On Structure and Leadership in Mentalization-based Group Therapy and Group Analysis

Introduction

Mentalization-based treatment (MBT), originally designed for the treatment of borderline personality disorder (BPD), has proved efficient in randomized controlled trials and naturalistic studies (Bateman and Fonagy, 2001; 2009; Bales et al., 2012; Rossouw and Fonagy, 2012; Kvarstein et al., 2014). MBT is a conjoint treatment, combining individual (MBT-I) and group therapy (MBT-G). Recently, manuals for MBT-I (Karterud and Bateman, 2010), MBT pscyhoeducative group therapy (Karterud and Bateman, 2011), and MBT-G (Karterud, 2012) have been published in Norwegian. These treatment manuals follow the recommended standards of Luborsky and Barber (1993), by including theoretical background, treatment principles and therapeutic techniques with specific examples, in addition to rating scales for therapist adherence and competence.

The main objective of MBT is to enhance patients’ inherent capacity for mentalizing for the purpose of improving their intersubjective competence, and thereby enhancing their sense of self-cohesion and identity. MBT requires the cultivation of a mentalizing stance in both therapist and patients, which entails a humble acknowledgement of not-knowing, patience in identifying different perspectives, as well as an active effort to make detailed accounts of experiences rather than explaining them (Bateman and Fonagy, 2004). The patients are ideally engaged in a mentalizing discourse where beliefs, feelings and interpersonal transactions are challenged to bring about changes in perspective, while solutions and answers play subordinate roles (Karterud and Bateman, 2010).

There is a large literature and a rich clinical tradition on treatment of borderline patients by group psychotherapy (Roth, Stone and Kibel, 1990). Such treatment is considered difficult and demands competent therapists and well-functioning groups (Marziali and Munroe-Blum, 1994). Typical clinical issues are how to deal with splitting and malignant projective identification, and in particular how to avoid that these phenomena infect the group as a whole (Pines, 1990). The textbook literature clearly advises therapists to compose their groups in a balanced way, by not adding more than a few borderline patients (Rutan, Stone and Shay, 2007; Yalom and Leszcz, 2005).

However, the group psychotherapy landscape has been changed in recent years by public mental health services obliging group therapists to accept poorly functioning patients, thereby forcing them to abandon previous politics of selecting patients according to established principles of suitability. We are increasingly witnessing groups where many of the members suffer severe personality disorders (Karterud and Wilberg, 2007).

The response to this situation has been a development of treatment techniques that might counteract the problems created when all the members of a group have serious problems of self-cohesion and interpersonal functioning. A common strategy has been to abolish the principle of free group association and providing more structure to the group processes. For example, the group component of dialectical behavioural therapy (DBT) is very structured and defined as a device for developing social skills (Linehan, 1993). The ‘Group schema therapy for borderline personality disorder’ is also a highly structured cognitive oriented group therapy, where the agenda is defined by the therapists, according to a predefined sequence of goals (Farrell and Shaw, 2012). However, MBT-G, which evolved through modifications of group analysis, still adheres to the principle of group psychotherapy through the group process (Karterud, 2011).

Both in the theoretical and empirical literature on MBT, the individual component of the treatment are well covered, while the group component has received comparatively less attention (Karterud and Bateman, 2011). Unfortunately, the recent and comprehensive manual by Karterud (2012) is only available in Norwegian. In this manual, general principles of MBT-G are formulated with an emphasis on the significance of the mentalizing stance in groups. It also includes detailed descriptions of 19 specified treatment principles, which constitute the core of a MBT-G adherence and competence rating scale. 10 of these principles are similar to items in the manual for individual MBT (Karterud et al., 2013): 1) Engagement, interest and warmth; 2) exploration, curiosity and not-knowing stance; 3) challenging unwarranted beliefs; 4) regulating arousal; 5) acknowledging good mentalization; 6) dealing with pretend mode (pretend mode refers to a mode of thinking and communicating which is overly abstract and aloof and not grounded in emotional reality; (Bateman and Fonagy, 2006); 7) dealing with psychic equivalence; 8) keeping an affect focus; 9) practising ‘stop and rewind’; and 10) focus on the relation between therapists and group members (‘transference’).

