Evidence-Based Psychotherapies for Bipolar Disorder

The core objective of psychoeducation is to foster a clear rationale for individuals with bipolar disorder to seek, adhere to, and remain in treatment (17–19). Whereas early psychoeducation was narrowly oriented toward simply improving medication adherence, today’s psychoeducation is more comprehensive (20). It often includes structured sessions that focus on empowering an individual to better understand his or her illness, recognize and manage symptoms, resolve stressful situations, and adhere to pharmacotherapy. It can be administered individually or in a group format—or, more recently, remotely through a telephone, smartphone, or Web platform (Table 1). Psychoeducation can be delivered by professionals of different backgrounds, including psychiatric nurse practitioners and trained peer supports (21, 22). Psychoeducation can be administered as a stand-alone treatment or combined with strategies from other evidence-based interventions. In fact, all the evidence-based psychotherapies include at least a modest amount of psychoeducation, suggesting that it is a core, common therapeutic element across interventions for bipolar disorder.

We identified 14 RCTs with in-person psychoeducation as the experimental condition. The number of psychoeducation sessions was as few as four and as many as 21. Here, we provide brief summaries of each report and describe particularly noteworthy findings.

Perry and colleagues conducted the first trial of individual psychoeducation (19). The goal of individual psychoeducation is to help patients identify prodromal symptoms and respond to prodromes with an “action plan.” The median number of one-hour sessions was nine. Compared with treatment as usual (i.e., pharmacotherapy alone), individual psychoeducation was associated with a significant extension of time to first manic relapse, reduction in the number of manic relapses, and improved social and occupational functioning over 18 months.

In another seminal trial, Colom and colleagues compared group psychoeducation to an unstructured support group (23). Group psychoeducation was a manualized, structured program of 21 sessions (18). The program targeted treatment compliance, illness awareness, early detection of prodromes, and lifestyle regularity. Compared with treatment as usual, group psychoeducation was associated with significantly longer time to recurrence, fewer relapses and hospitalizations, and shorter hospital stays at two- and five-year follow-ups (23, 24). Zaretsky and colleagues compared seven sessions of individual psychoeducation with or without 13 sessions of CBT (25). Compared with psychoeducation alone, psychoeducation with the addition of CBT was associated with significantly fewer days of depressed mood and fewer increases in antidepressant dosage over one year.

D’Souza and colleagues evaluated 12 sessions of group psychoeducation for companion and patient dyads compared with treatment as usual (26). Group psychoeducation was based on the same principles of Colom and Vieta (18) and was modified to reflect the premise that companions might recognize and detect prodromes too. Group psychoeducation was associated with significantly decreased likelihood of relapse, longer time to relapse, decreased manic symptoms, and improved medication adherence over 60 weeks compared with treatment as usual.

Candini and colleagues demonstrated that Colom and Vieta’s (18) 21-session group psychoeducation program significantly reduced risk of, time to, and length of hospitalization compared with treatment as usual over one year (27).

Javadpour and colleagues found that, compared with treatment as usual, eight sessions of group psychoeducation were associated with significantly greater improvement in all areas of quality of life, fewer relapses and affective symptoms, and greater medication adherence over 18 months (28).

Kessing and colleagues demonstrated that, compared with treatment as usual, 12 sessions of group psychoeducation provided in a specialized outpatient mood disorder clinic significantly reduced the rehospitalization rate and increased satisfaction with care over two years (29).

Parikh and colleagues found no significant group difference between six weeks of group psychoeducation and 20 weeks of individual CBT as an adjunct to naturalistic pharmacotherapy (30). This suggests that a brief course of group psychoeducation may be as effective as a full course of individual CBT. The group psychoeducation program was a manualized, structured program of six sessions based on Bauer and McBride’s life goals program (17). It was designed to teach illness recognition and coping strategies and assist with the creation of an action plan for an individual to use when experiencing prodromes or illness triggers. Individual CBT included traditional CBT techniques in addition to an emphasis on understanding the diagnosis and course of bipolar disorder, personal warning signs, and a “relapse drill” of actions to take to reduce full-episode recurrence.

