Group Cognitive–Behavioral Therapy in a Sample of Dutch Intimate Partner Violence Perpetrators: Development of a Coding Manual for Therapist Interventions

ABSTRACT

This qualitative study investigated which types of interventions are made by group therapists during group cognitive–behavioral therapy (GCBT) with individuals who have perpetrated intimate partner violence (IPV) in the “Not ‘Losing It’ Anymore”(NLIA) outpatient programs across the Netherlands. Audio recordings of group sessions (N = 60) facilitated by 13 co-therapist teams across 25 different groups were analyzed using the general inductive approach. Fifteen distinct therapist interventions were identified, categorized, and described using examples from the data. Two categories of interventions could be distinguished: CBT interventions and process-oriented interventions. In future research the coding manual can be tested for interrater reliability and might be used in research on the effectiveness of GCBT therapist interventions.

Intimate partner violence (IPV) is a worldwide public health problem. A systematic review by the World Health Organization (WHO) shows that 35% of women worldwide experience physical and/or sexual IPV (Garcia-Moreno et al., 2006). IPV results in significant psychological suffering, physical harm and societal costs (Bonomi et al., 2009). Women who have experienced IPV are at serious risk of developing adverse health conditions and health risk behaviors (Bosch et al., 2017; Centers for Disease Control and Prevention, 2008). The consequences of IPV, such as criminal justice and legal interventions, physical and mental health care, decreased work productivity, unemployment and even disability, result in high societal costs (Bonomi et al., 2009; Duvvury et al., 2013).

IPV perpetrators face the consequences of their aggressive behavior such as relationship problems, self-esteem issues, police and justice involvement, and costs deriving from material damage. These consequences can result in motivation for behavior change, in which intrinsic motivation appears to be related to stronger therapy engagement compared to external motivation (Deci & Ryan, 2008; Scott et al., 2011).

In the 1970s, intervention programs for male intimate partner violence perpetrators were developed as part of a wider movement addressing the rights and needs of battered women (Bowen & Gilchrist, 2004; Feazell et al., 1984; Pirog-Good & Stets-Kealey, 1985). These programs were largely based on the so-called Duluth model that considers IPV as a way for men to assert power and control over their female partners (Pence et al., 1993). The model has been successful in increasing awareness of the problems created by IPV, but Duluth-based treatments did not appear to succeed in reducing the intensity or frequency of IPV perpetration (Babcock et al., 2004; Dutton & Corvo, 2007). Several authors suggest their limited effectiveness is due to a lack of focus on the emotions and needs of IPV perpetrators, which results in low treatment motivation (Gondolf, 1993; Star, 1983).

In the 1990s, cognitive–behavioral therapy (CBT) for IPV perpetrators was developed (Murphy & Eckhardt, 2005). CBT focuses on the changing of cognitions, emotions, coping, and anger management, and uses a collaborative style to engage participants (Day et al., 2009; Feder & Wilson, 2005; Saunders, 2008). CBT programs are based on well-developed theory and extensive empirical research demonstrating small to medium-sized treatment effects (Babcock et al., 2004; Day et al., 2009; Eckhardt et al., 2013; Graham‐Kevan & Bates, 2020). The limited treatment effectiveness tends to be attributed to patient characteristics such as substance abuse, age, childhood mistreatment, antisocial personality, and motivational problems (Butters et al., 2021; Godbout et al., 2019; Stephens-Lewis et al., 2021). These findings stress the importance of therapists adequately adjusting to patients’ needs and personal characteristics (Butters et al., 2021; Taxman, 2020).

Studies on psychotherapy effectiveness have emphasized that not only do client characteristics and therapeutic interventions contribute to therapy outcomes, but so do therapist characteristics (Castonguay & Hill, 2017), regardless of treatment type and client characteristics (Barkham et al., 2017; Johns et al., 2019; Wampold et al., 2017). Therapist interventions affect the therapist–patient working alliance and therefore affect treatment outcomes (Flückiger et al., 2018; Horvath et al., 2011). Examples of therapist variables that are connected to good client outcomes include therapist responsiveness, therapist presence, supportiveness, humor, and creativity (Anderson et al., 2016; Barkham et al., 2017; Marshall & Burton, 2010). Research indicates that differences in therapist effectiveness are related to “facilitative interpersonal skills” (Anderson et al., 2016, p. 513), regardless of experience or training (Goldberg et al., 2016).

