More than one way home - Student raters’ impressions of interventions and group processes in mentalisation based group psychotherapy and group analytic psychotherapy

Abstract

In a study comparing mentalisation-based group therapy (MBT-G) and group analytic psychotherapy (GAP) in a day clinic, both group psychotherapy forms were found to be highly effective. But how did specific interventions and processes in both groups differ? The present article describes student raters’ impressions. Twelve psychology students listened to 100 audio recordings of 90 minutes group psychotherapy sessions of GAP and MBT-G. Each session was randomly assigned to two student raters, who were asked to write down their impressions. These were analysed. Group conductors in MBT-G used more questions, had short shares of speech, used group dynamics and fostered multiple perspectives on the issues discussed. Affect perception was stimulated by asking questions. In PDGT, conductors used more interpretations, confrontations and supportive interventions, and they had longer share of speech. Handling of affects was based on ‘allowing to get infected’. It is hypothesized that symptom reduction in both groups occurred via different ways: in GAP the pathic (affective contagion) function of interactions was more relevant, while in MBT-G it was the phatic (contact keeping) function. Results are also discussed in relation to previous findings on group processes and interventions.

Introduction

In a review on recent developments in group psychotherapy research, the authors positively conclude that the amount and diversity of research on group psychotherapy have significantly increased (Rosendahl et al., 2021). Yet, in spite of having the longest tradition in the field, only few studies in a recent systematic review for psychodynamic group psychotherapy meet the necessary standards for outcome research (Blackmore, 2009). A meta-analysis was therefore not possible. Some of the reasons included missing specification of group approaches and interventions (Janssen, 2018). For group analytic psychotherapy, Schultz-Venrath and Döring (2009) describe that, in spite of the attempt to find a common ground on scientific congresses, the many ‘group analytic dialects’ differ strongly. They argue in favour of a debate analogous to the discussion on pluralism and unity in individual psychoanalysis, which has been taking place more and more in recent group analytic discussions (“one group analysis or many?”, Lorentzen, 2011; Dalal, 2018).

An empirical link is thus desirable, as well as a bridging of the gap between clinical practice and research (Weber et al., 2013; Castonguay & Hill, 2017; Pries et al., 2019). To meet outcome research standards, it is essential to define and describe which psychotherapeutic approach has been used - and to test for treatment integrity. This includes therapists’ adherence to the specific psychotherapy form, their competence in delivering treatment and differentiation from other psychotherapy forms (Perepletchikova, 2011). Therefore, treatment manuals are a necessary basis - which leads to a dilemma for dynamically oriented therapists and researchers: the tension between this need for manualization and the risk of losing the dynamic aspects (Silverman, 1996). As a solution, Ogrodniczuk and Piper (1999) propose guidelines for psychodynamic psychotherapies which facilitate flexible reactions to the material brought into sessions by patients. The first and to date only guidelines for group analytic psychotherapy (GAP) were developed recently by Steinar Lorentzen (2013). In an RCT study comparing short (20 sessions) and long-term (80 sessions) group analytic psychotherapy, he found that patients with personality disorders benefit more from the long-term group psychotherapeutic approach. Furthermore, manuals for psychodynamic group psychotherapy have been developed by Sigmund Karterud (2015) for mentalisationbased group therapy (MBT-G), by Whittingham (2017) for an attachment theory-based ultra-short-group psychotherapy and by Tasca et al. (2006) for psychodynamicinterpersonal group psychotherapy. Strauß and Mattke (2018) stated that these advances and the corresponding studies could be interpreted in a hopeful view as a slow growing empirical revelation of the full potential of psychodynamic group psychotherapy.

Therapy effects of mentalisation-based and psychodynamic group psychotherapy in a randomized day clinic study

Brand et al. (2016) compared effects of MBT-G to GAP in a randomized controlled intervention study. In three main hypotheses, they tested whether MBT-G is superior in symptom reduction, improvement in quality of interpersonal relationships or in improvement of the ability to mentalize. The only difference between the two compared groups was the form of group psychotherapy. The group psychotherapy took place four times a week for 90 minutes, while treatment integrity was monitored by supervision. The influencing factors on outcome are depicted in the Table 1, sorted by a scheme proposed by Burlingame et al. (2004).

Group leaders were trained and supervised based on manuals for GAP (Lorentzen, 2013) and MBT-G (Karterud, 2015). This implicated guidelines for psychotherapeutic- technical handling of group processes. The applications of the Lorentzen guidelines have been adapted, as the average treatment duration of nine weeks is a high-dose short-term group psychotherapy. Therefore, interpretative and supportive interventions should have been used.

For both forms of group psychotherapy, Brand et al. (2016) found large effect sizes in symptom distress reduction, medium effect sizes for changes in interpersonal relationships, and small to medium effect sizes for mentalizing ability. A particular strength of the study is the detailed consideration of patient characteristics and the control of a large number of confounding variables that might influence the outcome in a naturalistic study. In addition to the variables reviewed, intervention style and theory of change are necessary test variables of highquality effectiveness research (Burlingame et al., 2013), which are captured in dissertation projects measuring treatment adherence and differentiation in both groups (Pries, 2021).

