Effectiveness and process variables in modularized CBT
for youth with social anxiety:
A single-case study examining therapist flexibility, child involvement,
therapeutic alliance, and anxiety
Ariënne Verveen
Studentnr.: 10581510
Master thesis
Clinical Developmental Psychology
University of Amsterdam
Supervisor Developmental Psychology: Marija Maric
External supervisor: Liesbeth Telman
Second reviewer: Helga Harsay
Date: 07-08-2017
Abstract
In this single-case study the effectiveness of modularized cognitive behavioral therapy (CBT) for youth with social anxiety disorder (SoAD) and relationships between the process variables therapist flexibility, child involvement, therapeutic alliance, and anxiety during treatment are investigated. Eleven youngsters (5 boys and 6 girls, 8 to 18 years old, Mage = 12.18, SD = 2.64) received modularized CBT with the therapists either receiving weekly feedback (N = 6) or no feedback (N = 7) about the child’s opinion on the session and level of anxiety. At pretreatment, post treatment, and follow up children and their parents were assessed on clinical diagnoses (SCID-junior) and anxiety symptoms (SCARED-71). Furthermore, children and parents weekly reported their opinion about the session (SRS) and their level of anxiety (TPM). Audiotapes of the sessions were used to code therapist flexibility (TTDAC-F), child
involvement (CIRS), and therapeutic alliance (WAI-O-S). At follow-up 67% of the children were free of their SoAD diagnosis. Mothers reported higher improvement in (social) anxiety symptoms than children and their fathers. Overall, there was no clear difference in social anxiety symptoms between children of which the therapists received feedback and children of which the therapists did not. Finally, for one participant there was a positive relationship between therapist flexibility and child involvement two weeks later during the treatment and for another participant a negative relationship between therapist flexibility and anxiety in the same week. Modularized CBT may be a promising treatment for children with SoAD. Future research is necessary to clarify the role of the process variables in this treatment.
Effectiveness and process variables in modularized CBT for youth with social anxiety: A single-case study examining therapist flexibility, child involvement, therapeutic alliance,
and anxiety
Anxiety disorders are one of the most common forms of psychopathology in children and adolescents (Costello, Mustillo, Erkanli, Keeler, & Angold, 2003). Research shows that approximately 10% of the adolescents in the Netherlands had an anxiety disorder in the 6 months prior to the
investigation (Verhulst, Van der Ende, Ferdinand, & Kasius, 1997). Among the anxiety disorders social anxiety disorder and specific phobia are the most common (Verhulst et al., 1997). Social anxiety disorder (SoAD), also known as social phobia, is the fear of social situations in which the individual may be criticized by others (APA, 2013). In order to diagnose children with SoAD the fear must occur in interactions with peers and not just in interactions with adults. The most important criteria in the Fifth Edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) are that the individual is afraid that others will negatively evaluate him or her and therefore fears or avoids social situations (APA, 2013). Children typically develop SoAD when they are between 8 and 15 years old(Kessler et al., 2005). The disorder is often comorbid with other anxiety disorders,
depression or substance abuse (Beesdo et al., 2007) and adolescents with SoAD have a strong risk for recurrent anxiety or depressive disorders in adulthood (Pine, Cohen, Gurley, Brook, & Ma, 1998). Moreover, SoAD is associated with elevated rates of school dropout and decreased well-being (Patel et al., 2002). It is important to treat youth suffering from SoAD disorders effectively, because of the early age of onset, high comorbidity, risk for recurrent anxiety or depressive disorders, and negative life consequences.
Cognitive behavioral therapy (CBT) is an efficacious method for treating anxiety. It consists of two central components: skill building and practice. The skill-building component is composed of psycho-education, somatic management, and cognitive restructuring and the practice component of gradual exposure to feared situations (Albano & Kendall, 2002). National and international studies show that 67% of the children treated for an anxiety disorder are free of their primary diagnosis after treatment (Bodden et al., 2008a; Bodden et al., 2008b; Ginsburg et al., 2011; In-Albon & Schneider, 2007; Van Steensel & Bögels, 2015; Van Steensel, Dirksen, & Bögels, 2014). However, CBT seems to be less effective in children and adolescents with SoAD. Youth with SoAD are less often symptom free at the end of treatment (Ginsburg et al., 2011; Hudson et al., 2015a), have a slower rate of change (Hudson et al., 2015b), and are more likely to continue to meet the criteria for their diagnosis than youth with other anxiety disorders (Crawley et al, 2009; Hudson et al., 2015a; Hudson et al., 2015b). Since SoAD is prevalent and CBT appears to be less effective for this group, it is important to investigate which factors contribute to a better treatment outcome for youth with SoAD.
Several suggestions have been proposed to improve the treatment of youth with SoAD. Firstly, because children with SoAD rate themselves as less socially competent (Spence, Donovan, &
Brechman-Toussaint, 1999) and show social skills deficits (Beidel, Turner, & Morris, 1999; Spence et al., 1999), several studies suggest adding or intensifying social skills training to treatment (Beidel et al., 1999; Ginsburg et al., 2011; Hudson et al., 2015a, Hudson et al., 2015b; Kendall, Settipani, & Cummings, 2012). Secondly, research shows that heightened self-focused attention is a predictor of and plays an important role in maintaining social anxiety (Hodson, McManus, Clark, & Doll, 2008; Kley, Heinrichs, Bender, & Tuschen-Caffier, 2012). Therefore, researchers propose adding attention-retraining techniques to treatment in which task concentration (Hudson et al., 2015a; Hudson et al., 2015b; Kley et al., 2012) or mindfulness is trained (Semple, Reid, & Miller, 2005; Van Bockstaele & Bögels, 2014). Mindfulness can be defined as paying attention in a particular way: on purpose, in the present moment, and non-judgmentally (Kabat-Zinn, 1994). Thirdly, because youth with SoAD show negative interpretation biases during social events (Vassilopoulos & Banerjee, 2008), it is suggested to integrate strategies to disconfirm negative beliefs (Hudson et al., 2015b). Finally, Kendall et al. (2012) recommend implementing exposure tasks with peers in therapy in order to resemble real social
situations. Some of these treatment strategies are not implemented in all CBTs with SoAD youth (e.g., task concentration training, mindfulness), other are used, but should be intensified (e.g., social skills training, cognitive therapy). However, the largest issue so far seems to be that in all previous studies all children receive the therapy protocol in the same way regarding order and dosage of the therapeutic strategies. As a strategy to enhance treatment outcomes of youths with SoAD, it has been suggested that these children and adolescents should receive therapy in a more individualized way, that is, tailoring the selection and implementation of therapy techniques to the personal needs of the clients (Crawley et al., 2008; Hudson et al., 2015b; Kendall et al., 2012;Kley et al., 2012).
