*Multi-Site RCT of Group Schema Therapy for Borderline Personality Disorder*

A recent meta-analytic study did not think it possible to draw conclusions
about the relative effectiveness of different psychotherapy approaches for BPD
(McMain & Pos, 2007). The just released UK National Institute for Health and
Clinical Excellence guidelines for BPD treatment and research (2009) suggest
that many questions remain unanswered regarding descriptions of sub-populations
of BPD, what comparator to use, and even which intervention is the most likely
candidate for further research investment. Most reviews emphasize that the
current RCTs of all models of BPD treatment are under-powered. Other problems
noted include: a lack of rigorous comparator controls, inadequate length of
follow-up periods to measure the maintenance of gains, confounding effects from
concurrent active treatment. In addition, research for most approaches lacks
adequate cost-effectiveness analysis, lacks replication or lacks feasibility of
implementation in public health settings where the majority of BPD patients are
treated.
Schema Therapy has demonstrated significant effects with all aspects of BPD in
two RCTs of individual treatment (Giesen-Bloo et al., 2006; Nadort et al.,
2008, 2009), one RCT of group treatment (Farrell et al., 2009), and one
case-series (Nordahl & Nysaeter, 2005). Drop-out from ST is very low (0% over
an eight month group and 20% over two years of individual therapy). Moreover,
ST has demonstrated positive effects on quality of life (Giesen-Bloo et al.,
2006; van Asselt et al., 2008; Nadort et al., 2008, 2009; Farrell, et al.,
2009) and decreases in societal costs (van Asselt et al., 2008). Schema Therapy
distinguishes itself by producing improvements in all nine criteria of BPD, as
well as functioning and quality of life. It also demonstrates cost
effectiveness, high retention of subjects, and high patient and therapist
satisfaction with the treatment. The range of effectiveness of ST, the size of
treatment effects demonstrated in RCTs, including the significant findings for
ST in groups,make a strong case for further testing of ST in clinical trials.

Our long-range goal is to develop a comprehensive, clinically and cost
effective treatment for BPD that produces both symptom remission and functional
recovery. We aim to determine the role that group ST can have in the
comprehensive treatment of BPD, and whether it is a cost effective treatment,
that can be made widely available in public and private healthcare settings.
Our central hypothesis is that eighteen months of group ST delivered with two
varied amounts of individual sessions will significantly outperform a control
condition of treatment as usual (TAU). We base this hypothesis on the results
of randomized controlled trials (RCTs) carried out and published by members of
our research team that have demonstrated the broad and pronounced effectiveness
of individual and group ST. The large effect sizes in an RCT of schema therapy
in group, suggest that group interventions for people with BPD may have
specific advantages and power that need to be further evaluated empirically. We
chose TAU as a comparator because it has not been used before in trials of ST
as a complete treatment, and it tests important questions regarding whether a
treatment is feasible and can be cost effective.

We plan to test our central hypothesis and accomplish the overall objective of
this application by pursuing the following specific aims:

1. Evaluate the effectiveness of group schema therapy as a core component of
comprehensive treatment for BPD. The working hypothesis for this aim is that a
combination of individual and group ST will be superior to TAU in ameliorating
core BPD symptoms, enhancing social and occupational or academic functioning,
and improving quality of life. Two combinations of individual and group
sessions were developed for this trial based upon preliminary studies and the
collective clinical experience of our research group. Comparison of the two ST
conditions will allow us to begin to determine the minimal *dose* of more
expensive individual therapy needed to support the effect of ST group
treatment. Our paradigm is that the use of group therapy, with individual
therapy as a limited adjunct, could fill a gap in public health care systems
that struggle to provide individual specialized treatments for BPD with limited
psychotherapy resources. In this process we will test a treatment manual for
group ST that will be refined, published and provide the standard for future
trials. We will also get an idea of the training and supervision needed for
psychotherapists already trained in individual ST to reach an acceptable
standard of fidelity in group ST. Our large sample, will allow us to begin to
evaluate the hypothesized active ingredients of ST (reduction in maladaptive
schemas and modes, establishment of secure attachment, and group cohesion).

2. Evaluate the cost-effectiveness of the group ST formats. The working
hypothesis for this aim is that ST in both formats will be more cost-effective
than TAU at all sites. A secondary hypothesis is that Format A will be more
cost-effective that Format B.

