1 THE THERAPEUTIC ALLIAN CE IN COGNITIVE-BEH AVIORAL TREATMENT OF PEDIATRIC OBSESSIVE-COMPULSIVE DISORDER By MARY LORRAINE KEELEY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009
2 2009 Mary Lorraine Keeley
3 To Frederick Shoucair
4 ACKNOWLEDGMENTS I would like to thank Gary Geffken, Ph.D. a nd Eric Storch, Ph.D. for their guidance and m entorship; my committee, Drs. Jim Johnson, Ken Rice, and Lori Waxenberg, for their encouragement and contributions; my lab colleag ues for their friendship and collaboration; and the children and families who participated in this study. Finally, I would like to thank my family and friends for their love and support.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ............................................................................................................... 4 LIST OF TABLES ...........................................................................................................................7 ABSTRACT ...................................................................................................................... ...............8 1 INTRODUCTION .................................................................................................................. 10 Introduction .................................................................................................................. ...........10 Pediatric Obsessive-Com pulsive Disorder .............................................................................11 Cognitive-Behavioral Treatment of Pedi atric Obs essive-Compulsive Disorder .................... 13 Therapeutic Alliance in the Psyc holo gical Treatment of Youth ............................................ 18 Study Objectives and Hypotheses .......................................................................................... 24 2 METHOD ........................................................................................................................ .......26 Participants .................................................................................................................. ...........26 Measures ...................................................................................................................... ...........27 Demographic Form ..........................................................................................................27 Childrens Yale-Brown Obsessive-Compulsive Scale .................................................... 27 Therapeutic Alliance Scale for Children ......................................................................... 27 Working Alliance Inventory ............................................................................................ 28 Homework Compliance Rating Form ............................................................................. 28 Child Behavior Checklist ................................................................................................29 Procedures .................................................................................................................... ...........29 Pre-Treatment Assessment .............................................................................................. 30 Treatment ..................................................................................................................... ....30 Treatment Assessments ................................................................................................... 31 Post-Treatment Assessment .............................................................................................31 3 RESULTS ....................................................................................................................... ........33 Sample Characteristics ............................................................................................................33 Hypothesis 1: TA Predicting Outcome ................................................................................... 33 Hypothesis 2: TA Discrepancy Scores Predicting Outcom e ..................................................35 Hypothesis 3: Alliance Shif ts Predicting Outcom e ................................................................36 Hypothesis 4: TA Predicti ng Hom ework Compliance ...........................................................38 Hypothesis 5: TA and Homework Com pliance Predicting Outcome ..................................... 38 Hypothesis 6: Associations among TA and Pre-Treatm ent Characteristics ........................... 39 Post Hoc Analyses ..................................................................................................................39
6 4 DISCUSSION .................................................................................................................... .....49 Hypothesis 1: TA Predicting Outcome ................................................................................... 49 Hypothesis 2: TA Discrepancy Scores Predicting Outcom e ..................................................52 Hypothesis 3: Alliance Shif ts Predicting Outcom e ................................................................54 Hypothesis 4: TA Predicti ng Hom ework Compliance ...........................................................55 Hypothesis 5: TA and Homework Com pliance Predicting Outcome ..................................... 57 Hypothesis 6: Associations among TA and Pre-Treatm ent Characteristics ........................... 58 Limitations ................................................................................................................... ...........59 Study Implications ............................................................................................................ ......61 Conclusions .............................................................................................................................63 LIST OF REFERENCES ...............................................................................................................65 BIOGRAPHICAL SKETCH .........................................................................................................76
7 LIST OF TABLES Table page 2-1 Assessment schedule ............................................................................................................32 3-1 Means (standard deviati ons) and ranges of study variables ................................................. 41 3-2 Session 1 TA predicting outcome: Hierarch ical regress ion analyses after controlling for pre-treatment symptom severity ..................................................................................... 42 3-3 Session 5 TA predicting outcome: Hierarch ical regress ion analyses after controlling for pre-treatment symptom severity ..................................................................................... 43 3-4 Session 5 TA predicting subsequent ch ange in sy mptom severity: Hierarchical regression analyses after controlling for prior change in symptom severity ........................ 44 3-5 TA discrepancy scores predicting outcom e: Hierarchical regression analys es after controlling for pre-treatment symptom severity ................................................................... 45 3-6 TA early alliance shift change scores predicting outcom e: Hierarchical regression analyses after controlling for pr e-treatment symptom severity ............................................ 46 3-7 TA and homework compliance predicting ou tco me: Hierarchical regression analyses after controlling for pre-treatment symptom severity .......................................................... 47 3-8 Correlations for session 1 TA and pre-treatm ent characteristics .......................................... 48
8 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE THERAPEUTIC ALLIAN CE IN COGNITIVE-BEH AVIORAL TREATMENT OF PEDIATRIC OBSESSIVE-COMPULSIVE DISORDER By Mary Lorraine Keeley August 2009 Chair: Gary R. Geffken Major: Psychology Although cognitive-behavioral therapy (CBT) fo r pediatric obsessive-compulsive disorder (OCD) is considered a fi rst-line treatment for the disorder, research reveals suboptimal remission and treatment non-responder rates, suggesting the need to investigate factors that may enhance or reduce treatment effects. Few studies have inve stigated predictors of treatment response in pediatric OCD, and there is an absence of studi es examining the influence of treatment process variables (e.g., therapeutic alliance [TA]) on tr eatment outcome. Using a multiple-informant and multiple-time point design, our st udy examined the role of the TA in family-based CBT for pediatric OCD. Analyses examined the predic tive value of the TA on homework compliance and OCD symptom reduction. Furthermore, analyses investigated the ro le of discrepant alliances (i.e., differences between child-therapist and pare nt-therapist alliances) on outcome and whether changes in the TA over time predict treatment re sponse. Our main findings were that (1) stronger child-rated, parent-rate d, and therapist-rated TA s were predictive of better treatment outcome, (2) large discrepancies in child-the rapist versus parent-therapist TAs were associated with poorer treatment outcome, (3) larger and more positive early alliance shifts (as rated by changes in child-rated TA between sessions 1 and 5) were predictive of better treatment outcome, (4) therapist-rated child-therapist TA and therapist-ra ted parent-therapist TA at session 1 predicted
9 homework compliance across treatment, (5) both TA and compliance ratings were uniquely predictive of treatment outcome, and (6) with the exception of child age being correlated with the therapist-rated parent-therapist TA, no measured pre-treatmen t child characteristics were predictive of TA ratings. Implications for the treatment of youth with OCD within family-based CBT are discussed based on study findings.
