Maintenance Model of Integrated Psychosocial Treatment in Pediatric Bipolar Disorder

Copyright @ 2007 American Academy of Child and Adolescent Psychiatry. Unauthorized reproduction of this article is prohibited.

Maintenance Model of Integrated PsychosocialTreatment in Pediatric Bipolar Disorder:

A Pilot Feasibility Study

AMY E. WEST, PH.D., DAVID B. HENRY, PH.D., AND MANI N. PAVULURI, M.D., PH.D.

ABSTRACT

Objective: The chronic and refractory course of pediatric bipolar disorder merits the study of adjunctive psychosocial

interventions designed to facilitate long-term improvements. The objective of this study is to conduct a pilot study of a

maintenance model of the child- and family-focused cognitive-behavioral therapy program (CFF-CBT), which comprises

psychosocial booster sessions and optimized pharmacotherapy, and to assess whether positive effects seen after the

acute phase of treatment could be sustained over time with the use of this model.Method: The study design was an open

trial with the goal of assessing feasibility of such a maintenance model over time. Thirty-four patients 5 to 17 years of age

who underwent CFF-CBT were delivered the maintenance model of treatment over a 3-year period and assessed for

symptom changes (Children_s Global Impressions Scale-Bipolar) and global functioning (Children_s Global Assessment

Scale). Results:Results indicated that participation in the maintenance model of CFF-CBT treatment was associated with

positive effects in symptoms and functioning over the 3-year follow-up period. There were no statistically significant

differences in postacute-phase treatment scores and scores at years 1, 2, or 3 on any study measures, indicating the

maintenance of clinically significant improvements. Conclusions: These findings suggest that maintenance treatment

models are feasible and may help facilitate the long-term management of symptoms. Controlled clinical trials that build on

this model will help advance treatments for pediatric bipolar disorder toward addressing the low recovery and high relapse

rates associated with the disorder. J. Am. Acad. Child Adolesc. Psychiatry, 2007;46(2):205Y212. Key Words: pediatric

bipolar disorder, psychosocial treatment, maintenance treatment.

Pediatric bipolar disorder (PBD) is characterized bymixed mood states, rapid cycling, excessive elation,prominent irritability, and frequent comorbid condi-tions (Birmaher et al., 2002; Findling et al., 2001;Geller et al., 1998; McClellan et al., 1999; Wozniaket al., 1995). This mood instability, not surprisingly, isassociated with behavioral and academic difficulties in

school, poor social skills, conflictual relationships withsiblings, and stressed parents (Geller et al., 2000).Furthermore, the complex presentation of PBD isencumbered with a chronic and refractory course, lowrecovery, and high relapse rates (Geller et al., 2003).Thus, a primary challenge facing treatment researchersin PBD is to develop treatments that target disorder-specific symptoms and functioning and promotesustained remission. This challenge calls for a treatmentapproach that integrates different psychotherapeuticmodalities to target specific problems affecting bothchildren with PBD and their family members across arange of individual, peer, and family domains. Thiskind of comprehensive approach is likely to facilitatelasting improvement.There is an unfortunate scarcity of research investi-

gating long-term maintenance in pharmacological orpsychosocial treatments for PBD. One preliminarypharmacological study of maintenance treatment with

Accepted August 29, 2006.All of the authors are with the Department of Psychiatry at the University of

Illinois at Chicago.Financial support for this study was provided by the Department of Psychiatry

at the University of Illinois at Chicago.The authors would like to acknowledge Julie Carbray, D.N.Sc., and Jodi

Heidenreich, LCSW, for their integral role in conducting this research.Reprint requests to Dr. Amy E. West, Department of Psychiatry, Institute for

Juvenile Research, 1747 W. Roosevelt Rd, Room 155, Chicago, IL, 60608;e-mail: [email protected].

0890-8567/07/4602-0205�2007 by the American Academy of Childand Adolescent Psychiatry.

