THE BEHAVIORAL TREATMENT OF SELF-MUTILATION

Self-mutilation, treatment & research symposium. XVI Congress of the
World Association for Social Psychiatry, Vancouver B.C., Canada.

Milton Brown

University of Washington


Address correspondence to:
Milton Brown
Department of Psychology, Box 351525
University of Washington
Seattle, WA 98195-1525
mibrown@u.washington.edu




Self-mutilation is a serious public health problem, yet there is very little empirical evidence little empirical evidence showing that treatments can reduce this maladaptive behavior. Self-mutilation is defined here as intentional and acute physical self-injury without intent to die, which includes various methods such as cutting and burning. In contrast, only a minority of individuals attempt suicide (i.e., intend to die) by cutting or burning (e.g., Wexler, Weissman, & Kasl, 1978), and suicide attempts are much more likely to be medically severe (Brown & Linehan, 1996). Despite these differences between self-mutilation and suicide attempts, many individuals who self-mutilate also attempt suicide or wish to die. It has been estimated that about half of individuals who self-mutilate also attempt suicide (e.g., Hillbrand et al, 1994). Collectively, this group of suicidal and nonsuicidal self-injury will be referred to as parasuicide (Kreitman, 1977).

Self-mutilation is an especially prevalent and chronic problem among those diagnosed with borderline personality disorder (BPD). BPD is a chronic debilitating disorder that is characterized by pervasive dysfunctions in emotions, behavior, and cognition. More BPD individuals engage in self-mutilation and attempt suicide than other diagnostic groups (e.g., Schaffer, Carrol, & Abromowitz, 1982; Linehan & Heard, 1997; Tanney, 1992), often with repeated acts over time. According to the classification scheme proposed by Favazza and Rosenthal (1993), self-mutilation in BPD can usually be described as superficial or moderate in contrast to major self-mutilation (i.e., infrequent acts of great severity, often associated with psychosis) or stereotypic self-mutilation (i.e., very frequent acts of usually fixed severity, often associated with mental retardation; cf. Schaffer et al., 1982). Indeed, most self-cutting or burning by BPD individuals involves little suicide intent or medical risk in contrast to other methods of parasuicide such as drug overdose (e.g., Brown & Linehan, 1996; Wexler et al., 1978). By and large, chronic self-mutilation and BPD have proven to be intractable conditions.

Although behavior therapy has been used for many years to treat many different psychological disorders, it has only recently been applied to the treatment of self-mutilation. Several studies suggest that cognitive behavior therapy can reduce suicide attempts, but I could locate only two randomized controlled treatment studies that may be relevant to self-mutilation. Linehan et al. (1991) found Dialectical Behavior Therapy (DBT) to be effective in reducing parasuicidal behaviors overall. The results of another study suggest that a manual-assisted short-term cognitive behavior therapy may reduce parasuicide (Evans et al., 1999). Neither study, however, reported outcomes specifically for self-mutilation.

This paper will present a behavioral conceptualization of parasuicide based on Linehan's biosocial theory of BPD (1993). Next, behavioral treatment strategies will be reviewed that are used to treat parasuicidal behaviors in DBT, a comprehensive treatment for BPD and parasuicidal behaviors based on this biosocial theory and social learning theory (Linehan, 1993). Although this paper will focus on self-mutilation, the etiology and treatment of both forms of parasuicide will be discussed. Although self-mutilation and suicidality are not the same, these two forms of parasuicide often co-occur in BPD, and it is likely that reducing the vulnerability to one can also lead to improvements in the other.


EMOTION DYSREGULATION AND SELF-MUTILATION

Several theorists have suggested that emotional dysregulation is an important risk factor for chronic self-mutilation. Specifically, psychodynamic theorists have stated that both self-mutilation and suicide attempts are the result of anger turned inward (Liebowitz, 1987; Friedman, et al. 1972) and results in emotional catharsis (Gardner & Cowdry, 1985). In contrast, Linehan (1993) has presented a biosocial theory that describes chronic negative emotions and self-invalidation as primary factors that predispose BPD individuals to self-mutilation and suicide attempts. This biosocial theory states that emotion dysregulation results from a combination of high sensitivity or reactivity to emotional stimuli and a deficit in emotion regulation skills (e.g., inability to distract). BPD individuals often engage in impulsive and maladaptive behaviors (including parasuicide) in response to intense emotions. These maladaptive behaviors are either automatic, mood-dependent responses to emotions or attempts to modulate or cope with emotions. Furthermore, individuals fluctuate between extremes of emotional experiencing and emotional inhibition. According to Linehan, self-invalidation is learned from environments that invalidate self-generated behaviors and communication of private experiences. These behaviors are often punished, trivialized, disregarded, pathologized, or criticized. The accuracy of one's self-description is rejected, and instead, behaviors are attributed to negative traits such as laziness, manipulative intent, negative attitude, and paranoia. Self-invalidation occurs when individuals have learned to disregard, punish, and invalidate themselves (and their emotions) in ways similar to how others have treated them (e.g., "internalization"). Such individuals blame and judge themselves harshly for their lack of control of behavior and emotions, and they treat their normal responses as invalid. Other forms of abuse that commonly occur in these environments may also explain why these individuals become highly emotionally reactive. Given an initial biological vulnerability, these traumatic experiences may sensitize these individuals with to react strongly to emotional stimuli and a variety of cues that become conditioned to early traumatic experiences.

According to Linehan's biosocial theory, negative emotions contribute to chronic parasuicide in BPD in three ways: 1) the reduction of emotional arousal following parasuicide negatively reinforces the behavior, 2) anger, contempt, and shame interfere with problem solving and emotional processing, and 3) shame-related emotions directly lead to self-punishment, or an extreme desire to hide or disappear (i.e., lose consciousness or die).

Parasuicide is reported to be a method individuals learn to relieve negative emotions, specifically, anxiety, tension, and possibly anger in some cases (evidence reviewed below). Drug overdose (and use), for example, becomes a principal way to reduce emotional tension. For reasons not currently well understood, self-mutilation often seems to effectively relieve aversive emotions (Simeon, Stanley, Frances, Mann, Winchel, & Stanley, 1992). Thus, when aversive emotions are relieved after parasuicide, parasuicide becomes negatively reinforced, and therefore, more likely to be repeated.