10 CHAPTER 1 INTRODUCTION Introduction Obsessive-com pulsive disorder (OCD) is an an xiety disorder characterized by the presence of recurrent obsessions and/or co mpulsions that are distressing, time consuming (more than one hour per day), and/or debilitati ng (interfere with normal func tioning) (American Psychiatric Association, 2000). With recent epidemiologi cal studies reporting prevalence rates of approximately 1-4% among children and adolescents (Douglass, Moffit, Dar, McGee, & Silva, 1995; Zohar, 1999), OCD is now considered on e of the most common childhood psychiatric illnesses (Stewart et al., 2004). Research has indicated a bimodal age of onset distribution, with the initial peak incidence occurring during th e pre-pubescent period (termed early-onset OCD) and the second during early adu lthood (termed late-onset OCD) (Pauls, Alsobrook, Goodman, Rasmussen, & Leckman, 1995). Approximately 80% of adults with OCD e xhibit initial symptom onset prior to 18 years of age (Riddle, 1998; Pauls et al., 1995), a nd individuals who are untreated experience a chronic and fluctuating course (Besiroglu et al., 2006; Zohar, 1999). High rates of comorbidity in pediatric OCD complicate the clinical presentation and exacerbate OCDrelated impairments in functioning (Masi et al ., 2006; Storch et al., 200 7a; Suchodolsky et al., 2005). Recent studies reported comorbidity rates as high as 80% (Barrett, Healy-Farrell, & March, 2004; Pediatric OCD Treatment Study [POT S], 2004; Storch, et al., 2007a), with major depression, anxiety disorders, tic disorders, attention deficit h yperactivity disorder (ADHD), and disruptive behavior disorders be ing among the most common comorbid disorders (Storch et al., 2008a).
11 Pediatric Obsessive-Compulsive Disorder The heterogeneous and idiosyncratic nature of obsessions and com pulsions, coupled with low levels of insight, complicates the clinic al presentation in youth with OCD (American Academy of Child and Adolescent Psychiatry [ AACAP], 1998). Obsessions are persistent and intrusive thoughts, ideas, impulse s, or images that result in in creased anxiety and accompanying distress. Compulsions are repetitiv e or ritualistic behaviors or me ntal acts aimed at reducing or preventing anxiety in response to the obsessive thought. The most common symptom subtypes include contamination/washing, doubting/checkin g, obsessions without overt compulsions (e.g., sexual, aggressive, or religious obsessions), and hoarding (Sookman, Abramowitz, Calamari, Wilhelm, & McKay, 2005). Although certain com pulsions are typically linked with certain obsessions, some cases do not follow this pattern, such as certain early-onset cases who exhibit compulsions without endorsing a ny identifiable obsession (Geller et al., 1998). Early-onset cases are more likely to be male, have a family history of OCD, present with tic-related compulsions (e.g., tapping, touching, rubbing), have a comorbid tic disorder, and engage in reassurance seeking behaviors (G eller et al., 1998). Cognitive models of OCD suggest that patients with OCD have a tendency to misinterpret the content of intrusive thoughts in a way that leads to a worsening spiral of symptoms (Salkovskis, 1999). Patients with OCD evidence difficulty in c ognitively inhibiting intrusive thoughts, and therefore they se lectively attend to intrusive thoughts (Tolin, Abramowitz, Przeworski, & Foa, 2002). Such selective attentio n increases the accessibil ity of these thoughts, which results in heightened anxiety and discom fort, which in turn l eads to counterproductive attempts (compulsions) aimed at decr easing these thoughts (Salkovskis, 1999). Behavioral models of the development and maintenance of OCD symptoms can be traced back to Mowrers (1939; 1960) tw o-stage theory, which posits that (1) fears are acquired due to
12 neutral stimuli becoming paired with feared s timuli via classical conditi oning, and (2) fears are maintained due to the engagement of behavi ors (i.e., compulsions) that function to reduce distress associated with the c onditioned stimuli via negative reinforcement. However, the reduction in anxiety resulting from the engagement in rituals is only temporary, due to the persistent, intrusive, and recurr ent nature of obsessive thoughts and images. Consequently, youth seeking relief from their obsessive-compulsive symptoms paradoxically become trapped in a time-consuming and incessant cycle of obsessions and compulsions that results in significant distress and impairment in func tioning (Carter & Pollock, 2000). Youth with OCD frequently i nvolve their parents and othe r family members in this pathological cycle, resulting in the further maintenance of obsessive-compulsive symptoms (Storch et al., 2008b). In addition to actively pa rticipating in rituals, family members may accommodate OCD symptoms by facilitating compulsive avoidance of feared stimuli, modifying family routines to accommodate rituals, and tole rating delays associated with ritual completion (Storch & Merlo, 2006). Approximately 80% of family members of patients with OCD