Treatment adherence and differentiation in psychodynamic group psychotherapy research

Although there is a research tradition of investigating leader behaviour that dates back to the 1950s in the context of process research (Bales, 1950), treatment integrity has been largely ignored in empirical research regarding psychodynamic group psychotherapies (Karterud, 2015). Only two of the five RCT studies in Blackmore’s systematic review (2009; 2012) provided information on adherence measures - the remaining three did not use manualized treatments. According to Perepletchikova (2011), procedures for addressing treatment integrity can be described on a continuum ranging from essentials (level 1) to optimal adequacy (level 5). While a specific treatment manual is part of a level 1 procedure, direct assessment of adherence (via recordings) would be an example of level 3, and the reduction of rater bias by blinding, multiple raters per target, consensus ratings and statistical correction of distortions would be an example of level 4. Each new level builds upon procedures at previous levels - to fully reach level 4, the difficulty for psychodynamic group psychotherapies lies in a partial lack of assessment measures with good psychometric properties. An exception is the MBT-G-AQS (Mentalisation-Based Group Therapy Adherence and Quality Scale) developed by Karterud (2015) and Folmo et al. (2017). It is a rating manual comprising 19 items based on intervention categories described in the MBTG manual (“1 managing group boundaries”, “2 regulating group phases”, “3 initiating and fulfilling turn-taking”, “4 engaging group members in mentalizing external events”, “5 identifying and mentalizing events in the group”, “6 care for the group and its members”, “7 managing authority”, “8 stimulating discussions on group norms”, “9 cooperation with co-therapist”, “10 engagement, interest and warmth”, “11 exploration, curiosity and not-knowing stance”, “12 challenging unwarranted beliefs”, “13 regulating emotional arousal”, “14 acknowledging good mentalizing”, “15 handling pretend mode”, “16 handling psychic equivalence”, “17 affect focus”, “18 stop and rewind”, “19 focus on the therapist-patient relationship”. It has been used to show treatment differentiation for mentalisation-based and psychodynamic group psychotherapy (Kalleklev & Karterud, 2018).

For group analytic psychotherapy, there is no such assessment measure available to date. This contrasts with the fact that group analytic theory by Foulkes (1975) can be viewed as the most influential concept in European group psychotherapy (Schlapobersky & Pines, 2009). Lorentzen has developed an adherence rating scale to distinguish between short-term and long-term GAP, and a scale for therapist competence in group analytic psychotherapy. The short-term adherence rating scale is based on MacKenzie’s developmental stages in groups (1997) and comprises: i) total activity of the therapist; ii) signs of a circumscribed focus (which is rated in five sub-items); iii) work in the here-and-now; and iv) work on termination.

The overall adherence in Lorentzen’s SALT-GAP study has been based on evaluating the presence of work with transference and resistances at group and individual levels, a process orientation of groups and whether therapists and group members worked with interpersonal issues, but the operationalisation of these items remains unclear. This may be due to the fact that the SALT-GAP study aimed at comparing short and long-term group analytic psychotherapy, and not so much to the operationalisation of adherence measures.

For the development of a global adherence assessment measure for group analytic psychotherapy, the operationalisation of the Foulkesian core concepts of configuration and localisation is still worked to be done - although there have been attempts for this in the first description of the interventions in the Lorentzen Manual by Garland et al. (1984).

Table 1.

Influencing factors on outcome of MBT-G and GAP (modified after Burlingame, 2004).

	MBT-G	GAP
Formal theory of change	Fostering mentalizing: improving ability to interpret human behaviour in terms of mental states (affects, thoughts, intentions) in oneself and others	Insight and emotional experiences, uncovering unconscious intrapsychic and interpersonal conflicts, ego-training in action
Small group processes	Cohesion in both group therapies comparable to closed groups
	Higher conflict-scores at the beginning of the group; fluctuations of group relationships much shorter, without a significant tendency; increase of cohesion at the end of therapy	More avoidance during the course of therapy; group conflicts of most successful patients undergo strong and ongoing fluctuations during therapy process; some of the most successful patients show a significant decrease in cohesion
Patient-variables
Affective disorders	64 (57.7%)	54 (54.4%)
Anxiety and somatoform disord ers	39 (35.1%)	43 (43%)
Personality disorders	53 (47.7%)	55 (55%)
Therapist-variables
Therapist behaviour	Participating > observing	Participating = observing
Activity	Active > passive	Active = passive
Prioritizing	Individual > group	Individual = group
Intervention techniques	MBT-G interventions based on Karterud manual (2015); Stop and rewind	GAP interventions based on Lorentzen manual (2013); Interpretation, guided facilitation, modelling, no immediate response
Current training	Individual analytic (3) Group analytic (1) Systemic family therapy (1)	Individual analytic (4)
Setting-variables*
Dose (average)	36 sessions (90 minutes, within 9 weeks)
Leaders	2 the rapists
Change of therapists after (average)	9 months	12 months
Group size	9 patients
Composition, style	Heterogenous diagnoses, slow open

*Group therapy was framed by a day clinic programme (5 days a week), which included individual chief and senior physician visits, primary nursing talks; dance, art and social therapy in groups; medication (if indicated); and others.

