Review of Outcome Studies of Psychotherapies for Bipolar Disorder
Randomized controlled trials (RCTs) have been completed evaluating 4 forms of psychotherapy for adjunctive treatment of bipolar disorder. These empirically validated interventions include: (1) psychoeducation, (2) cognitive behavioral therapy (CBT), (3) family focused therapy (FFT), and (4) interpersonal and social rhythm therapy (IPSRT). Despite some conceptual and procedural overlap, these interventions differ with respect to the frequency and duration of sessions and use of specific interventions. Thus, although all 4 interventions might be expected to capitalize on similar nonspecific therapeutic ingredients, they also might be expected to differentially affect selected aspects of bipolar illness.
The first study of psychoeducation was conducted by Perry and colleagues. These investigators evaluated the impact of 7 to 12 individual psychoeducational sessions addressing medication adherence, early recognition of prodromal symptoms and impending relapse, and rapid intervention to forestall relapse. The 69 remitted bipolar I patients taking mood stabilizers were randomized to receive either treatment as usual (TAU) or "add on" psychoeducational sessions. When compared with the TAU group, patients receiving psychoeducation had a significantly lower risk of manic relapse (18% vs 46%) and significantly better social and vocational functioning. Patients receiving therapy also were at somewhat lower risk for depressive relapse (ie, 31% vs 48%), although this difference was not statistically significant.
The second RCT evaluated a group model of psychoeducation, which consisted of up to 21 sessions over 5 months. A total of 120 outpatients with bipolar I or II disorder receiving ongoing therapy with mood stabilizers participated in the study. Participants were randomized to receive adjunctive treatment with either group psychoeducation or an attention-placebo control group focusing on social support and nondirective discussion. Across 2 years of follow-up, patients randomized to group psychoeducation experienced a significantly lower number of recurrences (67% vs 92%) as well as a significantly longer duration of sustained remission.
The third study, by Simon and colleagues, evaluated a disease management program added to TAU in a group of 440 outpatients with bipolar I or II disorder being treated within a health maintenance organization. The intervention included a psychoeducational group and individual sessions conducted by a psychiatric nurse that focused on care planning, monthly telephone monitoring of symptoms and medication, feedback, and coordination of care with the mental health treatment team. During the first 12 months of study participation, patients randomly assigned to the psychoeducational intervention experienced significantly lower mean mania ratings, as well as a significant reduction in days spent manic than did patients in the TAU condition. Depression ratings and days spent depressed were not significantly reduced by the experimental program, a finding that is consistent with the results of the smaller study of Perry and colleagues. The benefits of the experimental interventions were maintained across an additional year of follow-up and were found to offset the cost of the intervention.
The fourth study, a large multicenter trial conducted within the psychiatric clinics of the Veterans Administration health system, has been completed, although the final results are not yet published. This program, similar to the one studied by Simon and colleagues, blends group and individual psychoeducational interventions. The results of this study will be of great interest to those seeking to apply psychoeducational interventions to public mental health settings.
Cognitive Behavioral Therapy
Although CBT builds upon psychoeducation, the "meat" of the intervention involves helping patients use self-monitoring strategies in order to recognize the relationship between emotions, thoughts, and behaviors relevant to bipolar illness. Patients learn how to improve coping with step-wise behavioral assignments and cognitive restructuring exercises. So far, CBT is the best-studied form of psychotherapy of bipolar: results of 5 RCTs have been published and a paper describing the findings of a sixth study is currently under editorial review. Four of the 5 published RCTs of CBT have yielded unequivocally positive finding, the fifth tria yielded a positive result only among a subset of patients with less highly recurrent illness. As summarized below, the track record of adjunctive CBT for improved prophylaxis of bipolar disorder is compelling.
In the first trial, patients with bipolar I disorder who were receiving ongoing maintenance pharmacotherapy were randomly assigned to receive either 6 individual sessions of CBT or TAU. The intervention focused primarily on patients' thoughts and feelings about taking medication and identifying and overcoming roadblocks to medication adherence. Despite the relatively small size of the study and the short duration of the intervention, Cochran found a large effect favoring CBT, both in terms of improved adherence and reduced risk of hospitalization. For example, only 21% of the CBT-treated patients discontinued medication compared with 57% of the TAU patients.
