You are on page 1of 5


The Effectiveness of Group Metacognitive Therapy for

Generalised Anxiety Disorder: A Pilot Study
Colin van der Heiden Kim
Melchior Ella de Stigter
J Contemp Psychother (2013)
43:151157 DOI
Published online: 2 April 2013
Springer Science+Business Media New York 2013
Generalised anxiety disorder (GAD) responds well to individually delivered metacognitive
therapy (MCT). The current study investigated the effectiveness of MCT administered to groups
of patients with GAD in a general outpatient treatment centre. Thirty-three consecu-tively
referred adult patients with GAD were assessed before treatment (pretreatment), after the last
treatment session (posttreatment), and six months after treatment had ended (follow-up).
Analyses of treatment effects were conducted for all patients entering the study (i.e., intent-to-
treat analysis, using the last-observation-carried-forward procedure), and repeated for the
patients who completed treatment (completers analysis), using paired samples t-tests. Further,
effect sizes (ESs) and indices of clinical significance were calculated, and compared with those
reported in studies into the efficacy and effectiveness of individually delivered MCT for GAD.
Among both the intent-to-treat and the completers sample, large pre- to posttreatment reductions
in symptoms of worry, trait-anx-iety, and general psychopathology. The magnitude of change
and the degree of clinical significance were smaller than those reported in studies into the
effectiveness of individually delivered MCT for GAD, whereas the attrition rate (27 %) was
higher. It is tentatively concluded that MCT for GAD can be effectively delivered in a group
format in a heterogenous clinical practice setting. Further evaluation is clearly indicated.

Generalized anxiety disorder remains a challenging con-dition to treat. Although there is little
debate that cognitive-behavioral therapy (CBT) is an efficacious therapy for GAD (e.g., Hunot et
al. 2007), reviews of the clinical significance have indicated that only about 50 % of treatment
completers achieve high end-state functioning (Hunot et al. 2007) or recovery (Fisher 2006)
following treatment. These data show that there is room for improvement and a need for
augmentation of current CBT strategies. In an attempt to augment the efficacy of CBT for GAD
several new treatments have been developed, based on models of the mechanisms and factors
underlying pathological worry, the hallmark of this disorder (see Hoyer et al. 2011). One
prominent example is a new form of cognitive therapy, which is based on the metacognitive
model for GAD (Wells 1995). Within this model worry is regarded as an attempt to generate
ways of coping for potentially threat-ening events, based on positive beliefs about the usefulness
of worry (e.g., Worrying helps me cope). The resultant worrying increases distress before
eventually reducing distress as the (internal) goal of worrying is accomplished. As such, the
process of worry in itself is not considered to be unique for GAD nor sufficient to cause this
anxiety disorder. Instead, GAD develops when negative beliefs about the uncontrollability and
dangerousness of worry (e.g., If I cannot stop my worrying, I will go crazy) form.
Such negative metacognitions begin to develop when worrying becomes inflexible and
persistent. Once acti-vated, these beliefs about worrying result in meta-worry, in which patients
with GAD worry about the fact that they worry. Meta-worry increases anxiety and worrisome
thoughts, which ultimately lead to the use of counterpro-ductive mental regulation (e.g., thought
suppression) and avoidance strategies (e.g., reassurance seeking and keeping away from
worrisome situations). Following from this model, metacognitive therapy (MCT) focuses on
modifi-cation of erroneous negative metacognitive beliefs, positive beliefs about worrying as an
effective means of dealing with threat, and unhelpful strategies to avoid worry and to control
So far, MCT has been evaluated in one uncontrolled trial (Wells and King 2006), and two
randomized con-trolled trials (Wells et al. 2010; Van der Heiden et al. 2012), all of which
achieved quite favorable results. That is, MCT produced significant and sustained improvements
on all outcome variables, with large effect sizes (ESs) and high recovery rates. Further, MCT
was found to be superior to applied relaxation (Wells et al. 2010) and to another specific form of
CBT for GAD based on the intolerance-of-uncertainty model (IUT) (Van der Heiden et al. 2012)
on measures of worry, trait-anxiety, and metacognitions.
