Journal of Consulting and Clinical Psychology © 2015 American Psychological Association

2015, Vol. 83, No. 5, 951–963 0022-006X/15/$12.00 http://dx.doi.org/10.1037/ccp0000042

A Randomized Controlled Trial on the Efficacy of Mindfulness-Based

Cognitive Therapy and a Group Version of Cognitive Behavioral Analysis
System of Psychotherapy for Chronically Depressed Patients

Johannes Michalak Martin Schultze

Witten/Herdecke University Free University of Berlin

Thomas Heidenreich Elisabeth Schramm

Esslingen University of Applied Sciences University Medical Center Freiburg
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Objective: Mindfulness-based cognitive therapy (MBCT) has recently been proposed as a treatment
option for chronic depression. The cognitive behavioral analysis system of psychotherapy (CBASP) is the
only approach specifically developed to date for the treatment of chronically depressed patients. The
efficacy of MBCT plus treatment-as-usual (TAU), and CBASP (group version) plus TAU, was compared
to TAU alone in a prospective, bicenter, randomized controlled trial. Method: One hundred and six
patients with a current DSM–IV defined major depressive episode and persistent depressive symptoms for
more than 2 years were randomized to TAU only (N ⫽ 35), or to TAU with additional 8-week group
therapy of either 8 sessions of MBCT (n ⫽ 36) or CBASP (n ⫽ 35). The primary outcome measure was
the Hamilton Depression Rating Scale (24-item HAM-D, Hamilton, 1967) at the end of treatment.
Secondary outcome measures were the Beck Depression Inventory (BDI; Beck, Steer, & Brown, 1996)
and measures of social functioning and quality of life. Results: In the overall sample as well as at 1
treatment site, MBCT was no more effective than TAU in reducing depressive symptoms, although it was
significantly superior to TAU at the other treatment site. CBASP was significantly more effective than
TAU in reducing depressive symptoms in the overall sample and at both treatment sites. Both treatments
had only small to medium effects on social functioning and quality of life. Conclusions: Further studies
should inquire whether the superiority of CBASP in this trial might be explained by the more active,
problem-solving, and interpersonal focus of CBASP.

What is the public health significance of this article?

The results show that the group version of the cognitive behavioral analysis system of psychotherapy
(CBASP) is an effective treatment for chronically depressed patients. Results for mindfulness-based
cognitive therapy (MBCT) were more equivocal for this patient group.

Keywords: mindfulness-based cognitive therapy, cognitive behavioral analysis system of psychotherapy,

randomized controlled trial, chronic depression

Major depressive disorder (MDD) is among the most prevalent
psychiatric conditions and places enormous burdens on individuals,
their families, and on society (Richards, 2011; Murray et al., 2012).
Besides high rates of relapse/recurrence after remission or recovery
(Frank et al., 1990; Kupfer et al., 1992; Holtzheimer & Mayberg,
2011), a substantial minority of approximately 20 –26.5% of patients
develop chronic courses characterized by significant depressive
symptoms lasting for at least 2 years (Gilmer et al., 2005; Satyanaray-
ana, Enns, Cox, & Sareen, 2009). Among patients treated in mental
health care facilities, 47% suffer from some form of chronic depres-

This article was published Online First August 10, 2015.
Johannes Michalak, Department of Psychology and Psychotherapy,
Witten/Herdecke University; Martin Schultze, Department of Methods and Eval-
uation, Free University of Berlin; Thomas Heidenreich, Department of Psychology
for Social Work and Nursing, Esslingen University of Applied Sciences; Elisabeth
Schramm, Department of Psychiatry and Psychotherapy, University Medical Cen-
ter Freiburg.
This trial was registered (NCT01065311) and was funded by the German
Science Foundation (DFG: Mi 700/4 –1).
We are grateful to the patients who participated in the trial and the
research assistants at the two treatment sites; we especially want to thank
the study therapists Ruth Fangmeier, Anne Katrin Külz, Marc Loewer,
Petra Meibert, and Tobias Rathleff. Moreover, we thank the LWL-Klinik
Dortmund and the Kliniken Essen-Mitte for their support.
Correspondence concerning this article should be addressed to Jo-
hannes Michalak, Witten/Herdecke University, Alfred-Herrhausen-
Straße 50, D-58448 Witten, Germany. E-mail: johannes.michalak@uni-
wh.de

