Mechanisms of change: Exploratory outcomes from a randomised controlled trial of acceptance and commitment therapy for anxious adolescents

https://doi.org/10.1016/j.jcbs.2014.09.001Get rights and content

Highlights

  • Exploratory multiple mediation analyses from an RCT of ACT for adolescent anxiety.

  • Limited support for the full hexaflex model in mediating therapeutic outcomes.

  • Mixed specific effects observed for defusion, acceptance and mindfulness/self-as-context.

  • In contrast to predictions, mediation effects were predominantly treatment-common.

  • Despite their differences ACT and CBT may be underpinned by analogous mechanisms.

Introduction

Evaluation of the efficacy of interventions has been the mainstay of clinical research for decades, generating an increasingly complex knowledge foundation of the utility of various psychotherapeutic approaches for disorder and population-specific intervention (Arch et al., 2012, Kazdin, 2007). Despite this, we are some way from establishing an empirical account for the basis of therapeutic effects – why and how even our most well-researched psychotherapies work, the processes through which interventions foster positive outcomes – typically termed “the mechanisms of change” (Ciarrochi et al., 2010, Kazdin, 2007, Kraemar et al., 2002). Identification of treatment-specific mechanisms of change has been sought to support parsimonious clinical practice, optimising clinician–patient encounters to facilitate shorter term interventions delivered with improved sensitivity and specificity (Kazdin, 2007, Kraemar et al., 2002). While mediators of change, or variables that may statistically explain the relationship between therapy and outcome, are less specific than mechanisms of change – in that they may not account for the exact process through which change occurs – understanding the factors that mediate outcomes is an important precursor to identifying mechanisms (Kazdin, 2007, Kraemar et al., 2002).

Acceptance and Commitment Therapy (ACT) is a behavioural and cognitive psychotherapy that aims to foster psychological flexibility; or the ability to respond to present moment experience of psychological phenomena, with increasing awareness, whilst engaging in value-directed behaviour (S. C. Hayes, Levin, Plumb-Vilardaga, Villatte, & Pistorello, 2013). Described as a “third wave” behavioural and cognitive therapy, ACT reflects a synthesis and reformulation of concepts underpinned by prior waves including traditional cognitive behaviour therapy (CBT). Both ACT and CBT focus on the relationship of unhelpful thoughts and beliefs to psychological distress, utilise experiential learning as well as behavioural techniques and are underpinned by behavioural theory which explains, in part, the presence of psychopathology (Forman & Herbert, 2009; S. C. Hayes, Luoma, Bond, Masuda, & Lillis, 2006). However, these therapies have been distinguished on theoretical foundations, change processes, treatment methods, and primary outcome goals (Gaudiano, 2011). CBT views psychopathology as a consequence of distorted thought patterns that are addressed in treatment through cognitive change processes of cognitive disputation and restructuring, the primary aim being symptom remission or reduction (Beck, 2005, Forman and Herbert, 2009). In ACT, psychopathology is construed as a consequence of psychological inflexibility that occurs due to entanglement or fusion with thoughts and subsequent maladaptive efforts to control internal experience (“experiential avoidance”) that leads to a decreased capacity to modify or continue exhibiting behaviours that are in the service of personal values (S. C. Hayes et al., 2006, Luoma et al., 2007). Founded upon functional contextualism, ACT focuses on the historically and situationally-defined contexts in which psychological phenomena – thoughts, feelings and sensations – occur as the target of change interventions, in contrast to the first-order change of their form or frequency, exemplified by CBT (Blackledge et al., 2009, Flaxman et al., 2011; S. C. Hayes, 2004; S. C. Hayes et al., 2011, Ruiz, 2012). Rather than emphasising symptom remission, ACT aims to foster psychological flexibility via six interrelational core processes – mediators of change – that form a “hexaflex” model; acceptance, defusion, mindfulness, self-as-context, committed action and valued living (Luoma et al., 2007). These therapeutic techniques are adopted to support more flexible responding in relation to distressing thoughts, feelings or sensations, whilst simultaneously living one׳s values, thereby enhancing quality of life (QOL; Arch and Craske, 2008, Baer, 2003, Ciarrochi and Bailey, 2008; S. C. Hayes et al., 2006, O׳Brien et al., 2008).

