24 Preventive Interventions and Treatments for Suicidal Youth

The last category of youth suicide prevention efforts focuses on indicated interventions and treatments that target those who have already shown signs of suicidal ideation and behavior. Such efforts seek essentially to prevent suicide attempts and suicide completion. The interventions and treatments described in this chapter differ widely in terms of the groups they target, the methods they use, and the settings in which they have been implemented.

The central underlying assumption of school-based programs for suicidal students is that subsequent suicidal thoughts and behavior can be reduced by enhancing protective factors, in particular students’ personal and social support resources.

The most comprehensive school-based programs are those developed and tested by Eggert, Thompson, and their colleagues (1994, 1999, 2000, 2001), as part of the Reconnecting Youth (RY) Prevention Research Program. The interventions are directed to students who are deemed to be at risk of dropping out of high school, based primarily on school attendance data and observations of teachers, counselors, and other gatekeepers. Such students have been reported to have multiple co-occurring problems that, in addition to school performance difficulties, include depression, suicidality, drug involvement, and tendencies toward aggressive and violent behaviors (Eggert, Thompson, Herting, & Nichols, 1994; Lewinsohn, Rohde, & Seeley, 1993).

The interventions are based on a theoretical model that rests essentially on improving the student’s personal resources, leading to an enhanced sense of personal control and self-esteem, improved decision making, increased use of social support resources, and reduced suicidal behavior. The early research involved systematic evaluation of a semester-long, school-based small group intervention called the Personal Growth Class (PGC). The intervention included life skills training using strategies of group process, teacher and peer support, goal setting, and weekly monitoring of mood management, school performance, and drug involvement.

Evaluation studies by Thompson, Eggert, and their colleagues (Eggert, Thompson, Herting, & Nicholas, 1994, 1995; Thompson, Eggert, & Herting, 2000) involved approximately 100 high school students at risk for dropping out of high school, as determined by a set of defined criteria, who screened positive for suicidal behavior (as discussed in Chapter 22). The students were randomly assigned to one of three conditions: assessment protocol plus one semester of PGC; assessment protocol plus two semesters of PGC; and assessment protocol only. Participants were assessed at baseline and at 5 and 10 months after the intervention. Participants in all three groups showed significant declines in suicidal behavior. Unlike the students who received the assessment protocol only, PGC participants showed significant improvement in self-perceived ability to manage problem circumstances. Also reported was a significant positive impact of both teacher and peer support in decreasing suicide risk behaviors and depression.

Thompson, Eggert, and their colleagues (2001) subsequently tested two additional school-based prevention programs based on the PGC: a brief one-on-one intervention known as Counselors Care (C-CARE), and a small group skills-building intervention program, Coping and Support Training (CAST), derived directly from the PGC program. Both interventions, compared to a usual care control group, were found to reduce suicide risk behaviors and depression, even at the 9-month follow-up assessment; CAST was most effective in enhancing and sustaining protective factors such as problem-solving ability.

The CARE intervention, expanded to include a parent intervention component, P-CARE (Randell, 1999), is being studied to determine the added benefit of this component to further reduce depression, anger, and suicide risk behaviors. Preliminary results suggest that C-CARE, coupled with the parent intervention, is associated with more rapid rates of decline in suicidal ideation, direct suicide threats, depression, hopelessness, and anxiety when compared to usual care (Thompson, 2003).

CAST, originally designed for high school youth, has been found to be effective with young adults up to age 25 in school settings and other institutions, and as a community-based program. The results compared with treatment as usual showed a significant decrease in positive attitudes toward suicide and suicidal ideation and a significantly greater and swifter decrease in depressive symptoms, feelings of hopelessness, anxiety for female patients, and anger-control problems; there was a significant and greater increase in feelings of personal control and problem-solving and coping skills (Benchmark, 2009).

These programs for suicidal students at risk of dropping out of high school have demonstrated efficacy in reducing suicidal behavior and depression. There is some indication that prolonged intervention results in the most positive outcomes related to suicide, although it is not clear whether these effects are due to repeated contact with the treatment or to the nature of the treatment itself. As is often the case with programs involving multiple components, it is difficult to identify which component is most responsible for the outcomes reported by these programs. Preliminary reports suggest that including parents in the intervention is particularly effective.

The target groups addressed by the studies of Eggert and Thompson may limit the generalizability of the findings to other populations of suicidal youth. From the outset, the focus of these programs has been on students at risk of dropping out of high school as principally defined through attendance records. There is some evidence that high school dropouts may come from more deviant and neglecting families and thus may not be representative of suicidal adolescents overall. In addition, the inclusion criteria for these programs are somewhat idiosyncratic in their use of gatekeeper identification of problematic students, which may limit the exportability and testing of the model.

In addition, these interventions were designed and implemented by highly skilled, university-based professionals, who devoted considerable attention to ensuring program fidelity, evaluating program results, and making improvements based on empirical findings. Although the results appear promising, replication of the program in schools that do not have such resources may be difficult. The community-based dissemination of the CAST intervention implemented and evaluated in three sites (Randell, Eggert, & Pike, 2001), addressed this problem. Although the study widens the scope of the program, since the data for adolescents among this group are not presented, it limits the conclusions that can be drawn about them from the study.

A considerable number of adolescents who make suicide attempts obtain some form of medical intervention (Grunbaum, Kann, Kinchen, et al., 2002), typically beginning in a hospital emergency department (ED). This suggests that the ED may be a prime location for initiating treatment programs for them. Numerous studies have documented, however, that young suicide attempters’ adherence to outpatient treatment recommendations made in the ED is poor, with over 15% never attending any recommended outpatient sessions and fewer than half attending more than a few sessions (Spirito, Plummer, Gispert, et al., 1992; Stewart, Manion, & Cloutier, 2001; Trautman, Stewart, & Morishima, 1993). Poor adherence has been attributed to ED factors such as long waits, repetitive evaluations, and poor communication by ED staff, and also to cultural factors, including the perception that mental health treatment is shameful (Spirito, 2003). The primary assumption underlying ED interventions is that improved treatment adherence will result in decreased suicidal behavior; thus, their goal is to develop mechanisms for engaging suicide attempters in the treatment process.