Anger, contempt, and shame interfere with processes that could reduce parasuicide. Emotion dysregulation interferes with problem solving because of impaired cognitive control and avoidance of problems. Shame in particular motivates individuals to conceal their problems from those who could help them and to avoid focusing on problems enough to solve them. Anger interferes with the collaboration necessary for effective treatment. Anger, contempt, and shame may also interfere with emotional processing of fear and sadness-processing that may be necessary for improvement in BPD. Linehan proposes that the avoidance or inhibition of grieving is a fundamental process contributing to the maintenance of suicidal behaviors in BPD. One study suggests that anger and contempt interfere with recovery from grieving (Bonanno & Keltner, 1997). Similarly, contempt can interfere with improvement in major depression (Ekman, Matsumoto, & Friesen, 1997). High levels of anger can also interfere with the treatment of post-traumatic stress disorder by interfering with the processing of fear (Foa, Riggs, Massie, Yarczower, 1995). The shame experienced by Vietnam veterans over war atrocities may similarly explain why many often fail to benefit from exposure therapy. In previous maladaptive environments, individuals who self-mutilate have learned to cut off or invalidate their own primary emotional responses (often with other emotions modeled in the environment) as a means of regulating them. They believe that emotions cannot be tolerated. Since these individuals have not learned to effectively process the emotional material, they continue to re-experience precipitating events that elicit similar emotional responses. The presence of these emotions may explain why BPD has been associated with worse outcome in treatments of Axis I disorders such as major depression (Phillips & Nierenberg, 1994), OCD (Baer, Jenike, Black, Treece, Rosenfeld, & Greist, 1992), bulimia (Ames-Frankel, Devlin, Walsh, Strasser, Sadik, Oldham, & Roose, 1992; Coker, Vize, Wade, & Cooper, 1993) and substance abuse (Kosten, Kosten, & Rounsaville, 1989).

Finally, Linehan's theory states that in many cases parasuicide in BPD may be a form of learned self-invalidation that directly results from shame, self-hatred, or self-contempt. The intent to die may result from the basic urge to escape, hide and disappear invoked by intense shame, with suicide seen as a complete and irrevocable way to hide from the shameful scrutiny of self and others (cf. Baumeister, 1990/1991). Similarly, intoxication from drug overdose may serve a similar function by lowering aversive self-awareness (Hull & Young, 1983). Parasuicide can also function as self-punishment, and the injuries sustained viewed as deserved by the people who do it. Self-verification theory (Swann, 1992) would predict that individuals who hate themselves will seek experiences congruent with that self-attitude, desiring to harm themselves similar to how one might desire to harm any disliked person. People selectively seeking out experiences to confirm existing beliefs to maintain the stability of their basic beliefs and attitudes about themselves, which is necessary for making sense of the world and guiding behavior (Festinger, 1957). Lack of verification of our basic self-beliefs engenders feelings of tension and incoherence and a fear that our sense of self is fragmenting or disintegrating ("disintegration anxiety" or "cognitive dissonance"). When the stability of a person's conceptual system is threatened, anxiety mounts and that person may experience disorientation, depersonalization, and feelings of emptiness (akin to a psychological death). Failing to self-mutilate could threaten to change the person's self-concept. Thus, self-mutilation strongly confirms one's negative self-view and is able to relieve the intense anxiety, depersonalization, and emptiness that self-mutilators report before they harm themselves. At a minimum, a positive self-regard would inhibit any drive to harm oneself. Self-mutilators often report that their self-injury helps them form an identity and feel as if they exist (e.g., Leinbenluft et al., 1987).


EVIDENCE ON THE ASSOCIATION OF EMOTIONS TO SELF-MUTILATION

Empirical evidence for the emotion dysregulation hypothesis of parasuicide in BPD primarily comes from three lines of evidence: observed differences between parasuicidal and non-parasuicidal individuals, correlations between parasuicide and negative emotions, and patients' self-report that their parasuicide is due to emotion dysregulation. Caution is warranted in generalizing from some of these studies to chronic parasuicide in BPD, however, as many of these studies were not limited to BPD samples or involved suicide rather than parasuicide.

Self-mutilation and suicide attempts are often described as intentional efforts to escape distressing circumstances or to feel better (e.g., Bancroft, Skrimshire, & Simkins, 1976; Hawton, Cole, O'Grady, & Osborn, 1982; Parker, 1981). Although patients attribute their parasuicide to a wide variety of "reasons" and expect a variety of effects resulting from the behavior, emotion regulation is usually reported as the primary intent of parasuicide (e.g., Brown & Linehan, 1996). Tension relief was cited as a primary motive for self-cutting in 85% of 39 cutters (Jones, Congiu, Stevenson, Strauss, and Frei, 1979). Clinicians most often attribute parasuicide to hostile attempts to influence others although patients usually do not express this intent (Bancroft, Hawton, Simkin, Kingston, Cumming, Whitwell, 1979). Ninety-two percent of a sample of 63 parasuicidal subjects diagnosed with BPD reported intending to relieve negative emotions through self-mutilation or suicide attempts (Brown & Linehan, 1999). Self-mutilation, however, was rated as more effective than suicide attempts at relieving negative emotions. In some cases, parasuicide is described as behavior that is intentionally performed to regulate emotions, while in other cases parasuicide is described as impulsive and its consequences as automatic. For example, Wilkins and Coid (1991) reported that although only 40% of subjects intentionally engaged in self-mutilation to relieve negative feelings (the primary motive reported), 69% reported a relief of tension, dysphoria, and irritability, suggesting that the process is not necessarily intentional. Studies show that 80 to 94% of self-mutilators report emotional relief after self-mutilation (Bennum & Phil, 1983; Coid, 1993).