Scott and colleagues conducted the second study, a preliminary test of a more symptom-focused form of individual CBT. They randomly assigned 42 outpatients with type I or II bipolar disorder to either a waiting list control group or a 6-month, 20-session course of CBT in addition to concomitant therapy with mood stabilizers. Patients entered this RCT in various states of illness activity, with the majority of participants suffering from at least a subsyndromal level of depressive symptoms. At the end of the 6-month randomized experiment, patients in the waiting list control group were able to receive CBT; the entire treated cohort was then followed for 18 months. Compared with the waiting list control group, and in contrast to the studies of psychoeducation, CBT had little impact on manic symptoms, but produced a significant decrease in depressive symptoms, and patients experienced a significantly greater improvement in social and vocational functioning. For patients "crossed over" from the waiting list control group and followed for 18 months along with the 29 CBT-treated patients, outcomes before and after receiving therapy (a so-called mirror-image analysis) showed there was a 60% reduction in bipolar episodes.
The third study, conducted by Lam and colleagues, also examined a 6-month course of individual, symptom-focused CBT. In this trial, 103 outpatients with bipolar I disorder taking mood stabilizers were randomized to receive either CBT (an average of 14 sessions, followed by 2 booster sessions) or TAU. Although the patients were judged to be on a stable medication regimen, the majority were suffering from minor, residual depressive symptoms. During the first 12 months of follow-up, the patients treated with CBT experienced significantly fewer episodes of illness than those who received TAU (75% vs 44% ), as well as significantly fewer hospital admissions and days hospitalized. The follow-up of this study has now been extended out to 30 months, and a cost-effectiveness analysis has been completed. Results confirmed the sustained beneficial impact of CBT, particularly during the first year, and established its cost-effectiveness when added to medication maintenance therapy.
The fourth study evaluated a 6-month course of CBT in 52 patients with bipolar I or II disorder. Again, compared with TAU , there was a significant effect on depressive symptoms with CBT, and, during the subsequent year of follow-up, favorable trends for relapse prevention were noted.
As noted above, the study by Scott and colleagues is the "outlier" among the published studies in that the effect of CBT was not significant in the overall study group. In addition to its outlier status, this study is also noteworthy because it is the largest RCT of CBT completed to date (N = 253) and, in the coinage of these clinical trials, it is the only study that can be considered "pragmatic" (ie, it was conducted in community psychiatric clinics and employed relatively few inclusion and exclusion criteria). The study enrolled patients with bipolar I or II disorders who were receiving ongoing maintenance pharmacotherapy; consenting participants were randomly allocated to 6 months (22 sessions) of CBT or TAU and followed for a total of 18 months. Although the main effect for treatment was not statistically significant across the 18 months of follow-up, a significant treatment-by-illness course interaction was observed. Specifically, CBT significantly reduced the risk of relapse among patients with a past history of no more than 12 lifetime episodes; the opposite pattern was observed among the subset of patients who had experienced more than 12 prior episodes. Thus, CBT had significant therapeutic effects on the patients with less highly recurrent illness courses, but it appeared to have had an iatrogenic impact for those with the worse prognoses.
The sixth and most recent study of CBT was conducted by the investigators of the Systematic Treatment Evaluation Program for Bipolar Disorder (STEP-BD). STEP-BD is a large multicenter research program conducted under the auspices of the National Institute of Mental Health (NIMH) and, like the study by Scott and colleagues, the investigators aimed to conduct a pragmatic study of the role of psychotherapy that was open to a broad range of patients and was conducted in diverse clinical settings. In the STEP-BD psychotherapy study, patients presenting with an acute episode of bipolar depression (either type I or II) were treated with algorithm-based pharmacotherapy (ie, mood stabilizers alone or in combination with antidepressants) and randomized to receive either a minimal intervention (3 sessions of the psychoeducational program, Collaborative Care) or an intensive psychosocial intervention of up to 9 months. What was particularly novel about the STEP-BD psychotherapy study was that those patients allocated to receive intensive psychosocial intervention received CBT, IPSRT, or FFT, based on both the preference of each of the sites and random assignment. A paper describing the results of this study is now under editorial review. However, per agreement with NIMH, results cannot be disclosed prior to publication, so the results of this second large pragmatic trial of psychotherapy in bipolar disorder may not be available until later in 2006.
FFT aims to reduce the risk of bipolar relapses by strengthening social support and reducing expressed emotion and is appropriate for both the spouses and members of the nuclear family of people with bipolar disorder. To date, results of 2 RCTs of FFT have been published and a paper describing the findings of a third study, conducted by the STEP-BD investigators, is under editorial review.
The first RCT randomly assigned 101 recently remitted bipolar I outpatients to 9 months of treatment with either 21 sessions of FFT (n = 31) or a less intensive crisis management condition (n = 70). When compared to the psychosocial control group, patients receiving FFT experienced a significant increase in well days and a significantly lower rate of relapse across the 2-year follow-up (45% vs 73%). The patients who received FFT also experienced significantly lower levels of both depressive and manic symptoms than those in the control group across the 24-month follow-up.