Rotheram-Borus et al. (1996, 2000) designed an intervention that targeted both the ED staff and families of Latino adolescent females who attempted suicide and followed participants over 18 months. Using videotapes and therapists, this program involved ED staff and families with a focus on encouraging participation in outpatient treatment. In comparison with patients who received family therapy alone, participants who received both family therapy and the ED intervention were found to adhere more frequently to the recommendation to attend a first treatment session. Families receiving the combined intervention also had more favorable outcomes in terms of maternal depression and general psychopathology, patient ideation, and parent-reported family interaction.

Spirito, Boergers, Donaldson, et al. (2002) also developed an adherence enhancement intervention to improve adolescent suicide attempters’ engagement in therapy. Treatment expectations, misperceptions, and reasons for treatment dropout were separately presented to adolescents and parents, along with a 1-hour intervention to facilitate problem solving around factors that might impede treatment attendance. After this ED intervention, telephone contacts were made separately to adolescents and their parents at at 1, 2, 4, and 8 weeks. Many service barriers were reported, such as delays in getting an appointment, being placed on a waiting list, and insurance and out-of-pocket expenses. Family barriers to treatment included parental emotional problems, transportation difficulties, language difficulties, and scheduling problems. The adherence enhancement program increased the number of sessions attended, although premature termination of treatment continued to be a problem. The program developers emphasized the importance of reducing service barriers for adolescents who have attempted suicide.

Hospitalization for suicidal behavior, though often securing the safety of the suicidal individual, is quite costly and not always beneficial. In an effort to decrease hospitalization rates and suicidality and improve functioning, Greenfield, Larson, Hechtman, et al. (2002) implemented the Rapid Response (RR) ED intervention for suicidal adolescents who were not considered to require immediate medical or psychiatric hospitalization. The intervention included family therapy, medication, and community intervention, as indicated. Hospitalization rates were decreased and outpatient therapy was initiated more rapidly as a result of the RR intervention when compared with standard care. In addition, adolescents receiving the intervention were less likely to be readmitted to the hospital during the 6 months after their visit to the ED. Neither hospitalization nor RR was found to prevent subsequent suicidal behavior or ED visits.

Asarnow, Baraff, Berk, et al. (2011) conducted a randomized controlled trial in which 181 suicidal youths at two EDs (ages 10–18) were individually assigned to one of two conditions: (1) an enhanced mental health intervention involving a family-based cognitive-behavioral therapy session designed to increase motivation for follow-up treatment and safety, supplemented by telephone contacts after ED discharge, or (2) usual ED care enhanced by provider education. Assessments were conducted at baseline and approximately 2 months after discharge from the ED or hospital. The primary outcome measure was rates of outpatient mental health treatment after discharge. Intervention patients were significantly more likely than usual care patients to attend outpatient treatment (92% vs. 76%; p = .004). The intervention group also had significantly higher rates of psychotherapy (76% vs. 49%; p = .001), combined psychotherapy and medication (58% vs. 37%; p = .003), and psychotherapy visits (mean 5.3 vs. 3.1; p = .003). Neither the ED intervention nor community outpatient treatment (in exploratory analyses) was significantly associated with improved clinical or functioning outcomes.

Currier, Fisher, and Caine (2010) sought to determine whether a mobile crisis team (MCT) intervention would be more effective than standard referral to a hospital-based clinic as a means of establishing near-term clinical contact after ED discharge. This objective was based on the premise that increased attendance at the first outpatient mental health appointment would initiate an ongoing treatment course, with subsequent differential improvements in psychiatric symptoms and functioning for patients successfully linked to care. In a rater-blinded, randomized controlled trial, 120 participants who were evaluated for suicidal thoughts, plans, or behaviors, and who were subsequently discharged from an urban ED, were randomized to follow-up either in the community via MCT or an outpatient mental health clinic (OPC). Both MCTs and OPCs offered the same structured array of clinical services and referral options. Successful first clinical contact after ED discharge occurred in 39 of 56 (69.6%) participants randomized to the MCT versus 19 of 64 (29.6%) randomized to the OPC (relative risk =2.35, 95% confidence interval [CI] 1.55–3.56, p < .001). No significant differences were detected between the groups in symptom or functional outcome measures at either 2 weeks or 3 months after enrollment, nor were any significant differences found in outcomes between participants who did attend their first prescribed appointment via MCT or OPC versus those who did not. However, divided (MCT vs. OPC, present at first appointment vs. no show), groups showed significant improvements but maintained clinically significant levels of dysfunction and continued to rely on ED services at a similar rate in the 6 months after study enrollment. Community-based mobile outreach was a highly effective method of contacting suicidal patients who were discharged from the ED. However, establishing initial post-discharge contact in the community versus the clinic did not prove more effective at enhancing initial symptomatic or functional outcomes, nor did successful linkage with outpatient psychiatric care. Overall, participants showed some improvement shortly after ED discharge regardless of outpatient clinical contact but nonetheless remained significantly symptomatic and at risk for reported ED presentations.

The results of the programs implemented to date suggest that some improvement in outpatient treatment adherence by young suicide attempters, as well as reduced hospital admissions, can be achieved by concerted efforts in the ED. Such efforts, however, require educating the ED staff regarding the suicide risks and treatment needs of young suicide attempters. Barriers to outpatient treatment appear to remain significant and difficult to surmount, even for the most cohesive and well-functioning families. It seems essential that ED interventions provide some continuity of contact with the youth beyond the initial ED visit, which will require additional staffing. While this may seem costly, the cost reductions associated with decreasing immediate and future hospital admissions are significant. The fact that there was no significant reduction in suicidal behavior or functional outcomes indicates the need for improvements in community outpatient treatment.

Research has pointed to a lack of posthospital treatment adherence among the many youth who are hospitalized on inpatient psychiatric units following serious suicidal behavior (Cohen-Sandler, Berman, & King, 1982; Spirito, Brown, Overholser, & Fritz, 1989). One result is frequent readmission to the hospital for repeated suicidality (Greenfield et al., 2002; Stewart et al., 2001). The key assumption of posthospitalization programs is that providing consistent support and improving adherence to aftercare recommendations will help to prevent future suicidal behavior.

The only fully developed program of this sort is the Youth-Nominated Support Team (YST) intervention, developed by King et al. (King, 2003; King, Preuss, & Krammer, 2001). This program was an outgrowth of the developers’ finding that family dysfunction and parental psychopathology significantly impact treatment adherence by suicidal youth after hospitalization (King, Hovey, Brand, Wilson, & Ghaziuddin, 1997). Concentrating on the high-risk period for suicidality immediately following psychiatric hospitalization, the program specifically targets poor treatment adherence and negative perceptions of family support and helpfulness.