If therapist interventions can enhance effectiveness in individual psychotherapies, this raises questions about therapist effectiveness in group psychotherapy, as group therapists must deploy facilitative interpersonal skills in unique ways in order to address multiple participants while attending to the group climate (Hahn et al., 2022). Akin to alliance research in dyadic therapy, group psychotherapy research also emphasizes the importance of alliance (group therapist to group member) as well as the unique variable of group cohesion (group-as-a-whole climate) in group therapy (Marmarosh, 2019; Rosendahl et al., 2021; Shechtman, 2016). Attachment theory has often been used in this research as a theoretical basis for understanding the important role that the group plays for the group patient (Bowlby, 2012; Leszcz, 2017; Marmarosh, 2019; Tasca, 2014). Study outcomes show significant correlations between therapeutic alliance and treatment outcomes, and similar correlations are found for group cohesion and treatment outcomes, including CBT approaches to group therapy (Alldredge et al., 2021; Burlingame et al., 2018).

Considering the modest treatment effects found for CBT IPV perpetrator treatment, investigating the potential effects of therapist behaviors on therapeutic alliance and group climate may be promising. Studies acknowledge that therapist characteristics, participants’ perceptions of the therapist, therapeutic alliance and group climate are important factors in treatment effectiveness (Hamel et al., 2021; Marshall & Burton, 2010). Specifically, which therapist interventions can enhance treatment outcomes? Several categories of group therapy interventions have been studied empirically (Chapman et al., 2010; Yalom & Leszcz, 2020). In group therapy research, patient and therapist interpersonal behavior are often assessed using the structural analysis of social behavior (SASB; Lorentzen & Høglend, 2002; Tasca et al., 2011). The SASB was originally developed to assess psychopathology in terms of maladaptive interpersonal and intrapsychic patterns along three dimensions (Benjamin, 1974). Applied to group psychotherapy, the SASB provides insight into patient interpersonal behavior in groups and provides suggestions for more effective therapist interventions (Benjamin, 2000). In the literature on IPV perpetrator treatment, it is emphasized that research should not only assess therapist–patient interaction, but also how therapy techniques are delivered and to what extent therapists adhere to the treatment protocol (Eckhardt et al., 2006; Karakurt et al., 2019). To our knowledge, there is no coding system available that assesses both specific therapy interventions and process-oriented interventions concurrently.

The purpose of the present study is to make an inventory of the types of interventions used by group therapists during sessions of GCBT with IPV perpetrators, using a qualitative approach.

METHOD

Participants

Group Members

The GCBT groups are part of the “Not ‘Losing It’ Anymore” (NLIA) outpatient treatment program occurring at mental health treatment centers throughout the Netherlands (Van Dam et al., 2008, 2009). NLIA is an 18-session manualized GCBT treatment program for patients who have perpetrated IPV, who are seeking treatment voluntarily, and referred by their general practitioner. Apart from NLIA patients can receive medication, and they can get additional treatment like trauma treatment and partner counseling. A total of 155 group participants were recruited across 25 different NLIA groups, from four different mental healthcare centers, 133 of them actually participated in the treatment and were included in the final sample. Participants (patients as well as therapists) provided informed consent to participating in the treatment and in the research. Patients not wishing to participate in the research portion of the treatment protocol were still able to receive the treatment.

In addition to having a history of perpetrating IPV, most participants also carried a DSM-5 (American Psychiatric Association, 2013) diagnosis of intermittent explosive disorder. In addition, frequently co-occurring diagnoses among groups members were: cluster B personality disorder, PTSD, substance abuse disorder, depression, ADHD and anxiety disorder (for a typology of NLIA participants, see Serie et al., 2017). Demographic data for participants are displayed in Table 1. The NLIA groups are closed and typically comprise up to eight participants; in the current study, group size ranged from five to nine with an average of six participants and a modal number of six per group.