DOI: 10.1097/01.chi.0000246068.85577.d7

205J . AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 46:2, FEBRUARY 2007


mood stabilizer monotherapy found a median time todropout of 4 months. This was an 18-month follow-upstudy of children 5 to 17 years old with PBD type I or II(Findling et al., 2005). These findings suggest the needfor strategic pharmacotherapy regimens that incorpo-rate alternative methods for addressing the multitude ofsymptoms and complexity of symptoms in PBD.Pavuluri et al. (2004a) tested a systematic evidence-based pharmacotherapy algorithm in children withPBD type I. These children were studied closely for18 months, and if there was emergence of newsymptoms, worsening, or poor response, strategic medi-cation changes were made according to the algorithm.At the end of the study, there were significantdifferences between the algorithm group and thetreatment-as-usual group, with the children on thealgorithm showing greater symptom reduction andbetter overall functioning. However, the response ratewas limited to 68%. These results suggest that althoughpharmacotherapy for PBD is an iterative processrequiring careful follow-up to prevent the loss ofpositive effects, as many as 30% of patients still remainunimproved. Neither of these studies tested thecombined effect of pharmacotherapy with psychother-apy. More long-term follow-up studies and those thatcombine a psychosocial approach with pharmacother-apy are needed to determine whether the potential forpositive outcomes can be maximized.The lack of long-term studies of combination or

psychosocial treatment for PBD may be result of thepaucity of manual-based psychosocial treatments forPBD (e.g., Fristad et al., 2002; Miklowitz et al., 2003).However, long-term studies of evidence-based psy-chotherapies for childhood depression indicate thatestablishing continuation components after acute-phasetreatment may be important in maintaining treatmenteffects. Clarke et al. (1999) conducted a randomizedtrial of the Adolescents Coping With Depression(CWDA) that included booster sessions every 4 monthsand found that for those children who were still activelysymptomatic after treatment, booster sessions acceler-ated their recovery in a 24-month follow-up periodcompared with those who did not receive boostersessions. Another randomized trial of this interventiondemonstrated that initial posttreatment differences infavor of CWDA disappeared after 12 months with nobooster sessions (Rohde et al., 2001). A recent trial ofcognitive-behavior therapy (CBT) for adolescents with

depression (Brent et al., 1999) found that when anaverage of three booster sessions were included,treatment effects lasted at least 24 months beyond theacute treatment phase. However, at the 2-year follow-up, if booster sessions had been discontinued or notadministered, then there was no significant differencebetween CBT and the other treatment conditions.Kroll et al. (1996) demonstrated that the addition of6 months of CBT continuation treatment consisting ofbooster sessions from once every 2 weeks to once permonth, depending on patient need, produced lowerrelapse rates than CBT without the continuation phase.Findings from these studies collectively suggest theimperative need for maintenance models of therapy inchildhood mood disorders to prevent relapse and tosustain positive changes over time.The combination of promising results from main-

tenance psychotherapy in the treatment of depressionand the insufficiency of pharmacotherapy alone intreating some patients with PBD suggests that amaintenance model of treatment for PBD wouldideally include both pharmacotherapy and psychosocialtherapy. The feasibility of a maintenance pharma-cotherapy algorithm was documented in a previousstudy (Pavuluri et al., 2004a). Therefore, the presentstudy focused on developing and testing an adjunctpsychosocial component for use in a maintenancemodel of treatment for PBD. The purpose of testingmodels such as this is to determine whether adding apsychosocial component to maintenance therapy forPBD results in better outcomes for the patient beyondwhat medication accomplishes. The present study wasconducted as an open trial to investigate the feasibilityof a psychosocial treatment component in addition toan evidence-based pharmacotherapy algorithm in help-ing patients sustain healthy symptom status andfunctioning over time. This represents an initial stepin establishing the role of psychosocial treatmentintegrated with pharmacotherapy in maintaininglong-term remission in PBD, with larger randomizedclinical trials to follow.The maintenance phase of the child-and-familyY

focused CBT (CFF-CBT) was developed to meet theneed for a comprehensive and integrative long-termmodel of treatment for PBD. The CFF-CBT methods,in both the acute and maintenance models, integratepsychoeducation, cognitive-behavioral, and interper-sonal psychotherapeutic techniques to address the

WEST ET AL.

206 J . AM. ACAD. CHILD ADOLESC. PSYCHIATRY, 46:2, FEBRUARY 2007


intense psychosocial and interpersonal stress in childrenand families caused by affective instability in PBD. Inits acute treatment module, CFF-CBT combines phar-macotherapy with a 12-session manual-based treatmentcomprising seven core ingredients: routine, affect regu-lation, self-efficacy and coping, coping with depressivecognitions, social skill building, interpersonal problemsolving, and social support. The CFF-CBT acutetreatment model has shown promising effects in a pre-liminary open trial in reducing symptoms and improv-ing overall social and academic functioning at the end ofits acute treatment phase (Pavuluri et al., 2004b).However, it is not known whether these immediateeffects result in sustained remission over time, especiallygiven the chronicity and high rates of relapse in PBD.

