All treatments attempt to change or ameliorate the factors that are presumed to underlie or control clients’ problems. How this is done, however, can vary widely. There are two basic strategies for treating suicidal behaviors in clinical populations. The first strategy assumes that such behaviors are effects of some other underlying mental disorder. Treatment time and focus are allocated to treating the mental disorder, in the belief that its cure will in turn reduce the behaviors. Except to maintain life, no special modifications are made in the treatment of the underlying disorder. Reductions in suicidal behaviors are an indirect benefit of therapy. This approach is the model underlying most psychodynamic and biological approaches to treatment. The second strategy is to focus directly on the reduction of suicidal behaviors. In these treatments, reduction of these behaviors is an explicit treatment goal, and the specific behaviors are targets of intervention. When this strategy is pursued, therapy sessions focus on engaging the client in a discussion of current and immediately past suicidal behaviors (including suicidal ideation, threats, and communication), as well as of NSIB episodes. Explicit connections are then made to presumed underlying or controlling factors. This second model underlies most behavioral and cognitive-behavioral approaches to treatment. In this chapter we review the research on various treatments’ effectiveness for suicidal behaviors in both adolescents and adults. We group these approaches in the first two sections of the chapter by their strategy—indirect or direct targeting of the suicidal behaviors.
Unfortunately, data are very sparse regarding which treatments for primary mental disorders actually reduce the risk of completed suicide, suicide attempts, or suicidal ideation. The exclusion of highly suicidal individuals from most studies notwithstanding, investigators frequently include measures of suicidal behaviors in their outcome batteries, especially in studies of pharmacotherapy.
While lithium treatment has been correlated with reduced rates of suicide attempts and fatalities in adults diagnosed with bipolar disorders (Baldessarini & Jamison, 1999), and clozapine has been shown to reduce suicidal ideation and behavior among adults diagnosed with schizophrenia and schizoaffective disorder (Meltzer & Okayli, 1995), pharmacotherapy researchers have focused most of their attention on depressive disorders. Since mood disorders are the most common diagnoses related to suicide, researchers hypothesize that effective treatment of depression will reduce the incidence of suicide. Although this assumption makes intuitive sense, there are actually few or no empirical data to back up the assumption. Pharmacotherapy regimens that are more effective than placebo for reducing depression may or may not be more effective in reducing suicidal ideation or other suicidal behaviors (e.g., Beasley et al., 1992; Smith & Glaudin, 1992). To date, although there are many randomized clinical trials indicating that antidepressant medications reduce depression, there are no data from these trials suggesting that antidepressants reduce the incidence of either suicide attempts or completed suicide. For example, Buchholtz-Hansen, Wang, and Kragh-Sorensen (1993) followed 219 depressed inpatients who had previously been participants in multicenter trials of psychopharmacology. Not only were suicide rates higher than expected at follow-up, but there was no association between response to the antidepressant treatment and the suicide risk during the first 3 years of observation. Meta-analyses of pooled data from 17 doubleblind clinical trials comparing fluoxetine (n = 1,765) with a tricyclic antidepressant (n = 731), a placebo (n = 569), or both in the treatment of individuals with major depression showed no significant reductions in suicidal acts as a result of taking antidepressants (Beasley, Sayler, Bosomworth, & Wernicke, 1991; Beasley et al., 1992).
In studies where actively suicidal individuals were not enrolled, the failure to find significant treatment effects may be attributable to the low base rate of suicidal acts in a nonsuicidal population. That is, the treatment may not be powerful enough to make a large difference, or our present statistics may not be powerful enough to detect a small difference. However, the findings reported by Beasley et al. (1991) of a pooled incidence of suicidal acts of 0.3% for fluoxetine, 0.2% for placebo, and 0.4% for tricyclic antidepressants suggest that power may not be the problem. The general lack of follow-up, however, may be an important factor. In the studies analyzed by Beasley et al., results were reported for only the 5–6 weeks that individuals were active in the treatment protocol; no follow-up data were reported. Even if reducing depression does reduce risk of suicidal behavior, it is unlikely that effects would show up so quickly. An absence of significant effects could also be due to equivalent efficacy across many interventions with low-risk clients. That is, any active treatment may be equally effective at suicide prevention within a population at low initial risk for suicide.
