Group psychotherapy uses a ‘natural’ social setting—the small, the median, and the large group—to conduct its psychotherapeutic processes. We grow up in small groups (among family, peers, friends), we learn and work in median groups (in classrooms, committees, teams), and we engage in science, economic activity, and politics in large groups (through assemblies, networks, companies, political parties, parliaments, etc.). This chapter will discuss the conditions that are necessary for these natural groupings to become psychotherapeutic. The variety of different approaches to the psychotherapeutic use of groups, ranging from psychoeducation, psychoanalysis, psychodrama, group analysis, and humanistic psychology to cognitive-behaviorism, reflects the different historic roots of group psychotherapy.

So, let us first look at the different theoretical backgrounds, treatment principles, and conceptualization of processes, as well as at the different role played by the group therapist. The effectiveness and efficacy of group psychotherapy, and the indications and contraindications derived from empirical research findings, will be outlined below. As a group analyst, my main focus will be the group-analytic approach to group psychotherapy.

A group is more than the sum of its members, just like the meaning of a sentence is more than the line-up of various words. Therefore, we can use three different perspectives in order to understand whole group configurations:

• The personality of the different members. The intrapsychic world of an individual is made up of internalized networks of relationships (Laing, 1974, p. 16). This internalized network of relationships gets reactivated in a group.

• The interaction or the interrelatedness between different members. The process of interaction between different, unconscious networks of relationships, reactivated in a group setting, is a focus of attention in all analytically oriented group approaches. This interaction occurs even in other approaches, even when they refrain from using this process of interaction: for instance, cognitive-behavioral approaches tend to avoid it altogether. The sequence, the structure, and the emotional quality of this interaction is used to engender therapeutic processes in all analytically oriented approaches.

• The system of the group as a whole in its contexts. Each unique group develops through a clash of centripetal and centrifugal forces, which are balanced by its constantly changing structure. Centripetal forces consist of a shared goal and the cohesion or coherence of the group as a whole. While cohesion describes the attractiveness of a group to its members (Levin, 1951), coherence focuses on ‘an underlying sense of containment based upon differentiation and understanding’ (Pines and Schlapobersky 2000, p. 1452). Centripetal forces ensure the existence of the group and its stability over time. On the other hand, different norms, different roles and the individual's deviation from group norms are centrifugal forces that initiate change and development. Only if there is a balance between the two forces can the group develop a fluid structure. It is this structure that makes the group a safe enough place for its members to risk making changes.

In order to understand the group process, all three perspectives have to be taken into consideration. Each approach uses these perspectives differently. The political, cultural, and social context of each unique group creates the framework for an unfolding and ever expanding group process.

The first psychotherapeutic use of groups was made in the US by medical doctors with the aim of psychoeducation. Pratt (1908, 1922), Lazell (1921), and Marsh (1933) addressed lectures to different groups of patients. The aim was to increase self-control by providing them with more information given about their disease. These lectures were provided in small, median, or large groups. Pratt referred to them as ‘thought control classes’, Lazell as ‘etiology spiel’, and Marsh as ‘milieu therapy’ by social-educational groups (Ettin, 1999, p. 72–8). Similar developments in Vienna were described as ‘guidance groups’ by Adler. His follower Dreikurs (1932), working in the US, used the same term.

The psychoeducational approach is nowadays used in the application of behavior or cognitive-behavior therapy in groups (cf. Fiedler, 1996; Free, 1999).

In psychoeducational groups, the curative factors are defined as reeducation, socialization, the imbuing of an individual with hope, the raising of morale, and the emotional developments occurring during the teaching process in a group.

Psychoanalysis is the basic theory underpinning group psychotherapy approaches that focus not on teaching and education, but on insight. This insight-through-group experience is pursued through a number of variations on the psychoanalytic theme: group analysis, psychodrama, psychodynamic groups, and various forms of humanistic psychology such as Gestalt therapy (Perls) and transactional analysis (Berne), or encounter groups (Rogers).