A way to implement these various suggestions in a more individualized way in the treatment of youth with SoAD is with a modular CBT. A Modular CBT consists of evidence-based cognitive-behavioral techniques divided into separate modules that can be matched to the child’s individual strengths and needs (Chorpita, Taylor, Francis, Moffitt, & Austin, 2004; Reuther, Davis III, Moree, & Matson, 2011). This approach allows therapists to develop individualized treatment plans while remaining within the bounds of evidence-based practice (Reuther et al., 2011). Usually the treatment consists of a combination of compulsory modules, like psycho-education and relapse prevention, and optional modules, for example cognitive retraining and parent training (Chorpita et al., 2004; Galla et al., 2011; Reuther et al., 2011). Earlier research indicates the effectiveness of a modular treatment for children with anxiety disorders (Chorpita et al., 2004; Chorpita et al., 2013; Galla et al., 2011; Weisz et al., 2011). Chorpita et al. (2004) examined the efficacy of a modular CBT in 11 children with diverse anxiety disorders of whom two had (non-primary) SoAD. All seven completers were free of their diagnoses at the end of treatment and at 6 months follow-up. Furthermore, Galla et al. (2011) studied the efficacy of a modular CBT in children with separation anxiety disorder, generalized anxiety disorder or SoAD. The results indicated that 83,3% percent of children were free of their primary diagnosis at 1-year follow-up. Finally, Weisz et al. (2012) and Chopita et al. (2013) compared
the effectiveness of a modular CBT with a standardized CBT or usual care in children with anxiety disorders, depressive disorder, and/or conduct problems. Children receiving the modular treatment showed more improvement than children receiving the standardized treatment (Weisz et al., 2012) or usual care (Chorpita et al., 2013) and had fewer diagnoses than children receiving usual care (Weisz et al., 2012). However, these studies did not include mindfulness in the treatment protocol and the participants were only up until 12 years old. Moreover, to our knowledge no studies are conducted investigating the effectiveness of modular treatment specifically for youth with SoAD.
Most research on (modular) CBT for youth has focused on the effectiveness of treatment (Zack, Castonguay, & Boswell, 2007). Besides the knowledge that a specific treatment works, it is also important to study which processes contribute to this effectiveness. Although there is some research investigating the working mechanisms of specific CBT technique variables (Herres, Cummings, Swan, Makover & Kendall, 2015), research on therapist, child, or relationship variables in child therapy is sparse (Kendall & Ollendick, 2005; Zack, et al., 2011). In addition, to our knowledge there are no studies examining these process variables in modularized CBT for children. It is important to study process variables, such as therapist flexibility, child involvement and therapeutic alliance. Knowledge about which process variables contribute to treatment outcome may help us understand why treatment works (Hudson et al., 2014). Moreover insight in these variables will contribute in selecting and optimizing treatment for youth (Mcleod, 2011).
A way to investigate if individualizing modularized CBT contributes to treatment outcome for children with SoAD is by assessing the process variable therapist flexibility. Therapist flexibility is defined as treatment adaptation occurring within the boundaries of treatment fidelity (Chu & Kendall, 2009; Kendall & Chu, 2000; Kendall & Beidas, 2007). Earlier research examining adults with variable personal concerns or psychopathology indicates that therapist flexibility is related to a better
therapeutic alliance (Kivlighan, Clements, Blake, Arnzen, & Brady, 1993) and better treatment outcomes (Owen & Hilsenroth, 2014). However, few studies are conducted investigating therapist flexibility in the treatment of children with anxiety disorders and some authors argued that therapist flexibility should be investigated in combination with child involvement. Research indicates that child involvement during therapy is an important predictor of treatment outcomes in children with anxiety disorders (Chu & Kendall, 2004; Tobon et al., 2011). It is theorized that therapist flexibility could work as an engagement-enhancing strategy. In other words, if therapists respond better to the child’s individual needs this could lead to higher child involvement in treatment. Chu and Kendall (2009) tested this theory by examining the relationship between therapist flexibility, child involvement, and treatment outcome in a CBT for anxious youth. They found that therapist flexibility was related to child involvement later in therapy and that child involvement predicted improvement in diagnosis and impairment after treatment. Hudson et al. (2014) also found support for this relationship in their study on process variables. However, the authors of both studies used fixed instead of modular, more flexible, CBT protocols and did not examine if these relationships varied by anxiety disorder type.
Besides therapist flexibility, treatment can be personalized by adapting the content based on client feedback after each session. Client feedback is defined as providing the therapist with feedback on the client’s progress based on frequent assessment of the client’s well-being during therapy (De Jong et al, 2014). Research with adults indicates that when therapists receive feedback on the client’s well-being, this leads to better treatment outcomes for clients who are not progressing well (De Jong et al., 2014; De Jong, Van Sluis, Nugter, Heiser, & Spinhoven, 2012). The relationship between client feedback and treatment outcomes was also found in research investigating children (Bickman, Kelley, Breda, de Andrade, & Riemer, 2011). Bickman et al. (2011) studied the effect of weekly feedback to the therapist on treatment outcomes in a randomized trial with youth with various mental health problems. Results indicated that youth in the feedback condition improved faster than youth in the no-feedback condition. However, to our knowledge no studies are conducted investigating the effect of client feedback on treatment outcome for youth with (social) anxiety disorders.
Finally, because children with social anxiety are usually withdrawn, it is suggested that in order to enhance treatment outcomes it is necessary to invest extra time in building a good therapeutic alliance between the child and the therapist (Crawley et al., 2008; Hudson et al., 2015b). Therapeutic alliance consists of three components: the bond between the therapist and client, the agreement on goals, and the agreement on tasks in therapy (Bordin, 1979). The bond between the therapist and client can be defined as the feeling of attachment, trust, and mutual engagement. Agreement on goals refers to agreement on the areas in which change should occur. Finally, agreement on tasks can be described as agreement on the activities that should be performed in order to realize the predefined goals
(Stickens, Ulburghs, & Claes, 2009). Studies investigating the association between therapeutic alliance and treatment outcome in youth with anxiety disorders yielded mixed results, with some studies finding that a stronger alliance is associated with a better treatment outcome (Cummings et al., 2013; McLeod & Weisz, 2005; Hudson et al., 2014; Hughes & Kendall, 2007) and others not (Chiu, McLeod, & Wood, 2008; Kendall, 1994; Kendall et al., 1997). However, to our knowledge there are no studies investigating this relationship for youth with SoAD.
Several studies suggest that children with SoAD benefit less from CBT than children with other anxiety disorders (Crawley et al., 2008; Ginsburg et al., 2011; Hudson et al., 2015a; Hudson et al., 2015b). Therefore, applying CBT in a modular and individualized manner in order to target specific SoAD symptoms more intensively might be extra beneficial for this group (Crawley et al., 2008; Kendall, et al, 2012;Kley et al., 2012; Hudson et al., 2015b). It is important to investigate this by examining the relationship between therapist flexibility, child involvement, and the level of anxiety in youth with SoAD. Furthermore, since client feedback may provide a way to personalize treatment, it is important to investigate the effect of client feedback on treatment outcome in youth with SoAD. Finally, there are no studies investigating the association between therapeutic alliance and anxiety symptoms in youth with SoAD and building a good therapeutic relationship may be necessary to treat youth with SoAD effectively. Therefore, it is important to investigate the relationship between
therapeutic alliance and the level of anxiety in youth with SoAD. Clarification of the effectiveness of a modular CBT for youth with SoAD and the relationship between child involvement, therapist
flexibility, therapeutic alliance, and treatment outcome may benefit therapists in choosing the most beneficial treatment for youth with SoAD.