Morris and colleagues reported no significant differences between Colom and Vieta’s (18) 21-session group psychoeducation and unstructured peer-support groups (31), although attendance was higher for group psychoeducation and was associated with improved outcomes for individuals with fewer episodes. Group psychoeducation was more costly than peer support (32).

Kallestad and colleagues (33) compared three sessions of individual psychoeducation and 10 sessions of group psychoeducation, on the basis of Colom and Vieta’s (18) model. Although there were no significant group differences, there was a significant interaction between group and diagnosis. Compared with individuals with bipolar disorder type I, individuals with bipolar disorder type II benefited less from either intervention, and individuals with bipolar disorder type II receiving individual psychoeducation fared significantly worse than those receiving group psychoeducation.

Not all investigations have reported positive effects of psychoeducation. de Barros Pellegrini and colleagues (34), de Azevedo Cardoso and colleagues (35), and Gumus and colleagues (21) all compared group psychoeducation to treatment as usual and found no significant group differences. Of note, de Barros Pellegrinelli et al. reduced Colom and Vieta’s program from 21 weekly sessions to 16 sessions held twice weekly, resulting in a shorter intervention duration (34). They noted that, although the intervention did not result in a decreased risk of recurrence, the interventions were perceived as beneficial by both clinicians and patients. Gumus and colleagues decreased psychoeducation to only four sessions (21).

In aggregate, there is a strong level of evidence for psychoeducation as a maintenance treatment (4), and psychoeducation may be one of the most effective psychotherapies for bipolar disorder, especially type I (36). Psychoeducation appears to be most effective when administered in a group format (23, 26, 30, 33), when delivered over several months (23, 34), and when administered early in the course of illness (31). When resources are limited, psychoeducation may be a reasonable alternative to more costly individual modalities (30). Readers are referred to Colom and Vieta’s (18) and Bauer and McBride’s (17) manuals for full descriptions of psychoeducation.

The core objectives of CBT are to identify and change maladaptive thoughts, beliefs, and behaviors that contribute to and escalate symptoms (37). CBT for bipolar disorder rests on the premise that thoughts, feelings, and behaviors are interconnected and that shifts in mood and cognitive processes during affective episodes influence behavior, all of which can lead to a vicious cycle that contributes to illness burden. Modification of automatic thoughts and elimination of distorted thinking interrupt the problematic affective cycle. CBT includes psychoeducation, thought records, mood diaries, and activity schedules. In CBT, the clinician helps an individual link mood and thoughts, recognize and monitor symptoms and prodromes, develop behavioral strategies for symptoms, learn basic CBT techniques, improve sleep and activity routines, improve medication adherence, and resolve psychosocial problems (38, 39) (Table 1). CBT can be administered individually or in a group format (CBT-G) and has been adapted in many ways to include other strategies, such as MBCT, or to target comorbid conditions, such as integrated group therapy (IGT) for substance use in bipolar disorder.

We identified 13 RCTs with in-person CBT or cognitive therapy (CT) as the experimental condition. Two studies were described earlier in the section on psychoeducation and are not repeated here (25, 30). For this review, we did not include variants of CBT, such as MBCT or IGT. Below, we provide brief summaries of each report and detail particularly noteworthy findings. We also discuss results of a recent meta-analysis, which included CBT and modifications.

In the earliest trial of CBT for bipolar disorder, Cochran and colleagues compared six sessions of CBT with treatment as usual (40). At six months, compared with treatment as usual, CBT was associated with significantly lower rates of discontinuing lithium against medical advice, fewer hospitalizations, and fewer episodes precipitated by lithium noncompliance.

Scott (41) investigated 25 sessions of individual CT versus a wait-list (WL) control group. At six months, CT was associated with significantly greater improvements in symptoms and functioning, compared with those of the WL group. In the WL group, relapse rates in the 18 months after commencing CT were 60% lower than those in the 18 months before commencing CT.