The acute phase of therapy is followed by amaintenance phase. This second phase consists ofpsychosocial booster sessions, along with medicationmanagement, delivered in a systematic yet flexible waybased on the patient_s mental status, level of function-ing, and access to the clinic. The hypothesis exploredwas that the delivery of a maintenance treatment model,constituted by booster sessions carefully designed toaddress key challenges specific to PBD, would be afeasible adjunctive treatment and would show promis-ing preliminary results in the maintenance of symptomreduction and functioning. The data analytic planincluded preYpost comparisons of mean levels ofsymptom and function measures.

METHOD

Entry Criteria

Entry criteria for this study included patients with PBD whocompleted the initial CFF-CBT treatment in the Pediatric MoodDisorders Program at the University of Illinois at Chicago (Pavuluriet al., 2004b). Patients were eligible to participate if they receivedadequate initial treatment (defined as participation in at least 8of 12 sessions) and were receiving medication management inour clinic.

Participants

Patients were recruited through the Pediatric Mood DisordersProgram using an institutional review boardYapproved protocol andconsent/assent procedure. Participants included 24 boys and 10 girlsranging in age from 5 to 17 years (mean = 11.33 years; SD = 3.06years). This sample represented 100% of the children who hadparticipated in the original open trial of the acute-phase treatmenteffects. Participants were from predominantly middle-class house-holds; 23 (68%) were European American, 8 (23%) were AfricanAmerican, 2 (6%) were Latino, and 1 (3%) was Asian. Although

CFF-CBT is designed primarily for and is well suited for 8- to12-year-olds, a wider age range of subjects was included as anexploratory exercise. During the maintenance phase of the study,there were no dropouts, although two patients transferred clinicianswithin the clinic and one patient aged out and moved into adult careafter 27 months.

Measures

All of the participants were diagnosed with PBD using theWashington University in St. Louis Kiddie Schedule for AffectiveDisorders and Schizophrenia (WASH-U-KSADS; Geller et al.,1998). The WASH-U-KSADS is a semistructured interview used toestablish a DSM-IV diagnosis (American Psychiatric Association,1994). Most participants (n = 28) received a primary diagnosis ofPBD type 1, three were considered to have PBD type 2, and threewere classified as PBD not otherwise specified (AmericanPsychiatric Association, 1994). Of those who had PBD type I,990% had a most recent episode of mania or hypomania. All butfive participants had other Axis I conditions, with an average of 2.76T 1.0 additional conditions. The most prevalent comorbidconditions were attention-deficit/hyperactive disorder (ADHD;73.5%), oppositional defiant disorder (35%), and learningdisorders (32%).Outcome measures included the Clinical Global Impressions

Scale for Bipolar Disorder (CGI-BP; Spearing et al., 1997) and the

TABLE 1Pharmacotherapy at End of Maintenance Phase (Year 3)

No. (%)Current Daily Dose

(Average), mg

Combinationpharmacotherapya

31 (91)

Type of medicationMood stabilizer 27 (79)Lithium 10 (29) 1,013Lamotrigine 6 (25) 267Divalproex sodium 4 (12) 1,000Oxcarbamazepine 7 (21) 900

Second-generationantipsychotics

26 (76)

Risperidone 14 (41) 0.9Quetiapine 6 (18) 342Aripiprazole 6 (25) 12.5

PsychostimulantsMethylphenidate 4 (12) 19

Antihypertensives 2 (0.06)Guanfacine 1 (0.03) 2Clonidine 1 (0.03) 1

SSRI (Lexapro) 5 (15) 15OtherAtomoxetine 3 (0.09) 62Trazadone (for sleep) 1 (0.03) 25Chlorpromazine 1 (0.03) 300Clonazepam 2 (0.06) 2

Note: SSRI = selective serotonin reuptake inhibitor.aAverage number of medications per patient = 2.5.

MAINTENANCE MODEL FOR BIPOLAR DISORDER



Children_s Global Assessment Scale (CGAS; Shaffer et al., 1983).The CGI-BP items cover a variety of symptom dimensions, such asmania, depression, ADHD, psychosis, aggression, and sleepdisturbances. The CGAS measures children_s general social andacademic functioning. Measures were completed by a studyclinician at the initiation of treatment, at the end of acute-phasetreatment, and in June or July of each summer during the follow-upperiod. Study clinicians were not blinded to treatment, but aseparate blinded rating on the CGI-BP was completed at thebaseline of the maintenance study (acute treatment follow-up) by aboard-certified child psychiatrist outside our specialty clinic todetermine rater reliability. The interrater reliability was excellentfor the overall improvement scale and the six subscales, rangingfrom 0.83 to 1.0, respectively, with a . of 0.96.