In a recent quantitative research approach, the intervention categories described in the manual of Lorentzen have been used for testing treatment adherence and differentiation. While showing promising results with a substantial inter-rater reliability of 0.724 and more than 60% group analytic interventions over all sessions (Pries, 2021), the wideness of intervention categories made rater training a challenge.

In the present study a qualitative approach to research on treatment adherence and differentiation is used. The aim is to explore differences between GAP and MBT-G shown in raters’ free associative impressions while and after listening to group psychotherapy sessions.

The following research questions were formulated: i) Do student raters’ impressions show differences which are in line with treatment adherence and differentiation? ii) Are there differences in emotional experiences of the rating process for the two different group psychotherapy forms? iii) Are differences between both group psychotherapy forms visible in the audio material?

Materials and methods

The present study is part of the controlled health service research study in a day clinic described above (Brand et al., 2016) based on audio material recorded for the purpose of adherence ratings. In a doctoral dissertation (Pries, 2021), a mixed-methods design was used to test for treatment adherence and differentiation. The present article focuses on the results of the qualitative part - results of the quantitative part are not yet published in a journal.

Material and origin of material

The material comprises a total of 169 student rater impressions - 74 for MBT-G and 95 for GAP. They provide for the quantitative research on adherence and differentiation: Twelve psychology students from the University of Witten/Herdecke took part in a project entitled ‘Listening to group psychotherapy’. They were randomly assigned to training sessions for recognizing specific interventions based on the above-mentioned GAP treatment manual and the MBT-G-AQS. The training consisted of weekend workshops with a total duration of 14 hours. Student raters were introduced to the intervention categories through a PowerPoint presentation. Examples of interventions for each category were given and markers to rate interventions belonging to the category were discussed. Afterwards, sample audio recordings were rated, and the reasons for each rating were discussed. The inter-rater-reliability was based on Cohen’s kappa ≥0.60, calculated over one session of 90 minutes, each intervention being rated separately. For the qualitative part of the project described in this article, raters were asked to write down their associations, impressions, feelings and thoughts for each session they rated. They wrote them down in the same Excel-rating sheet which was used for the quantitative part of the research project, while or after listening to the session.

After completing the training, student raters listened to a total of 100 group psychotherapy sessions of 90 minutes duration, blinded for the specific group psychotherapy form. These 100 sessions were 50 MBTG sessions randomly drawn from a pool of 443 recorded MBT-G sessions, and 50 GAP sessions randomly drawn from a pool of 445 recorded GAP sessions. Each session was rated by at least two raters. Eight of the psychology students already had experience with group psychotherapy, e.g. participating in a psychotherapy group from internships or conducting groups on their own in clinics. Each session was rated by at least one student with previous experience in group psychotherapy. The numbers of rater impressions differ for MBT-G (74) and GAP (95), as not every rater commented on each session.

Structuring and summarizing content analysis of rater impressions

A structuring content analysis (according to Mayring, 1983) was conducted. First, the categories “therapeutic style”, “therapist behaviour”, “group climate”, “structure/focus”, “group process”, “therapeutic process”, “group members behaviour” and “rater’s emotional experience” were defined deductively based on the research hypotheses above. Second, typical examples of the categories were chosen; third, coding rules were formulated. Next, tests were conducted to establish whether the categories fit with the material. Therefore, find-spots were marked with different colours. After approximately 50% of the material, the categories were revised and reduced to three main categories (Table 2). This was due to the fact that some categories were too narrow and not helpful enough to structure the material.

Afterwards, a summarizing content analysis was conducted by paraphrasing parts of the rater impressions carrying content, generalizing them on an abstraction level based on group psychotherapy-specific considerations, reducing them by selecting and discarding paraphrases with the same meaning, and finally putting contents together as a category system (Table 3). As the impressions differed markedly in length, it was possible that aspects of one impression were put into two categories. Finally, within these categories, frequencies were counted for specific words which were found particularly often.

Experimental approach

Furthermore, for an experimental research approach, 10 minutes of each group psychotherapy were turned into audio events using the audio software Logic. Next, patients and therapists were each assigned one tone on a virtual keyboard (therapists in low tones, patients in high tones) and assigned different colours (therapists in green, patients in other colours). The 10 minutes were speeded up to one minute in order to make the differences hearable.

Results

The structuring content analysis of a total of 169 rater impressions (95 for GAP and 74 for MBT-G) yielded three main categories “therapist behaviour”, “group process and group dynamic” and “emotional experience of students”. The subsequent summarizing content analysis resulted in four subcategories for the first main category (“intervention technique and style”, “handling of affect”, “praise” and “criticism”) and three subcategories for the second main category (“most frequent process”, “praise” and “criticism”; Table 4). The category “intervention technique and style” refers to how group leaders intervened, while “handling of affect” contains raters impressions of how they handled their own and the patients affects. The two categories “praise” and “criticism” contain positive and negative remarks on how therapists behaved. For “group process and group dynamic”, the subcategory “most frequent process” contains impressions on processes often found - while “praise” and “criticism” again refer to positive and negative connotations in raters impressions.

