Adapting Dialectical Behavior Therapy for the Treatment of Dissociative Identity Disorder

Dissociative identity disorder (DID) is a chronic post-traumatic condition (Dell, 2009) characterized according to the DSM-5 by “disruption of identity characterized by two or more distinct personality states”, with “marked discontinuity in sense of self. . .accompanied by related alterations in affect, behavior, consciousness, memory, perception, cognition, and/or sensory-motor functioning” (American Psychiatric Association, 2013, p. 292). Other common symptoms include amnesia, hearing voices of other personality states, and depersonalization (Dell, 2006). Dissociative identity disorder is not a rare condition, with estimates of the prevalence in inpatient and outpatient settings ranging between 2% and 11% (Dell, 2009, p. 410). However, there are no published randomized controlled trials investigating treatments for DID (Brand, Classen, McNary, & Zaveri, 2009; Sar & Ross, 2009). The goals of the current paper are to suggest an empirically informed behavioral approach to the treatment of DID and the management of its associated symptoms, in an effort to promote empirical investigations of treatments for the disorder, and to provide assistance to clinicians working with patients who have DID until the much-needed research is conducted and disseminated.

Although randomized controlled trials are absent from the DID literature, case reports and uncontrolled trials suggest that psychotherapy with DID patients results in symptom reduction and improved functioning (for reviews, see Brand et al., 2009; Brand, Loewenstein, & Spiegel, 2014). A recent naturalistic study used a 30-month follow-up design to examine symptom levels and functioning of more than 100 patients treated with psychotherapy for dissociative disorders. The results indicated decreases in dissociation, PTSD symptomatology, depressive symptomatology, and need for hospitalization, with corresponding increases in measures of positive functioning (Brand et al., 2013). Treatment guidelines developed by the International Society for the Study of Trauma and Dissociation (2011), based on empirical outcome studies and expert clinical consensus, recommend a staged treatment approach for DID. In this approach, safety issues are addressed at the outset before moving on to a phase centering on the amelioration of trauma symptoms—so-called “working through” of trauma.

The concept of a distinct first stage, where a safe and collaborative treatment frame is established and dangerous behavior is addressed, was emphasized in Judith Herman’s (1992) seminal Trauma and Recovery. Herman’s description, in turn, resembles what Linehan (1993) designates as “Stage One” of Dialectical Behavior Therapy (DBT), the treatment which Linehan created for the treatment of suicidal behavior and borderline personality disorder (BPD). The treatment approach in DBT involves four stages, with Stage One establishing safety and reducing suicidal and self-harm behaviors, followed by a phase to improve quality of life and emotional experiencing (“Stage Two”), including amelioration of trauma symptoms when present (Linehan, 1993). In multiple randomized controlled trials, DBT has been shown to be effective in the treatment of emotion dysregulation, self-harm, and suicide risk, including in patients with BPD, eating disorders, and comorbid substance use disorders (Linehan et al., 2006; Miller, 2015; Stoffers et al., 2012). We argue that the repertoire of clinical interventions in Stage One DBT, which have contributed so much to the successful treatment of patients with BPD, can be usefully adapted to the treatment of DID. Over several years, the authors have heard a number of presentations at national meetings describing the application of DBT to DID, and there is a growing literature about the use of DBT for PTSD (Becker and Zayfert 2001, Lanius and Tuhan 2003, Wagner, Rizvi and Harned 2007, Harned, Korslund and Linehan 2014) and even specifically for PTSD with dissociation (Granato, Wilks, Miga, Korslund and Linehan 2015) but to our knowledge there is no published literature which describes adapting DBT for use with DID patients.

The principal reason we and others have attempted this adaptation is the considerable overlap between BPD and DID, beginning with the prominence of self-harm and suicidality in both conditions. Of course, BPD is associated with self-harming and suicidal behavior, both by diagnostic definition (American Psychiatric Association, 2013) and by a long-standing clinical literature (e.g., McGirr, Paris, Lesage, Reynaud and Turecki 2009). Suicidal behavior is one major area of commonality between BPD and DID: Foote, Smolin, Neft, and Lipschitz (2008) found that outpatients with dissociative disorders, including DID, are 15 times more likely to have reported a history of multiple suicide attempts compared to outpatients without dissociative disorders, above and beyond the contribution of borderline personality disorder, alcohol abuse, and post-traumatic stress disorder. Multiple attempter status is a strong predictor of eventual lethal behavior (e.g., Suominen et al., 2004), underlining the urgency of identifying effective treatment for DID.