Anxiety disorders are among the most ubiquitous post-modern psychiatric afflictions. ACT has been found to be effective in the treatment of the range of anxiety disorders in a systematic review of 38 studies (Swain, Hancock, Hainsworth, & Bowman, 2013). A recent metaanalysis of nine ACT randomised controlled trials (RCTs) for the anxiety disorders also observed significant large effect sizes (ES) in favour of ACT relative to waitlist control and no significant ES difference relative to alternative manualised treatments (including traditional CBT) across outcome measures (Bluett, Homan, Morrison, Levin, & Twohig, 2014).

Despite the common misperception that ACT is too complex for children, it has been argued that the experiential and metaphorical delivery of ACT processes may be more suitable for children than traditional therapeutic methods such as cognitive disputation (Coyne, McHugh, & Martinez, 2011). Developmental adaptation of ACT processes has been undertaken. A systematic review of ACT in the treatment of problems among children found ACT to produce improvements in symptoms, QOL outcomes and/or psychological flexibility, with many studies demonstrating further gains at follow-up assessment (Swain, Hancock, Dixon, & Bowman, Submitted for publication). This was true for both adolescents and children as young as 6 years. This supports the conclusions of Coyne et al. (2011) – from an earlier review of the ACT literature for children – that ACT processes operate in a similar way among children and adults. Since the conduct of the most recent review, further evidence for the effectiveness of ACT in the treatment of anxiety among children has emerged. In a recent RCT of ACT versus CBT for mixed anxiety disorders, Hancock et al. (Submitted for publication) found ACT produced significant change of equivalent magnitude on clinician, parent and self-report anxiety outcome measures compared to CBT, as well as superior outcomes to waitlist control. However, relative to CBT there are comparatively fewer studies examining proposed mechanisms of change underpinning therapeutic effectiveness among anxious populations for ACT and, to date, none of the existing studies involve child populations. Despite this, one study found a significant relationship between acceptance and defusion and anxiety disorders among 111 inpatient adolescents (Venta, Sharp, & Hart, 2012). This is also in line with Coyne and colleagues׳ conclusion that child-focused studies generally support ACT׳s conceptual model in children, adolescents and parents, and that targeting processes such as acceptance and defusion are the indicated next step in research. In addition, given that children and adolescents are typically subsumed within a family system, the influence of specific factors such as family environment, parenting and emotion regulation that may impact these processes are also in need of investigation.

Laboratory-based component studies provide a controlled method of evaluating therapeutic processes of change. A recent metaanalysis of 66 studies was conducted of single-session ACT component conditions versus inactive and/or distinct alternative comparisons on a range of ACT theoretically specified outcomes (e.g. persistence/willingness to engage in a difficult task, belief in distressing cognitions and behavioural outcomes such as academic results) and other outcomes not theoretically postulated to change (Levin, Hildebrandt, Lillis, & Hayes, 2012). Results indicated some support for each of the core processes that make up the ACT hexaflex. The model as a whole was found to have a significantly greater impact on theoretically specified outcomes than inactive conditions, a finding of medium effect size. Whilst support was also identified for the hexaflex model in terms of impact on outcomes related to the intensity and frequency of negative thoughts/feelings, larger effect sizes were observed for theoretically postulated outcomes such as QOL (Levin et al., 2012).