Table 1. Patient Gender, Age, and Marital Status (N = 155)

	Category	N	Percentage of total
Gender	Men	155	100%
Age	≤ 20	7	4.5%
	21–30	42	27.1%
	31–40	42	27.1%
	41–50 51–60 ≥ 60	49 12 3	31,6% 7.7% 1.9%
Marital status	Single	41	26.5%
	Married	50	32.3%
	Cohabiting In relationship, living apart Divorced Unknown	25 28 9 2	16.1% 18.1% 5.8% 1.3%

Group Therapists

A total of 18 group therapists participated in the current study. Each of the 25 groups were led by a co-therapist team; there were 13 different co-therapist teams. All therapists completed a two-day training program to learn the NLIA protocol (see Table 2). In the final sample, data from nine of the 13 co-therapist teams were analyzed through random selection. This sample of nine co-therapist teams comprises six licensed mental health professionals (12 unlicensed), two with over 15 years of experience, three with six to 10 years of experience, and one with less than five years of experience. Moreover the sample comprises six social workers, five with over 15 years of experience, one with six to 10 years of experience, and a nurse with over 15 years of experience. The first author, who is a licensed mental health psychologist in the Netherlands for over 15 years, led 10 of the 60 selected sessions with three different co-therapists, while the second author, who is also a licensed mental health psychologist for over 15 years, led nine of the 60 selected sessions with two different co-therapists.

Table 2. Therapist Gender, Years of Experience, and Discipline (N = 18)

	Category	N	Percentage of total
Gender	Women	10	55.6%
	Men	8	44.4%
Years of experience	0–5 years	2	11.1%
	6–10 years	5	27.8%
	11–15 years	0	0%
	> 15 years	11	61.1%
Discipline	Psychologist	7	38.9%
	Social worker	9	50%
	Nurse	2	11.1%

Procedure

NLIA is a manualized CBT approach to group work with those who have histories of perpetrating IPV. Its goals are to help members with the identification and recognition of triggering stimuli, monitoring of emotions, changing dysfunctional cognitions, response prevention, stress reduction skills, relapse prevention, social skills and problem-solving skills (Karakurt et al., 2019; Murphy & Eckhardt, 2005). Treatment consists of 18 two-hour sessions, once a week. Participants use a workbook in which theory, techniques, and homework assignments are recorded during and in between sessions. A therapist treatment manual covers the different CBT techniques, and therapists are also instructed to intervene on participant and group interactions in sessions. One chapter in the manual describes general group therapy processes and emphasis on the group process in GCBT referring to Bieling et al. (2009). To meet the emotional needs of participants, group therapists are instructed to actively invite participants to express underlying emotions and experiences (e.g., separation anxiety, humiliation, etc.). The therapists and participants acknowledge painful underlying emotions without justifying the resulting violent behavior, as well as discuss how to handle these emotions constructively (Andrews et al., 2011; Dutton, 2006). In addition, for every treatment session, participant interactions that could occur related to the specific content of the session are described and therapists are offered suggestions for how to respond to these interactions.

We collected 343 audiotaped sessions across 387 sessions from 25 different groups. A random selection of audio recordings was made to prevent selection bias (e.g., choosing only our own sessions; choosing sessions from a small number of groups, etc.). We selected a final sample of 60 sessions as we determined this number of sessions would provide enough heterogeneity of interventions across therapists, groups and time points in the treatments, while being of a manageable size that we could reasonably reach data saturation. To increase the trustworthiness of the data and limit bias (Levitt et al., 2017), the qualitative data analysis was performed by an investigator (the third author) who is a licensed mental health psychologist who was trained and supervised in qualitative research methods in post-master's education. She was familiar with the treatment program, but not participating as a therapist. In addition, an experienced group psychotherapist unfamiliar with the treatment program coded three randomly assigned sessions independent of the third author and audited the final list of intervention types.

