Short and long term effects of clinical DBT-PE for patients with severe posttraumatic stress disorder and co-morbid psychopathology

Until only a few years ago, we had little to offer in terms of evidence-based
treatment to those suffering from severe posttraumatic stress disorder (PTSD)
and co-morbid conditions such as personality disorders or substance use
disorder. There was a trend to stabilize these patients before giving them
trauma-focused treatment and a period of five years of ambulatory treatment was
not uncommon. In 2012 the Complex Trauma Task Force of the International
Society of Traumatic Stress Studies (ISTSS) proposed a stepped model with three
treatment phases in the Expert Consensus Guidelines for Complex PTSD (Cloitre,
et al., 2012): 1. Stabilization and strengthening: patient safety,
strengthening one*s capacities for emotional awareness and expression,
increasing positive self-concept, addressing feelings of guilt and shame, and
increase interpersonal and social competencies. 2. Review and reappraisal of
trauma memories: re-experiencing traumatic events in the context of an actual
and subjectively experienced safe environment. 3. Consolidation of the gains of
the treatment so far.

Some experts have questioned this somewhat conservative approach to trauma
treatment. Bicanic, De Jongh and Ten Broeke (2015), for example, argue that
there is no convincing empirical evidence justifying the stabilization phase
prior to an evidence-based trauma treatment (EMDR or Trauma-focused cognitive
behavioral therapy). In the Netherlands, a number of experts (Van Minnen, 2009;
Jongh, et al., 2016) have also suggested that trauma-focussed therapy without
prior stabilization is feasible and clinically beneficial for patients with
severe PTSD. In an intensive, residential treatment program of 2 x 4 successive
days consisting of a combination of exposure and Eye Movement Desensitisation
and Reprocessing (EMDR), Wagenmans, Van Minnen, Sleijpen and De Jongh (2018)
show that patients with severe PTSD benefit from treatment without a
stabilization phase. Additionally, differentiating in type of trauma showed
that this treatment was also beneficial for patients with a history of
childhood sexual abuse. Another study of the group proved that treatment was
effective for patients with the dissociative subtype of PTSD as well (Zoet,
Wagenmans, Van Minnen and De Jongh, 2018). Clearly, patients with severe PTSD
and dissociation benefit from trauma-focused treatment without a stabilization
phase.

In some patients, especially those with personality disorders and profound
emotional instability, deliberate self-harm and suicidal behaviour, stability
as a pre-treatment condition may be too much to ask for. Moreover, often the
symptoms and problems associated with the PTSD in turn complicate the treatment
of the personality disorder. With PTSD and personality disorders complicating
and frustrating each other's treatment, therapists of these patients had the
least to offer to those patients who needed it most, that is, until recently.

In 2013, Bohus and his co-workers from the Central Institute of Mental Health
in Mannheim published the results of a 12-week residential treatment program
that was specifically designed for patients with severe PTSD complicated by
comorbid conditions such as a Borderline Personality Disorder (BPD). The
program combines dialectical behavioural therapy (DBT), one of the best
documented interventions for BPD, with prolonged exposure (PE) to trauma
memories (Steil, 2011). Bohus et al. (2013) compared their intervention to a
waiting list condition in a single blind randomized design. The results of the
Mannheim residential program were very promising, with significant and
clinically relevant changes in the patient group that was randomized to the
active condition, a large effect size (a Hedges' g of 1.6), and little dropout.
The treatment results on the PTSD severity seem to be independent of severity
of the personality disorder; neither a diagnosis of BPD nor the severity of the
number of BPD symptoms was significantly related to treatment outcome.

The research group of Marsha M. Linehan, founder of the DBT method, published
the results of a similar combined DBT-PE outpatient treatment program. However,
with a more linear design than the Mannheim program and a treatment duration of
12 months instead of 12 weeks (Harned, Korslund, Foa & Linehan, 2012; Harned,
Korslund & Linehan, 2014; Harned, Wilks, Schmidt & Coyle, 2018). In this
program, if DBT leads to a complete disappearance of deliberate self-harm,
patients are allowed to start with prolonged exposure, while continuing with
the DBT programme. Patients reported significant reductions in PTSD severity
after treatment and there was no evidence of exacerbations of deliberate
self-harm (Harned, et al., 2012; Harned, et al., 2014). In addition, compared
to DBT alone, DBT-PE was more effective in improving health-related quality of
live and global (social) functioning (Harned et al., 2018)

In contrast to this linear model, in the Mannheim protocol exposure starts
regardless of the presence or absence of deliberate self-harm. This seems more
in line with recent insights about offering trauma-focussed treatment without a
long stabilization phase. Inspired by the promising results of the Mannheim
program, GGZ Friesland started a similar residential treatment program in
Leeuwarden. The complete staff of our department was trained by Martin Bohus
before the first patients were included in March 2014. Results of Routine
Outcome Measurements data of the residential DBT-PE program show that patients
improved significantly on PTSD symptoms after treatment (effect size: d=.73,
p<0.001).