DEPRESSION AND PARASUICIDE

.Evidence linking depression to parasuicide supports the importance of emotion vulnerability. High levels of depression are experienced by many individuals who self-mutilate (Bennum & Phil, 1983) and attempt suicide (Maris, 1981). A vast literature shows that major depression increases the risk of suicide ideation, attempted suicide, and completed suicide (e.g., Tanney, 1992; Gould et al., 1998). Among BPD individuals, major depression is related to an increased likelihood and severity of suicide attempts (Comtois, Brown, & Linehan, 1998; Fyer, Frances, Sullivan, Hurt, 1988). Research suggests that it may not be the presence of a depression diagnosis per se (Soloff, Lis, Kelly, Cornelius, & Ulrich, 1994a) but rather the experienced affects that predict suicidal behavior in BPD. For example, both Soloff et al (1994) and Mann (Mann, 1997) found that depressive moods and experiences, rather than comorbid affective disorder or severity of disorder, are associated with suicidal behavior in individuals with BPD. Depressed mood is associated with increased risk of non-fatal suicide attempts (Lewinsohn, Rohde, & Seeley, 1994) and frequent non-suicidal self-mutilation (Dulit, Fyer, Leon, Brodsky, & Frances, 1994). Although depression is probably a risk factor for parasuicide, it is not clear which emotions are most implicated since depression can occur with high levels of sadness (Maris, 1981), anxiety (Maser & Cloninger, 1990; Ohring, Apter, Ratzoni, Weizman, Tyano, & Plutchik, 1996), shame (Andrews, 1995; Maris, 1981).

SPECIFIC NEGATIVE EMOTIONS RELATED TO PARASUICIDE

.Many studies suggest that parasuicidal behavior is associated with high levels of specific negative emotions. Many studies demonstrate the importance of anger in parasuicide. These studies have found that suicide attempts are associated with high levels of irritability, anger, and hostility (Crook, Raskin, & Davis, 1975; Weissman, Fox, & Klerman, 1973; Paykel & Dienelt, 1971). Similarly, studies have found that self-mutilators have high levels of anger and aggression (Bennum, & Phil, 1983; Simeon et al., 1992; Hillbrand, 1995). One third of violent individuals have been estimated to engage in self-destructiveness, and 10-20% of self-destructive individuals engage in violent behaviors (Plutchik & van Praag, 1990). Some individuals report that their self-mutilation temporarily reduces anger (Wilkins & Coid, 1991; Kemperman, Russ, & Shearin, 1997).

A small number of rigorous studies test the association of parasuicide to anxiety. Anxiety disorders (in particular, post-traumatic stress disorder and panic) increase the risk of suicide attempts in the general population (Gould, et al., 1998) and specifically in individuals with comorbid BPD (Comtois, Brown, & Linehan, 1998). Severe anxiety/agitation and panic attacks are common experiences in the days preceding suicide, and the experience of panic may predict suicide better than panic disorder itself (Fawcett, Scheftner, Fogg, & Clark, 1990). Similarly, self-mutilators typically report unbearable anxiety and tension (that usually accompanies depersonalization or emptiness) that is often relieved soon after self-mutilation (Simeon et al., 1992; Wilkins & Coid, 1991, Kemperman, Russ, & Shearin, 1997; Simpson, 1975). One psychophysiological study supports the view that repetitive self-mutilation is maintained by the quick reduction of negative emotional arousal (Haines, Williams, Brain, & Wilson, 1996). Some individuals report that calmness is induced after their decision to commit suicide.

Some studies suggest that relief from anxiety is more strongly associated with self-mutilation than suicide attempts. In one study, tension preceded nonsuicidal cutting for 85% of patients whereas depression preceded cutting for 26%. In contrast, 52% reported high tension preceding overdose suicide attempts and 66% of patients reported that attempts were preceded by depression (Jones et al., 1979). Most of these patients that reported tension or depression prior to the act reported the emotion to be absent following the act. Similarly, overdose with low intent to die has been reported as primarily a method to escape tension, in contrast to high intent overdose (Parker, 1981).

Some studies have compared the association of various emotions to parasuicide. For example, in one study, patients reported that anger and anxiety were the emotions most relieved by self-mutilation, but that sadness also decreased (Kemperman, Russ, & Shearin, 1997). Subjects from two studies reported anger at self and tension relief as primary reasons for self-mutilation (Bennum & Phil, 1983; Roy, 1978). Considered together, two studies showed that at least 57% of overdoses are reported to occur when feeling lonely or unwanted (e.g., sadness and shame), 45% when feeling like a "failure" (e.g., shame), 44% when worried, and 38% when angry (Bancroft et al., 1976; Hawton et al., 1982). In the sample of parasuicidal subjects studied by Brown and Linehan (1999), approximately equal numbers of subjects reported intending to relieve anxiety, anger, and shame through parasuicide (each emotion endorsed by about half. A majority of self-mutilators explained their self-mutilation as self-punishment (e.g., shame), whereas only a minority of suicide attempters did so. Parasuicide was reported to be effective at relieving anxiety and moderately effective at reducing anger. Self-mutilation was rated as more effective than attempting suicide for relieving negative emotions. These subjects rated their parasuicide as very ineffective at reducing self-hatred and shame and highly effective at punishing themselves. Self-verification theory would predict that, unlike anxiety, self-mutilation would not relieve shame.

Among nonviolent psychiatric inpatients, suicide risk (as measured by the Suicide Risk Scale, with risk of violence partialed out) was correlated highly with sadness and fear/anxiety, but not with anger (Apter, Kotler, Sevy, Plutchik, Brown, Foster, Hillbrand, Korn, van Praag, 1991). In contrast, among violent incarcerated patients suicide risk was correlated with anger and fear/anxiety, but not with sadness. The only published study to measure facial expressions of basic emotions (but not shame) found that contempt was displayed much more in suicidal patients than depressed patients (Heller & Haynal, 1997).

