Randomized clinical trial of adapted mindfulness-based stress reduction versus group cognitive behavioral therapy for heterogeneous anxiety disorders

Abstract

Objective

To compare a mindfulness-based intervention with cognitive behavioral therapy (CBT) for the group treatment of anxiety disorders.

Method

One hundred five veterans (83% male, mean age = 46 years, 30% minority) with one or more DSM-IV anxiety disorders began group treatment following randomization to adapted mindfulness-based stress reduction (MBSR) or CBT.

Results

Both groups showed large and equivalent improvements on principal disorder severity thru 3-month follow up (ps < .001, d = −4.08 for adapted MBSR; d = −3.52 for CBT). CBT outperformed adapted MBSR on anxious arousal outcomes at follow up (p < .01, d = .49) whereas adapted MBSR reduced worry at a greater rate than CBT (p < .05, d = .64) and resulted in greater reduction of comorbid emotional disorders (p < .05, d = .49). The adapted MBSR group evidenced greater mood disorders and worry at Pre, however. Groups showed equivalent treatment credibility, therapist adherence and competency, and reliable improvement.

Conclusions

CBT and adapted MBSR were both effective at reducing principal diagnosis severity and somewhat effective at reducing self-reported anxiety symptoms within a complex sample. CBT was more effective at reducing anxious arousal, whereas adapted MBSR may be more effective at reducing worry and comorbid disorders.

Introduction

Several decades ago, Kabat-Zinn et al. conducted the first investigation of Mindfulness-Based Stress Reduction (MBSR; 1990) for anxiety disorder patients (1992). The small open trial (n = 22) demonstrated significant reductions in anxiety, panic, and depressive symptoms, and a subsequent report showed maintenance of treatment gains three years later (Miller, Fletcher, & Kabat-Zinn, 1995). These nascent findings inspired interest in mindfulness-based interventions for anxiety disorders.

Despite its early promise, only two studies have investigated MBSR for anxiety disorders within randomized clinical trial designs. The first one randomized generalized social anxiety disorder patients to MBSR or group cognitive behavioral therapy (GCBT) (Koszycki, Benger, Shlik, & Bradwejn, 2007). Findings supported the general efficacy of both treatments, although GCBT outperformed MBSR on social anxiety outcomes, response and remission rates. The second study compared MBSR to a waitlist control for heterogeneous anxiety disorder patients (Vollestad, Sivertsen, & Nielsen, 2011). Among completers, results showed large effect size improvements in anxiety and depression that endured through 6-month follow up for MBSR relative to the waitlist control group. Several additional studies (Craigie, Rees, Marsh, & Nathan, 2008; Evans et al., 2008; Roemer, Orsillo, & Salters-Pedneault, 2008) have assessed mindfulness-based interventions other than MBSR for a variety of anxiety disorders; results have indicated moderate to strong success. A recent meta-analysis (Hofmann, Sawyer, Witt, & Oh, 2010) demonstrated that mindfulness-based interventions, including MBSR, reduced anxiety symptoms across a variety of psychiatric and medical populations (effect size = .63), and especially in the subgroup of patients with anxiety and mood disorders (effect size = .97), supporting the notion that MBSR may be particularly effective at reducing anxiety.

Other than Koszyski et al. (2007), however, no studies have compared a mindfulness-based intervention for anxiety disorders to another active treatment. Cognitive behavioral therapy (CBT) represents the most evidence-based psychotherapy for anxiety disorders in both individual and group formats (Butler, Chapman, Forman, & Beck, 2006; Hofmann & Smits, 2008; Norton & Price, 2007; Tolin, 2010), and therefore, the gold standard against which to assess the efficacy of an alternative intervention such as MBSR. Directly comparing MBSR and CBT provides an opportunity to investigate different approaches to anxiety-related thoughts (mindfully observing and accepting thoughts in MBSR versus reappraising and modifying thought content in CBT) and emotions (mindfulness observing and making space for emotions in MBSR versus controlling and reducing emotions in CBT). Whereas MBSR encourages moving toward uncomfortable internal experiences by promoting a stance of openness, curiosity and acceptance, it does not utilize formal behavioral exposure procedures, as in CBT. Further, CBT focuses on treating anxiety disorder symptoms whereas MBSR focuses more broadly on redirecting participants' attention toward the present moment and shifting their overarching relationship with thoughts, feelings, and current experience. Thus, MBSR represents a broader set of strategies for dealing with internal experience that might consequently impact broader symptom outcomes (beyond anxiety). Moreover, outside of group CBT, clinicians have few evidence-based group treatments for anxiety disorders from which to choose. If MBSR is effective for anxiety disorders, then clinicians will have more options for evidence-based group treatments.

On these bases, we compared an adapted version of MBSR and group CBT for the treatment of heterogeneous anxiety disorders. We integrated several features of hybrid efficacy-effectiveness study designs to maximize external validity (Chambless & Hollon, 1998) by conducting the study within a real-world treatment clinic and utilizing minimal exclusion criteria. Therefore, our study differed from Koszyski et al.'s (2007) in that we targeted all anxiety disorders rather than social anxiety disorder specifically, and did so within a less narrowly screened, more clinically severe patient sample. For example, in our sample 70% of patients had comorbid psychiatric diagnoses versus 19% in Koszyski et al. sample; further, half of our patients were unemployed or disabled. The real-world context, severe and complex patients, and minimal inclusion criteria make this study strongly relevant to clinicians practicing in community, hospital, and VA settings.

Based on the anxiety-specific focus of CBT¹ versus broader focus of adapted MBSR, we predicted that CBT would reduce reported anxiety symptoms to a greater degree than adapted MBSR, whereas adapted MBSR would reduce broader symptoms (e.g., depression symptoms and co-occurring emotional disorders) to a greater degree than CBT. Given the very few studies in this area, however, these hypotheses were relatively exploratory.