Borderline personality disorder is also highly comorbid with DID. In four cohorts of patients with BPD, the prevalence of any comorbid dissociative diagnosis ranged from 50% to 72%, with prevalence of DID between 11% and 52% (Sar et al., 2003; Sar & Ross, 2006; Foote, Smolin, Legatt, & Lipschitz, 2006; Ross, 2007); conversely, in populations of patients with DID, BPD prevalence varied between 31% and 83% (Foote, 2015). Therefore, any DBT practitioner treating a population of patients with BPD will predictably encounter a substantial minority of patients who also suffer from DID. The disorders share many clinical features, including prominent emotion dysregulation, identity disturbance, and dissociation. Additionally, highly elevated rates of childhood trauma have been found in both disorders, with likely etiological significance (Herman, Perry and van der Kolk 1989; van der Kolk, Hostetler, Herron, & Fisler, 1994; Foote 2015; Dalenberg et al., 2012). The extremely high degree of comorbidity, combined with the etiological and phenomenological overlap, have led some authors to propose that BPD and DID are not fully distinct disorders, but belong on a continuum of developmental trauma disorders (Classen, Pain, Field, & Woods, 2006; van der Kolk, 2005). At the very least, the multiple areas of overlap would seem to represent strong a priori grounds for proposing that a treatment known to be effective in BPD might be useful in DID as well.

We propose that given the considerable similarity in clinical presentation between BPD and DID, there are many ways in which the DBT model can be applied to treatment of DID without significant changes. However, the clinician will also encounter clinical features of DID that demand alterations to the DBT approach. In this paper we provide theoretical principles for adapting DBT for DID and illustrate these principles with practical clinical strategies. In order to provide concrete examples of the principles and challenges inherent to this approach, we begin by providing a case example that will be interwoven into our subsequent discussion.

Case Illustration

Patient Description

The patient, whom we will call Mariella, is a Dominican-born Hispanic female in her late thirties who is supported by public assistance and lives with her three youngest children in public housing. The patient was referred to the DBT program at our hospital after being discharged from her previous outpatient clinic due to non-compliance with treatment; there, she had been treated with medication, supportive counseling, and inpatient services. Mariella presented with a diagnosis of DID that had been made at her previous clinic. She reported a history of extensive childhood trauma, including severe physical and sexual abuse which began before the age of six and continued intermittently thereafter. She began having children during her mid-teenage years, and eventually married the father of her children while he was incarcerated.

Case Formulation

A DBT case formulation includes a delineation of problematic behaviors at the four treatment target levels to create a treatment hierarchy that is used both within and across sessions to guide therapeutic actions. Mariella presented with several life-threatening behaviors, which became Target 1: multiple nearly lethal suicide attempts, chronic suicidal ideation, and severe non-suicidal self-injury (NSSI). At the start of treatment in our clinic, Mariella reported cutting herself with knives or razors multiple times per day. Mariella also presented with several therapy-interfering behaviors (Target 2), including attending fewer than half of her scheduled individual and group sessions in a given month. Behavioral analyses (also referred to as “BAs”, or as chain analyses) (Linehan 1993) of this noncompliance yielded evidence that interpersonal skills deficits (e.g., inability to act assertively with family members) and ambivalence about reducing self-harm behaviors were at the root of her frequent missed sessions. Mariella often refused to conduct behavioral analyses of cutting behaviors (“willfulness,” in DBT terms) because she believed BAs would lead her to self-harm after the session. Activation of this belief during a therapy session was a common precipitant for the patient dissociating and switching to another personality state who did not have access to memories of NSSI behaviors, thereby short-circuiting the BA—dissociative “switching” which interfered with treatment. As this is a therapy-interfering behavior specific to DID, it is addressed in more detail below.

Mariella also presented with numerous quality-of-life-interfering behaviors (Target 3), including severe depression, substance abuse, dissociative episodes with disremembered behavior, and bingeing/purging. Mariella presented with severe deficits in all four of the DBT skill domains (Target 4): mindfulness, interpersonal effectiveness, emotion regulation, and distress tolerance. The secondary treatment target most salient for Mariella’s treatment was the dialectical dilemma of unrelenting crises versus inhibited grieving. The lack of synthesis of this dialectic frequently manifested as switching behaviors in session, as the emergence of overwhelming affects would prompt a switch, which would abruptly terminate cognitive and affective processing.

Course of Treatment

For eighteen months, Mariella was treated with weekly DBT-informed individual psychotherapy, medication management, and, initially, weekly DBT group skills training. The therapists, predoctoral psychology interns, were both supervised weekly by the lead author. The percentage of kept appointments for individual therapy ranged from 50% per month (common) to 100% (rare). The patient’s attendance at skills group was so sporadic that this modality was ultimately dropped from her treatment plan. She was hospitalized once during this course of treatment for suicidal intent and plan.