In addition there are nine group specific items. These items address specific structural elements, which are of vital importance for the group as a whole:

The purpose of these principles is to provide the therapists with tools that will make it possible to conduct a dynamic group even with poorly functioning patients. MBT-G structuring principles can be understood as a response to the inherent dangers of chaos and destructivity associated with including many borderline patients in groups.

MBT-G has its roots in group analysis and psychodynamic group psychotherapy. These group therapies all share a common therapeutic principle in that significant intersubjective transactions unfolding in the group are actively brought to attention and explored—a hallmark of dynamic therapy. However, MBT-G parts ways with some other long-held dynamic principles. The most noteworthy deviation is that the therapists take responsibility for structuring each session, explicitly invite exploration of interpersonal events, and make an active effort to regulate current emotional arousal. Furthermore, the roles and tasks of the patients and the therapists, the objective of the group and the way to reach this objective, are more clearly defined and explicated for patients in MBT-G.

A session typically starts out with a short recapitulation of the previous session (‘building bridges’ and ‘minding the group’). Then, for the sake of structuring the present session, patients are asked to signal their need for time and space in the group, typically centred on significant (interpersonal) events from the last week. Thus, MBT-G encourages turntaking as a central therapeutic and organizing principle, which puts it in contrast to the typical unstructured psychodynamic group therapy. It is important to underline that this turntaking is not equal to ‘individual therapy in groups’. It is still dynamic group psychotherapy, however focused in turns on designated patients and their intersubjective and emotional problems. The manual contains detailed discussions on how to conduct such interpersonally focused turntaking, without creating a dependency group in a Bionian sense (Bion, 1961) or resorting to a kind of individual therapy with spectators (Karterud, 2012).

A typical MBT-G session will revolve around a constant effort to understand specific situations from the patients’ life, or situations unfolding in the group here-and-now. Interpersonal events are probably more highly prioritized in MBT-G than in other psychodynamic group therapies. The therapists’ chief task is to stimulate patients to mentalize these events, and further, to create and sustain a mentalizing discourse in the group. A mentalizing discourse is hallmarked by an emphasis on details of events that allows for constructive imagination and realization of what goes on in the minds of the subjects. Thus, it is necessary to be specific about what happened, at what time, who was involved, what the involved persons were doing or saying, and what the narrator was thinking and feeling. The patients should learn to sketch interpersonal situations so that both therapists and the other patients are able to tune in on the transactions. For that purpose, it is advisable to dive into single events instead of sketching many events superficially.

By a skillful practice of the above mentioned principles, the aim is to construe a mentalizing group culture. As patients in MBT-G conjointly meet with their individual therapists, the group therapists can intensify the interpersonal emphasis in the group sessions, leaving in-depth intrapersonal exploration and social support to the individual sessions. In principle, the group situation is an excellent arena for exploring interpersonal events and enhancing the capacity to mentalize. The challenging task for the therapists is to realize this potential. With a group consisting of persons with severe personality disorders, this is difficult.

Managing authority is a therapeutic principle specific to MBT-G, not being part of the individual treatment manual. The group manual discusses in length how the therapists’ authority may be challenged by different aspects of borderline pathology, e.g. by devaluating or aggressive patients, and in general by powerful projective identifications. It is important for therapists to remember the fact that they have invited the patients to participate in a specifically designed project, with a particular purpose and a set of ground rules. The therapists should lead the group with a firm hand and make sure that issues of relevance for the therapeutic project are attended to. MBT-G is far from the group analytic ideals of ‘trust the group’ and ‘leave it to the group’. When patients derail the process, the therapist should interrupt the ongoing transaction and get the group back on track. In contrast to more typical psychodynamic group therapies, MBT group therapists should provide the group with structure and guidance, and be more explicitly responsible for the therapeutic process.