Preliminary research in community settings offers mixed support for the ACT hexaflex model of psychological flexibility and its core component processes as mechanisms for change for the anxiety disorders (Ciarrochi et al., 2010, Forman et al., 2007; S. C. Hayes et al., 2006). Bluett et al.׳s (2014) meta-analysis of 63 studies examined the relationship between anxiety and measures of psychological flexibility. Results showed a significant medium correlation between psychological flexibility and anxiety disorder symptoms among both non-clinical and clinical samples (Bluett et al., 2014). The analysis found modest support for psychological flexibility as a mediator of change. However, mediation effects were treatment-common with no significant differences between ACT and other manualised programs (CBT) identified. For example, in one study defusion was found to be a treatment-common mediator of change in clinical worry, avoidance and QOL for ACT and CBT, but not post-treatment anxiety severity (as measured by the Anxiety Disorders Interview Schedule-IV-Revised) across treatment (Arch et al., 2012). Some evidence for treatment-specific mediation was obtained in the largest formal evaluation of mediation effects treated with ACT or cognitive therapy (CT), among 174 outpatients with anxiety/depression (Forman et al., 2012). Repeated measures of several putative mediator and outcome variables were taken with the Before Session Questionnaire (Forman et al., 2012) – a brief self-report measure that collects ratings on a Likert scale continuum with one pole reflective of CT and the other of ACT putative processes/outcomes – ahead of each therapy session. Results showed an emphasis on acceptance approaches in response to distressing psychological phenomena mediated change in symptom intensity ratings for ACT, but not CT participants (Forman et al., 2012). A movement from an emphasis on cognitive change approaches to that of acceptance across sessions was associated with reduced symptom intensity (Forman et al., 2012). Defusion and committed action were observed to be treatment-common change mediators in this study (Forman et al., 2012). Processes proposed to mediate change in CBT alone have also been found to be treatment-common to ACT such as anxiety sensitivity, dysfunctional thinking, as well as defusion in some studies (Arch et al., 2012, Forman et al., 2012). These findings highlight the need for further research examining an overarching mechanism of change across cognitive behavioural approaches for anxiety disorders.

The existing ACT mediation literature for anxiety is subject to several methodological limitations. Substantial heterogeneity has been observed in study design, sample, data collection schedule, outcomes and measurement tools, treatment protocol and statistical techniques; factors that impact the capacity to draw meaningful conclusions. Few studies have compared ACT to another active psychotherapy to determine whether proposed processes are ACT-specific (Arch et al., 2012). It is also unclear whether particular elements of ACT are more critical in terms of therapeutic outcome or whether specific techniques are more effective for disorder- or population-specific samples (Ciarrochi et al., 2010). Furthermore, little is known about whether these processes are equivalently observable among child populations or whether therapy works to affect change differently in young people. To effectively assess mediation relationships, multiple measures completed at various time points are required. Thus, the challenge for researchers is to balance the need for a breadth of psychometrically reliable and valid outcome/process measures with considerations of the acceptability and possible participant burden created by multiple repeated assessment batteries (S. A. Hayes, Orsillo, & Roemer, 2010).

The current exploratory study aimed to examine the ACT hexaflex model as a mediator for therapeutic change among adolescents. The specific indirect effects of the core processes – acceptance and defusion, mindfulness/self-as-context and valued living/committed action (valued action) – collected at multiple time points, using measures with established psychometric reliability/validity, were also explored in terms of their actual and relative contribution to mediation effects. Finally, the specificity of observed mediation effects to ACT were identified through comparison to CBT. ACT has been purported to foster psychological flexibility, thereby enhancing QOL, via the aforementioned core processes. In line with this it was hypothesised that the ACT hexaflex, and its core component processes, would operate as mediators for change in across QOL, with mediation effects expected to be treatment-specific to ACT. While clinical outcomes such as symptom remission or amelioration are not the focus in ACT, research indicates that ACT also produces change in these outcomes. As such, it was hypothesised that both QOL and clinical outcomes (main outcomes) would also be mediated by the ACT hexaflex model and its core component processes for ACT, but not CBT or waitlist control (WLC) participants. To the researchers׳ knowledge, this study is the first of its kind to simultaneously examine all core processes as putative mediators and to extend the anxiety mediation research to a sample of young people.

Section snippets

Method

Data for the present investigation were collected as part of a larger RCT of ACT versus CBT in the treatment of anxiety disorders among children aged 7–17 years. As the full methodology of this study has been previously reported (Swain et al. 2013), methodological components of the trial relevant to the present investigation are presented below.