Looking at the relationship of reducing depression to reducing suicidal behavior from the reverse direction, Linehan and colleagues (Linehan, Armstrong, Suarez, Allmon, & Heard, 1991; Linehan et al., 2006b) showed that DBT resulted in a significant reduction in suicide attempts and other intentional self-injury, compared to either treatment as usual (TAU) or community treatment by experts, did so despite being no more effective in reducing depression than the control conditions. (Depression improved in all treatments.) Similarly, Sakinofsky, Roberts, Brown, Cumming, and James (1990) found that improvement in depression, hostility, locus of control, powerlessness, self-esteem, sensitivity to criticism, and social adjustment was not related to reduced risk for repeated intentional self-injury over the next 3 months.
Recently the U.S. Food and Drug Administration ordered a “black box” label to be placed on all antidepressant medications, to indicate the potential for increased suicidal ideation and other suicidal behaviors among children and adolescents using these medications (Delate, Gelenberg, Simmons, & Motheral, 2004). These concerns have led many families of depressed and suicidal youth to pursue nonpharmacological interventions. Despite preliminary findings from the Treatment for Adolescents with Depression Study (TADS) that for non-suicidal depressed youth, the combination of fluoxetine and CBT was more effective at reducing depression than fluoxetine alone, CBT alone, or a placebo (TADS Team, 2004), concerns remain about prescribing antidepressant medications for depressed adolescents. Clinicians are divided as to whether antidepressants should be recommended as a primary intervention for suicidal multiproblem youth. Thus there is a glaring need for effective behavioral interventions for these youth.
Despite the incidence of completed suicide, suicide attempts, and suicidal ideation, there is remarkably little research on whether therapeutic interventions aimed directly at reducing suicide risk and associated behaviors are effective in achieving these aims. There are many books, articles, professional workshops, and legal precedents concerning treatment of suicidal behaviors, but very few of the recommended or required interventions have been subjected to controlled clinical trials. Thus, although there are standards of care for intervening with individuals at high risk for suicidal acts, there are few or no empirical data confirming that these standards of care are effective in preventing suicide or reducing the frequency or medical severity of suicide attempts.
Emergency inpatient admission for suicidal individuals has been examined in only two studies; neither found the inpatient admission to be more beneficial than outpatient treatment (Water-house & Platt, 1990; Huey et al., 2004). Among inpatient programs, no differences have been found between behavioral and insight-oriented treatments (Liberman & Eckman, 1981), or among cognitive therapy, problem-solving therapy, and supportive treatment (Patsiokas & Clum, 1985). A number of studies have examined whether access to emergency treatment without requiring current suicidal behaviors might be effective. No beneficial effects were found for adolescents ages 12–16 years who were given access to emergency inpatient admission (Cotgrove, Zirinsky, Black, & Weston, 1995), or for adults who were given access to on-call psychiatrists (Morgan, Jones, & Owen, 1993; Evans, Morgan, Hayward, & Gunnell, 1999).
Three studies evaluated the coordination of care following inpatient admission for self-inflicted injury, with negative results. Two (Möller, 1992; Torhorst et al., 1987) examined the impact of continuing with the same clinician from inpatient to outpatient treatment compared to TAU. Contrary to predictions, in both studies the proportion of repeated self-inflicted injuries was higher in the continuity condition. A third study examined coordination of care with a patient’s general practitioner. This involved sending a letter following emergency care for self-inflicted injury to notify the general practitioner of this emergency care; enclosed in this notification were expert consensus guidelines for appropriate follow-up, as well as a letter to send to the patient asking him or her to schedule an appointment. This was compared to a TAU condition. There were no differences in the rates of repeated self-injury.
Studies comparing clinical TAU to clinical outreach efforts have had mixed results. Three out of four studies have found in-person outreach effective in reducing repeated self-injury (van Heeringen et al., 1995; Termansen & Bywater, 1975; Welu, 1977; Chowdhury, Hicks, & Kreitman, 1973). Two studies found no beneficial effect of telephone outreach, whether this consisted of a simple “befriending” call (Wold & Litman, 19763) or a call to motivate suicidal individuals to attend or stay in treatment (Cedereke, Monti, & Öjehagen, 2002). Metaanalyses of intensive treatments plus outreach also did not find them to be effective (Hawton et al., 1998; van der Sande, Buskens, Allart, van der Graaf, & van Engeland, 1997).