The primary focus of Mariella’s treatment was on suicidal and self-harm behaviors, as well as therapy attendance. At the start of treatment in our clinic, Mariella engaged in NSSI multiple times per day.

During each session, Mariella’s therapist assessed for the presence of suicidal and self-harm behaviors since the previous session by reviewing Mariella’s DBT diary card. Per the DBT hierarchy, when suicidal behaviors or self-harm behaviors were present (i.e., Target 1 life-threatening behaviors), behavioral analyses were conducted to determine the triggers, functions, and consequences of the behaviors. This allowed Mariella and her therapist to conduct a “solution analysis” to formulate a behavioral plan to prevent recurrence of the dangerous behavior. These steps are fundamental parts of Stage 1 DBT. However, practicing Stage 1 DBT also required Mariella’s therapists to manage challenges posed by DID using a principled approach, to which we turn in the next section.

Principles in Adapting DBT to Treatment of DID

The treatment literatures for DBT and DID both propose staged models of treatment. However, the DBT literature describes Stage 1 in more detail than the later stages of treatment, while the DID literature emphasizes Stage 2, the therapeutic working-through of traumatic experience, which is conceptualized as central to the treatment of DID (International Society for the Study of Trauma and Dissociation, 2011). In both treatments, however, therapeutic focus on trauma is deferred until a number of other issues have been addressed. These issues, which constitute the focus of Stage 1 DBT, include self-harm and suicidality, as well as other impulsive behaviors (such as angry outbursts or substance use) that have a deleterious effect on the patient’s safety, therapy, or quality of life. In adapting DBT for DID, we relied on the following three principles.

Principle One. Given that patients with a DID diagnosis tend to closely resemble patients with BPD in the frequency of Stage 1 target symptoms, the first principle in adapting DBT for DID was to utilize the DBT model of stabilizing symptoms of behavioral dysregulation (e.g., suicidality, impulsivity, etc.) before addressing traumatic material. Our experience is that the Stage 1 techniques are necessary, and for the most part, transfer nicely for use with patients who have DID.

Principle Two. Where the treatment of patients with DID will differ from those with BPD is in the presence of extensive dissociative behaviors, which complicate both Stage 1 and Stage 2 work. Chaotic presentations caused by dissociative symptoms usually require at least partial stabilization prior to engaging traumatic material. These dissociative symptoms include the presence of different personality states (“alters”) that may cause problems by impulsivity or self-harm; often the patient reports amnesia for these behaviors, which greatly complicates the process of addressing them. One part of the patient may be motivated for treatment, while other personality states may be mistrustful or actively opposed to treatment. The patient may be unwilling to take responsibility for harmful behaviors because “it wasn’t me.” Different alters may attend different sessions, and this may lead to amnesia for skills taught. A pattern may emerge in which the patient switches to another self-state when a session becomes difficult, disrupting the behavioral analysis.

These special issues often require attention because they are associated with potentially life-threatening behaviors or because they disrupt the usual process of treatment (therapy-interfering behavior). Thus, our second principle: to the extent that dissociative behaviors function as therapy-interfering behaviors or are functionally related to life-threatening behaviors, they must be targeted in Stage 1.

Principle Three. The phrase “dissociative behaviors” covers a wide range of phenomena. These include depersonalization, numbing, “spacing out” or trance-like behavior, memory distortion, or amnesia-all of which may be found in any of the dissociative disorders, as well as other disorders such as PTSD. However, the specific symptom of “switching” from one personality state to another personality state which is subjectively experienced as separate is the hallmark of DID, and for the most part does not occur in other contexts. Although these may all be labeled as “dissociative behaviors,” it is our position that switching in DID requires special consideration.

The phenomenon of switching presupposes the existence of alternating, discontinuous personality states—described in the DSM-5’s diagnostic criteria for DID as “disruption of identity characterized by two or more distinct personality states,” with “marked discontinuity in sense of self and sense of agency” (American Psychiatric Association, 2013, p. 292). These personality states, which develop in childhood as a reaction to extensive trauma, harbor separate traumatic memories, as well as different aspects of the individual’s overall personality (e.g. one state may be assertive and forceful while another is timid, another sexual, etc.). Usually, patients are unhappy with the existence of the other personality states, and frequently express the wish to be rid of them. In response, the therapist explains that the others, though seemingly separate, each contain valuable aspects of the whole, are all in fact parts of the whole person, and cannot be eliminated. This perspective is crucial as we discuss the therapist’s approach to the patient’s switching behaviors, as compared with other dissociative behaviors.