Participants

Participants were 49 adolescent outpatients (12–17 years) diagnosed with a DSM-IV anxiety disorder and their parent/caregivers (for more details see Hancock et al., Submitted for publication). Participants were randomised into ACT (n=16), CBT (n=10) or waitlist control (WLC; n=23). Inclusion criteria required participants to have completed a minimum of 70% of treatment sessions, as well as complete data for a minimum of one anxiety outcome measure and 75% of process measures (see Measures) to

Measures

Measures included main and process outcome measures. Main outcome measures included clinical severity ratings for the principal diagnosis, anxious symptoms, total behavioural/emotional problems, depression, and QOL. Process outcome measures were incorporated on the basis of ACT putative mediator hypotheses (see Table 1). Assessment of all outcome measures were completed pre-therapy, with repeated measures undertaken immediately post (or after 10 weeks for WLC) and 3MFU for the intervention

Anxiety disorder diagnosis clinical severity (primary outcome): Anxiety Disorders Interview Schedule (ADIS-IV; Albano & Silverman, 1996)

The ADIS-IV is a structured diagnostic interview that assesses a range of DSM-IV disorders among children aged 7–17 years, incorporating the perspectives of both child and parent (Albano & Silverman, 1996). Participants endorse symptoms as either present or absent and if symptom count is sufficient to meet diagnostic criteria, a clinical severity rating (CSR) from 0 to 8, where 0 indicates no impairment and 8 indicates significant impairment, is gathered (Silverman, Saavedra, & Pina, 2001). The

Process measures

The six core processes that make up the ACT hexaflex model of psychological flexibility are interrelated and overlap (Baer, 2010; S. C. Hayes et al., 2006). Arguably, mindful awareness of psychological phenomena is required to support acceptance and defusion. Values identification is indicated in order to determine what committed actions are required to live a meaningful life, and acceptance of distressing phenomena is a precursor to foster the willingness to exhibit them. Self-as-context is an

Treatment

Participants assigned to ACT and CBT received 10×1.5 h weekly group therapy sessions using a manualised treatment programme, in accordance with the relevant therapy, delivered by psychologists. Psychologists were trained and experienced in delivering ACT and CBT (1–3 years for ACT, 2–10+ years for CBT) and provided both forms of therapy, with treatment adherence, credibility and therapist competency also assessed. Treatment fidelity was examined via analysis of videorecorded therapy sessions in

Data analysis

Data coding and analysis was conducted using the IBM SPSS Statistics v.21 software program. Preliminary linear mixed model analyses were undertaken with the Least Significance Differences method to examine group-related change in main outcome and process measures across time (pre-, post- and 3MFU). Ordinary Least Squares (OLS) regression with bootstrapping was employed to conduct exploratory mediation analyses. Residualised change scores were utilised in order to control for measurement error

Results

A comparison of pre-treatment differences between groups revealed no significant differences for any sociodemographic or anxiety outcome measure (p>.15) with the exception of ADIS-IV clinical severity ratings (CSR; p<.05). The WLC obtained a CSR that was .7 higher than the treatment groups. While this result was statistically significant it was not a clinically meaningful difference as all groups evidenced ADIS-IV pre-treatment scores in the severe range (for full details see Hancock et al.,

Changes in main outcomes across time and group

Mixed model analyses for main outcome measures were conducted among the full sample of children involved in the RCT, with age entered as a covariate. The only significant effects for age were for the ADIS-IV CSR. However, the differences were not clinically meaningful. Both treatment groups were found to produce equivalently statistically significant change for pre-post measures and in comparison to WLC of large-to-very large effect size for most measures (Hancock et al., Submitted for

Acceptance and cognitive defusion

Mixed model ANOVAs for the AFQ found significant main effects for group (F 2181.15=5.31, p<.01), time (F 2172.37=42.69), and the interaction (F 3220.54=6.59, p<.001). While there were no significant differences between ACT and CBT, both treatments produced improvements in acceptance and defusion compared with WLC (p<.001, d=.61 for ACT, d=.80 for CBT). ACT and CBT both resulted in significantly more acceptance and defusion at post (p<.001, Δ=.50 for ACT, Δ=.79 for CBT), with no significant

Summary

ACT and CBT both evidenced increased acceptance and defusion (AFQ-Y) at post, with ACT evidencing further significant within-group improvements post to 3MFU. Within group changes on mindfulness/self-as-context (CAMM-20-OBS/AWA) at post-treatment were non-significant across groups, but ACT and WLC evidenced higher scores than CBT at this time point on mindful observing (CAMM-20-OBS). No significant changes were observed for valued action (VLQ).