Aim of the Study
In this study the effectiveness of CBT for youth with SoAD is investigated using a CBT protocol (Thinking + Doing = Daring - adapted modularized version; Bögels, 2008) that allows the therapist to select modules that best fit the child’s needs. Participants who receive treatment are assigned to a feedback or a control condition. Therapists with clients in the feedback condition will have the opportunity to personalize the treatment based on client feedback, whereas therapists with clients in the control condition will not receive the feedback. It may be possible that this protocol treats children with SoAD more effectively than traditional CBT protocols, because the therapist receives feedback (De Jong et al., 2012; De Jong et al., 2014) and is able to adapt the protocol more to the child’s unique problems (Crawley et al., 2008; Kendall, et al, 2012;Kley et al., 2012; Hudson et al., 2015b). Moreover, several of the earlier proposed suggestions to improve treatment for youth with SoAD are included in the optional modules. The protocol consists of a mindfulness module and a task concentration exercise to train attention. To the best of our knowledge, no studies are conducted that included mindfulness in modularized treatment for youth with anxiety disorders. In addition, therapists can choose to focus extra on disconfirming negative beliefs with the cognitions module or implement exposure with peers in the exposure module. Furthermore, the relationship between therapist
flexibility, child involvement, and the level of anxiety during therapy is examined. A high level of therapist flexibility may be extra beneficial for youth with SoAD, because they improve less from CBT than youth with other anxiety disorders. The modular character of the protocol enables the therapists to be more flexible in their treatment. This may lead to higher therapist flexibility than in the previous study, which may lead to stronger relationships between therapist flexibility, child
involvement, and treatment outcome. Finally, the relationship between therapeutic alliance and the level of anxiety during treatment is examined. Since this is the first study investigating therapist flexibility, child involvement, and therapeutic alliance in youth with SoAD a range of single-case studies were conducted with a detailed analysis of changes in main variables per session.
Based on the previous findings the following hypotheses are composed: 1. Effectiveness
a. Modularized CBT will lead to a significant decrease in the level of (social) anxiety after treatment as compared to before treatment (Bodden et al., 2008a; Bodden et al., 2008b; Crawley et al., 2008; In-Albon & Schneider, 2007; Kendall, et al, 2012;Kley et al., 2012; Hudson et al., 2015b; Van Steensel & Bögels, 2015; Van Steensel et al., 2014).
b. The level of (social) anxiety of youth in the control condition will improve less during and after the treatment than that of youth in the feedback condition (De Jong et al., 2012; De Jong et al., 2014).
2. Process variables
a. Therapist flexibility is positively related to child involvement during treatment (Chu & Kendall 2009).
b. Child involvement is negatively related to the level of anxiety during treatment (Chu & Kendall 2009).
c. Therapist flexibility is negatively related to the level of anxiety during treatment (Chu & Kendall 2009).
d. Therapeutic alliance is negatively related to the level of anxiety during treatment (Crawley et al., 2008; Hudson et al., 2015b).
___________________________________________________________________________
Method
Participants
Children and adolescents with anxiety disorders and/or symptoms were registered for mental healthcare at participating treatment centers in the Netherlands. If therapists decided the treatment protocol was suitable for a child, children (N = 97) and at least one of their parents were asked to participate in a larger research project, investigating the working factors in the treatment of anxiety. In line with previous studies, youth that are intellectually disabled or have psychotic symptoms,suicidal tendencies, autism spectrum disorder or bipolar disorder (N = 1)were excluded (Bodden et al., 2008a; Bodden et al., 2008b; Van Steensel & Bögels, 2015; Van Steensel, Dirksen, & Bögels, 2014). From this larger group participants with a primary diagnosis of SoAD (N = 19) were selected as participants for this study. See figure 1 for an overview of the selection procedure. It was required that at least 10 participants participated in this study, because this was the first study investigating therapist flexibility and child involvement in youth with SoAD and for each participant a detailed assessment per session was done. This minimum of 10 participants was chosen, because the Task Force of the Society of Clinical Psychology of the American Psychological Association requires at least 9 single case studies in order to give a treatment the label effective (APA, 1995). Eleven children between 8 and 18 years old (6 girls and 5 boys, Mage = 12.18, SD = 2.64) and 22 parents participated in the present study. An overview of the data per participant is displayed in Table 1.
Figure 1. An overview of the participants’ selection procedure.
Note. The participant that was excluded had Autism Spectrum Disorder.
Table 1
An Overview of the Condition, Sex, Age, Type of SoAD (General or Performance Only SoAD), Comorbidity, Type of Mental Health Institution and which Parents are Participating for each Participant
Participant Condition Sex Age Type of
SoAD
Comorbidity Type of mental
healthcare
Parents participating
1 Feedback Female 8 G - Sp Both
2 Feedback Male 12 PO GAS Sp Both
3 Feedback Female 10 G - Sp Both
4 Feedback Male 13 G - Gen Both
5 Feedback Female 14 G SpP Gen Both
6 Feedback Female 15 G GAS, Ins Sp Both
7 Control Male 11 G SAD, SpP, GAS,
ADHD-C, Dyst
Sp Both
8 Control Male 9 PO - Sp Both
9 Control Male 17 G - Gen Both
10 Control Female 12 G SpP, GAS Sp Both
11 Control Female 13 G GAS, ADHD-I, Dep,
Ins
Note. G = general SoAD, PO = performance only SoAD
SAD = Separation anxiety disorder, SpF = Specific phobia, GAS = Generalized anxiety disorder, ADHD-C = Attention deficit hyperactivity disorder - combined type, ADHD-I = ADHD – Inattentive, Dyst = Dysthymia, Dep = Depression, Ins = Insomnia.
Sp = Specialized mental healthcare, Gen = General mental healthcare.
Modular treatment
The children receive modularized CBT based on the Thinking + Doing = Daring protocol (Bögels, 2008). To facilitate individually based treatment and therapist flexibility, this version of the protocol consists of seven optional modules (cognitive restructuring, coping, mindfulness, exposure, experiments, learning the child to talk to his or her parents and parental guidance) and two compulsory modules (psycho-education at the start of treatment and relapse prevention at the end of treatment). After receiving an initial training in therapy protocol, the therapist is free to choose the optional modules that he or she thinks are best suited to treat a particular child. This enables the therapist to tailor the treatment to the individual needs of the child and include components that he or she thinks may improve the treatment of youth with SoAD. The therapist could for example give more attention to negative beliefs, task concentration or mindfulness (Beidel et al., 1999; Ginsburg et al., 2011; Hudson et al., 2015a, Hudson et al., 2015b; Kendall et al., 2012; Kley et al., 2012; Semple et al., 2005; Van Bockstaele & Bögels, 2014). In general, the aim is to complete the treatment in 10 weeks.
Therapy protocol adherence has been warranted via examinations of audiotaped sessions and checklists filled in by the therapists at the end of each session.
Instruments
Outcome assesment
.
To diagnose youth at the beginning of the treatment and in order to investigate the course of the diagnoses during treatment the Dutch version of the Structured Clinical Interview for DSM-5 Disorder for Children (SCID-junior, Braet, Vandevivere, & Wante; unpublished manuscript) is used. The SCID-junior is a semi-structured interview to diagnose children andadolescents between 8 and 18 years old with DSM-5 disorders. The most prevalent psychiatric disorders are divided into 17 sections, of which 14 are assessed in children and their parents and three in parents only. For each section a screening question is asked. If this question is answered affirmative, the rest of the questions in the module are asked. If a child meets the criteria for a certain disorder an impairment score is asked. This score ranges from 0 to 8 and indicates how impaired a child is by his or her symptoms, 8 being very bothered and 0 being not bothered at all. A diagnosis is given, when the impairment score is higher than 4. The diagnoses of the child and the parent are combined. A primary anxiety diagnosis is determined by first giving priority to a diagnosis that both have reported and then looking at the highest impairment score. The SCID-junior developed for the DSM-5 as an adaptation
of the Structured Clinical Interview for DSM-IV diagnoses (Kid-SCID, Hien et al., 1994; Dutch translation by Dreessen et al., 1998). Since the DSM-5 has been recently published, there is ongoing research investigating the psychometric properties of the SCID-junior by Wante and colleagues (C. Braet, personal communication, July 12, 2017). The Kid-SCID is tested to be reliable and valid (Roelofs, Muris, Braet, Arntz, & Beelen, 2015). The inter-rater reliability varied between reasonable and excellent (𝜅𝜅 = 0.63-1.00) for the majority of the Kid-SCID diagnoses and the internal consistency was satisfactory to good (α = 0.70 – 1.00) for most diagnoses (Roelofs et al., 2015).