Category 1 - Therapist behavior

There were remarkable differences for intervention technique and style. In MBT-G, the word “question” appeared seven times, e.g.:

• “the therapist gives new impulses by questions”,

• “the therapist lets processes deepen through questions”.

In GAP, on the other hand, the word “interpretation” appeared five times, e.g.:

• “the therapist gives interesting interpretations”,

• “there were many interpretations from the therapist”.

Table 2.

Categories, typical examples and coding rules used in the first attempt at structuring, and the resulting three final main categories.

Category	Typical examples	Coding rules	Final main category
“Therapeutic style”	“Sometimes the therapist switches in style to being rather provocative”	Style is directly or indirectly described	Therapist behaviour
“Therapist behaviour”	“Therapist talked very, very much, interrupted patients multiple times and appeared very teacherly”	Observations on the behaviour therapists display
“Therapeutic process”	“A lot of rationalisation, dwelling on generalisations - little here-and-now”	References to therapeutical intentions of the process	Group process and - dynamic
“Group climate”	“Quick-tempered therapy session”	Atmosphere or climate of the group session is described
“Group process”	“The session is unspectacular, but processes seem to gain momentum”	Observations on the presence of group processes
“Group members’ behaviour”	“Fellow patients acting very therapeutically and at the same time confrontational”	Behaviour of group members is described
“Structure/focus”	“Very confused, muddled, distanced, unfocused, almost trivial, silly”	The dimensions of structuring and focusing vs Chaos are addressed
“Rater’s emotional experience”	“Pleasant”, “arduous”, “enjoyable”, “tiring”	Any statement relating to how raters experienced listening to the session	“Raters emotional experience”

Table 3.

Examples for the summarizing content analysis.

Rater	Session number	Group type	Paraphrase	Generalisation	Reduction
1	03	MBT-G	“Group by itself is very active and humorous”	Good group dynamics and climate	Dynamic and independent working of the group
2	32	MBT-G	“The group has worked well on its own”	Group working independently
3	64	GAP	“Very lively conversation between patients”	Lively dynamic
4	45	GAP	“Group by itself having an animated conversation”

The word “confront” appeared six times among the impressions of GAP sessions, e.g.:

• “the patient was very clearly confronted”,

• “the therapist used much confrontation”,

Based on the raters’ impressions, the style of group therapists in MBT-G can be quite well characterized through following up with interest, using group dynamics, fostering change of perspective and multiple opinion, for example:

• “the therapist often followed up, wanted to reveal implicit assumptions”,

• “the therapist mainly works by getting the group moving and stimulating exploration”,

• “the therapist encourages other opinions and change of perspective”.

Therapeutic style in GAP can be described more as supportive, confrontative and teacherly-educative, for example:

• “it was pleasant how the therapist provides support in a minimalist way”,

• “the therapist gives many examples... and you could have said: xyz”,

• “the therapist was very clearly confrontative”,

• “acting a very teacherly way” (instructional).

Parts of speech of the MBTG-therapists were rather short and came quickly one after the other, for example:

• “intervening very little,”

• “small interventions, which appear very effective” while in GAP, therapists tended to use long parts of speech in succession, for example:

• “big parts of speech all at once”,

• “talks at great length continuously ”.

In the therapeutic style MBT-G and GAP group therapists had in common that they provoked on the one hand and held back/let group processes run on the other, for example:

• “therapist very provocative” (MBT-G),

• “sometimes the therapist switches in style to being rather provocative” (GAP);

• “therapists held back a lot and did not say much”,

• “I like that therapist let the discussion run” (MBT-G) and

• “therapist scarcely intervenes ”,

• “therapist holds back a lot” (GAP).

Table 4.

Rater impressions, similarities and differences for MBT-G and GAP (N=100).

	MBT-G	GAP
Therapist behaviour
Intervention-technique and style	- Asking questions	- Interpretive
	- Following up with interest	- Confrontational
	- Using group dynamics	- Supportive
	- Encouraging change of perspective and multiple opinions	- Teacherly, didactic
	- Intervening very little	- Big, long parts of speech in a row
	- Short, quick parts of speech one after the other	- Interrupting
	- Provocative
	- Holding back/letting things run
Handling of affects	- Stimulating affect perception through questions	- Allowing oneself to be infected by affects
Praise	- Addressing the group as a whole	- Emotionally devoted
	- Using group processes	- Supportive - Empathy - Multiple patients get a chance to speak
Criticism	- No depth reached	- Out of tune
	- Text book-like/“right-wrong”	- Avoiding/little in the “Here & Now”
	- “The method is annoying”	- Too much “individual therapy in the group”
Group process and - dynamic
Most frequent process	- Involvement of all participants	- Individual therapy in the group (individuals in the foreground, group in the background)
		- Figure-ground-constellations
Praise	Dynamic and independent working of the group
	- Participation and inclusion of all participants	- Multiple participants “get something”
	- Dissent allowed	- Balance between needs of individuals
Criticism	- Defensive/inauthentic	- Empathy, support - Too much space/time for individuals
	- Participants not protected enough	- Monologizing
Rater’s emotional experience	- Stagnation	- Dyadic arguments
	Chaos/not enough structu	re and focus
Positive	28	19
Negative	24	28

In MBT-G, an asking approach was observed, for example:

• “interesting how much the group members are often asked about their own feelings”.

