Introduction

Transference-focused psychotherapy (TFP) is a modified psychodynamic psychotherapy designed to be used for patients suffering from severe personality disorders, most prototypically borderline and narcissistic personality disorders [1]. Otto Kernberg, based on his experiences with the Menninger psychotherapy research project, began modifying standard psychodynamic psychotherapy. Initially, he referred to this therapy as “exploratory psychotherapy,” in an effort to distinguish it from more supportive psychotherapies. These modifications were based on Kernberg’s articulation of the developmental psychopathology underlying severe personality disorders and the clinical realities of treating those with severe personality disorders. Over the subsequent decades, Kernberg and colleagues, particularly, Frank Yeomans and John Clarkin, at the Personality Disorders Institute of Cornell University, further articulated and developed the treatment in a series of treatment manuals [2]. Support for TFP’s efficacy has been shown in several studies for both symptom and personality change. A number of recent meta-analyses and Cochrane Collaboration reviews have also concluded that TFP is efficacious and has found no differences in the effect sizes between TFP and other empirically supported treatments [3••]. Given these findings, a number of prominent treatment guidelines, including the Society of Clinical Psychology Committee on Science and Practice, the United Kingdom’s National Institute for Health and Care Excellence (NICE) guidelines, the German Society for Psychiatry Treatment Guidelines for Personality Disorders, Australia’s NHMRC Clinical Practice Guidelines, the Swiss Association for Psychiatry and Psychotherapy, and the Netherlands’ Multidisciplinary Directive for Personality Disorders recognize TFP as an empirically supported treatment. The general consensus considers TFP among the “big four” specialized therapies, along with dialectical behavioral therapy, mentalization based therapy, and schema-focused psychotherapy, for treating borderline personality disorder (BPD). In this article, we describe the theory underlying TFP, the goals, structure, and techniques of the treatment, as well as its indications. We additionally summarize the existing body of research regarding symptom and personality change in TFP. Lastly, we discuss recent developments in and applications for TFP.

Goals of the Treatment

The overarching goals of TFP are to improve self-control, reduce impulsivity, increase emotion regulation abilities, increase intimacy in relationships and relationship satisfaction, and improved capacity to realize life goals (that are consistent with the patient’s abilities and desires). More specific goals include improvements in the symptoms central to BPD, especially suicidal and parasuicidal behaviors, angry outbursts, and impulsive behavioral difficulties. Improvements in these areas are hypothesized to lead to reduction of emergency service use, hospitalizations, and difficulties in relationship. These changes are posited to follow from the integration of disparate, contradictory, and incoherent internal mental representations of self and others.

Fundamental to the TFP model is that BPD derives from a failure to develop internal representations of self and others that are complex and realistic and characteristic of healthy psychological maturation. These fragmented representations of self and others impede the person’s capacity to reflect on interactions with others as well as one’s own beliefs and to behave in a thoughtful and consistent goal-directed manner. Additionally, this lack of integration leads to fluctuations between extreme positive or negative emotions that impairs an individual’s perception of day-to-day interactions (“black and white thinking”) [1]. The inconsistent sense of self and others is called “identity diffusion” in the TFP model and is analogous to identity disturbance defined in DSM-5 as well as psychological processes regarding identity formation described by Blatt [4], Erikson [5], Marcia [6], and McAdams [7]. In the TFP model, identity diffusion is considered the source for emotion dysregulation seen in BPD. Thus, the treatment focuses on the integration of one’s sense of self and others and the emotions linking them. This integration is hypothesized to lead to representational and affective experiences becoming more nuanced, enriched, and modulated. The increased differentiation and integration of these internal representations result in the patient developing the capacity to think more flexibly and positively about the therapist, significant others, and themselves. The integration of these internal representations is achieved by exploring and understanding the patient’s contradictory experiences of self and others, but particularly of the therapist [1].