SHAME AND PARASUICIDE

.Many clinicians and researchers have theorized that shame, guilt, self-hatred, self-blame, and self-punishment are important reasons for self-mutilation and suicide (e.g., Friedman, et al., 1972; Maris, 1981; Baumeister, 1990). Others have speculated that self-mutilation results from shame stemming from sexual abuse (Anderson, 1981; Shapiro, 1987; Leinbenluft et al., 1987; Walsh, 1987; Walsh and Rosen, 1988). Several converging lines of evidence suggest an association of shame to parasuicide. One study showed self-mutilators to have very high "intropunitive hostility" (i.e., anger at self; Brittlebank, Cole, Hassanyeh, Kenny, Simpson, & Scott, 1990), but some studies have failed to show this. Half of the subjects in two studies cited anger at self as a primary reason for their self-mutilation, citing anger at others much less often (Bennum & Phil, 1983; Roy, 1978). Patient's often describe their self-mutilation as self-punishment (Brown & Linehan, 1999). In one study (Rosen, 1976), near-fatal suicide attempters mostly gave reasons such as feeling alienated, rejected, worthless, and hopeless. Several prospective studies suggest a link of negative self-concept to suicidal ideation (Kaplan & Pokorny, 1976) and suicide attempts (Lewinsohn, Rohde, & Seeley, 1994). Suicide attempters hold a particularly negative self-concept (Neuringer, 1974), and negative self-concept has predicted suicide independently of depression and hopelessness (Beck & Stewart, 1989). In one study, guilt was not prospectively associated with suicide attempts (Paykel & Dienelt, 1971). Unlike shame, self-directed anger, and low self-esteem, however, guilt is not strongly associated with a negative self-concept (e.g., Tangney, Wagner, Fletcher, & Gramzow, 1992).

One study showed that for 38 subjects with BPD, high initial self-reported shame predicted (above and beyond anger and anxiety) increased urges to self-harm when discussing one's recent parasuicidal behavior (Brown & Linehan, 1996). These shame ratings obtained before treatment also predicted whether they engaged in self-mutilation during the first four months of therapy, although anxiety and anger did not predict parasuicide (Brown, Levensky, & Linehan, 1997). Eighty-four percent of high-shame subjects repeated parasuicide (mostly acts of self-mutilation) in contrast to 45% of those not high in shame. This difference was highly significant even when statistically controlling for the number of pretreatment parasuicide episodes and self-reported anger and anxiety. Furthermore, nonverbal expressions of shame and contempt before treatment predicted the number of subsequent acts of self-mutilation during therapy.

The association observed between self-invalidation and suicidal behavior implicates shame. The construct of perfectionism overlaps with Linehan's description of self-invalidation and predicts suicidal ideation and suicide threats (Hewitt, Flett, & Turnbull-Donovan, 1992/1994; Hewitt, Flett, Weber, 1994; Baumeister 1990/1991). Individuals who hold extreme and rigid standards of performance often experience an extreme sense of failure and shame after even minor mistakes or setbacks.

Shame may explain the relationship observed parasuicide and other disorders. Depression, for example is associated with shame (Andrews, 1995), and depression predicts suicide attempts in BPD (Comtois, Brown, & Linehan, 1998; Fyer, Frances, Sullivan, Hurt, 1988). Depression and shame both often result from blaming one's character deficits, in contrast to when one blames one's own specific behavior which more often leads to guilt (Janoff-Bulman, 1979; Tangney at al., 1992). Body shame, in particular, is a vulnerability to major depression. Furthermore, on self-report depression scales, self-mutilators, suicide attempters, and BPD individuals endorse high numbers of items reflecting shame and self-contempt, suggesting a shame-based depression (Bennum & Phil, 1983; Maris, 1981). Shame can also explain the consistent relationship found between anger and parasuicide. Studies suggest that shame and humiliation often initiate anger toward others that are perceived as disapproving of (or "shaming") the individual (Tangney at al., 1992). Similarly, shame may explain why self-mutilation is associated with eating disorders (body shame; Walsh, 1987; Favazza & Conterio, 1989) and sexual abuse (e.g., Walsh, 1987).

Although prior studies seem to suggest that almost every negative emotion is in some way related to parasuicide, it is difficult to know which studies have identified factors that are necessary for chronic parasuicide, factors that add to risk but are not necessary, or (sets of) factors that are sufficient for parasuicide to occur. The factors that explain the emergence of parasuicide may also differ from those that explain the repetition and chronicity. Similarly, the emotions that have been shown to differentiate parasuicidal and nonparasuicidal groups may not be useful in predicting repetition for individuals within a parasuicidal group. Unfortunately, most of these studies are based only on self-report and many did not specify diagnosis. Variations in definitions and sampling strategies, choice of comparison groups, timing of interviews in relation to the parasuicide episode, and the measures used also create difficulty in interpretation.


DIALECTICAL BEHAVIOR THERAPY FOR SELF-MUTILATION

Dialectical Behavior Therapy (DBT) utilizes behavioral and cognitive strategies to reduce self-mutilation and other parasuicidal behaviors (Linehan, 1993). These strategies include practical methods designed to directly suppress or preclude parasuicidal behavior as well as more indirect methods that attempt to reduce vulnerability to parasuicidal urges. The practical methods to suppress the behavior include such methods as helping patients to avoid events that trigger parasuicidal impulses (e.g., stimulus control), blocking access to lethal means, increased supervision by others, and other external controls. In DBT, emotional dysregulation is conceptualized as a significant proximal risk factor for parasuicide, and is therefore a primary target to reduce vulnerability to parasuicide. Unlike some risk factors, emotion dysregulation can be identified through verbal and nonverbal behavior and is potentially amenable to change.

Standard DBT is a one-year outpatient treatment including weekly primary individual psychotherapy and skills training sessions. The primary therapists work on the major maladaptive behaviors targeted in DBT (e.g., parasuicide). Skills training follows a structured psychoeducational format (typically in a group) to address deficits in the ability to regulate emotions, tolerate distress, and interact effectively with others. Whereas skills trainers aim to increase patients' behavioral repertoire, primary therapists work to strengthen and generalize patients' skillful behavior, often through telephone consultation with the patient. Therapists receive ongoing supervision and support from a DBT consultation team. The individual psychotherapy in DBT provides the relationship and context in which patients use new skills to gain control over self-harm and suicidal behaviors.