3. Assess the opinions of the major stakeholders (BPD patients and their
therapists) about the treatment using established qualitative methods. With
in-depth interviews and focus-groups we will get information about what
patients and therapists think about the treatment, to what degree it meets
their needs, and what is missing in their opinion from the treatments. We will
use this information as an additional source of information upon which to
decide what ST-format should be used and to further adjust the preferred
treatment to better meet the needs of patients and therapists, before its
implementation in general practice is tested.

4. Identify predictors of treatment response from the assessment measures and
repeated outcome measures employed. Our large sample size and psychometrically
sound measures will allow us to test preliminary hypotheses about predictors of
recovery and drop-out derived from the earlier RCT of individual ST. Moreover,
we can conduct post hoc analyses of other potential predictors that can be
tested a priori in a subsequent study.
Collectively, the empirical validation of our working hypotheses will add to
the progress of BPD treatment research by further evaluating ST as a promising
treatment with potential as a manualized, comprehensive, cost-effective, and
replicable treatment for the devastating disorder of BPD.

5. Investigate changes in underlying brain processes and cognitive biases.
Previous studies have found that BPD patients have increased amygdala and
hippocampal responses and reduced ACC and medial prefrontal cortex responses to
negative emotional stimuli, probably reflecting emotion regulation problems and
easily triggered fight/flight responses. Similarly, BPD patients showed
specific ACC response patterns associated with emotion-related impulsivity.
Preliminary studies suggest that successful psychological treatment leads to
normalization of these responses. The present study aims to investigate in more
detail and with more power how brain responses in pertinent area change in
association with the treatment. Other studies have documented cognitive biases
towards threat, i.e. attentional and interpretational biases. Changes in these
biases during treatment will be assessed and related to recovery of BPD and to
changes in emotional brain responses. To interpret the baseline fMRI and
threat-bias findings the BPD-group will be compared to clinical controls
(Cluster-C PD patients) and nonpatients. This fundamental substudy will take
place in 3 academic centres (Maastricht, Lübeck and Freiburg). In each center
at least N=10 of each control group will be assessed.

To further evaluate the effectiveness of group ST with two schedules of
individual ST we will conduct a multi-center RCT with adequate power to reach
the following aims:
Specific Aim 1: Evaluate the comparative contributions of group and individual
schema therapy to comprehensive treatment that leads to symptom remission and
functional recovery from BPD, including meaningfully improved quality of life.
This study will expand the Giesen-Bloo et al. (2008) study and the Farrell et
al. (2009) study of group ST, by comparing two group formats to TAU, as well as
comparing the two group formats with each other. We will compare the treatment
and cost effectiveness of the two combinations of group and individual ST for
two years (see table) versus TAU at two universities with multiple clinical
sites each. The clinical sites are:
1) RIAGG Maastricht, Nederland
2) Mondriaan zorggroep, Heerlen, Nederland
3) GGZ Oost Brabant, vestiging Helmond, Nederland
4) GGZ Centraal, vestiging Hilversum, Nederland
5) De Viersprong, Amsterdam, Nederland
6) Vincent van Gogh Instituut Venray & Venlo, Nederland
7) Universitätsklinikum Freiburg, Freiburg, Duitsland
8) Institut für Verhaltenstherapie-Ausbildung, Hamburg, Duitsland
9) Klinik für Psychiatrie und Psychotherapie, Universität zu Lübeck, Duitsland
10) South Metropolitan Area Health Service, Mental Health Rockingham, Perth,
Australia
11) South Metropolitan Area Health Service, Mental Health, Peel, Perth,
Australia
12) Bradford District Care Trust, Bradford, UK.
13) South London and Maudsley NHS Foundation Trust, London, UK.
14) Department of Psychiatry, Eginition Hospital, Medical School, Athens
University, Athens, Greece.

The range of clinical settings included * e.g., community mental health
centers, university outpatient clinics, institutes, psychiatric hospital
clinics - allows an initial evaluation of the generalizability of the treatment
and lays the groundwork for future implementation studies. This is the first
international multi-site trial of a BPD treatment and it makes good use of the
combined resources and sites of the international researchers and experts in
Schema therapy.