Lam and colleagues investigated 12–20 sessions of individual CT over six months versus treatment as usual (42). At 12 months, compared with treatment as usual, CT was associated with significantly fewer affective episodes, higher social functioning, fewer mood fluctuations, and better medication compliance. In a larger follow-up study, Lam and colleagues again compared individual CT with treatment as usual (38, 43). CT was administered in 12 to 18 sessions over six months, followed by two booster sessions. At 12 and 30 months, the risk of relapse was significantly lower in the CT group compared with the treatment-as-usual group. At 30 months, CT was associated with a significantly longer time to depressive but not manic or hypomanic relapse.

Ball and colleagues found that, after six months of 20 sessions of CT or treatment as usual, individuals receiving CT had lower depression scores compared with individuals in treatment as usual, but the benefits of CT diminished somewhat over 12 months (44).

Miklowitz and colleagues conducted STEP-BD, a multisite project that compared the efficacy of three individual psychotherapies—CBT, IPSRT, and FFT—to a three-session psychoeducation control intervention as treatment for acute bipolar disorder depression (39). For the psychoeducation control condition, participants received an informational video and companion workbook, and sessions focused on a review of these materials. After controlling for site, family involvement, and bipolar disorder subtype, assignment to any one of the three intensive psychotherapies—compared with psychoeducation—was associated with significantly higher rates of recovery, shorter time to recovery, and greater improvement in functioning. CBT, IPSRT, and FFT were equally efficacious for both bipolar disorder types I and II.

Scott and colleagues compared 22 sessions of CBT to treatment as usual (45). They reported that, by 18 months, there were no group differences in recurrence, duration of illness episodes, or mean symptom severity. CBT was significantly more effective than treatment as usual in delaying recurrences for individuals with fewer than 12 previous episodes and significantly less effective for individuals with more than 12 episodes.

Gomes and colleagues investigated the effects of 18 sessions of CBT-G and treatment as usual over a period of 12–24 months (46). There were no group differences in time to recurrence or number of episodes, although median time to relapse was shorter in the treatment-as-usual group than in the CBT-G group.

Meyer and Hautzinger compared 20 sessions of CBT and supportive therapy (47). There were no significant group differences. Bipolar disorder subtype was a nonspecific predictor of outcomes. Those with bipolar disorder type II had a higher risk of recurrence and increased risk of depressive relapse. That is, CBT and supportive therapy were both less effective for the management of bipolar disorder type II compared with bipolar disorder type I.

Costa and colleagues found that, compared with treatment as usual, 14 sessions of CBT-G significantly reduced mood symptoms and number of episodes and improved quality of life (48, 49).

Similarly, González Isasi and colleagues investigated the effects of group psychoeducation with CBT-G and treatment as usual over five years (50). Psychoeducation with CBT-G was associated with fewer hospitalizations at 12 months and lower rates of depression and anxiety at 12 months and five years. Over five years, 89% of the treatment-as-usual group had experienced persistent affective symptoms, compared with only 20% for group psychoeducation with CBT-G.

In sum, there is a strong level of support for CBT as an acute treatment for bipolar depression and as a maintenance treatment. Our findings are consistent with Chiang and colleagues’ meta-analysis, which concluded, with small to medium effect sizes, that CBT lowers relapse rates, improves depressive symptoms, reduces mania severity, and improves psychosocial functioning (51). They included 19 RCTs evaluating CBT-based treatments for bipolar disorder, including variants such as MBCT and IGT. They found that CBT was more effective in reducing relapse rates in bipolar disorder type I groups than in combined bipolar disorder type I and type II groups and when the session duration was 90 minutes or longer. Readers are referred to Basco and Rush (37) for a full description of CBT for bipolar disorder.