951
 952 MICHALAK, SCHULTZE, HEIDENREICH, AND SCHRAMM
sion (Torpey & Klein, 2008). The lifetime prevalence of chronic
depression in the population ranges between 2.7% and 4.6% (Murphy
& Byrne, 2012; Satyanarayana et al., 2009; Young, Klap, Shoai, &
Wells, 2008); rates depend on the inclusion/exclusion of individuals
with dysthymic disorder.
Research has shown substantial differences between chronic and
nonchronic forms of MDD (Murphy & Byrne, 2012), which led to the
inclusion of a new diagnosis of persistent depressive disorder in the
fifth edition of the Diagnostic and Statistical Manual of Mental
Disorders (DSM–5; American Psychiatric Association, 2013). Two of
the most important features of chronic MDD are the slower rate of
improvement and a poorer treatment response (Klein, Shankman &
Rose, 2006). Approximately 50% of chronically depressed patients
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
fail to respond to antidepressant medication or psychotherapy and
This document is copyrighted by the American Psychological Association or one of its allied publishers.
another 20% do not achieve complete remission (Harrison & Stewart,
1995; Keller et al., 1998; Kocsis et al., 1988; Thase et al., 1996).
Accordingly, efforts to improve treatment outcome in this highly
challenging population are needed.
One psychotherapeutic approach to have recently been proposed as
a treatment option for chronic MDD is mindfulness-based cognitive
therapy (MBCT; Segal, Williams, & Teasdale, 2002). MBCT was
originally designed and evaluated for relapse prevention of remitted
MDD patients. A meta-analysis by Piet and Hougaard (2011) showed
that MBCT plus treatment as usual reduced risk of relapse by 43% for
formerly depressed patients with three or more previous episodes in
comparison to treatment as usual alone.
MBCT is a group-based program combining intensive training
in mindfulness (Kabat-Zinn, 1990) with cognitive– behavioral el-
ements targeting depression. In previous trials, MBCT was usually
delivered in eight weekly group sessions with a group size ranging
from 12 to 15 patients. The core skill that MBCT aims to teach is
the ability to recognize and disengage from mind states character-
ized by self-perpetuating patterns of ruminative, negative thought
that escalate and maintain depressive symptoms. In contrast to
getting lost in negative modes of mind and being preoccupied with
the past or future scenarios, MBCT fosters a nonjudgmental aware-
ness of the present moment experience, including one’s sensations,
thoughts, bodily states, consciousness, and the environment, while
encouraging openness, curiosity, and acceptance (Bishop et al.,
2004; Kabat-Zinn, 2003). The mindful awareness of the here-and-
now encompasses a decentered view on cognitions: By intensive
training in mindfulness, patients are supported to recognize that
(even long-held) thoughts and feelings are events in the mind and
not self-evident truths or aspects of the self (Teasdale, Moore,
Hayhurst, Pope, Williams, & Segal, 2002). Moreover, mindfulness
helps patients to turn towards rather than escape or avoid experiences
(e.g., dysphoric mood) without getting lost in ruminative, negative
thoughts. In addition, antidepressive cognitive– behavioral therapy
elements such as exercises on taking alternative viewpoints on cog-
nitions (e.g., just watching thoughts come and go, without feeling that
one has to follow them) or activity scheduling are introduced. How-
ever, it should be highlighted that MBCT focuses on changing the
relationship to thoughts rather than changing the cognitive content.
Besides its effects in relapse prevention it was hypothesized that
MBCT might also assist chronically depressed patients to facilitate
decentered and self-compassionate views of typical maladaptive
core beliefs. By training in mindfulness (e.g., during sitting med-
itation), chronically depressed patients might learn to recognize
when they lose contact with the here-and-now and develop the
ability to step out of mind states characterized by ruminative
negative thinking. Besides increases in mindfulness skills, research
on mechanisms of change has shown that the effects of MBCT are
mediated by enhanced self-compassion (Neff, 2003), the ability to
relate to oneself with kindness when encountering pain and per-
sonal shortcomings (Kuyken et al., 2010).
Results of three smaller or uncontrolled studies have suggested
that chronically depressed patients benefit from MBCT (Barnhofer
et al., 2009; Eisendrath et al., 2008; Kenny & Williams, 2007);
however larger randomized controlled trials testing the efficacy of
MBCT in chronic depression are lacking.
The only psychotherapeutic model to have been specifically
developed for the treatment of chronic depression is the cognitive
behavioral analysis system of psychotherapy (CBASP, Mc-
Cullough, 2000). In contrast to the cognitive focus of MBCT,
CBASP mainly has an interpersonal orientation. It is a highly
structured learning approach integrating behavioral, cognitive, and
predominantly interpersonal treatment strategies. Assuming that
early interpersonal trauma leads to dysfunctional mechanisms of
derailed affective and motivational regulation and to a reduction of
perceived functionality, the main objectives in CBASP are to learn
to recognize the consequences of one’s own behavior on other
persons and to develop social problem solving skills and empathy.
Based on the assumptions about relationships patients have formed
following experiences in their early history with being maltreated
by significant others, patients formulate a proactive transference
hypothesis (e.g., “I have nothing to expect from my therapist”). In
the CBASP group version used in the present study, patients also
formulated transference hypotheses regarding the other group
members (e.g., “If I make myself vulnerable, I will be ridiculed”).
By means of interpersonal discrimination exercises, these hypoth-
eses are contrasted with the actual behavior of the therapist and the
group members to modify dysfunctional expectations. In this way,
the patient learns to read other persons by focusing on their overt
behavior and the effects he or she has on them.
The therapist gives the patient direct feedback by expressing his
or her personal reactions to the patient’s dysfunctional behavior
patterns and offers alternatives (disciplined personal involvement).
However, in a group format, disciplined personal involvement is
used less frequently when it comes to dysfunctional behavior in the
group and is often replaced by using the Kiesler Interpersonal
Circle model in an educative and structuring way. Kiesler (1982)
operationalized the way people interact as reciprocal interpersonal
transactions carried out within two domains: (a) power (dominance
vs. submission) and (b) affiliation (hostile vs. friendly) using a
circumplex design with power and affiliation serving as
perpendicular-intersecting diameters in the circle. The patient
chooses his or her position in the circle (e.g., friendly dominant)
according to the outcome he or she desires for the interpersonal
situation (e.g., “I want to tell the group that I am disappointed”)
and compares it with the impact of his or her usual behavior
position (e.g., hostile-submissive). Most of the group sessions
comprise step-by-step situational analyses in which patients learn
to formulate a desired outcome in an interpersonal situation and
how to achieve this outcome by role playing goal-oriented behav-
ior. The percent distribution of situation analyses and the other
techniques is about 70:30 in both individual and group sessions.
The efficacy of CBASP as a principal treatment (as opposed to
an augmentation strategy in the context of a pharmacotherapy
 MBCT AND CBASP FOR CHRONICALLY DEPRESSED PATIENTS
algorithm) was shown in a large American study (Keller et al.,
2000) and a smaller German study that comprised 22 individual
outpatient sessions over 16 weeks (effect size of d ⫽ 1.4; Schramm
et al., 2011). These findings were confirmed in a recent network
meta-analysis (Kriston, von Wolff, Westphal, Hölzel & Härter,
2014), which concluded the CBASP approach was recommended
over interpersonal psychotherapy for chronically depressed pa-
tients. Adapted to a group modality, CBASP showed encouraging
results in treatment-resistant depression in a single-arm pilot study
(Sayegh et al., 2012).
We report a bicenter randomized clinical trial (RCT) that treated
chronically depressed patients with either MBCT plus treatment as
usual (TAU), CBASP (group version) plus TAU or TAU alone.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This study is the first RCT using MBCT in chronically depressed
This document is copyrighted by the American Psychological Association or one of its allied publishers.
patients with an adequate sample size to detect a medium-sized
effect. Moreover, although previous studies compared MBCT with
pharmacological (Kuyken et al., 2008; Segal et al., 2010) or
cognitive psychological education conditions (Williams et al.,
2014), the present study is the first to compare MBCT with a gold
standard alternative psychotherapeutic approach relying on differ-
ent mechanisms of change. Whereas the primary focus of MBCT
is to facilitate a mindful and compassionate relationship to inner
experiences, the primary focus of CBASP is to improve patients’
interpersonal functioning.
Method
Design
At each trial site, patients were randomly assigned to either
TAU alone or—in addition to TAU— either MBCT or CBASP.
After patient eligibility was assessed and informed consent was
obtained, patients were formally enrolled in the study. For
randomization, the trial sites mailed the patient study number to
a central allocator who was independent of the staff involved in
the recruitment, assessment, and management of study partici-
pants. Information was sent for groups of eligible patients at a
time. Block randomization (block size ⫽ 6) to the three condi-
tions was performed by the independent allocator using a
computer-generated list of random numbers. The central allo-
cator then mailed the allocations back to the treatment sites. If
insufficient patients for each condition (i.e., 12) could be re-
cruited during a recruitment period, the numbers used for block
randomization were lower. Thus, sample size was not a multiple
of the block size.
Patients were recruited at two sites in Germany: Site A was located
at the Ruhr megalopolis (population 5 million); Site B in the area of
Freiburg im Breisgau (population more than 230,000). Site A was
particularly experienced in MBCT, Site B in CBASP, thus making it
possible to control for allegiance effects (Luborsky et al., 1999).
Written informed consent from participants was obtained after
the procedure had been fully explained. The Research Ethics
Committee of the German Psychological Association approved the
study (JM 072009).
Participants
At Site A all patients were recruited by media announcements;
at Site B patients were recruited from community health care
953
facilities or private practices. All individuals interested in partici-
pating in the study took part in a telephone screening based on the
mood disorder module of the Structured Clinical Interview for
DSM–IV (SCID, Wittchen, Wunderlich, Gruschwitz, & Zaudig,
1997). Patients who seemed eligible were invited to an extended
diagnostic interview. Diagnoses were assessed using the SCID
Axis I and Axis II disorders. All diagnostic evaluations were
conducted by trained and certified clinical psychologists and were
reviewed by senior study investigators (Johannes Michalak and
Elisabeth Schramm).
All patients had a current major depressive episode (MDE) as
defined by the Diagnostic and Statistical Manual of Mental Disorders
(4th ed.; DSM–IV; American Psychiatric Association, 1994) and had
experienced depressive symptoms for more than 2 years without
remission. We included three subtypes of depressed patients: (a)
patients with chronic major depression (i.e., current MDE lasting for
more than two years); (b) patients meeting criteria for double depres-
sion (current MDE superimposed on an antecedent dysthymic disor-
der), or (c) patients with current MDE as part of a recurrent major
depression with incomplete recovery between episodes during the last
two years (i.e., depressive symptoms present during the entire two
year period). We used exclusion criteria that corresponded to those
used in previous studies on MBCT (e.g., Teasdale, Segal, Williams,
Ridgeway, Soulsby & Lau, 2000): history of schizophrenia or schizo-
affective disorder, current substance abuse, eating disorder, organic
mental disorder, borderline personality disorder, and inability to en-
gage in treatment for physical, practical, or other reasons. We ex-
cluded patients with eating disorders because they frequently experi-
ence depression secondary to eating disorders and the MBCT
program was not designed to deal with the primary eating disorder.
Patients with borderline personality disorders were excluded because
their style of interaction might be too difficult to deal with within the
group format of MBCT or CBASP. Current substance abuse is a
contraindication for the meditation exercises used in MBCT (Segal et
al., 2002).
MBCT and CBASP groups were set up at each site (four of each
at Site A, and two of each at Site B). All MBCT groups at Site A
were conducted by a clinical psychologist and certified MBCT
therapist with 20 years of mindfulness practice (female, age 51; 12
MBCT courses before the start of the study). At Site B, there were
two MBCT therapists: one MBCT group was conducted by a
psychiatrist certified in MBCT with 20 years of mindfulness
practice (male, age 38; seven MBCT courses before the start of the
study), and the other group by a clinical psychologist and psycho-
therapist with 5 years of mindfulness practice (female, age 38; two
MBCT courses before the start of the study). All MBCT groups
were supervised by Johannes Michalak. A licensed clinical psy-
chologist and certified CBSAP therapist conducted the four
CBASP groups at Site A (male; age 29; no experience in conduct-
ing CBASP groups before the beginning of the study). At Site B,
both groups were directed by a licensed clinical psychologist and
certified CBASP therapist (female, age 47; four CBASP groups
before the beginning of the study). All CBASP groups were
