Anxiety disorders are one of the most common types of psychopathology during childhood . In Europe, 12-20% of children experience anxiety . More than 12% of the children is diagnosed with an anxiety disorder. In the Netherlands the prevalence of anxiety disorders is even higher than 20% [3, 4]. Various forms of anxiety disorders exist. The most common anxiety disorders in childhood are the specific phobia, social anxiety disorder, separation anxiety and generalized anxiety .
A characteristic feature of all anxiety disorders in children is the preoccupation with danger. Although the type of stimulus eliciting fear may change over time due to developmental transformations, anxiety and fear are a chronic problem. In an attempt to cope with the perceived threat of the outside world children with anxiety disorders learn to avoid potentially threatening situations. As a result, the social and academic functioning of children with anxiety is jeopardized. Children with anxiety have fewer friends and receive lower grades at school . Additionally, the importance of parental factors (e.g. genetic transmission, anxious modeling, over-controlling parenting style) in the etiology and maintenance of childhood anxiety is well established . In turn, as a result of a child's anxiety family life can be impaired such that families with anxious children perceive more stress and participate less in social activities . Finally, clinical levels of anxiety during childhood present great risks for future development, such as an increased risk for substance abuse and suicidality during adolescence  and higher rates of psychopathology and educational underachievement in adulthood [9, 10]. Because of the high prevalence of anxiety in children and the detrimental effects on socio-emotional and academic functioning, which bear great challenges for future development, much research has been devoted to identifying effective interventions to target childhood anxiety.
Cognitive behavioral therapy (CBT) - with or without parental involvement - is consistently being identified as the most effective treatment for childhood anxiety. A recent meta-analysis among 24 studies found an overall posttreatment remission rate of anxiety disorders of 55.4% and showed a mean overall effect size of CBT is .86 . All CBT treatments share similar ingredients such as exposure to anxious situations, cognitive restructuring of dysfunctional thoughts, relaxation before and during anxious situations and positive self-talk. In the Dutch context a protocollized CBT treatment for anxious children and adolescents called 'Thinking + Doing = Daring' (TDD) has been developed by Bögels  based on these principles. Importantly, the intervention integrates parents by teaching them how to communicate with their child about anxious situations and how to motivate and support their child in overcoming its fear. Also the parent's own fears and anxieties are being discussed. The treatment consists of twelve weekly sessions with the child and three sessions with the parents. Three months after therapy, a follow-up session takes place.
In a study by Bodden, Bögels and colleagues  the effectiveness of the TDD-treatment was tested with a randomized controlled trial including three different conditions. Children were between eight and 17 years old and either received the TTD (individual CBT with little parental involvement), a family CBT or were put on the waitlist for eight weeks. Some of the families in the waitlist condition received the treatment after eight weeks. These post waitlist results were included in the effect calculation. At post-treatment 41% of the children was free of all anxiety disorders and 56% was free of their primary anxiety diagnosis. All waitlist children still had anxiety disorders after the waitlist period. At three-month follow up these percentages were 52% and 67% respectively. Concerning the difference between the TTD and the family CBT, the study found better treatment outcomes for the TDD (56% recovered from anxiety) compared to the family CBT (28% recovered from anxiety). The effect size for the TDD was 1.39 and 1.03 for the family CBT (as measured with the parent version of the dutch version of the Screen for Child Anxiety Related Emotional Disorders (SCARED-NL). In families where parents had an anxiety disorder, children also benefited more from the TDD (46% recovered from anxiety at post treatment) compared to family CBT (19% recovered from anxiety at post treatment). Also, more dropouts were found in the family CBT-condition (19%) than in the TDD (3%). In sum, the effect sizes of the TDD program are promising and based on this study's results the TDD, that is an individual CBT with little parental involvement, seems to be more beneficial than a family CBT. The child and therapist manuals of the TDD are published including assessments and treatment integrity forms, and the program is now widely used in the Netherlands.
The primary aim of this study is to replicate and extend the findings of the Bodden et al study  in a randomized controlled trial. In contrast to the Bodden et al study  two conditions will be used: the experimental condition which will consist of the TDD program and the control condition which will consist of treatment as usual (TAU). The ultimate proof of the effectiveness of a treatment program is when it exceeds the effects of the treatment that families and children normally receive. Furthermore, to test the long-term effects of treatment six-month and 12-month follow up assessments will be conducted. Importantly, this study will take place in the real-world context, where co-morbidity is the rule rather than the exception . In this way the study's results can be generalized to the context where the intervention may eventually be delivered, that is in mental health institutions in the Netherlands.
Finally, despite the promising results of CBT so far, variability in treatment outcomes remain. Not all children with anxiety profit from therapy. It is not clear why some children fail to show improvement in therapy, since we have little understanding of the underlying mechanisms and processes of change. Randomized controlled trials inform us if a certain interventions works but they do not tell us by what mechanism, information that is essential in order to further improve and tailor intervention efforts. Two potential mediators of change will be examined in the current study: therapeutic alliance and parenting.
Many researchers have confirmed the importance of alliance in adult therapy. Stronger therapeutic alliance predicts better outcome [15, 16]. However, the role of alliance in child-therapy has received little attention so far and results are mixed. Kazdin  found that strong therapeutic alliance predicted more improvement in the child. However, in a study by Liber  alliance and treatment outcome were only moderately related. In two studies by Kendall [19, 20] no significant association between alliance and treatment outcome were found. In the adult literature, alliance at one month after treatment has started is usually used to predict treatment outcomes . However, alliance is likely to fluctuate across the treatment period. In order to test when alliance best predicts treatment outcomes, alliance will be assessed at multiple time points: one and two months after treatment has started and at post treatment.
The second potential mediator is parenting. Childhood anxiety is more common in families with anxious parents, suggesting a familial transmission of anxiety . Numerous studies in previous years have focused on the influence of family interactions in the development, maintenance, and improvement of childhood anxiety (e.g. [22, 23]) and found several potential parenting behaviors influencing childhood anxiety. The dimensions rejection and control received a great amount of consideration in the parenting literature (e.g. [24–26]), but with different definitions and mixed results. Recently, McLoad, Wood and Weisz  conducted a meta-analysis on both dimensions and found that both rejection (small effect) and control (medium effect) were associated with childhood anxiety. If treatment for childhood anxiety is effective, parent-child interactions are likely to change. In the current study, we will track changes in the dimensions rejection and control.
Aim and hypotheses
The primary aim of the study is to evaluate the effectiveness of the CBT-program Thinking, Doing and Daring (TDD) by comparing it to treatment as usual (TAU). Our primary outcome, children's anxiety, will be assessed through parent reports and children's self-reports. To measure long-term effects of treatment, follow-up assessments at three month, six month and one year follow up will be conducted. The second aim is to analyze whether there are secondary positive outcomes beside recovery of anxiety. Secondary outcomes (e.g. depression, aggression) will be assessed through parent and teacher reports. The third aim is to analyze the potential mediating influence of alliance and parenting on positive treatment outcomes.
More specifically, we expect that a) children in the experimental condition will have significantly less anxiety symptoms after treatment and at the follow-up measurements than children who received treatment as usual, b) children who recover from an anxiety disorder will also show a significant reduction in secondary problem behavior (e.g. depression, aggression), c) children who form a strong alliance with their therapist will have less anxiety symptoms than children who form a less strong alliance, d) parent-child dyads for those children who improve through therapy will show less parental control and rejection after treatment than at the start of treatment.