The one exception to this conclusion is a very important intervention developed by Motto (1976), which consisted of sending nondemanding letters to 320 individuals who did not enter follow-up care after a hospitalization for depressive or suicidal states. Each letter was signed by the staff member who had interviewed the patient in the hospital; it consisted of a simple expression of concern, and invited a response if the patient wished to send one. Subsequent letters were individualized and included responses to comments previously received. Results showed that for the first 2 years, when contact with the participants was greatest, there was a lower percentage of suicides in the contact group than in the no-contact group; over the full 5 years of the study, the difference was not significant. It is important, however, that this difference was significant during the first 2 years following hospitalization, which is when suicides are most likely to occur (Motto & Bostrom, 2001). To date, this is the only intervention ever shown in a controlled trial to have a significant effect on completed suicide.
Various forms of CBT have been examined in randomized clinical trials. Two brief CBT interventions have been found more effective than TAU in reducing subsequent self-injury: 5 sessions of problem-solving CBT (Salkovskis, Atha, & Storer, 1990) and 10 sessions of CBT plus case management (Brown et al., 2005). However, a manual-assisted version of CBT, consisting of a brief CBT self-help manual plus up to 7 in-person sessions, was not more effective than TAU in reducing repeated self-inflicted injury (Evans et al., 1999; Tyrer et al., 2003). Only 60% of the individuals had at least 1 in-person session, however. Finally, no differences were found in two studies comparing skills training therapies to problem-solving therapy (Mc-Leavey, Daly, Ludgate, & Murray, 1994) and to supportive therapy (Donaldson, Spirito, & Esposito-Smythers, 2005). Several trials have evaluated various forms of brief crisis-oriented problem-focused treatments, which share many similarities with CBT. These approaches have not been found more effective than TAU (Hawton et al., 1981, 1987; van der Sande, van Rooijen, et al., 1997; Gibbons, Butler, Urwin, & Gibbons, 1978).
DBT is the only treatment to date with more than one clinical trial demonstrating effectiveness in reducing suicide attempts and NSIB. When compared to TAU for suicidal women meeting criteria for BPD, the proportion of clients in DBT who made suicide attempts and engaged in NSIB in the following year and a 1-year follow-up was lower than for those in the TAU condition (Linehan et al., 1991) and (Linehan, Heard, & Armstrong, 1993a). A replication trial (Linehan et al., 2006b) found that among chronically suicidal women meeting criteria for BPD, the proportion of individuals with a suicide attempt during the 1 year of treatment and 1 year of follow-up was half that of individuals randomly assigned to nonbehavioral community expert psychotherapy. In addition, the medical risk of those suicide attempts and NSIB that did occur were lower in DBT than in TAU or in treatment by experts. Emergency room and inpatient admissions for suicidal behaviors were also lower in DBT than in treatment by experts. Similar outcomes were found in a third 12-month randomized clinical trial in the Netherlands evaluating DBT for women with BPD, compared with TAU (Verheul et al., 2003). Reductions in suicide attempts and self-injury were found in four other randomized clinical trials conducted by three different research groups (van den Bosch, Koeter, Stijnen, Verheul, van den Brink, 2005; Koons et al., 2001; Turner, 2000; Verheul et al., 2003).
Two meta-analyses across all CBT trials have reached different conclusions about the efficacy of CBT for suicidal individuals. One meta-analysis (Hawton et al., 1998) included crisis intervention but not DBT as a problem-solving therapy and did not find that therapy to be effective. In contrast, a second meta-analysis separated crisis interventions from CBT and included DBT in the CBT category. This meta-analysis found CBT, but not psychosocial crisis intervention, to be effective (van der Sande, Buskins, et al., 1997).