The first psychoanalyst to bring together patients suffering from neurotic symptoms in a group using psychoanalytic techniques was T. Burrow (1928), again in the US. His starting point was the perception of humans as a social being and of the group as the natural focus of treatment. His aim was to make conscious both latent and repressed meanings through the here-and-now interaction within a group. Burrow named his method group analysis. He relied on Freud's study, Group psychology and the analysis of the ego (Freud, 1921).

Jacob Moreno used the theatre stage to create a scenic understanding of intrapsychic conflictual life. From 1928 onwards, he offered psychodrama demonstrations at Carnegie Hall. Moreno used psychodrama in a small group setting for psychotherapy at the Mount Sinai Hospital in New York City. He developed the method of sociometry (1938) and in 1942 founded the American Society for Group Psychotherapy and Psychodrama. The Society in 1951 became the International Association for Group Psychotherapy, an umbrella organization for all approaches.

In the 1940s, several psychoanalysts used their psychoanalytical understanding to work with patients in a group setting: Lauretta Bender (1937), Louis Wender (1940), Paul Schilder (1940), and Alexander Wolf (1949) worked with resistances and transference processes. They defined the therapist as a symbolic parent and the other patients as representing siblings with the aim of providing social insights through interpersonal exchange: the group ‘removes’ the problem from the sphere of the individual's symptom formation and suffering. Through group interaction the isolation of the individual, which is seen as an important part of psychoneurosis, is opened up (Schilder, 1936, pp. 612–14).

All these psychoanalytically based group therapy approaches had one thing in common: the method they used was the application of individual psychoanalysis to a group setting with the aim of social integration. The therapeutic emphasis was kept on the individual patient in a small group setting. Samuel Salvson founded the American Group Psychotherapy Association in 1942 and created the International Journal of Group Psychotherapy in 1951.

Moreno, Schilder, and Wolf worked mainly with adults suffering from neurotic illnesses and recommended the exclusion of psychopaths, alcoholics, hypomanic patients, and hallucinating psychotics from groups. Bender and Slavson worked with disturbed children and used puppet play (Bender, 1937, p. 1161) and activity groups primarily to treat overaggressive and excessively withdrawn children. Salvson invented the term group dynamics in 1933.

Three different perspectives can help us to differentiate among the various psychoanalytically based approaches:

• psychoanalysis in the group

• psychoanalysis of the group

• psychoanalysis by the group.

Each of these has a different focus when it comes to the process, the task of the therapist and the curative factors in group psychotherapy.

Agazarian (1997) implemented general system theory in group psychotherapy by focusing the therapist's and the group's attention on boundary issues and subgrouping factors. The main task, she argued, was to increase communication across boundaries: ‘How the group communicates is always more important than what it is communicating about’ (Agazarian, 1989, p. 176). In her technique the group leader clarifies subgroup boundaries and encourages very actively interactions that cross boundaries in a respectful way.

In all approaches, the selection and composition of the group is crucial in terms of therapeutic efficiency. Most approaches, except for the psychoeducational or cognitive-behavioral ones, follow the general rule that the selection of patients and the composition of the group determine the quality of interaction among group members. These factors, in turn, then provide the foundations for effective treatment.

In all analytical approaches, it is assumed that the composition of the group should be heterogeneous as regards the patients’ psychopathology and personality structures. It should also vary in terms of social class, age, and gender, i.e., patients’ culturally specific ways of interacting. However, the group should be homogeneous when it comes to the level of frustration its members can cope with or the conflicts they can work through productively (measures of ego strengths). Only for some groups of patients does homogeneity have advantages: these are patients with drug or alcohol addiction, severe psychosomatic illnesses or personality disorders with destructive acting out, as well as psychotic patients or forensic psychiatric patients (Knauss, 1985; Dies, 1993).

In contrast, cognitive-behavioral and psychoeducational approaches prefer groups to be homogeneous groups in terms of symptoms in order to teach patients with similar symptoms about the psychological background of their deviant behavior.

The ‘Noah's Ark Principle’ (‘the animals came in two by two’) applies to all approaches:

The group size recommended by analytic approaches is between five and nine, a size that encourages the development of trust and intimacy. Cognitive-behavioral approaches recommend a group size of between four and six for the training of new skills. However, there may be up to 20 group members present for the teaching part.