Procedure for Maintenance CFF-CBT

The initial acute-phase 12-session treatment was administeredto study participants from July 2000 through January 2001.Participants then transitioned into the maintenance phase oftreatment, which consisted of medication management with astudy psychiatrist and psychosocial booster sessions with one of fourstudy clinicians (a social worker, an advanced practice nurse, andtwo child psychiatrists). Although pharmacological maintenancetreatment was not the focus of this study, Table 1 includes asummary of pharmacological treatment at the end of themaintenance phase for descriptive purposes. All of the maintenancestudy clinicians either had been part of the preliminary study teamor were trained in CFF-CBT. The follow-up study team met weeklyfor 6 months before the initiation of the maintenance model todiscuss issues and problems that arose with their patients in follow-up treatment. The common themes were identified and organizedaround their fit within the framework of the seven core ingredientsof CFF-CBT. These themes were conceptualized as barriers to themaintenance of treatment effects, and the study team carefullydesigned specific interventions for implementation to address thesebarriers to continued treatment success. Table 2 summarizes themost commonly encountered barriers to recovery addressed bybooster session interventions. As is apparent from the table, follow-up issues spanned the range of treatment modules included in theinitial therapy but were geared toward elements of individualpsychosocial and interpersonal functioning that tend to reemergeduring the course of the disorder, such as behavioral difficulties,helplessness and low self-esteem, social impairment, and familystress. As described previously (Pavuluri et al., 2004b), CFF-CBTincludes sessions for parents alone, children alone, and the familytogether, with an emphasis on individual psychotherapy withchildren, parent training and support, and family therapy. Themaintenance model continues to incorporate individual, parent,and family sessions.During the first, second, and third years of the follow-up study,

booster sessions were administered from once per week to onceevery 3 months as determined by patient and family need and accessto the clinic. Each booster session comprised a standard 50-minutepsychotherapy session in which the CFF-CBT therapist readdressedelements of the initial treatment program with the child and thefamily. The amount, timing, and emphasis of treatment varied, butall of the children and parents participated in booster sessionscovering the seven core ingredients of the CFF-CBT programduring the course of the maintenance phase. Follow-up data onpatient_s symptom experience and functioning were collected andrecorded in years 1, 2, and 3 during the maintenance phaseof treatment.

RESULTS

Three years after the initial acute phase of treatment,those patients who received CFF-CBT maintenancetherapy had sustained the positive effects of the initialintervention. Single-sample t tests (two-tailed) indi-cated that those who received the maintenance therapymodel were functioning significantly better than theywere before their participation in the initial interven-tion and maintenance phase. Single-sample t tests wereused because, according to institutional review boardprotocol, all information linking subject data to a namewas destroyed at the end of the preliminary acute-phasetrial. Thus, we were not able to link follow-up data toposttreatment data within single subjects. A single-sample t test design was used as the closest approxima-tion to a paired-sample t test, which would have beenmost appropriate for data. Thus, these results are likelyan underestimation of true effects.Statistical findings are summarized in Table 3.

Compared with pretreatment scores, the follow-upscores on each scale and subscale maintained signifi-cance at the p < .0001 level. In contrast, there was nosignificant difference between posttreatment scores andfollow-up scores 3 years later, indicating that treatmenteffects were maintained over the 3-year period after theintervention. CGAS and CGI-BP data were collectedfor each patient at points during years 1, 2, and 3 of themaintenance phase. The 95% confidence intervals(CIs) for the posttreatment means on the CGAS andCGI-BP were calculated. For both measures, each ofthe follow-up means remained within the 95% CI forthe posttreatment score, indicating no significantchange over time, showing that patients were able tomaintain improvements consistently over time.These findings demonstrate not only statistically

significant differences but also clinically significantdifferences in the severity of symptoms and the abilityto function in everyday life. At year 3, using the clinicalcutoff score for the CGI-BP Severity Index of 3.0, wefound that 83% of participants had responded totreatment and were experiencing no or minimalsymptoms. Specifically, on the CGI-BP, across allsymptom scales (Overall Psychiatric Disorder, ADHD,Mania, Depression, Aggression, and Psychosis) thesechildren went from Bmildly ill[ to Bmarkedly ill[ (onaverage) at baseline to Bnot at all ill[ to Bmildly ill[ (onaverage) after treatment and at follow-up. On theCGAS, children went from an average score indicatingBmajor impairment in functioning in several areas and