To measure the amount of anxiety symptoms in children the Screen for Child Anxiety and Relational Emotional Disorders Child-version and Parent-version (SCARED-71, Bodden, Bögels, & Muris, 2009) are used. These questionnaires consist of 71 items measuring the symptoms on the different anxiety disorders, such as panic disorder, separation anxiety disorder, social anxiety disorder, specific phobia and generalized anxiety disorder. The Child-version and the Parent-version differ in the perspective from which the questions are asked. The items are answered using a 3-point Likert scale ranging from 0 (almost never) to 2 (often). An example item is: “I am afraid to ask questions in the classroom.” For the parents and the child a total score and a score for the subscale social anxiety disorder will be computed. The total score ranges from 0 to 142 and the score for the SoAD subscale ranges from 0 to 18, with higher scores indicating a higher level of anxiety. The SCARED-71 has a good discriminative validity and a high internal consistency for the total score (α > 0.95) and SoAD subscale score (α > 0.85) for the Child-version as well as the Parent-version (Bodden et al., 2009).
To measure the level of anxiety during treatment, the overall suffering-item of the Top Problem Measure (TPM) is assessed in children and their parents. This instrument consists of 4 items (thoughts, arousal, avoidance, overall suffering) assessing the weekly anxiety symptoms and these are answered on a 5-point scale ranging from 0 (never) to 4 (always). The item used for this study is: “How often did your anxiety impair you this week” The score ranges from 0 to 4, with higher scores indicating a higher level of anxiety. The items are based on items of the Dimensional Anxiety Scales (Möller, Majdandzic, Craske, & Bögels, 2014), which has high internal consistency (α > 0.78) and moderate to high levels of convergent validity (rs = 0.29–0.73).
Process Assessment. The Child Involvement Rating Scale – revised and corresponding
manual (CIRS-R and CIRS-R manual, Hudson et al., 2014) are used to assess behavioral indices of child engagement during a session. This instrument and manual were translated into Dutch. The CIRS-R consists of 12 items, which are divided into 8 positive involvement items and 4 negative
involvement items. The items are coded on a 6-point scale ranging from 0 (not at all present) to 5 (a great deal present) using audiotapes of the therapy sessions. An example item is: “Does the child initiate discussion or introduce new topics?” The four negative involvement items are reversed scored and their sum is added to the sum of the positive involvement items. The total scores ranges from 0 to 60 with higher scores indicating higher involvement. The CIRS has a strong internal consistency (α > 0.73) and good inter-rater reliability (ICC = 0.61) (Chu & Kendall, 2004; Chu & Kendall, 2009).
Therapist flexibility is assessed with the Thinking Doing Daring Adherence
Checklist-Flexibility Scale (TDDAC-F) with the use of a corresponding manual. The TDDAC-F is based on the Coping Cat Adherence Checklist-Flexibility Scale and manual (CCPAC-F and CCPAC-F manual; Southam-Gerow, Jensen Doss, Gelbwasser, Chu, & Weisz, 2001), which has a fair inter-rater reliability (ICC = 0.40). This instrument has 23 items, each containing a critical session technique of the Coping Cat protocol. These items are ordered per session of the protocol. The CCPAC-F and manual were first translated into Dutch and then adapted for the Thinking Doing Daring protocol. Because of the modular character of this protocol there is no fixed program for each session. This made it impossible to use the fixed session format of the CCPAC-F. Therefore a list of all possible session techniques was created. Coders identified the used techniques from each session and noted these on the TDDAC-F form. The manual distinguishes two types of flexibility: content flexibility and structural flexibility. Content flexibility can be defined as the use of examples during therapy that make the session content relevant to the child. Structural flexibility can be described as the
modifications that a therapist makes to treatment activities to meet the needs of a child. Coders rated therapist flexibility on a 6-point scale ranging from 0 (not at all flexible) to 5 (extremely flexible) using audiotapes of the therapy sessions. Cases of structural flexibility received higher scores (3 or higher) than cases of content flexibility (1 or 2). The level of flexibility was weighed against the duration of the used technique by multiplying the flexibility for each technique with the corresponding duration. These weighed scores for each technique were added and subsequently divided by the total duration of the techniques.
To measure therapeutic alliance the Session Rating Scale (SRS; Boon, De Boer, & Ravestijn, 2012) and the Working Alliance Inventory Observer Version – Short Form (WAI-O-S, Tichenor & Hill, 1989) are used. The SRS consists of 4 questions (listening, how important, what we have done, everything together) assessing the child’s satisfaction of the session. The items are answered by marking agreement with the questions on a visual analogue scale ranging from 0 (low) to high (10). An example item is: “The therapist listened to me.” The score on the SRS ranges from 0 to 40 with higher scores indicating higher satisfaction. The SRS has high internal consistency (α > 0.90) (Hafkenscheid, 2010).
The WAI-O-S consists of 12 items assessing the working alliance between the child and the therapist. The items are divided into three scales: one assessing the presence of a therapeutic bond, the second agreement on therapeutic tasks, and the third agreement on the goals of therapy. Coders rate the working alliance on a 7-point scale ranging from 1 (strong evidence against) to 7 (strong evidence) using audiotapes of the therapy sessions. An example item is: “There is consensus about the utility of the current tasks in therapy.” The total score ranges from 12 to 84 with higher scores indicating a better working alliance. The WAI-O-S has high internal consistency (α = 0.91) and good inter-rater reliability (ICC = .78) (Mcleod, 2011).
Design
This study involves a single-case design research paradigm. Before the start of treatment children were randomized to the feedback (N = 6) or control condition (N = 5). In the feedback condition therapists received weekly feedback on the child’s well-being and his or her opinion about the session. In the control condition therapists received no feedback. There were three large
assessment points: At the start of the treatment (pretreatment), at the end of treatment (post treatment) and 10 weeks after the end of treatment (follow-up). At these assessment points children and their parents received the SCARED-71 (Bodden et al., 2009) via email and the SCID-junior (Braet, et al.; unpublished manuscript) is carried out with the child and one of his or her parents face-to-face or via phone. Furthermore, the children and their parents received the SRS and TPM weekly via e-mail. Therapists of children in the feedback condition received feedback on the results of the SRS and TPM via e-mail. Furthermore, therapists in the feedback condition were asked if and how they used the client feedback during therapy. See Figure 2 for a visual representation of the design.
Figure 2. Visual representation of the research design, which displays time, the large and weekly assessment points and the used instruments at each point.
Procedure
Coding training. Coders received training for scoring therapist flexibility and child
involvement (N = 2) or working alliance (N = 4). They listened to training sessions of participants that were not included in this study and used manuals to code each session. During this process they discussed their results with each other and came to a consensus on each item. The coders for therapist flexibility and child involvement were a trained PhD student and a trained graduate student. They listened to 20 training sessions and reached a good reliability, ICC = .73 over the last 11 sessions (Cicchetti & Sparrow, 1981). The coders for working alliance were trained graduate students. They listened to 7 training sessions, which were extensively discussed before they could start coding the actual sessions.
Coding involvement, therapist flexibility, and therapeutic alliance. Coders were randomly
assigned to sessions of the same participant. It was decided to listen to consecutive sessions of the same participant to improve the reliability of detecting flexible choices of the therapist between sessions. It was chosen to listen to entire sessions instead of fragments of the session in order to prevent missing fluctuation in child involvement or cases of flexibility during the sessions. These sessions were approximately 40 to 60 minutes long. Coders made notes of examples of flexibility and child involvement or therapeutic alliance while listening to the audiotape. After the tape was finished they rated the items for either flexibility and child involvement or for working alliance. Coders rated the audiotapes individually with the use of the scoring manuals.