Each of the 60 group sessions were transcribed verbatim by the third author by listening to the sessions’ audio recordings and noting what was said and registering the moments of silence (McLellan et al., 2003). The general inductive approach (GIA; Thomas, 2006) was used to analyze the data. In the GIA, textual data from sessions or interviews is closely read to become familiar with the content. The meaningful units are labeled and described. In our research, verbatim transcripts of the entire audiotaped sessions were read so no details were missed, and every identified therapist intervention was coded. The investigator coded 15 transcripts included in the research (it turned out not all 60 sessions needed to be coded) generating a list of therapist interventions. Three of the 15 transcripts were read and coded independently by an experienced group psychotherapist naïve to the protocol, in order to determine interrater reliability. Once this step of coding was complete, in accordance with the GIA, the investigator grouped the codes into categories that could be defined, and examples were extracted from the data to illustrate and further refine these categories. Finally, these codes and corresponding categories and examples were entered into a user-friendly table, and the table was sent to the same group therapist who coded previously transcripts for auditing. At this stage the auditor was asked to give feedback on the description of the different intervention types in the coding list. The feedback was processed and the descriptions were refined accordingly.

RESULTS

In the first three transcripts, 226 therapeutic interventions were identified. We considered therapeutic interventions as both manualized treatment techniques and interpersonal interventions aimed to create an effective therapeutic environment. Not every fragment in which a therapist is talking is considered a therapeutic intervention, for example, when therapists provide practical information such as the opening hours of the mental health center. After interrater reliability was determined ($\kappa$ = .91), the few differences of opinion were resolved by reading these text segments together and deciding which label was appropriate by taking the patient context, i.e., the patient information and reaction in the discussion, into account. An example of a disagreement was a text fragment in which a therapist talked to a patient about stress reduction. This could be coded as a behavioral technique, but also as psychoeducation. By taking the context into account, it could be determined the patient was sharing their experiences of the prior week and the therapist was not referring to the participant’s workbook, so the therapist intervention was coded as psychoeducation instead of behavioral technique. During analysis, categories were often combined and relabeled. After coding 15 of the 60 therapy sessions, it became clear that no new interventions had been identified since the eleventh session. The four sessions in which no new techniques were coded assured data saturation had been reached (Guest et al., 2006). Fifteen different types of interventions could be distinguished. Among the final list of intervention types, a further categorization was made between interventions we classified as “CBT” and “process” interventions. Our definition of a “CBT” intervention was: any intervention described in the treatment manual or that is used to influence cognitions, modify behavior or improve motivation. Our definition of “process” intervention was: any intervention not described in the treatment manual, but that is part of the therapeutic process (Bieling et al., 2009; Yalom & Leszcz, 2020). The scoring list is displayed in Table 3.

Table 3. Therapist Interventions in Group CBT for IPV Perpetrators: Not “Losing It” Anymore

