Bipolar disorder is characterized by a course where recurrence of manic and depressive episodes is common (Maj, 1999; Tohen et al., 2003), and consequently, functioning and quality of life are often chronically impaired (Goldberg & Harrow, 1999; Tohen et al., 2003; Kessler, Chiu, Demler, Merikangas, & Wakers, 2005). Psychiatric comorbidity is common, with a 25% prevalence rate of personality disorder (George, Miklowitz, Richards, Simoneau, & Taylor, 2003; Vieta et al., 2001) and with anxiety disorders co-occurring in 31% to 95% of bipolar individuals (Vieta et al., 2001; Kessler, 1999). These outcomes occur despite quality psychiatric medication management. Controlled and non-controlled studies show that as an adjunctive treatment for persons with bipolar disorder group psychotherapy may benefit relapse rates, symptom levels, adherence to medication, and quality of life, up to 2 years posttreatment (Shakir, Volkmar, Bacon, & Pfefferbaum, 1979; Volkmar, Bacon, Shakir, & Pfefferbaum, 1981; Kripke & Robinson, 1985; van Gent, Vida, & Zwart, 1988; Cerbone, Mayo, Cuthbertson, & O'Connell, 1992; van Gent & Zwart, 1993; van Gent & Zwart, 1994; Graves, 1993; Hallensleben, 1994; Palmer, Williams, & Adams, 1995; Colom, Vieta, Martinez, Jorquera, & Gasto, 1998; Patelis-Siotis et al., 2001; Colom et al., 2003).
Group psychotherapies may be indicated for various reasons. One reason is related to the stigma associated with bipolar disorder, that is, community perceptions as well as social network reactions to psychiatric illnesses (Michalak, Yatham, Kolesar, & Lam, 2006; Hayward, Wong, Bright, & Lam, 2002). The supportive aspects of group and the experience of "universality" (that others share similar experiences) (Yalom, 1995) may be particularly relevant in bipolar disorder. Second, group psychotherapy may be indicated given the salience of psychoeducation in bipolar disorder management. Education communicated by professionals and group members may be experienced as "imparting of information" and "instillation of hope" (Yalom, 1995). The "recapitulation of the primary family group", which likely occurs in group psychotherapy (Yalom, 1995), could indicate a psychodynamic approach that can putatively address this group experience. Psychodynamic group process techniques have been applied in bipolar disorder with benefits reported in uncontrolled studies (Shakir et al., 1979; Volkmar et al., 1981; Kripke et al., 1985; Cerbone et al., 1992; Graves, 1993). Individual psychodynamic psychotherapy treatments for other Axis I and Axis II disorders, varying in length from 2 months to 2 years, demonstrate effectiveness (Burnand, Andreoli, Kolatte, Venturini, & Rosset, 2002) and improved outcomes (Dekker et al., 2005; Svartberg, Stiles, & Seltzer, 2004; Vinnars, Barber, Noren, Gallop, & Weinryb, 2005; Crits-Christoph et al., 2001; Crits-Christoph et al., 1999).
Longer-term psychotherapy treatments for complicated mood disorders may be indicated (Lam, Hayward, Watkins, Wright, & Sham, 2005). Randomized controlled trials with unipolar depression and bipolar disorder have shown that short-term psychotherapy (10 to 25 sessions) can have relapse rates up to 2 years post treatment averaging 54% (Frank et al., 2005; Colom et al., 2003; Miklowitz, George, Richards, Simoneau, & Suddath, 2003; Lam et al., 2005; Scott et al., 2006; Perry, Tarrier, Morriss, McCarthy, & Limb, 1999). In a well-conducted study of a 22-session cognitive therapy for more complicated bipolar disorders (i.e., more than 12 prior episodes), the therapy was not more effective than treatment as usual (Scott et al., 2006). Psychiatric comorbidity can also decrease treatment effectiveness, with greater relapse occurrences or longer time to stabilization in bipolar disorder common (Swartz, Pilkonis, Frank, Proietti, & Scott, 2005; Colom et al., 2004). Relapse rates are problematic in standard pharmacotherapy treatments that are discontinued, for example, a 45% relapse rate within 5 months of treatment discontinuation (Maj, 1999).
In this study, we report initial results of a case study with a matched, non-random comparison group of individuals with bipolar disorder referred for adjunctive psychotherapy treatment. We sought to investigate the benefit of a longer term approach to the treatment of bipolar disorder, using psychodynamic and group process techniques incorporated into psychoeducational and illness management approaches. We adapted the Integrative Outpatient Model (IOM), developed by Nick Kanas for bipolar disorder (further described below, Kanas, 1993; Kanas, 1999). We predicted that individuals who participated in the IOM would have fewer manic and depressive symptoms, more days well per week, and improved overall general functioning, as compared to a control group, at treatment termination and one year follow-up.