Diagnostic DSM-IV in/exclusion criteria will be assessed with SCID-I and
SCID-II interviews executed by trained SCID interviewers. A complete medical
history and, if indicated, a physical evaluation will be conducted at this
time. (In addition, all subjects entering the study will have their saliva
collected at Baseline 2 in order to create a *bank* for the subsequent analysis
of DNA in this large sample. All collected saliva will be sent to Maastricht
University for this process and will be part of a separate study.) The
Borderline Personality Disorder Severity Index (BPDSI) will be administered at
the first assessment session (baseline 1). All centers have a regular stream of
BPD patients and it is expected that quotas at each site will be filled within
one year. No formal or paid advertising is planned. All patients with BPD or
suspicion of it will be asked to participate in the screening process. After an
independent central research assistant has checked in/exclusion criteria, the
informed consent procedure and the qualifying BPDSI have been completed,
patients will wait for a 3 month qualification period, at the end of which the
BPDSI will be administered again (baseline 2). The structured interviews and
self-report questionnaires of our assessment battery will be scheduled on a
weekly basis during the qualification period. These sessions will be with
research assistants, conducted in a supportive and validating manner to engage
the patient and to provide some support. Individual supportive sessions,
utilizing TAU and not a specialized model of BPD treatment, will also be
available to patients during this pre-randomization phase, delivered by
psychotherapists who are not part of the study treatment team. The
qualification period will filter out patients poorly motivated for the study,
and control for early changes in the primary outcome measure (BPDSI). After the
second baseline BPDSI assessment, patients will be randomized over two
conditions: group-ST or TAU. The central research assistant responsible for
randomization (located at the main Netherlands site) will randomize blocks of 2
patients over the two conditions using a computerized randomization program. In
other words, for each site this assistant waits until two new patients are
included, then prompts the computer to randomize the patients over conditions,
and communicates the condition assignment to the site coordinator. In this way
condition assignment is unpredictable up to and including the very last patient
of each site. The type of TAU patients will receive will be decided by regular
clinical practice at each site, thus representing common clinical practice. The
amount and kind of this naturalistic TAU provided will be tracked and described
in the trial results. Half of the sites will first offer ST format A (twice a
week group), and then ST format B (once a week group, once a week individual).
The other sites will offer the formats in reverse order. Thus, order of group
format type will be balanced over sites, controlling for therapist learning and
any other time effects.

1. Group Schema Therapy A (Group-focused): 124 group sessions plus 0-18
individual sessions

2. Group Schema Therapy B (Group and individual combined): 74 group sessions
plus 62 individual sessions

3. Treatment-as-usual: session number varying, individual, group, day therapy,
inpatient, pharmacotherapy

Assessments take a total of about 20 hrs. over 3 years, and will be executed by
research assistants trained to relate to the participants in a supportive way.

For Maastricht, Heerlen, Freiburg & Lübeck participants an extra 3 hours fMRI
assessment is planned, and an extra 3 hours of threat-bias assessment.

There are no direct risks involved in the interventions and in the assessments.

However, suicidality, self-injury and crisis are considered behavioral
hallmarks of BPD. The emergency procedure of each clinical site will be
followed for these emergencies. If needed, emergency hospitalization will take
place. Individual crisis management sessions will be available before
self-injury or suicide attempts occur. If contact is made with the individual
or group therapists after these behaviors, contact will only be long enough to
arrange any needed emergency treatment. The *bank* of individual sessions in
the primarily group treatment condition allows for crisis management as well as
a patient*s request for an individual session. If a patient has used all the
credit in their *bank*, in an emergency situation they can be seen by ER, the
Crisis Intervention Unit of the clinic, or a therapist at the site, who is not
part of the study. Any additional treatment, whether individual sessions or
inpatient hospitalization will be monitored and examined as a possible
predictor of adequacy of this treatment for a particular subgroup of patients.
Patients will only be dropped from the study at their request.

Participants either receive treatment as usual, or ST that is known as one of
the most effective treatments, and that does not have specific risks. Any
treatment can be emotionally confonting for BPD patients. Patients receive
information about the treatments, from which it becomes clear that they might
get group-ST. Participants are also told that ST involves processing of adverse
childhood experiences. Potential participants that don't want group treatment
or don't want a treatment partially focusing on their childhood can therefore
decide not to participate.

As to the additional threat-bias and fMRI study : there are no specific risks
with the computer task to assess threat-bias, and with the fMRI assessment
given the fact that we use the usual exclusion criteria for MRI assessments.
Because of the extra time involved in the additional study, and the extra
burden to travel to anaother building, a small financial compensation is
offered.