	Intervention type	Description and purpose of intervention	Verbatim examples
CBT-based interventions	Psychoeducation and advice	Explaining the problems in the broadest general principles and making recommendations regarding behavior. Self-disclosure may form part of this process. Purpose: Increase awareness of choice points and better understanding of the problem.	“What we would definitely like to see … before you’re in the escalation phase, is that you start looking for alternatives. That you start thinking: how can I prevent myself from snapping? Even if you feel like, it’s going very quick, that escalation, you’re about to snap, you still have choices.” “You often see that when you’re really angry, you can’t think clearly anymore. That’s something people often say afterwards: you make decisions and later you think: yeah, right, that wasn’t so smart.”
	Enhancing motivation	Discussing participant’s motivation and the consequences of certain behaviors. Purpose: Enhance motivation and readiness to change.	“Going back to the bigger picture: what do you think, what is your problem that we need to work on?” “So you’d really like to learn some more ways to handle it better.”
	Self-reflection	Encouraging the participants to reflect by means of concretizing questions. Purpose: Improve self-reflection and self-awareness.	“I get the impression that you’re trying to take back control, right, about that sense of powerlessness you have, by launching a counter-attack.”
	Behavioral techniques	Conducting manualized behavioral exercises. Purpose: Enhance response prevention, alternative behavior and stress reduction.	“I would like to introduce the time-out procedure.” “What kind of actions can you apply when stress is rising?’
	Role-playing	Role-playing based on challenging day-to-day situations experienced by participants. Purpose: Improve social skills.	“But the situation you just discussed, I think it would be good to practice with that, because I notice that you’re quick to revert to your old pattern. Let’s replay the situation, who would you like to join in as your boss in this situation?”
	Cognitive therapy	Identifying and evaluating maladaptive beliefs, using an ABC model or applying the Socratic dialogue. Purpose: Reduce dysfunctional cognitions and attune negative emotions.	“Can you actually know what someone is thinking?” “Can we investigate what cognitions are preceding your anger?”
	Homework assignments	Give and explain homework assignments.	“For next week, we would like you to read chapter 5 and fill out exercises 5.1 to 5.3”
Process-based interventions	Structuring	Structuring by keeping an eye on the time and task, providing information about the structure of the session, taking breaks, discussing group rules, letting people take turns and holding participants accountable for their behavior in the session, indicating boundaries. Purpose: Keep the structure of the sessions to ensure that participants work on the task, are involved as much as possible and prevent or stop undesirable behavior.	“Shall we continue where we left off?” “Let’s agree to this: whatever we discuss here stays right in this room … ” “And F … also said: if I’m no longer doing well I’m just going to walk out. And we would prefer you to do the same here. So if you think: If I stay in the group I’m going to lash out and attack someone that instead you say: I’m leaving and I’m not coming back but I’ll ring you later. Because I’d like to stay in touch with you afterwards.”
	Answering questions	Responding to questions from the group and providing information about matters not related to the treatment protocol. Purpose: Inform participants, put them at ease and model an open approach.	“Because of our holiday leave, the group will be interrupted the first two weeks of August.” “It’s not necessary to email us when you’ve filled the online questionnaires, we will receive a notification you did.” “I’m not divorced, no.”
	Involving the group	Promoting interaction by asking participants for recognition, noticing agreements, involving the group in the discussion and the decision-making process. Purpose: Enhance interpersonal cohesion, interpersonal learning through interaction with and observation of others, and adjust individual perceptions.	“Is that familiar to others?” “How else do you think John can benefit from being in this group?”
	Identifying emotions	Noticing what someone feels—consciously or subconsciously—in the here-and-now, including nonverbal signals. Acknowledging someone’s perceptions. Purpose: Increase awareness and contain personal feelings and displays of emotion.	“Yes, that’s pretty annoying. Absolutely!” “You feel it in your body and I can see it happening, too, when you talk about it, you can see the tension is rising. You’re getting a bit red in the face. And you, like, you kind of puff yourself up.”
	Rewarding	Showing interest, enthusiasm and empathy when participants report positive behavior in the context of the treatment objectives, giving recognition, giving compliments, thanking someone for his input, wishing someone positive things, talking about the group atmosphere, putting someone at ease, social small talk. Purpose: Reinforce positive behavior and motivate individuals and the group-as-a-whole to increase positive behavior, provide support, increase self-confidence and regulate tension.	“I think you did a really good job, because somewhere along the line you decided: I’m not going to be aggressive anymore.” “It seems quite logical to me that that bothers you. If you get that kind of message and you don’t know what to expect. Not surprising, really. I’d find it pretty strange if you could just shake that off, that would definitely be strange.” “I’m glad you were willing to share that here. Best of luck, okay? It’s a good thing you told us this. I’m happy you’re still here.”
	Humor	Using humor, self-deprecation or laughing about a joke the patient has made. Purpose: tension regulation, being able to present difficult interventions and promoting cohesion in a lighthearted manner.	“But you can’t control your son, right? It’s not like you have a ‘kid remote control.’ That would be easier!”
	Co-therapist interaction	A group therapist addresses the co-therapist or explains his/her behavior, reaction or absence. Purpose: Consult about the structure, contact the co-therapist or model collaboration.	“It takes a bit of getting used to if [my co-therapist] is not there.” “Are you working that Monday?” “Shall we … take a break?”
	Listening	Giving the patient room to talk, listening carefully and making it clear that you are paying attention to what he is saying.	“ … … ” “Hm. Aha.” “Yes … Yes … ”