The treatments in those three studies were all individ-ually delivered, whereas group therapy
might be an attractive alternative for several reasons, such as a more effective use of therapists
time, peer modeling and peer support (Himle et al. 2003; Yalom and Leszcz 2005). To date, only
two studies have been dedicated to determining the efficacy of group behavioral and cognitive
therapies for GAD. In a first study, 109 patients with GAD were allo-cated to either cognitive,
behavioral, cognitive-behavioral or placebo group therapy, or to a waiting list control con-dition
(White et al. 1992). The four treatment conditions all showed significant change when compared
to the waiting list control group, but only few differences emerged between the conditions.
Moreover, recovery rates (on the STAI-T) were only poor, ranging from 17 to 30 % at
posttreatment and 2033 % at six months follow-up (Fisher 2006). In a second, more recent study
(Dugas et al. 2003), using intolerance of uncertainty (IU) as the target variable of a group CBT
approach, 52 patients were randomly assigned to either group-IUT or a wait-list control group.
Results showed that the treatment group had greater improvement on all dependent variables.
Recovery rates were comparable to individual CBT-approaches, with both treatments achieving
recovery rates of approximately 50 %. However, group-IUT showed a marked increase in
recovery rates over the 1-year follow-up phase, from 48 to 64 %.
As the large ESs and high recovery rates suggest that MCT is a highly effective treatment for
GAD, and there is a paucity of research into the effectiveness of group ther-apy for GAD, the
primary goal of this study is to examine the effectiveness of MCT when administered to groups
of patients with GAD. To evaluate the potential of group MCT for GAD, treatment outcome data
are compared with data collected in an earlier study conducted at our center that utilized
individual MCT (Van der Heiden et al. 2012). Due to the sequential delivery of individual and
group treatment, the results of the two studies cannot be com-bined. However, using ESs and
recovery rates, the out-comes of group MCT will be compared to the individual outcomes of Van
der Heiden et al. (2012) as the treatment manuals, procedures and lead therapists were the same.
This methodology allows for an estimation of the effec-tiveness of group MCT, without needing
to use experimental controls (e.g., control conditions, randomization procedures), which are
impractical, or even impossible, in routine mental health care. Additionally, we will also
compare the ESs and recovery rates of both other studies on the efficacy and effectiveness of
MCT for GAD, and the ESs of the two earlier studies on group-CBT for GAD.

Thirty-three patients were found to meet the study inclusion criteria. All potential patients were
willing to participate in the study and none had to be excluded. The sample con-sisted of 12 men
and 21 women. Their mean age was 31.33 years (SD = 8.96; range 1966 years). Twenty-four
patients (73 %) had an additional diagnosis (panic disor-der = 8, major depressive disorder = 7,
social phobia = 7, undifferentiated somatoform disorder = 1, obsessive compulsive disorder =
1). Twenty patients (61 %) were employed and the rest was unemployed or retired. Sixteen
patients (48 %) were taking psychotropic medication (anti-depressants).
Of the 33patients entering the study,9 (27 %) dropped out in the active treatment phase, and
another 7 patients (21 %) were lost to follow-up. Further, of the 17 patients who com-pleted the
study, 10 received additional treatment (for other problems) during the 6-month follow-up period.
Therefore, the follow-up data were not included in the analyses, as they wouldbeinfluenced by the
additional treatments. Analyses of treatment effects were conducted for all patients entering
treatment (i.e., intent-to-treat analysis, using the last-obser-vation-carried-forward procedure),
and repeated for the patients who completed treatment (completers analysis).
MCT focuses on modifying metacognitive beliefs about worrying and increasing the flexible
use of alternative strategies for dealing with triggers, instead of modifying the worry itself or
teaching ways to control worrying (as is the case in most CBT approaches for GAD).