One precaution that is given special attention in both individual and group MBT is that the therapist should not stand out as an omnipotent expert with a privileged knowledge of what is ‘really’ going on in the group or in the mind of its members. Instead, the therapeutic stance should be curious, active, empathic, and at times challenging. The therapists should be focused on the patients’ mind and their ongoing intersubjective transactions in an engaged, questioning, and not-knowing way (Bateman and Fonagy, 2006). The principle of not-knowing is believed to stimulate the curiosity and engagement of the patients, and ultimately to foster mentalization. It is considered potentially anti-therapeutic to take on an expert attitude with respect to the content of the mind, as the patients may accept a therapist’s point of view indiscriminately, following his lead while leaving their own feelings of what is important behind.

The manual of MBT-G was developed by this author in co-operation with Anthony Bateman and several skilled group analysts from prominent treatment units in the Nordic countries that had extensive experience with BPD patients (Karterud, 2012). The development was based upon extensive examinations of video-recordings of ongoing groups and experiments with different techniques and treatment principles, in particular groups from the Department of Personality Psychiatry, Oslo University Hospital. The manual was revised several times before reaching an acceptable consensus. MBT-G as implemented at the Department of Personality Psychiatry has been very successful. In a recent study Kvarstein et al. (2014) compared data from the MBT programme, gathered since 2008 (n = 64 BPD patients), with the previous psychodynamic programme (n = 281 BPD patients). The MBT programme outperformed the psychodynamic programme on most outcome variables by effect sizes that were nearly twice as large (e.g. EZ 1.8 vs 0.9 for symptom distress). In the psychodynamic programme borderline patients had a very high dropout rate (42%) during the first six months, and we were disturbed to find that they related this to their experiences with the group therapy (Hummelen et al., 2007). In the MBT programme these problems vanished, and the six month dropout rate was down to 5%.

During recent years, courses in MBT-G have been arranged by the Norwegian Institute for Group Analysis (IGA) and the Norwegian Institute for Mentalization (IM). Training at the IGA used to consist of 10 days spread over two years, in concert with the block training courses in group analysis. Training at the IM consists of eight days spread over one year, with no parallel self-experiential component. The candidates are supervised during training based upon video recordings of their ongoing groups. However, time does not allow for more than some short sequences of the respective groups. Consequently there has been little knowledge about what might be the most important challenges when MBT-G was to be implemented in clinical practice at ordinary Mental Health Centres (MHC) in Norway. We thus were very curious when we succeeded to perform a video-based study of three consecutive sessions of a mentalization-based group at a MHC (see Inderhaug and Karterud, 2015). Did the therapists succeed with their project and what were the main obstacles?

In the above mentioned study we found rather chaotic group processes, of a type which is well known from the literature (Pines, 1990; Marziali and Munroe-Blum, 1994; Hummelen, Wilberg and Karterud, 2007). The group was conducted by two therapists who had previous training in group analysis and MBT. We could observe that one of them (T1) consistently tried to apply MBT principles. However, the effect upon the group seemed minimal with respect to constructing scenarios that were useful for conjoint exploration and mentalizing. The therapists in the group we observed struggled to manage their authority in an efficient fashion, and also exhibited what we conceived as an exaggerated not-knowing attitude towards their therapeutic work. That is, the therapists were seen to adhere little to the item of managing authority, while the not-knowing stance was overplayed.

The therapists seemed to be trapped between two models. With one leg in each camp, they did not practise group analytic psychotherapy, nor did they practise MBT-G. Their dilemma possibly reflects the conceptual and therapeutic history of group analysis and MBT-G.