DISCUSSION

The present study aimed to investigate the therapist interventions applied in GCBT with individuals who have perpetrated IPV. The inductive nature of this study helped us to explore what kinds of interventions group therapists use during therapy sessions. We distinguished 15 types of therapist interventions.

Our observation that group therapists used both CBT techniques as well as what has been referred to as the “common factors” or “nonspecific factors” in psychotherapy, is in line with a vast body of research demonstrating that all effective psychotherapy practice combines therapist application of specific (to the treatment approach) and nonspecific (common to all effective psychotherapy) factors (Cuijpers et al., 2019). Several authors contend that therapists who build a positive working alliance, make strategic use of interventions, and manage group dynamics will achieve better therapy outcomes than therapists who are less skilled (Leszcz, 2018; Norcross & Wampold, 2011). Other authors argue that the application of a treatment protocol with specific techniques will improve treatment effectiveness by providing a coherent approach that generates consistency, confidence and hope (Bieling et al., 2009; Gallagher et al., 2020). Probably a combination of manual-based interventions and strong therapist skills create a positive working alliance, which enables change (Burlingame et al., 2013; McAleavey & Castonguay, 2015; Wampold et al., 2017).

We identified as much as eight “process-oriented” interventions. Initially, CBT group treatment focused on techniques aimed at symptom reduction with an emphasis on psychoeducation, CBT techniques and homework assignments, making GCBT protocols quite similar to individual CBT protocols. As a result, there was limited attention to the group process and process-oriented group therapy interventions (Hofmann & Barlow, 2014; Nezu & Nezu, 2015). According to a more contemporary point of view, GCBT should preferably be a blend of CBT techniques and group supportive elements (Dozois et al., 2019; Söchting, 2014). Furthermore, group CBT clearly differs from individual CBT therapy because evidently group dynamic processes will occur (Kleinberg, 2011; Whitfield & Scott, 2018) and therapy outcome is affected by the group process (Burlingame et al., 2013; Söchting, 2014).

Our findings lend additional support to the notion that process-oriented interventions in group CBT are highly important. Contemporary approaches to GCBT outline that in order to enhance treatment effectiveness the merging of both manualized treatment techniques and common group therapy interventions is crucial (Hahn et al., 2022; Paquin, 2017; Söchting, 2014; Tasca, 2014). They also state that group therapist skills are defined by a conjunction of knowledge, experience, personality characteristics, and behavior. With our findings we were able to distinguish and describe the different process-oriented interventions as well as the manual-based interventions that jointly cover how NLIA was delivered in our sample.

The use of process-oriented or “extra-protocol” group therapeutic techniques next to manualized interventions shows the NLIA therapists deliver the treatment as is recommended in studies regarding (G)CBT with IPV perpetrators (Butters et al., 2021; Hamel et al., 2021). IPV perpetrators experience complex psychopathology; besides aggression regulation deficiencies, personality disorders, PTSD, depression, anxiety disorders, and substance use problems are often comorbid (Crane & Easton, 2017; Elklit et al., 2018; Semiatin et al., 2017; Siria et al., 2021). These underlying psychological problems presumably derive from early childhood trauma and insecure attachment relationships (Barbaro & Shackelford, 2019; LaMotte & Murphy, 2017). Several authors suggest that childhood trauma and insecure attachment require specific attention in treatment. The use of process-oriented interventions next to CBT techniques may be essential from an attachment perspective (Marshall & Burton, 2010; Tasca, 2014) and enhances treatment effectiveness in manualized group therapies (Paquin, 2017). It has been suggested that in IPV perpetrator treatment, therapists should provide a secure base environment by structuring therapy sessions, collaborating with participants and showing empathy (Bowen & Gilchrist, 2004; Day et al., 2009; Marshall & Burton, 2010). In a secure and supportive group climate, the group itself can be used to practice more adaptive interpersonal responses and thereby provide participants with positive and corrective feedback to their behavior (Lawson & Brossart, 2009; Velotti et al., 2018). In addition, paying close attention to group dynamics in IPV perpetrator group treatment may be an important factor in reducing treatment attrition (Edleson, 2012; Hamel et al., 2021).