Six clinical trials have evaluated the effectiveness of psychotherapies other than CBT. Bate-man and Fonagy (1999) demonstrated that an 18-month psychodynamically oriented partial hospitalization program for adults diagnosed with BPD was more effective than TAU at reducing suicide attempts and self-mutilation. These results, however, have not yet been replicated in a second trial or tested by an independent research team (Lieb et al., 2004).
Four sessions of interpersonal therapy provided in the home (Guthrie et al., 2001), and a “developmental group therapy” for adolescents (based on CBT, DBT, and psychodynamic therapies; Wood, Trainor, Rothwell, Moore, & Harrington, 2001), have been found to be more effective than TAU in reducing deliberate self-harm. It is unclear whether any of these “deliberate self-harm” behaviors were actually suicidal, however. Treatments not found to be more effective than TAU include 18 sessions of psychotherapy (including one home visit; Allard, Marshall, & Plante, 1992), four sessions of home-based family therapy for adolescents following self-injury (Harrington et al., 1998) and unspecified long-term therapy (i.e., one session per month for 12 months) (Torhorst, Möller, & Schmid-Bode, 1988).
Only four randomized treatment studies have explicitly focused on adolescent samples (Cotgrove et al., 1995; Harrington et al., 1998; Wood et al., 2001; and Huey et al., 2004). In the Cotgrove et al. (1995) study of adolescents age 15 and older, the investigators found no significant differences between the experimental (who were given access on request to inpatient care) and control (TAU) groups on measures of repeated self-injurious behavior.
In the second study, Harrington et al. (1998) randomly assigned subjects age 16 and greater to TAU or to TAU plus a short-term, action-oriented, home-based family intervention for adolescents who deliberately poisoned themselves. Investigators found that this intervention did not significantly reduce self-injurious behaviors. Thus brief family therapy alone for suicidal teens and families seems inadequate as a stand-alone intervention.
In a study of 63 adolescents ages 12–16 years, Wood et al. (2001) randomly assigned subjects either to “developmental group therapy” (containing components of CBT, DBT, and psychodynamic group psychotherapy) plus TAU or to TAU alone. These investigators employed Hawton and Catalan’s (1982), definition of deliberate self-harm: “any intentional self-inflicted injury, irrespective of the apparent purpose of the act” (Wood et al., 2001, p. 1247). The interesting finding here was that adolescents who participated in the experimental group therapy were less likely to be “repeaters” of self-harm at the end of the study than adolescents who received TAU alone (2/32 vs. 10/31; odds ratio = 6.3). Interestingly, those adolescents who participated in the experimental group were also less likely to use TAU, had better school attendance, and had a lower rate of behavioral disorders than adolescents given TAU alone. More sessions of the group therapy were associated with a better outcome, whereas more sessions of TAU were associated with a worse outcome. Hence more of the “right” type of treatment may sometimes be better. Lastly, the interventions did not differ in terms of their effects on depression, suicidal ideation, or global outcome.
In the fourth study, Huey et al. (2004) evaluated whether an intensive family- and community-based treatment called “multisystemic therapy” (MST; Henggeler, Schoenwald, Rowland, & Cunningham, 2002) could serve as a safe and effective outpatient intervention, compared to inpatient hospitalization of adolescents presenting with mental health emergencies. Participants were 156 males and females 10–17 years of age who were approved for psychiatric hospitalization because of suicidal ideation/planning or attempted suicide, homicidal ideation or behavior, psychosis, or other threat of harm to self or others. Subjects were randomly assigned to either the MST or inpatient psychiatric hospital condition. MST was delivered in each family’s natural environment, including the home, the school, and the community, by intensively trained therapists.
The study compared group differences from pretreatment to posttreatment and at a 1-year follow-up. Relative to inpatient psychiatric hospitalization, MST was efficacious at reducing the frequency of attempted suicide, although it did not seem to have a greater impact on suicidal ideation, hopelessness, and depressive affect. As the investigators point out, it is possible that the superior effects of MST may reflect a regression to the mean, since the group receiving MST had significantly higher rates of attempted suicide at pretreatment. In addition, two limitations of this study should be noted here. First, it appears as though the “attempted suicide” measure included NSIB; if so, this would call into question the interpretation of the results. Second, 44% of the MST-treated youth were also admitted for psychiatric hospitalization during the course of treatment, because of behavioral emergencies that could not be managed in a community setting. Hence nearly half of those in the MST condition additionally received the comparison treatment, which ultimately confounds the results. The investigators did not report the results after controlling for those subjects who received both treatments.