When a patient presents with problematic dissociative behavior other than switching, the DBT approach can be fairly straightforward. For example, if a patient tends to go into a trance-like state in response to feeling anxious, and this state interferes with therapy or with day-to-day life, this behavior can be labeled as a treatment target, and the usual techniques (behavioral analysis, solution analysis) can be applied to it. However, we propose that the therapist and patient cannot similarly adopt “stopping switching” as an uncomplicated treatment goal. If the patient experiences more than one personality state, the states must necessarily be accessed by switching from one state to another. Thus to aim for “stopping switching” would mean either joining the patient in the impossible wish that some parts of the personality are eliminated while one part is preserved, or else setting up the unrealistic expectation that the patient should access all different personality states without switching–i.e., that she achieve the therapeutic goal of integration instantaneously, prior to starting therapy. In terms of the patient’s in-the-moment experience of the therapy, asking the patient not to switch would mean that the therapist would be accepting one part of the patient (the one currently speaking to the therapist) as valid, and invalidating other parts of the patient.

Our solution is to apply the DBT principle that it is the functions and consequences of behaviors that determine whether they are maladaptive and need to be targeted. Within this framework, the fact that a patient may “switch” from one personality state to another—the defining symptom of DID—is not necessarily seen as a problem, and eliminating switches may or may not be a treatment goal. In fact, the successful therapy of a patient with DID will inevitably include countless switches, which with time become more volitional and well-controlled. Switches may lead to dangerous behavior, or therapy-interfering or quality-of-life-interfering behavior, in which case they would be targeted as problematic; or switches may be part of a successful adaptation, without harmful consequences, in which case they would not be targeted. Thus, our third principle is that “switching” is not assumed to be maladaptive, but rather, can serve numerous functions, which should be assessed with behavioral analysis. Specifically, when we identify a maladaptive pattern of switching, the treatment aims at decreasing that particular maladaptive switching behavior, not at stopping all switching. We believe this approach to be completely consistent with general principles of DBT, although to our knowledge the specific issue of switching, as distinct from other dissociative behaviors, has not been addressed in previous papers from a DBT perspective. (For excellent discussions of managing dissociative symptoms other than switching in DBT, see papers by Wagner & Linehan, 1998; Granato et al., 2015).

Illustration of the Principles

A recurring challenge for Mariella and her therapist was the fact that Mariella became highly emotionally dysregulated by the process of conducting BA’s on self-harm behaviors. During these BA’s, she would frequently switch to an alter who prided herself on emotional strength and berated Mariella for cutting (“Ivette”), or a child alter who was not aware of cutting behaviors (“Cassie”). Both forms of switching represented “therapy-interfering behaviors”, as neither alter engaged in cutting and neither could provide the details or solutions needed for a useful BA. Thus, in line with our second principle, Mariella’s therapists targeted in-session switching during behavioral analyses as therapy-interfering behavior and sought to determine its common triggers, functions, and consequences in order to develop a solution for this problematic behavior.

In the case example below; we enter the session as Mariella and her therapist begin a chain analysis about a self-harm episode that occurred on the Sunday before the session, per Mariella’s diary card. The therapist is discussing the diary card with Mariella, and during the exchange below also speaks with Ivette.

This case example illustrates Principle Two of targeting dissociation when it functions as an in-session “therapy-destroying behavior” (Linehan, 1993) because in this instance, switching kept therapist and patient from adhering to the DBT hierarchy of addressing life-threatening behaviors. In this example, the therapist acknowledged Ivette, referring to the patient by this name and expressing interest in speaking with her, while at the same time expressing the need to speak with Mariella to complete the behavioral analysis. The therapist shifts focus to complete a behavioral analysis of the in-session switching behavior so that a solution analysis can be completed to allow completion of the original behavioral analysis; this aspect of our approach is consistent with standard DBT.

Following our approach, not all instances of switching were targeted by Mariella and her therapist as problem behaviors (i.e., Principle Three). For example, during one suicide risk assessment, Mariella was tearful and repeating, “I’m fine, I don’t want to go to the hospital.” A switch occurred, and Ivette told the therapist, “I don’t know if can keep her safe right now. She keeps blocking me.” The therapist then spoke with Ivette, as doing so allowed the therapist to gather valuable information about the patient’s suicide risk: namely, that the suicide urges were getting stronger, and the part of the patient who wanted to live (i.e., Ivette) was becoming weaker—akin to a patient without DID expressing reduced ambivalence about suicide. In this instance, switching was not targeted as therapy-interfering, as switching served the function of helping promote the patient’s safety.