Before leaving the hospital, program participants nominate specific adults from their home, school, or community to support them when they are released. YST conducts a psychoeducation session with these adults, then engages them in weekly consultations designed to improve their understanding of the suicidal youngster and how he or she can be effectively supported. A social network is encouraged among the adults, who typically come from diverse settings. The program is designed to supplement usual treatments.

The response to YST by participating youth and the nominated adults has been positive (King, 2003), with 80% of those nominated actually participating in the program. Positive effects have been reported for adolescent females, including reduced suicidal ideation and mood impairment. Similar benefits were not evidenced among male participants, although some described YST as having beneficial effects.

Since this intervention has only recently been implemented, it is too early to know whether the positive effects found among the suicidal girls will be translated into reductions of suicide attempts and repeat hospital admissions. It will also be important to identify the reasons underlying the lack of clear effects among male participants, and to incorporate the necessary programmatic changes. The fact that the program has been manualized will likely encourage its replication, while permitting independent assessment of specific program components.

As noted in Chapter 21, previous suicidal behavior is the most important factor associated with suicide risk among both adults and youth. Because recognizing that repetition of a suicide attempt vastly increases the risk of a fatal outcome (Arensman et al., 2001), considerable effort has been directed toward developing psychotherapeutic and other psychosocial treatment modalities to prevent subsequent suicidal behavior among identified individuals.

Most of the psychotherapeutic approaches for treating suicidal youth have employed variations of cognitive-behavioral therapy (CBT). The underlying assumption is that the primary focus of treatment should be the suicidal behavior itself, rather than the underlying psychopathology (Brent, Holder, Kolko, Birmaher, Baugher, Roth, et al., 1997; Harrington, Kerfoot, Dyer, McNiven, Gill, Harrington, et al., 1998; Stanley et al., 2009).

Rudd et al. provided the first description of a cognitive-behavioral skills group intervention designed to treat young adults with suicidal ideation or suicidal behavior (Rudd, Rajab, Orman, Stulman, Joiner, & Dixon, 1996). The intervention, an intensive 2-week program that participants attended for 9 hours per day, included an experiential affective group, psychoeducational classes with homework, and a problem-solving and social competence group. A variety of strategies such as behavioral rehearsal, role playing, and modeling were used to improve basic social skills and effective coping. Participants (N = 264) were randomly assigned to either the experimental intervention or a treatment-as-usual condition involving long-term outpatient treatment. In a 2-year follow-up, Rudd et al. found that participants in both groups showed significant reductions in suicidal ideation and behavior and experienced stress, and improvements in self-appraised problem-solving ability. The intensive time-limited intervention was found to be more effective than long-term treatment in retaining the highest-risk participants. Subsequent analyses showed that patients with psychiatric symptomatology experienced the most improvement in response to this intervention (Joiner, Voelz, & Rudd, 2001). The rate of suicide attempts at follow-up was not reported for either the experimental or control group, however, and conclusions were limited by high attrition rates in both the experimental and control group.

Harrington and his colleagues in Great Britain (Byford, Harrington, Torgerson, Kerfoot, Dyer, Harrington, et al., 1999; Harrington et al., 1998; Harrington, Kerfoot, Dyer, M, McNiven, Gill, Harrington, et al., 2000) developed a home-based family intervention for adolescents with a history of deliberate self-poisoning. This intervention used a cognitive-behavioral approach to address family dysfunction assumed to be related to the suicide attempt (Kerfoot, 1988; Kerfoot, Dyer, Harrington, Woodham, & Harrington, 1996), and to improve adherence to treatment by bringing it into the home. The intervention consisted of five highly structured sessions focusing on goal setting, reviewing the self-poisoning episode, communication, problem solving, and issues related to the family. The program included a treatment manual and videotape for training. This brief intervention was found to be effective primarily among those adolescents who were not seriously depressed and had less severe suicidal ideation, who made up about one third of the 85 participants (Harrington et al., 1998, 2000). Adherence and parental satisfaction with treatment were better for participants in this treatment relative to treatment as usual. The intervention was found to be no more costly than routine care alone (Byford et al., 1999).

Wood et al. have developed an additional psychotherapeutic variant, using Developmental Group Therapy (DGT) as an alternative to usual care for adolescents who have repeatedly attempted to harm themselves (Wood, Harrington, & Moore, 1996; Wood, Trainor, Rothwell, Moore, & Harrington, 2001). The group therapy format was hypothesized to be useful in providing an arena for working on social problem-solving and relationship skills that are often considered core to suicidal behavior. Using a developmental approach to address issues unique to adolescents, the intervention combines problem-solving and cognitive-behavioral interventions (Harrington, et al., 1998), Dialectical Behavior Therapy (DBT; Linehan et al., 1991), and psychodynamic approaches. An acute phase focusing on core themes (family and peer relationships, school problems, anger management, depression, self-harm, and hopelessness) is followed by a longer phase that concentrates on group processes. In interviews conducted about 7 months after treatment began, DGT participants reported engaging in less self-harm than did adolescents who received routine care, although depression did not appear to improve. Episodes of self-harm became less frequent as participants attended more sessions of the group therapy, whereas among those in usual care, self-harm behaviors were found to increase compared to baseline. DGT participants, particularly youth who had made multiple suicide attempts, also showed reductions in conduct problems. While DGT shows promise, Hazell et al. (2009) attempted to replicate its effectiveness in a sample of 72 12- to 16-year-olds by comparing this treatment to routine care and assessing youth 6 and 12 months after treatment. Adolescents in the DGT sample reported more self-harm at both follow-up periods. Changes specific to suicide attempts could not be determined since the groups were not equal at baseline. Given the inconsistent findings, further study of this intervention is warranted.

As part of the multisite Treatment of Adolescent Suicide Attempters study, Stanley et al. (2009) developed Cognitive-Behavior Therapy Suicide Prevention (CBT-SP). The goal of this intervention is to prevent future suicidal behavior among adolescents with depression and a previous suicide attempt. The treatment is manual-based and focuses specifically on suicidal behavior and therefore can be applied to adolescents with recent or imminent suicidal behavior. It can accompany other forms of treatment focused on other mental health issues. The stress-diathesis model serves as the theoretical basis for this treatment. The assumption is that an individual is at potential risk for suicidal behavior as the result of a variety of individual and environmental factors, and the individual’s risk for suicidal behavior may increase as part of his or her response to a significant stressor. The goal of the treatment is to identify triggers for suicidal behavior and apply cognitive and behavioral tools rather than suicidal behavior when the triggers are present. Parents are involved to the extent that they can learn specific strategies to help reduce suicidal behavior in their adolescent.