The results of controlled studies as a whole indicate that outpatient psychosocial treatments targeting suicidal behaviors directly, particularly CBT interventions, are effective in reducing the risk of future such behaviors in individuals identified as at high risk for them. However, only DBT has been replicated to date, as mentioned earlier in this chapter. In a previous review, Linehan, (1998) noted that 45% of clinical trials for self-inflicted injury excluded high-risk individuals. However, among the trials reported here that did not exclude such individuals, a significant reduction in self-injurious behavior was found. This finding highlights that individuals at high risk for suicidal behaviors and NSIB are able to benefit from outpatient treatments. In contrast, there are no data suggesting that inpatient treatments are effective at reducing suicidal behaviors and NSIB. Our review further suggests that the existing treatment research does not seem to support the premise of targeting suicidal behaviors indirectly by treating associated disorders. Rather, the treatment of suicidal behaviors and NSIB must be direct, although it must take into account the complexity and severity of co-occurring disorders in the patient.
DBT was originally developed specifically for chronically suicidal patients. These are patients who are unremittingly high in suicidal ideation, frequently threaten suicide or talk about taking their own lives, have difficulty articulating any reasons for living or staying alive, and may attempt suicide or engage in NSIB on multiple occasions. Although the treatment manuals describing DBT (Linehan, 1993a, 1993b) label it as a treatment for BPD, in fact the first drafts of these manuals never even mentioned BPD. The treatment and its theoretical underpinnings were originally developed to apply to suicidal individuals. The metamorphosis of the treatment into one aimed at BPD was due almost entirely to the substantial overlap between BPD and suicidal behavior (see Chapter 1).
DBT clinical trials have typically included older adolescents; however, to date there have been no analyses of treatment outcomes for adolescents alone. The success of DBT in reducing suicidal behaviors in adults diagnosed with BPD has led many clinicians to use it with adolescents who are also at high risk for such behaviors. As one of the few treatments to date showing efficacy for suicidal behaviors, the use of DBT for adolescents—either in its standard form or with adaptations for adolescents specifically—is widespread (Miller et al., 1997; Miller, Rathus, Dubose, Dexter-Mazza, & Goldklang, in press). However, promising preliminary research exists, and clinical experience is accumulating for implementing DBT with suicidal adolescents.
Two of us (Rathus & Miller, 2002) studied a version of DBT modified specifically for adolescents. Initial outcome data on our 12-week DBT program yielded promising results. In this quasi-experimental pilot investigation of suicidal adolescent outpatients with borderline personality features, we compared DBT (n = 29) to TAU (n = 82). The pilot sample was composed primarily of an ethnic minority (largely Hispanic) population. This is important, given the general dearth of behavior therapy and research with culturally diverse populations, as well as Hispanic adolescents’ status of having a higher suicide attempt rate than adolescents from other ethnic groups. Results indicated significant differences between groups: 13% of adolescents receiving TAU versus 0% receiving DBT were psychiatrically hospitalized during treatment, and 40% of those receiving TAU versus 62% receiving DBT completed treatment. There were no significant differences between groups in number of suicide attempts (9% for those receiving TAU and 3% for those receiving DBT). However, because the group receiving DBT was initially classified as more impulsive, was diagnosed with a greater number of Axis I disorders, and had a greater number of prior hospitalizations than the group receiving TAU, it is possible that the adolescents receiving DBT were actually at higher risk for suicidality. Thus the fact that they were no more suicidal than the control group during treatment is noteworthy. Unfortunately, we did not formally measure NSIB at that time. When we examined pre–post change within the DBT group, there were significant reductions in suicidal ideation and Axis I and II symptomatology.
H. Fellows (personal communication, December 11, 1998) employed an adaptation of DBT for adolescents similar to ours (Rathus & Miller, 2002). This adaptation resulted in significant reductions in adolescents’ use of costly treatment services. For example, prior to treatment the group had 539 inpatient psychiatric days, compared to 40 days during DBT treatment and 11 days during the 6-month posttreatment period.