Modifications from Current Treatments

Having delineated some general principles, we turn our attention to specific issues that arise when using a DBT approach with a DID population, and propose additions and modifications to both standard DBT and to what we will call “consensus DID treatment.” There is not one universally agreed-on psychotherapeutic approach for the treatment of DID, so we will use the Treatment Guidelines mentioned in the Introduction (International Society for the Study of Trauma and Dissociation, 2011) as representative of mainstream practice and will refer to this as “consensus DID treatment.” Many other valuable accounts of DID treatment exist, and were used as references in assembling a picture of consensus practice (Boon, Steele, & van der Hart, 2011; Chefetz, 2015; Cloitre, Cohen, & Koenen, 2006; Howell, 2011; Kluft, 1999; Loewenstein, 2006; Putnam, 1989; Ross & Halpern, 2009). Our adaptation of DBT for DID introduces several therapeutic interventions and strategies which are drawn from consensus DID treatment and would be applicable in Stage 1, including “mapping the alters,” “grounding,” “containment” and “safe place” imagery. Before reviewing specific techniques, however, we begin with the overarching issue of addressing personality states.

Addressing Separate Personality States

The “biosocial theory” of BPD serves both as a theoretical underpinning and an important clinical touchstone for the DBT treatment approach. Patients find it enormously helpful in understanding their condition, and clinicians turn to it for guidance with case formulation and treatment decisions. The treatment of DID is perhaps even more reliant on a central explanatory concept—sometimes referred to as the “trauma model of DID” (Dalenberg et al., 2012). In this model, dissociation is seen as both a spontaneous reaction to trauma and a defense against overwhelming traumatic affects. We posit that when a child faces chronic trauma from an early age, her repeated experiences of traumatic dissociation can lead to the development of seemingly separate personality states separated by amnestic barriers. This concept is essential in helping the patient bring together bafflingly disconnected pieces of her experience, as well as framing them in a sympathetic light. For the therapist, the fundamental understanding that patients with DID have dissociated in response to overwhelming traumatic affects, and that the personality structure is based on avoiding these affects, informs all clinical decisions.

The trauma model serves as the foundation for a crucial dialectical maneuver that begins the therapy of DID and continues throughout. The therapist tells the patient that she (the patient) experiences herself as if she were many different people, and that the therapist embraces this reality and will respectfully attend to each part of her individually. Simultaneously, the therapist tells the patient that despite her conviction to the contrary, the patient is in fact only one person, and all of her seemingly separate aspects are actually parts of one person. Both facets of this seemingly contradictory message are essential for the therapy to succeed. This is akin to the fundamental DBT dialectic of accepting the patient where she is, and at the same time, expecting her to change, all in the service of building a life worth living. Just as DBT therapists avoid the word “but”—replacing it with “and”—the DID therapist will say to the patient’s alters: “I am going to engage with each of you individually, and I respect your sense of yourselves as separate people—AND, I intend to continually push you to realize that you are only one person, and to begin to experience yourself as such.” This stance leads to the therapist’s persistent urging that the patient improve communication and cooperation between parts of herself.

Some patients with DID desire “integration” (the process by which alter personalities gradually merge into one, so that the patient ultimately does not manifest separate personality states), but some do not—so this may or may not be a goal of the therapy. What is always a goal is to reduce the ways in which the existence of separate personality states causes problems in living. Successful therapy always involves working toward better cooperation and communication between personality states, which reduces symptoms such as amnesia and internal conflicts that express themselves as self-damaging behavior, as explicated below. At the start of treatment many patients with DID will not understand their situation well enough to be able to agree to this goal, so we do not suggest it as a prerequisite for treatment. As therapy proceeds, however, the therapist brings increasing pressure to bear on this central issue, because the patient’s DID will not resolve without these efforts.

DBT includes a number of strategies to improve commitment—both the patient’s general investment in the therapy process, and her commitment to specific changes in behavior. In addition to techniques such as “devil’s advocate” and “foot-in-the-door/door-in-the-face,” we recommend adding a commitment strategy which we colloquially refer to as “sweet-talking the alters” but which can be seen in DBT terms as a validation strategy. At the beginning of treatment, while some personality states engage with the therapist and evince motivation for treatment, there are invariably other personality states taking an opposite position—of mistrust of the therapist, of unwillingness to discuss traumatic memories, or of simply wishing to be left alone to die. Often, parts of the patient are convinced that revealing information about trauma history or DID symptoms would be highly dangerous. This opposition may be voiced explicitly, but more commonly there are personality states in the background, quietly resisting therapeutic involvement. Ultimately, treatment success depends on engaging all parts of the patient—especially the “negative” parts—so that the therapist seizes every opportunity to reach out to them. For instance, Mariella complained that “Ivette” was making life impossible with constant angry outbursts, and asked, “Why can’t I just get rid of her?” Ivette had been unwilling to participate in the therapy, asserting, “None of her mess is my fault.” However, as the therapist acknowledged the difficulties presented by Ivette’s anger, she made sure to offer Mariella observations about her alters such as “I have heard about how you were treated as a child, and although it may seem like Ivette is overreacting, I am certain that she has good reasons to feel so angry”; or “yes, her anger creates problems sometimes, but she is also able to stand up and be assertive in a way that is very difficult for you.” Statements like these are second nature to DBT therapists as validation strategies; what is different here is that the therapist is proactively validating the emotions of personality states with whom she has not yet spoken, with the aim of making these parts more receptive to entering the therapy arena.

