Mediation Effects of Group Cohesion in Group-Based Cognitive Behavioral Therapy for Youth Anxiety Disorders

ABSTRACT

There is limited research on group cohesion as a potential outcome facilitator in group-based cognitive-behavioral treatment (GCBT) for youth. We examined if group cohesion mediated the relation between the temperamental trait behavioral inhibition and posttreatment outcomes following GCBT for youth with anxiety disorders. The sample comprised 88 youth (M age = 11.2 years) from a randomized controlled effectiveness trial. The outcomes were posttreatment clinical severity and treatment satisfaction. Group cohesion fully mediated the relation between behavioral inhibition and posttreatment severity. Higher group cohesion was associated with lower posttreatment clinical severity. There was no significant association between behavioral inhibition and treatment satisfaction, hence no mediation. We conclude that group cohesion is a factor that can be targeted by clinicians to potentially enhance GCBT outcomes.

Higher group cohesion, defined as the bond between treatment group members, is associated with better outcomes of group-based cognitive behavioral therapy (GCBT) in a review of studies for adult clients with anxiety disorders (21% explained variance; Luong et al., 2020). Compared to the adult field, group cohesion has been less studied with youth clients with anxiety. One study found that higher youth-rated group cohesion was associated with larger pre/post anxiety reduction in support groups in a nonclinical sample of youth exposed to traumatic events (Shechtman & Mor, 2010). A recent study of youth with anxiety disorders showed that higher observer-rated group cohesion measured late (i.e., > session 6) in a 10-session GCBT program predicted lower posttreatment clinical severity ratings and higher treatment satisfaction, but not anxiety symptoms (Fjermestad et al., 2023). Thus, there is some, albeit very limited evidence that group cohesion influences GCBT outcomes for youth with anxiety problems.

In the current report, we examined the outcomes that were predicted by group cohesion in Fjermestad et al. (2023), i.e. clinical severity and client satisfaction. Examining more than “pure” group cohesion and posttreatment outcome associations, we added data on behavioral inhibition, which refers to children’s temperamental pattern of behavioral and emotional responses when facing new and unfamiliar people, objects, places or situations (Kagan et al., 1984). The evidence for a link between the behavioral inhibition and anxiety outcomes is based on preschool samples (Hirshfeld-Becker et al., 2010; Morgan et al., 2018) or specific phobias only (Capriola et al., 2017). Based on longitudinal studies following children from toddlerhood to later childhood, behavioral inhibition is considered a moderately stable trait (Degnan & Fox, 2007; Muris et al., 2011). High behavioral inhibition has been linked to anxiety and has for example been identified as one of the largest single risk factors for developing social anxiety disorder (Clauss & Blackford, 2012). In a longitudinal study, Hirshfeld et al. (1992) found that children with stable high behavioral inhibition had a higher prevalence of multiple anxiety disorders.

Given the evidence that behavioral inhibition and group cohesion both are likely to influence outcomes, in the current study we move beyond pure associations between these variables and focus on potential mediation effects, to provide the field with more comprehensive knowledge on outcome predictors. More mediation studies have been called for in the pursuit of factors that can enhance outcomes for youth with anxiety disorders (Creswell et al., 2020). We take advantage of the timing of measures that allows testing of mediation. That is, behavioral inhibition was measured before treatment, group cohesion was measured during treatment, and the outcomes were measured posttreatment. This original approach helps move the field forward by disentangling the role of pretreatment potential predictors versus processes measured during treatment. An additional original feature is the inclusion of outcome data other than symptoms, i.e., the inclusion of treatment satisfaction, which is an important implementation element in evidence-based practice and personalized treatment (Stumpp & Sauer-Zavala, 2022). The research question was: Does group cohesion mediate the association between behavioral inhibition and posttreatment outcomes? Given that group cohesion was measured closer in time to outcomes than behavioral inhibition, we expected full or partial mediation based on evidence from adult studies (Luong et al., 2020).

METHODS

Sample

The sample represents the full GCBT arm of a randomized controlled trial (RCT) comparing GCBT and individual CBT to waitlist in community youth mental health outpatient clinics (Wergeland et al., 2014). The sample comprised 88 youth (M age at pretreatment = 11.7 years, SD = 2.1, age range 8–15, 55% girls, 45% boys). The youth identified as European White (90.7%) and Asian (1.6%; 7.7% missing ethnicity). Family social class was high 35.2%, medium 51.1%, low 5.7%, and missing 8.0%; (Currie et al., 2008). The inclusion criteria in the RCT were that the child should be (a) regularly referred to the outpatient clinic and (b) meet the criteria for a primary DSM-IV (American Psychiatric Association, 1994) diagnosis of a separation anxiety disorder (SAD; 33.0% in the current GCBT sample), social phobia (SoP; 46.5%), or generalized anxiety disorder (GAD; 20.5%). The exclusion criteria were intellectual disability, pervasive developmental disorder, psychotic disorder, and/or mental retardation. In the current GCBT sample, 78.4% had at least one registered comorbid disorder, of which the most common were specific phobia or other anxiety disorders (17.0%), depression (15.4%), tic disorder (6.8%), and/or attention deficit hyperactivity disorder (5.5%).