continuously supervised by Elisabeth Schramm.
All group sessions were videotaped (Site A) or audiotaped (Site
B) for therapist supervision. Sessions 4 and 7 were used for ratings
of treatment adherence and competence.
 954 MICHALAK, SCHULTZE, HEIDENREICH, AND SCHRAMM
Treatments
TAU. All patients were instructed that they should be in
individual treatment by either a psychiatrist or a licensed psycho-
therapist (not a member of the study team) during the study period.
If patients were already in psychiatric or psychotherapeutic indi-
vidual treatment at study intake, they continued their treatment
with this psychiatrist or psychotherapist. Patients were encouraged
to continue any current medication and to attend appointments
with their psychiatrist or psychotherapist. There were no restric-
tions on other forms of supplementary treatment.
MBCT. The treatment protocol followed the MBCT manual
developed by Segal, Williams and Teasdale (2002). The program
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
consisted of an individual preclass interview and eight weekly 2.5-hr
This document is copyrighted by the American Psychological Association or one of its allied publishers.
group sessions. In contrast to most previous studies on MBCT, group
size was restricted to six patients per class in the present trial to equate
group size in the two psychotherapy conditions.
The eight sessions included guided formal mindfulness practices
(i.e., body scan, sitting meditation, yoga), inquiry into patients’
experience of these practices and review of weekly homework
(i.e., daily 40 min of formal mindfulness practice and generaliza-
tion of session learning). Moreover, informal mindfulness practice
(i.e., exercises designed to integrate the application of awareness
skills into daily life) and cognitive– behavioral skills (e.g., activity
scheduling, skills in dealing with cognition) were taught and
discussed. Some minor alterations were made to adapt the program
to chronically depressed patients. For example, possible suicidal
tendencies were carefully assessed during the preclass interview.
Moreover, instead of exploring early warning signs for relapse,
early warning signs for a further deterioration of mood were
explored and possible functional coping strategies were discussed.
CBASP. The CBASP treatment protocol followed the manual
developed by McCullough (2000), and modified for the group
setting by Schramm, Brakemeier, and Fangmeier (2012). The
CBASP program consisted of two individual treatment sessions (to
derive transference hypotheses) and eight weekly 2.5-hr group
sessions. The main modifications to the individual format included
the derivation of the transference hypothesis (described above)
with regard to the group. Based on the patient’s transference
hypothesis regarding the group, the actual and observable behavior
of the group members were contrasted with this hypothesis in the
presence of a transference hotspot (e.g., the patient makes a mis-
take and expects to be ridiculed but experiences support from the
group members instead) by the end of the sessions and the learning
experience is explicitly expressed (e.g., “I can make a mistake and
still be accepted”). Moreover, as described above, disciplined
personal involvement of the therapist is used less frequently in the
group version of CBASP. Instead, the impact of the patients’
dysfunctional behavior is identified and described using Kiesler’s
interpersonal circle model (described above). However, the thera-
pist maintains a pronounced personal-authentic stance toward the
participants throughout treatment. Because of the 2.5-hour dura-
tion of each group session, in some sessions two situational anal-
yses were performed, thus the number of situational analyses was
comparable to other studies using individual CBASP. To enable
situational analyses of all group members, we limited the number
of participants to six. These modifications should not impact
CBASP’s theoretical mechanisms of change.
Measures
Primary outcome measure. The primary outcome measure
was the 24-item Hamilton Depression Rating Scale (HAM-D;
Hamilton, 1967; Guy, 1976), a widely used interview-based mea-
sure of the severity of depressive symptoms covering a range of
affective, behavioral, and biological symptoms. The HAM-D was
administered at baseline and posttreatment after the 8-week treat-
ment phase by five trained doctoral-level psychologists. To main-
tain rater blindness, patients were instructed at the beginning of
each interview not to mention their treatment condition or their
psychotherapist. To enhance reliability of the assessment, the
Structured Interview Guide for the HAM-D (Moberg et al., 2001)
was applied. A sample of 36 interviews from the baseline assess-
ment were assessed by an independent rater, yielding an interrater
correlation of r(34) ⫽ .97, p ⬍ .001.
Secondary outcome measures. We used the BDI (Beck,
Steer & Brown, 1996; German version by Hautzinger, Keller, &
Kühner, 2006) to assess depressive symptoms by self-report. The
BDI is a widely used 21-item measure covering affective, cogni-
tive, motivational, behavioral, and biological symptoms of depres-
sion with good psychometric properties (Beck, Steer, & Carbin,
1988). Moreover, we assessed social functioning with the Social
Adaptation Self-Evaluation Scale (SASS) (Bosc, Dubini & Polin,
1997; German version by Duschek, Schandry & Hege, 2003).
SASS is a 21-item scale for the evaluation of patient social
motivation and behavior in depression and has been shown to be
reliable, valid, and sensitive to change. In addition, we measured
quality of life with the Short Form Health Survey (SF-36; Ware &
Sherbourne, 1992; German version by Morfeld, Kirchberger &
Bullinger, 2011). The SF-36 assesses mental and physical health
with 36 items and shows good psychometric properties. We fo-
cused on mental health in our analysis, using the vitality, mental
health, social functioning, and role emotional (assessing role lim-
itations due to emotional problems) subscales of the SF-36.
Statistical Analysis
The two outcome measures related to depressive symptoms
were analyzed together using a structural equation modeling
(SEM) approach. This approach was chosen because it allows for
the explicit modeling of measurement errors, thus enabling the
analysis of intervention effects on latent representations of the
outcome measures. The analysis of true (i.e., void of measurement
error) scores in intervention studies has the advantage that the
possible confounds of changes in errors being attributed to changes
in the underlying construct is avoided. Following preliminary
analysis, HAM-D and BDI were each split into two halves and
parcels were computed.
To analyze both outcome measures simultaneously, a multi-
method analysis using the CTC(M-1) (correlated trait-correlated
method minus one, Eid, 2000) was conducted. In this approach one
assessment method is chosen as a reference and other methods
assessing the same construct are contrasted against this reference.
The HAM-D, being the primary outcome measure in this study,
was chosen as the reference method and the BDI was contrasted
against it. The contrasting is carried out by regressing the BDI
indicators on the latent state St, generating residuals of the BDI
that represent components that are not shared with the HAM-D.
This results in the assessments made via HAM-D being disaggre-
 MBCT AND CBASP FOR CHRONICALLY DEPRESSED PATIENTS
gated into two components (the latent state St and the measurement
error εikt), whereas the assessments made via the BDI are disag-
gregated into three components (the latent state St, the method
factor Mt, and the measurement error εikt). Within this modeling
approach the latent state St represents the occasion-specific de-
pressive symptoms assessed via the HAM-D (occasion: pre- and
posttreatment assessment). The method factor Mt represents the
occasion-specific component of the assessment made via the BDI
that is not shared with the assessment of the same occasion made
via the HAM-D. Using this approach, consistency coefficients can
be computed that represent the amount of variance in the BDI that
is shared with HAM-D assessment, indicating the degree of con-
vergent validity between the two assessment methods.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
To incorporate the inherent longitudinal nature of the study in
This document is copyrighted by the American Psychological Association or one of its allied publishers.
the analysis, the true-change approach proposed by Steyer, Eid and
Schwenkmezger (1997) was chosen. In this approach intraindi-
vidual changes are represented by the latent variable SC, which is
defined as the difference between the latent states of two different
measurement occasions: SC ⫽ St ⫺ St=. Incorporating the change
between two occasions as a latent variable into the model allows
for the direct analysis of the intervention effects. This is achieved
by regressing the latent change variables of the reference measured
states (representing HAM-D measurements) as well as those of the
latent method variables (representing the discrepancy between
HAM-D and BDI measurements) on dummy-coded predictors
indicating the received treatment. For this TAU was chosen as the
reference.
Beyond the intervention evaluation, combining the CTC(M-1)
and the true-change approaches allows for the computation of
reliability and consistency coefficients for the assessment of oc-
casion specific states as well as the assessment of change (Eid,
2000), providing a way to inspect the quality of BDI and HAM-D
at a given occasion as well as in a longitudinal context.
To account for the study implementation at two different sites
and to investigate possible allegiance effects the model was ex-
tended to a multiple-group SEM. Equality assumptions were tested
using the Wald test. To counteract the possibility of inflated Type
I errors due to the group-centered implementation of the interven-
tions (Baldwin, Murray, & Shadish, 2005) cluster-based standard
error corrections were performed in accordance with the approach
for complex survey data proposed by Asparouhov (2004). To
handle missing data due to nonresponse as well as drop-out a
saturated multiple-group imputation model including all primary
and secondary outcome measures, gender, age, implementation
site, and treatment condition was used to generate 50 imputed data
sets with the MCMC simulation procedure described by Asp-
arouhov and Muthén (2010).
Power analysis revealed that for analyses using the CTC(M-1)
approach with three treatment conditions at two treatment sites a
sample size of 114 patients is required to detect intervention
effects of at least medium effect size (d ⬎ .5) with Type I error
rates set at .05 (two sided) and Type II error rates set at .2. To
obtain this optimal sample size estimation, parallel test halves as
well as equal group sizes were assumed. Because of some prob-
lems in patient recruitment, the actual sample size was slightly
smaller than the optimal sample size estimated in the power
analysis.
955
We also computed proportions of remitters. A remitter was
defined as a participant with a HAM-D score of less than 8 at
posttreatment.
The secondary outcomes assessing mental health and social
adaptation (i.e., SF-36 and SASS) were investigated with a single-
indicator true-change model for each outcome separately. The
CTC(M-1) approach used for the simultaneous analysis of
HAM-D and BDI was not used for these outcomes because a
comparison of two assessment methods was not required. In ad-
dition, a SEM with all five relevant scales (four mental health
subscales of SF-36 as well as SASS) was not feasible with the
obtained sample size.
The true-change models were analyzed using Mplus Version 7
(Muthén & Muthén, 1998 –2013) and all additional analyses were
conducted using R Version 3.0.1 (R Development Core Team,
2013) or SPSS16.0.
Results
Description of Sample
Patient flow is illustrated in Figure 1. One hundred and six
patients agreed to participate, met the inclusion criteria, and were
randomized to TAU only (N ⫽ 35), or in addition to TAU to
MBCT (N ⫽ 36) or CBASP (N ⫽ 35).
Patient characteristics are shown in Table 1. The majority of
patients (83%) met DSM–IV criteria for chronic major depression
and 88% of the sample fulfilled the criteria for persistent depres-
sive disorder as defined by DSM–5. We found no significant
differences between conditions in any of the sociodemographic or
clinical variables.
Descriptive statistics of the pretreatment scores of the HAM-D
and BDI are shown in Table 2. An analysis of variance for the
baseline data revealed that neither the main effects of treatment
condition, HAM-D: F(2, 100) ⫽ 0.79, p ⫽ .46; BDI: F(2, 98) ⫽
0.33, p ⫽ .73, the main effect of implementation site, HAM-D:
F(1, 100) ⫽ 3.79, p ⫽ .05; BDI: F(1, 98) ⫽ 1.59, p ⫽ .21, nor their
interaction, HAM-D: F(2, 100) ⫽ 0.89, p ⫽ .41, BDI: F(2, 98) ⫽
1.66, p ⫽ .20, were statistically significant.
We found no significant differences between treatment sites in
the number of patients treated with medication (Site A: 48; Site B:
26; ␹2 ⫽ 0.01, df ⫽ 1, p ⫽ .93) or individual psychotherapy (Site
A: 17; Site B: 13; ␹2 ⫽ 1.31, df ⫽ 1, p ⫽ .25). Moreover, we found
no significant differences between sites in any of the demographic
characteristics except education, with a higher proportion of well-
educated patients at Site B (secondary school only: Site A: 31%;
Site B: 53%; ␹2 ⫽ 4.24, df ⫽ 1, p ⫽ .04). However, we found
differences between sites in a number of clinical variables. We
observed a statistical trend for higher HAM-D scores at Site A
(Site A: M ⫽ 24.15, SD ⫽ 6.18; Site B: M ⫽ 21.94), SD ⫽ 6.62,
t(104) ⫽ 1.95, p ⫽ .05); no significant differences between sites
were observed for the BDI. More patients at Site A had comorbid
Axis I (Site A: 74%; Site B: 44%; ␹2 ⫽ 8.25, df ⫽ 1, p ⫽ .004)
and Axis II diagnoses (Site A: 24%; Site B: 6%; ␹2 ⫽ 5.67, df ⫽
1, p ⫽ .02). Moreover, there were significant differences in type of
depression between sites (Site A: chronic major depression: 77%;
double depression: 4%; recurrent depression: 19%; Site B: chronic
major depression: 97%; double depression: 3%; recurrent depres-
sion: 0%; ␹2 ⫽ 8.11, df ⫽ 2, p ⫽ .02).
 956 MICHALAK, SCHULTZE, HEIDENREICH, AND SCHRAMM