EMPIRICAL SUPPORT

The first randomized controlled clinical trial of DBT (Linehan, Armstrong, Suarez, Allmon, & Heard, 1991; Linehan, et al., 1994) demonstrated its efficacy for a sample of 47 chronically parasuicidal women meeting criteria for BPD. Subjects were assigned to either 12 months of DBT (n=24) or to treatment-as-usual in the community (TAU; n=23). Compared to TAU, fewer DBT subjects reported having parasuicided during the treatment year; they also reported fewer and less medically severe parasuicide episodes. DBT subjects also spent less time in psychiatric hospitals, dropped out of therapy less often, and were better adjusted interpersonally. Although DBT subjects reported significantly less anger, levels of depression did not differ between the two treatments. DBT patients also maintained significantly higher Global Assessment Scale scores than TAU subjects (Linehan, Heard & Armstrong, 1993c). DBT subjects showed greater improvement than TAU even when DBT was compared to only those TAU subjects who received stable individual psychotherapy during the treatment year and when number of hours of psychotherapy and of telephone contacts were controlled (Linehan & Heard, 1993). The superiority of DBT was largely maintained during the year following treatment. Results are presented fully in Linehan et al. (1991, 1994). A number of other studies evaluating DBT have been completed or are currently in progress at sites in the United States and in Europe (Koons & Robins, 1998; van den Bosch, 1991; Barely, Buie, Peterson, Hollingsworth, Griva, Hickerson, Lawson & Bailey, 1993; Miller, Rathus, Linehan, Wetzler, & Leigh, 1997; Telch, 1997). DBT has also been adapted to other settings such as inpatient psychiatric units (Swenson, Sanderson & Linehan, in press), day treatment (Swenson, Sanderson, Hoffman & Linehan, in press), and forensic psychiatric settings (Ball, McCann, Linehan, & Ivanoff, 1996, unpublished manuscript).


TREATMENT STAGES AND TARGETS

DBT involves a particular set of treatment stages and targets aimed at creating a life worth living for parasuicidal individuals. There are four stages of treatment reflecting roughly four stages of disorder or readiness for treatment. The pretreatment stage involves orienting the patient to the treatment, reaching agreement on the treatment targets, and getting patients committed to treatment. There are four fundamental sets of treatment strategies that remain conceptually intact across all applications of DBT: dialectical strategies, core strategies (validation and problem-solving), communication strategies (irreverent and reciprocal communication) and case management strategies (consultation-to-the-patient, environmental intervention, and supervision/consultation with therapists).

Patients must agree that reducing all types of parasuicide, including self-mutilation, is the first priority in therapy. Patients are also asked to commit to not killing themselves while they are in DBT. Once an individual enters the treatment phase, the patient's current behavior and goals and the DBT hierarchy of targets determines the treatment agenda. When patients fail to stay committed to stopping parasuicide, the "pre-treatment" phase may be reinstated to enhance commitment.

The first treatment stage in DBT focuses on getting self-mutilation and suicidal behaviors under control. Although the dialectic that parasuicidal behaviors and wishes to live can co-exist is understood within DBT, treatment cannot progress until this behavior stops. Patients' explicit agreement to stop parasuicide must be maintained during this stage. Any parasuicidal behaviors, significant intent or urges to parasuicide, and significant changes in suicidal ideation are addressed in individual therapy immediately following their occurrence. Self-mutilation and suicide attempts are always the highest priority in session and are never ignored. Other primary behaviors to decrease in stage one include violent behavior, behaviors that interfere with therapy, and other behaviors that interfere with quality of life. Other goals of stage one are forming and maintaining a strong therapeutic relationship, and increasing skills to regulate emotion and behavior.

Weekly diary cards are used to monitor targeted problems including parasuicide, urges to parasuicide, suicidal ideation, drug use, binge eating, and general misery, as well as to monitor use of DBT skills. The session agenda is set at the beginning of each individual session based on the diary card and the patient's current level of parasuicidal urges. Using diary cards is likely to improve the accuracy of reported behaviors and feelings compared to traditional narrative reports that rely on memory. Diary cards provide specific detail about the timing and relationship between self-mutilation and other dysfunctional behaviors and emotions, and the level of mood that can be tolerated without resorting to self-mutilation. Further, the diary card helps teach patients to be objective observers of their behavior and emotions, and help monitor their progress.

By the end of Stage 1, patients describe their state as one of quiet desperation. Primary targets of stage 2 involve increasing patients' tolerance for emotional experiencing, and reducing their vulnerability to negative emotions. Post-traumatic stress responses resulting from early life events are often treated in stage 2 with exposure therapy to facilitate emotional processing of traumatic events. These targets are not addressed until the patient demonstrates behavioral control because intense emotions generated by exposure may re-activate major dysfunctional behaviors. A strong commitment to not parasuicide and the basic skills to cope with emotion dysregulation are necessary for Stage 2 work. Stage 3 targets include self-respect, achievement of individual goals, and ordinary problems in living. Stage four targets issues such as a sense of incompleteness, lack of sustained joy, connection to others, and spirituality.


THERAPY STRATEGIES

Based on the primary assumption that the lives of BPD individuals with BPD are currently unbearable, DBT places great emphasis on change. Cognitive-behavior therapy procedures are the main methods of change in DBT. Numerous research studies have shown that these procedures effectively change maladaptive thought, emotions, and behavior for a variety of psychological disorders. Without such powerful change strategies it is unlikely that therapy will help these severe multi-problem patients make significant or lasting changes. Just as standard cognitive therapy for depression relies more heavily on behavior therapy than cognitive therapy when patients show severe behavioral impairment during severe episodes of major depression, DBT relies more on behavior therapy than cognitive therapy. The exact balance between cognitive and behavior strategies, however, depends largely on the controlling variables of the problematic behavior for each particular patient.

In DBT acceptance strategies are also considered very important when treating severely parasuicidal patients. These strategies help keep patients from dropping out of therapy prematurely, and help teach patients self-validation and acceptance. Furthermore, if therapists fail to validate sufficiently many patients fail to comply with cognitive-behavior therapy procedures that require effort or involve discomfort. Patients are also taught that to make important changes in life it is necessary to accept (though not approve of) those things that need to be changed. Pure acceptance and validation in therapy, however, would not likely be sufficient to change such serious problems. The conceptual division between change and acceptance strategies illustrates the fundamental dialectic that the DBT therapist must balance in treating chronically parasuicidal individuals. This balance is not necessarily achieved through equal amounts of acceptance and change strategies. Strategies are balanced according to the needs of the current situation. For example, therapists may choose to only use enough acceptance strategies as is necessary to keep the patient collaborating with the change process.

