Psychosocial Approaches to the Treatment of Bipolar Disorder

Clinical Context

Bipolar Disorder: Prevalence, Course, and Disability

Bipolar disorder is a highly disabling illness, with average onset at 18 years, low academic and social achievement in childhood, high rates of drug or alcohol abuse in adolescence and adulthood, and high rates of suicide and cognitive dysfunction throughout the lifespan. The lifetime prevalence in the United States is 1.0% for bipolar I disorder, 1.1% for bipolar II disorder, and 2.4% for subthreshold bipolar disorder (1), with lower rates in a cross-national epidemiologic study (2). Approximately 2% of adolescents have had manic episodes (3). Across clinical studies conducted in different countries, lifetime rates of bipolar disorder in childhood converge on 1.8% (4), suggesting that rates may be increasing in more recent birth cohorts.

Pharmacological regimens for bipolar disorder include mood stabilizers (commonly, lithium, valproate, or lamotrigine) or second-generation antipsychotics (i.e., quetiapine, olanzapine, aripiprazole, risperidone, ziprasidone, and lurasidone). Antidepressants, when given at all, are usually prescribed adjunctively to mood stabilizers. Even with optimal medication regimens, however, only 21%–27% of patients with bipolar depression recover fully from episodes within 1 year (5) and 37%–49% have recurrences (6, 7). Five-year relapse rates in naturalistically treated samples typically range from 60% to 85% (8).

Even when treated with best-practice pharmacotherapy, many bipolar patients have subthreshold depressive symptoms that persist after acute episodes of mania or depression. Subthreshold symptoms have considerable prognostic significance: even mild levels of residual depression are associated with low levels of psychosocial functioning and high risk for recurrences (9, 10). Lack of full remission—indicating that the illness is still active—is associated with a shorter time to recurrences of depression or mania (11). Moreover, many patients have residual psychosocial impairments following episodes that are not targeted by medications (9, 10). For example, lack of social support following acute mood episodes was associated with more severe depressive symptoms at 1-year follow-up, especially among patients who did not fully recover from the their acute episodes (12). Correspondingly, treatments that bring the patient to a full and durable remission and enhance social support during the postepisode period are most likely to prevent or delay recurrences.

Psychosocial Treatments as Adjunctive to Pharmacotherapy

When evaluated against the backdrop of the growing number of pharmacological options for bipolar disorder, progress in psychological interventions has been relatively slow. Yet, adding psychotherapy to pharmacotherapy has been found to reduce rates of recurrence by 50% or more compared with usual care (13). Additionally, 31 of 35 controlled studies examining the efficacy of psychosocial interventions have shown efficacy when compared with control conditions (typically, nonspecific treatments or usual care) (14). Most treatment guidelines for bipolar disorder recommend adding psychotherapy (at minimum, psychoeducation sessions focused on illness management) to medications to speed recovery from mood episodes, delay or minimize recurrences, and enhance psychosocial functioning (e.g., reference 15, 16.).

Therapy modalities with evidence of efficacy in bipolar disorder include cognitive-behavioral therapy (CBT), group psychoeducation, family-focused therapy (FFT), and interpersonal and social rhythm therapy (IPSRT). This article reviews recent randomized trials of psychosocial interventions, emphasizing major strategies, putative therapeutic ingredients, and limitations. We conclude with clinical and research recommendations, with particular emphasis on community implementation studies.

The Role of Stress in Bipolar Disorder

What psychological characteristics of individuals with bipolar disorder—or their life contexts—would one seek to modify in psychotherapy? When evaluated in remission, bipolar patients do not appear to have consistent personality types or patterns of dysfunctional thinking that distinguish them from individuals with other disorders. There is evidence, however, that many bipolar patients are highly stress-sensitive (i.e., prone to rapid recurrences following life stressors; 17, 18) and that childhood adversity (i.e., sexual, physical, or emotional abuse) potentiates this stress sensitivity (19, 20). Life events that cause changes in sleep/wake rhythms (21) or increases in goal-directedness (22) may precipitate manic symptoms in otherwise stable patients. Negative life events, such as the loss of a significant other, are more closely associated with depressive episodes, especially in the absence of adequate social supports (18).