Data Analysis
To examine if there was a significant change in the level of anxiety after treatment as compared to before treatment, the reliable change index (RCI) was calculated. The RCI is a reliable and
commonly used method to calculate clinically significant symptom changes in case studies (Jarret & Ollendick, 2012; Versluis, Maric, & Peute, 2014). Jacobson and Truax (1991) recommend an RCI cutoff of 1.96 in standard error of the difference units to meet the criteria of “improved”. A cross-lagged correlation analysis was used to examine the relationship between therapist flexibility and child involvement, the relationship between child involvement and level of anxiety, the relationship
between therapist flexibility and level of anxiety, and the relationship between therapeutic alliance and level of anxiety during treatment. Furthermore, percentages of participants that improved or had a score under the cut off score were used to compare participants in the feedback condition to participants in the control condition and to investigate if CBT led to a decrease in the number of children with a diagnosis of SoAD.
___________________________________________________________________________
Results
Effectiveness
General effectiveness.
To assess if modularized CBT led to a significant decrease in thelevel of (social) anxiety for each participant a reliable change index (RCI) of his or her scores on the SCARED-71 was calculated. The RCI was calculated per informant (child, mother, and father) for the total score and social score on the SCARED-71 comparing pretreatment to post treatment and
pretreatment to follow-up. These results are shown in Table 2. See Table s1 in the supplementary for an overview of the mean scores and standard deviations for the total and social scale of the SCARED-71 per informant pretreatment, post treatment, and at follow-up.
Table 2
Reliable Change Index (RCI) Scores Reported for each of the Informants of each Participant comparing the Total Scores and Social Scores of the SCARED-71 between Pretreatment and Post Treatment and between Pretreatment and Follow-up
Participant Informant Pre-post Pre-FU
Total score Social score Total score Social score
1 Child Mother 1.62 -0.54 2.87 1.07 Father 3.85 0.97 6.26 0.97 2 Child 2.13 2.34 1.28 2.34 Mother 4.13 3.75 3.77 4.29 Father 2.41 2.42 3.53 2.91 3 Child Mother Father 8.67 3.39 4 Child 2.27 2.84 2.41 2.84 Mother 3.59 1.61 4.85 3.21 Father 3.85 2.91 3.05 0.97 5 Child Mother 1.79 1.61 2.87 3.5 Father 0.32 0 0.96 0 6 Child -1.56 -1.22 -0.99 -0.41 Mother 2.33 1.61 3.59 2.68 Father 0.64 0.97 7 Child 1.56 0.41 2.41 1.22 Mother 10.59 5.89 11.13 6.96 Father 8 Child Mother 1.26 1.61 2.87 3.21 Father 0.96 0.48 2.09 1.45 9 Child 1.70 0.81 2.55 0.81 Mother Father 0.16 -0.48 0.80 0 10 Child 0 -0.41 0.57 1.22 Mother 4.49 2.14 2.69 3.21 Father 4.01 1.94 6.42 3.39 11 Child 3.83 2.84 Mother 2.69 1.61 Father 2.57 5.33
Note. Bold scores indicate an improved RCI-score (RCI > 1.96).
The spaces are left blank if the RCI could not be calculated because of missing values. Pre-post = Comparison between pretreatment score and post treatment score.
Pre-FU = Comparison between pretreatment score and follow-up score.
The percentages of participants that improved on the RCI and the percentages of participants that scored below the cutoff-score of the SCARED-71 were calculated to assess the effectiveness of the treatment based on (social) anxiety symptoms. For the total scores on the SCARED-71 cutoff-scores of 30, 25, and 22 were used for children, mothers, and fathers respectively and for the social
scores cutoff-scores of 8 and 7 for children and parents were used (Bodden, Bögels, & Muris, 2009). These percentages are shown in Table 3. The results showed large differences between informants. Overall, mothers reported higher improvement on total scores and social scores of the SCARED-71 than children and fathers. At follow-up both parents had a higher percentage of total scores below the cut off score. Children had the highest percentage of social scores below the cut off score.
Table 3
Percentage of Improved Participants based on the RCI and Percentage of Participants that Scored below the Cutoff-score Post Treatment and at Follow-up
Informant Total score Social score
Percentage improved
Percentage under cut off score
Percentage improved
Percentage under cut off score
Post FU Post FU Post FU Post FU
Child 42.9 50.0 30.0 40.0 42.9 33.3 40.0 50.0
Mother 66.6 100.0 44.4 66.6 33.3 87.5 22.2 44.4
Father 55.5 75.0 20.0 66.6 33.3 37.5 30.0 44.4
Note. Improved = RCI > 1.96.
Post = post treatment, FU = follow-up.
To assess the course of the clinical diagnoses the impairment scores for SoAD on the SCID-junior pretreatment were compared to the scores post treatment and at follow-up. An overview of the impairment scores is shown in Table 4.
Table 4
Impairment Scores for Social Anxiety Disorder on the SCID-junior Reported by the Child and one of the Parents of each Participant Pretreatment, Post Treatment, and at Follow-up
Participant
Informant
Pre
Post
FU
1
Child
5
7
1
Parent
6
6.5
2
Together
6
7
2
2
Child
7
0
0
Parent
4
0
0
Together
7
0
0
3
Child
4
4
Parent
6
4
3.5
Together
6
4
4
4
Child
6
1
0
Parent
8
1
2
Together
8
1
2
5
Child
4
No diagnosis No diagnosisParent
7
No diagnosis No diagnosisTogether
7
No diagnosis No diagnosisParent
6.5
6.5
4
Together
6.5
6.5
6.5
7
Child
4
0.5
0
Parent
8
0
0
Together
8
0.5
0
8
Child
8
5
Parent
6.5
5
4
Together
6.5
8
5
9
Child
5.5
6.5
2.5
Parent
Together
5.5
6.5
2.5
10
Child
5
3
2
Parent
5.5
2
2
Together
5.5
3
2
11
Child
6
3
Parent
7
4
Together
7
4
Note. Pre = pretreatment, Post = post treatment, FU = at follow-up.
Impairment scores range from 0 to 8 with scores under 4 indicating a subclinical level of impairment and no social anxiety disorder diagnosis. The subclinical scores are displayed in bold font.
The spaces are left blank if the value was missing.
For participant 5 the impairment score was not asked during the interview, but the child did not meet the criteria for a diagnosis of social anxiety disorder.
The percentages of participants that rated their impairment score for SoAD on the SCID-junior below 4, which is the cutoff-score for a clinical diagnosis (Silverman, Saavedra, & Pina, 2001), were calculated post treatment and at follow-up to assess the effectiveness of the treatment based on the course of SoAD diagnoses. These percentages are shown in Table 5.
Table 5
Percentage of Participants that Rated their Impairment Score for SoAD on the SCID-junior as Subclinical Post Treatment and at Follow-up
Informant
Percentage subclinical (Impairment score < 4)
Post
FU
Child
45.5
70.0
Parent
50.0
70.0
Together
30.0
66.7
Note. Post = post treatment, FU = at follow-up.