Following case conceptualisation and socialisation to the metacogni-tive model, MCT proceeds
with the direct modification of metacognitions by means of both verbal cognitive restructuring
strategies (e.g., questioning the evidence supporting these beliefs, inducing cognitive dissonance
by eliciting both positive and negative beliefs about worrying, employment of a mismatch
strategy in which patients are asked to compare their worrisome predictions with the actual
outcome of situations) and worry experiments. An example of the latter is a pushing-worry-to-
the-limit experiment, in which negative beliefs about the specific dangers of worrying are tested
by asking patients to worry more during distressing worry episodes. Positive meta-cognitions can
be tested by worry modulation experiments, in which patients are asked to reduce worrying on
one occasion and increase it on another occasion in order to determine whether greater worry
indeed leads to more positive outcomes. The modification of beliefs about worry follows a
particular sequence: first negative beliefs about the uncontrollability of worrying are targeted,
then danger metacognitions are challenged, and finally positive beliefs concerning the usefulness
of worry are addressed. In the final stage of MCT, alternative strategies for processing
threatening triggers are explored (e.g., letting go of thoughts instead of trying to deal with them).
Treatment comprised of 14 weekly 90-min sessions of MCT, which were administered to three
groups of 1014 individuals. The mean number of sessions was 12.94 (SD = 0.66; range: 1214)
for patients who completed the study, and 10.76 (SD = 3.71; range: 114) for the intent-to-treat
sample. All sessions were highly structured and manualized to ensure treatment integrity.
Treatment fol-lowed the manual by Van der Heiden et al. (2012), which was in turn based on the
MCT treatment outline as developed by Wells (1997). The manual is described in detail
elsewhere (Van der Heiden 2009).
Discussion and Conclusion
This is the first study into the effectiveness of MCT for GAD in a group format. In a general
outpatient setting, group MCT produced large pre- to postreatment decreases in symptoms of
worry and trait-anxiety among patients with GAD. The same was true for negative beliefs about
worrying. As for the the datas clinical significance, results point to good outcomes. In
conclusion, the results of this pilot study suggest that group MCT is effective in the treatment of
patients with GAD, and support the continued evaluation of group MCT, which should be
compared to both individually delivered MCT and other group treatments for GAD.

2. Interpersonal Emotion Regulation Model of Mood and Anxiety Disorders
Cogn Ther Res
DOI 10.1007/s10608-014-9620-1 - Springer Science+Business Media New York 2014
Published online: 23 May 2014Stefan G. Hofmann(&) Department of Psychology, Boston
University, 648 Beacon Street, 6th Floor, Boston, MA 02215-2002, USA

Although social factors are of critical impor-tance in the development and maintenance of
emotional disorders, the contemporary view of emotion regulation has been primarily limited to
intrapersonal processes. Based on diverse perspectives pointing to the communicative function of
emotions, the social processes in self-regulation, and the role of social support, this article
presents an interpersonal model of emotion regulation of mood and anxiety disorders. This model
provides a theoretical framework to understand and explain how mood and anxiety disorders are
regulated and maintained through others. The literature, which provides support for the model, is
reviewed and the clinical implications are discussed.
Experiencing and regulating emotions are essential human qualities. As humans, we have the ability
to empathize with another persons emotional state, because we know and sense what another
person must feel like. Some of us are better able to empathize and experience another persons
emotional state than others. This ability has been described with various terms, including empathy
(Preston and de Waal 2002), theory of mind (Leslie 1987), emotional intelligence (Salovey and
Mayer 1990), and alexithymia (Taylor et al. 1997).
Emotional experiences are complex and differ on many levelsintensity, valence, duration,
controllability, com-plexity, and action tendency (Barrett et al. 2007). To some extent, we do
have control over our emotional experience, either by avoiding the situations, people, or triggers
that elicit distress, or by changing our view of the situation. This ability has become known as
emotion regulation (Gross 2002, 1998; Thompson 1994) and coping (Lazarus 2000; Lazarus and
Folkman 1984). Coping researchers primarily examine responses to general stress, whereas
emotion reg-ulation researchers examine strategies an individual pos-sesses to deal with specific
positive and negative emotions.