FFT is a manualized treatment developed by Miklowitz to decrease overall stress for individuals with bipolar disorder by improving family and primary relationship functioning (52) (Table 1). FFT posits that unsupportive and negative interactions and high expressed emotion within the family or primary relationships increase an individual’s stress—which, in turn, increases vulnerability to developing affective symptoms and episodes. FFT involves three core phases: psychoeducation, communication enhancement training, and problem solving. First, the clinician promotes an understanding of the illness, the vulnerability-stress model, and the importance of medication adherence with the individual and his or her family member or other support. Next, the clinician helps them improve their communications through role plays and skill practice. Finally, the clinician strengthens the problem-solving abilities of both the individual and the family member or other support. FFT is typically delivered to a single individual and family (vs. multifamily).

We identified four RCTs evaluating in-person FFT as an acute treatment, a maintenance treatment, or both. One trial, which found similar benefit across active psychotherapies (IPSRT, CBT, and FFT), was discussed within the CBT results section and is not repeated here (39). In the following text, we provide brief summaries of each RCT and findings.

Miklowitz and colleagues conducted the first trial of FFT and compared 21 sessions of FFT to a comparison condition involving two psychoeducation sessions and crisis management delivered over nine months (53, 54). At one and two years, compared with crisis management, FFT was associated with significantly fewer relapses, longer time before relapse, decreased depressive symptoms, and better medication adherence. Families with high expressed emotion benefited most.

In another study, Rea and colleagues compared nine months of FFT to individual treatment (55). At two years, compared with individual treatment, FFT was associated with significantly fewer hospitalizations and recurrences; both treatments had similar effects on time to first relapse.

Fiorillo and colleagues compared an adaptation of FFT for nontertiary settings in Italy to treatment as usual (56). FFT was associated with significantly greater improvement in social functioning and reduction of family burden, compared with treatment as usual.

There is a strong level of support for FFT as an acute treatment for depression and as a maintenance treatment (57). FFT is an efficacious and effective treatment that reduces illness burden for not only the individual but also the family. Readers are referred to Miklowitz’s (52) manual for a full description of FFT.

IPSRT is a manualized treatment developed by Frank to improve mood and stability by resolving interpersonal problems and regulating social rhythms (58). Social rhythms are those daily activities—such as time getting out of bed, first contact with another person, the start of daily activity, dinner, and time going to bed—that are thought to exert an effect on underlying biological rhythms and serve as anchors for dysregulated circadian rhythms. IPSRT is a hybrid treatment, combining the core elements of interpersonal psychotherapy for unipolar depression (IPT) (59) with a behavioral intervention, social rhythm therapy, targeting disrupted social rhythms. IPSRT includes psychoeducation and emphasizes mood and activity monitoring with the Social Rhythm Metric (60) (see Table 1). IPSRT rests on an “instability model” that defines three interconnected pathways to bipolar recurrences: stressful life events, medication nonadherence, and social rhythm disruption. Accordingly, life events contribute to unstable or disrupted daily routines, which lead to circadian rhythm instability and, in turn, the initiation, maintenance, or worsening of affective symptoms (61, 62). In IPSRT, the clinician helps an individual link mood and life events, identify and manage symptoms, mourn the loss of the healthy self ( i.e., who the person would have been without bipolar disorder), resolve a primary problem area (e.g., role transitions, role disputes, interpersonal sensitivities, or grief), maintain regular daily rhythms, and predict and troubleshoot potential precipitants of rhythm dysregulation (e.g., interpersonal triggers). IPSRT typically is administered individually but can be provided in a group format (63), and the therapy has been modified for bipolar disorder type II (64). Changes include increased attention to the rationale for making changes to social rhythms; identification of mood states; regulation of levels of stimulation; and management of grandiosity, emotional dysregulation, and comorbid substance use.

We identified five RCTs evaluating in-person IPSRT as an acute treatment, a maintenance individual treatment, or both. One trial, which found similar benefit across active psychotherapies (IPSRT, CBT, and FFT), was discussed within the CBT results and is not repeated here (39). Here, we provide brief summaries of each RCT and findings.