Eleven outpatients with bipolar disorder (type I, II or NOS), ages 21 to 70 years old, were recruited from 250 patients enrolled in one site of the National Institute of Mental Health Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD). The STEP-BD was a national, multi-site, long-term efficacy and effectiveness study of bipolar disorder. The STEP-BD assessed patients at each psychiatric visit using a structured rating instrument that included information on mood symptoms, overall clinical status, and overall functioning. For further information on STEP-BD methods and design, see Sachs et al (2003). Integrative Outpatient Model participants were referred by a psychiatrist or psychiatric nurse, and were interviewed by the group psychotherapist (JG) to ascertain fit for group psychotherapy, as a routine clinical procedure. That is, we provided information to patients about group psychotherapy functioning, particularly how it differs from a support group or individual psychotherapy. Described to the patient were group psychotherapy goals and objectives, in particular, goals to provide feedback to other members, openness to receipt of feedback, and sharing of past and within group experiences. Also discussed was the long-term commitment of the treatment. We asked each potential group member about his or her goals for treatment, and there was mutual discussion about whether a group setting might address these goals.
Diagnosis was verified with a structured clinical interview administered by certified psychiatrists, using the Affective Disorders Evaluation (ADE) (1990; Sachs, 1990). The ADE includes a modified version of mood and psychosis modules from the Structured Clinical Interview for Diagnostic and Statistical Manuals (SCID) (1990; Spitzer, Williams, Gibbon, & First, 1990). Inclusion criteria were any form of bipolar disorder, at least 21 years of age, and viewed as a good fit for IOM from the clinical interview. We did not exclude individuals with Axis I and II comorbidity from the study.
Study Design and Treatment
The group, with 7 to 11 active members, was led by a psychologist and a doctoral-level psychology intern, and met weekly for 90 minutes for 16 months, then biweekly for 2 months. There were two phases: psychoeducation, which lasted for 5 sessions, followed by IOM group psychotherapy approach.
Primary goals for the IOM group sessions included: (a) offering psychoeducation to participants about bipolar disorder, (b) discussing illness management issues in bipolar disorder, and (c) focusing on the quality of relationships through group discussion, feedback, active leader participation and discussion of psychodynamic and interpersonal issues. Techniques and interventions following the IOM model guided the treatment. Content for a particular session, after the initial 5 sessions, was not predetermined. Content reflected group member interests, from education and illness management (e.g., medications and side effects, psychiatry visits, diagnostic issues, prognosis and outcome, suicidality and crisis) to interpersonal issues (e.g., current situations with significant others, social support issues), and psychodynamic themes (e.g., group interactions as symbolic of previous relationships, defense mechanisms, processing group dynamics, transference and countertransference). We provide three case examples in the Appendix. The examples include symptoms stemming from psychiatric comorbidity commonly seen in bipolar disorder, such as substance use disorder, anxiety disorder, and/or personality disorder.
Participants and controls were assessed at baseline prior to treatment, termination of treatment (18 months), and one year follow-up (30 months). Controls were selected by matching from the pool of 250 STEP-BD participants on age, gender, and the Global Assessment of Functioning (GAF) score at the baseline visit. Matched controls had no additional psychosocial treatments or visits as part of this study nor were they restricted in receipt of psychosocial care. Thus, the control subjects may or may not have received psychosocial treatment. All participants, including controls, signed consent forms prior to baseline assessments for this Institutional Review Board approved study. All study participants were receiving standard psychiatric treatment as part of the STEP-BD. That is, expert-guided pharmacological treatments for bipolar disorder and any comorbidities. Patients enrolled in the group psychotherapy were not enrolled in any STEP-BD related randomized controlled trials during the study period.
Baseline assessment data were extracted from the psychiatric visit closest to commencement of treatment. Follow-up assessment data were extracted from the psychiatric visit nearest group termination, and then the visit nearest 1-year post treatment. The vast majority of psychiatric visits occurred within one month of the assessment time point. Psychiatrists were not blinded to whether the patient was in a treatment or control condition. We utilized data collected from the STEP-BD project. Psychiatrists were not apprised of which data would be utilized as outcome data, nor were they told at what time during the 30-month period we would extract data (e.g., 18 months, 30 months).
The Global Assessment of Functioning scale and Clinical Global Impression Scale for Bipolar Disorder (CGI-BD) were used to measure overall psychological, social, and occupational functioning. The GAF and CGI scores were assessed by the psychiatrist, rating functioning of the previous month.