Much like patients with other psychiatric disorders, bipolar patients are quite sensitive to family conflict and distress. Bipolar patients who have caregivers who are high in “expressed emotion” when discussing the patient with a clinician (i.e., express highly critical, hostile, or overprotective attitudes or beliefs) or when interacting directly with the patient are more likely to relapse over 9 months than those in less affectively charged (low expressed emotion) caregiver/patient relationships (23). Most recently, the role of childhood adversity has been highlighted as a predictor of initial and recurrent mood episodes (20). In the Stanley Foundation multisite study of bipolar disorder, sexual and physical abuse were associated with earlier age at onset, more medically serious suicide attempts, more illness episodes, and substance abuse comorbidity (24, 25).

Thus, stress plays an important role in the onset of mania and depression among individuals with bipolar disorder, through both genetic and environmental risk pathways. Conversely, teaching patients strategies for coping with stressful events, even if they do not directly address childhood contexts, may reduce the impact of such stressors on relapse and psychosocial impairment.

Treatment Strategies and Evidence

Common Elements of Psychotherapy

Psychosocial treatments are intended as adjuncts to pharmacotherapy and are typically offered during the postepisode “continuation” phase or the maintenance phase of treatment. They are present-focused and emphasize learning skills for managing the disorder (psychoeducation). Although the modalities have common objectives (Table 1), there are distinct elements as well: format (e.g., whether treatment is given individually, in groups, or in family units), length (minimum 3–6 sessions and up to 30 or more), and whether treatment is intended for patients in remission, those recovering from an acute episode, or both.

In the following sections, we review randomized trials of psychotherapy. Further details of these trials, and coverage of psychotherapy methods that have only been tested in uncontrolled trials, are available elsewhere (26, 27).

Cognitive-Behavioral Therapy (CBT)

CBT approaches to bipolar disorder consist of three core strategies: behavioral activation (i.e., helping patients to increase activity levels when depressed and “dial it down” when their moods escalate), relapse prevention (identifying prodromal symptoms of new episodes and implementing preemptive plans), and cognitive restructuring (modifying automatic negative thoughts and core dysfunctional beliefs). Behavioral activation may include reducing daily activities to avoid overstimulation, even if these events are pleasurable. Cognitive restructuring may involve challenging “hyperpositive” thinking (e.g., “I cannot lose…I’m in complete control of my fate”) as well as overly pessimistic thinking.

Cognitive-behavior therapy has the most extensive record of randomized trials in bipolar disorder, although CBT protocols vary from study to study. In the U.K. study of Lam et al. (28), 103 euthymic patients with bipolar disorder were randomly allocated to CBT (12–18 sessions) plus pharmacotherapy or usual care (pharmacotherapy alone). CBT emphasized psychoeducation, challenging dysfunctional cognitions, and medication adherence monitoring. At 1 year, 44% of the patients in CBT had relapsed versus 75% of those in usual care. At 30 months, patients in CBT no longer differed from patients in usual care on time to overall relapse, but they did have fewer depressive relapses and days in mood episodes.

Scott and associates (29) examined 22 sessions of CBT compared with treatment as usual (TAU) among 253 bipolar patients treated in five U.K. mental health centers. Unlike the Lam et al. (28) study, patients began in a variety of symptom states. No differences were found between CBT and TAU on time to recurrence over 18 months. However, a post hoc analysis revealed that patients who had fewer prior episodes were more likely to have recurrences in TAU (55%) than in CBT (41%), whereas recurrences were more likely among those who had many prior episodes in CBT (81%) than TAU (66%). These results suggest two possibilities: CBT is best suited to the earliest phases of the disorder, or CBT may be unsettling and agitating to patients who are unstable, have a more refractory illness, or have more cognitive impairment.

In a well-designed Canadian trial of 204 patients in full or partial remission, participants were randomly assigned to 20 sessions of individual CBT or 6 sessions of group psychoeducation (30). No differences emerged over 1.5 years in symptom severity or recurrence rates. Given that group psychoeducation cost an average of $180 per participant whereas CBT cost $1,200 per participant, group psychoeducation would appear to be the more cost-effective alternative.