Limitations and Strengths of Present Study

An important limitation of our study is that nonverbal behaviors, such as facial expressions, gestures and body language could not be considered as we relied on audio recordings of sessions. It is unclear how this could have affected our findings as there is some disagreement as to what extent therapists’ verbal and nonverbal behaviors are in alignment (Del Giacco et al., 2019; Wachtel, 2014; Westland, 2015).

In addition, we did not take the stage of the group treatment into account when selecting our sample. Specifically, time of group session was not a variable we considered (our final sample included six sessions from the first third of treatment, six from the middle third, and three from the last third), therefore our results are mostly based on the beginning and middle part of the treatment and we did not learn as much about the last third section of the treatment. Thus, we cannot fill the gap in the literature linking therapist interventions with stages of group development (Kaklauskas & Greene, 2020; Markin & Marmarosh, 2010; Yalom & Leszcz, 2020). Bieling et al. (2009) describe the different stages of group development in GCBT and provide examples of interventions to be applied during the various stages, to optimize group processes. Although we did not take group stages into account, the different therapist interventions we distinguished and describe in our list correspond with the interventions described by Bieling et al. (2009). Therefore we think that it is unlikely that the interventions in the third part are substantially different than in the first parts, but it can of course not be ruled out.

A significant strength of our study is that the list of interventions was derived using audiotaped transcripts from different co-therapist teams at four different mental health centers. Thus, the coding manual is likely generalizable to GCBT across settings and therapists (at least in the Netherlands and/or in other culturally similar regions).

Finally, the interventions we describe in our manual are commonly used CBT techniques and process-oriented interventions. While our list of therapist interventions was derived using data from our group work with participants who have perpetrated IPV, our coding manual might be useful for coding GCBT interventions with other populations, but this is obviously an empirical question.

Recommendations for Future Research

The purpose of this research was to develop a coding manual to better understand the different types of therapist interventions made in the course of group therapists delivering this treatment, in actual practice. Connecting these therapist interventions with patient outcomes in this patient population is the next important step for research in this area. However, future research must first examine the reliability of our coding manual. In addition, the coding manual should also be tested for reliability when group session data come directly from different sources, including listening to audio recordings and/or watching video recordings when possible rather than reading verbatim transcriptions of sessions, so that subtle nonverbal details such as tone of voice can be taken into account. Subsequent research may then address the question of whether therapist interventions affect treatment adherence and effects. Furthermore, questions concerning the relationship between group development phase and therapist interventions could be investigated, and their relationship to therapy outcomes.

Acknowledgments

We would like to thank Annelies Claeys, MSc, for providing us with her independent feedback during the coding process, which was an essential contribution to our research.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the authors.

Additional information

Notes on contributors

Carola A. van Tilburg

Carola A. van Tilburg is a psychotherapist at GGZ Westelijk Noord Brabant and a Ph.D. student in the Department of Psychology and Neuroscience of Maastricht University. Arno van Dam is a clinical psychologist at GGZ Westelijk Noord Brabant and a Professor of Antisocial Behavior, Psychiatry and Society at Tranzo, Tilburg University. Edith de Wolf-Jacobs is a clinical psychologist at GGZ Westelijk Noord Brabant. Corine de Ruiter is a Professor of Forensic Psychology at the Department of Psychology and Neurosciences of Maastricht University. Tom Smeets is a Professor of Clinical Psychology at the Tilburg School of Social and Behavioral Sciences of Tilburg University.