accommodate the patients symptoms to some degree, and higher rates of accommodation are associated with higher levels of family dysfunc tion and family distress (Calvocoressi et al., 1995; Calvocoressi et al., 1999; Storch et al., 2008b). Research investig ating the effects of OCD on the family has found that parents of children with OC D, compared to parents of children with other clinical disorders or healthy cont rols, exhibit higher levels of e xpressed emotion, lower levels of warmth, and decreased encouragement of child autonomy or problem-solving (Barrett, Shortt, & Healy, 2002). Researchers have posited that pe diatric OCD and family functioning influence each other in a bi-directional patt ern (Freeman et al., 2003), which leads to increased tension and frustration within family relationships (Geffk en et al., 2006; Waters & Barrett, 2000). Thus,
13 researchers emphasize the importance of including fa mily members in the treatment of pediatric OCD to mitigate the impact that OCD has on the entire family (Storch & Merlo, 2006). In addition to impairments in family func tioning associated with OCD, children with OCD suffer detrimental effects in both academic and social functioning as well (Flament et al., 1988; Leonard et al., 1993; Piacentini, Bergman, Keller, & McCracken, 2003), although the severity of impairments is variable across individuals (Piacentin i et al., 2003). These impairments may have an accumulating negative impact over time, with early OCD-related impairments in functioning linked to later in terpersonal and psychological difficulties in adulthood (Thomsen & Mikkelsen, 1995; Flamen t, Koby, Rapoport, & Berg, 1990). Given the significant and long-lasting effect s of OCD on overall functioning a nd quality of life for both the patient and his or her parent(s), participat ion in effective treatment is essential. Cognitive-Behavioral Treatment of Pe diatric Obsessiv e-Compulsive Disorder Cognitive-behavioral therapy (CBT) for OCD in youth involves four primary components: psychoeducation, cognitive training, mapping of OCD, and exposure with response prevention (E/RP; AACAP, 1998; March & Mulle, 1998; Storch et al., 2007a; Storch et al., 2006). Treatment lasts approximately 14 sessions, and can be delivered in a weekly or intensive format (Lewin, Storch, Adkins, Murphy, & Geffke n, 2005). Psychoeducation reviews the neurobehavioral etiology of OCD and assists patients in labeling and monitoring symptoms. Cognitive training teaches patients to identify and restructure anxiety-provoking cognitions, engage in constructive self talk, cultivate nonattachment, a nd externalize blame. Cognitive training is adapted to the childs developmental level; training with younger children tends to focus on teaching constructive self talk aimed at increasing the patients sense of self-efficacy and predictability, whereas training with older ch ildren who have the cognitive ability to identify and challenge cognitions tends to focus more on cognitive restructur ing techniques. Mapping
14 OCD involves constructing a hierarchy of OCD symptoms that will be used for E/RP. In E/RP, patients are gradually exposed to anxiety-provoking stimuli while refraining from engaging in compulsive behaviors or ritu als (Meyer, 1966). Extended exposur e allows for habituation of anxiety and provides data for youth to disconfirm distorted fear representations, which are characterized by inaccurate expectations of ha rm and responsibility (Foa & Kozac, 1986). Parent involvement in therapy is an important aspect of succe ssful treatment (Barrett et al., 2004; Freeman et al., 2003; Knox, Albano, & Ba rlow, 1996; Storch et al., 2007a). During treatment, parents are provided with psychoeducation regarding the neurobehavioral etiology of OCD to reduce unnecessary blame and criticism as well as to decrease parental accommodation of child symptoms (Freeman et al., 2003). Parents are also taught operant procedures to reinforce effort and adaptive coping skills through praise and rewards, as we ll as to extingu ish reassurance seeking behaviors through removal of parental attention (March, Franklin, Nelson, & Foa, 2001). Furthermore, parent involvement in treatment is essential for generalizati on of therapeutic gains to other settings (e.g., home) (March, Frankli n, Nelson, & Foa, 2001). Finally, parents learn the importance of modeling, and they are instructed to decrease the extent to which they model anxious interpretations of even ts or avoidance behaviors. Recent research examining the effectivene ss of CBT for pediatric OCD has revealed promising results for the treatmen t of youth. A recent meta-analysi s of 18 studies identified CBT and selective serotonin reuptake inhibitors (SSR Is) as evidence-based treatments for pediatric OCD (Abramowitz, Whiteside, & Deacon, 2005). Fi ndings from this meta-analysis revealed larger effect sizes for CBT when compared to SSRIs and placebo. Additionally, youth receiving CBT tended to have a higher percentage (62%) of clinically significant improvement when compared to youth receiving SSRIs (38%).