Finally, investigators at the University of Tubingen, Germany randomized 76 patients to 20 sessions of CBT or 20 sessions of individual supportive therapy, both with pharmacotherapy (31). The patients had subthreshold manic or depressive symptoms, but none were in an acute episode. No differences were observed in relapse rates over 33 months (overall rate, 64.5%). The authors point to the common elements of the two approaches (i.e., provision of information, systematic mood monitoring) in explaining the lack of differences on relapse. Additionally, the length of the two treatments proved informative: risk for relapse decreased by 10% with each therapy session that patients attended, regardless of the treatment condition.

Group Psychoeducation

Several research teams have evaluated group treatment in conjunction with pharmacotherapy for relapse prevention. Group treatments take advantage of the social support provided by other patients, who may enliven psychoeducation with real-life examples. Colom and associates (32) at the University of Barcelona tested a 21-session group treatment that included exercises to promote greater awareness of illness states, early detection and intervention with prodromal symptoms, the importance of medication adherence, and enhancing stability through lifestyle (e.g., sleep/wake cycle) regularity. Although this group treatment had elements in common with CBT, it made minimal use of cognitive restructuring or behavioral activation (pleasant events) schedules (33). Colom et al. found that, over 2 years a 21-session structured psychoeducation group was associated with fewer recurrences (67% versus 90%) and better psychosocial functioning than a 21-session support group (32). Over 5 years, patients who had received the structured groups had far fewer days of acute illness (mean 154 days) compared with those who received the unstructured group (586 days) (34).

Two randomized trials examined the effectiveness of group psychoeducation within the context of multicomponent care plans. In 11 Department of Veterans Affairs sites (35), 306 patients received mood monitoring from a nurse care coordinator and group psychoeducation (5 weekly followed by twice monthly groups for up to 3 years) to improve relapse prevention skills. Over a 3-year period, patients in the multicomponent program spent fewer weeks in manic episodes than patients who received usual care. Patients in multicomponent care also had significant improvements in social functioning and quality of life.

The largest randomized trial in bipolar disorder (N=441) tested a similar 2-year multicomponent care intervention—with group psychoeducation at the same frequency—within the Kaiser Permanente health network (36). Patients in the multicomponent intervention had lower mania scores and spent less time in manic or hypomanic episodes than patients in a usual care condition. Neither this study nor the Bauer et al. study found effects of the multicare program on depressive symptoms. “Dismantling” design studies, in which modules of multicomponent treatments are tested with and without each other, will provide one avenue for determining the unique contribution of group psychoeducation to symptom outcomes.

An adaptation of structured psychoeducation groups called “functional remediation treatment” emphasizes patients’ cognitive functioning, with exercises designed to improve memory, attention, problem solving, and organizational skills. In a 10-site randomized trial in Spain, 268 patients were assigned to 21 weekly group sessions of functional remediation, 21 sessions of standard group psychoeducation, or TAU (37). Patients in the functional remediation groups showed greater improvements in psychosocial functioning than those in TAU, but fared only slightly better than patients in the standard psychoeducation groups.

Hence, group psychoeducation has been shown to be an effective and, in all probability, cost-effective adjunct to pharmacotherapy for patients with bipolar disorder. Its role in treating and preventing manic symptoms is more consistent than for depressive symptoms. Research on the processes that mediate the effects of group psychoeducation—for example, whether being treated alongside of others leads to decreased stigmatization and a greater willingness to adopt illness management strategies—may lead to the development of even more powerful group approaches. Group treatments may be more difficult to implement in public or private settings where treatment is dispersed across different providers and locations, but the availability of Skype and other online communication tools may minimize these limitations.

Family-Focused Treatment (FFT)

Given in up to 21 weekly followed by biweekly sessions during the postacute (continuation) period, FFT aims to hasten stabilization and reduce the likelihood of recurrences of bipolar disorder (38). In the initial treatment phases, patients and family caregivers (usually spouse or parents) are instructed in how to recognize early warning signs of mania or depression and develop prevention strategies (e.g., how best to alert the patient to changes in his or her moods or behavior, rehearsing what to say to the attending psychiatrist). Assisting patients in stabilizing their sleep/wake cycles and staying adherent to medications are also key strategies of family psychoeducation; however, clinicians encourage family members to recognize their roles in these problems (e.g., a parent who sets no limits on a 15-year old who stays up all night playing video games, a spouse who supports the patient in believing that marijuana is a good substitute for mood stabilizers). In later stages of FFT (6–9 months), clinicians assist families in skills training for enhancing communication (i.e., learning to listen actively, request changes in each other’s behavior, offer both positive and constructive negative feedback) and problem-solving. At this point, patients are often able to return to tasks that were on hold during and following acute episodes (e.g., parenting of young children).