Dropout was relatively scarce (23/241, 9.5%). This data is limited for several
reasons. It is retrospective data, with a lot of missing values. Our database
contained only 55 valid pre-post assessments of disease severity (DTS) in 241
patients. Of all patients that entered the program since 2014 only a small
group filled in the questionnaires regularly because of administrative changes
at the treatment centre.

To date, there is only one published clinical trial, with less than a total of
80 randomized patients, to substantiate the claim of success of this new
intervention (Bohus, 2013). It has been shown in numerous studies that the
results in first trials of new interventions are not necessarily replicated in
follow-up studies. Apart from publication bias that selects trials with
positive findings, the relation between effect and protocol adherence, the
effects of a particular psychological intervention may also be somewhat
dependent of the personal qualities of the founder and his or her team.
Implementation of the method by a different group, regardless of protocol
adherence, may not be as effective as the original one. To further solidify the
evidence base of this promising new method, follow-up studies are absolutely
necessary.

Furthermore, the cost-effectiveness of clinical DBT-PE has not been
investigated before. Clinical DBT-PE is a costly intervention, as it consists
of 12 weeks intensive residential care, including trauma therapy delivered by
highly skilled psychotherapists. This is a costly type of care; on the other
hand: this integrated treatment program may still prove less costly compared to
other strategies. For example, traditional outpatient treatment (treatment as
usual) with long periods of stabilization an numerous treatment sessions spread
over a couple of years may in the end not be cheaper after all. The same goes
for a hypothetical do-nothing strategy, where patients do not receive a disease
specific treatment, but nevertheless make healthcare costs that are related to
the disease, such as hospital admissions in times of crisis etcetera.

The current study aims to determine the effect of a 12-week residential DBT-PE
program on a range of different outcomes, including the severity of PTSD,
dissociation, personality disorders, and social functioning; The study also
aims to predict which patients benefit most from the program and to explore
cost-effectiveness of the program as compared to other strategies, including
treatment as usual and a do-nothing strategy.

We differentiate between short and long term effects; short term effects are
defined as the change in symptoms over the 12 weeks that the treatment program
lasts, whereas long term effects are the changes at 6 and 12 months
respectively. Unless specified otherwise, we look at both short and long term
effects.

The short term pre vs. post changes will be compared to changes that may be
seen during the waiting list period; thus participants of the study may serve
as their own controls.

The main questions of the study are:
1) What is the effect of clinical DBT-PE on the severity of PTSD?
2) Which patient characteristics predict the effect of clinical DBT-PE?
Possible predictors include: severity of PTSD symptoms at T0, degree of
dissociation, type of trauma, borderline personality disorder, cluster C
personality disorder.

Secondary questions are:
3) What is the effect of clinical DBT-PE on PTSS related phenomena like
dissociation and frequency of and urge for non-suicidal self-injury (NSSI) and
suicidal ideation?

4) What is the effect of clinical DBT-PE on the severity of co-morbid disorders
including: borderline and cluster C personality disorders?

5) How stable are the effects of clinical DBT-PE after the clinical treatment
program has stopped (i.e. comparing results at 12 weeks with those at 6 and 12
months posttreatment)?

6) What is the effect of clinical DBT-PE on social functioning?

7) What is the incremental cost-effectiveness of clinical DBT-PE as compared to
other treatment strategies including outpatient treatment as usual and a
hypothetical do-nothing strategy? We will use standard cost-effectiveness
simulation techniques, such as Markov modelling integrating data collected in
this study with literature data and date from other sources, to answer this
question. See statistical methods for details.

We propose a pre- to posttreatment observational study, in which patients are
additionally assessed at the start of the regular waiting list period. To
examine the short term effects of the treatment program, we have three
measurements: before waiting list period (at intake), immediately pretreatment
and posttreatment (12 weeks after start of treatment). Thus, we can compare the
changes during the program with the propensity of the outcomes to change over
time in the absence of a disease specific intervention (i.e., waiting list). We
plan additional follow up measurements at 6 and 12 months posttreatment to
study long term effects. For two of the variables, we propose a more frequent
sampling during the 12 week program, to be able to study the changes associated
with the treatment program in more detail: The Davidson Trauma Scale (DTS) is a
weekly self-report PTSD severity. Participants also use daily diary reports on
non-suicidal self-injury and suicidal ideation (DBT diary). Both measures are
already part of the current routine of the treatment program.