Process measures as mediators of treatment outcome

Results of the OLS regression with 95% bias-corrected bootstrap confidence intervals (CI) are presented in Table 3. For the primary outcome, anxiety disorder diagnosis and clinical severity as measured by the ADIS-IV CSR, the total indirect effect of all process variables varied across treatment groups. The hexaflex model mediated the relationship between treatment and CSR for ACT [point estimate (PE)=−.54; lower limit (LL)=−1.14; upper limit (UL)=−.16], but not CBT. A significant indirect

Discussion

The current study examined the ACT hexaflex model and its component core processes as putative mediators for treatment-related change in a sample of adolescents with anxiety disorders. Participants were drawn from a larger RCT of ACT versus CBT to allow assessment of the treatment-specificity of mediation effects. Both treatments produced significant changes across main outcome measures over time and in comparison to WLC (Hancock et al., Submitted for publication).

In terms of putative

First page preview

First page preview
Click to open first page preview

References (69)

  • S.C. Hayes et al.

    Acceptance and commitment therapy: Model, processes and outcome

    Behaviour Research and Therapy

    (2006)
  • R.G. Klein

    Parent–child agreement in clinical assessment of anxiety and other psychopathology: A review

    Anxiety Disorders

    (1991)
  • M.E. Levin et al.

    The impact of treatment components suggested by the psychological flexibility model: A meta-analysis of laboratory-based component studies

    Behavior Therapy

    (2012)
  • H.J. Lyneham et al.

    Interrater reliability of the anxiety disorders interview schedule for DSM-IV: Child and parent version

    Journal of the American Academy of Child and Adolescent Psychiatry

    (2007)
  • H.J. Lyneham et al.

    Psychometric properties of the children׳s anxiety life interference scale (CALIS)

    Journal of Anxiety Disorders

    (2013)
  • J.S. March et al.

    The multidimensional anxiety scale for children (MASC): Factor structure, reliability and validity

    Journal of the American Academy of Child and Adolescent Psychiatry

    (1997)
  • J.S. March et al.

    Test–retest reliability of the multidimensional anxiety scale for children

    Journal of Anxiety Disorders

    (1999)
  • C. Metzler et al.

    A randomized controlled trial of a behavioural intervention to reduce high-risk sexual behaviour among adolescents in high-risk sexual behaviour among adolescents in STD clinics

    Behavior Therapy

    (2000)
  • P. Norton

    A randomized clinical trial of transdiagnostic cognitive-behavioral treatments for anxiety disorder by comparison to relaxation training

    Behavior Therapy

    (2012)
  • M.A. Rynn et al.

    The psychometric properties of the MASC in a pediatric psychiatric sample

    Anxiety Disorders

    (2006)
  • W.K. Silverman et al.

    Test–retest reliability of anxiety symptoms and diagnoses with the anxiety disorders interview schedule for DSM-IV: Child and parent versions

    Journal of the American Academy of Child and Adolescent Psychiatry

    (2001)
  • J. Swain et al.

    Acceptance and commitment therapy in the treatment of anxiety: A systematic review

    Clinical Psychology Review

    (2013)
  • R.K. Wicksell et al.

    Exposure and acceptance in the rehabilitation of adolescents with idiopathic chronic pain – A pilot study

    European Journal of Pain

    (2007)
  • T.M. Achenbach et al.

    National survey of problems and competencies among four- to sixteen-year-olds: Parents׳ reports for normative and clinical samples

    Monographs of the Society for Research in Child Development

    (1991)
  • A.M. Albano et al.

    Anxiety Disorders Interview Schedule for DSM-IV Child Version: Clinician׳s manual

    (1996)
  • J.J. Arch et al.

    Acceptance and commitment therapy and cognitive behavioral therapy for anxiety disorders: Different treatments, similar mechanisms?

    Clinical Psychology: Science and Practice

    (2008)
  • R. Baer

    Mindfulness training as a clinical intervention: A conceptual and empirical review

    Clinical Psychology: Science and Practice

    (2003)
  • R. Baer

    Assessing mindfulness and acceptance processes in clients: Illuminating the theory and practice of change

    (2010)
  • R. Baer et al.