15 In addition to several unc ontrolled studies supporting the effectiveness of CBT for pediatric OCD in individual, family, and gr oup formats (e.g., Franklin et al., 1998; March, Mulle, & Herbel, 1994; Piacentini, Bergman, Jacobs McCracken, & Kretchma n, 2002; Storch et al., 2006; Thienemann, Martin, Cregger, Thom pson, & Dyer-Friedman, 2001), four randomized controlled trials (RCTs) comparing CBT, pha rmacological treatment, and placebo/waitlist conditions have established strong support for the efficacy of CBT, pharmacological treatment, and their combination (Barrett et al., 2004; de Haan, Hoogduin, Buitelaar, & Keisjers, 1998; POTS, 2004, Storch et al., 2007a). In an RCT co mparing E/RP to clomipramine in youth with OCD, results revealed that both treatments resu lted in significant re duction of symptoms as assessed by the main outcome measure in pedi atric OCD (Childrens Yale-Brown Obsessive Compulsive Scale [CY-BOCS]). Fu rthermore, results indicated that the group receiving CBT ( n = 12) had a greater response rate (defined as pre to post improvement on CY-BOCS > 30%) (66.7% versus 50.0%) and a greater mean reduc tion in symptom severity on the CY-BOCS (59.9% versus 33.4%) than patients treated with clomipramine ( n = 10), providing support for the superiority of CBT over clomipramine (de H aan et al., 1998). Effect sizes were 1.58 and 1.45 for the behavior therapy and cl omipramine groups, respectively. In an RCT study of individual family-based CBT ( n = 24), group family-based CBT ( n = 29), and waitlist control ( n = 24), Barrett et al. (2004) reporte d an 88% response rate (as defined by no longer meeting diagnostic crit eria for OCD) for children in the individual CBT condition, a 76% response rate for children in the group CBT condition, and a 0% response rate in the waitlist control condition. In a long-term follow-up study, Barrett, Fa rrell, Dadds, and Boulter (2005) indicated that in bo th the individual ( n = 22) and group ( n = 26) conditions, treatment gains were maintained at 12-month and 18-month follow-ups, with a total of 70% of participants in
16 individual therapy and 84% of pa rticipants in group therapy dia gnosis-free at follow-up (Barrett et al., 2005). In a randomized trial of weekly ( n = 20) versus intensiv e (daily sessions; n = 20) familybased CBT, Storch et al (2007a) re ported that both modes of tr eatment resulted in clinically significant reductions in OCD symptoms at post-treatment and 3-month follow-up. Response analyses (as assessed by a post-treatment CY-B OCS score < 10) indicated a high rate of responders in both conditions, with 75% of youth in the intensive group and 50% of youth in the weekly group meeting remission status criteria at post-treatment. At 3-month follow-up, 72% of youth in the intensive group and 77% of youth in the weekly group met remission status criteria. Effect sizes conducted in this study revealed post-treatment and 3-month follow-up effect sizes of 2.62 and 1.73 for the intensive group, and 2.20 and 2.33 for the weekly group. The authors concluded that both intensive and weekly CBT are efficacious treatments for pediatric OCD, with the intensive group displayi ng slight immediate advantages in symptom improvement at post-treatment (Storch et al., 2007). Finally, in a recent multi-site RCT, i nvestigators compared CBT alone ( n = 28), sertraline alone ( n = 28), the combination of CBT and sertraline ( n = 28), and placebo ( n = 28) in a sample of youth with OCD (POTS, 2004). Response analyses (assessed by a pos t-treatment CY-BOCS score < 10) indicated that the combination group had a significantly higher rate of responders (53%) compared to the sertraline alone group (21%) and the placebo group (3%). The combination group and the CBT alone group did not differ significantly with regard to response rates, with 39% of the CBT alone group classified as responders. Effect sizes conducted in this study revealed effect sizes of 1.14, .97, and .67 fo r the combination group, CBT alone group, and
17 sertraline alone group, respectively, suggesting the relative superi ority of the combination of CBT and sertraline over monothera py, and suggesting the superior ity of CBT over sertraline. Taken together, findings suggest that CBT de livered alone or in combination with SSRIs should be the first-line treatment of pediatric OCD (POTS, 2004). However, despite the relative effectiveness of CBT for OC D, many patients do not res pond following a full course of treatment, and others remain symptomatic de spite some therapeutic gains. Findings from controlled and open trials of CBT demonstrate sub-optimal remission rates ranging from 25% to 54%, and pre-post-treatment change scores indi cate that between 14% and 40% of patients do not significantly improve (i.e., are non-res ponders) following treatment (de Haan, 2006). In an effort to enhance treatment response rates, researchers have begun to investigate possible predictors of poor response to identify specific targets for additional or adjunctive interventions. Treatment outcome research in pediatric OCD has indicated greater OCD symptom severity (Piacentini et al., 2002), higher levels of OC D-related academic impairment (Piacentini et al., 2002), higher levels of fa mily accommodation (Storch et al, 2007b), higher levels of family dysfunction (Barrett et al., 2005), and the presence of a comorbid non-anxiety disorder diagnosis (Storch et al., 2008a) as pred ictors of worse outcome. Overall, these results suggest that those children with a more complex and severe presentation are at particular risk for poor response to CBT. Given findings of severa l family-related predictors of poor outcome, involving family members in treatment seems critical for optimal reduction in symptoms. There has been considerably more work conducted in the examin ation of treatment response predictors in the adult OCD literature. Ho wever, much of this research is characterized by inconsistent findings. Despite th is, several pre-treatment variab les have emerged as relatively reliable indicators of worse treatment respons e. Symptom severity (Franklin, Abramowitz,