Unlike CBT or group psychoeducation, FFT sessions always involve family members, and skills training focuses on improving family relationships. Cognitive restructuring is not a key component of treatment except in cases where, for example, patients’ or caregivers’ attitudes are based on misinformation about the illness (e.g., “lithium destroys brain cells”; “bipolar disorder is no different than just being moody”).

Table 2 summarizes the randomized trials of FFT in adults and adolescents with bipolar disorder. Overall, FFT and pharmacotherapy have been associated with a 35%–40% reduction over 2 years in recurrence rates compared with brief psychoeducation and pharmacotherapy, with numbers needed to treat (NNTs) ranging from 5 to 10 (39–48). In several trials, effect sizes for FFT (compared with brief treatment) have been stronger in patients from families with high expressed emotion than from families with low expressed emotion, suggesting that patients in high-intensity/high-conflict families may benefit most from FFT (40, 41, 47).

Family Interventions for Pediatric Bipolar Spectrum and High-Risk Conditions

Despite the greater uncertainty about the diagnostic boundaries of pediatric bipolar conditions, there is a more extensive empirical basis for family interventions in this age group. Fristad and colleagues (49) examined 8-session multifamily psychoeducational groups in which parents of bipolar children had the opportunity to interact with one another. The psychoeducational material included information about mood management, communication skills, and coping strategies to avert mood escalation. In the largest pediatric study to date, 165 bipolar (70%) and depressed (30%) children (ages 8–12 years) were randomly assigned to immediate 6-month group treatment or a delayed group treatment in which treatments were given from study months 12–18. Over 1 year, children whose families participated in the immediate groups showed greater improvement in affective symptoms than children whose families were waitlisted (49). The clinical benefits of psychoeducation were mediated by improvements in parents’ ability to advocate for their child’s mental health care and the higher quality of services utilized. In turn, quality of services was associated with improved mood symptoms in children over 1 year (50).

The 12-session child and family-focused cognitive behavioral therapy program (also known by the acronym “RAINBOW”) incorporates single-family psychoeducation sessions with individual parent psychoeducation and CBT for the child (cognitive restructuring, behavioral activation [pleasurable events scheduling], and mindfulness meditation) (51). In a randomized trial with 69 children and adolescents (ages 7–13, mean 9 years) who also received medication management, greater improvements were observed over 6 months for mania symptoms, depressive symptoms, and global functioning scores compared with an equally intensive psychosocial TAU condition (51).

FFT has been examined in two trials with bipolar adolescents (FFT Adolescent version, or FFT-A). In the first, adolescents with bipolar I, II, or not otherwise specified disorder who received 21 sessions of FFT and pharmacotherapy had more rapid recovery from depressive episodes at study entry, less time in depressive episodes at follow-up, and more time well over 2 years compared with adolescents in brief psychoeducation (“enhanced care”) and pharmacotherapy (40). A second trial involving 145 adolescents with bipolar I or II disorder treated over three sites did not replicate these results: adolescents in FFT-A and those in enhanced care (both administered with best practice pharmacotherapy) were equivalent in time to recovery and time to recurrence. Adolescents in FFT-A, however, had fewer severe manic symptoms in the second study year than those in enhanced care (39).

Children and teens who are at high risk for bipolar disorder, typically defined as those with bipolar disorder not otherwise specified or major depressive disorder who have at least one first-degree relative with bipolar I or II disorder, also have positive responses to FFT. In a 1-year randomized controlled trial, genetically predisposed children and adolescents (ages 9–17 years) who received 12 sessions of FFT (high-risk version) with or without pharmacotherapy recovered more rapidly from their initial depressive symptoms, had more weeks in remission, and showed greater improvement in hypomania symptoms over 1 year than those who received brief psychoeducation with or without pharmacotherapy (47). Studies currently underway will examine whether family intervention is effective in delaying or preventing the onset of bipolar I or II disorder in high-risk children.