Clinicians and Treatment

Clinicians (N = 15, M_age = 49.8 years; SD = 9.4; 93.3% female; 100% European White) volunteered to participate and were employed at one of the seven clinics. On average, clinicians had 12.0 years of experience (SD = 6.0). Two clinicians led each group, with one designated as the group leader and the other as the facilitator. The sample comprised 16 treatment groups, with 4 to 7 youth in each group (median = 5).

The treatment program was the FRIENDS for Life program (FRIENDS; 4th ed., Barrett, 2004), which is targeting emotional awareness and regulation, cognitive restructuring, and exposure tasks. The program is theoretically and empirically based on the basic principles of CBT, proposing that changed behavior is facilitated by emotion regulation (e.g., breathing exercises) and cognitive restructuring (e.g., thinking “I can try” instead of “This will never work”). The program comprised 10 weekly 60-min sessions and two booster sessions. Clinicians showed adequate adherence and competence (Wergeland et al., 2014).

Original Trial Outcomes

The percentages of participants who no longer met criteria for their primary anxiety diagnosis in the intent-to-treat sample was 35.2% at posttreatment, 46.6% at 12 months follow-up, and 59.4% at 4 years follow-up. Clinical severity scores (scored on a 0–8 scale) were significantly reduced with 2.3 points from pre- to posttreatment, 3.2 points from pre- to 12 months follow-up, and 4.4 points from pre- to 4 years follow-up (Kodal et al., 2018).

MEASURES

Observer-Rated Measures

The Therapy Process Observational Coding System for Child Psychotherapy—Group Cohesion Scale (TPOCS-GC; Fjermestad et al., 2023; Lerner et al., 2013). The observer-rated TPOCS-GC was used as a measure of group cohesion. We used a four-item version of the TPOCS-GC that has been psychometrically evaluated with a youth sample (Fjermestad et al., 2023). The items cover the extent that the group member demonstrate positive affect toward the other group members (affect); shares their experiences with the other group members (share), and to what extent the interaction between the group member and the other group members is lively and energetic (energy). One item is reverse coded, i.e., extent that the group member appears anxious or uncomfortable with the other group members (anxious). The items are rated on a 6-point scale ranging from 0 (not at all) to 5 (a great deal). Coders watch entire treatment sessions, shifting their main focus between each group member every five minutes but providing an average score per session per client. The TPOCS-GC has evidence of internal consistency and coder agreement, e.g. internal consistency of α = .80 and mean interrater agreement of ICC (1, 2) = .75 to .88 in parent groups (Lerner et al., 2013; Luong et al., 2021). The current sample represents the first use of the TPOCS-GC with youth clients. The coders were two advanced clinical psychology students who were trained by one of the TPOGS-GC developers (MDL). The average inter-rater agreement was ICC (2, 1) = .61. The internal consistency across the four items was (α = .72).

Outcome Measures

Clinicians’ Severity Ratings (CSR; Silverman & Albano, 1996). The clinician-rated CSR is based on the Anxiety Disorders Interview Schedule—Child and Parent versions (ADIS-C/P; Silverman & Albano, 1996). For each ADIS-C/P section (i.e., SAD, SOP, and GAD)., a clinician’s severity rating (CSR) ranging from 0 to 8 was assigned based on combined parent and child report (ADIS-CSR). A higher score reflects a higher impact on personal, academic, social, and family functioning. Intraclass correlations for the CSR were 0.82 (ADIS-C) and 0.82 (ADIS-P). The ADIS-C/P is widely used in clinical and research settings (Creswell et al., 2020). The interrater agreement estimated by kappa (κ) for the presence of an inclusion anxiety diagnosis was 0.84 (ADIS-C) and 0.86 (ADIS-P) based on masked rescoring of 20% of tapes conducted by clinical experts trained by the ADIS-C/P developer (WKS).

The Client Satisfaction Scale (CSS; Ollendick et al., 2009). The youth-rated CSS was used to measure satisfaction with treatment with 10 items (e.g., How successful do you think this program was in teaching you to deal with your fears?) rated on a Likert scale ranging from 1 to 5. The CSS has demonstrated discriminant validity between various treatment forms (Ollendick et al., 2009). In the current trial, the CSS was administered at posttreatment and internal consistency was α = .78.

Procedures

The study took place in Western Norway (2014–2018) and was approved by the Western Norway Regional Committee for Medical and Health Research (# 3.2007–2322). Parents and youth above 12 years provided written informed consent/assent and youths below 12 years provided verbal assent. No honorarium was given.