WEST ET AL.



unable to function in one area (i.e., at home, at school,with peers)[ to an average score both after treatmentand at follow-up indicating Bvariable functioning withsporadic difficulties in several but not all social areas.[According to the CGAS, before treatment, thesechildren were likely to have been described asBrequiring a special school or hospitalization,[ whereas

after treatment and at follow-up, these children werelikely to be described as Bhaving disturbances thatwould be apparent to those who encounter the child ina dysfunctional setting or time, but not to those who seethe child in other settings.[ These preliminary findingssuggest that the intervention and medication manage-ment/booster sessions may stabilize the disorder so that

TABLE 2Booster Intervention Targets

CFF-CBT Component Barriers to Recovery Booster Intervention Strategies

Routine Difficulties making transitions Prioritizing activities to target in establishing routine; coming up withBmust stick[ times, such as dinner, homework, and sleep timesResidual oppositional behavior

Focus on school attendancePushing limits with bedtimeRe-establishing sleep routine

Affect regulation Uncontrolled aggression Parent education on chronic and ingrained behaviors and howto influence changeExcitability

Focus on quality of lifeIsolative and withdrawnEmpathy training for parents and patients and reminders of effectivecommunication skills

Sarcasm toward family

Strategies for achieving balance between compassion and firmboundaries

I can do it Feelings of helplessness Emphasis on hope and elucidating emotional growth and maturityLow self-esteem Assist in finding positive outlets and activitiesMaladaptive lying and risk-takingbehavior to cover up feelings ofinsufficiency

Introduce the phrase Bislands of competence[

No negative thoughts/livein the now

Continued struggle with academicachievement

Emphasize acceptance of lifelong struggles while normalizingeveryday struggles

Feelings of social inadequacy Identifying and nurturing valuable friendshipsIncorporating new strategies, such as tutoring

Be a good friend/balancedlifestyle

Feelings of grandiosity,bossiness with peers

Encouraging patients to focus on small circle of positive friendshipsrather than aspiring for global popularity

Parents feeling overwhelmed andoverextended

Empathy training and focus on Bgiving[ in friendships, introducingthe concept of Brespect[

For parents, revisiting need for self-care and carving outpersonal time; Beverything in moderation[

Oh, how can we solve it? Regression to old problem-solvingpatterns

Parents are retaught effective problem-solving skills and encouragedto use persuasive influence rather than overt firmness to reducereactivity in children and themselvesFeeling teased and scapegoated

Pep talks to patients by both parents and therapists; validating parent_sefforts and commitment, highlighting any source of conflict andcreatively see strengths in each other among parents

Parents feel it is a losing battleMarital conflict

Parents encouraged to see strengths in children to facilitate collaborativeproblem-solving approach; the magic word is Bcompassion[

Getting the child to apologize after mishapWays to get support Reaching out can become strenuous

process or it gets stale; childrenand/or parents may not continueto nurture relationships(Blet the ball drop[)

Encouraging parents and children to expect and use social supportsystems

Helping them to think of Bplanning[ to connect with support systemsand keeping it alive even in between crisis points

Note: CFF-CBT = child-and-familyYfocused cognitive-behavioral therapy.




these children can function adequately in most socialsettings.

DISCUSSION

The present study explored the feasibility of amaintenance model of CFF-CBT for PBD thatcomprised a psychotherapy adjunct to optimal pharma-cotherapy to address the ongoing psychosocial strugglesin PBD. Challenges that arose during the initial6 months of the follow-up period were consolidatedinto a structured, yet flexible, model that incorporatedthemes from the seven main ingredients of the initialacute-phase program of treatment to address emergingand recurring barriers to the continuity of treatmentsuccess. This model was then used in a systematic wayduring the maintenance phase of the study. Resultsindicated that participants in the maintenance modelof treatment were able to sustain the symptomreduction and improvements in functioning that wereobserved immediately after their participation in the12-session acute phase of CFF-CBT.