Effectiveness in feedback and control condition. The therapists that treated children in the
feedback condition were asked if and how they used the feedback in their treatment. Five out of six therapists responded of which three were not able to use the feedback because the participants rarely filled out the weekly measurements. The other two therapists discussed the feedback with the child if he or she rated the session more negatively than usual. The therapists asked the child what he or she did not like about the session and changed their program based on this information. One therapist covered a topic that was not in the manual because the child wanted to discuss this. The other therapist skipped assignments that the child did not like or took it more slowly if the child was not ready for an assignment.
The percentages of participants that improved on the RCI and the percentages of participants that scored below the cutoff-score of the SCARED-71 were calculated for the feedback and control condition to compare the effectiveness of the treatment in each condition. The percentages for the total scores on the SCARED-71 are shown in Table 6 and the percentages for the social scores in Table 7. Overall, children reported higher percentages in the feedback condition than in the control condition, except for the improvement percentages on the total score at follow-up. However, parents showed no clear pattern of higher percentages in the feedback condition.
Table 6
Percentage of Improved Participants based on the RCI and Percentage of Participants that Scored below the Cutoff-score with the Total Scores of the SCARED-71 for the Feedback and Control Condition Post Treatment and at Follow-up
Informant
Percentage improved total scores
Percentage under cut off score total scores
Post
FU
Post
FU
Feedback Control Feedback Control Feedback Control Feedback Control
Child
66.7
25.0
33.3
66.7
40.0
20.0
50.0
25.0
Mother
60.0
75.0
100.0
100.0
60.0
25.0
66.7
66.7
Father
60.0
50.0
80.0
66.7
20.0
20.0
60.0
75.0
Note. Improved = RCI > 1.96.
Post = post treatment, FU = follow-up.
Table 7
Percentage of Improved Participants based on the RCI and Percentage of Participants that Scored below the Cutoff-score based on the Social Scores of the SCARED-71 for the Feedback and Control Condition Post Treatment and at Follow-up
Informant
Percentage improved social scores
Percentage under cut off score social
scores
Post
FU
Post
FU
Feedback Control Feedback Control Feedback Control Feedback Control
Child
66.7
25.0
66.7
0.00
60.0
20.0
67.7
25.0
Mother
20.0
50.0
80.0
100.0
20.0
25.0
50.0
33.3
Father
40.0
25.0
25.0
33.3
20.0
40.0
40.0
50.0
Note. Improved = RCI > 1.96.
Post = post treatment, FU = follow-up.
Feedback = feedback condition, Control = control condition
Process variables
Inter-rater reliability. Inter-rater reliability was assessed for the scores on the CIRS-R,
TTDAC-F, and WAI-O-S. For the CIRS-R and TTDAC-F both coders rated 30 percent of the sessions. A two-way mixed, single measure intraclass correlation coefficient (Shrout & Fleiss, 1979) was
calculated. Coders reached excellent reliability for the CIRS-R, ICC = .88 and fair reliability for the TTDAC-F, ICC = .46 (Cicchetti & Sparrow, 1981). The fair reliability of the TTDAC-F is not unreasonable in the exploratory phase of research and corresponds to the reliability found in the study of Chu and Kendall (2009). For the WAI-O-S 12% of the sessions was rated by all coders. They reached good reliability, ICC ranges between .69 and .76 (Cicchetti & Sparrow, 1981).
Therapist flexibility and child involvement. Cross-lagged correlation analyses were used to
assess the relationship between therapist flexibility and child involvement during treatment for each participant. Since a minimum of seven measurement points is maintained in earlier research
(Bochardt, Calhoun, Uhlmann, & Julier, 2006; Jarret & Ollendick, 2012) and there were weekly measures and audiotapes missing, the analysis could only be carried out for four of the participants. The results are presented in Table 8. The results indicate a significant positive relationship for participant 11 between therapist flexibility and child involvement 2 weeks later. See Figure s1 in the supplementary for a graphical representation of the course of therapist flexibility and child
Table 8
Cross-Lagged Correlation Analyses for CIRS-R and TTDAC-F
Participant
Cross-lagged correlations
5
-.59 (-1)
7
.33 (-3)
10
.50 (0)
11
.47 (2)*
Note. Parenthetical values indicate the significant lag or the lag with the highest correlation. * p > .05
Child involvement and anxiety. Cross-lagged correlation analyses were used to assess the
relationship between child involvement and anxiety during treatment for each participant. The analyses could be carried out for three of the participants. The results are presented in Table 9. There are no significant cross-lagged correlations between child involvement and anxiety.
Table 9
Cross-Lagged Correlation Analyses for CIRS-R and TPM
Participant
Cross-lagged correlations
5
m-.48(-3)
5
f.78 (0)
10
c.62 (4)
11
c-.27(1)
11
m-.44 (-7)
Note. c = TPM reported by the child, m = TPM reported by the mother, f = TPM reported by the father. Parenthetical values indicate the significant lag or the lag with the highest correlation.
* p > .05
Therapist flexibility and anxiety. Cross-lagged correlation analyses were used to assess the
relationship between therapist flexibility and anxiety during treatment for each participant. The analyses could be carried out for three of the participants. The results are presented in Table 10. The results indicate a significant positive relationship for participant 5 between therapist flexibility and anxiety reported by the mother in the same week. See Figure s1 in the supplementary for a graphical representation of the course of therapist flexibility and anxiety reported by the mother of participant 5 over time.
Table 10
Cross-Lagged Correlation Analyses for TTDAC-F and TPM
Participant
Cross-lagged correlations
5
m-.82(0)*
5
f.57 (1)
10
c.42 (-1)
11
c-.35(0)
11
m-.47 (-6)
Note. c = TPM reported by the child, m = TPM reported by the mother, f = TPM reported by the father. Parenthetical values indicate the significant lag or the lag with the highest correlation.
* p > .05
Therapeutic alliance and anxiety. Cross-lagged correlation analyses were used to assess the
relationship between therapeutic alliance and anxiety during treatment for each participant. The analyses could be carried out for two of the participants. The results are presented in Table 11. There are no significant cross-lagged correlations between therapeutic alliance and anxiety.
Table 10
Cross-Lagged Correlation Analyses for SRS or WAI-O-S and TPM
Participant
Cross-lagged correlations
9
c.59 (2)
10
c.43 (-6)
10
o&c-.43 (3)
Note. c = SRS and TPM reported by the child, c&o = WAI-O-S reported by the observer and TPM reported by the child. Parenthetical values indicate the significant lag or the lag with the highest correlation.
* p > .05
_________________________________________________________________________
Discussion
This study investigated the general effectiveness of modularized CBT for children with SoAD and the effectiveness of treatment if therapists received feedback. Furthermore, the relationship between therapist flexibility, child involvement, therapeutic alliance, and anxiety during treatment was examined.
To investigate the general effectiveness of the treatment the social and general anxiety
symptoms as well as the SoAD diagnoses were assessed. Overall, mothers report higher improvement in general and social anxiety symptoms than children and fathers. At follow-up both parents report subclinical general anxiety symptoms more often than children, although children report more
subclinical social anxiety symptoms than parents. These results partially support hypothesis 1a. Furthermore, approximately 2/3 of the participants were free of their SoAD diagnosis at follow-up, which supports hypothesis 1a. To assess the effectiveness of treatment if therapists received feedback, the (social) anxiety symptoms were compared between the feedback and control condition. In general, a higher percentage of children in the feedback condition reported less social anxiety symptoms than children in the control condition. These results largely support hypothesis 1b. However, parents did not report less (social) anxiety symptoms in the feedback condition compared to the control condition, which does not support this hypothesis.
For one out of four participants a positive relationship between therapist flexibility and child involvement two weeks later was found. For one out of three participants a negative relationship was found between therapist flexibility and anxiety reported by the mother in the same week. However, no significant relationships were found for the remaining participants or for the relationships between child involvement or therapeutic alliance and anxiety. These results partially support hypotheses 2a and 2c, but do not support hypotheses 2b and 2d.