Concerning authority, Foulkes emphasized that the therapist(s) was not the leader of the group, but its conductor. He repeatedly argued against leadership in a traditional sense (Foulkes, 1964). The reasons were partly ideological, but mainly theoretical. We have to remember Foulkes’ embeddedness in the anti-authoritarian Frankfurter School and the profound anti-authoritarian mentalities in Europe in the 1960s and 1970s. Perhaps more important was his theory of man: being profoundly social and ‘groupish’ by nature, man would naturally strive for co-operation and group membership. However, when hampered in his development, man would develop symptoms due to thwarted basic needs. In becoming member of a therapeutic group, the need for co-operation and group affiliation would be reactivated, accompanied with the individual’s idiosyncratic ways of dealing with these needs. The symptoms would appear as group problems and the main task of the group analyst would be to assist the group in articulating these problems in the verbal discourse. Basically, these ideas are compatible with modern theories of the evolution of human thinking (Tomasello, 2014). Abstract and reflective social cognition are basically group phenomena which new members of the Homo sapiens group are predisposed to learn in order to become competent cooperative members of adult societies. Foulkes labeled the learning process ‘ego-training in action’, while the corresponding MBT slogan is ‘the group as a training ground for mentalizing’.

The overriding pragmatic question for both group analysis and MBT-G is now: What characterizes a ground suited for proper training of the ego or mentalizing? Foulkes maintained that almost any kind of group (with respect to its members) conducted according to group analytic guidelines could become a proper training ground. I consider this view erroneous. To support this view one should demonstrate empirically both high process qualities and favorable outcomes of groups composed of members who all share profound problems of social cognition/mentalizing, e.g. severe personality disorders. To my knowledge, there exists no report from group analytic practice, which offers such support. Foulkes (and his generation of group analysts) seems to have underestimated the constituents of (successful) social cognition (Sharp et al., 2008). Successful social cognition (or good mentalizing abilities) is a late developmental achievement and dependent on the child being socialized by trustworthy adults and peers in safe situations. Furthermore, children seem to develop an ‘epistemic trust’ when they encounter adults who perform a kind of ‘pedagogical stance’, conveying basic properties and values of the social world (e.g. the inherent ethics of cooperative communication and concern for truth; Csibra and Gergely, 2009). Development of social cognition is thus dependent on a range of communicational qualities of the environment.

These prerequisites are likely to become violated when the communicational reeducation takes place in a group composed of personality disordered individuals who receive minimal instructions on how to proceed. Such experiments likely end up in disconnected speech and utterances, aborted stories conveying fragments of experience, emotional outbursts and activation of psychic equivalence, malign projective identifications, sequences of pseudomentalization, experiences of not being heard and disrespected, as well hopelessness and disillusionment, as demonstrated in several of the citations in the article of Inderhaug and Karterud (2015). Such situations call for therapists, not as conductors, but as competent leaders who model a kind of parental-like authority. This parental-like authority is less concerned with what is right or wrong, as how to reach truth about oneself and others. Philosophically, such an attitude is grounded in communicational ethics (e.g. Grice, 1991; Habermas, 1990) which concerns the implicit ethos of communication between humans, e.g. a wish for communication and conveying honest messages, a concern for truth, a respect for the dialogue partner, active listening and an effort to understand the partner’s mental state and point of view, etc. This kind of communication does not come by nature. It is the hard won achievement of cultural (group) evolution. The concept of mentalizing stance belongs to this tradition of ideas.