Depression and manic symptoms were rated by the treating psychiatrist using a Likert-type scale with a range of 0 (no symptoms) to 2 (severe). Twelve symptoms were rated for depression (e.g., decreased interest, sleep disturbance); nine symptoms for hypomania/mania (e.g., self-esteem, less need for sleep).
Clinical status was assessed at each psychiatric visit using the DSM IV criteria for manic, hypomanic, mixed, or depressive episodes. Clinical status was coded "0" for recovered (with minimal or no symptoms) for at current episode of mania, hypomania, mixed mania, or depression.
Number of Days Well Per Week
To determine the number of days well in any given week, clinical status of well days was assessed by the number of days well between two psychiatric visits. A patient was considered well at a visit if clinical status was "recovered" or "subsyndromal", and the patient was considered sick if clinical status was "manic", "hypomanic", "mixed mania," or "depressed". If the patient was well on two consecutive visits, the number of days well in between visits was all days. If the patient was well at one visit and sick the next, only half the days between visits were coded as well. If the patient was sick at both visits, the number of days well between visits was none. By standardizing the periods between visits, we calculated the number of days well per week to assess change in status throughout treatment. The number of days well (as coded above) was averaged by quarter (3 months).
Descriptive statistics were calculated using chi-square for categorical variables and analysis of variance for continuous variables. Group participants were matched to a control on age, gender, and baseline GAF score prior to treatment. We examined matched pairs to determine if we would conduct analyses with a matching technique or between groups. There were no significant differences among the pairs, so we did not adjust for pair-to-pair differences, and remaining analyses were conducted between groups. For outcome analyses, we conducted between-group and within-group comparisons, repeated measures analyses of covariance were conducted, with one covariate of baseline depressive symptoms. For days-well-per-week analyses, we used a mixed model analysis of covariance with one repeated-measures factor (time) and one between-groups factor (treatment versus control) and an interaction (time by group), and one covariate of baseline depressive symptoms. For follow-up comparisons, treatment termination scores and 1-year follow up scores were averaged. The three dropouts were not included in outcome analyses, as there was no follow-up data available. For between-group analyses, the p value was <0.05. For days well per week, which included multiple comparisons, the p value was set at p<0.005. We conducted statistical analyses with SPSS version 12.0 and SAS version 9.1.2.
Twenty three participants signed Consent to participate; one participant dropped out prior to completing baseline assessments or finding a matched control, thus analyses include the remaining 22 individuals. At baseline, IOM participants had worse depression-symptom scores than controls. Three patients terminated the group early, at 2, 3, and 6 months, with a dropout rate of 27% (3/11). One person terminated at one year and was counted as a completer, the remaining 7 individuals completed 18 months of treatment. In terms of dropout characteristics of the three early terminators, there was no indication that these persons were more symptomatic or had worse functioning overall. Two were lost to follow up in the parent study. Two of the control subjects were also lost two follow-up.
For the one year follow-up period, group members were less likely to be in any mood episode, as compared to controls (F( 1, 11)=6.13, p=0.03, E.S. = 1.08). Group members had fewer depressive symptoms during the follow-up period compared to controls (F( 1, 11)=5.03, p<0.05, E.S. = 0.99). There were no significant differences in manic symptoms, GAF or CGI during the follow-up period. The CGI showed an effect size of 0.66, in favor of the psychotherapy treatment. There were two psychiatric hospitalizations per group; one occurred during the treatment period and one during the follow-up period; this was the case for both groups.
The number of days well per week improved significantly in the psychotherapy group, but not in the control group. (F=2.24, p=0.025). Group members began treatment averaging 3.5 days well per week. By the beginning of the second year, the group improved to an average of 5.9 (quarter 5), 6.3 (quarter 6), 6.4 (quarter 7), 6.1 (quarter 8), 6.1 (quarter 9), and 5.8 (quarter 10) days well per week. Comparatively, the control group showed no significant improvements in days well per week. The case examples describe other outcomes (intrapsychic and interpersonal), possibly a result of the IOM psychotherapy.
Individuals who completed a long-term psychotherapy group demonstrated improvements in clinical status, depression symptoms, and the number of days well per week, up to one year post termination. The group was feasible in terms of acceptability of the modality and approach by the members, with a dropout rate of 27% over a year's time. At baseline, individuals referred for psychotherapy were more likely to have significant depressive symptomatology, even when matched to controls on age, gender, and GAF score. Referring persons with bipolar disorder, who have chronic depression and do not respond satisfactorily to standard pharmacological treatments, is consistent with treatment guidelines.