The percentages of participants that do not have a SoAD diagnosis post treatment and at follow-up correspond to the percentages found in earlier studies studying the effectiveness of CBT for the overall treatment of anxiety disorders (Bodden et al., 2008a; Bodden et al., 2008b; Ginsburg et al., 2011; In-Albon & Schneider, 2007; Van Steensel & Bögels, 2015; Van Steensel, Dirksen, & Bögels, 2014). Since youth with SoAD are usually treated less effectively with CBT than children with other anxiety diagnoses, the percentages found in this study indicate a higher effectiveness of modular CBT for youth with SoAD compared to regular CBT (Crawley et al, 2009; Ginsburg et al., 2011; Hudson et al., 2015a; Hudson et al., 2015b). However, the improvement percentages of children and their fathers are lower when they report (social) anxiety symptoms. A possible explanation for this finding may be that children learn to cope with their social anxiety and therefore report low levels of impairment although they still experience anxiety symptoms. This is consistent with the aim of the Thinking Doing Daring Protocol (Bögels, 2008). Moreover, when looking at the used modules in treatment, all children received the cognitive module and coping module and almost all children received
mindfulness training. For an overview of the used modules per participant in each session see Table s2 in the supplementary. The cognitive module consists of cognitive restructuring. Children learn to disconfirm negative beliefs. In the coping module children learn helpful behaviors, such as task concentration, to cope with their anxiety. Finally, in the mindfulness module children learn to non-judgmentally focus their attention to the present moment. The content of these modules does not focus on decreasing anxiety, but on coping with anxiety or anxious thoughts. The addition of an extra module (mindfulness) that focuses on coping with rather than decreasing anxiety and the freedom of the therapists in selecting the modules, may explain why this modularized CBT is more effective in decreasing impairment in children with SoAD than CBT in previous studies (Crawley et al, 2009; Ginsburg et al., 2011; Hudson et al., 2015a; Hudson et al., 2015b) although children still experience
anxiety symptoms. Another possible explanation for the different outcomes on the questionnaire (SCARED-71) and interview (SCID-junior) may be that children give more socially desirable answers in an interview than in a questionnaire, thereby underreporting their anxiety symptoms (Schniering, Hudson, & Rapee, 2011). Earlier research indicates that children with SoAD are especially likely to respond in socially desired ways and that high levels of social desirability correlate with lower levels of self-reported anxiety (Hagborg, 1991).
Besides the difference between change in impairment versus change in anxiety symptoms, the discrepancy in the reports of the child, mother, and father on anxiety symptoms stands out. On average mothers report higher improvement on total and social anxiety symptoms than children and their fathers. Earlier studies support the informant discrepancy in the reporting of anxiety disorders (Bodden et al., 2009; De Los Reyes 2011; Pereira, Muris, Barros, Goes, Margues, & Russo, 2015). Children usually describe more anxiety symptoms than their parents (Pereira et al., 2015). This could be explained by the theory that children report internalizing disorders, such as SoAD, usually more accurately because these disorders are less visible for parents than externalizing disorders (Stallings & March, 1995). Although this could clarify the differences in improvement on anxiety symptoms and differences in the mean level of anxiety symptoms between informants post treatment and at follow-up, it does not explain the fact that the mean pre-treatment levels of anxiety symptoms of all informants are closely related or the fact that fathers report lower improvement. Bodden et al. (2009)
hypothesized that informant discrepancy in clinically anxious children differs from informant discrepancy in a non-clinical sample, since children in a clinical sample are usually referred to treatment by their parents who observed anxiety symptoms. A possible explanation could be that parents accurately observe the anxiety symptoms before treatment, but mothers overestimate the effectiveness of treatment because the seeking of or involvement in treatment created an expectancy bias. If this would be the case, mothers should have been more involved in treatment than fathers were. The data supports this assumption. In total, mothers attended twice the amount of sessions fathers did and filled in ten more weekly measurements than fathers. Moreover, parents could choose which of them participated in the interviews (SCID-junior). Only for one out of eleven participants the father instead of the mother took part in the interviews. Finally, the finding that mothers are more involved than fathers in the treatment of children with anxiety is also supported by earlier research (Duhig, Phares, & Birkeland, 2002; Lazar, Sagi, & Fraser, 1991). It could be the case that the stronger involvement of the mother in treatment created an expectancy bias for the treatment outcome, which led to mothers reporting higher improvement than children and their fathers did.
The results in this study do not provide much support for the hypothesis that the level of (social) anxiety of youth in the control condition would improve less during and after treatment than that of youth in the feedback condition. This hypothesis was built on the assumption that therapists in the feedback condition would use the feedback to adapt their treatment to the needs of the child. Only two of the six therapists with clients in the feedback condition reported that they could use the feedback in
their treatment due to missing reports. This could explain the fact that the level of anxiety in the feedback condition did not improve more during or after treatment than that of children in the control condition.
Only for one of the participants there was a significant relationship between therapist flexibility and child involvement and for another participant a significant relationship between therapist
flexibility and anxiety. There were no significant relationships either between child involvement and anxiety or therapeutic alliance and anxiety. These results do not correspond with earlier research from Chu and Kendall (2009) and Hudson et al. (2014). However, in this study the process variables were analyzed on an individual level, whereas earlier studies analyzed the variables on a group level. It may be possible that the flexibility of a therapist with a particular child or the therapeutic alliance between the child and therapist remains stable during treatment, whereas anxiety decreases, therefore the finding no significant cross-lagged correlations. This corresponds to finding of Hudson et al. (2014) that therapist flexibility and therapeutic alliance remained approximately consistent over time. These results do not contradict the possibility that children who are treated by therapists that are on average more flexible or children who have a better therapeutic alliance show higher improvement in anxiety symptoms. Another possibility for the insignificant relationships between the process variables is that therapists in this sample were not very flexible during treatment. When looking at the data, most flexibility scores approximated the mean, resulting in less variation in flexibility between the sessions, which could possibly explain the insignificant relationships.
Prior to treatment, therapists did not receive information about which modules and tasks are especially suitable for treating children with SoAD. This knowledge could have been beneficial in adapting the treatment based on the needs of children with SoAD, thereby improving the effectiveness of the treatment. However, therapists could adapt treatment based on their clinical expertise and on the needs of the children. In future research therapists should be informed beforehand on the benefits of particular tasks and modules for children with SoAD, allowing therapists to make more informed decisions in individualizing treatment. This could lead to a higher effectiveness of modularized CBT for children with SoAD,
Some of the data for the effectiveness as well as the process variables was missing. Therefore, it was not possible to involve every informant of each participant when calculating the effectiveness of treatment. Moreover, this study was a single-case study including eleven participants. Although the task force of the society of clinical psychology of the American Psychological Association advises at least 9 single case studies to give a treatment the label effective (APA, 1995), the effectiveness conclusions of this study should be considered as preliminary, because of the few participants and missing values. Future research is needed to further support the effectiveness of this treatment.
Furthermore, the cross-lagged correlation analyses could only be carried out for some of the participants, as a minimum of seven measurement points is maintained following earlier research (Bochardt, Calhoun, Uhlmann, & Julier, 2006; Jarret & Ollendick, 2012). Therefore, the conclusions
on the process variables may not be representative for all participants. Moreover, in the analyses that were carried out data points for particular sessions were still missing. Thus, the data points in the analysis were not equally divided over time. This complicates the detection of relationships between variables over time. It is possible that more significant relationships would have been found if these data points were not missing. Although it is really difficult to prevent missing data in clinical research (Yang, Li, & Shoptaw, 2008), future research with less missing values would contribute to drawing more reliable conclusions about the relationship between the process variables.