Indications

TFP is indicated for the outpatient treatment of severe personality disorders including borderline, histrionic, narcissistic, and antisocial personality disorders as described in section II of DSM-5. From the perspective of the DSM-5 Alternative Model of Personality Function (AMPF), TFP is indicated for severities above 0 (little or no impairment) from level 1 (some impairment) to level 4 (extreme impairment) with borderline, narcissistic, antisocial, schizotypal, and avoidant personality disorder types. When using the ICD-11, TFP is suitable for those with moderate or severe personality disorders, particularly when the borderline pattern qualifier is specified. Additionally, when using the Psychodynamic Diagnostic Manual Version 2 (PDM-2), TFP is suitable for those in the borderline level of personality organization and with personality syndromes of borderline, narcissistic, psychopathic, and histrionic. Irrespective of nomenclature, these disorders are considered to fall within a borderline organization of personality structure that is characterized by identity disturbance, intact reality testing that can become impaired with severe levels of stress, and an inclination to defend against intense emotion with “black and white” thinking and feeling (“splitting”) [1]. An adapted form of TFP is more suitable for individuals with low-level severity disorders such as obsessive-compulsive (sections II and III, DSM-5; PDM-2), dependent (section II, PDM-2), or avoidant personality disorders (section II, PDM-2) [8, 9•].

Theory Underlying TFP

In TFP, BPD is conceptualized as a disorder of impaired and distorted psychological or representational structures that can be characterized as relatively undifferentiated and unintegrated. As such, those with BPD have difficulty evoking internal representations of one’s self and others that are complex, integrated, and realistic. Such representations result in impaired psychological functioning, particularly under times of stress or during ambiguous interpersonal situations. Kernberg [10] referred to this undifferentiated and unintegrated representational states as “identity diffusion”, which is close to the DSM concept of identity disturbance. Vacillations in split and polarized (exclusively negative or exclusively positive) representations are hypothesized to drive vacillations between extreme negative and positive emotions, both of which impair an individual’s interpretation of daily situations.

TFP attempts to integrate these representations by observing, tolerating, exploring, and working through the patient’s contradictory experiences of others, and especially the patient’s experience of the therapist in session [1]. Such a process is assumed to promote increased awareness of and insight into the fragmentation, and, in turn, an eventual reduction in identity diffusion. This is reflected in representations of self and others that are differentiated and integrated which enables patients to have views of themselves and significant others that are more flexible, positive, and realistic. As a result, patients’ emotional experiences grow increasingly rich and regulated, and relationships become more satisfying and less conflictual. In addition, reductions in symptoms and self-destruction tendencies as well as increased functioning promote greater capacity for independence and intimacy.

Structure and Techniques of TFP Treatment

TFP is designed to be a twice-weekly, individual, face-to-face, outpatient psychotherapy. Like most therapies for personality disorders, TFP is conceptualized as a long-term treatment typically lasting 12–18 months at minimum. TFP should be delivered by an experienced therapist that is trained and certified by the International Society for TFP (ISTFP) or is training and under supervision by an approved ISTFP supervisor.

Structure, Strategies, and Techniques

TFP occurs in three treatment phases: assessment, the establishment of the treatment frame, and the active treatment phase.

Assessment

Before the start of therapy proper, TFP begins with a thorough assessment of diagnostic issues, the patient’s difficulties, and life-structure. This is typically carried out over one to three sessions. This assessment is used to establish a diagnostic understanding and to provide the patient with an understanding of his or her difficulties that includes his or her perception of self and others and to collaboratively set the treatment goals, frame, and explicate the roles and responsibilities of both the patient and therapist in the treatment.

Establishment of a treatment frame

In TFP, this verbal agreement is often referred to as the treatment contract or treatment frame; it establishes the conditions or frame of the therapy in a way that emphasizes the experience of emotions within the therapy and curbs the expression of emotions in the form of acting out (cutting, taking overdoses, substance abuse, unsafe sex, etc.). This contract or treatment frame, in the case of inactive or socially isolated individuals, also establishes the therapeutic value of the patient becoming involved in some form of structured activity that involves other people (bringing in issues of work, responsibility, and interpersonal relations). Collaboratively, the therapist and patient establish specific treatment goals that are as attainable as possible and capture issues of work, responsibility, relationships, and hobbies. A treatment frame is established that articulates roles and responsibilities within the treatment [1]. The patient’s responsibilities typically include regular attendance to session, working towards the treatment goals, reduction of self-destructive and impulsive behaviors, bringing thoughts and feelings into the therapy room without censorship, and reflecting on such thoughts and feelings, along with the therapist’s remarks.