In GAP, there were six rater impressions which can be summed up as “allowing oneself to be infected by affects”, for example:

• “the therapist sometimes also let herself be affected by the groups’ anger”,

• “therapists let themselves get carried away by the prevailing mood of the group”.

In MBT-G, affect was rather handled by stimulating perception of affect through questions, for example:

• “interesting, how much group members are often asked for their own feelings”.

Positive comments were made on addressing the whole group and using group processes in MBT-G, for example:

• “once and again activates the group”,

• “what does the group think’”,

• “therapists work more with group processes”.

In GAP, there was praise for emotional devotion, security, support, empathy and room for multiple patients, for example:

• “emotionally warm and rationally engaged”,

• “everything in a validating and comprehensible way”,

• “therapist protects [the] patient in spite of inadequate behaviour and stays neutral towards the group”,

• “the therapist responded very empathically to patients and considered other patients even if one issue was very time consuming”.

Critical impressions of raters can be summarized as “no depth reached”, “text book-like/right-wrong’”, and “the method is annoying”, for example:

• “therapist who includes the group, but always in the same way, thus not going deep”,

• “a kind of teacher-student atmosphere prevailed, because one was searching for the right answer to the question”,

• “the therapist has something particular in mind, which makes the therapeutic session very, very suffocating to follow”,

• “why is the question always being asked `how do you feel, when xy?”

In GAP, critical rater impressions can be summed up as “out of tune”, “avoiding/little in the “Here & now”, too much “individual therapy in the group”, for example:

• “She also puts the patient down while he is talking about situations with the friend”,

• “I miss the relation to the “Here & now” - instead there were, very long-winded and evasive remarks of particular group participants”,

• “therapist is again hardly working with the group as an instrument”.

Category 2 - Process and dynamic

Most frequent process: The most frequent process found in MBT-G consisted in an involvement of all participants, for example:

• “Often involving the whole group”,

• “as the group participants enthusiastically participated in the conversation”,

• “as the group had a lot of space and engaged with one another”.

In impressions from GAP sessions, the processes which dominated can be described as “individual therapy in the group” and “individual in the foreground, group in the background”, for example:

• “many conversations with individual patients, sometimes rather seems like an individual conversation”,

• “for a long time rather the character of an individual therapy”,

• “most of the time only one patient was subject to cross-examination by therapists”.

There were also some indicators of figure-ground-constellations in the sense of observations, where individuals are described in terms of their relationship to the group, for example:

• “he still doesn’t seem to have found access to the group”,

• “moving theme in the beginning, which the group responds to empathically”.

Dyadic interactions were experienced as important parts of the process, for example

• “the session was productive for the group; I also experienced the rather dyadic parts as adding to the group process”.

In both group therapy forms, raters positively commented on dynamics and independent working of the group, for example for MBT-G:

• “the group has worked well on its own”,

• “group by itself is very active and humorous”, and for GAP:

• “very lively conversation between patients”,

• “group is working constructively”.

In MBT-G impressions, the group was more often thought of as an entity of its own, for example:

• “group is searching for opportunities of integration”.

This is reflected in a quantitative difference of the appearance of the word “group”: in a total number of 95 impressions from GAP, the word “group” was found 23 times (which amounts to 24%), while in 74 MBT-G, the word “group” was found 31 times (42%).

Positive comments on the group process for MBT-G included the involvement of all participants and the allowance of dissent, for example:

• “the group got a good chance to speak, many different longer parts of speech, not only understanding of participants but also discussions”,

• “a lot of participation, a lot of courage to also express criticism or present a different view”.

In GAP, there were quite positive remarks that multiple participants had room and could work on issues in parallel, that a balance between needs of individuals was reached and that empathy and support dominated, for example:

• “beautiful that multiple patients can talk about their concerns”,

• “interesting to see the parallels that have been shown between patients”,

• “different persons dive deeper into their topic”,

• “nonetheless also the issues of the other patients got a chance to be addressed (balance)”.

Criticism: Critical remarks for the impressions of MBT-G raters were that a lot of defensiveness dominated and the group treated one another inauthentically, for example:

• “palpable low ability or willingness to reflect in a huge part of the group/marked defensiveness”,

• “as the group was very inauthentic - lying to each other”,

• “seems like a functional, maybe rather rationalizing group, but anyway many feelings are experienced - and are also voiced ”.