Katz and colleagues (Katz & Cox, 2002; Katz, Gunasekara, Cox, & Miller, 2004) and Trupin and colleagues (2000) have begun adapting and evaluating DBT for adolescents in different settings, including inpatient and forensic programs. Preliminary results appear promising, although none of the DBT programs modified for adolescents have yet established efficacy through randomized controlled trials. It is clear that such trials are indicated to establish with full certainty whether DBT with adolescents is more effective than TAU. But in the meantime, we believe that DBT with appropriate modifications is highly promising for helping these adolescent clients.
Given the empirical evidence of DBT’s efficacy with adults, applying it to high-risk suicidal adolescents makes sense, as long as attention is paid to the developmental issues pertaining to this age group. Regardless of diagnosis, DBT skills directly target significant problems in emotion regulation and behavioral control. These are the same problems that characterize high-risk suicidal youth who are typically referred to emergency rooms and inpatient units.
As we discuss in greater detail in Chapter 3, DBT views suicidal behaviors as learned methods of coping with acute emotional suffering when no other coping options are available. The emotional picture of suicidal individuals is one of chronic, aversive emotion dysregulation. Those individuals who commit suicide are characterized by extreme dysphoria, often combined with high anxiety and panic (Fawcett, 1990). Generally, suicidal individuals are unlikely to have the ability to ameliorate or tolerate the emotional, interpersonal, and behavioral stresses in their lives. The cognitive difficulties found in studies of suicidal (primarily suicide-attempting) individuals include cognitive rigidity (Levenson & Neuringer, 1972; Neuringer, 1964; Patsiokas, Clum, & Luscomb, 1979; Vinoda, 1966), dichotomous thinking (Neuringer, 1961), and poor abstract and interpersonal problem solving (Goodstein, 1982; Levenson & Neuringer, 1971; Schotte & Clum, 1982). Impairments in problem solving may be related to deficits in specific (as compared to general) episodic memory capabilities (Williams, 1991), which have been found to characterize suicidal versus nonsuicidal individuals with mental disorders. In work at the University of Washington, we found that individuals with self-inflicted injuries exhibited a more passive (or dependent) interpersonal problem-solving style (Linehan, Camper, Chiles, Strosahl, & Shearin, 1987). Hopelessness is a strong predictor of both attempted suicide and eventual suicide (see Weishaar & Beck, 1992, for a review of this literature). Those who complete suicide are further characterized as indecisive and as having difficulties concentrating (Fawcett, 1990).
Suicidal behaviors can be viewed as problem-solving behaviors that function to remediate negative emotional arousal and distress either directly (e.g., by ending life [and presumably pain], putting an individual to sleep, or distracting him or her from emotional stimuli), or indirectly (e.g., by eliciting help from the environment), or as “inevitable” outcomes of unregulated and uncontrollable negative emotions. Although suicidal behaviors are not logically inevitable outcomes, paradigms of escape conditioning suggest that strong urges to escape or actual escape behaviors can be learned so completely that they become automatic for some individuals when faced with extreme and uncontrollable physical or emotional pain. Suicide, of course, is the ultimate escape from problems in one’s life.
Suicidal behaviors, from a DBT perspective, are a result of two interacting conditions: (1) Individuals lack important interpersonal, self-regulation (including emotion regulation), and distress tolerance skills and capabilities; and (2) personal and environmental factors inhibit the use of those behavioral skills the individuals may already have. These personal and environmental factors also interfere with the development of new skills and capacities, in addition to reinforcing the inappropriate self-injurious behaviors.
DBT directly addresses these conditions by (1) teaching suicidal clients specific skills for interpersonal effectiveness, self-regulation (including emotion regulation) and distress tolerance; (2) structuring the treatment environment to motivate, reinforce, and individualize appropriate use of the skills; (3) identifying and breaking up learned behavioral sequences that precede clients’ dysfunctional behaviors, and removing reinforcers for these behaviors; (4) providing treatment mechanisms to encourage the generalization of new skill capabilities from therapy to the life situations where they are needed; and (5) providing support for therapists treating high-risk suicidal individuals.