18 Kozak, Levitt, & Foa, 2000; Ke ijsers, Hoogduin, & Schaap, 1994; Mataix-Cols, Marks, Greist, Kobak, & Baer, 2002), comorbid severe depres sion (Abramowitz, Franklin, Street, Kozak, & Foa, 2000; Abramowitz & Foa, 2000; Foa, 1979; Steketee, Chambless, & Tran, 2001), symptom presentation characterized by religious/sexual obsessions (Alonso et al., 2001, Mataix-Cols et al., 2002; Rufer, Fricke, Moritz, Kloss, & Ha nd, 2006) or hoarding (A bramowitz, Franklin, Schwartz, & Furr, 2003; Rufer et al., 2006), comorb id personality disorder (Fricke et al., 2006; Moritz et al., 2004; Minichiell o, Baer, & Jenike, 1987; Steket ee et al., 2001), and family dysfunction (Chambless & Steketee, 1999; Emme lkamp, Hoekstra, & Visser, 1985; Steketee, 1993) have been the most consistent predictors of poor response. In addition, two treatmentrelated variables have emerged as reliable pred ictors of treatment response: the therapeutic alliance (TA; Hoogduin, de Haan, & Schaap, 1989; Keijsers et al., 1994; V ogel, Hansen, Stiles, & Gotestam, 2006) and patient compliance (Abr amowitz, Franklin, Zoellner, & DiBernardo, 2002; De Araujo, Ito, & Marks, 1996; Tolin, Ma ltby, Diefenbach, Hannan, & Worhunsky, 2004; Whittal, Thordarson, & McLean, 2005). These latter tw o variables seem particularly relevant as intervention targets, as it may be more feasible fo r clinicians to focus on enhancing aspects of the TA or patient compliance than to attempt to mo dify pre-treatment characteristics, such as symptom subtype. However, no study of CBT ou tcome in pediatric OCD has examined these two variables and their potential influen ce on treatment response in youth with OCD. Therapeutic Alliance in the Psychological Treatment of Youth Although no study has exam ined the TA in pediatric OCD, research has examined the influence of the TA on treatment outcome for a numb er of other psychiatric disorders. The TA is a more sophisticated form of the therapeutic relationship, as it encompasses cognitive (e.g., commitment to goals) and behavioral (e.g., collabor ative engagement in the tasks of treatment) aspects of the interaction in addition to affective (e.g., bond) aspects of the therapist-patient
19 interaction (Shirk & Saiz, 1992). Researchers have purported that the examination of the TA in treatment outcome studies may be especially im portant for youth treatment, as children rarely refer themselves to treatment, often do not ackno wledge the existence of problems, and may be at odds with parents regardi ng the goals of ther apy (Shirk & Saiz, 1992). This may be particularly true for anxiety-pr ovoking treatments, such as CBT fo r OCD, as children with OCD often lack insight into their symptoms (Geller et al., 1998) and are less likely to understand the long-term benefits of engaging in a therapeutic task that caus es discomfort, because children, cognitively speaking, are less future-orien ted than adults (P iacentini, 1999). A recent meta-analysis of the TA indicated that few ( n = 14) child treatment outcome studies assess the TA (Karver, Handelsma n, Fields, & Bickman, 2006), despite the large magnitude ( n = 1500+) of controlled studies of child psychological treatment (Kazdin, 2002). These studies have demonstrat ed the utility of assessing the TA, with studies reporting significant effects on treatment outcome (Karve r et al., 2006). Child, parent, and therapist perspectives of the therapeutic relationship and a lliance are uniquely related to outcome (Shirk & Karver, 2003), and these effects are evident ac ross developmental levels (child versus adolescent), service types (research versus co mmunity based), types of treatment (individual, parent training, family therapy), modes of tr eatment (behavioral versus nonbehavioral, and manual-based versus non-manual based), and treatm ent contexts (inpatient versus outpatient) (Shirk & Karver, 2003). Research has indicated that the type of presenting problem is a clinical moderator of the alliance-outcome relation, with the relation having a str onger effect for patients presenting with externalizing versus internalizing problems (Shirk & Karver, 2003). Studies examining the specific influences of the TA on outcome have indicated that a strong TA is positively associated with symptom improveme nt (Hawley & Weisz, 2005; Hogue, Dauber,