Also there were concerns as to whether participants are protected enough, for example:

• “a pity that one patient was pilloried by the group”,

• “in parts I find it a bit hard how the other patients voice their own opinions toward the one patient towards the end, as the session is already finishing and I’m not sure how well it can be absorbed”.

Finally, there were critical remarks on stagnation of processes, for example:

• “long circulating around the same issue, little development in a group process that seemed tough”,

• “issues are treated superficially and insistently”,

• “in total the group has sometimes moved on the spot”.

In GAP sessions, critical impressions rather concerned the view that too much space and time were used for individual patients, for example: “the patient took up a lot of attention and energy”. Furthermore, it was remarked that there were long monologues of individuals, for example:

• “patient with personality accentuation dominates (the first part of the session), which is unnerving me as listener, also because I ask myself how the rest of the group is doing“,

• “during the whole session one patient is the centre of attention, so the group essentially becomes invisible”.

Finally, raters made critical comments on dyadic conflict conversations: “the conflict between two patients takes over the whole conversation and excludes most of the other patients”.

For both group psychotherapy formats, the critical observation was made that there was too much chaos with too little structure and focusing, for example:

• “chaotic, a lot of talking all at once”,

• “very confused, muddled, distanced, unfocused, almost trivial, silly”,

• “the session seems ‘untidy’ in total, no topic is finished”.

In MBT-G, there was more often criticism on having no clear ending of the session, for example:

• “irritating also: No welcoming and farewell”,

• “no final end is found”.

In GAP, there were more often critical remarks on interrupting each other, for example: “especially in the last fifteen minutes people often talked in parallel”, “a lot of talking at once and interrupting”, “at times it bothers me when the patient interrupts the therapist, and also other patients do it at the end of the session”.

Category 3 - Rater’s emotional experience of sessions

There were clear differences: in the impressions for MBT-G, there were 28 positive adjectives (for example: “pleasant”, “exciting”) and 24 negative ones (e.g. “lengthy”, “stressful”). In GAP, there were 19 positive adjectives (e.g. “interesting”, “pleasant”) compared with 28 negative ones (e.g. “uncomfortable”, “frustrating”).

Experimental approach

The differences in the length of interventions and utterances of each group member are depicted in Figure 1. This is also exemplified in Figure 2, where two randomly chosen group psychotherapy sessions were coloured and sliced to show differences. In GAP, in the first box one patient (light brown) had a long share of speech, followed by a silence (marine blue). Another patient (dark brown) had a longer share of speech, to which the therapists replied (light green). After a silence (marine blue), another patient (dark violet) talked for a longer period of time; this was also followed by silence (marine blue). In the second box, after a long share of speech by another patient (light violet), therapists had short shares of speech; then another patient came in (purple) who had longer shared of speech. Another patient came in (pink), replying to him. After a long silence (blue, after the second box) this patient (pink) talked for a long time during the last third of the session, only interrupted by the therapists (light green).

Figure 1.

Ten minutes of MBT-G in one minute and ten minutes of GAP each speeded up to one minute. In each upper row, therapists’ voices are visible in green; in each lower row, patients’ voices are visible in brown, orange, red and more. A video with sound can be requested from the author.

Figure 2.

Randomly chosen group psychotherapy sessions exemplifying the differences between group processes in MBT-G and GAP.

In MBT-G, in the second third of the session three patients interacted with shorter parts of speech, and the therapist was also involved. There was almost no silence (marine blue). In the last third of the session, a lot of group interaction took place with five patients and the therapist - each one had short shares of speech.

Discussion

Based on the rater impressions, therapist behaviour can be summarized as follows: In MBT-G, therapists frequently used short interventions articulated as questions to activate the group. Most often, these were directed to the group as a whole, and all group members were involved. It was positively remarked that therapists used the here and now of the group, worked a lot with group dynamics and all patients were included. Critical remarks focussed on therapists not going into depth and acting in at text book-like-manner, being busy with right or wrong. In some cases, raters also showed subjective irritation over the intervention technique.

By contrast, in GAP, the way of intervening involved a lot of interpretation and confrontation - some therapists being provocative, allowing themselves to be infected by the affects of patients. There were positive remarks on therapists’ supportiveness, emotional devotion and empathy - and also on the way they make room for multiple individuals. However, criticism was based on therapists being out of tune, not enough focused on the here and now in the group and doing too much individual therapy in the group.

In both groups, raters described the behaviour of therapists as provocative at times and holding back to let processes run. The latter can be viewed as an analytic therapeutic attitude of leaving room for processes to evolve, which recent studies examining speech in psychotherapy have found to be linked to a positive therapeutic relationship (Steinert et al., 2022). The differences in intervention technique are especially found in the subcategory of intervention technique. In working with affect, the way GAP therapists allow themselves to be infected by the patients’ affects can be interpreted as working through transference processes.