The treatment strategies in DBT can be conceptually divided into the following basic categories: dialectical strategies, core strategies (validation and problem-solving/behavior therapy), communication strategies (irreverent and reciprocal communication) and case management strategies (consultation-to-the-patient, environmental intervention, and supervision/consultation with therapists).

DIALECTICAL STRATEGIES.

Dialectics involves the reconciliation of apparent opposites. A dialectical philosophy of treatment is central to DBT and permeates DBT at many levels. Dialectics function as a frame for therapy by providing a worldview, a theory of disorder, a style of persuasive communication. At one level, dialectics involves balancing acceptance of patients as they are now with a push for change. Self-mutilation, for example, may be viewed as helping patients cope with their painful emotions (i.e., valid or understandable), while also treated as a destructive act that contributes to long-term suffering and needs to be changed. Dialectics also involves offering both acceptance and change-focused solutions for patients' problems (e.g., environmental change, acceptance of adverse situations, regulation of emotional reactions, and acceptance of emotions). Additionally, giving help to the patient is balanced with requiring the patient to help herself/himself.

Another basic dialectic in DBT involves alternating between a communication style that is reciprocal (i.e., warm and empathic), and one that grabs attention with irreverence, confrontation, or humor. At another level, dialectics describes the use of stories, metaphors, paradox, and philosophizing. An aim of using dialectical philosophy in therapy is to increase patients' comfort with inconsistency, ambiguity, and change. DBT therapists consistently highlight paradoxes and contradictions with statements such as "on the one handand on the other hand." Thus, dialectics can function as cognitive restructuring to counteract dichotomous thinking. Metaphors are often used for persuasion (e.g., accepting responsibility for pursuing help-"if you were hit by a car and broke your leg, would you refuse to have it set because it wasn't your fault the car hit you?") for validation (e.g., the difficulty of pursuing treatment-"climbing out of hell on an aluminum ladder"), and for didactics. Dialectical persuasion refers to deliberately confronting patients with the contradictions between their behavior and ultimate ends or goals. The devil's advocate is a paradoxical technique of persuasion used to strengthen the patient's commitment to therapy.

Dialectical philosophy emphasizes interconnectedness and wholeness; all things are seen as inherently heterogeneous and comprised of opposing forces that synthesize to produce change and to facilitate the construction of truth over time. Change is therefore viewed as a fundamental aspect of reality. Analyzing individual components of a system is useful only if each component is related to the whole. To fully understand a behavior, it is necessary to understand the context in which behavior occurs.

VALIDATION STRATEGIES.

Validation Strategies include active listening, empathy, and explicit verbal statements (Linehan, 1997). Explicit validation includes verbal statements that patient's behavior is normal or effective in relationship in a particular context as well as direct acknowledgement that something is true or valid. Implicit validation occurs when therapists' behavior indicates that they accept something about the patient as valid (e.g., "Tell me more about that. Let's figure this out"). Facial expressions can also communicate empathy and validation. Such behaviors function as validation yet does not validate through explicitly verbal statement. Encouraging and believing in the patient (i.e., "cheerleading") can function to communicate that the patient is a valid person (e.g., if a patient says "I can't do it", the therapist replies, "I know you think you can't do it, but I have complete faith that you really can do it"). The therapist can search for and communicate the valid elements of even the most extreme dysfunctional behaviors (i.e., the "kernel of truth"). The dialectical dilemma is to find how the apparently unreasonable is in some way reasonable. By validating in this way patients are more likely to not feel blamed and therefore to collaborate in therapy.

It is important to remember, however, DBT therapists are proscribed from validating aspects of patients that are not truly valid. Although major dysfunctional behaviors may be validated in some ways, these behaviors are still treated as dysfunctional. Several situations illustrate this dialectical dilemmas: when a dysfunctional behavior makes sense historically, but not currently; when behaviors makes sense in terms of short-term, but not long-term, goals; when patients' behaviors reflect valid conclusions from invalid assumptions; and when a response is normal or understandable, but not effective. Similarly, a focus on the patient's capabilities is balanced with a focus on the patient's limitations and deficits.

The most basic forms of validation are active listening (Level one) and accurate reflection and description of patients' statements and behaviors (i.e., paraphrasing; Level 2). A more difficult form of validation involves articulating experiences that the patient is not verbalizing or expressing directly (usually termed "empathy"; Level 3). Another form of validation involves explaining the unique responses of the patient by identifying learning experiences, beliefs, and/or biological factors that make the those current responses inevitable, and therefore understandable (Level 4). Validating the present, that is identifying events in the current environment that make current responses "make sense" (e.g., serve a function or are normal), is considered a very useful form of validation (Level 5).

Radical Genuineness (Level 6) involves treating the patient as a valid person rather than a mental patient (i.e., an invalid). Radically genuine therapists respond to the patient as a person of equal status rather than responding to the patient from a prescribed therapist role. Lack of radical genuineness also involves treating patients as fragile, through an overly soothing voice tone (i.e., "sounding like a therapist"), by treating them as if they are unable to solve their problems, or by failing to fully acknowledge obviously dysfunctional behaviors as problematic. This "fragilizing" may encourage (and reflect) a perception of the patient that militates against the patient becoming anything other than a mental patient (e.g., a self-fulfilling prophecy). Thus, the DBT therapist does not actively intervene when patients are able to solve their own problems. Radical genuineness stems from a belief in the patient's ultimate capacity to change and move towards ultimate life goals.

Radically genuine therapists are likely to respond in ways that naturally reinforce or punish behaviors. When radically genuine therapists observe dysfunctional behaviors or improvements in session they are likely to respond spontaneously by amplifying their natural responses (e.g., their excitement or disappointment). When therapists establish strong relationships with their patients these natural reactions are likely to strongly influence their patients' behavior (cf. Kohlenberg & Tsai, 1991). Since validation is usually very reinforcing to patients, it can be used to strengthen the therapeutic relationship to establish the therapist as a potent reinforcer. Additionally, radical genuineness can facilitate generalization of learning by creating a therapeutic relationship that is more like a "real" relationship, and thus providing a context for learning that more resembles life outside of therapy.