The central question remains: Which patient can benefit from which group? ‘A central issue is whether or not a particular treatment group is suitable for a specific patient at a given point in the manifestation of symptomatology and the group's current level of development’ (Dies, 1993, p. 487).

To prepare patients for the often anxiously awaited experience of group psychotherapy is as important for a positive outcome as the selection and composition of the group (Salvendy, 1993). The point of preparing patients is to reduce anxiety and to motivate them by cognitively and positively prestructuring the group experience that they want to undergo in order to get better. Anxiety to disclose one's so far private fantasies, needs, weaknesses, and traumatic experiences, often connected with guilt and shame (Seidler, 1997, 2000a,b) can make a patient dread the encounter with that public body, the group. But so can the social isolation caused by the patient's symptom formation. Therefore, it is crucial that the therapist devotes several individual sessions or a waiting group to the patient's preparation and motivation. This strategy is an empirically proven predictor of a successful outcome (Tschuschke and Dies, 1994).

One aspect of patient preparation involves familiarizing the patient with the group rules. All analytically oriented approaches operate with the following rules: confidentiality, free-floating association with no conscious exclusion of any fantasy, memory or thought arising in the group process, tolerance towards every verbal communication, the exclusion of body action, and no meetings outside the group session. If such meetings do occur, they have to be discussed in the following group session. The group boundaries have to be clear to all patients: they must know who belongs to the group, the time, place, and duration of group sessions, and the honorarium for each session. It is helpful to mention the possibility of symptom aggravation at the beginning of therapy, which is a sign of resistance to change, because this may prevent patients from dropping out early on. Meetings outside the group and subgroupings are interpreted as resistance to open communication and therefore have to be discussed inside the group.

In order to allow the group to work through a separation, members must announce the termination of their therapy to the group well in advance.

‘The group, the community, is an ultimate primary unit of consideration, and the so-called inner processes in the individual are internalisations of the forces operating in the group to which he belongs’ (Foulkes, 1971, p. 212). This means that conscious and unconscious processes in groups are deeply structured by the social unconscious: the group is the primary psychological unit, the individual the primary biological one.

In order to understand all analytic approaches the above theoretical statement by Foulkes, based on the social philosophy of Norbert Elias (1987) and the dynamic psychoneurology of Kurt Goldstein, should to be taken into account.

In those approaches that focus on psychoanalysis in groups, the analysis of the individual in a group context, which contributes to individual analysis, is placed at the center.

In approaches that focus on the psychoanalysis of the group, as in the Tavistock tradition, the main objectives of analysis are the shared assumptions, tensions, or conflicts of the group as a whole.

In group-analytic psychotherapy the shared experience of developing communications within the group is analyzed by the group, including its conductor, at four levels.

• The current level, which is the working alliance.

• The transference level.

• The projective level.

• The primordial or archaic level.

The aim is not only analysis, but also the translation of unconscious symptom formation into conscious conflicts within and between group members. Group-analytic psychotherapy seeks to create an ever expanding, increasingly complex process of communication. The individual is perceived as a nodal point in the network of group relations, and transference as well as countertransference processes form his or her link to the outside world. The here-and-now conflictual interaction resembles a figure hovering in the background of the dynamic matrix of the group as a whole. ‘In learning to communicate, the group can be compared to a child learning to speak’ (Foulkes and Anthony, 1968, p. 263).

The healing effect of group communication is defined by Foulkes as one of the basic laws of group dynamics: ‘The deepest reason why patients… can reinforce each others’ normal reactions and wear down and correct each others’ neurotic reactions is that collectively they constitute the very norm from which, individually, they deviate’ (Foulkes, 1948, p. 29). It is the deviation of different individual members from the norm, with each of them going off in a different direction, rather than their submission to the norm, which brings about the development: we seek ‘to replace submission by co-operation on equal terms between equals’ (Foulkes, 1964, p. 65).