We assume that the voice of the therapists had different functions in both groups. For the development of the self, Anzieu (1996, p. 208) describes a “soundshell” which leads to the “introjection of the acoustic world”. He describes it as “auditive-phonic [...] skin [...], which plays a crucial role in the attainment of the ability of giving meaning to objects, later in the acquisition of the symbolizing function of the psychic apparatus”. Felsberger (2017) emphasizes that the voice “maybe more than the eye can be seen as an interface, as “joint embodiment”. In this manner, joint embodiments between two persons have taken place much longer in succession in GAP, while in MBTG there were shorter joint embodiments of individuals with other group members. Due to the shorter interventions and the many instances of turn-taking in MBT-G on a musical level, the rhythm of interactions played a greater role - in such a way practising affective attunement as between mother or father and child in the group. The effect of the voice of therapists might then have been different in both groups. In MBT-G, therapists’ voices had the function of activating the group again and again to establish and maintain emotional contact (phatic function, Felsberger, 2017). In GAP, the function of the therapist’s voice was more on the pathic function, transporting affect in long dialogues between individuals and group therapists. The audio-transformation exemplifies these differences in the structure of communication. Beneath the differences in turn-taking, especially the different lengths of shares of speech are visible.

Schultz-Venrath and Felsberger (2016) warn against a “teacher-position”, if group therapists define themselves as experts based on turbulent situations in groups or fear. Rater impressions suggest that this has sometimes happened in both group psychotherapy formats.

The critique of a text book-like right or wrong attitude for MBT-G can be understood as a bias known from previous psychotherapy research, where trained study therapists give up some of their flexibility for adherence to the treatment protocol (Vanderbilt II, Henry et al., 1993). Critical comments on lack of depth in MBT-G may mirror the fact that the psychotherapy form is not designed to reach towards deeper unconscious phenomena as GAP is, but rather to strengthen pre-conscious functions of the mind.

The observations on group processes show differences. In MBT-G, the most commonly observed dynamic was an inclusion of all participants while utterances were rather short. In GAP, the focus was more on individual participants with longer parts of speech by group members and therapists. The group itself often built a background for the discussed topics of one individual in the foreground. The process in MBT-G is in line with the goal of the manual (Karterud, 2015) to use the group as a training ground for mentalizing, while balancing turn-taking. Longer silences between group members have been avoided, and an active attitude of therapists hindered the unfolding of transference processes. In contrast, the process model in GAP with Schultz-Venrath (2012) can be described as analysis of the individual in the group (Schilder, 1936) and of the individual through the group. Rater impressions and unpublished quantitative evaluations suggest that therapists have often offered themselves as a transference object, but rather neglected transferences between group members and the analysis of the group through the group as described by Foulkes (1975) and Lorentzen (2013). The experimental-exploratory observations of randomly chosen audio material point in the same direction. However, this interpretation has to be read with caution, as there are to date no operationalisations of the central group analytic concepts of localisation and configuration.

For now, it remains a desideratum to develop an instrument able to measure adherence to group analytic psychotherapy based on free group communication, that is group dialogue. The type of communication displayed in the present GAP audio recordings was largely dominated by monologues of individual patients, thus not displaying the playfulness characteristic of well-functioning group analytic groups. In this light, the fact that dependence on the group therapist was higher for GAP (Hecke et al., 2016) is to be seen as part of the particular group psychotherapeutic model used and cannot be generalized for group analytic psychotherapy. It is in line with the finding of Sabel (2007) that the individual personality and style of intervention of group therapists are as important for the outcome as the specific form of psychotherapy - maybe even more influential than the latter.

The fact that, in central phases of group development, the group as a whole was more important in MBT-G, while the group leader was more important in GAP (Hecke et al., 2016) fits well with the impressions of raters. However, they may be due to the fact that none of the study therapists were in group analytic training, thus implicitly deviating from group analytic principles by using many interpretations directed towards individuals in the group (Pries, 2021). There were positive comments for both group psychotherapy forms on the independent working of the group and group dynamics. This may mirror the fact that both were part of a group analytically informed clinic which gives group dynamics a core position in treatment. The worry about the protection of patients can be connected to the intervention technique of often asking questions to the whole group, thus re-enforcing potential confrontations through the group as a whole and individual members. It seems that, in this kind of process, raters were more worried about individual patients being potentially consigned to the aggression of the group. Critical comments on too little structuring and too much chaos can be used as an important hint for future research, supervision and foci in training. It seems as though, in MBTG, not enough attention has been paid to the structuring frame in some sessions - which would suggest a neglect of the intervention ‘regulating group phases’ in the manual (Karterud, 2015). In GAP, a group culture of verbal aggression (Alder & Buchholz, 2017) seems to have crept in during some sessions, a culture in which a form of subtle communicative violence reveals itself in the superficial structure of communication as not allowing others to finish speaking.

Finally, the emotional experience of raters shows the difference of both group psychotherapy forms. The higher number of negative affects while listening to the sessions can be understood as an effect of negative transference aspects. In MBT-G, negative affects combined with negative transferences onto the therapists and the group was prevented from unfolding by brief questions which helped the group digest negative affect. In contrast, the unfolding of negative affects in the group, with group therapists repeatedly offering themselves as transference figures to then confront infantile wishes and unconscious imaginations, has led to more discomfort in raters while listening.