Therapists who view patients as fragile often validate and soothe patients more following self-mutilation. Doing so, however, can inadvertently strengthen self-mutilation. Historically, many individuals have learned that self-mutilation is one of the few behaviors that effectively elicits validation or help from others. For similar reasons, increasing sessions because of self-mutilation behavior or threats can reinforce the behavior. One solution to this problem is to make sure validation and warmth (and therapist availability) are abundant before patients get to the point of self-mutilation or to make these things contingent upon adaptive behaviors.

Finally, validation is also useful in treating self-mutilating patients because it often helps to regulate patients when emotions interfere with therapy. Specifically validation helps the patient regulate anger, contempt, and shame. Validation also teaches patients how and when to validate, accept, and trust themselves (e.g., how to reduce excessive self-blaming).

PROBLEM-SOLVING STRATEGIES.

Problem solving strategies are applied to each instance of self-mutilation and other major dysfunctional behaviors. Behavioral analysis is used to test hypotheses about the factors contributing to self-mutilation. The first step in a thorough behavioral analysis is to assess suicide intent. Sometimes patients cut themselves with the intent to die. The next task is to identify the internal and external events preceding self-mutilation, the context, and the events following the problem behavior. A "chain analysis" is a very precise behavioral analysis that is used to obtain a detailed step-by-step description of the events, emotions, cognitions, and overt behaviors that precede a single episode of self-mutilation. The consequences of self-mutilation are also examined, to understand its function for the patient.

Once the therapist understands for a particular patient the factors that trigger self-mutilation, increase the vulnerability to the trigger, and reinforce the self-mutilation, the task becomes to determine how the basic change strategies can address those factors. The therapist can help the patient structure the environment to make sure that 1) the events that elicit problematic emotions, thoughts, or behaviors get prevented or avoided, 2) dysfunctional behaviors get blocked and not reinforced, or 3) adaptive alternative behavior gets reinforced rather than punished. Sometimes prolonged and repeated behavioral analyses are sufficient aversive consequences to reduce self-mutilation. Skills training is used to address lack of ability to produce more adaptive responses to skillfully resolve, accept, or cope with problem situations. Cognitive modification addresses dysfunctional beliefs and assumptions that contribute to self-mutilation. Exposure-based strategies are used to reduce vulnerability to anxiety, shame, or other emotional responses that directly lead to self-mutilation or interfere with adaptive problem-solving. A great emphasis is placed on directly treating relevant dysfunctional emotions, cognitions, and behaviors that occur in the session (Kolhenberg and Tsai, 1991). Thus, if a therapist notices a patient looking down to avoid eye contact, excessively using euphemisms, and speaking softly, the therapist can ask if the patient is experiencing shame, and if so, determine the cue by identifying the exact moment that shame was elicited. When these behaviors are identified, the therapist attempts to get the patient to emit a low-probability adaptive behavior in place of the dysfunctional behavior (i.e., "dragging out new behavior" or "opposite action"). Doing so provides chances for the behavior to get reinforced and allows opportunities to practice in the presence of relevant cues (e.g., emotion). Thus, if a patient self-invalidates during the session, the therapists instructs the patient to "redo" the sequence of behavior by self-validating or by speaking in a nonjudgmental and non-invalidating way.

Exposure Strategies.

Because chronic emotion dysregulation is so strongly associated with self-mutilation, exposure strategies are used in DBT to reduce emotion vulnerability. Considerable empirical evidence demonstrates that exposure strategies effectively reduce high-arousal emotions such as anxiety and anger (e.g., Foa & Rothbaum, 1998). Most of these studies have shown that strategies such as systematic desensitization lead to long-term reductions in anxiety and fear. Behavioral analysis is used to determine for an individual which emotions elicit the urge to self-mutilate, and which cues elicit the emotions. If the extreme emotional reaction is the result of a correctly estimated severe consequence (i.e., an "unconditioned" stimulus such as the actual current threat of being assaulted) then the therapists would seek to reduce the threat rather than attempt to reduce the emotion through exposure. However, if the extreme emotion results from a cue that is actually associated with very little threat (i.e., a "conditioned" stimulus) then exposure strategies would be used to reduce the emotion (termed "non-reinforced exposure"). Exposure therapy involves ensuring that patients maintain prolonged attention on the emotion-eliciting cues so that patients can learn that no ultimate feared consequences result, and that they can successfully cope with the cues and the emotions (see Foa & Rothbaum, 1998). Therapists usually present the cues by getting patients to 1) imagine events that resemble avoided situations, 2) engage in avoided behaviors, or 3) approach and stay in actual situations that they frantically try to avoid. Therapists can also expose patients to the cues by discussing in detail the emotion-eliciting events (e.g., conducting a behavioral analysis) and the ultimate feared consequences (catastrophes) of the feared situations. To ensure that exposure actually occurs and that patients actually process the corrective information about the true threat, therapists must be sure that the emotion is elicited when the cue is presented. DBT therapists enhance exposures by coaching patients to practice mindfulness and acceptance of both cues and responses. Ultimately, exposure therapy will be most effective if all forms of avoidance are prevented including distraction and dissociation and the patient perceives some self-control over exposures. Exposure therapy is ended when the patient responds to the emotion cues with manageable levels of emotion.

After a bout (or session) of exposure to cues, the therapist can elicit opposite-to-emotion action, that is, get the patient to practice strategies to reduce the emotions elicited during the exposure (cf., Foa & Rothbaum, 1998). Thus, the patient can practice emotion regulation skills such as relaxation. Or the patient can practice reciting adaptive thinking to challenge any dysfunctional thinking that occurred during the exposures. Practicing opposite-to-emotion action following exposures provides opportunities to learn skills in the presence of relevant cues (e.g., emotion).

Exposure strategies are used informally during Stage 1 to decrease emotional responses to current generalized cues (e.g., loud noises, darkness, men) and to increase tolerance to current negative emotions. The use of exposure is more structured and prolonged during Stage 2. Exposure to cues directly associated with traumatic childhood events is usually always postponed until Stage 2. During these exposures, the patient is presented with the conditioned cues that elicit the emotions related to the traumatic experience. These cues could include memories, thoughts, emotions, behaviors, or physical reminders.