443 Patients contacted trial

centers (A: 343; B: 100)

Enrollment 350 assessed for eligibility

(A: 302; B: 48)
244 Patients were excluded (A: 231; B:
13)
174 not meeting inclusion criteria
66 refused to participate
106 met the inclusion criteria
and were randomized (A: 69;
B: 37)
Allocation
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

36 were allocated to MBCT + 35 were allocated to CBASP + 35 were allocated to TAU

TAU intervention (A: 24; B: 12) TAU intervention (A: 22; B: 13) intervention (A: 23; B: 12)
This document is copyrighted by the American Psychological Association or one of its allied publishers.

10 Patients discontinued 10 Patients discontinued

1 Patient withdrew (A: 0; B: 1)
Treatment

intervention (A: 7; B: 3) intervention (A: 9; B: 1)

Phase

10 Patients did not attend ≥ 4 6 Patients did not attend ≥ 4

sessions (A: 7; B: 3) sessions (A: 4; B: 2)
Assessment

28 Patients participated in Post 25 Patients participated in Post 32 Patients participated in Post

Assessment (A: 17; B: 11) Assessment (A: 13; B: 12) Assessment (A: 22; B: 10)
Post

36 Patients in ITT analysis (A: 35 Patients in ITT analysis (A: 35 Patients in ITT analysis (A:
Analysis

24; B: 12) 22; B: 13) 23; B: 12)

26 Patients in Per-protocol 28 Patients in Per-protocol 32 Patients in Per-protocol

analysis (A: 17; B: 9) analysis (A: 18; B: 10) analysis (A: 22; B: 10)

Figure 1. CONSORT flow diagram. MBCT indicates mindfulness-based cognitive therapy; CBASP ⫽
cognitive behavioral analysis system of psychotherapy; TAU ⫽ treatment as usual. A indicates the Ruhr
megalopolis site; B indicates Freiburg site.

Treatment Compliance
TAU. Of the 85 patients participating in the posttreatment
assessment from whom data on treatment during the 8-week study
period are available, 46 (54.1%) had one or more individual
appointments with a psychiatrist (average number of appoint-
ments ⫽ 2.22; SD ⫽ 1.86) and 31 (36.5%) had sessions with a
licensed psychotherapist (average number of sessions ⫽ 3.68;
SD ⫽ 2.00). There were no significant differences between pa-
tients exclusively treated with TAU and patients additionally
treated with MBCT or CBASP in the number of individual ap-
pointments with a psychiatrist (␹2 ⫽ 1.66, df ⫽ 2, p ⫽ .44), in the
number of psychotherapy sessions (␹2 ⫽ 0.38, df ⫽ 2, p ⫽ .83), or
in the average number of sessions they received—number of
appointments with a psychiatrist: F(2, 82) ⫽ 2.17, p ⫽ .12;
number of individual sessions with a psychotherapist: F(2, 82) ⫽
0.18, p ⫽ .84. Moreover, we found no significant differences
between conditions in the proportion of patients treated with
antidepressant medication.
However, it should be noted that on a descriptive level the
number of patients receiving antidepressant medication was
lower in patients exclusively treated with TAU (n ⫽ 17, 53.1%)
than in patients who received TAU plus MBCT (n ⫽ 21, 75%)
or TAU plus CBASP (n ⫽ 19, 76%) (see Table 1). Medications
were mostly selective serotonin and norepinephrine reuptake
inhibitors (n ⫽ 29, 34.1%) and selective serotonin reuptake
inhibitors (n ⫽ 17, 20%). Five patients (5.9%) received tricyclic
antidepressants whereas four (4.7%) took Agomelatin and three
(3.5%) received MAO inhibitors (note that a small number of
patients took antidepressant medication from two of these sub-
stance classes). Nine patients (10.6%) took neuroleptic medi-
cation, whereas one (1.2%) was on lithium. No reliable infor-
mation on medication dose was available. There were no
significant differences in AD intake between study sites (␹2 ⫽
0.30, df ⫽ 1, p ⫽ .58). A minority of 15 patients (17.6%) did
not report any treatment (no sessions with a psychotherapist or
psychiatrist and no antidepressant medication) within the
8-week study period.
MBCT and CBASP. Of the 71 patients allocated to one of the
group therapies, 20 patients (28.2%) discontinued therapy or at-
tended fewer than four sessions (10 MBCT patients [27.8%] and
10 CBASP patients [28.6%], ␹2 ⫽ 0.003, df ⫽ 1, p ⫽ .96).
Separate analyses for both sites revealed no significant difference
in the rate of discontinuation of MBCT and CBASP in the two
centers (Site A: MBCT ⫽ 7, CBASP ⫽ 9; ␹2 ⫽ 0.34, df ⫽ 1, p ⫽
 MBCT AND CBASP FOR CHRONICALLY DEPRESSED PATIENTS 957

Table 1
Sample Characteristics

MBCT CBASP TAU

Characteristic (n ⫽ 36) (n ⫽ 35) (n ⫽ 35) p value

Age, M (SD) 48.4 (11.5) 50.2 (10.5) 54.0 (13.24) .14a

Female sex, n (%) 21 (58.3) 22 (62.9) 23 (65.7) .81b
Secondary school (Abitur), n (%) 13 (36.1) 13 (37.1) 11 (31.4) .58b
Subtypes of depression, n (%) .80b
Chronic major depression 28 (80.0) 30 (85.8) 30 (85.8)
Double depression 1 (2.9) 1 (2.9) 2 (5.8)
Recurrent depression 6 (17.1) 4 (11.4) 3 (8.6)
Antidepressant medication, n (%)c 21 (75.0) 19 (76.0) 17 (53.1) .11b
Individual Psychotherapy, n (%)3 9 (32.1) 10 (40.0) 12 (37.5) .83b
Any additional Axis I diagnoses, n (%) 21 (58.3) 21 (60.0) 26 (74.3) .31b
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Any Axis II comorbidity, n (%) 8 (22.2) 5 (14.3) 6 (17.1) .68b

This document is copyrighted by the American Psychological Association or one of its allied publishers.

Marital status: .68b

Single, n (%) 12 (33.3) 8 (22.6) 11 (31.4)
Married, n (%) 17 (47.2) 19 (54.3) 12 (34.3)
Divorced, widowed, n (%) 5 (13.9) 7 (20.0) 6 (17,1)
Note. MBCT ⫽ mindfulness-based cognitive therapy; CBASP ⫽ cognitive behavioral analysis system of
psychotherapy; TAU ⫽ treatment-as-usual.
a
By analysis of variance. b By ␹2 test. c Percentage was calculated for patients staying in the study and
participating in the post-treatment assessment.

.56; Site B: MBCT ⫽ 3, CBASP ⫽ 1; ␹2 ⫽ 1.01, df ⫽ 1, p ⫽ .32).
Patients who attended four or more therapy sessions did not differ
on any of the demographic or baseline clinical characteristics from
those who discontinued therapy or attended fewer than four ses-
sions. There were no suicides, suicide attempts, or other serious
adverse events.
Treatment Adherence and Competence
Adherence to the MBCT manual was rated with the Mindfulness-
Based Cognitive Therapy Adherence Scale (Segal, Teasdale, Wil-
liams & Gemar, 2002) by a licensed psychotherapist with experience
in conducting MBCT groups. Adherence in both centers was con-
firmed by a score of 1.88 for therapies conducted at Site A and 1.71
at Site B (scale range: 0 –2). Adherence to the CBASP manual was
rated with the CBASP Adherence Scales (McCullough, 2003) by
Elisabeth Schramm. Adherence in both centers was confirmed by an
average overall score of 4.62 for therapies conducted at Site A and
4.90 at Site B (scale range: 1–5).
Competence of delivery of MBCT was rated with the Mindfulness-
Based Interventions—Teaching Assessment Criteria (Crane et al.,

Table 2
Means and Standard Deviations (in Parentheses) of Depressive Symptoms at Baseline
and Posttreatment

MBCT CBASP TAU

HAM-D Score
Overall
Baseline 23.03 (6.27) 24.71 (6.69) 23.87 (6.33)
Posttreatmenta 17.86 (10.37) 14.64 (8.85) 21.16 (8.16)
Site A
Baseline 24.63 (5.63) 24.91 (6.55) 23.87 (6.58)
Posttreatmenta 20.18 (10.24) 15.14 (8.19) 20.82 (8.52)
Site B
Baseline 19.83 (6.49) 24.33 (7.22) 21.67 (5.80)
Posttreatmenta 14.27 (9.97) 14.00 (10.01) 21.90 (7.67)
BDI-Score
Overall
Baseline 31.31 (9.55) 30.26 (7.52) 29.76 (7.42)
Posttreatmenta 22.69 (12.11) 22.28 (12.54) 28.34 (10.27)
Site A
Baseline 33.42 (8.03) 30.00 (7.08) 30.00 (7.73)
Posttreatmenta 23.71 (11.59) 22.93 (12.82) 27.68 (9.34)
Site B
Baseline 27.08 (11.24) 30.75 (8.58) 29.33 (7.14)
Posttreatmenta 20.78 (13.54) 21.45 (12.73) 29.80 (12.49)
Note. HAM-D ⫽ Hamilton Rating Scale for Depression; BDI ⫽ Beck Depression Inventory.
a
No imputation of missing data.
 958 MICHALAK, SCHULTZE, HEIDENREICH, AND SCHRAMM

2013) by Thomas Heidenreich. Competence was good, as indicated Table 3

by an average overall score of 5.10 for MBCT therapies conducted at Results of the Regression Analyses Predicting the True-Changes
Site A and 4.71 at Site B (scale range: 1– 6). No specific instrument of the HAM-D (Reference-Measured States; SC) and BDI
for assessing CBASP competence is currently available. As a proxy to Specific Components (Method Factors; MC) by
assess CBASP competence we separately analyzed the Quality of the Intervention Type
Interpersonal Relationship subscale of the CBASP Adherence Scales
Standardized
(McCullough, 2003). This subscale comprises items assessing the Estimate estimatea SE p value
therapeutic competence of CBASP therapists (e.g., to form a collab-
Predicting SC
orative rapport, to listen effectively, and to appropriately control the (Intercept) ⫺0.13 0.04 ⬍.001
session). Competence was good, as indicated by an average overall MBCT ⫺0.09 ⫺0.29 0.07 .18
score of 4.77 for CBASP therapies conducted at Site A and 4.94 at CBASP ⫺0.26 ⫺0.82 0.07 ⬍.001
Site B (scale range: 1–5). Predicting MC
(Intercept) 0.02 0.03 .39
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

MBCT ⫺0.11 ⫺0.40 0.08 .18

Treatment Effects on Depressive Symptoms ⫺0.01 ⫺0.04
This document is copyrighted by the American Psychological Association or one of its allied publishers.