Clinical Implications

The most compelling implication of these findings isthat the refractory nature and low recovery ratesassociated with PBD may be preventable through theuse of ongoing integrated pharmacotherapy and

psychosocial treatment. In this case, a model thatincluded an initial injection of intensive psychotherapy,followed by a maintenance phase of psychosocialbooster sessions adjunct to pharmacotherapy, appearedto help patients maintain optimal symptom manage-ment and healthy functioning over a 3-year period.Although preliminary, these results set a precedent forfuture studies investigating this type of integrated long-term model to address this chronic illness.The results of this study have several important

clinical implications in the treatment of PBD. The firstis that there may be need for a maintenance phase oftreatment after any intensive short-term treatmentprogram to sustain treatment effects. This is importantbecause even with substantial improvement initially,patients and parents in this study demonstrated theintermittent recurrence of symptoms and interpersonaldifficulties related to the disorder. To prevent thesedifficulties from becoming barriers to continued successin treatment, clinicians needed to administer periodicbooster sessions to reemphasize themes or skills learnedin the initial acute phase of treatment.Second, these results confirm the importance of

integrated psychosocial and pharmacological treatment.Few scientists or clinicians would argue with theimportance of continued medication management inPBD. In fact, medication management is the currentstandard of care for PBD. However, the tendencytoward relapse characteristic of the disorder, even withcareful medication management (Pavuluri et al.,2004a), necessitates the study of other methods ofintervention designed specifically to prolong treatmenteffects and sustain remission.Whether the integration ofpsychosocial therapy with medication helps to sustainremission by increasing adherence or works by addressingseparate but important psychosocial themes has yet to bedetermined empirically but will be an important area forfuture treatment mechanism studies. For example, itcould be that continued psychosocial intervention overtime promotes the maintenance of treatment gains byincreasing the chances that patients persist in takingtheir medication and remain in treatment. Alternatively,it could be that continued psychosocial interventionworks primarily to promote remission by addressingpsychological and interpersonal themes, such as thedevelopment of a positive self-concept, adaptive copingstrategies, and healthy social relationships, which mayimprove the quality of life and may protect from

TABLE 3Pretreatment, Posttreatment, and Follow-up Scores on CGI-BP

and CGAS

Measure PretreatmentPostacuteCFF-CBT

Follow-up(Year 3)

CGI-BPADHD 4.53 (1.85) 2.41 (1.13)** 2.26 (1.82)Aggression 4.88 (1.53) 2.12 (1.15)** 2.03 (1.72)Mania 5.24 (1.52) 1.85 (1.02)** 2.20 (1.98)Psychosis 2.71 (2.01) 1.59 (1.13)* 1.29 (1.10)Depression 4.26 (1.69) 2.00 (1.91)** 1.66 (1.71)Sleepdisturbance

3.79 (1.73) 1.35 (0.77)** 1.60 (1.31)

CGAS 46.21 (6.92) 57.67 (8.26)** 58.57 (14.89)

Note: n = 34. CGI-BP = Clinical Global Impressions Scales forBipolar Disorder; CGAS = Children_s Global Assessment Scale;CFF-CBT, child-and-familyYfocused cognitive-behavioral therapy;ADHD = attention-deficit/hyperactivity disorder. On the CGI-BP,higher scores indicate more severe symptoms. On the CGAS, higherscores indicate better overall functioning. Standard deviations are inparentheses.

* p < .01; ** p < .0001.

WEST ET AL.



recurrence of the disorder. In addition, maintenancemodels of psychosocial intervention also may work byaffecting symptoms of comorbid disorders, such asanxiety and ADHD, and developmental dysfunctions,such as the poor regulation and control of impulses thatcommonly occur in these patients.

The notion that psychosocial treatments for PBDwork, at least partially, through facilitating the devel-opment of self-efficacy, social skills, and familycommunication and coping is consistent with findingsfrom the depression treatment literature that have foundthat addressing these kinds of psychosocial difficulties infollow-up treatment can prevent relapse (Brent et al.,1999; Kroll et al., 1996). However, the model adoptedin this study is preliminary and does not incorporatemethodology to assess the mechanisms of treatmenteffect. Future controlled clinical trials and treatmentmechanism studies of maintenance models will help toadvance our knowledge of the actual role of psychosocialintervention in long-term follow-up.

The model espoused in this study was intended formaintenance treatment after acute-phase treatment forPBD. However, a topic beyond the scope of this study,but another important direction for future research,may be whether psychosocial impairment and inter-personal impairment not only are consequences of PBDsymptoms but also may be precursors or risk factors forillness onset. In this case, it could be that psychosocialinterventions could be used for children at risk forbipolar disorder to prevent onset or to alleviate severityof illness.