A version of dialectical behavior therapy (DBT) has been developed for adolescents (ages 12–18 years) with bipolar I, II, or not otherwise specified disorder (52). DBT was modeled as a 1-year treatment consisting of 18 family skills training (conducted with individual family units) and 18 individual skills sessions. DBT is a cognitive-behavioral therapy that incorporates components of Eastern philosophy (e.g., mindfulness meditation) to enhance emotion regulation, mindful awareness, distress tolerance, and interpersonal skills. In a 20-subject trial with a 2:1 randomization ratio, adolescents were randomly assigned to DBT (N=14) or a less-intensive psychosocial treatment condition (N=6). All participants received medication management as well. Adolescents who received DBT had less severe depressive symptoms and more improvement in suicidal ideation over the year; they also evidenced more weeks in remission (52).

There are some clues as to what variables are associated with a positive response to family interventions versus usual care in pediatric bipolar patients, including family expressed emotion (53), more severe parental depression (51), and greater child impairment at baseline (54). Currently, we do not know what patient or family attributes predict a stronger response to family versus individual CBT or group psychoeducation, a fertile area for future research.

Interpersonal and Social Rhythm Therapy (IPSRT)

The interpersonal psychotherapy of depression, originally developed as a comparison to CBT in the Treatments for Depression Collaborative Research Program (55) has been adapted for bipolar disorder. In both traditional IPT and IPSRT, clinicians assist patients in resolving issues related to grief, role transitions (e.g., divorce or separation), role disputes (e.g., marital or family conflict), or interpersonal deficits (e.g., repetitive, self-defeating behavior patterns in relationships). In IPSRT, strategies to enhance social and circadian rhythm regularity are integrated into interpersonal problem-solving (56). Indeed, psychosocial events that disrupt daily or nightly routines such as when a person sleeps, wakes, eats, or exercises have repeatedly been found to precipitate episodes of mood disorder (21, 57).

In the largest trial of IPSRT (58), 175 patients were randomly assigned during an acute episode of mania, depression, or mixed illness to IPSRT or intensive clinical management (ICM, a psychoeducational control therapy), both with protocol-based pharmacotherapy. Once patients had stabilized (minimum 4 weeks of stability) from their acute episode, they were rerandomized to IPSRT or ICM, with biweekly and then monthly sessions for up to 2 years. Thus, four treatment strategies were formed. The 2-year recurrence rates were: 41% for IPSRT followed by IPSRT, 41% for IPSRT followed by ICM, 28% for ICM followed by ICM, and 63% for ICM followed by IPSRT. IPSRT in the acute phase was associated with a longer time to recurrence in the maintenance phase than ICM (58). Moreover, patients in IPSRT showed better occupational functioning during acute treatment than those in ICM (59).

Interestingly, patients who received IPSRT acutely were more able to stabilize their social routines and sleep/wake cycles during acute treatment than those in ICM. Thus, acute stabilization of sleep/wake rhythms may have downstream effects on the prevention of future mood instability (60). It is less clear why rates of recurrence were highest in those patients who switched from ICM to IPSRT for the maintenance phase.

IPSRT may have “stand-alone” effects for patients with bipolar II disorder. In a small trial for acute bipolar II depression (60), 25 patients were randomly assigned to quetiapine monotherapy (beginning at 25 mg and increasing to 300 mg) or IPSRT monotherapy. Over 12 weeks, both groups improved equally in depression scores, although absolute response rates were low (27%–29%). Future studies should examine whether there is an additive effect of combining IPSRT with a second-generation antipsychotic or mood stabilizer in the acute treatment of bipolar depression.