CBASP 0.07 .86

Descriptive statistics of posttreatment HAM-D and BDI scores
can be found in Table 2. To ensure the adequate interpretability of
the model parameters retrieved from the SEM used to analyze
HAM-D and BDI, longitudinal measurement invariance was tested
using the Wald test to compare nested models. This revealed that
the model assuming strict factorial invariance was not inferior to
the model assuming only configural invariance (W ⫽ 2.83, df ⫽ 5,
p ⫽ .73). The final model incorporating strict invariance across
measurement occasions showed good overall model fit across the
50 imputations (average ␹2 ⫽ 25.82, SD ␹2 ⫽ 3.06, df ⫽ 50,
average root square error of approximation ⫽ .001, SD root mean
square error of approximation (RSMEA) ⫽ .005, average compar-
ative fit index ⫽ 1.000, SD CFI ⫽ .001).
The results from the final model were used to compute the
model-implied reliabilities as the proportion of the manifest vari-
ance attributable to true-score variance.1 These indicate good
reliabilities for HAM-D in the assessment of states (reliabilities ⫽
.79) as well as of changes across two occasions (reliabilities ⫽
.72). Results for BDI indicate even better reliabilities with a
reliability of .87 in state-assessment and .80 in the assessment of
change. The consistencies of BDI indicate that roughly half of the
reliable variance of states was shared with the assessment via
HAM-D (consistencies ⫽ .52). A slightly higher consistency was
found for the assessment of change (consistencies ⫽ .59). These
results concerning reliability of HAM-D and BDI as well as their
convergent validity are in line with previous findings (e.g., Uher et
al., 2008) with the addition that roughly the same holds true when
assessing change using these scales.
The results of the regression analysis predicting the change
measured by the reference method (HAM-D) as well as the change
of the method effects of the BDI are shown in Table 3. TAU
showed significant mean changes between pre- and posttreatment.
MBCT did not show a significantly stronger effect than TAU
(effect size ⫽ .29). In contrast, a significant intervention effect of
CBASP indicates that CBASP led to a significantly greater de-
crease in HAM-D scores than TAU. As indicated by the second
column of Table 3, the average change in CBASP participants was
roughly .82 SDs larger than that of the TAU group. It should be
noted that these effect sizes are likely to be larger than those of
comparable studies because they pertain to latent variables, mean-
ing that variability due to measurement error is not included in
their computation.2 Although the direct comparison of MBCT and
CBASP revealed no significant difference between the effects of
the two intervention approaches, we observed a statistical trend
Note. HAM-D ⫽ Hamilton Rating Scale for Depression; MBCT ⫽
mindfulness-based cognitive therapy; CBASP ⫽ cognitive behavioral anal-
ysis system of psychotherapy.
a
Only dependent variables were standardized.
favoring CBASP over MBCT (estimate ⫽ 0.17, SE ⫽ 0.09, p ⫽
.06).
The analyses concerning the method factors revealed no signif-
icant effects on the BDI above and beyond those found for the
HAM-D. It should be noted, however, that analyses implementing
the traditional HLM approach for HAM-D and BDI separately,
revealed some differences. As expected, the results for the
HAM-D very closely approximated those of the SEM analysis
reported here (bCBASP ⫽ ⫺0.28, SE ⫽ 0.07, p ⬍ .001;
bMBCT ⫽ ⫺0.08, SE ⫽ 0.07, p ⫽ .22). The HLM investigating the
BDI independently of the HAM-D revealed a very similar effect
for CBASP (bCBASP ⫽ ⫺0.24, SE ⫽ 0.09, p ⫽ .01). The effect of
MBCT however, was also significant when investigating the BDI
alone (bMBCT ⫽ ⫺0.23, SE ⫽ 0.11, p ⫽ .04) indicating that
MBCT was superior to TAU.
The first state and the change in the reference-measured states
were not significantly correlated when controlling for intervention
effects, r ⫽ .20, p ⫽ .20.
Site Differences for Treatment Effects on
Depressive Symptoms
To ensure the interpretability of the parameter’s loadings, inter-
cepts and residual variances were assumed to be equal across sites
1
State assessment reliabilities were computed as rel共Y兲 ⫽
␭2Svar共St兲 ⫹ ␭2M var共M t兲
and change reliabilities were computed
␭2Svar共St兲⫹␭2M var共M t兲⫹var共⑀兲
␭Svar共SC兲⫹␭ M var共M C兲
2 2
rel共Y C兲 ⫽ 2 for indicators of the BDI.
␭Svar共SC兲⫹␭2M var共M C兲⫹var共⑀1⫹⑀2兲
For indicators of the HAM-D (the reference method) the terms pertain-
ing to the method factors M were omitted.
2
These effect sizes are comparable to Cohen’s d, in that they also depict
differences between groups in relation to a pooled estimate of the standard
deviation. The average of the classical Cohen’s d across all 50 data
imputations when investigating only the HAM-D was d ⫽ 0.25 for the
comparison between MBCT and TAU and d ⫽ 0.85 for the comparison
between CBASP and TAU. For the BDI the manifest effect size estimates
were d ⫽ .51 for the comparison between MBCT and TAU and d ⫽ .54 for
the comparison between CBASP and TAU.
 MBCT AND CBASP FOR CHRONICALLY DEPRESSED PATIENTS 959

in addition to the strict factorial invariance across measurement Table 5

occasions. The assumptions were tested using the Wald test (W ⫽ Fixed Effects of the Multilevel Analyses Investigating the Four
5.16, df ⫽ 3, p ⫽ .16), which showed no significant reduction in Mental Health Subscales of the Short Form Health Survey and
model fit. the SASS
The Wald test for the overall equality of the intervention effects
did not reveal significant site differences in the overall regression Estimate SE Effect sizea
models. However there was a statistical trend toward site effects Vitality
(W ⫽ 12.55, df ⫽ 6, p ⫽ .051). Therefore, we conducted a more (Intercept) 0.40 0.48
detailed comparison of the sites (see Table 4), which revealed MBCT 1.12 0.90 0.31
significant intervention effects in the TAU group at Site A, but not CBASP 1.32 0.70 0.36
Social functioning
at Site B. Moreover, the analyses revealed that MBCT did not (Intercept) 0.05 0.28
show significant intervention effects above and beyond those of MBCT 0.27 0.43 0.12
TAU at Site A, but did at Site B. CBASP showed significant CBASP 0.74 0.44 0.34
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

effects at both sites. Role-emotional

(Intercept) 0.00 0.12
This document is copyrighted by the American Psychological Association or one of its allied publishers.

MBCT 0.52ⴱ 0.26 0.48

Remission Rates CBASP 0.46 0.24 0.42
Mental health
In the MBCT group, 16.7% (n ⫽ 6) of the patients remitted (Site (Intercept) 0.56 0.69
A: 8.3%, n ⫽ 2; Site B: 33.3%, n ⫽ 4), in the CBASP group 25.7% MBCT 0.63 0.95 0.14
CBASP 1.25 0.89 0.28
(n ⫽ 9) (Site A: 21.7%, n ⫽ 5; Site B: 33.3%, n ⫽ 4), and in the
SASS
TAU group 5.7% (n ⫽ 2) (Site A: 4.3% [n ⫽ 1]; Site B: 8.3%, n ⫽ (Intercept) ⫺1.17 0.74
1). Although the remission rates in the MBCT and TAU condition MBCT 2.43ⴱ 1.01 0.57
did not differ significantly (␹2 ⫽ 2.13, df ⫽ 1, p ⫽ .15), the CBASP 1.68 0.93 0.40
remission rate in the CBASP condition was significantly higher Note. MBCT ⫽ mindfulness-based cognitive therapy; CBASP ⫽ cogni-
than in TAU (␹2 ⫽ 5.86, df ⫽ 1, p ⫽ .02). Remission rates of tive behavioral analysis system of psychotherapy; SASS ⫽ Social Adap-
MBCT and CBASP did not differ significantly (␹2 ⫽ 0.87, df ⫽ 1, tation Self-Evaluation Scale.
a
p ⫽ .35). Only dependent variables were standardized.
ⴱ
p ⬍ .05.

Treatment Effects on Social Functioning

and Quality of Life
Table 5 shows the intervention effects on social functioning
(SASS) and quality of life (Mental Health subscales of SF-36) as
well as the model implied effect sizes (difference in standard
deviations of change) for each contrast to TAU. Effect sizes for
these secondary outcomes were small to medium for both MBCT
and CBASP. In the overall sample MBCT differed significantly
from TAU in changes on the Role-Emotional subscale of SF-36
and SASS. We observed no significant effects of CBASP on any
of these secondary outcomes.
Discussion
The results of the present study suggest that with an effect size
of 0.82 an eight-session group version of CBASP as a supplement
to standard treatment is effective for patients suffering from
chronic major depression. In the overall sample as well as at both
treatment sites, CBASP was significantly more effective than TAU
alone in reducing clinician-rated depressive symptoms. Results for
MBCT were more mixed. In the overall sample, MBCT was no
more effective than TAU in reducing depressive symptoms as-
sessed by HAM-D. Although the overall site effect was not sig-

Table 4
Site Comparisons From the Multiple-Group SEM With the Regression Predicting the True-
Changes of the HAM-D (Reference-Measured States; SC) and BDI Specific Components (Method
Factors MC) by Intervention Type