Strengths of this study include the fact that it isprotocol driven and includes multiple points ofmeasurement to assess change systematically over time.In addition, this study used multiple clinicians to delivertreatment to help control for individual therapist effectsin treatment outcomes. Raters were trained to assessreliability on study measures, so data are assumed to beconsistent and valid across raters. Finally, the mostinnovative feature is that this study puts forth a modelfor integrated maintenance treatment in PBD. Thereare, to the best of our knowledge, no maintenancemodels of integrated psychosocial and pharmacologicaltreatment in the field of PBD treatment. Given thechronic and refractory nature of the disorder and thehigh rates of relapse, models that attempt to sustainsymptoms and functional improvements over the courseof development are imperative.

Limitations

The present study does have limitations thatpotentially detract from its impact. First, the studysample is relatively small; therefore, power to detectsmall and medium effects was minimized. In addition,there is no control group. The present study wasdesigned to be an initial open-trial pilot study, andplans for a controlled clinical trial are underway.Third, the therapist rating the outcome measures wasnot blinded to treatment. Nonblinded ratings ofclinical data are troublesome because they allowinflated estimates of efficacy. Multiple raters andblind ratings by another board-certified child psychia-trist to establish rater reliability at the baseline of themaintenance treatment study could mitigate rater biasresulting from unblinded ratings. However, the factremains that follow-up measures were completed byclinicians who were in therapeutic relationships withthe study participants, which could bias both theclinician_s and the participant_s reports. Fourth,because the acute-phase study and follow-up studywere proposed as separate studies, the institutionalreview board required the destruction of data from theinitial acute-phase study immediately after data werecollected. Thus, it was not possible to look at changesover time compared with baseline measurementswithin single individuals because data from themaintenance study could not be linked to data fromthe acute-phase study, except in the examination ofoverall means on the study measures. Fifth, thedifference in frequency of the dosing of the treatmentfrom once per week to once every 3 months makes itdifficult to determine in any systematic way whatintensity and/or dose of follow-up treatment arenecessary to maintain effects. The model was flexibleto accommodate individual follow-up needs, but itsflexibility may hinder the methodical study of therelationship between maintenance treatment and out-comes. Finally, a causal relationship between treat-ment and outcomes cannot be established with thepresent study design. All study participants weretreated concurrently with medication, and the studydesign did not allow us to discern the varianceaccounted for by medication versus psychotherapy inimprovements. Improvement could be the result ofthe natural history of the disorder, stabilization overtime on medication, or additional attention andstructure rather than the specific intervention. It is




possible that for some of these patients, psychosocialissues were the primary presenting problem (ratherthan symptom severity); therefore, these patients mayhave reported positive outcomes when psychosocialimpairments improved, regardless of the impact oftreatment on symptom experience. In this case, thepsychosocial intervention may benefit the patient bystabilizing psychosocial functioning, but not in theway that we hypothesized, which was that PBD wouldbe stabilized through addressing psychosocial themes.Future studies can incorporate a control group,multiple informants on outcome measures, andtreatment mechanism designs to address the limita-tions in the current open-trial design.Despite these limitations, this study represents an

innovative venture into an imperative direction forfuture research. There is a need for research intopsychosocial treatments for PBD that include astructured maintenance model for the integration ofpharmacotherapy and psychosocial therapy to promotesustained treatment gains after initial acute-phasetreatment. These kinds of integrated long-term treat-ments, if empirically validated and well disseminated,have the potential to transform PBD from a chronicdebilitating illness to a disorder that can be managedeffectively while healthy functioning is maintained.

Disclosure: Dr. Henry serves as a consultant on methodology andstatistics for United Biosource Corp. Dr. Pavuluri’s work is supported byGlaxoSmithKline-NeuroHealth, Abbott, and Janssen Research Foun-dation. She is a consultant to Shire, Janssen, Abbott, andGlaxoSmithKline-NeuroHealth. She served on the speakers’ bureausof AstraZeneca, Janssen, Abbott, and GlaxoSmithKline-NeuroHealth.Dr. West has no financial relationships to disclose.