The 6-month treatment includes acute and continuation phases. The acute phase includes initial, middle, and end phases and may take up to six family sessions. During the initial acute phase of treatment a chain analysis is conducted to understand the precipitants for the suicidal behavior, a safety plan is developed that includes strategies the patient can use to stay safe, and psychoeducation is provided to the patient and family regarding suicidal behavior, depression, and removal of lethal means. The middle phase of acute treatment takes place after the immediate crisis is over and includes skills development for the individual and the family. Sessions have a specific structure, and the goal is to help the patient and family develop an effective approach to managing stressors and suicidal ideation and behavior if it emerges. The end of the acute treatment phase and the continuation phase are focused on preventing relapse. The suicide attempt and the chain of events, thoughts, feelings, and behaviors that led to it are reviewed, and the patient discusses how he or she can use the new skills under similar circumstances. This reinforces the newly developed skills and problem solving. The 12-week continuation phase focuses on titrating treatment, continued presentation of new skills, and practice of all learned skills.

The investigators reported that 72.4% of participants completed the acute treatment phase but only 28.6% completed the continuation phase. Based on a qualitative interview conducted with 42 participants, the treatment was found to be helpful by 100% of patients, 42% stated that they would not change the treatment if given the opportunity, and 86% said they would recommend the treatment to others. Those who recommended changes focused primarily on procedures such as rewards, assessments, and the research protocol itself. While most found repeated assessments to have no impact (30%) or a positive effect (19%), 30.9% found the repeated assessments to be mildly negative and almost 12% found the repeated assessments aversive. No participant reported increased suicidal ideation of the repeated assessments.

Therapist adherence was also assessed and varied depending on the module being considered. Overall, therapist adherence was good, although modules related to mobilizing supports and social skills for the patients were less likely to be presented.

A 6-month treatment similar to CBT-SP was developed and pilot tested in London (Taylor et al., 2011). Of an original sample of 37 12- to 18-year-olds, 16 adolescents with deliberate self-harm (13 of whom also had an incident of self-harm with suicidal ideation) completed the treatment. The treatment modules were (1) Getting Started; (2) Feelings, Thoughts, and Behaviors; (3) Coping; and (4) On You Go. A parent module was offered, but only two parents agreed to participate. The investigators were only able to obtain follow-up information on the treatment completers, thereby barring comparison with those who did not complete the program. Self-harm and suicidal behavior were greatly reduced, with only one participant having made a suicide attempt at 3-month follow-up. In sum, CBT with a focus on suicidal behavior has demonstrated potential for treating the adolescent with suicide attempts.

In another variation, Miller et al. (1997; Rathus & Miller, 2002) used a modification of DBT in their treatment of adolescent suicide attempters who demonstrated at least three features of borderline personality disorder. The intervention they developed consisted of 12 weeks of twice-weekly individual and family skills training. In one trial (Rathus & Miller, 2002), participants in the DBT group were found to have better adherence to treatment and fewer hospitalizations than those receiving treatment as usual, even though they had greater psychiatric comorbidity than control subjects. DBT treatment was also found to be associated with reduced suicidal ideation, symptom severity, and distress. Although suicide attempts were less likely in the DBT group compared to the controls, this difference was not found to be significant.

The feasibility of DBT with adolescent inpatients admitted because of a suicide attempt or strong preoccupation with suicide was tested with a 2-week inpatient adolescent DBT program and compared with patients with treatment as usual, which used a traditional psychodynamically oriented assessment and treatment model (Katz et al., 2004). Assessments of depressive symptoms, suicidal ideation, hopelessness, parasuicidal behavior, hospitalizations, ED visits, and adherence to follow-up recommendations were conducted before and after treatment and at 1-year follow-up for both groups. In addition, behavioral incidents on the units were evaluated. DBT significantly reduced behavioral incidents during admission when compared with treatment as usual. Both groups demonstrated highly significant reductions in parasuicidal behavior, depressive symptoms, and suicidal ideation at 1 year. The authors concluded that DBT can be effectively implemented in acute-care child and adolescent psychiatric inpatient units (Katz, Cox, Gunasekara, & Miller, 2004).

Fleischhaker et al. (2011) conducted a pilot study of DBT with 12 adolescent females with non-suicidal self-injury, eight of whom also had a suicide attempt. Participants received 16 to 24 weeks of DBT at an outpatient mental health clinic in Germany and were assessed 4 weeks and 1 year after treatment. Non-suicidal self-injury was reduced but persisted. No suicide attempts occurred during the 1-year follow-up period.

Interpersonal Psychotherapy for depressed adolescents (IPT-A; Mufson et al., 2004) has been shown to be equally effective as CBT in treating depression in adolescents (Horowitz et al., 2007). IPT-A works to reduce suicidal behavior by helping adolescents and their families resolve problems between them and deals with the affects that underlie the problems. Its approach is different than that of CBT, which treats maladaptive cognitions as the root of the problem. It is being tested for its ability to reduce suicidal behavior in adolescents.

Mentalization, an important component of DBT, served as the core of an intervention developed by Rossouw and Fonagy (2012) to treat adolescents who engaged in self-harm behavior, including suicide attempts. In a randomized controlled trial, Mentalization-Based Treatment for Adolescents (MBT-A) was compared with treatment as usual among 12- to 17-year-olds presenting at a community health center or hospital ED. To qualify, the youth had to have engaged in self-harm behavior (80% reported at least one suicide attempt). MBT-A was a year-long manualized treatment program that combined attachment theory–based psychoanalytic therapy with monthly family therapy sessions (MBT-F) that focused on impulsivity and affect regulation. The goals of the treatment include more accurate assessment of their own and others’ emotions and improved functioning in emotionally challenging situations. The assumption is that reduced self-harming behavior would result. Eighty adolescents were randomized to MBT-A or treatment as usual, and 97% met the criteria for at least one mental disorder. One third of the MBT-A families refused to participate in treatment. While both groups improved, the MBT-A group showed greater reduction in self-harm. By 12 months, depression was reduced for both groups, more so for those receiving MBT-A. It is noteworthy that the rate of self-harm behavior during treatment (MBT-A 43% vs. treatment as usual 68%) and the presence of depression (MBT-A 49% vs. treatment as usual 70%) were still significantly high. Thus, this approach showed reduction in self-harm behavior and depression but not remission.