Discussing commitment work leads us to another challenge encountered in working with DID: namely, the patient’s sequential presentation of discontinuous and often diametrically opposing views. For instance, if a patient says that she is committed to therapy, and in the next session the same patient presents as a different alter who states that she has no intent whatsoever of engaging in therapy, is the patient committed to therapy? The question arises even more pointedly during safety assessment, when a patient may assure the clinician that she is not feeling suicidal “but I can’t speak for my other parts.” We cannot offer any technique that entirely eliminates the clinical uncertainty which arises when therapist and patient do not have easy access to all personality states–it is at times an unavoidable part of treating DID–but we offer several suggestions.

First, the therapist persistently urges the patient to improve communication with other parts of herself. The therapist might suggest that the patient “ask inside,” or use other techniques, such as journaling or “the dissociative table technique” (Fraser, 2003) to gradually improve internal communication.

Second, the therapist understands that this situation is not so different from a safety assessment with a non-DID, BPD-only patient, where the patient may make commitments about safety but where the therapist worries that these commitments will not hold up when the patient is in a different emotional state. The therapist includes this possibility as a risk factor in the safety assessment, and the DID therapist makes a similar calculation (“Is the patient likely to switch to a more dangerous alter? Under what circumstances?” etc.) and strategizes as to how to manage these possibilities. Ultimately, in both scenarios, the therapist makes her best assessment by weighing the relative strength of pro-safety and proharm aspects of the patient.

Third, experienced DID therapists convey the message that the patient is ultimately responsible for her safety, despite feeling that she lacks control over other parts of herself. When a patient says that she can’t guarantee safety because she cannot control other parts, we may say “That may be, but we need to find a way to be sure you’ll be safe anyway—how can we do that?” Possibilities include asking if other pro-safety alters can commit to intervening, planning ways for safe parts to stay in control, and making backup plans in case the patient feels that control is slipping. We suggest that the patient examine her own mixed feelings about keeping safe, although the patient disavows the suicidal feelings as “other.” In this vein, we may tell a patient that “while it may seem like the suicidal feelings are someone else’s, we believe that if you are 100% committed to not acting on them, then you won’t–it is your own ambivalence about suicide that opens the door for other parts to engage in dangerous behavior”. This intervention helps to bring the split-off feelings slightly closer, and gives the patient an empowering message of responsibility and control.

Other DID-Specific Techniques

A technique that can be integrated into DBT behavioral analysis is “mapping the alters.” This involves determining under which circumstances different parts of the patient are most prominent, as well as the ongoing status of communication and cooperation between the alters. Mapping the alters continues throughout the course of treatment, but would typically begin in Stage One as part of establishing a safe treatment frame. As behavioral analysis involves a detailed functional assessment of patient behaviors, it is ideally suited to allow DBT therapists and patients to learn about patients’ alters and the functions these alters serve. Especially important is assessing the patient’s state of awareness of her DID and whether parts realize that they share the same body or believe that they inhabit different bodies, as this reality distortion can confound behavioral analyses and increase risk.

Another technique from the DID literature, which fits comfortably among DBT’s “crisis survival skills,” is the development of a “safe place” or “healing place” (Turkus & Kahler 2006). This technique involves guided imagery in which the therapist helps the patient develop an image of a place, either real or imaginary, where she feels safe. After practice with the therapist, once the patient can establish a feeling of safety in the imagined place, this technique can be used both within and outside of sessions when the patient begins to experience overwhelming emotions, especially trauma-related terror. Patients are more willing to engage in the hard work of behavioral analyses that often leaves them feeling vulnerable and afraid if they know they have skills to help them manage the strong emotions that will likely result.