REFERENCES

American Psychiatric Association (1994), Diagnostic and Statistical Manualof Mental Disorder, 4th edition (DSM-IV). Washington, DC: AmericanPsychiatric Association

Birmaher B, Kennah A, Brent D, EhmannM, Bridge J, Axelson D (2002), Isbipolar disorder specifically associated with panic disorder in youths?J Clin Psychiatry 63:414Y419

Brent DA, Kolko D, Birmaher B, Baugher M, Bridge J (1999), A clinicaltrial for adolescent depression: predictors of additional treatment in theacute and follow-up phase of the trial. J Am Acad Child Adolesc Psychiatry38:263Y271

Clarke GN, Rohde P, Lewinsohn PM, Hops H, Seeley JR (1999),Cognitive-behavioral treatment of adolescent depression: efficacy ofacute group treatment and booster sessions. J Am Acad Child AdolescPsychiatry 38:272Y279

Findling R, Gracious B, McNamara N et al. (2001), Rapid, continuouscycling and psychiatric co-morbidity in pediatric bipolar I disorder.Bipolar Disord 3:202Y210

Findling RL, McNamara NK, Youngstrom EA et al. (2005), Double blind18-month trial of lithium versus divalproex maintenance treatment inpediatric bipolar disorder. J Am Acad Child Adolesc Psychiatry 44:409Y417

Fristad M, Goldberg-Arnold J, Gavazzi S (2002), Multifamily psychoeduca-tion groups (MFPG) for families of children with bipolar disorder.Bipolar Disord 4:254Y262

Geller B, Warner K, Williams M, Zimerman B (1998), Prepubertal andyoung adolescent bipolarity versus ADHD: assessment and validity usingthe WASH-U-KSADS, CBCL, and TRF. J Affect Disord 51:93Y100

Geller B, Bolhofner K, Craney JL, Williams M, DelBello MP, Gundersen K(2000), Psychosocial functioning in a prepubertal and early adolescentbipolar disorder phenotype. J Am Acad Child Adolesc Psychiatry39:1543Y1548

Geller B, Craney JL, Bolhofner K et al. (2003), Phenomenology andlongitudinal course of children with a prepubertal and early adolescentbipolar disorder phenotype. In: Bipolar Disorder in Childhood and EarlyAdolescence, Geller B, DelBello M, eds. New York: The Guilford Press

Kroll L, Harrington R, Jayson D, Fraser J, Gowers S (1996), Pilot study ofcontinuation cognitive-behavioral therapy for major depression inadolescent psychiatric patients. J Am Acad Child Adolesc Psychiatry35:1156Y1161

McClellan J, McCurry C, Snell J, DuBose A (1999), Early-onset psychoticdisorders: course and outcome over a 2-year period. J Am Acad ChildAdolesc Psychiatry 38:1380Y1388

Miklowitz D, George E, Richards J, Simoneau T, Suddath R (2003), Arandomized study of family-focused psychoeducation and pharma-cotherapy in the outpatient management of bipolar disorder. Arch GenPsychiatry 60:904Y912

Pavuluri MN, Henry DB, Devineni B, Carbray JA, Naylor MW, Janicak PG(2004a), A pharmacotherapy algorithm for stabilization and main-tenance of pediatric bipolar disorder. J Am Acad Child Adolesc Psychiatry43:859Y867

Pavuluri MN, Grazyk PA, Henry DB, Carbray JA, Heindenrich J,Miklowitz DJ (2004b), Child- and family-focused cognitive-behavioraltherapy for pediatric bipolar disorder: development and preliminaryresults. J Am Acad Child Adolesc Psychiatry 43:528Y537

Rohde P, Clarke GN, Lewinsohn PM, Seeley JR, Kaufman NK (2001),Impact of comorbidity on a cognitive-behavioral group treatment foradolescent depression. J Am Acad Child Adolesc Psychiatry 40:795Y802

Shaffer D, Gould MS, Brasic J et al. (1983), Children_s Global AssessmentScale (C-GAS). Arch Gen Psychiatry 40:1228Y1231

Spearing M, Post R, Leverich G, Brandt D, Nolen W (1997), Modificationof the Clinical Global Impressions (CGI) Scale for use in bipolar illness(BP): the CGI-BP. Psychiatry Res 73:159Y171

Wozniak J, Biederman J, Kiely K et al. (1995), Mania-like symptomssuggestive of childhood-onset bipolar disorder in clinically referredchildren. J Am Acad Child Adolesc Psychiatry 34:867Y876

WEST ET AL.


Documents

Maintenance Model of Integrated Psychosocial Treatment in Pediatric Bipolar Disorder