Family psychotherapeutic approaches for adolescents who have made a suicide attempt have also been developed and are currently being tested. Attachment-Based Family Therapy (ABFT; Diamond et al., 2002), originally developed for the treatment of adolescents with depression, has been found to be effective in reducing suicidal ideation among adolescents who have made a suicide attempt and have at least one parent available for family treatment (Diamond et al., 2010). It has also been adapted and shows promise among lesbian/gay/bisexual/transgender youth who have made a suicide attempt (Diamond et al., 2012). ABFT assumes that problems with attachment, high levels of parental criticism, and lack of emotional attunement in families lead to the development of ineffective coping skills for dealing with stressors that may trigger depression and suicidal ideation and behavior. ABFT also assumes that these disrupted functions are available for reparation, and improved coping and lower risk for depression and suicide are possible. The targets of treatment include parental criticism and hostility, low motivation and engagement in the adolescent, ineffective parenting, family disengagement, and negative self-concept. Specific tasks are provided to assist family members in developing more functional and effective social and attachment skills. The development of these skills allows for improved coping and decreased depression and suicidal ideation and behavior.

A 12-week randomized control trial of ABFT compared with enhanced usual care was conducted with adolescents between the ages of 12 and 17 recruited from emergency and primary care hospital settings who were assessed to have significant suicidal ideation and depression and who had a parent available for participation in treatment (Diamond et al., 2010). A follow-up assessment was conducted at 24 weeks after baseline. While both groups demonstrated improvement with treatment, adolescents who participated in ABFT demonstrated quicker and greater decreases in suicidal ideation and depression, and the reduction of symptoms persisted at follow-up.

An analysis of the data from 18 families for whom treatment videotapes were available (Shpigel et al., 2012) provided evidence that parental factors improved early in treatment and adolescents’ attachment, anxiety, depression, and suicidal ideation improved over the entire course of treatment. These findings suggest that parental improvement plays a role in the suicidal teen’s improvement. ABFT has been adapted for LGBT adolescents with the addition of a module just with parents that focuses on their feelings about their child’s LGBT status, working toward acceptance, and reducing rejecting statements by parents (Diamond et al., 2012). In a pilot study with 10 LGBT adolescents with significant suicidal ideation, the teens reported significantly reduced suicidal ideation after 12 weeks of treatment. There is strong evidence that ABFT is an effective treatment for reducing suicidal ideation in adolescents with parents willing to engage in treatment.

Pineda and Dadds (2013) examined a parent intervention for parents of adolescents between the ages of 12 and 17 who were initially seen in the ED or mental health program in a hospital and had been identified as experiencing suicidal ideation or behavior or non-suicidal self-injurious behavior. The treatment, Resourceful Adolescent Parent Program (RAP-P), is part of a larger program that includes treatment components for the adolescent as well as his or her school. This study examined the impact of the parent program for 24 adolescents treated with RAP-P compared with 24 adolescents who were treated with usual care. The RAP-P program consisted of usual care for the adolescent plus a psychoeducational program for parents, followed by one to three family sessions. The data suggested a strong effect for the RAP-P program for decreasing scores on the Adolescent Suicide Questionnaire (ASQ; Pearce & Martin, 1994), and a mediation analysis suggested that the reduction in suicidal ideation was a function of change in family functioning as measured by the Family Assessment Device (FAD; Epstein, 1983).

The family approach to treating adolescents with suicidal ideation and behavior demonstrates great promise for families with parents willing to engage in treatment. Psychoeducation and effective family functioning appear to be important components for preventing suicide attempts and reducing suicidal ideation.

Results reported to date suggest the effectiveness of CBT and DBT interventions in improving social functioning and reducing suicidal ideation among suicidal adolescents, particularly those with mild to moderate depression and those with borderline features. In some cases, however, the outcomes of experimental treatments have not been substantially better than those obtained by comparison or standard care treatments. Treatments focusing specifically on suicidal behavior, its precipitants, and strategies for managing suicidal crises seem to fare the best. In addition, when families are available and willing to participate, family treatments have demonstrated efficacy in reducing suicidal ideation and behavior. However, many studies found that most parents are not willing or available to participate in treatment. Long-term effects of psychotherapy interventions on suicidal behavior, particularly suicide, have not yet been reported. Given that maladaptive cognitions and behaviors have likely developed over a long period of time, it is not clear that short-term psychotherapies will ultimately be found to be effective in reducing suicidal behavior.

As was noted in a previous chapter, there has been marked reluctance among college and university officials to specifically identify suicidal students or offer treatment services that specifically address this problem. One university-based treatment program has been based on the assumption that students who engage in suicidal threats or behavior will not voluntarily submit to a clinical assessment, and thus that such assessment must be mandated as a condition of the student’s continued enrollment at the university.

In the fall of 1984 the University of Illinois instituted a program designed to reduce the rate of suicide among its students. The program has a policy that requires any student who threatens or attempts suicide to attend four sessions of professional assessment—specifically, identified students are required to attend four weekly sessions at the University Counseling Center, during which the student receives a comprehensive clinical assessment and referral for additional treatment if needed (Joffe, 2003). Students who persist in suicidal behavior and refuse to accept counseling can be obliged to withdraw from the college.

Figures for the suicide rates for the 8 years before the project was instituted (1975–1983) were compared with the suicide rates for the 21 years from 1984 to 2004 (Joffe, 2008). The rate dropped 45% during the project period, from 6.9 per 100,000 students to 3.78 per 100,000. That rate is only half what it is at other Big Ten schools. Despite over two thousand reported incidents during this period (96 per year), only one student who participated in the Illinois program was forced to withdraw. Students who resist coming for the four sessions, however, are told that remaining at the university requires that they do so. Almost 10% of the students known to have threatened or attempted suicide withdrew voluntarily. Others have withdrawn for a variety of reasons, often in response to pressure from parents to do so. Few students who withdraw return to the college.