While DBT skills focus on managing emotion dysregulation of all sorts, the skills used in consensus DID treatment tend to focus specifically on managing post-traumatic symptoms such as flashbacks, since these symptoms are so ubiquitous in DID. In adapting DBT for DID, it is important to augment the standard DBT crisis survival skills with these additional skills specifically aimed at post-traumatic symptoms. “Safe place,” outlined above, is used for this purpose; two other sets of consensus DID techniques which are particularly DID-useful crisis survival skills are “containment” imagery and “grounding.” In containment imagery, the therapist and patient develop a set of images to contain traumatic memories and affects (Turkus & Kahler, 2006). “Grounding” refers to a variety of techniques used to reestablish a patient’s orientation in the here and now when traumatic memories intrude and leave the patient unable to distinguish past from present. Both grounding and containment are taught briefly in the “distracting” component of the DBT distress tolerance module, though not named as such. Containment is similar to the “pushing away” DBT skill in which patients are taught to “put the pain on a shelf. Box it up and put it away for a while” (Linehan, 2015, p. 441). Grounding is also used as one form of “distracting” in DBT: patients are taught to distract for the purpose of “surviving crises” by focusing on other thoughts (e.g., naming the colors in a painting) or sensations (e.g., holding an ice cube). However, DBT does not focus on the use of these skills in flashback regulation, while the DID literature has provided detailed discussions and clinical illustrations of these techniques (Turkus & Kahler, 2006; Vermilyea, 2013).

Challenges

While mindfulness exercises, such as following the breath, might be used to manage post-traumatic intrusions with patients who do not have DID, we and others have found that patients with DID may experience such mindfulness exercises as invitations to dissociate. As an illustration of how some mindfulness exercises may make switching more likely, we continue with the case illustration. In response to the therapist’s question about which skill Mariella would like to use, she selects mindfulness to a sensory experience, namely, Play-Doh^TM. The therapist hands Mariella the Play-Doh^TM.

The therapist attempts to help Mariella prevent dissociation by practicing mindfulness with Play-Doh^TM; however, it serves as a trigger for the patient’s child alter to emerge, disrupting the BA, so that the therapist switches to a grounding mindfulness exercise involving naming colors in the room. This exercise is more concrete and guided by the therapist. By attending to the patient’s current affect and dysregulation, and engaging the patient with in-vivo mindfulness skill practice that does not prompt dissociation, the therapist is able to help Mariella to identify the sensations that immediately precede switching, as well as the triggers for the recent self-harm incident.

This example also illustrates another dilemma faced by DID thera-pists—how to complete behavioral analyses when a patient is amnestic for the event being analyzed. In this example, the therapist repeatedly asked to speak to the part of the patient with the necessary information (i.e., Principle Three, encouraging adaptive switching).

Another challenge that often arises is the need to generalize skills between parts. During another behavioral analysis of a particularly severe episode of self-harm, Mariella’s therapist determined that another alter, “Ashley,” engaged in the self-harm behavior because she was angry at Mariella (i.e., an expression of self-hatred, a common trigger for the patient’s self-harm). In speaking with Ashley, the therapist determined that this part of the patient had no desire to prevent further episodes of self-harm, did not know any distress tolerance skills, and would not commit to learning and using skills to prevent self-harm. Mariella stated that she was not able to communicate with Ashley and would be unable to stop Ashley from self-harming, but that she could notice becoming sleepy before switching. Thus, the solution analysis involved using mindfulness skills to notice sleepiness and then to “call for Ivette” because Ivette reported being able to communicate with Ashley and also reported being stronger than Ashley and able to keep Ashley from engaging in self-harm. In this case, switching was used as a crisis survival skill, allowing the therapist and patient to use commitment strategies to increase Ashley’s motivation to use skills to reduce self-harm while keeping the patient safe in the meantime. The therapist will typically need to return to DBT’s arsenal of commitment techniques repeatedly, with different alters, hoping to increase motivation among all alters, but working selectively with more motivated alters, as necessary, until this can be achieved.

The amnestic barriers of DID often interfere with skills generalization. For instance, one alter may typically attend skills training group, because she feels more comfortable in the group setting, but this may not be the alter typically involved in dangerous behaviors, so the learned skills may not be accessible when they are needed. Various strategies may be used here, including encouraging “co-consciousness” either when learning skills or when called upon to use them; encouraging planful internal dissemination of skills as part of the general agenda of improving internal communication and cooperation; or recommending adaptive switching to an alter who is able to mobilize the requisite skills. Finally, patients with DID often require extra coaching to be able to attend group therapy without maladaptive switching, which either confuses or frightens the patient or interferes with the functioning of the group. For instance, a patient who switches into a conspicuously child-like state during a group session is likely to be very distracting to other group members, and this will need to be addressed as therapy-interfering behavior.

Discussion

The case example provided illustrates how a behavioral approach can be used with patients with DID to target and reduce self-harm and suicidal behaviors. After two years of DBT-informed treatment for DID, Mariella substantially reduced the frequency of self-harm behaviors, including a six-week interval with no NSSI or suicidal behaviors—the longest period she could remember ever abstaining from this behavior. This represented a significant treatment milestone for Mariella, as it contradicted her belief that she would kill herself if she did not cut herself every week. (In longer-term followup, Mariella continued in DBT for four additional years after the treatment described in this article, and by the time patient terminated treatment because she moved out of the area, self-harming and suicidal behavior had been completely eliminated during the last three years of treatment, and Mariella had achieved her “life-worth-living” goals of earning her GED and graduating from beauty school.) The authors strongly believe that Mariella’s success would not have occurred without both essential treatment elements: the behavioral management skills and treatment structure derived from Stage 1 DBT, and the full engagement of her alters in the treatment, following the DID consensus approach.

