown as research has not addressed whether parental attributions predict outcomes associated with parenting training. The current study examined whether problematic pre-treatment and change resistant parental attributions during treatment uniquely predict child behaviour outcomes.

Method

Participants were 250 families with children aged from 3 to 16 referred to specialist clinics for the treatment of conduct problems. Measures of family demographic information, parental attributions, and severity of child conduct problems were collected as part of pre-treatment, post-treatment, and 3-month follow-up assessments.

Results

Linear regression showed that mothers' pre-treatment parental attributions uniquely predicted severity of conduct problems at the post-treatment and 3-month follow-up assessments. Fathers' pre-treatment parental attributions uniquely predicted severity of conduct problems at the 3-month follow-up assessment. Reductions in problematic pre-treatment parental attributions were recorded at post-treatment for both parents. However, smaller reductions or increases in mothers’ problematic attributions uniquely predicted worse child behaviour outcomes, a result not replicated for fathers.

Conclusions

Findings that pre-treatment and change resistant parental attributions predict poorer child behaviour outcomes after controlling for other predictors and treatment effects recommend that parental attributions should be assessed prior to and after treatment and possibly included in treatment to maximise treatment gains for children with conduct problems.

Introduction

Child conduct problems are associated with significant risk for a life course of antisocial behaviour and poor mental health (Copeland, Shanahan, Costello, & Angold, 2009). Research evidence indicates that Behavioural Parent Training (BPT) is an effective form of intervention for child conduct problems and curtailing the risk for poor mental health outcomes (Kaminski & Claussen, 2017). Follow-up outcomes for BPT in the treatment of conduct problems diagnosed as Oppositional Defiant Disorder or Conduct Disorder suggest two-thirds of cases are in the nonclinical range 1-year after treatment with ongoing benefits for social and emotional adjustment a decade after treatment completion (McMahon, 1994; Webster-Stratton, Rinaldi, & Reid, 2011). This impressive data, however, is counterbalanced by indications that one-third of cases fail to maintain the benefits of BPT. Therefore, a major issue in the dissemination of BPT is to maximise its effectiveness by being beneficial to the diversity of families seeking out and participating in treatment.

In one line of research addressing the issue, researchers have identified family, parent, and child predictors of poor outcomes in efforts to modify BPT to accommodate for risk factors. Those least likely to benefit are families with multiple dysfunctions characterised by low socioeconomic status, parental psychopathology, marital distress, and children with severe conduct problems, callous-unemotional (CU) traits, and/or comorbidity (Beauchaine, Webster-Stratton, & Reid, 2005; Hawes, Price, & Dadds, 2014; Reyno & McGrath, 2006). However, some factors are less susceptible to modification (e.g., socioeconomic status) or not the primary focus of treatment (e.g., parental mental ill-health). One factor that is modifiable and worthy of increased attention is problematic parental attributions defined as internal, stable, and controllable causal explanations for child problem behaviours, which is hypothesised to influence how parents accept, engage, and benefit from BPT (Hoza, Johnston, Pillow, & Ascough, 2006; Mah & Johnston, 2008; Sawrikar & Dadds, 2018). The current study examines whether problematic parental attributions uniquely predict child behaviour outcomes as evidence to include a focus on parental attributions in BPT.

In the recent Sawrikar and Dadds (2018) review, mixed evidence was found for parental attributions having a direct effect on child behaviour outcomes after controlling for other predictors of outcomes. Previous longitudinal studies have concentrated on problematic pre-treatment parental attributions and demonstrated their utility in predicting worse child behaviour outcomes (Hoza et al., 2000; Mattek, Harris, & Fox, 2016). However, there was little evidence from clinical studies suggesting that targeting parental attributions using cognitive therapy improves outcomes beyond the standard program (Sanders et al., 2004). While the latter argues against including components focusing on parental attributions, it was suggested that the longitudinal studies provided indirect evidence that a subset of parents show pre-treatment parental attributions that interfere with outcomes. Parents posited to benefit were those demonstrating problematic, change resistant causal explanations for problem behaviours in treatment.

Hoza et al. (2006) presented the role of change resistant parental attributions in predicting outcomes in their multistage social cognitive model for BPT. Hoza et al. (2006) hypothesised that changes in parental attributions could occur experientially within the improvement phase of treatment where parents, who may have initially presented with problematic parental attributions but recognise behavioural changes in their child from implementing the strategies, change their perceptions of the child. This suggests that the effects of parental attributions on outcomes change as parents’ progress through treatment. As such, longitudinal research is required to go beyond asking whether pre-treatment parental attributions uniquely predict outcomes, to include an examination of whether changes in parental attributions during treatment also predict child behaviour outcomes.

Unfortunately, there is a dearth of available research evaluating the improvement phase of treatment. Only two studies were found to have studied whether BPT was associated with changes in causal explanations for child behaviour, both indicating that parents make less child-internal attributions for problematic behaviour post-treatment (Sanders et al., 2004; Whittingham, Sofronoff, Sheffield, & Sanders, 2009). Neither study was conducted with families referred for the treatment of child conduct problems. Additionally, only one explorative study demonstrated that pre-to post-treatment changes in parental attributions were associated with positive treatment outcomes and that a lack of change in parental attributions was associated with a similar lack of child behaviour improvement in treatment (Wilson & White, 2006). These findings were only descriptive as the study consisted of five parents where the analysis was not subjected to statistical testing and did not control for other predictors of outcomes. In sum therefore, research has thus far been unable to clearly ascertain whether pre-to post-treatment changes in parental attributions are a unique predictor of child outcomes. Thus, their importance as a focus of treatment remains unknown.

It is noted that Katzmann et al. (2017) implicitly examined the predictive utility of changes in parental attributions while investigating the parenting and cognitive mechanisms of a self-help BPT program. Changes in hostile intent attributions were modelled to predict and mediate child behaviour improvements, while controlling for changes in parenting behaviour and pre-treatment scores on child behaviour symptoms. Changes in hostile intent attributions were found to predict and mediate improvements in child adjustment. However, caution in generalising the findings to clinic-based BPT appears warranted as the assumed mechanism of change, parenting behaviours, was not found to mediate outcomes. Their analysis was also correlational as the assessment of mediators and outcomes occurred at the same time, thus their findings may only reflect a concomitant effect of change in parent and child behaviour. This limits the extent to which inferences of prediction can be made. A more valid test of prediction dictates temporal ordering of variables to ensure measures of pre-to post-treatment changes in parental attributions precede assessment of child behaviour outcomes.

The current study tests whether parental attributions are a unique predictor of child outcomes in parenting programs, and thus should play a key role in evidence-based interventions. It addresses the key limitation of existing longitudinal research having conducted little research into whether pre-to post-treatment changes in parental attributions predict subsequent child behaviour outcomes. It should be noted that, as with all single group treatment prediction studies, changes in parental attributions and other variables measured in this study are attributable to both treatment effects and other concurrent changes in child and family functioning occurring during the treatment period which is consistent with anticipated outcomes of BPT (Colalillo, Miller, & Johnston, 2015; Kazdin & Wassell, 2000). Based on Hoza et al. (2006) model, it was hypothesised that (i) pre-treatment parental attributions uniquely predict child behaviour outcomes after controlling for other predictors of outcomes, and (ii) changes in parental attributions uniquely predict child behaviour outcomes after controlling for other predictors of outcomes and changes in child behaviour.