Groups are held together by the need to belong, which is basic when it comes to the development of cohesion. At first, all differences within the group tend to be denied. The regressive process can take two directions: either the group turns into a fused mass, which identifies with an idealized, omnipotent leader, and his or her ideology, or it becomes the sum of isolated individuals who cannot find any meaning in relating. Therefore, submission, idealization, and dependency, which all initially contribute to the cohesion of the group, are simultaneously interpreted as possible defenses against development.

The conductor's analytic attitude frustrates the group members’ dependency needs by then setting into motion a process of conscious differentiation. ‘The group analyst has to accept the unconscious fantasy of the group which puts the therapist in the position of a primordial leader image and one who is omnipotent, and the group expects magical help from him. But instead of fulfilling this regressive need, the conductor uses it in the best interest of the group, which means that he has to change from a leader of the group to a leader in the group, replacing thereby… the leader's authority by that of the group’ (Foulkes, 1964, p. 61).

By working through these dependency needs, the deviations from the unifying norms of the group are brought to the fore, and so can be experienced and discussed. This differentiation process runs parallel with the very painful and slow process of accepting the otherness of the other, of being separated. It is advanced by destructive fantasies, which, according to Winnicott (1980), create a differentiation between the self and the other. The major defense against listening to the otherness of the other is a fear of the stranger, of a different world view. ‘A functioning group could be seen as a communication process in which competing discourses come into conflict with the aim to free each group member from being stuck in ones own, private discourse, ones own experience of the self and the world and initiates a process of opening up to communicate with other discourses, other ways of being and experiencing which one did not have previous access to’ (Dalal, 1998, p. 177).

To achieve this process of differentiation, the group makes use of mirroring (Pines, 1982). Mirroring combines empathy and the notion of being different from the other into a single emotional reaction to one's perception of the otherness of the other. Through mirroring, similarities and differences between group members are explored. If the projective part of the process in mirroring is not fully understood, it can deteriorate into malignant mirroring (Zinkin, 1983). This happens when one group member attempts to change the other, while trying to deny his or her otherness. In case of malignant mirroring, the conductor needs to act with the aim of containing the unbearable pain of separateness.

As the following example illustrates, destructive fantasies play an important part in this differentiation process (cf. Knauss, 1999):

A process was set in motion that went from a lack of communication, through resonance and mirroring, to open communication, i.e., from monologue to dialogue. The analytic attitude, the setting and the careful selection and composition of the group all serve to protect the group from actual destruction. They also encourage the creation of a safe place for the communication of destructive fantasies. This means ego development and not only ‘ego training in action’ (Foulkes, 1964, pp. 82, 129).

Morris Nitsun (1996) has described destructive group forces, naming them the ‘antigroup’. He showed that actual destruction is only acted out if the destructive fantasy cannot be verbalized and if the object of the destructive fantasy seeks retaliation. In that case, developmental processes of differentiation between me and not me cannot be set in motion. The creative potential of verbalized, destructive fantasies encourages a differentiation process (Knauss, 1999). It fosters the group members’ perception of one another as different subjects with their own rights and needs. The social unconscious of the group, understood as a harbor of all the denied heterogeneity within the group, becomes conscious and brings power relations within and between groups into a shared process of communication. This can also be seen as a process of democratization.

Cohesion, coherence, regression, mirroring, imitation, identification, internalization, resonance, condensation, exchange, sharing, socialization, and polarization, as well as projective identification are group-specific therapeutic factors. The dynamic links between the group structure, the contents of the discussion and the form it takes can become conscious by an ongoing process of mutual interpretations among group members. Only when the communication process gets stuck does the group conductor need to step in with an interpretation or clarification. The conductor therefore follows the group process with curiosity, empathy, and tact, and should not interrupt communication as long as it flows freely. He intervenes solely if the communication process stagnates as a result of overwhelming defense mechanisms.

In each approach, the therapeutic attitude of the group conductor or group leader is an essential factor. A culture of basic trust can develop only if he takes a largely positive, sincere, and curious attitude towards each patient and the group. This attitude is a precondition for the success of all therapeutic alliances. In analytic approaches, the abstinence and neutrality of the conductor (Knauss, 1994), combined with a sincere effort to understand what is going on, but not to act, is an additional requirement. As mentioned, in a group-analytic approach, the term ‘leader’ is replaced by the term ‘conductor’, so as to make clear that the conductor must be ‘free from the temptation to play this god-like role, to exploit it for his own needs’ (Foulkes, 1964, p. 60). Instead, he must seek to ‘wean the group from this need for authoritative guidance…’ (Foulkes, 1964, p. 61) and leadership.