Although communicational ethics is a cultural (group) product, being legitimized and sanctioned by cultural codes and practices, the pragmatics of communication is most often discussed with references to a two-person kind of dialogue (‘I—Thou’). When it comes to group situations things get more complicated, and culture has invented a series of procedures for securing rational reasoning. There is an implicit and explicit cultural knowledge (‘common ground’) that group situations easily activate mental phenomena that run counter to rationality. The pioneering work of Bion (1961) conceptualized these difficulties in the basic assumption theory, contending that rationality was a property of the ‘work group’, while prementalistic cognition flourished in the ‘basic assumption group’. The position of group analysis has been that abolishing ordinary procedures for group cooperation, e.g. by advocating free group associations, would ultimately liberate the group from harsh superego control and rigid cultural artifacts, if the group analyst adhered to certain therapeutic principles. MBT-G, on the other hand, maintains that, depending on the group member’s level of personality functioning, a certain communicational structure is necessary for reason to survive and thrive. This requires leadership. The therapists have to uphold a communicational ideal, to model it, strive for it, defend it, explicate its raison d’etre, and try to recruit other group members so that the group self becomes embodied not merely by the therapists, but by other group members as well (Karterud, 1998). There is a lot of ‘knowing’ embedded in this undertaking with respect to communicational processes and how to intervene, e.g. how to identify poor versus good mentalizing and their effects on the group process. It is easy to lose this knowledge when being bombarded by projective identifications. It is all the more important therefore, that the therapist pair is firmly grounded in the ethos of their therapeutic project.

According to this logic, the not-knowing stance comes second to communicational structuring. In order to train in mentalizing, in a more limited sense, one needs scenes (either in the group, or in the narrative) with protagonists in time-sequential events. Through such scenes group members can imaginatively explore mental states, their intersubjective connectedness, contextual dependency and implications for the subject(s) who ‘own’ the scenes. In such scenes, the therapists have to refrain from the temptation to ‘know’ the ‘hidden’ motives behind the transactions, and to facilitate the group member’s open-minded exploration of minds.

Basically, the idea is that by the individual group members striving to become a member of the mentalizing group self, i.e. the work group aspect of the group, one has to develop one’s social cognition (Karterud and Stone, 2003). Or phrased in another way: the development of social cognition (narration, attention, focusing, inhibitory control, affect consciousness, metacognition, mentalizing, integration, etc.) is likely to be a result of ones efforts to become a member of the group self. This logic is supported by modern theories of evolution of thinking, self-consciousness and culture (Bogdan, 2010; Tomasello, 2014).

The findings of our companion study (Inderhaug and Karterud, 2015) illustrate some difficulties therapists encounter when trying to establish themselves as leaders for a certain communicational structure when the group members easily resort to primitive modes of mentalizing. The therapists in this study were overheard, not paid attention to, interrupted, and at times ridiculed. As explained in previous paragraphs, no development of social cognition can take place in such a discourse. The therapists were vaguely aware of the discrepancies between their therapeutic ideals and the chaotic group discourse. However, they feared aggressive outbursts in the group, felt helpless when they experienced the absence of therapeutic alliance, and lost confidence in MBT-G principles. Nevertheless they tried to practise a kind of not-knowing stance. However, these attempts failed since the structure could not capture and contain adequate scenes long enough for joint exploration to take place. The result was pseudomentalization. Phrased in an object-relational language, many of the therapist’s difficulties can be conceived as countertransference strategies for coping with severe ‘attacks on linking’ (Bion, 1970). Bion coined this term as a slogan for destructive mental processes of non-integration and resistance towards construction of meaning. Such attacks are as expected in groups with borderline patients. In order to deal with attacks of this magnitude (augmented by group processes), there is an obvious need for professional support and supervisory assistance. The lack of such support may be one explanation as to why MBT-G, which has proved very efficient in an academic/university setting with a high professional standard (Kvarstein et al., 2014), failed to do so in an ordinary MHC. Supervisory services thus have to be strengthened for MBT-G. There already exist in Norway a MBT quality assurance laboratory (http://www.oslo-universitetssykehus.no/omoss_/avdelinger_/mbt-kvalitetslaboratorium_) that performs such services for individual MBT. This year we hope to extend those services to include also MBT-G. There is also an obvious need to publish qualitative studies on successful MBT-G. By comparing successful and failed attempts at practising MBT-G we might obtain more knowledge about the pragmatics of communication within this particular type of group discourse. Such studies are under way.