Site A Site B Difference

Estimate p value Estimate p value Estimate p value

Predicting SC
(Intercept) ⫺0.16 ⬍.001 ⫺0.07 .27 ⫺0.08 .25
MBCT ⫺0.02 .76 ⫺0.22 .01 0.20 .06
CBASP ⫺0.17 .01 ⫺0.43 ⬍.001 0.26 .06
Predicting MC
(Intercept) 0.03 .37 0.02 .68 0.01 .90
MBCT ⫺0.16 .06 0.00 .99 ⫺0.16 .29
CBASP ⫺0.02 .80 0.02 .81 ⫺0.04 .68
Note. HAM-D ⫽ Hamilton Rating Scale for Depression; BDI ⫽ Beck Depression Inventory; MBCT ⫽
mindfulness-based cognitive therapy; CBASP ⫽ cognitive behavioral analysis system of psychotherapy.
 960 MICHALAK, SCHULTZE, HEIDENREICH, AND SCHRAMM
nificant, the more fine-grained analyses revealed that this result
was largely attributable to very small improvements in HAM-D
scores and large TAU effects for MBCT patients treated at Site A,
whereas MBCT was more effective than TAU at Site B. Remark-
ably, this result cannot be attributed to allegiance effects since Site
A was more specialized in MBCT. Though the difference between
the effects of MBCT and CBASP was not statistically significant
(possibly due to insufficient power to detect differences) there is a
preliminary indication of a larger effect of CBASP.
Self-reported depressive symptoms (BDI) and clinicians’ ratings
(HAM-D) showed good consistency in state assessments as well as
in their assessments of changes. When analyzing our data with a
SEM approach, we found no indication that the effects differed
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
within any of the three intervention approaches by method of
This document is copyrighted by the American Psychological Association or one of its allied publishers.
assessment (as seen by the absence of significant regression
weights on the method effects Mc). This indicates consistent in-
tervention effects, the detectability of which does not depend on
the assessment method but is evident across self-reports (BDI) and
clinicians’ ratings (HAM-D).
However, the supplementary analysis with the more traditional
HLM approach revealed that when treatment effects were assessed
by self-report with the BDI, MBCT was superior to TAU. All other
HLM analyses revealed similar effects to those of the SEM ap-
proach. Although these indicate some differences between the
effects for BDI and HAM-D, overall CBASP showed more con-
sistent effects.
The effect size for the reduction in depressive symptoms was
particularly large at Site B, the center with more experience in
CBASP, and comparable to effect sizes reported in previous re-
search on CBASP. In the study by Keller and colleagues (2000),
the remission rates for individual CBASP (33%) after 12 weeks
were similar to our results (33% in the CBASP group at Site B,
26% overall across both sites) after 8 weeks. However, it should be
noted that the majority of patients in our study were on stable
medication. The combined nefazodone and CBASP condition in
the Keller et al. study achieved a higher remission rate of 48%. The
results suggest that CBASP in a group modality is a useful and
potentially more economical alternative to individual CBASP
treatment.
In contrast, the effects of MBCT in our study were smaller than
those reported in previous trials with chronically depressed pa-
tients (Barnhofer et al., 2009; Eisendrath et al., 2008; Kenny &
Williams, 2007). For example, in the Kenny and Williams study
with treatment-resistant depressed patients, the prepost effect size
for BDI was d ⫽ 1.04. However, effect size estimates in previous
studies were based on smaller samples or derived from uncon-
trolled designs. In our study, adherence and competence were
good, making it unlikely that suboptimal application of MBCT
caused the smaller effects. One difference, however, was that with
up to six patients per group, the MBCT condition in the present
study had a smaller group size than typically used in MBCT (12 to
15 patients). It might be that the relatively small number of patients
in the MBCT groups sometimes hindered optimal exchange be-
tween patients in the groups, because only a few patients contrib-
uted to the group discussions. However, it also seems possible that
the avoidant style of chronically depressed patients might reduce
the efficacy of lager groups. Because the relationship between
group size and outcome in group therapy is empirically unexplored
(Burlingame, Fuhriman & Mosier, 2003), particularly the effect of
group size in MBCT, this issue should be explored in future
research.
One difference between treatment sites was the effect of TAU.
Although the differences were not significant in the direct com-
parison, at Site B TAU had nearly no effect, but at Site A TAU
resulted in significant improvements in depressive symptoms as-
sessed with both HAM-D and BDI. This difference might be
attributable to differences in recruitment strategies between the
centers. Patients at Site A were recruited by media announcements,
at Site B they were recruited from health care facilities or private
practices. Although we found no significant differences with re-
spect to medication or supplementary individual psychotherapy in
the TAU condition between the two centers, it could be spec-
ulated that patients at Site B might have had a longer history of
ineffective TAU. This might have motivated health care pro-
viders at Site B to refer patients to our study with alternative
treatment approaches and might have had the effect that the
patients were highly resistant to TAU. The differences in sub-
type of depression observed between the two sites, namely the
higher proportion of patients with chronic major depression at
Site B and more patients with recurrent depression with incom-
plete remission (i.e., a more fluctuating type of depression) at
Site A, support this assumption. Notably, differences in efficacy
between treatment sites have been reported repeatedly in psy-
chotherapy RCTs in depression, for example, in the NIMH
Treatment of Depression Collaborative Research Program (El-
kin et al., 1989) and in a recent dismantling study on MBCT
(Williams et al., 2014).
The difference in TAU between treatment sites affected com-
parisons with CBASP and MBCT and resulted in larger estimates
of between-groups effect sizes for the treatment groups at Site B.
In particular, the effect size of MBCT was dependent on treatment
site. These differences between centers could not be attributed to
differences in adherence or competence. We found good adherence
and competence ratings for both MBCT and CBASP at both
treatment sites. Moreover, therapists at both centers had intensive
training in MBCT and CBASP. In particular, the MBCT therapist
at Site A had considerable experience in conducting MBCT
groups. During the trial the quality of each treatment was moni-
tored by the senior investigators. Moreover, interrater correlations
for HAM-D were excellent, making it unlikely that systematic
differences in rating procedures were responsible for the differ-
ences in HAM-D effects between centers.
The more consistent effect of CBASP on depressive symptoms
in our study might indicate that besides common factors like
alliance or group cohesion some specific factors are important in
the treatment of chronically depressed patients. It could be spec-
ulated that the more active, problem-solving, and interpersonal
focus of CBASP might have some benefit for this highly avoidant
group of patients (Brockmeyer, Kulessa, Hautzinger, Bents, &
Backenstrass, 2015). It might assist them to form a representation
of the effects of their behavior in an interpersonal context and
might help them to develop new and more satisfying ways to
establish and maintain important relationships. The active and
problem-focused style of CBASP might be beneficial because it
provides patients with immediate feedback on the impact their
behavior has on others. Moreover, focusing their attention on
present-moment interaction might prevent them from drifting into
ruminative states of mind. In contrast, because the mindfulness
 MBCT AND CBASP FOR CHRONICALLY DEPRESSED PATIENTS
exercises during the sessions and the regular mindfulness practice
at home as core components of MBCT require a relatively high
degree of attentional and motivational self-regulation, some pa-
tients with chronic depression might be overchallenged by the
requirements of MBCT. This might have resulted in a more limited
efficacy of MBCT in our study. If these differences in the efficacy
of MBCT and CBASP can be replicated in future studies, this
might call into question the assumption of the so-called dodo-bird
verdict (Luborsky, Singer, & Luborsky, 1975) that all psychother-
apies, regardless of their specific components, produce equivalent
outcomes.
Effects of both psychotherapeutic approaches on social func-
tioning and quality of life were moderate. In contrast to the effect
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
of CBASP on depressive symptoms we observed no significant
This document is copyrighted by the American Psychological Association or one of its allied publishers.
effect of CBASP on social functioning or any subscale of SF-36.
Research on the sequencing of change during psychotherapy has
shown that an enhancement of life functioning usually occurs in
later phases of treatment after a phase of reduction in symptom-
atology (see phase model of psychotherapy by Howard, Lueger,
Maling, & Martinovich, 1993). Correspondingly, it might be that a
higher dose of treatment would be needed before an enhancement
of life functioning could be observed in this difficult to treat group
of chronically depressed patients.
Dropout rates were relatively high with approximately 28% of
patients discontinuing treatment in both group treatment condi-
tions. These numbers are considerably higher than those reported
in previous studies of MBCT in chronically depressed patients
(e.g., Barnhofer et al., 2009 12.5%; Eisendrath et al., 2008, 7%;
Kenny & Williams, 2007, 1%) but are comparable to those re-
ported in previous trials of CBASP (e.g., Keller et al., 2000: 24%;
Wiersma et al., 2014; 25%). A meta-analysis of dropout rates in
psychotherapy reported an average rate of discontinuation of
19.7% for therapies conducted in group settings (Swift & Green-
berg, 2012), lower than the rate in our study. It might be speculated
that chronically depressed patients might be especially prone to