The data on college student suicides are limited at Illinois and all colleges by the lack of knowledge of students who committed suicide off campus. Students who die by suicide at home or elsewhere during the 4 years after the time they entered college are essentially dropouts and would need to be counted to be able accurately to determine the suicide rate. The Illinois program attempted to track these students with no success, although a few were reported by parents. This limitation of college student figures confirmed an earlier study of suicide rates at the same colleges (Silverman, Meyer, Sloane, et al., 1997). Nevertheless, this program seems ahead of other university programs in its methodology and its ability to provide evidence that it has an impact in preventing suicidal behavior.

Although the utilization figures are an achievement, the program is not in a position to tell us how effective the treatment is that students are receiving after their referral. The drop in suicide rates is even more impressive, However, almost 10% of students known to have threatened or attempted suicide withdrew voluntarily because they did not want to participate in the program or because they did not want to wait until they were asked to leave.  Their leaving makes it difficult to evaluate the significance of the suicide figures.

More colleges in the country are developing mental health services for students. Many, however, lack the financial resources or are unwilling to spend them on mental health. The problem is aggravated by the numbers of students now seeking those services. The programs discussed have the resources and are able to evaluate utilization of their services, but neither one is in a position to tell us how effective the treatment is that students are receiving after their referral. In addition, when failure to report suicidal behavior is punishable, it may encourage reporting by some and the opposite behavior by others. Anonymous surveys of students would be needed to ascertain that information.

The neurobiological underpinnings of suicidal behavior are the subject of considerable research, and new information has broadened our understanding of this complex area. Researchers have identified serotonergic dysfunction, noradrenergic dysfunction, dopaminergic dysfunction, and hypothalamic–pituitary–adrenal axis hyperactivity as the key neurobiological correlates of suicidality (Mann, 2003; Nemeroff, Compton, Berger, et al., 2001).

The most extensively replicated studies have focused on the role of serotonergic dysfunction. Studies have reported that depressed patients who have made suicide attempts have lower levels of 5-hydroxyindoleacetic acid (5-HIAA) in the brain stem and in cerebrospinal fluid compared to depressed non-attempters (Åsberg, Lil Traskman, & Thoren, 1976; Nordstrom & Åsberg, 1992). Decreased CSF 5-HIAA is hypothesized to be a marker of the impulsive, aggressive, and violent nature of suicide and appears to correlate with a high degree of suicidal planning and a high level of lethality of suicide attempts (Stanley, Molcho, & Stanley, 2000). Central nervous system serotonergic dysfunction has also been associated with suicidal behavior. For patients with unipolar depression, selective serotonin reuptake inhibitors (SSRIs) are considered superior to other antidepressants for treating both suicidal behavior and suicidal ideation (Oquendo, Malone, & Mann, 1997).

Although no large randomized controlled trials of SSRIs have included outcomes related to suicide due to the past exclusion of suicidal patients from most pharmaceutical-sponsored trials, there is considerable evidence pointing to the positive effects of such medications on the central nervous system serotonergic dysfunction noted above to be associated with suicidality in adults (Oquendo, Malone, & Mann, 1997).

Subsequently, several European studies reported inverse correlations between use of SSRIs and suicide deaths, suggesting their potential significance for reducing suicide risk in adults (Barbui, Campomori, D’Avanzo, et al., 1999; Carlsten, Waern, Ekedahl, & Ranstam, 2001; Gunnel, Middleton, Whitley, et al., 2003; Isacsson, 2000). Although large randomized controlled trials need to be conducted in order to determine causative linkages, SSRIs appear to be a potent means of treating suicidality. SSRIs have been reported to have less inherent toxicity than the previously widely used tricyclic antidepressants and are thus less likely to be related to death from overdoses. Although side effects such as gastrointestinal upset, insomnia, and sexual dysfunction are fairly common, most SSRIs appear to be well tolerated.

There was accumulating evidence for the effectiveness of fluoxetine, the one SSRI approved by the U.S. Food and Drug Administration in the treatment of young people who are suicidal. An analysis by Olfson, Shaffer, Marcus, and Greenberg (2003) reported an inverse relationship between regional change in the use of antidepressants among youth ages 10 to 19 and suicide mortality. The relationship was found to be significant specifically among male youth ages 15 to 19 and in geographic regions with lower family median incomes. Although this study does not establish that the use of antidepressants is causally linked to decreases in suicide deaths, other studies provide more direct evidence of their effectiveness. An intervention by Cornelius et al. used fluoxetine (Prozac) to treat adolescents with comorbid major depression and an alcohol use disorder, including some who demonstrated suicidal ideation at baseline (Cornelius, Bukstein, Birmaheret, et al., 2001). The intervention was based on findings that reducing depression and problem drinking in adults resulted in a reduction of suicidal behavior (Dinh-Zarr, Diguiseppi, Heitman, & Roberts, 1999). Cornelius et al. also found fluoxetine to be effective in treating suicidal adults with an alcohol use disorder. Such treatment improved, but did not completely eliminate, both depressive symptoms (including suicidal ideations) and level of drinking (Cornelius, Salhoum, Lynch, et al., 2001). In their studies involving youth, all patients receiving fluoxetine improved with respect to depressive symptoms, and over half improved in symptoms of alcohol dependence. Among participants with suicidal ideation at baseline, ideation decreased and these decreases remained 1 year after treatment (Cornelius, 2003). Cornelius reported no serious adverse effects of fluoxetine among youth.

A definitive study supported by the National Institute of Mental Health known as the Treatment of Adolescents with Depression Study (TADS) has provided the strongest evidence of the effectiveness of fluoxetine in treating adolescent depression and suicidality. This study randomly assigned 439 youths ages 12 to 17 diagnosed with moderate to severe depression to one of four treatment conditions for a period of 36 weeks: fluoxetine therapy alone, cognitive-behavioral therapy (CBT) alone, fluoxetine and CBT, and a placebo drug treatment. Based on the results obtained during the first 12 weeks of the study, the highest rate of clinical improvement (71%) was found among those receiving the combination treatment, followed by 61% in those who received fluoxetine alone, 43% in those who received CBT alone, and 35% in those who received the placebo drug treatment (March, Silva, Petrycki, et al., 2004). The most seriously suicidal adolescents were excluded from the TADS sample, and thus only 29% of participants reported having clinically significant suicidal ideation at baseline. This percentage decreased to 10% by week 12. Although no suicides occurred during the trial, the risk of a suicide attempt among study participants during the first weeks on fluoxetine was reported to be twice that for participants not receiving the medication. The study investigators concluded, however, that the benefits of the medication far outweighed any associated risk.