The Active Phase of Treatment

Once the assessment is complete, feedback is provided to the patient, and the frame is collaboratively set and agreed upon, the therapy can begin in earnest. As part of the treatment frame, the patient takes responsibility for their treatment by beginning session, with a focus on those issues that are most relevant to the difficulties they are experiencing. As the patient speaks, a hierarchy of priorities guides the therapist in choosing themes to focus on (a) attending to threats to the patient or therapist safety or of the treatment (such as suicidality, homicidality, and treatment interfering behaviors); (b) understanding the inner experience of the patient as it is reflected in relation to the therapist (“transference”); (c) addressing other emotionally charged themes in the patient’s life outside of therapy.

In sessions, the therapist focuses on the main patterns of relationships as they are manifested in relation to the therapist. The therapist monitors the patient’s emotion and helps articulate their experience, especially disparate facets of their experience of which they may not be conscious. Within a single session, or spanning over several sessions, a typical sequence of interventions may involve the therapist first helping the patient clarify their subjective experience [11]. This technique is called clarifying or making clarifications. The therapist then gently brings into the patient’s awareness discrepancies between what the patient is saying, doing, or communicating non-verbally, and encourages the patient to reflect on those discrepancies. In TFP and many psychodynamic therapies, this is called a confrontation. Obviously, this is not an “in your face” confrontation. It is called a confrontation because the bringing into awareness of disparate information, confronts the patient’s defensiveness or splitting, and allows for exploration and integration. As mentioned, it should be done tactfully so as to increase the likelihood that it is taken in by the patient. Then, the therapist utilizes interpretations of the patterns or themes and their manifestations in the relationship with the therapist [1]. It is important to provide interpretations in a timely, clear, and tactful way and address the patient’s hypothesized causes for maintaining those inconsistencies and split representations of self and others. Such causes may include, for instance, a desire to protect positive aspects of self and other by keeping them apart from negative ones. The goal of interpretations is to facilitate awareness of and tolerance for representations of self and other and emotional states that were previously experienced as inacceptable and kept outside of awareness. Although interpretations should be delivered in a timely manner characterized by immediacy, interpretations take time to build and should not be experienced as abrupt or an attack. Additionally, although interpretations should be delivered with a certain confidence and without unnecessary trepidation, they are done so as a hypothesis with an openness to feedback and correction. Such a process promotes improved reflective ability, richer and more positive perception of self and others, and improvement in intimate relationships. In this way, interpretations are not so much an isolated intervention, but part of an interpretative process that includes clarifying and confrontation. Only after the material is close to the patient’s awareness, is the interpretation delivered.

Empirical Support

Evidence for the Underlying Theoretical Assumptions of TFP

From a TFP perspective, identity diffusion is the central mechanism that underlies BPD. The traits and symptoms typical to BPD, including suicidality and emotion dysregulation are hypothesized to stem from and rooted in identity diffusion. There is also a burgeoning literature on the importance of identity in the difficulties experienced by those with BPD.

In a series of studies, Levy and colleagues have focused on the role of identity and views of self and other [12, 13, 14••]. Specifically, Levy et al. [13] found that vacillations in mental states about the self and others (including the therapist) predicted the quality of the therapeutic alliance as rated by observers. Beeney and colleagues [12] found that those with BPD showed abnormalities in the neural circuitry central to self and other processing. Those with BPD had greater activation in the precuneus and posterior cingulate for both self-reflection at baseline and other-reflection at baseline contrasts, while the control group evidenced greater activation in the self-other contrast in the angular gyrus (all results p < .005, k = 24, equivalent to FWE < .05, corrected). Behaviorally, BPD patients showed more fragmented, unintegrated, fluctuating, and negative self-concept on a self-concept card sorting task completed twice over a 3-h period.

In another study, those with BPD were compared with anxiety-disordered patients in a 21-day intensive repeated measurement design collecting both random and event contingent data via smartphones [14••]. On average, patients provided data 12 times per day. We found, as many might predict, that affect regulation deficits in terms of the experience of negative affect predicted suicidal urges. However, this association was only found when patients were in identity diffuse mental states. Although BPD patients scored significantly higher and experienced more identity disturbance, negative affect, and suicidal urges than those with anxiety disorders, consistent with an RDoC approach, the process worked similarly across both groups.

