Acceptance and Commitment Therapy group for treatment-resistant participants : A randomized controlled trial

Acceptance and Commitment Therapy (ACT) is a theoretically coherent approach addressing common processes across a range of disorders. The aim of this study was to investigate the effectiveness of a group-based ACT intervention for “treatment-resistant” participants with various diagnoses, who had already completed at least one psychosocial intervention. Of 61 individuals randomized into a servicebased trial comparing ACT and Treatment as Usual based on Cognitive Behavior Therapy (TAU-CBT), 45 provided data (ACT n1⁄426; TAU-CBT n1⁄419). Primary outcomes were measures of psychological symptoms. All participants showed reduced symptoms immediately after intervention but improvements were more completely sustained in the ACT group at 6-month follow-up. More elaborate and more fully controlled evaluations are required to confirm the findings, improve understanding of ACT processes and assess health economic benefits. & 2014 Published by Elsevier Inc. on behalf of Association for Contextual Behavioral Science.

Clients with these multi-diagnostic presentations may also be categorized as "treatment-resistant" if they meet certain criteria for relapse or chronicity. For example, Kenny and Williams (2007) defined treatment resistance in relation to depression as having had three or more previous episodes, or one chronic episode lasting 1 year or more. Such clients consume a disproportionate amount of clinical resources (Amsterdam, Hornig, & Nierenberg, 2001;Crown et al., 2002;Russell et al., 2004) but, ironically, they are sometimes excluded from clinical trials to reduce variability and thus increase internal validity (Persons & Silberschatz, 1998;Westen, Novotny, & Thompson-Brenner, 2004;Zarin, Young, & West, 2005).
Fortunately, there is some evidence that "third wave" (Hayes, 2004) forms of behavior therapy incorporating principles of mindfulness can successfully treat complex and intransigent clinical problems such as chronic or recurrent depression and personality disorder (e.g., Lynch, Trost, Salsman, & Linehan, 2007;Ma & Teasdale, 2004;Segal, Williams & Teasdale, 2002). Acceptance and Commitment Therapy (ACT; Hayes, Strosahl & Wilson, 1999) has figured prominently amongst these new forms. ACT's central focus, distinguishing it from other psychotherapies, is the notion that a broad range of psychological difficulties, typically viewed as distinct under DSM-IV, emerges from our capacity for human language (Hayes et al., 2004a;Wilson, Hayes, Gregg, & Zettle, 2001). ACT uses several therapeutic techniques to increase psychological flexibility by undermining unhelpful verbal representations of experience, encouraging a present moment focus, and promoting action consistent with long-term values (Hayes, Luoma, Bond, Masuda, & Lillis, 2006). Because ACT aims to teach generic positive psychological skills, rather than targeting specific unwanted experiences and feelings (Lundgren, Dahl, & Hayes, 2008), it is inherently transdiagnostic (Hayes et al., 2006). The central role that ACT accords to values may facilitate motivation in treatment-resistant clients with long-standing problems. This shift in focus may be especially salient to clients more familiar with forms of therapy that emphasize symptomatic relief.
Following Hayes et al. (2006), some recent studies have attempted to evaluate ACT with heterogeneous client groups. For example, Fledderus, Bohlmeijer, Pieterse, and Schreurs (2012) compared two versions of an ACT self-help intervention (minimal vs. substantive email support) with a waiting list control, using a sample of individuals experiencing depression and/or anxiety. Improvements in each experimental condition exceeded those for controls, and were sustained at 3 month follow-up. Forman, Herbert, Moitra, Yeomans, and Geller (2007) also compared ACT with Cognitive Therapy for individuals with severe anxiety and/or depression. They kept their "exclusion criteria purposefully broad for maximum external validity" (p. 8). Both groups improved on a range of measures, but the mechanisms of action appeared to differ in ways consistent with the underlying therapeutic models. Finally, Lang et al. (2012) have described a protocol for an ongoing randomized controlled trial designed to compare ACT with psychotherapy for military veterans exhibiting high levels of comorbidity.
We recently conducted a treatment development trial to test the utility of ACT with a heterogeneous group of clients in a naturalistic clinical setting (Clarke, Kingston, Wilson, Bolderston, & Remington, 2012). Ten participants presented: All had clinical disorders (Axis I diagnoses), and half met criteria for co-morbid personality disorders (Axis II diagnoses). The inclusion criteria specified that participants had "already received at least one previous episode of therapy, for which they attended at least eight sessions, and … were being re-referred with significant residual mental health concerns" (p. 562; emphasis in the original). This criterion for treatment resistance differs from that of Kenny and Williams (2007) because (a) it is not specific to depression, and (b) it requires the client to have had psychological treatment. In fact, group members had on average attended a mean of 3.5 previous psychological interventions. After 16 ACT-based group sessions using a protocol adapted from Hayes and Smith (2005), the group showed significant improvements in self-reported depression, overall symptomatology and quality of life, with medium to large effect sizes on all measures. Individual analyses showed clinically significant and reliable change in up to 70% of participants. Moreover, significant improvements over baseline were maintained at 6 and 12 month follow-up. These findings, although promising, are tentative because the study was uncontrolled and the small sample of clients limited the power to detect differences. To increase external validity, replication using a randomized controlled trial (RCT) design is required.
Thus, building on previous research, the aim of the present study was to assess the effectiveness of ACT for a heterogeneous group of treatment-resistant clients. Following the recommendations of Smout, Hayes, Atkins, Klausen, and Duguid (2012), we used an active CBT-based control group for comparison purposes. We designated this control condition as Treatment as Usual based on CB (TAU-CBTT) because CBT is the most widely utilized and researched psychotherapy (Norcross, Hedges, & Castle, 2002). Although most CBT research is disorder-specific, the use of unified treatment protocols for mood and anxiety disorders is now being explored by some investigators (McEvoy, Nathan, & Norton, 2009). CBT-based TAU was therefore considered an appropriate active comparison condition against which to evaluate ACT.
In keeping with previous ACT research, we chose to conduct a small-scale RCT utilizing a pragmatically acquired sample. We chose our primary outcome measures to reflect the heterogeneous and treatment-resistant characteristics of this sample, and the goals of ACT-based intervention. Because our selection criterion did not specify any diagnosis, we assessed participants' overall symptomatology. Because treatment resistance is strongly associated with depression (Blom et al., 2007;Fournier et al., 2009;Joyce et al., 2002;Sotsky et al., 1991), we used a robust measure of depressive symptomatology. Our secondary outcome measures were chosen to monitor changes in complex personality disorder symptomatology and quality of life; the latter is a nonsymptomatic measure expected to vary with personal adjustment. Assessments were made post-therapy and at 6 month follow-up, to assess whether any benefits had been sustained. Furthermore, because inappropriate care can worsen symptoms and personality pathology for treatment-resistant clients (Clarke, Thomas, & James, 2013;Tyrer & Simmonds, 2003), we assessed changes in participant functioning on both an individual and a group basis. We further assessed preliminary theory-driven process variables for both conditions.
We hypothesized that participants receiving ACT would show greater improvements in primary outcome measures across time than those receiving TAU-CBT-. We also predicted that a greater proportion of ACT participants would improve and that less would deteriorate than CBT-based TAU participants.