This group-analytic attitude contributes fundamentally to the development of a group culture and a climate of tolerance and differentiation within a process of free-floating communication. The transference processes are not only involving the conductor, but also all other group members in a process of multiple, mutual transferences. The same applies to the process of interpretation. The various countertransference processes focus on the dynamic of the unconscious network of relationships existing within the group, creating in the conductor a countertransference to the dynamic matrix of the group as a whole, which he is a part.

The conductor may abandon his role as a ‘participant observer’ when, and only when, the communication process breaks down due to unconscious conflicts. If the group process does get stuck, however, the conductor must intervene by addressing the process rather than the contents of the communication. When the group stopped communicating after the report of the first patient about his violence towards his wife and was waiting for my reaction to it, I said: ‘Do you keep silent and just wait to see how I would react?’ This remark on the process opened a space for resonance, chain and mirror reactions in other group members and the communication process, which was blocked by separation anxiety, could continue.

In this commenting-on-process intervention the conductor needs to take into consideration (1) the past outside the group—the ‘social history’; (2) the past inside the group—the group's history; (3) the present outside the group—the actual context; and (4) the present inside the group—the actually developing dynamic matrix. The conductor might localize mirror reactions, which can be helpful when it comes to recognizing aspects of oneself and others and to accept the viewpoints of others on oneself. He can also make the group aware of group role configurations, especially if a scapegoating process is taking place, and should always keep in mind the aim of fostering the process of individualization and relating. ‘The therapist's task is to follow the interaction, to use interventions sparingly and strategically, to cultivate a reflective curiosity’ (Pines and Schlapobersky, 2000, p. 1455). This will include gesture, behavior, body language, and other nonverbal communications that convey feelings when emotions cannot be put into words.

Appropriate training is of course required in order to carry out these tasks. Group-analytic training will include theory and a long-term personal therapy by the therapist in a group, as well as long-term and intensive clinical supervision of group-analytic processes.

The effect size (ES) in 23 studies using RCTs to compare directly individual and group psychotherapy showed no difference, while both treatment modalities show a big difference in effect size, compared with the control group: for individual psychotherapy it is 0.76 ES and for group psychotherapy 0.90ES (McRoberts et al., 1998).

A review of the literature on the efficacy of group psychotherapy in the treatment of bulimia nervosa by McKisack and Waller (1997) shows that improvement was associated with long-term groups.

Valbak (2003) undertook an empirical study of long-term group-analytic psychotherapy homogeneously composed with severely disturbed bulimic patients. Its positive results demonstrate that the technique of the treatment has to include the following elements:

• A careful assessment interview.

• Consistent monitoring of eating habits and of the connection between self-esteem and outlook.

• An active response to any ruptures of the therapeutic alliance to prevent dropping-out.

• An emphasis on continuity, attunement and timing of supportive and confronting interventions.

• The sustaining of the group matrix as a carrier of hope.

Externalizing patients tend to do better in cognitive-behavioral group psychotherapy, while internalizing patients do better in supportive group psychotherapy (Beutler et al., 1991, 1993).

Severely alcohol abusing patients show more improvement in cognitive-behavioral group psychotherapy, while less severe abusers do better in psychodynamic group psychotherapy (Kadden et al., 1989, 2001; Sandahl et al., 1998). In the case of personality disorders, no difference has been shown between the two treatment modalities (Kadden et al., 2001). Steuer et al. (1984) showed that when it comes to depression in the elderly, psychodynamic and cognitive-behavioral group psychotherapy are equally effective in reducing levels of depression. It appears that it is the group that works for these patients, rather than any specific approach.