In other work from our lab, we used the reverse correlation method to examine the capacity of those with BPD traits to form accurate visual self-images. Participants were tasked with choosing, in a series of 300 presentations, between two faces as to which one looked more similar to themselves. Two images were generated using a reverse correlation method—one based on all the images selected as more similar to the participant and the other, the rejected image. Using a computer, the self-relevant generated image was then compared with the participants’ actual image, and a similarity score was generated. BPD traits were negatively related to the accuracy of the generated self-image. The greater the number of BPD traits and symptoms, the more diffuse the self-image (b = − .29, p < .02). In contrast, although narcissistic traits as measured by the Pathological Narcissism Inventory [15] also predicted less accurate self-images, generated images tended to be rated by independent assessors as better looking than the participants’ actual photograph. Thus, BPD traits and symptoms were related to the creation of more diffuse self-images, and narcissistic traits and symptoms were related to more enhanced self-images.

Treatment Effectiveness and Efficacy

TFP’s effectiveness and efficacy has been demonstrated in multiple pre-post studies [16,17,18,19, 20•], a quasi-experimental study [21], and three independently carried out randomized controlled trials across four countries (RCTs) [22,23,24].

In the first RCT [22], Clarkin and colleagues compared TFP with dialectical behavioral therapy (DBT) and supportive psychodynamic therapy (SPT). Although all three of the treatments showed improvements in depression, anxiety, functioning, and adjustment, TFP showed improvements in a larger number of symptom domains, particularly anger and aggression. Doering and colleagues [23] compared TFP with treatment in the community by therapists experienced in treating BPD. They found that TFP led to greater decreases in suicide attempts and psychiatric admissions and also had a lower dropout rate. Additionally, TFP resulted in greater improvements in borderline symptoms, psychosocial functioning, and personality organization. Patients treated in either group showed similar improvements in depression and anxiety.

Arntz and colleagues’ RCT [24] compared schema-focused therapy (SFPT) with TFP conceptualized as a control group. Both treatments were effective with large effect sizes. Although SFPT showed better results in the intent-to-treat analyses, the completer analyses showed no differences between the treatments [24]. This discrepancy between SFPT and TFP was due to the difference in non-random dropout. Additionally, there are indicators that the randomization had failed. For instance, patients in the TFP group scored higher on measures of self-destructiveness than patients in the SFPT group and where thus more severely disturbed in a manner that has been shown to be related to worse outcome.

Meta-analytic Studies

There have now been a number of meta-analytic reviews that have examined TFP in relation to other treatments [3, 25,26,27,28]. These studies have consistently found that there are no reliable differences in overall effect sizes between TFP treatments and other treatments, including DBT. This finding is not surprising given the effect sizes found in individual studies and those with direct comparisons of DBT with TFP [22] and other treatments.

Changes in Personality

Although symptom change is important, especially for disorders like BPD, where the symptoms are so serious, TFP aspires for more ambitious and deeper change, namely personality change. Therefore, psychotherapy research in TFP has focused on assessing changes in deep personality structures. Change in personality is especially important for disorders, such as BPD, where the fundamental difficulties are hypothesized as being rooted in personality proper. This notwithstanding, the assessment of personality change has been largely neglected as an outcome in BPD research. However, personality change has consistently been examined in RCTs of TFP, and this is one area of assessment that TFP has reliably demonstrated specific benefits in. Significant increases in attachment security and mentalizing (the ability to make sense of one’s and others’ mental states) were shown in two of the RCTs and in a pre-post study [29,30,31,32]. These improvements occurred only in the TFP conditions and not in control groups. Levy and colleagues [(31] found that in the TFP condition, of the patients who were classified as insecurely attached when entering treatment, almost a third (29%) were classified as securely attached by the end of treatment. By comparison, none of the patients in the control treatments achieved secure attachment status. Similarly, Buchheim and colleagues [29•] found that 12 of the 38 (31%) patients in TFP showed improved scores in terms of attachment security, whereas none of the comparison group participants did. Levy and colleagues [32] also found that 30% of the patients treated in TFP were securely attached after 1 year of treatment, representing a significant change from baseline.

With regard to mentalizing, the pre-post effect size improvements in all three studies [30,31,32] were similar, despite patients and therapist being drawn from three different countries, three different cohort times, involving different therapists, interviewers, and coders. These findings suggest improved relationships and insight into both one’s own motivations and inferences about other people’s intentions and behaviors, as well as increased capacity for satisfying interpersonal relationships. These findings are important because such changes in personality are posited to be related to sustained changes in improved coping with difficulties with internal conflict, relationships, and major life challenges in love and work.