Data Analytic Plan

To test potential mediation effects, we tested two sets of models, with the independent variable (behavioral inhibition) measured at baseline, the potential mediator (group cohesion) measured during treatment, and the dependent variable (clinical severity and client satisfaction, respectively) measured at posttreatment. We used linear regression models in SPSS version 29. We first examined the direct associations between the independent variable, the mediator, and the dependent variable, before examining the associations between the independent variable and the dependent variable while controlling for the mediator. In terms of missingness, we used all available data for each participant. Missing values ranged from 0.0% to 18.2% across the variables. Analyses showed these were missing completely at random (p = .185; Little, 1988). The main reason for missingness was dropout from GCBT (n = 9; 10.2%), defined as missing at least three consecutive sessions, or not being present in the sessions coded with the TPOCS-GC (n = 5; 5.7%).

RESULTS

Figure 1a shows that there were significant associations between (a) behavioral inhibition and group cohesion, (b) group cohesion and clinical severity, and (c) behavioral inhibition and clinical severity, thus meeting the requirement for potential mediation. Higher behavioral inhibition was associated with lower group cohesion and higher clinical severity. Higher group cohesion was associated with lower clinical severity

Figure 1a. Direct Effects of Behavioral Inhibition and Group Cohesion on Clinical Severity at Posttreatment.

Figure 1b shows the mediation model. The significant association between behavioral inhibition and clinical severity disappeared when group cohesion was controlled for, indicating full mediation.

Figure 1b. Mediation Model: Controlling for Group Cohesion on the Effect Between Behavioral Inhibition and Clinical Severity at Posttreatment.

Figure 2 shows that there were no significant association between behavioral inhibition and client satisfaction. Hence, the requirement for potential mediation was not met and the model for client satisfaction as outcome was not examined further.

Figure 2. Direct Effects of Behavioral Inhibition and Group Cohesion on Client Satisfaction.

BIQ = Behavioral Inhibition Questionnaire. GC = group cohesion. CSR = clinicians´ severity rating (from the Anxiety Disorders Interview Schedule). CSS = Client Satisfaction Scale.

DISCUSSION

We built on the emerging evidence that group cohesion and behavioral inhibition are associated with anxiety treatment outcomes (e.g., Fjermestad et al., 2023; Morgan et al., 2018). For the first time, we included both factors in one analysis, examining if group cohesion mediated the relation between behavioral inhibition and outcomes. We found that group cohesion mediated the relation between behavioral inhibition and clinical severity at posttreatment. This is an important finding, in that it shows that treatment-related processes such as group cohesion can “overrule” the effects of more stable temperamental traits associated with anxiety. This is a promising finding, as group processes may be easier to amend than internal traits.

There was no direct association between behavioral inhibition and the client satisfaction outcome, hence a potential mediation model for this outcome was not supported. This may be explained by the fact that client satisfaction is conceptually different from symptom-related outcomes, like clinical severity, which the limited existing evidence is based on (Batbaatar et al., 2015). Whereas treatment effects can be considered part of a wider treatment satisfaction concept, an important difference in the current study is that whereas clinical severity was clinician-derived based on combined youth and parent report, the satisfaction measure was purely youth-rated. The fact that there was no direct association between behavioral inhibition and client satisfaction may be that other variables, such as practicalities, location, timing of sessions, and/or other factors we did not measure, are more strongly linked to satisfaction as a wider outcome than clinical severity.

The current study’s main strength is that the mediator was measured between the independent and the dependent variable in time. The limitations include the fact that the RCT nevertheless not was designed as a mediation study, that we only included anxiety disorders, and that the sample was homogeneously European-White.

The main implication of the current report is that group cohesion appears to be a treatment-related factor that can be targeted by clinicians to enhance outcomes. There is, however, currently no available evidence of specific techniques that GCBT clinicians can apply to enhance outcomes, and research into this represents an important next step for the field. This research could be inspired by work that has been done in the alliance field, where some papers have identified specific therapeutic behaviors that negatively (e.g., over-structuring sessions) or positively (e.g. using positive reinforcement, emphasizing collaboration) influences the alliance (Creed & Kendall, 2005; Fjermestad et al., 2021; Karver et al., 2008). Other therapist behaviors may be involved in enhancing group cohesion, but the behaviors identified by alliance research represent a useful starting point for observing therapist behaviors that could potentially enhance group cohesion. The multirelational element of group cohesion relative to the dyadic alliance concept also warrants research into how other group members act to facilitate (or not facilitate) group cohesion. Such research represents means toward enhanced knowledge about optimal group-based interventions for youth with anxiety disorders.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the Western Norway Health Authority under Grant numbers [911366 and 911253].

Notes on contributors

Krister W. Fjermestad

Krister W. Fjermestad is a Clinical Psychology Professor at the Department of Psychology, University of Oslo, Norway, and a practicing clinician at Frambu resource centre for rare disorders. Frederike Naujokat and Malin Wallin are practicing clinical psychologist who undertook this research work during their professional training at the Department of Psychology, University of Oslo, Norway. Gro Janne Wergeland is a Psychiatry Professor at the Department of Medicine, University of Bergen and a practicing Child and Adolescent Psychiatrist at Åsane Child and Adolescent Mental Health Services, Haukeland University Hospital, Norway.