Design
We used an RCT to compare the effectiveness of a 16 week, group-based ACT intervention with a group-based TAU-CBT intervention of the same duration at a specialist personality disorder clinic in a public health setting (ISRCTN17801606). The Dorset Research and Development Support Unit assigned participants to treatments using block randomization (block sizes 2-4) to ensure the numbers allocated to each intervention were always closely balanced. Outcome and process measures were obtained at baseline, post-therapy and follow-up. The protocol was approved by the UK National Health Service Research Ethics Committee (Dorset: 06/Q2201/170).

Participants
Participants were recruited from referrals to a Community Mental Health Team and a specialist outpatient service for people with a personality disorder. Consistent with our earlier definition of treatment resistance (Clarke et al., 2012), eligible participants had received at least one previous 8-session episode of psychological therapy and had been re-referred. No independent data were available on the quality or fidelity of previous treatments. Owing to the group-based nature of the intervention, and the relative vulnerability of the client group, exclusion criteria (based on DSM-IV, 1994), were intellectual disability, schizophrenia or other psychotic illness, or any of the following high risk behaviors: (a) current drug or alcohol dependency; (b) a current eating disorder and a BMI of o16; and (c) deliberate self-harm in the past 6 months (defined using Kreitman's (1977) criteria). In keeping with the service protocol, clients who engaged in selfharming behavior monthly or more were referred directly to an established Dialectical Behavior Therapy program.

Primary outcome measures
The Symptom Check List-90 Revised (SCL-90-R; Derogatis, 1993), a 90-item self-report measure, was used to measure psychiatric distress. Given the symptomatic heterogeneity of the conditions, we