Although fluoxetine is the only antidepressant medication approved by the FDA as effective for the treatment of major depression in children and adolescents, it did not forbid the use of other SSRIs such as sertraline and citalopram. Both have been shown to be effective in the treatment of depression (Papanikolau, Richardson, Pehlivanidis, et al., 2006; Pösel & Hauzinger, 2006; Wagner, 2005) and are used by clinicians (“off-label”) where fluoxetine has been ineffective (Taurines, Gerlach, Warnke, et al., 2011).

For depressed adolescents who do not respond to an SSRI, drugs that are noradrenergic as well as serotonergic receptor inhibitors have been found to be effective. These medications reduce anxiety as well as depression; a combination of both symptoms increases the risk of suicide. Venlafaxine (Effexor), which affects both receptors, was the first such drug and is independently approved for the treatment of anxiety disorders (Bailly, 2008; Courtney, 2004; Emslie, Waslick, Weller, et al., 2007; Mandoki, Tapia, Tapia, et al., 1997).

A controlled study of augmentation treatments for treatment-resistant depression enrolled patients who met the DSM-IV criteria for major depression but were not psychotic. Atypical (second-generation) antipsychotics, such as risperidone, olanzapine, and quetiapine, which add the stimulant effects of dopamine to the brain, were shown to cause rapid improvement in the mood of patients with treatment-resistant depression, reflected in dramatic drops in their Hamilton Depression Scale scores (Ostroff & Nelson, 1999). Studies of the effectiveness of atypical antipsychotics in adolescents with treatment-resistant depression have been limited in terms of the number of cases seen because of the danger of side effects, mainly weight gain and drowsiness. In adolescents, these medication are used only when SSRIs have not helped.

A small open-label study of 10 adolescents with treatment-resistant depression treated with adjunctive use of quetiapine showed a significant improvement in mood; three who had engaged in cutting had relief from that symptom. Drowsiness was seen in 40% but usually resolved in the first few weeks of treatment. Eighty percent of patients had insignificant weight gains, but in two of the patients weight gain was significant (Pathak, Johns, & Kowatch, 2005).

Long-term treatment with clozapine, another atypical antipsychotic, has been shown to reduce symptoms among treatment-resistant adolescents with schizophrenia. Its side effects—neutropenia, akathisia, and weight gain—make it a useful line of defense, but only after other atypical antipsychotics have been tried (Gogtay & Rapoport, 2008; Sporn, Vermani, Greenstein, et al., 2007).

Suicidality is decreased among mood-disordered adult patients receiving long-term lithium treatment (Sharma, 2003; Tondo, Jamison, & Baldessarini, 1997). Lithium has been effective in treating severe and persistent aggression among children and adolescents (Malone, Delaney, Luebbert, et al., 2000). Although it is being used to augment SSRIs in treatment-resistant depression, there is no evidence that by itself or in combination it reduces suicidal behavior in adolescents.

Since 2003, concerns have been raised about the safety of the use of SSRIs by children and adolescents, based initially on unpublished data from drug company studies linking use of SSRIs by children and adolescents to suicidal ideation and self-harm behaviors. In late 2003, these reports led the British drug regulatory agency to recommend against the use of all SSRIs except fluoxetine in treating depression among youth under age 18 (Goode, 2003). In 2004, the FDA undertook a review of 23 clinical trials involving the use of nine different antidepressant medications by children and adolescents. There were no completed suicides among the 2,200 children treated with SSRIs. The results of this analysis, however, found that the medications increased the risk of suicidal thinking and behavior (suicidality) in children and adolescents with major depressive disorder or other psychiatric disorders (Hammand, 2004). Specifically, 4% of all youth taking medication reported an “adverse event” (i.e., thoughts of suicide and/or potentially dangerous behavior) compared to 2% of those taking a placebo drug.

On October 15, 2004, the FDA directed pharmaceutical companies to label all antidepressant medications distributed in the United States with a “black box” warning of the danger to young people who are prescribed these medications. The warning states that the increased risk of suicidal thinking and/or behavior occurs in a small proportion of youth and is most likely to occur during the early phases of treatment. In 2007, further review led the FDA to recommend that the warning be extended to include young adults up to age 25. Although the FDA did not prohibit the use of antidepressants by children and adolescents, it called upon physicians and parents to closely monitor youth who are taking the medications for a worsening in symptoms of depression or unusual changes in behavior.

On February 1, 2005, the American Psychiatric Association (APA) and the American Academy of Child and Adolescent Psychiatry (AACAP) released detailed fact sheets for physicians and parents on the use of medications in treating childhood and adolescent depression. The fact sheets were jointly developed by the two organizations and endorsed by more than a dozen leading medical, mental health, and advocacy organizations because of their concern that the FDA black-box warning could have the unintended effect of limiting necessary, appropriate, and effective treatment of depression and other psychiatric disorders in youth. The APA/AACAP statement was particularly critical of the FDA’s measurement of suicidality following antidepressant use among young people, which essentially used thoughts of suicide or potentially dangerous behaviors that had been spontaneously shared by the young participants and subsequently recorded in the researchers’ “adverse events reports.” Although the FDA analysis showed more such spontaneous reports among those taking an antidepressant medication as compared to placebo (4% vs. 2%), this finding was not supported by data from 17 of the 23 studies examined that had systematically asked all participants about their suicidal thoughts and behaviors, using standardized forms and not clinicians’ reports. The FDA’s analysis of these data concluded that medication neither increased suicidality that had been present before the treatment, nor induced new suicidality in those who were not thinking about suicide at the start of the study. All studies collecting such data reported a reduction in suicidality over the course of treatment. The APA/AACAP statement noted that while the FDA reported both sets of findings, it did not comment on the contradiction between them. The statement further questioned the reliability of the 2% and 4% spontaneous report rates, noting findings from numerous community samples that as many as half of adolescents with major depression were thinking of suicide at the time of diagnosis and 16% to 35% had reported making a suicide attempt. The statement included suggestions for physicians and parents in monitoring youth receiving antidepressant medication and called for the development of a readily accessible registry of clinical trials that could aid in resolving the controversy and conflicting information surrounding the prescribing of antidepressants to children and adolescents.