In another RCT of FFT, 53 recently hospitalized patients with bipolar I disorder receiving ongoing pharmacotherapy were randomized to receive either family intervention or a similarly intensive course of individual supportive therapy. Although there was no between-group difference in relapse risk during the first year of the study, a significant difference in hospitalization rates emerged during the second year favoring FFT over individual supportive therapy (60% vs 12%). When the results of the STEP-BD psychotherapy study are published, a third study of FFT will be added to the literature.
Interpersonal and Social Rhythm Therapy
IPSRT focuses on helping patients with bipolar affective disorder develop more regular daily rhythms and to identify and resolve interpersonal difficulties. To date, 2 studies of IPSRT have been completed: a single-site, 2-phase RCT conducted at the University of Pittsburgh Medical Center and the aforementioned pragmatic RCT conducted by the STEP-BD investigators.The study conducted by the Pittsburgh group enrolled 175 bipolar I outpatients experiencing acute depressive, manic, or mixed episodes. All patients were treated by a psychiatrist/psychotherapist team, and pharmacotherapy was conducted according to a detailed algorithm. During the first phase of the study, patients were randomly assigned to receive either adjunctive IPSRT or intensive clinical management (ICM). The initial phase continued until the patient either remitted for 4 consecutive weeks or withdrew from the study; a total of 125 patients achieved stable remission and entered into the second phase of the study, with an average time to stabilization of 19 weeks. The second phase of the study,which extended for up to 2 years, began with a re-randomization to either IPSRT or ICM. Across the 2 phases, patients thus were treated in 1 of 4 sequences: IPSRT followed by IPSRT; IPSRT followed by ICM; ICM followed by IPSRT; ICM followed by ICM. Preliminary analyses demonstrated that, compared with ICM, IPSRT resulted in significantly greater stabilization in social rhythms and sleep-wake cycles, but had no greater effect on treatment adherence or clinic attendance. Final results indicated that the patients who received IPSRT during the first phase of the study were no more likely to recover from the index episode than patients who did not, but subsequently were significantly less likely to relapse during the 2-year follow-up. Of interest, the addition of IPSRT to ongoing pharmacotherapy after stabilization slightly worsened outcomes (albeit not significantly) compared with the patients who received ICM during both study phases.
As only one positive study of IPSRT has yet been published, it has the weakest level of empirical support among the 4 forms of focused psychosocial intervention for bipolar disorder. The forthcoming results of the STEP-BD study are therefore of considerable interest to ascertain whether IPSRT has replicable therapeutic benefit for patients with bipolar affective disorder.
The results from a large number of RCTs that have examined the efficacy of psychotherapy for the treatment of bipolar affective disorder provide a remarkably clear and compelling answer to the question about clinical utility: Patients who receive adjunctive focused psychotherapy are likely to experience a significant reduction in relapse risk and a significant increase in well-time. The fact that 4 distinctly different models of intervention -- psychoeducation, CBT, FFT, and IPSRT -- have shown significant therapeutic effects is strongly suggestive of an overall therapeutic effect. The fact that on 3 different occasions, the magnitude of this effect was greater than that observed in patients randomized to control groups intended to equate exposure to the so-called nonspecific effects of psychotherapy suggests that the added benefit of these focused therapies goes beyond social support and increased therapeutic contact.
Although it is sometimes hazardous to compare the magnitude of therapeutic effects across studies, the magnitude of the effects of adjunctive psychotherapy are generally moderately large and -- with the exception of the second study by Scott and colleaguesthe 4 different approaches described here seem to offer comparable benefits. In fact, the advantage in relapse prevention conveyed by psychoeducation, FFT, CBT, and IPSRT in RCTs compares quite favorably to the effect sizes observed in contemporaneous pharmacotherapy studies.
Whereas it is clear that adjunctive psychotherapies reliably improve longer-term outcomes of bipolar patients treated with pharmacotherapy, the potential benefit for speeding recovery, or actually increasing the likelihood of recovery, remains underdeveloped. This state of affairs is puzzling, given the established utility of CBT and interpersonal psychotherapy for the acute phase treatment of major depressive disorder and the remarkably dismal performance of most first-line antidepressant therapies in studies of bipolar depression. All that can be said at this time is that the results of the preliminary study of Scott and colleagues suggest that CBT may indeed significantly reduce bipolar depressive symptoms, whereas the findings of Frank and colleaguesfailed to confirm such an effect for IPSRT. Again, the findings of the STEP-BD psychotherapy study will be most informative, as all patients entered are in acute bipolar depressive episodes.