One interpretation of the outcomes might be that patients simply recovered due to the STEP-BD pharmacotherapy treatments or perhaps, as a natural course of the disorder. In the one year follow-up period, patients' recovery, in terms of days well, occurred about 1 year into group treatment (5.9) and was sustained during the follow-up period (ranging from 5.8 to 6.4 days well per week). That the group therapy patients sustained their recovery as a group over one year of follow up suggests a treatment effect. Additionally, between-group effect sizes for the follow-up period for three various indicators were 1.08, 0.99, and 0.66 in favor of the treatment group. This also suggests that the group simply did not recover to baseline scores similar to the control group, but improved over and above the symptom levels of the control group.
In this study, patients' manic symptoms and global level of functioning did not differ from the control group, a finding consistent with recent psychotherapy studies of bipolar disorder (Lam et al., 2005; Miklowitz et al., 2003), but contradictory to other studies of bipolar disorder (Colom et al., 2003; Perry et al., 1999). There are several potential explanations for these findings. One relates to patient recruitment and inclusion criteria in a particular clinical trial. In studies consistent with our findings, patients recruited had a recent bipolar mood episode or were currently in an episode (Lam et al., 2005; Miklowitz et al., 2003). These entry criteria were consistent with our criteria. In the studies that found more clinical benefits with manic symptoms, euthymic patients (Colom et al., 2003) or generally less chronic patients (Perry et al., 1999) were entered into the study. Psychosocial treatments with a potentially less ill group may have differential effects on outcomes for depression and mania, similar to how pharmacotherapy strategies also differ based on the symptom profile. It may be that psychoeducation, such as the need for greater adherence to antimanic agents, has lasting effects on the prevention or curbing of manic episodes. Psychotherapies with intensive strategies addressing cognitive, behavioral, family, and relationship difficulties may see their greatest benefit in depression, the more residual, chronic and chronically disabling aspect of bipolar disorder (Bauer, Kirk, Gavin, & Williford, 2001; Gao & Calabrese, 2005).
There are several limitations of this small pilot study. The small sample size may have lessened the capability to detect meaningful group differences. We recruited patients from a sample of 250 STEP-BD patients, thus generalization of these findings to all persons with bipolar disorder is limited. Patients were not randomized to treatment, and those assigned to group had more symptoms. Randomization may better identify the effects of an active treatment. The measure of days well per week was a broad assessment of clinical status and was not systematically measured across time. The control group did not receive a weekly intervention, thus we cannot attribute effects to the aspects of the IOM model. In future studies, manualization of the IOM should further specify its active ingredients, or components, and the treatment approach. Finally, the raters (psychiatrists) were not blind; however, the data used to collect outcomes were based on standardized psychiatric visits and psychiatrists were not apprised of which data were extracted over a 2.5 year period.
The IOM model allows for the flexibility of content and theoretical approaches varying from psychoeducation, group process, psychodynamic, and interpersonal techniques. Psychodynamic psychotherapy, which has shown effectiveness in cocaine dependence, major depression, personality disorders, and Axis I disorders comorbid with personality disorders (Bond & Perry, 2006; Dekker et al., 2005; Svartberg et al., 2004; Vinnars et al., 2005; Burnand et al., 2002; Crits-Christoph et al., 2001; Crits-Christoph et al., 1999), should be investigated for individuals who may have not only a chronic illness, such as bipolar disorder, but interpersonal and intrapsychic conflicts exacerbating the illness. One concern about psychodynamic psychotherapy treatment is it considered by some to be too lengthy a treatment modality. However, our perspective differs: Pharmacological treatment of bipolar disorder--the primary intervention--is long-term and typically indefinite. It seems counterintuitive that an adjunctive treatment, such as psychotherapy, would be recommended at a lower dose or intensity than the primary treatments in prevention of relapse and response to acute illness.
-Gonzalez, Jodi M.; Prihoda, Thomas J.. American Journal of Psychotherapy, 2007, Vol. 61 Issue 4
Reflection Exercise Explanation
Goal of this Home Study Course is to create a learning experience that enhances
your clinical skills. We encourage you to discuss the Personal Reflection
Journaling Activities, found at the end of each Section, with your colleagues.
Thus, you are provided with an opportunity for a Group Discussion experience.
Case Study examples might include: family background, socio-economic status, education,
occupation, social/emotional issues, legal/financial issues, death/dying/health,
home management, parenting, etc. as you deem appropriate. A Case Study is to be
approximately 150 words in length. However, since the content of these Personal
Reflection Journaling Exercises is intended for your future reference, they
may contain confidential information and are to be applied as a work in
progress. You will not
be required to provide us with these Journaling Activities.
The article above contains foundational information. Articles below contain optional updates.
Reflection Exercise #1
The preceding section contained information about psychodynamic group psychotherapy for bipolar disorder. Write three case study examples
regarding how you might use the content of this section in your practice.
Psychiatric comorbidity occurs in what percentage of bipolar individuals? Record the letter of the correct answer the