The treatment model described here utilizes a principle-guided adaptation of Stage One DBT for the treatment of emotion and behavioral dysregulation in patients with DID before initiating formal trauma work, either in a DBT framework or in a consensus DID treatment framework. The DBT therapist helps the patient find a “middle path” (Miller, Rathus, & Linehan, 2007) between extreme emotional dysregulation in the face of distress and experiential avoidance or denial of distress. For the patient with DID, the “unrelenting crisis” is the repetitive intrusion of overwhelmingly painful, traumatic memories, while “inhibited grieving” presents as uncontrollable dissociation, with amnesia and severe disruption of identity. Linehan (1993) has characterized borderline “splitting” as dialectical failure, highlighting the rigid cognitive style that leaves the patient unable to resolve contradictory beliefs or views of herself. In DID, the amnestic separation between alters represents an even more severe dialectical failure, in which the patient inhabits first one reality, followed by another, conflicting one, with the different parts experienced as separate and incompatible. The therapy proceeds steadily and dialectically, usually over several years, towards what DBT calls “synthesis and transcendence” and a DID therapist calls cooperation, “co-consciousness,” or integration. We suggest that both DBT case conceptualization and treatment structure, including a hierarchy of behavioral targets, are well-suited to the treatment of DID and that structured skills training in DBT provides an optimal set of procedures for achieving the safety and stabilization goals of the initial phase of DID treatment. However, DID treatment involves unique clinical challenges that necessitate modifications of the standard DBT protocol.

Our approach differs from consensus DID treatment in several ways. First, the consensus treatment model is less structured than DBT and does not specifically delineate how to address safety issues (i.e., suicide and self-harm) and emotion regulation skill deficits. Second, the consensus treatment model emphasizes contextual and psychodynamic conceptualization and intervention, including a focus on attachment and transference issues, rather than behavioral conceptualization and intervention.

The proposed approach also differs from standard DBT, which has not described working with alters. In our approach, as in consensus DID treatment, alters are engaged directly in therapy, the alters’ existence is not automatically seen as a problem, and the therapist does not try to discourage them from existing, or prevent the patient switching from one to another (Kluft, 2006). This overarching philosophy would seem to contrast with, for example, the approach described by Wagner and Linehan (1998) or Wagner et al. (2007), where there is not a distinction made between dissociative symptoms in general and the specific phenomenon of switching in DID. Our approach differs even more sharply from that of Harned (2013). In her report of the treatment of a PTSD patient with dissociative symptoms, the patient was strongly discouraged from switching, and the therapist punished dissociative behavior by withdrawing warmth and expressing irritation. We propose that directly engaging alters in treatment is not antithetical to DBT principles, and in fact is likely necessary for the successful implementation of DBT with DID patients. Our view is that dissociative behaviors, including switching, may involve a variety of functions, some adaptive, some maladaptive, thus necessitating a principled and flexible approach to the clinical management of dissociative behaviors in patients with DID. We suggest that this approach fits naturally with a DBT therapeutic approach and that engaging alters in treatment in a principled, structured manner is a necessary element of “meeting the client where she is” to bring about lasting and generalizable behavioral change.

The treatment of DID is difficult and complex. Dialectical Behavior Therapy is an elaborate, multifaceted treatment model. Therefore, adapting DBT for the treatment of DID is bound to be an enormously complicated proposition. This introductory paper only begins to address the multiple issues which arise in this adaptation, but we have attempted to outline several guiding principles and to provide orientation to some basic clinical issues. A longer publication would better detail the overlap between the two clinical populations, would describe other challenging situations which arise when applying DBT to the DID population, would outline approaches to pervasive negative schemas such as self-blame which are central to successful DID treatment, and would provide a longer list of specific trauma-related skills that are commonly used in DID treatment and which would be well-incorporated here—most likely in the form of an additional, trauma/dissociation-related skills module. In our view, the use of DBT, with modifications to address the various functions of severe dissociative symptoms, represents a promising and empirically-informed behavioral approach to the treatment of DID warranting further investigation in a research protocol. The comorbidity between BPD and DID, as well as the similarity of presenting problems, makes DBT a natural fit for treating patients with DID. We hope that our article will provide clinical guidance as well as generating discussion and providing a platform for the conduct of the much-needed research into psychotherapy for this painful, debilitating, and at the same time, treatable disorder.