We are aware of the fact that an observational study to assess the effects of
the DBT-PE program has some methodological disadvantages as compared to a
randomized controlled trial (RCT). We seriously considered a RCT design and a
wide range of different possible comparators. A RCT, however, has its own
disadvantages and possible sources of bias, especially if the participants
cannot be blinded to the condition and if intervention that is studied requires
commitment of the participant, which is the case in our study. If less than
100% of potential candidates are willing to participate and many participants
have strong opinions about the treatment alternatives, the results of a
classical RCT may be severely biased or not representative for the target
population. In a pilot study we asked 14 patients on the waiting list or during
the DBT-PE treatment program whether they would have been willing to be
randomized to a different treatment center with a short, intensive, clinical
program purely focused on PTSD treatment (PsyTrec, Bilthoven). Only two of
these patients expressed their willingness to participate in such a design. We
concluded that an RCT would not be feasible on the basis of recruitment
problems alone, apart from the methodological and logistical problems that we
would face.

The intervention consists of a 12-week residential program of trauma-focused
treatment, in which patients are offered dialectical behavioural therapy
combined with prolonged exposure (DBT-PE). DBT is developed by Linehan (Harned,
2010), specifically for people with a borderline personality disorder (BPD).
The main goal of the therapy is to increase emotional and cognitive regulation
by increased insight in triggers that lead (undesired) reactions and applying
appropriate coping skills. DBT has been found effective in patients with BPS
(Kliem, 2010) and, among others, in survivors of childhood sexual abuse
(Decker, 2010).

The other important component of the trauma therapy is prolonged exposure,
which is specifically targeted at decreasing PTSD-burden. The therapy mainly
consists of imaginary exposure, which is a specific type of cognitive
behavioural therapy developed for the treatment of PTSD. During the imaginary
exposure, the traumatic experience is repeatedly described and re-experienced.
By being confronted with the traumatic experience in this way, rather that
avoiding feelings and thoughts about it, emotions are weakened and patients can
start to process the traumatic experience. Prolonged exposure is a very
effective treatment for chronic and severe PTSD.

The 12-week residential program is an intensive modular therapy program that is
specifically tailored to the individual patient. It consists of individual
treatment within a group setting; in addition to individual trauma therapy
there are numerous group modules that are targeted at increasing the general
coping skills of the participants. The program allows for a total of 25
psychotherapy sessions, lasting 60 minutes each.

There is a strong holding environment for the patient thanks to the 24-hour
care. The treatment entails three different phases. The first phase covers the
first three weeks during which patients prepare themselves for trauma therapy
by determining their individual goals during their therapy and learning skills
to regulate their arousal. During the second phase, from week 4 until week 9,
the focus is on skills-based prolonged exposure. Skills-based prolonged
exposure implies that patients use their skills to regulate their arousal
during the trauma therapy, allowing the patient to process his trauma. In the
last three weeks, the aim is to finalize processing of the trauma through
radical acceptance of the trauma as part of history with its consequences for
the future and the main focus will be resocialization in order to prepare to
return home.

In each phase there is a variety of treatment modules to suit the purpose of
the relevant phase. Some modules are standard, others are optional. From this
complete set of treatment modules, a subset is selected that is tailored to the
specific needs of the participant. This individual program can be adjusted
during the course of the therapy.

Medication
Since there is no established farmacotherapeutic treatment for PTSD, nor for
BPD, psychiatrists are free to follow their clinical experience. However,
benzodiazepines should be avoided, because of their negative influence on
cognitive functioning like memory, attendance and information processing (see
for instance Hendriks 2012).

The burden associated with participation consists of in total 13,5 hours of
interview time (T0: 5 hours, T1: 1 hours, T2: 0,5, T3 and T4: 3,5 hours each).
We are aware this may be considered burdensome for participants. However, the
questionnaires can be filled in at home on the computer in participants* own
time and schedule. To furthermore decrease the burden, we will split up the
interviews in multiple sessions, and the follow-up interviews will be done by
telephone. Earlier studies with extensive interview time, such as the Frysian
Trauma Study (Swart et al, 2017), encountered hardly any problems even though
they had 8 hours of interview time per measurement.
We expect no risk of participation, as an earlier study into the effectiveness
of this therapy showed no increase in self-harming behaviour or in suicidal
behaviour. DBT-PE is a highly promising intervention of which participants are
likely to benefit, in a much faster pace compared to care as usual.
Furthermore, most questionnaires applied in this study are part of regular
diagnostics or Routine Outcome Monitoring procedures, albeit we carry out
additional measurements, i.e. before the waiting list period and 6 and 12
months posttreatment.