A Comparison of Therapy Approaches: The STEP-BD Study

A significant limitation of the bipolar psychotherapy literature is the lack of controlled comparisons of one specialty treatment to another. The Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD) recruited 293 patients in a depressed phase of bipolar I or II disorder from 15 sites, and randomly assigned these patients to 1) one of three intensive psychosocial treatments (up to 30 sessions of FFT, IPSRT, or CBT over 9 months) or 2) a 3-session control treatment called collaborative care (CC). All patients received protocol pharmacotherapy (mood stabilizers or antipsychotics with or without an SSRI or buproprion). Over 1 year, being in any of the intensive psychotherapies was associated with more frequent (and more rapid) recovery from depression and better psychosocial functioning than being in the CC treatment, but there were no statistical differences among the intensive treatments (see Table 2 for recovery rates) (44, 45). Patients in the intensive treatment were 1.6 times more likely than patients in CC to be clinically well in any given month of the study.

A reanalysis of STEP-BD data (61) revealed that having a lifetime history of an anxiety disorder was a significant predictor of differential response to intensive therapy versus brief treatment. Among patients with lifetime anxiety disorders (N=177), the number needed to treat to observe a difference in intensive therapy versus CC was 5.88 (small-to-medium effect). Among patients without a history of anxiety disorder (N=92), the NNT to observe a difference in intensive therapy versus CC was 50.0 (minimal effect). This is a clinically useful finding in that pharmacotherapy for comorbid anxiety disorders usually includes SSRIs that can theoretically increase the risk of affective switches in bipolar patients. The STEP-BD study suggests that psychotherapy is a vital part of the effort to stabilize episodes of bipolar depression, particularly among patients with anxiety comorbidity. Furthermore, patients with acute depression and anxiety may require more intensive psychotherapy than is typically offered in community mental health centers.

Questions and Controversies

Despite the increasing number, variety, and sophistication of trials of psychotherapy for bipolar disorder over the past decade, we are left with fairly simple conclusions. First, psychotherapy is an effective adjunct to pharmacotherapy in the postepisode or remitted phases of bipolar disorder, with significant evidence for several forms of family intervention, group psychoeducation, IPSRT, and CBT. These treatments focus on illness management (psychoeducation) and, to a lesser extent, interpersonal skill training. Treatments found to be more effective than usual care are usually 12 or more sessions and last at least 4–6 months.

Psychotherapies that have an effect on recurrence rates also reduce hospitalization days, suggesting economic benefits. For example, one study of FFT found that patients with bipolar disorder were less likely to relapse over 2 years than patients in comparably intensive individual psychoeducation and were less likely to require hospitalization when they did relapse (42). In the first Barcelona study (32), structured psychoeducation groups were associated with a cost savings of approximately $6,500 per patient over 5 years (62). The “price point” at which psychotherapies pay for themselves in terms of illness or treatment cost savings (e.g., reductions in costs of medications, lost days from work, insurance copays) will be of interest to patients, clinicians, and health care administrators, but we are far from determining how this price point differs across settings, age groups, or clinical presentations.

We do not know what forms of psychotherapy are the most effective for different phases of illness. Studies of group psychoeducation (e.g., 32) or CBT (28) specify up to 6 months of remission as an entry criterion, which would significantly reduce the number of eligible patients in many settings. Both FFT and IPSRT include patients who have subthreshold levels of illness and make use of current symptoms as a teaching tool for defining prodromal symptom states. Some patients may only need a brief period of psychoeducation and support to help make sense of their recent mood episode and do not need longer-term therapy; patients who respond quickly to medications may be in this group. The role of comorbid disorders other than anxiety, including substance abuse or personality disorders, in specifying the type and frequency of treatment deserves study.

We know relatively little about “mediating variables” or change mechanisms responsible for why patients improve in one treatment versus another. Ideally, a study of mediating mechanisms would compare two or three forms of intensive psychotherapy after an acute episode and measure presumed mediators at baseline, midtreatment, and after treatment to determine whether changes in the mediator precede changes in symptoms or functioning. An example in FFT is the study of Simoneau et al. (63), who measured family interactional behavior in laboratory problem-solving tasks before and after FFT or after brief psychoeducation. Adult bipolar patients showed greater increases in positive verbal and nonverbal behavior from pretreatment to posttreatment in FFT than in the brief treatment. Moreover, these changes in interactional behavior predicted the degree of improvement in mood symptoms among patients over a 9-month treatment interval. Because it only measured family interactions at two time points, this study falls short of showing that changes in family behavior are causally related to changes in patients’ symptoms. Nonetheless, identifying correlates of symptom change at the cognitive, emotional, or interpersonal levels may give us clues as to what treatments are the most powerful in bringing about meaningful clinical change and how to make these treatments more efficient. In the same vein, the ability of pharmacological agents to change specific biological markers (e.g., brain-derived neurotrophic factor) may eventually guide our choice of drug treatments.