Potential use of ratings for training

For a long time there have been demands to include more research in psychotherapeutic training (Whitaker, 1992; Piper, 2004; Schultz-Venrath & Döring, 2009) - also attempts at bridging the gap between clinical practice and psychotherapy research (Weber et al., 2013). Fonagy and Luyten (2019) argue that a focus on adherence and flexibility is necessary in psychotherapeutic training to facilitate psychotherapeutic adaptation to individual patient characteristics. Yalom and Leszcz (2007) emphasize the central role of participating observation for group psychotherapeutic training. Rating processes can have a similar effect by enabling identification with therapists. Identification is one of the mechanisms of change in group psychotherapies (Mattke, 2018) and at the same time the basis for the formation of a psychotherapeutic identity (Leichsenring et al., 2019). Using ratings of group psychotherapy sessions as part of psychotherapeutic training can give the opportunity to identify and to develop inner representations for one’s own psychotherapeutic interventions. Also, student raters can get an idea of which therapeutic styles might attract them for their own future interventions with patients. The following quotations from a survey at the end of the research internship Listening to group psychotherapy illustrate this:

To the question “Please describe your learning experiences in the research project ‘Listening to group psychotherapy’. Answer: “Through the participation of therapists and the following reactions of the group I could often make up my mind on which comments of therapists seemed useful or sometimes inadequate to me. I learned from the positive and negative examples of therapists. But I also often felt that the therapists only have limited influence on the group process and furthermore often didn’t have much time to think about their participation in the group process, as such a group often moves quickly into different directions, difficult to steer.”

To the question “Did something change in your way of understanding psychodynamic group processes during the training or during the rating? If so, what?” Answer: “While at the beginning of the rating process I felt that I can hardly learn anything from therapists’ comments, which often seemed random to me, this feeling has changed with listening to more sessions. In the end I could better anticipate the reactions of the group and I now believe that leading such a group psychotherapy also improves with experience. Even if patients in the one group were different from those in the other, there were similar role patterns.”

Strengths and limitations

Given the above-mentioned dilemma of manualization and the lack of global adherence scales for group analytic psychotherapy to date, a strength of the present study lies in a view on treatment adherence and differentiation based on a phenomenological bottom-up approach. The view of student raters who are blind to the specific psychotherapy form is a fresh one and can be helpful in not endangering the positive developments in psychodynamic group psychotherapies through ideological debates as pointed out by Strauß (2021).

The approach may help complement quantitative views based on the theory-based definitions of intervention categories (Garland et al., 1984), which are often “inconsistent, abstract, ambiguous, unspecific and heterogenous” (Gumz et al. 2017). The empirical findings relativize the “ideal group analysis” with which every group leader (and maybe every group analytic apprentice even more so) is in conflict (Wilke, 2015).

However, free associations were not as free as they might have been: raters will have been influenced by the particular training on specific interventions and the implications of the model. Despite two-thirds of raters having experience with group processes, as the quotes above illustrate, raters may often have felt lost in group processes - thus potentially influencing the emotional experience of ratings. Although each session was rated by one rater with experience in group psychotherapy, their clinical and empirical experience was low. Furthermore, they only rated one session without having the context of the other three sessions and the ongoing group process, which may have added to difficulty in describing what actually happened. Differences between the two impressions by group psychotherapy format were quantified based on keywords found particularly often, for example “questions” in MBT-G and “interpretation” in GAP. Previous research showed that, at least for this core specific intervention category, which is generally more typical for individual than for group analytic psychotherapy, experts expressed a high degree of agreement in a survey that it is being (Pries et al., 2019). The keywords for GAP again lead back to the problem of defining intervention categories: What is it that may be defined as an interpretation?

Another limitation concerns the qualitative content analysis, which has only been conducted by a single person (JP). Therefore, inter-rater reliability was not tested and the reliability of this method is unclear.

For the interpretation of the results in combination with the study of Brand and colleagues, a limitation of the original study is that study therapists were not completely trained in the specific psychotherapy forms MTB-G and GAP, but received supervision based on the specific guidelines - some of them training at the same individual analytic institute. The many individual interventions and the few comments to the group as a whole in GAP may be connected to the fact that study therapists were only in individual analytic training. No evaluation has been conducted to establish how competent group therapists delivered the specific interventions. Also, it is possible that the lower activity of GAP therapists and the higher activity level of MBT-G therapists which are in line with the manuals contributed to raters’ emotional experience and the perceived group processes.

Conclusions

A research contribution on similarities and differences between two psychodynamic group psychotherapy forms. Rater impressions by students who were trained based on manuals for GAP and MBT-G show that processes and interventions in both psychodynamic group psychotherapies differed, while therapy effects in symptom reduction was equal for both groups. It may be interpreted as another account for the ‘Dodo bird verdict’ (Rosenzweig, 1937) - for improvement of patients in a day clinic, there is more than one way home. Adherence research may be a helpful learning tool for psychotherapy training.