Since shame is often an emotion contributing to self-mutilation, the exposure model has been adapted in DBT to reduce vulnerability to shame. The general exposure strategy for shame is to get the patient to stop hiding personal characteristics or behaviors that elicit shame. If there is actually a significant threat that the patient will get rejected by her reference group or if the patient's shame-eliciting behavior has violated her own moral values, then the patient is instructed to repair the behavioral transgression, correct the personal characteristic, or find a more accepting reference group. Specific exposure procedures include: 1) talking about the things that make the patient feel ashamed, 2) having the patient physically reveal personal characteristics that elicit shame, and 3) having the patient engage in behaviors that elicit shame. If shame is detected in the session the therapist can determine the cue that elicited the shame and then repeatedly present this cue until the patient responds to it with less shame. Additionally, the patient is instructed to not look ashamed while in the presence of shame cues. Avoiding eye contact, speaking in a barely audible voice, and describing shame-eliciting personal characteristics with euphemisms are all forms of avoidance to stop during exposure procedures. Instead, the patient is encourage to "act as if" she is not ashamed while doing what makes her feel ashamed (i.e., "opposite action"). Psychology experiments have shown that when individuals with low self-esteem convincingly role-play having high self-esteem, self-esteem can actually increase (Rhodewalt & Agustsdottir, 1986). Similarly, the patient must refrain from self-invalidation during shame exposure since self-invalidation is a shame-congruent behavior can strengthen shame reactions.

It is critical for the therapist to obtain a strong commitment to exposure therapy since it is such an aversive form of therapy. Thus, it is important to make sure the patient fully understands what the exposure procedures will entail and the rationale for them. It is also helpful to provide ample encouragement and validate the difficulty of exposure tasks. Similarly, it is important to reinforce the patient for engaging in exposures that approximate full exposures when first beginning exposure therapy for a particular set of emotion cues. A strong therapeutic alliance can also help ensure that patient will agree to and comply with exposure therapy.

Skills Training.

DBT assumes that many of the problems experienced by patients who self-mutilate are due to a combination of high emotion vulnerability and behavioral skill deficits; that is, the skills to regulate negative emotions were never originally learned. For this reason, DBT emphasizes building skills to facilitate behavior change and acceptance. In standard DBT, skills are taught weekly groups following a structured psychoeducational format. Groups use a standard behavioral skills-building format and procedures, including modeling, instructions, behavioral rehearsal, feedback and coaching, and homework assignments. It is usually necessary to conduct skills training separate from the primary individual therapy since BPD patients frequently arrive for individual sessions in crisis leaving little time to learn skills. The individual therapist primarily serves as a skills coach to facilitate the generalization of skills. The four DBT skills training modules that directly target the behavioral, emotional, and cognitive dysfunction of self-mutilators are mindfulness, interpersonal effectiveness, emotion regulation, and distress tolerance.

Mindfulness is a psychological and behavioral translation of meditation skills usually taught in eastern spiritual practices. The goal of this module is attentional control and awareness. Wise mind is taught as a mindfulness skill to develop self-trust and good decision making. Patients are first taught to just observe, and then to describe external and internal stimuli. Learning to observe and label emotional states is a major goal of mindfulness training. Patients are taught to fully enter into positive experiences and effective behavior (i.e., "participate"). Finally, mindfulness is teaches patients how to be non-judgmental and effective, and how to stay focused on one thing at a time. These Zen ideals of mindfulness are taught in simple language.

Distress Tolerance focuses on the ability to accept both oneself and the current environmental situation in a non-judgmental manner. It is particularly useful in situations where nothing can be immediately done to change the environment and the patient needs a way to avoid engaging in impulsive behavior such as self-mutilation. Although distress tolerance encourages acceptance of reality it is very important to explain to patients that acceptance does not equal approval. Patients are taught to engage in activities that are effective distractions, improve the current moment, self-soothe through sensations, and consider the advantages and disadvantages of tolerating distress rather than acting impulsively.

Interpersonal effectiveness skills are similar to standard interpersonal problem-solving and assertion training. These skills include effective strategies for asking for what one needs and saying no to unwanted requests. Interpersonal effectiveness involves obtaining desired changes or objectives, keeping the relationship, and building and maintaining self-respect. Learning to effectively interact with others can help reduce emotion dysregulation by reducing interpersonal events that elicit negative emotions. Negative interpersonal interactions are common precipitants of parasuicidal behavior.

Emotion Regulation skills are defined as behaviors that: 1) directly increase or decrease physiological arousal associated with emotion; 2) reorient attention; 3) inhibit mood-dependent actions; 4) facilitate effective emotional experiencing; and 5) organize behavior in the service of external non-mood dependent goals. Emotion regulation usually begins with identifying and labeling current emotions. Patients are taught to identify emotions by observing and describing events that prompt their emotions, their interpretations of these events, their physiological responses, their emotionally expressive behaviors, and the effects of their emotions. Reducing vulnerability to negative emotions, increasing positive emotions, and decreasing emotional suffering is also targeted. Patients identify the emotions that precipitate their self-mutilation, and learns to monitor her specific vulnerabilities, e.g., sleep, eating, or substance use, activities. The patient is taught to develop pleasant, goal-oriented, competence-enhancing activities that reduce the emotions that precedes self-mutilation.

Telephone contact with the primary therapist between sessions is important for coaching the patient in using skills during crisis situations when the patient is experiencing very high emotional arousal. Patients are often offered skills coaching when they are not certain which skill to use or exactly how to apply the skill in a particular situation. DBT patients are typically encouraged to call before suicidal crises, or at least before they harm themselves. Consistent with the therapist's role as coach for adaptive behavior, this contact must occur prior to the self-mutilation or other parasuicidal behavior. If the patient has already engaged in self-mutilation, the "24-hour rule" stipulates that patients cannot have supportive phone contact with the therapist for 24 hours after self-mutilation with phone contact being limited to basic medical management only. This rule provides reinforcement for adaptive coping and aversive consequences for maladaptive behavior. Phone calls are also sometimes used as an opportunity to resolve misunderstandings and conflicts that arise during therapy sessions, instead of waiting until the next session to deal with the emotions.

SUMMARY.

DBT is an empirically-derived treatment for self-mutilation, suicide attempts, and other behaviors associated with BPD. Further studies are needed substantiate and improve the effectiveness of specific strategies within DBT for treating self-mutilation.