As we will see in the chapters that follow, DBT flexibly addresses multiple problems and suicide risk factors concurrently (see Table 2.1). Most other empirically supported treatments developed for adolescents (e.g., other forms of CBT, interpersonal therapy) only deal with one major problem at a time (e.g., depression, school avoidance, interpersonal problems); in fact, they typically exclude teens who present with suicidality or multiple problems (Miller et al., 1997). DBT focuses on adolescents who have comorbid mental disorders (often excluded from treatment research studies) and who are currently suicidal (often excluded from outpatient treatment programs). DBT with adolescents employs multiple modes—concurrent individual therapy, multifamily skills training groups, family therapy as needed, between-session telephone consultations (with both teens and parents), and consultation to therapists—to achieve its numerous functions. These include increasing clients’ motivation, skill acquisition, and skill generalization, and providing support and skill enhancement for therapists (to name a few). Also, DBT employs a target hierarchy; that is, the treatment is structured to address target behaviors according to their priority within each treatment mode, while at the same time allowing for flexibility within each session. For example, the therapist may begin the session by targeting a client’s serious threats to commit suicide during the past week, and simultaneously addressing the binge-drinking episode related to the threats. Later in the same session, as time permits, the therapist and client may then target mood lability and episodic hopelessness. Clearly, this treatment can address a multitude of suicide risk factors for adolescents.
TABLE 2.1. BPD Characteristics, Suicide Risk Characteristics in Adolescents, and Corresponding DBT Skills Training Modules
DBT also targets treatment noncompliance—a substantial problem with suicidal adolescents. In one study, Trautman et al. (1993) reported that 77% of adolescents who attempted suicide and presented to an emergency room subsequently failed to attend or complete traditional outpatient treatment. Other researchers have found that the average number of outpatient visits attended by adolescents who have attempted suicide is about five (Spirito et al., 1989). In addition to high treatment dropout rates, other forms of treatment noncompliance are rampant. Some adolescents attend sessions late; some are erratic with their therapy homework compliance; and some become noncollaborative during sessions, replying to all questions with “I don’t know” or “I don’t care.” DBT is unique in its attention to such “therapy-interfering behaviors,” ensuring that a therapist and client cannot overlook these behaviors when they arise.
As we will discuss later, DBT for adolescents can be adapted to target family dysfunction in addition to adolescents’ individual difficulties. Many family members of suicidal multiproblem adolescents have their own problems. At a minimum, communication problems exist between parents and teens. Outpatient adolescent DBT teaches skills to family members in multifamily skills training groups or in family skills training sessions. Family therapy sessions can also be woven throughout the treatment. These often include family behavioral analyses, and they target invalidation, ineffective use of contingency management, and skills deficits (particularly in the interpersonal realm) (Miller, Glinski, Woodberry, Mitchell, & Indik, 2002; Woodberry, Miller, Glinski, Indik, & Mitchell, 2002).
DBT is a cost-effective outpatient treatment, since clients receiving this treatment typically require fewer psychiatric hospitalizations and emergency room visits (H. Fellows, personal communication, December 11, 1998; Linehan & Heard, 1999; Miller et al., 2002; Potenza, 1998). Inpatient settings that previously exhausted numerous resources on suicidal clients with BPD, because they had little to no guiding treatment philosophy, find DBT feasible and useful in providing a principle-driven treatment that offers the necessary structure (Katz et al., 2004).
Relatedly, DBT helps providers maintain a compassionate stance toward their adolescent clients and their families by employing a biosocial theory to explain the etiology and maintenance of BPD. This theoretical framework has far-reaching implications for clients and staff alike. First, inpatient milieu staff members employing this framework approach adolescents in a less judgmental manner, which inevitably provides a more therapeutic context for the adolescents’ treatment. Moreover, the staff members interact more positively among themselves, which fosters a more positive work environment. Some adolescent programs indicate that they have had reductions in staff burnout after implementing DBT (Katz et al., 2004).
In sum, DBT with multiproblem suicidal adolescents makes sense, in the absence of established efficacy for any one treatment for this population. In the next chapter and throughout the rest of this book, we explain DBT and discuss its applications to work with multiproblem suicidal adolescents in a range of settings under a range of conditions.