This study gives some insights into the possible future treatment of children with SoAD. Future research with larger samples and less missing data is necessary to clarify the role of the process variables in this treatment. Knowledge about which processes contribute to the effective treatment of SoAD may improve the future education of therapists. Furthermore, this study shows preliminary evidence for the effectiveness of modular CBT for children with SoAD. If these results are supported in future studies, this therapy may contribute to the effective treatment of children with SoAD in the future.
_________________________________________________________________________
References
Albano, A. M., & Kendall, P. C. (2002). Cognitive behavioural therapy for children and adolescents with anxiety disorders: Clinical research advances. International Review of Psychiatry, 14(2), 129-134.
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Washington, DC: Author.
Beesdo, K., Bittner, A., Pine, D. S., Stein, M. B., Höfler, M., Lieb, R., & Wittchen, H. U. (2007). Incidence of social anxiety disorder and the consistent risk for secondary depression in the first three decades of life. Archives of General Psychiatry, 64(8), 903-912.
Beidel, D. C., Turner, S. M., & Morris, T. L. (1999). Psychopathology of childhood social phobia. Journal of the American Academy of Child & Adolescent Psychiatry, 38(6), 643-650.
Bickman, L., Kelley, S. D., Breda, C., de Andrade, A. R., & Riemer, M. (2011). Effects of routine feedback to clinicians on mental health outcomes of youths: results of a randomized trial. Psychiatric
Services, 62(12), 1423-1429.
Bochardt, O., Calhoun, R., Uhlmann, J. K., & Julier, S. J. (2006, July). Generalized information representation and compression using covariance union. In 2006 9th International Conference on Information Fusion, (pp. 1-7). IEEE.
Bodden, D. H., Bögels, S. M., & Muris, P. (2009). The diagnostic utility of the screen for child anxiety related emotional disorders-71 (SCARED-71). Behaviour Research and Therapy, 47(5), 418-425.
Bodden, D. H. M., Bögels, S. M., Nauta, M. H., De Haan, E., Ringrose, J., Appelboom, C.,
Brinkman, A. G., & Appelboom-Geerts, K. C. M. M. J. (2008a). Child versus family cognitive-behavioral therapy in clinically anxious youth: An efficacy and partial effectiveness study. Journal of the American Academy of Child and Adolescent Psychiatry, 47, 1384-1394.
Bodden, D. H. M., Dirksen, C. D., Bögels, S. M., Nauta, M. H., De Haan, E., Ringrose, J., Appelboom, C., Bri nkman, A. G., & Appelboom-Geerts, K. C. M. M. J. (2008b). Costs and cost-effectiveness of family CBT versus individual CBT in clinically anxious children. Clinical Child Psychology and Psychiatry, 13, 543-564.
Bögels, S. M. (2008). Behandeling van angststoornissen bij Kinderen en Adolescenten met het cognitief-gedragstherapeutisch protocol Denken + Doen = Durven. Houten: Bohn Stafleu van Loghum. Boon, A., De Boer, S., & Ravestijn, E. (2012). De Child outcome rating scale (C-ORS) en de Child session
rating scale (C-SRS). Tijdschrift voor Psychotherapie, 38(2), 73-87.
Bordin, E. (1979). The generalizability of the psycho-analytic concept of the working alliance. Psychotherapy: Theory, Research and Practice, 16, 252-260.
Braet, C., Vandevivere, E., & Wante, L. (2016). SCID-Junior. Ongepubliceerd manuscript van Universiteit Gent
Chiu, A. W., McLeod, B. D., Har, K., & Wood, J. J. (2009). Child–therapist alliance and clinical outcomes in cognitive behavioral therapy for child anxiety disorders. Journal of Child Psychology and
Chorpita, B. F., Taylor, A. A., Francis, S. E., Moffitt, C., & Austin, A. A. (2004). Efficacy of modular cognitive behavior therapy for childhood anxiety disorders. Behavior Therapy, 35(2), 263-287.
Chorpita, B. F., Weisz, J. R., Daleiden, E. L., Schoenwald, S. K., Palinkas, L. A., Miranda, J., ... & Ward, A. (2013). Long-term outcomes for the Child STEPs randomized effectiveness trial: A comparison of modular and standard treatment designs with usual care. Journal of Consulting and Clinical Psychology, 81(6), 999.
Chu, B.C., & Kendall, P.C. (1999). Therapist Flexibility Scoring manual: Coping Cat Protocol Adherence Checklist–Flexibility Scale. Unpublished scoring manual. (Available from Brian C. Chu, Department of Clinical Psychology, GSAPP, Rutgers University, Piscataway, NJ 08854.)
Chu, B.C., & Kendall, P.C. (2004). Positive association of child involvement and treatment outcome within a manual-based cognitive–behavioral treatment for children with anxiety. Journal of Consulting & Clinical Psychology, 72, 821–829.
Chu, B. C., & Kendall, P. C. (2009). Therapist responsiveness to child engagement: flexibility within manual‐ based CBT for anxious youth. Journal of Clinical Psychology, 65(7), 736-754.
Cicchetti, D.V., & Sparrow, S.A. (1981). Developing criteria for establishing interrater reliability of specific items: Applications to assessment of adaptive behavior. American Journal of Mental Deficiency, 86, 127–137.
Costello, E.J., Mustillo, S., Erkanli, A., Keeler, G. & Angold, A. (2003). Prevalence and development of psychiatric disorders in childhood and adolescence. Archives of General Psychiatry, 60, 837–844.
Crawley, S. A., Beidas, R. S., Benjamin, C. L., Martin, E., & Kendall, P. C. (2008). Treating socially phobic youth with CBT: Differential outcomes and treatment considerations. Behavioural and Cognitive Psychotherapy, 36(04), 379-389.
Cummings, C. M., Caporino, N. E., Settipani, C. A., Read, K. L., Compton, S. N., March, J., ... & Ginsburg, G. (2013). The therapeutic relationship in cognitive-behavioral therapy and pharmacotherapy for anxious youth. Journal of Consulting and Clinical Psychology, 81(5), 859.
De Jong, K., Timman, R., Hakkaart-Van Roijen, L., Vermeulen, P., Kooiman, K., Passchier, J., &
Busschbach, J. V. (2014). The effect of outcome monitoring feedback to clinicians and patients in short and long-term psychotherapy: A randomized controlled trial. Psychotherapy Research, 24(6), 629-639.
De Jong, K., Van Sluis, P., Nugter, M. A., Heiser, W. J., & Spinhoven, P. (2012). Understanding the differential impact of outcome monitoring: Therapist variables that moderate feedback effects in a randomized clinical trial. Psychotherapy Research, 22(4), 464-474.
De Los Reyes, A. (2011). Introduction to the special section: More than measurement error: Discovering meaning behind informant discrepancies in clinical assessments of children and adolescents. Journal of Clinical Child and Adolescent Psychology, 40, 1–9.
Dreessen, L., Stroux, A., & Weckx, M. (1998). Nederlandse vertaling van het Gestructureerde Klinisch Interview voor DSM-IV-kind versie (KIDSCID; Versie 1.0) (Dutch translation of the Structured Clinical Interview for DSM-IV-Child edition). Maastricht, The Netherlands: Maastricht University.
Duhig, A. M., Phares, V., & Birkeland, R. W. (2002). Involvement of fathers in therapy: A survey of clinicians. Professional Psychology: Research and Practice, 33(4), 389.