A comprehensive review of pediatric trials conducted between 1998 and 2006, partly funded by National Institute for Mental Health (NIMH), concluded that the benefits of antidepressant medications likely outweigh their risks to children and adolescents with major depression and anxiety disorders (Bridge, Iyengar, & Salary, 2007). In a meta-analysis published in 2012 (Gibbons, Brown, Hur, et al., 2012), investigators obtained complete longitudinal data for randomized controlled trials conducted by Eli Lilly on their treatment of depressed adolescents with fluoxetine. The patient-level data allowed them to (1) estimate rates of change in suicidal thoughts and behavior during the course of treatment; (2) study the relationship between depression severity and suicidal thoughts and behavior; and (3) determine the extent to which changes in depressive symptoms mediate the relationship between antidepressant treatment and suicidal thoughts and behavior. These analyses substantially expand the original analyses conducted by the FDA in youth by including the complete longitudinal record of each patient, which was not previously reported by the FDA. An unanswered question noted by the investigators was the absence of any statistically significant effect of fluoxetine on the suicide risk of young people despite a large overall reduction in their depressive symptoms. They point out that we do not know why some youths whose depressive symptoms are reduced continue to have suicidal ideation and behavior. Perhaps other psychological factors play a more important role in suicidal behavior and ideation, such as aggressive, impulsive traits in youths (Brent, Oquendo, Birmaher, at al., 2002). Thus, even if antidepressants reduce depressive symptoms overall in youths, it is possible that a subset may retain some level of risk for suicidal ideation or attempt. The overall rate, however, was not significantly greater than that found for youths randomized to placebo. They found no evidence that fluoxetine increased the risk of suicidal thoughts or behavior.

In a comprehensive release of information to parents and caregivers, available online, NIMH supported and referred parents to the 2007 review by Bridges et al. indicating that the benefits of SSRIs outweighed the risks to children and adolescents with major depression and anxiety disorders. That controversy is essentially resolved, at least among mental health professionals.

Until recently, seriously suicidal adolescents (and adults) were excluded from randomized controlled trials conducted by all pharmaceutical companies. In 2010 the FDA mandated that pharmaceutical companies include suicidal patients in all drug trials with depressed patients, a widely welcomed decision that should stimulate needed additional pharmacological research. Many of the medications that are useful in reducing adult suicidal behavior, and have proven successful in treating depression in adolescents, need to be tested for their effect on adolescent suicidal behavior. The avoidance of adolescents at serious risk for suicide has been true of clinical research in medical centers as well. A recent suicide attempt has often been the basis for exclusion from a treatment research project.

Two multisite projects have addressed this problem. The Treatment of Adolescent Suicide Attempters (TASA) study was designed to examine the course of depression during the treatment of adolescents who had made a suicide attempt in the past 90 days (Brent, Greenhill, Compton, et al., 2009; Vitiello, Brent, Greenhill, et al., 2009). At five academic sites, these adolescents entered a 6-month program of treatment with antidepressant medication; a modified CBT focused on suicide prevention; or a combination of both. Treatment assignment was either random or chosen by the study participants. Most patients chose their treatment assignment and, overall, three quarters received the combined treatment.

Among these adolescent attempters, the factors that predicted another suicide attempt or acute suicide ideation requiring emergency referral were higher self-rated depression, suicidal ideation, family income, greater number of previous suicide attempts, lower maximum lethality of previous attempt, history of sexual abuse, and lower family cohesion.

Vitiello et al. reported that, with combined medication and CBT, which was chosen by the majority of patients, rates of improvement and remission of depression were comparable to those in non-suicidal depressed adolescents. Brent et al. reported that allocation to CBT, medication, or the combination lowered the 6-month risks for suicidal events and reattempts. These promising results need to be interpreted in light of the fact that the TASA trial was largely nonrandomized and uncontrolled; that patients with substance abuse (common among suicide attempters) were excluded; that the dropout rate—albeit in the range of other treatment studies and hardly surprising with this adolescent sample—was not small (31%); and that approximately 20% of the subjects displayed suicidal behavior during treatment (Walter, 2009).

The Treatment of SSRI-Resistant Depression in Adolescents study (TORDIA) was designed to inform clinicians as to the best next step in treatment after a depressed adolescent has not responded to an adequate trial with an SSRI. In this six-site study, conducted from 2000 to 2006, 334 treatment-resistant adolescents with major depressive disorder were randomized to one of the four treatment strategies: (a) switch to another SSRI; (b) switch to venlafaxine; (c) switch to another SSRI and CBT; or (d) switch to venlafaxine and add CBT (Emslie et al., 2010). Poorer response across treatments was predicted by greater severity of self-rated depression, hopelessness, suicidal ideation, a history of non-suicidal self-injury, and family conflict. CBT continued to be associated with an increased likelihood of a response, even after controlling for the most salient adverse predictors of response. SSRIs were superior to venlafaxine in those participants with higher self-reported depression. The combination of CBT and medication was superior to medication alone for those with multiple comorbidities, particularly anxiety or attention-deficit/hyperactivity disorder. Participants with a history of substance abuse showed a lower response rate to combination treatment than to medication monotherapy. Those who received pharmacological treatment for sleep difficulties showed a poorer response rate than those who did not receive medication for sleep. Participants whose self-reported drug and alcohol use was high at baseline and remained so, or whose substance use accelerated during the trial, showed a poorer response rate compared to those whose drug and alcohol use either remained or became low.

The TORDIA study found no difference in the rate of suicidal events among treatments, perhaps because the proportion of seriously suicidal participants was much higher than in most studies. Surprisingly, TORDIA found a lower rate of non-suicidal self-harm in the medication-only group. Given that the medication time to a suicidal event in TORDIA was 3 weeks, it is unlikely that CBT as currently delivered could effectively protect against these early events. The TORDIA study demonstrated that highly suicidal individuals can be included and managed in the context of clinical trials. The authors stress the importance of studying high-risk populations if we are going to improve the treatment of these suicidal adolescents.

In the TASA study, the rates of improvement and remission seen in adolescent suicide attempters treated with CBT, medication, or a combination of both are encouraging. That the study was not randomized or controlled should stimulate efforts to build on it.

The TORDIA study provided guidelines for therapists in choosing options available for treatment-resistant depression in adolescents depending on their symptoms. Still, matching patients to appropriate treatments remains largely a matter of guesswork. The fact that seriously suicidal adolescents could be included in the trials without a problem will encourage further research. The medication options were limited to SSRIs and the noradrenergic receptor inhibitor venlafaxine. There is a need for randomized controlled testing of the atypical antipsychotics that have shown effectiveness in treatment studies of adolescents whose depression is resistant to SSRIs.