20 Stambaugh, Cecero, & Liddle, 2006; Florsheim, Shotorbani, & Guest-Warnick 2006; Hughes & Kendall, 2007; Kazdin, Marciano, & Whitley, 20 05; Kazdin & Whitley, 2006; Kazdin, Whitley, & Marciano, 2006; McLeod & Weisz, 2005; Tetzlaff et al., 2005), treatment retention (Garcia & Wiesz, 2002; Hawley & Weisz, 2005; Kazdin, Holland, & Crowley, 1997), treatment satisfaction (Hawley & Weisz, 2005), and treatme nt acceptability (K azdin et al., 2005; DeVet, Kim, CharlotSwilley, & Ireys, 2003). Despite these significant findi ngs, researchers have highli ghted a host of methodological inadequacies that characterize the research on th e TA in child treatment outcome studies. First, most studies of the alliance-outcome relation in ch ild treatment studies fail to utilize a multipleinformant approach. Green (2006) encourages the assessment of the TA from multiple informants (i.e., child, parent, therapist, obs erver) to decrease the influence of shared measurement inflation bias on the interpretation of results. Additionally, assessing the TA from multiple informants may yield interesting data pertaining to how discrepant alliances (e.g., differences between parent-rated alliances a nd child-rated alliances) potentially influence outcome. For example, research examining the in fluence of discrepancies between child-rated and parent-rated TAs on treatment outcome has demonstrated that greater discrepancies are predictive of dropout (Robbins, Turner, & Al exander, 2003) and less symptom improvement (Shelef, Diamond, & Guy, 2005). Identifying discrepa ncies in alliances may be particularly relevant for the treatment of pediatric OCD, in which parental involve ment is a significant component (Barrett et al., 2004). Children ma y be hesitant to be gin an anxiety-provoking treatment, especially if they lack insight into their symp toms, and thus, their willingness to develop a strong TA may be limited. On the ot her hand, parents of children with OCD often present to treatment motivated to engage in th erapeutic tasks aimed at reducing their childs
21 symptoms, and therefore may be more likely to de velop a strong TA. If the parent-therapist TA is substantially greater than the child-therapist TA, the degree to which symptoms are reduced may be negatively affected, as children may recognize the unbalanced TAs and react in an oppositional manner by refusing to comply with therapeutic tasks. A second methodological issue concerns the d earth of studies that assess the TA at multiple time points during treatm ent (Kazdin & Nock, 2003). As a result, few studies have evaluated the directionality of e ffects. Therefore, it cannot be know n for certain whether it is the strength of the TA that predicts symptom improvement, or it is symptom improvement that predicts the strength of the TA. Measuring the TA at multiple time points will permit further investigation into how the quality of the TA changes over time, and such information will be influential for the development of interven tions aimed at building or improving the TA (Diamond, Diamond, & Liddle, 2000). Research in this area has indi cated that changes in the TA over time (i.e., improvements or declines) are more predictive of outcome than early alliance levels (Hogue et al., 2006). Within CBT for OC D, it may be that the introduction of anxietyprovoking exposures exercises in th erapy affects the nature of th e established TA. However, no study of CBT for OCD has examined how the TA changes over time and whether these potential changes in the nature of the alliance affect outcome. Third, researchers have underscored the need for studies to include other variables (e.g., pre-treatment characteristics, treatment-related fact ors) that may influence or may be influenced by the strength of the TA. The addition of other variables in studies of the alliance-outcome relation permits the examination of the relative contributions of posited predictors of treatment outcome (Green, 2006), and may shed light on the ways in which the TA influences outcome (Kazdin & Whitley, 2006). In an adult study of the influence of the TA, motivation, and
22 treatment expectancies on treatment outcomes in CBT for OCD, only the TA significantly predicted treatment outcome, sugges ting that the strength of the TA has greater predictive power than either motivation or expe ctations (Vogel et al., 2006). Although several studies identified patient comp liance as a reliable pr edictor of treatment outcome in CBT for OCD, no research has examin ed both the TA and patie nt compliance within the same study. Researchers have examined these two variables within th e context of CBT for non-OCD anxiety disorders, and results indicate d that the therapistrated child-therapist relationship was uniquely predictiv e of anxiety disorder severity at post-treatment and 1-year follow-up whereas therapist-rated homework compliance was not (Hughes & Kendall, 2007), suggesting the relative importance of the ther apeutic relationship within CBT for anxiety disorders. Although informative, this study only incorporated therapist ratings of the childtherapist relationship, and did not assess the th erapist-parent relations hip. Furthermore, both measurement of the relationship and of complian ce were based on a 1-item questions (i.e., How would you describe the quality of the therapeuti c relationship during the session? and Rate the childs degree of compliance w ith the homework task.) rather than on already established measures of the TA and homework complian ce (Hughes & Kendall, 2007). More research is needed in this area to bett er understand these factors within CBT for pediatric OCD. Such research would allow investigator s to evaluate the relative predic tor power of these two variables as well as to examine whether the strength of the TA influences the degree of patient compliance, or vice versa. Knowledge of these influences has implications for the manner in which therapists approach the establishment of rapport, the formation of the TA, and the discussion of the importance of co mpliance with therapeutic tasks.