    Using self-report assessment methods to explore facets of mindfulness

    Assessment

    (2006)
  • A.T. Beck

    The current state of cognitive therapy: A 40-year retrospective

    Archives of General Psychiatry

    (2005)
  • J.T. Blackledge et al.

    Acceptance and commitment therapy: Contemporary theory, research and practice

    (2009)
  • C.A. Burke

    Mindfulness-based approaches with children and adolescents: A preliminary review of current research in an emergent field

    Journal of Child and Family Studies

    (2010)
  • M.P. Carey et al.

    Children׳s depression inventory: construct and discriminant validity across clinical and nonreferred (control) populations

    Journal of Consulting and Clinical Psychology

    (1987)
  • J. Ciarrochi et al.

    A CBT practitioner׳s guide to ACT: How to bridge the gap between cognitive behavioural therapy and acceptance and commitment therapy

    (2008)
  • Cited by (29)

    • A Review of Acceptance and Commitment Therapy for Adolescents: Developmental and Contextual Considerations

      2024, Cognitive and Behavioral Practice
      Citation Excerpt :

      Short (8 items) and long (17 items) forms of the AFQ-Y are available and established as youth-friendly measures of psychological flexibility. In terms of specific ACT processes, the Valued Living Questionnaire (VLQ) was adapted for assessment of values in adolescents using two 10-item subscales of importance and consistency around valued behaviors in the week prior to assessment (Swain et al., 2015). Present-moment awareness and self-as-context have been measured in previous youth samples using the Child and Adolescent Mindfulness Measure (CAMM)—this 20-item self-report has subscales of Observing and Acting With Awareness for a more specific assessment of present-moment awareness (Ciarrochi et al., 2011).

    • In the shadow of COVID-19: A randomized controlled online ACT trial promoting adolescent psychological flexibility and self-compassion

      2023, Journal of Contextual Behavioral Science
      Citation Excerpt :

      A promising approach in promoting psychological flexibility and self-compassion and offering adaptive skills to deal with adverse events is acceptance and commitment therapy (ACT; Hayes et al., 2012). According to meta-analyses, ACT may improve mental and behavioral problems, quality of life, and psychological flexibility in adolescents (Fang & Ding, 2020; Swain et al., 2015). In addition, before the COVID-19 pandemic, our studies suggest that web-based ACT can help increase adolescents’ academic buoyancy and life satisfaction and reduce stress and depressive symptoms (Lappalainen et al., 2021; Puolakanaho et al., 2019).

    • The efficacy of group-based acceptance and commitment therapy on psychological capital and school engagement: A pilot study among Chinese adolescents

      2020, Journal of Contextual Behavioral Science
      Citation Excerpt :

      However, the research on ACT in children demonstrates promise. It appears ACT is effective in treating children's anxiety (Hancock et al., 2018; Swain, Hancock, Dixon, et al., 2015; Swain et al., 2015), depression (Ataie Moghanloo, Ataie Moghanloo, & Moazezi, 2015; Hayes, Boyd, & Sewell, 2011) and other mental and behavioral disorders (Azadeh, Kazemi-Zahrani, & Besharat, 2016; Livheim et al., 2015; Whittingham, Sanders, McKinlay, & Boyd, 2014; Wicksell, Melin, Lekander, & Olsson, 2009). Fewer studies focused on the positive mental factors and behaviors of children with randomized designs or control comparisons.

    • Acceptance and Commitment Therapy in group format for anxiety and depression. A systematic review

      2020, Journal of Affective Disorders
      Citation Excerpt :

      With regard to the mechanisms responsible for change, two studies report that changes in psychological flexibility mediate the improvement in anxiety/depression (Bohlmeijer et al., 2011; Zettle et al., 2011). Swain et al. (2015) find that changes in psychological flexibility mediate improvements in anxiety, both in ACT and in CBT. In contrast, Zettle et al., 2011 find that the greater change in cognitive defusion (a component of psychological flexibility) in ACT, in comparison to CT, mediates better results in depression.

    View all citing articles on Scopus
    View full text