premature termination because of low levels of energy and moti-
vational deficits, and a high tendency toward interpersonal avoid-
ance. However, the discrepancy between the dropout rates in our
study and those reported in previous MBCT studies remains puz-
zling.
The present study has several limitations. First, the TAU con-
dition was relatively unspecified. The majority of patients received
antidepressant medication and some received treatment with addi-
tional individual psychotherapy. However, we did not control for
the quality of supplementary treatments. Correspondingly, TAU
here represents the kind of treatment that chronically depressed
patients in Germany usually receive (which is overall of high
quality), but it is unlikely that it represents the optimal treatment
for chronically depressed patients according to current clinical
guidelines. Second, although the number of patients receiving
antidepressant medication did not differ significantly between
treatment conditions, descriptively a smaller proportion of patients
exclusively treated with TAU received antidepressant medication.
Although, by definition, chronically depressed patients show a
relatively stable level of symptoms it cannot be ruled out that this
difference affected our results. Third, we did not perform a non-
inferiority analysis (Piaggio et al., 2006) to assess whether MBCT
is less effective than CBASP, the treatment approach specifically
developed for the treatment of chronically depressed patients.
961
Although in the present study the effects of CBASP were more
consistent than those of MBCT, our trial was underpowered to
reliably detect differences between the two treatment conditions.
Fourth, the results in the TAU condition suggest that recruitment
of patients might have affected the findings. Hence, future research
should use optimized recruitment strategies (e.g., searching in
computerized practice databases in primary care settings to iden-
tify patients; see Kuyken et al., 2008) to improve generalizability
of the results. Fifth, it cannot be ruled out that the greater effects
of CBASP were attributable to a higher dose of treatment. Patients
in the CBASP condition received two individual preclass sessions
due to the necessity of assessing the significant other history
before the groups started, compared to one preclass interview in
the MBCT condition. Moreover, the individual preclass interview
in the MBCT condition had a focus on informing patients about the
program, whereas in individual sessions in the CBASP condition
the transference hypothesis was derived. Thus, in addition to the
difference in dose, the individual CBASP sessions might have had
a stronger therapeutic impact.
In conclusion, the results of the present study demonstrate the
efficacy of CBASP in a group format for the treatment of chronically
depressed patients. MBCT was effective in one of the study centers
but not in the other. It is unlikely that potential allegiance or differ-
ences in adherence between conditions were responsible for these
effects. Although these findings need replication, the results of this
study might inform future research on the possible role of the more
active, problem-solving, and interpersonal focus of CBASP for the
successful treatment of chronically depressed patients.
References
American Psychiatric Association. (1994). Diagnostic and statistical man-
ual of mental disorders (4th ed.). Washington, DC: Author.
American Psychiatric Association. (2013). Diagnostic and statistical man-
ual of mental disorders (5th ed.). Washington, DC: Author.
Asparouhov, T. (2004). Stratification in multivariate modeling (Mplus
Web Notes: No. 9). Retrieved from http://www.statmodel.com/download/
webnotes/MplusNote921.pdf
Asparouhov, T., & Muthén, B. O. (2010). Multiple imputation with Mplus:
Technical appendix. Retrieved from www.statmodel.com/download/
Imputations7.pdf
Baldwin, S. A., Murray, D. M., & Shadish, W. R. (2005). Empirically
supported treatments or type I errors? Problems with the analysis of data
from group-administered treatments. Journal of Consulting and Clinical
Psychology, 73, 924 –935. http://dx.doi.org/10.1037/0022-006X.73.5
.924
Barnhofer, T., Crane, C., Hargus, E., Amarasinghe, M., Winder, R., &
Williams, J. M. G. (2009). Mindfulness-based cognitive therapy as a
treatment for chronic depression: A preliminary study. Behaviour Re-
search and Therapy, 47, 366 –373. http://dx.doi.org/10.1016/j.brat.2009
.01.019
Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Manual for the Beck
Depression Inventory (2nd ed.). San Antonio, TX: The Psychological
Corporation.
Beck, A. T., Steer, R. A., & Carbin, M. G. (1988). Psychometric properties
of the Beck Depression Inventory: Twenty-five years of evaluation.
Clinical Psychology Review, 8, 77–100. http://dx.doi.org/10.1016/0272-
7358(88)90050-5
Bishop, M., Lau, S., Shapiro, L., Carlson, N. D., Anderson, J., Carmody,
J., . . . Devins, G. (2004). Mindfulness: A proposed operational defini-
tion. Clinical Psychology: Science and Practice, 11, 230 –241. http://dx
.doi.org/10.1093/clipsy.bph077
 962 MICHALAK, SCHULTZE, HEIDENREICH, AND SCHRAMM
Bosc, M., Dubini, A., & Polin, V. (1997). Development and validation of
a social functioning scale, the Social Adaptation Self-evaluation Scale.
European Neuropsychopharmacology, 7(1, Suppl 1), S57–S70. http://dx
.doi.org/10.1016/S0924-977X(97)00420-3
Brockmeyer, T., Kulessa, D., Hautzinger, M., Bents, H., & Backenstrass,
M. (2015). Differentiating early-onset chronic depression from episodic
depression in terms of cognitive-behavioral and emotional avoidance.
Journal of Affective Disorders, 175, 418 – 423. http://dx.doi.org/10.1016/
j.jad.2015.01.045
Burlingame, G. M., Fuhriman, A., & Mosier, J. (2003). The differential
effectiveness of group psychotherapy: A meta-analytic perspective.
Group Dynamics: Theory, Research, and Practice, 7, 3–12. http://dx.doi
.org/10.1037/1089-2699.7.1.3
Crane, R. S., Eames, C., Kuyken, W., Hastings, R. P., Williams, J. M. G.,
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Bartley, T., . . . Surawy, C. (2013). Development and validation of the
This document is copyrighted by the American Psychological Association or one of its allied publishers.
mindfulness-based interventions: Teaching assessment criteria (MBI:
TAC). Assessment, 20, 681– 688. http://dx.doi.org/10.1177/
1073191113490790
Duschek, S., Schandry, R., & Hege, B. (2003). Soziale Aktivität
Selbstbeurteilungs-Skala (SASS). Diagnostik sozialer Funktionsstörun-
gen bei depressiven Störungen [Social Activity Self-report Scale
(SASS). Diagnostics of deficits in social functioning in depressive
disorders.]. Göttingen, Germany: Beltz Test.
Eid, M. (2000). A multitrait-multimethod model with minimal assump-
tions. Psychometrika, 65, 241–261. http://dx.doi.org/10.1007/
BF02294377
Eisendrath, S. J., Delucchi, K., Bitner, R., Fenimore, P., Smit, M., &
McLane, M. (2008). Mindfulness-based cognitive therapy for treatment-
resistant depression: A pilot study. Psychotherapy and Psychosomatics,
77, 319 –320. http://dx.doi.org/10.1159/000142525
Elkin, I., Shea, M. T., Watkins, J. T., Imber, S. D., Sotsky, S. M., Collins,
J. F., . . . Parloff, M. B. (1989). National Institute of Mental Health
Treatment of Depression Collaborative Research Program: General ef-
fectiveness of treatments. Archives of General Psychiatry, 46, 971–982.
http://dx.doi.org/10.1001/archpsyc.1989.01810110013002
Frank, E., Kupfer, D. J., Perel, J. M., Cornes, C., Jarrett, D. B., Mallinger,
A. G., . . . Grochocinski, V. J. (1990). Three-year outcomes for main-
tenance therapies in recurrent depression. Archives of General Psychi-
atry, 47, 1093–1099. http://dx.doi.org/10.1001/archpsyc.1990
.01810240013002
Gilmer, W. S., Trivedi, M. H., Rush, A. J., Wisniewski, S. R., Luther, J.,
Howland, R. H., . . . Alpert, J. (2005). Factors associated with chronic
depressive episodes: A preliminary report from the STAR-D project.
Acta Psychiatrica Scandinavica, 112, 425– 433. http://dx.doi.org/
10.1111/j.1600-0447.2005.00633.x
Guy, W. (1976). ECDEU Assessment Manual for Psychopharmacology.
Washington, DC: U.S. Department of Health, Education, and Welfare.
Hamilton, M. (1967). Development of a rating scale for primary depressive
illness. British Journal of Social and Clinical Psychology, 6, 278 –296.
http://dx.doi.org/10.1111/j.2044-8260.1967.tb00530.x
Harrison, W. M., & Stewart, J. W. (1995). Pharmacotherapy of dysthymic
disorder, in diagnosis and treatment of chronic depression. In J. H.
Kocsis & D. N. Klein (Eds.), Diagnosis and treatment of chronic
depression (pp. 124 –145). New York, NY: Guilford Press.
Hautzinger, M., Keller, F., & Kühner, C. (2006). Beck Depression Inven-
tory (2nd ed.) (BDI- II). Bern, Switzerland: Hogrefe.
Holtzheimer, P. E., & Mayberg, H. S. (2011). Deep brain stimulation for
psychiatric disorders. Annual Review of Neuroscience, 34, 289 –307.
http://dx.doi.org/10.1146/annurev-neuro-061010-113638
Howard, K. I., Lueger, R. J., Maling, M. S., & Martinovich, Z. (1993). A
phase model of psychotherapy outcome: Causal mediation of change.
Journal of Consulting and Clinical Psychology, 61, 678 – 685. http://dx
.doi.org/10.1037/0022-006X.61.4.678
Kabat-Zinn, J. (1990). Full catastrophe living: How to cope with stress,
pain and illness using mindfulness meditation. New York, NY: Dell
Publishing.
Kabat-Zinn, J. (2003). Mindfulness-based interventions in context: Past,
present, and future. Clinical Psychology: Science and Practice, 10,
144 –156. http://dx.doi.org/10.1093/clipsy.bpg016
Keller, M. B., Gelenberg, A. J., Hirschfeld, R. M. A., Rush, A. J., Thase,
M. E., Kocsis, J. H., . . . Harrison, W. M. (1998). The treatment of
chronic depression, Pt. 2: A double-blind, randomized trial of sertraline
and imipramine. Journal of Clinical Psychiatry, 59, 598 – 607.
Keller, M. B., McCullough, J. P., Klein, D. N., Arnow, B., Dunner, D. L.,
Gelenberg, A. J., . . . Zajecka, J. (2000). A comparison of nefazodone,
the cognitive behavioral-analysis system of psychotherapy, and their
combination for the treatment of chronic depression. The New England
Journal of Medicine, 342, 1462–1470. http://dx.doi.org/10.1056/
NEJM200005183422001
Kenny, M. A., & Williams, J. M. (2007). Treatment-resistant depressed