23 Finally, the investigation of the TA in he terogeneous populations and diverse treatment approaches threatens the internal validity of the results and limits the degree to which results can be interpretable for a specific patient population. A ssessing the TA within more homogeneous populations is especially important given findings that type of presenting problem moderates the alliance-outcome relation (Shi rk & Karver, 2003). Unfortunate ly, few studies have been conducted that examine the TA within the cont ext of anxiety-provoking treatments for children. Kendall (1994) examined the influence of the therapeutic relationship (affective component only) on outcome in CBT for youth with non-OCD anxiety disorders. Results indicated that for the subset of individuals who ra ted the therapeutic relationship as one standard deviation below the mean, the therapeutic relationship was pred ictive of worse outcome on measures of child anxiety and depression. However, because Kendall (1994) restri cted the assessment of the therapeutic relationship to one time point (pos t-treatment) and one informant (child), the interpretation of these findings is limited. Furthe rmore, the study did not assess the cognitive and behavioral components of the alliance, which may be of particular importan ce in treatments that require children to engage in anxiety-provoking tasks. In another study of the th erapeutic relationship within CBT for non-OCD anxiety disorders, Hughes and Kendall (2007) found that th e therapist-rated child-therapist relationship, averaged across all treatment sessions, was a si gnificant predictor of outcome. In a study of alliance-building behaviors, Creed and Kendall (2005) reported that ther apists attempts at collaborating with the patient (i.e., setting goals with child, agreeing on therapeutic tasks) was the behavior most predictive of a positive TA, as judged by all th ree raters (observers, children, and therapists). Thus, assessing the larger construct of the TA versus only the therapeutic relationship, may yield more in teresting findings regarding its influence on outcome in the
24 treatment of child anxiety disord ers. However, because none of th e participants in these studies had a primary diagnosis of OCD, the field has no information regarding th e influence of the TA in the treatment of pediatric OCD. Study Objectives and Hypotheses The prim ary objective of this study was to examin e the role of the TA in CBT for pediatric OCD, and this objective was divided into 5 specific aims. First, this study aimed to examine whether the TA, as rated separately by the child, parent, and therapist, would predict treatment outcome (as defined by reduction in OCD symptoms). Given findings from adult studies that have indi cated the TA as a significant predictor of CBT outcome (Hoogduin et al., 1989; Keij sers et al., 1994; Vogel et al., 2006), it was hypothesized that stronger child-rated, parent-ra ted, and therapist-ra ted TAs would predic t better treatment outcome. Second, this study sought to investigate whet her discrepancies betw een child-rated and parent-rated TAs would predic t treatment outcome. In accordance with preliminary data suggesting an association between greater discrepancies in allia nce ratings and worse outcomes (Robbins et al., 2003; Shelef et al., 2005), this study hypothesized that greater discrepancies between child-rated and parent-rated TAs w ould predict worse treatment outcome. Third, this study aimed to examine whether ch anges in the strength of the TA over time would predict treatment outcome. Based on studies that have repor ted on the therapeutic benefits of strengthened versus weakened alliances in samples of deli nquent adolescents (Hogue et al., 2006; Florsheim et al., 2000), it was hypothesized that improvements in the TA over time would be predictive of better outcome, whereas declines in the TA over time would be predictive of worse outcome.
25 Fourth, this study sought to examine the relation between the TA and homework compliance. Given that one aspect of the TA involves developing the motivation and commitment to engage and collaborate in ther apeutic tasks, it was hypothesized that stronger TAs would lead to higher leve ls of homework compliance. Finally, this study aimed to assess the rela tive contributions of the TA and homework compliance on treatment outcome. This study hy pothesized that both the TA and patient compliance would uniquely and positivel y predict treatment better outcome. A secondary objective of this study was to investigate pre-treatment characteristics that are associated with the TA. Identifyi ng those pre-treatment characteris tics that are associated with weak TAs will aide clinicians in recognizing specific patient presentations that may be at risk for poor treatment response. Thus, based on past studies analyzing the correlates of weak TAs (Eltz, Shirk, & Sarlin, 1995; Kazdin & Whitley, 2006), this study hypothesized that high levels of child social problems and oppositionality w ill be correlated with weaker TAs.
26 CHAPTER 2 METHOD Participants Twenty-five youth aged 7 to 17 years old ( M = 13.16, SD = 2.69) with a principal diagnosis of OCD and their parent s participated in our study. Partic ipants (11 females; 14 males) were recruited from families who presented to the University of Florida (UF) OCD Clinic. Eighty percent of the sample was Caucasian, 8% was African American, 8% was Hispanic, and 4% was biracial. The average ag e of onset for the sample was 10.48 years of age. Inclusion criteria were as follows: Principal diagnosis of OCD CY-BOCS Total Score 16 Able to read study measures (as determ ined by having participant read measure instructions out loud) No change in psychotropic medication (if appl icable) for at least 8 weeks prior to study entry Children were excluded if they me t any of the following criteria: History of and/or current psychosis, bi polar disorder, or current suicidality Autism or mental retardation A current diagnosis in the careg iver of mental retardation, psychosis, or other psychiatric disorders or conditions that would limit their ability to understand CBT Three of the 25 participants dropped out of th e study prematurely. Completers and dropouts did not differ based on age, gender, or ethnicity.