Sandahl et al. (2000) found that there was a significant difference in the way cognitive-behavioral group conductors communicate in comparison with the group analysts’ way of communicating in a group: ‘cognitive-behaviorally oriented therapists talked more than twice as much as the group-analytically oriented therapists’ while ‘group members talked 85 percent of the time in the group-analytically oriented groups and 60 percent in the cognitive-behavior therapy’ (p. 343). The contents of the communication differs significantly in two categories: while group analysts and patients in group-analytic psychotherapy communicate more on the contents.

Piper et al. (1984) showed that long-term group therapy (average 76 sessions) is more effective than short-term group psychotherapy (average 22 sessions). Lorentzen (2000) documented significant progress in the scale of symptoms, interpersonal problems, target complaints, and psychosocial functioning after 100 sessions of outpatient group-analytic psychotherapy. Similar preliminary results are reported by Tschuschke and Anbeh, (2000) who studied a large number of patients (more than 600) in a natural setting with a prepost design, comparing long-term, outpatient analytic group psychotherapy with psychodrama groups. He found for patients in group-analytic psychotherapy an ES of 0.97 for the Global Assessment Functioning Scale and an ES of 2.35 for Target Complaints. The Inventory of Interpersonal Problems showed an ES of 0.62 and the Global Severity Index-SCL90-R showed an ES of 0.67.

This study compared psychodynamic and supportive group psychotherapy in a RCT design on various levels:

• The group as a whole

• The individual patient

• The different styles of conducting

• The interpersonal dynamic between patients in the group.

The authors found no significant difference between the two approaches on the level of the group as a whole. The styles of conducting were clearly different.

Patients with more stable object relations (QOR = Quality of Object Relations) benefit more from a psychodynamic approach. Patients with a lower capacity to understand psychodynamic processes (low psychological mindedness) benefit more from other patients. Psychological-mindedness and a high esteem of the group as a whole are the best predictors for a good individual outcome in group psychotherapy.

In this study, five process variables were used to predict who would be the successful and the unsuccessful inpatients in group-analytic psychotherapy. These factors were:

• cohesion

• self-disclosure

• feedback

• interpersonal learning

• reenactment of early family conflicts in the group.

The authors found that:

• cohesion and the feeling to belong to the group are good predictors for success;

• early self-disclosure produces a better outcome;

• more critical feedback was received and given by successful patients;

• insight into the reenactment of infantile conflicts in groups changes the internalized relational network for the better and is linked to an improvement in the quality of object relations.

An intense and positive way of emotionally relating to co-members, which can be fostered by preparing patients for the group process, promotes the capacity to disclose and leads to more frequent and intense feedback from fellow patients. On the other hand, the patient who has a negative emotional relationship to other group members will disclose little and will receive relatively little meaningful feedback. Tschuschke and Dies (1994) conclude that there is ‘a complex interdependency among the three therapeutic factors of cohesiveness, self-disclosure and feedback, which promotes a working-through process that is also apparent in the improvement of interpersonal patterns (interpersonal learning-output) within the group and produces enduring intrapsychic changes in objects and self-representations (family re-enactment)’.

Early cohesion and the development of coherency, early disclosure to and confrontation with others, a largely positive alliance with others, and an increasingly noninterventionist group therapist are all linked to a positive outcome for the individual patient (Soldz et al., 1992; Strauß, 1992; Marziali et al., 1997).

Seidler (2000a,b) has found a significant correlation between increased self-relatedness and the reduction of psychosomatic symptoms among inpatients undergoing analytically oriented group psychotherapy. In the beginning of therapy he could observe a shift from somatic symptoms to neurotic symptom formation.

Liebermann (1971) has empirically shown that a well functioning group is able to establish a group culture in which the group members identify with the therapist's therapeutic attitude and thus become more and more therapeutically active.

Kordy and Senf (1992) have shown that being isolated with a specific symptom in a group leads to premature drop-out (as per the Noah's Ark Principle, see above).

A review of the literature dealing with the empirical research of group psychotherapy by Dies (1993) and by Burlingame et al. (2001), 2002) argued that group psychotherapy is effective in cases of alcoholism, anxiety disorders, bereavement, eating disorders, depression, schizophrenia, and sexual abuse.