Frank and colleagues randomly assigned acutely ill individuals to pharmacotherapy with IPSRT or intensive clinical monitoring (ICM) (65). Once stabilized, individuals were randomly reassigned to two years of monthly IPSRT or ICM maintenance treatment. Individuals assigned to IPSRT during the acute phase survived significantly longer without a new episode, regardless of maintenance treatment assignment.

Swartz and colleagues compared 12 sessions of IPSRT to flexibly dosed quetiapine for depression in an unmedicated bipolar disorder type II sample (66). Over 12 weeks, both groups experienced improvements in depressive and manic symptoms, and there were no significant differences between groups. In a follow-up study, Swartz and colleagues compared IPSRT plus placebo to IPSRT plus quetiapine for treatment of bipolar disorder type II depression (7). IPSRT plus quetiapine was associated with significantly faster improvements in depression and manic symptoms, albeit with more side effects than IPSRT alone.

Inder and colleagues compared IPSRT and specialist supportive care for a sample of young adults (ages 15 to 36 years) (67, 68). Specialist supportive care was a manualized treatment of psychoeducation with supportive psychotherapy. After 78 weeks, there were no significant group differences. Both groups had improved depressive symptoms, social functioning, and manic symptoms.

There is a strong level of support for IPSRT as an acute treatment for depression and as a maintenance treatment. Evidence shows that IPSRT is an efficacious treatment that may be most potent during the acute phase of illness (39, 65). When adapted for bipolar disorder type II, IPSRT may be an effective monotherapy for a subset of patients (7, 66). Readers are referred to Frank’s manual (58) and www.ipsrt.org for more information about IPSRT.

Peer support includes group and one-on-one support and encompasses a variety of heterogeneous interventions. It is founded in the belief that a peer’s experiential knowledge and example of recovery is a valuable resource for an individual with a serious mental illness (69, 70). Peer support follows a resource- versus deficit-oriented approach (see Table 1). Of late, the field has been moving peer support from a supportive, user-led program to a structured, supported, and trained intervention (71). Peer support may be administered as a stand-alone adjunctive treatment or combined with other strategies (72).

We identified four RCTs evaluating in-person or virtual peer support with an identified bipolar population. To clarify, there are reviews and meta-analyses of other RCTs of peer support (73, 74), but these interventions are for severe mental illness in general and not necessarily bipolar specific. One peer-support trial was previously discussed in the context of psychoeducation trials (discussed earlier) and is not repeated here (31). In the following text, we provide brief summaries of the RCTs and findings.

Proudfoot and colleagues compared an attentional control condition and an eight-week online psychoeducation program completed alone or with the virtual support of a peer (72). Virtual support was provided entirely by e-mail and focused on self-management across three domains: medical, emotional, and role. E-mails from the peer were restricted to two 300-word communications each week. Outcomes did not differ across the three conditions. Adherence, however, was significantly higher for women, individuals over age 30, and those who had peer support.

Mahlke and colleagues compared one-on-one in-person peer support and treatment as usual over one year among individuals experiencing severe mental illness for the past two years (70). Individuals with peer support attended structured training totaling 192 hours. At one year, compared with treatment as usual, peer support was associated with greater improvements in self-efficacy. There were no group differences for quality of life, social functioning, or number of hospitalizations.

Salzer and colleagues evaluated the effectiveness of trained peer support in comparison with that of treatment as usual in a community living center for individuals with schizophrenia spectrum and affective disorders (75). Over 80% of individuals made at least one contact with peer support, but only about half had at least a second contact. Despite limited participation, engagement with peer support led to a new resource (e.g., employment, education, leisure information, referral, or support) for over half of the participants.

Overall, evidence supports the utility of peer support as a maintenance treatment; however, there are not enough data to evaluate its efficacy as an acute intervention (4). Peer support may be most helpful for improving self-efficacy and reducing isolation and stigma. When resources are limited, peer-support groups may be a reasonable alternative to more costly clinician-led group psychoeducation (31). Peer support may work best when peers receive ongoing training and support. Emerging models of peer support include online resources sponsored by peer advocacy groups such as the Depression and Bipolar Support Alliance (www.dbsalliance.org).