A related issue is the importance of measuring changes in neural processes from before to after psychosocial treatments. For example, using a repeated measure neuroimaging design, one could examine whether patients with bipolar disorder show decreases in amygdala activation and increases in dorsolateral or ventrolateral prefrontal cortical activation when viewing negative facial stimuli from before to after psychotherapy. Integrating the study of psychotherapy with brain imaging techniques may also help determine what patients are the best candidates for intensive therapy. A pilot study found that amygdala hyperactivation when viewing fearful faces predicted the degree of response to FFT versus TAU in children at high risk for bipolar disorder (64).

Recommendations from the Authors

In conducting this review, we have been struck by the lack of evidence for dissemination of evidence-based psychotherapies in clinical practice with bipolar patients. Few of the available treatments are being implemented at the community level, in part because of the difficulty in accessing training and supervision from experts. Treatment manuals that are easy to obtain and digest, followed by low-cost supervision of training cases, will be needed before treatments can be disseminated on a larger scale. Computer-assisted learning methods, such as webinars (instead of weekend workshops), online methods of supervision (e.g., chat rooms), and clinician- or patient-administered measures of treatment fidelity (rather than supervisory tape viewing) will all be useful in reducing training costs (65). These methods, however, may be less satisfying to learners and may affect their motivation.

In some community mental health centers, training one highly motivated clinician to “champion” the treatment and train others (the “train the trainer” model) can be of immense help in encouraging the broader adoption of novel psychotherapy methods (65). In community care, administrators will have to provide individual clinicians with release time to obtain this specialized level of training.

Clinicians need to adapt the existing treatment manuals to their practice settings, taking into account the treatment framework normally used in that setting. So, for example, a clinic in which the majority of practitioners are psychoanalytically trained may more easily adopt IPSRT than CBT or FFT. Moreover, clinicians may work in settings in which patients are not fluent in English, structured diagnostic interviews are not considered cost-effective, psychiatric medications are dispensed by a general practitioner, or therapy protocols that exceed 6–8 sessions are not economically feasible. Fortunately, certain treatments, including FFT and IPSRT, have modules that can be given separately from the full protocols (e.g., prodromal symptom monitoring, relapse prevention planning, social rhythm tracking and stabilization, family communication training) and implemented as stand-alone strategies.

Web-centered treatment, in which all components of an intervention are delivered through an interactive website, is becoming increasingly available (65). An online self-care program, “Living with Bipolar,” successfully engaged patients and was associated with higher quality of life scores than a waitlist control (66). Another program, bipolarcaregivers.org, was useful to caregivers in navigating the complexities of bipolar disorder and the mental health system. It was less useful to caregivers of highly chronic patients or those who had complex family problems (67). Further evaluation of online psychoeducation, either as an adjunct to evidence-based psychosocial treatments or as a substitute for them, is clearly needed.

Finally, future research must determine the best point in illness development to begin intervening with psychosocial therapy, and at what level of intensity. Treatment focused on the earliest symptom phases may interrupt neurotoxic processes of the illness and enhance long-term outcomes (68). Early intervention may be most effective if it is successful in modifying biomarkers or psychosocial risk processes that are dysregulated prior to illness onset.

The timing and duration of early interventions, however, should not be solely based on their costs or presumed efficacy at the group level. Early interventions must also be personalized. Children or adolescents with early signs of bipolar disorder are not always motivated for treatment, nor are their parents necessarily invested in preventing a disorder that may not develop anyway. Targeted interventions that focus on disorders that herald the development of bipolar disorder in genetically susceptible children, such as anxiety disorders, ADHD, conduct disorder, depression, or substance/alcohol abuse, may be seen as more relevant and acceptable to patients and parents. Finally, early interventions may achieve considerable effects by building on resilience factors in patients, families, or even communities, such as by increasing community awareness of treatment options for depression and bipolar disorder (69).