Changes in Brain Functioning

We conducted a pilot study to investigate neural activation associated with and predictive of clinical improvement in emotional and behavioral regulation in BPD following TFP. Using a within-subject design, we have examined pre-post treatment neural changes using fMRI scans in ten women reliably diagnosed with BPD who were treated in 1 year of TFP by trained, supervised, and adherent therapists. During the scans, a BPD-specific emotional-linguistic go/no-go task was used to assess the relation between negative emotional processing and inhibitory control. We found fMRI measured brain changes that were associated with variation in outcome. These findings demonstrate symptom change in TFP is related to relevant brain changes. Specifically, we found relative increased dorsal prefrontal (dorsal anterior cingulate, dorsolateral prefrontal, and frontopolar cortices) activation and relative decreased ventrolateral prefrontal cortex and hippocampal activations following treatment. Clinical improvement in constraint correlated positively with relative increased left anterior-dorsal anterior cingulate cortex activation. Clinical improvement in affective lability correlated positively with left posterior-medial orbitofrontal cortex/ventral striatum activation, and negatively with right amygdala parahippocampal activation.

Recent Clinical and Theoretical Developments and Advances

Kernberg first began to articulate a treatment approach for working with patients diagnosed with BPD back in the mid-1960’s based on his experiences with the Menninger psychotherapy project [33]. Since that time, he and his colleagues have continued to explicate and elaborate the theory and techniques of TFP. The first manual for TFP was published in 1989 [34], followed by updates in 1999 [35], 2006 [36], and 2015 [2]. Along the way, important supplements were published on contracting [37], dealing with difficult patients and clinical situations [38] and a primer [39].

Although TFP the manuals were first developed specifically for treating BPD, Kernberg always saw the principles articulated in TFP as having broader relevance to a range of clinical disorders associated with borderline personality organization (BPO). BPO is a concept that predates and was intended to be broader than the narrower definition of BPD later adopted by the DSM-III and subsequent versions. As such, it includes BPD but also a number of other personality styles or disorders, most notably narcissistic personality disorder. Relatedly, Kernberg has a long history of writing about narcissism, and there is some preliminary evidence, described earlier, that TFP is uniquely efficacious when compared with DBT and a supportive psychotherapy for narcissistic patients [40]. Nevertheless, in recent years, Kernberg and a number of clinical researchers from our group have written about technical modifications to TFP when treating those with significant narcissistic pathology to accommodate differences in the pathology between these disorders [41,42,43,44,45,46,47].

Because TFP was originally conceptualized for a range of personality difficulties and in the TFP model, the underlying developmental psychopathology rooted in impairments in self and other representations that transverse personality disorder categories, a perspective further supported by recent research suggesting a large general factor of personality pathology [48, 49], principles of TFP have been extended to working with a range of personality pathology [8, 9•]. The application of TFP principles have now been described for working with college students [50], medication prescription [51], in acute clinical care settings [52, 53], with traumatized patients [54], complex depression [55], and training psychiatric residents [56]. Furthermore, consideration has been given to the integration of TFP with other treatments such as good psychiatric management [57], supportive psychotherapy [58], behavioral activation [59, 60], and modular treatments [61, 62].

Finally, based on the pioneering work of Paulina Kernberg [63], there have also been articulation and modification of TFP principles for adults as applied to child and adolescent treatments [64,65,66,67,68].

Conclusion

TFP is a psychodynamic psychotherapy that has been empirically shown to be effective for individuals diagnosed with personality pathology, particularly BPD. TFP is an ambitious treatment that aims to integrate diffuse identity, increase emotion regulation, and improve relationships. Such changes are seen as important in facilitating the patient’s capacity for developing intimate and satisfying love relationships and developing sustained interests, commitments, and capacity for investments needed for a successful work life. Therapeutic change occurs primarily through the relationship with the therapist through exploration and consolidation of the patient’s split and fragmented representations of self and others. Such exploration results in increased awareness and understanding of disparate experiences and representations. Empirical support for TFP has been shown in numerous studies thus far, with results demonstrating improvements in general symptom domains of BPD, as well as more specific personality change such as security of attachment and capacity for mentalizing.