In addition to these intrapersonal processes, social pro-cesses also appear to be important
aspects in the experience and expression of emotions. However, these processes have largely been
ignored in contemporary theories of emotion regulation. This article will introduce an interpersonal
model of emotion regulation for mood and anxiety disorders. This model offers a complementary
framework to the popular intrapersonal emotion regulation model (Gross 2002). More specifically,
the objective of this article is to: (1) briefly review the contemporary intrapersonal model of emotion
regulation and discuss its limitations; (2) introduce an interpersonal model of emotion regulation;
and (3) discuss the relevance of this interpersonal model of emotion regulation for the main-tenance
and treatment of mood and anxiety disorders.
Intrapersonal Emotion Regulation Models
Recently, authors have explored the role of emotion reg-ulation and dysregulation in emotional
disorders (Aldao et al. 2010), and especially mood and anxiety disorders (Amstadter 2008; Cisler
et al. 2010; Berking et al. 2013, 2014; Hofmann et al. 2012; Mennin et al. 2005; Wirtz et al.
2014). Experimental studies have shown that participants with anxiety and mood disorders
generally judge their negative emotions in response to a distressing film as less acceptable and
tend to suppress their emotions to a greater extent than nonanxious participants (Campbell-Sills
et al. 2006a). However, when instructed to accept their emotions, individuals with clinical
diagnoses of anxiety or depression report less subjective distress and lower autonomic arousal
than when asked to suppress their emotions in response to a distressing film (Campbell-Sills et al.
Similar effects have been observed in individuals who were asked to undergo a social stress task
(Hofmann et al. 2009). In this study, participants were randomly assigned to reappraise, suppress,
or accept their anticipatory anxiety prior to an impromptu speech. The instructions to suppress
anxiety were associated with greater increase in physio-logical arousal than the instructions to
reappraise and accept. Furthermore, the suppression group reported more subjective anxiety than
the reappraisal group. However, the acceptance and suppression groups did not differ in their
subjective anxiety response.
The intrapersonal model of emotion regulation (e.g., Gross 2002; Gross and John 2003; Gross and
Levenson 1997) has made an important contribution to the clinical field (Aldao et al. 2010).
However, the intrapersonal model does not consider the communicative function of emotions
and the self-regulatory processes that include social aspects. Social factors do not only act as
input or moderating variables of the model. Rather, the model presented here suggests that
emotions happen within a social context and are partly regulated through other people. Based on
a recently pro-posed framework (Zaki and Williams 2013), I distinguished intrinsic versus
extrinsic response-independent versus response-dependent interpersonal emotion regulation
strat-egies for mood and anxiety disorders. Depending on the context, interpersonal strategies can
be adaptive if they serve as a buffer of emotional stress and maladaptive if they contribute to the
maintenance of the problem. An example of a maladaptive response-dependent and independent
intrinsic emotion regulation strategy is the presence of a safety person for an individual with
panic disorder and agoraphobia. Frequent or habitual use of interpersonal emotion regulation
strategies can conceivably reduce the patients sense of control of his/her own emotion experience.
Therefore, it is quite possible that interpersonal emotion regulation can become maladaptive if a
patient becomes dependent on specific individuals or social groups in order to regulate ones own
emotions. This model offers a transdi-agnostic perspective of emotional disorders by considering
the broader social context of an individuals behavior and emotional experience.
Despite these advantages, an interpersonal model of emotion regulation shows a number of
weaknesses. First and foremost, this is a new model and there are no instruments available to
measure interpersonal emotion regulation strategies. Therefore, the direct empirical evi-dence for
the impact of these strategies on emotional dis-tress, including mood and anxiety disorders, is
relatively weak. Any assessment instrument will need to consider the influence of the cultural
context, because interpersonal emotion regulation strategies are directly related to social
standards and expectations. Finally, it remains unknown how interpersonal and intrapersonal
emotion regulation strategies interact and the relative importance of these strategies are
Although this review was primarily focused on mood and anxiety disorders, very similar
issues probably also apply to other disorders. A fuller understanding of these issues can
significantly advance our understanding of emotional disorders and can lead to new and
improved treatment strategies.