patients show a good response to mindfulness-based cognitive therapy.
Behaviour Research and Therapy, 45, 617– 625. http://dx.doi.org/
10.1016/j.brat.2006.04.008
Kiesler, D. J. (1982). The 1982 Interpersonal Circle: A taxonomy for
complementarity for human transaction. Psychological Review, 90, 185–
214. http://dx.doi.org/10.1037/0033-295X.90.3.185
Klein, D. N., Shankman, S. A., & Rose, S. (2006). Ten-year prospective
follow-up study of the naturalistic course of dysthymic disorder and
double depression. The American Journal of Psychiatry, 163, 872– 880.
http://dx.doi.org/10.1176/ajp.2006.163.5.872
Kocsis, J. H., Frances, A. J., Voss, C., Mann, J. J., Mason, B. J., &
Sweeney, J. (1988). Imipramine treatment for chronic depression. Ar-
chives of General Psychiatry, 45, 253–257. http://dx.doi.org/10.1001/
archpsyc.1988.01800270071008
Kriston, L., von Wolff, A., Westphal, A., Hölzel, L. P., & Härter, M.
(2014). Efficacy and acceptability of acute treatments for persistent
depressive disorder: A network meta-analysis. Depression and Anxiety,
31(8). http://dx.doi.org/10.1002/da.22236
Kupfer, D. J., Frank, E., Perel, J. M., Cornes, C., Mallinger, A. G., Thase,
M. E., . . . Grochocinski, V. J. (1992). Five-year outcome for mainte-
nance therapies in recurrent depression. Archives of General Psychiatry,
49, 769 –773. http://dx.doi.org/10.1001/archpsyc.1992.01820100013002
Kuyken, W., Byford, S., Taylor, R. S., Watkins, E., Holden, E., White, K.,
. . . Teasdale, J. D. (2008). Mindfulness-based cognitive therapy to
prevent relapse in recurrent depression. Journal of Consulting and
Clinical Psychology, 76, 966 –978. http://dx.doi.org/10.1037/a0013786
Kuyken, W., Watkins, E., Holden, E., White, K., Taylor, R. S., Byford, S.,
. . . Dalgleish, T. (2010). How does mindfulness-based cognitive therapy
work? Behaviour Research and Therapy, 48, 1105–1112. http://dx.doi
.org/10.1016/j.brat.2010.08.003
Luborsky, L., Diguer, L., Seligman, D. A., Rosenthal, R., Krause, E. D.,
Johnson, S., . . . Schweizer, E. (1999). The researcher’s own therapy
allegiances: A “wild card” in comparisons of treatment efficacy. Clinical
Psychology: Science and Practice, 6, 95–106. http://dx.doi.org/10.1093/
clipsy/6.1.95
Luborsky, L., Singer, B., & Luborsky, L. (1975). Comparative studies of
psychotherapies: Is it true that “everyone has won and all must have
prizes?” Archives of General Psychiatry, 32, 995–1008. http://dx.doi
.org/10.1001/archpsyc.1975.01760260059004
McCullough, J. P. (2000). Treatment for chronic depression: Cognitive
behavioral analysis system of psychotherapy (CBASP). New York, NY:
Guilford Press.
McCullough, J. P. (2003). Treatment for chronic depression: Cognitive
behavioral analysis system of psychotherapy (CBASP). New York, NY:
Guilford Press.
Moberg, P. J., Lazarus, L. W., Mesholam, R. I., Bilker, W., Chuy, I. L.,
Neyman, I., & Markvart, V. (2001). Comparison of the standard and
 MBCT AND CBASP FOR CHRONICALLY DEPRESSED PATIENTS
structured interview guide for the Hamilton Depression Rating Scale in
depressed geriatric inpatients. The American Journal of Geriatric Psy-
chiatry, 9, 35– 40. http://dx.doi.org/10.1097/00019442-200102000-
00006
Morfeld, M., Kirchberger, I., & Bullinger, M. (2011). Fragebogen zum
Gesundheitszustand (SF-36) [State of Health Questionnaire (SF-36)]
(2nd ed.). Göttingen, Germany: Hogrefe Verlag.
Murphy, J. A., & Byrne, G. J. (2012). Prevalence and correlates of the
proposed DSM–5 diagnosis of chronic depressive disorder. Journal of
Affective Disorders, 139, 172–180. http://dx.doi.org/10.1016/j.jad.2012
.01.033
Murray, C. J. L., Vos, T., Lozano, R., Naghavi, M., Flaxman, A. D.,
Michaud, C., . . . Memish, Z. A. (2012). Disability-adjusted life years
(DALYs) for 291 diseases and injuries in 21 regions, 1990 –2010: A
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
systematic analysis for the Global Burden of Disease Study 2010.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Lancet, 380, 2197–2223. http://dx.doi.org/10.1016/S0140-
6736(12)61689-4
Muthén, L. K., & Muthén, B. O. (1998 –2013). Mplus user’s guide (7th
ed.). Los Angeles, CA: Muthén & Muthén.
Neff, K. D. (2003). The development and validation of a scale to measure
self-compassion. Self and Identity, 2, 223–250. http://dx.doi.org/
10.1080/15298860309027
Piaggio, G., Elbourne, D. R., Altman, D. G., Pocock, S. J., Evans, S. J. W.,
& CONSORT Group. (2006). Reporting of noninferiority and equiva-
lence randomized trials: An extension of the CONSORT statement.
Journal of the American Medical Association, 295, 1152–1160. http://
dx.doi.org/10.1001/jama.295.10.1152
Piet, J., & Hougaard, E. (2011). The effect of mindfulness-based cognitive
therapy for prevention of relapse in recurrent major depressive disorder:
A systematic review and meta-analysis. Clinical Psychology Review, 31,
1032–1040. http://dx.doi.org/10.1016/j.cpr.2011.05.002
R Development Core Team. (2013). R: A language and environment for
statistical computing. Vienna, Austria: R Foundation for Statistical
Computing. Retrieved from http://www.R-project.org
Richards, D. (2011). Prevalence and clinical course of depression: A
review. Clinical Psychology Review, 31, 1117–1125. http://dx.doi.org/
10.1016/j.cpr.2011.07.004
Satyanarayana, S., Enns, M. W., Cox, B. J., & Sareen, J. (2009). Preva-
lence and correlates of chronic depression in the Canadian community
health survey: Mental health and well-being. The Canadian Journal of
Psychiatry/La Revue canadienne de psychiatrie, 54, 389 –398.
Sayegh, L., Locke, K. D., Pistilli, D., Penberthy, J. K., Chachamovich, E.,
McCullough, J. P., & Turecki, G. (2012). Cognitive behavioural analysis
system of psychotherapy for treatment resistant depression: Adaptation
to a group modality. Behaviour Change, 29, 97–108. http://dx.doi.org/
10.1017/bec.2012.2
Schramm, E., Brakemeier, E. L., & Fangmeier, R. (2012). CBASP in
Gruppen: das Kurzmanual. [CBASP in groups: the short manual.].
Stuttgart, Germany: Schattauer.
Schramm, E., Zobel, I., Dykierek, P., Kech, S., Brakemeier, E. L., Külz, A.,
& Berger, M. (2011). Cognitive behavioral analysis system of psycho-
therapy versus interpersonal psychotherapy for early-onset chronic de-
pression: A randomized pilot study. Journal of Affective Disorders, 129,
109 –116. http://dx.doi.org/10.1016/j.jad.2010.08.003
Segal, Z. V., Bieling, P., Young, T., MacQueen, G., Cooke, R., Martin, L.,
. . . Levitan, R. D. (2010). Antidepressant monotherapy vs sequential
pharmacotherapy and mindfulness-based cognitive therapy, or placebo,
for relapse prophylaxis in recurrent depression. Archives of General
Psychiatry, 67, 1256 –1264. http://dx.doi.org/10.1001/archgenpsychiatry
.2010.168
963
Segal, Z. V., Teasdale, J. D., Williams, J. M., & Gemar, M. C. (2002). The
Mindfulness-Based Cognitive Therapy Adherence Scale: Inter-rater re-
liability, adherence to protocol and treatment distinctiveness. Clinical
Psychology & Psychotherapy, 9, 131–138. http://dx.doi.org/10.1002/cpp
.320
Segal, Z. V., Williams, J. M. G., & Teasdale, J. D. (2002). Mindfulness-
based cognitive therapy for depression: A new approach to preventing
relapse. New York, NY: Guilford Press.
Steyer, R., Eid, M., & Schwenkmezger, P. (1997). Modeling true intrain-
dividual change: True change as a latent variable. Methods of Psycho-
logical Research—Online, 2, 21–33.
Swift, J. K., & Greenberg, R. P. (2012). Premature discontinuation in adult
psychotherapy: A meta-analysis. Journal of Consulting and Clinical
Psychology, 80, 547–559. http://dx.doi.org/10.1037/a0028226
Teasdale, J. D., Moore, R. G., Hayhurst, H., Pope, M., Williams, S., &
Segal, Z. V. (2002). Metacognitive awareness and prevention of relapse
in depression: Empirical evidence. Journal of Consulting and Clinical
Psychology, 70, 275–287. http://dx.doi.org/10.1037/0022-006X.70.2
.275
Teasdale, J. D., Segal, Z. V., Williams, J. M. G., Ridgeway, V. A., Soulsby,
J. M., & Lau, M. A. (2000). Prevention of relapse/recurrence in major
depression by mindfulness-based cognitive therapy. Journal of Consult-
ing and Clinical Psychology, 68, 615– 623. http://dx.doi.org/10.1037/
0022-006X.68.4.615
Thase, M. E., Fava, M., Halbreich, U., Kocsis, J. H., Koran, L., Davidson,
J., . . . Harrison, W. (1996). A placebo-controlled, randomized clinical
trial comparing sertraline and imipramine for the treatment of dysthy-
mia. Archives of General Psychiatry, 53, 777–784. http://dx.doi.org/
10.1001/archpsyc.1996.01830090023004
Torpey, D. C., & Klein, D. N. (2008). Chronic depression: Update on
classification and treatment. Current Psychiatry Reports, 10, 458 – 464.
http://dx.doi.org/10.1007/s11920-008-0074-6
Uher, R., Farmer, A., Maier, W., Rietschel, M., Hauser, J., Marusic, A., . . .
Aitchison, K. J. (2008). Measuring depression: Comparison and integra-
tion of three scales in the GENDEP study. Psychological Medicine, 38,
289 –300. http://dx.doi.org/10.1017/S0033291707001730
Ware, J. E., Jr., & Sherbourne, C. D. (1992). The MOS 36-item short-form
health survey (SF-36): I. Conceptual framework and item selection.
Medical Care, 30, 473– 483. http://dx.doi.org/10.1097/00005650-
199206000-00002
Wiersma, J. E., Van Schaik, D. J., Hoogendorn, A. W., Dekker, J. J., Van,
H. L., Schoevers, R. A., . . . Van Oppen, P. (2014). The effectiveness of
the cognitive behavioral analysis system of psychotherapy for chronic
depression: A randomized controlled trial. Psychotherapy and Psycho-
somatics, 83, 263–269. http://dx.doi.org/10.1159/000360795
Williams, J. M., Crane, C., Barnhofer, T., Brennan, K., Duggan, D. S.,
Fennell, M. J., . . . Russell, I. T. (2014). Mindfulness-based cognitive
therapy for preventing relapse in recurrent depression: A randomized
dismantling trial. Journal of Consulting and Clinical Psychology, 82,
275–286. http://dx.doi.org/10.1037/a0035036
Wittchen, H.-U., Wunderlich, U., Gruschwitz, S., & Zaudig, M. (1997).
SKID I. Strukturiertes klinisches Interview für DSM–IV [Structured
Clinical Interview for DSM-IV.]. Göttingen, Germany: Hogrefe.
Young, A. S., Klap, R., Shoai, R., & Wells, K. B. (2008). Persistent
depression and anxiety in the United States: Prevalence and quality of
care. Psychiatric Services, 59, 1391–1398. http://dx.doi.org/10.1176/ps
.2008.59.12.1391
Received October 24, 2014
Revision received June 5, 2015
Accepted June 10, 2015 䡲