27 Measures Demographic Form Parents completed a general dem ographic form to assess demographic variables, including childs age, gender, ethnicity, car egiver education and occupation, family income, marital status, and childs age of OCD onset. Childrens Yale-Brown Obsessive-Compulsi ve Scale (CY-BOCS; Scahill, Riddle, McSwiggen-Hardin, & Ort, 1997) The CY-BOCS is a 10-item clinician-rate d measure designed to assess OCD symptom severity and response to treatment in children. It ems are anchored on a 0 to 4 point scale, and they assess the degree to which obsessions and co mpulsions occupy the patients time, interfere with functioning, cause subjective distress, are actively resiste d, and can be controlled by the patient. The CY-BOCS has demonstrated good to ex cellent levels of in ternal consistency ( = .87) and inter-rater reliab ility (rs ranging between .68 .91), and evaluations of divergent and convergent validity have indicated that the measur e is valid (Scahill et al., 1997; Storch et al., 2005). Additionally, the CY-BOCS has been shown to be sensitive to therap eutic change (Scahill et al., 1997). Cronbachs alpha for the CY-BOCS total score in our study was .81. Therapeutic Alliance Scale for Ch ildren (TASC; Shirk & Saiz, 1992) The TASC is a 12-item measure of the childs perception of the quality of the TA in child psychological treatment. Items are anchored on a 0 to 4 point scal e, and they assess the childs affective experience of therapy as a positive bond, negative affective experience to therapy itself, and collaboration with tasks of therapy. Child (e.g., I like spe nding time with my therapist) and therapist (e.g., The chil d likes spending time with you, the ther apist) versions exist to capture the unique and subjective perspectives of both child and ther apist (Shirk & Karver, 2003). The original measure was designed to measure the alliance across treatment, and it has since been
28 revised (TASC-R; Creed & Kendall, 2005) to assess the TA after individual treatment sessions. The TASC-R has demonstrated acceptable intern al consistency across st udies, with Cronbachs alpha coefficients ranging from .85 to .96 (C reed & Kendall, 2005; Kazdin & Whitley, 2006). In our study, Cronbachs alpha for the child-rated and therapist-rated versions of the TASC-R were .91 and .98, respectively. Working Alliance Inventory (W AI; Horvath & Greenberg, 1989) The W AI is a 36-item measure of the quality of the TA in adult psychological treatment. The measure was adapted to be a suitable assessme nt of the parent-therapi st alliance in child psychological treatment (Kazdin et al., 2005). Items are anchored on a 1 to 7 scale, and they assess the degree of positive pe rsonal attachment, acceptance, and trust in the therapeutic relationship, the extent to which parent and ther apist agree on and endorse therapeutic goals, and the extent to which parent and therapist agree on and accept responsibility for therapeutic tasks. Parent (e.g., My childs therapis t and I are working towards mu tually agreed upon goals) and therapist (e.g., My patients pa rent and I are working towards mutually agreed upon goals) versions exist to capture the uni que and subjective perspectives of both parent and therapist (Shirk & Karver, 2003). The WAI is a widely used measure, with proven reliability and validity (Horvath & Bedi, 2002). It has recently been us ed in the context of child therapy, and has demonstrated acceptable psychometric properties in this context (e.g., median Cronbachs alpha of .92) (Kazdin et al., 2005; Kazdin & Whitley, 2006). In our study, Cronbachs alpha for the parent-rated and therapist-ra ted versions of the WAI were .87 and .93, respectively. Homework Compliance Rating Form The Hom ework Compliance Rating Form was cr eated according to Abramowitz et al.s (2002) protocol for assessing homework comp liance in CBT for adult OCD. The measure assesses the therapists ratings of patient comp liance with homework on a scale ranging from 0
29 (did not complete any assigned homework) to 6 (completed all homework and made efforts above and beyond assignments). Child Behavior Checklist (CBCL; Achenbach & Rescorla, 2001) The CBCL is a 118-item parent report measure designed to assess a wide range of child emotional and behavioral problems. Using a 0 (not true) to 2 (very or often true) rating scale, parents are asked to describe their childs behavior and emotional functioning over the past 6 months. Factor analyses have yielded reliable an d valid individual narrow subscales and broadband Internalizing and Extern alizing subscales, in addition to a Total Problems composite score. For the purposes of this study, the Social Problems syndrome subscale and the Oppositional Defiant Problems DSM-IV orie nted subscale were used. The CBCL has demonstrated acceptable to excellent psychometri c properties across a variety of clinical and non-clinical populations, with Cr onbachs alpha coefficients fo r subscales ranging from .79 to .94 (Achenbach & Rescorla, 2001). Cronbachs alpha for the Oppositional Defiant Problems DSM-IV oriented subscale in our study was .70. Cronbachs alpha for the Social Problems syndrome subscale in our study was .73. Procedures Participants were enrolled in part of a larg er study examining predictors of CBT outcome at the UF OCD Clinic. During the initial clin ical interview conducted by a clinical child psychologist, families who met study criteria were asked if they would like to hear about a study examining CBT for pediatric OCD. If they answ ered yes, families completed human subjects informed consent procedures approved by the UF institutional review board and were scheduled for their pre-treatment assessment.
30 Pre-Treatment Assessment A trained research assistant, under the supervision of a clinical child psychologist, adm inistered the CY-BOCS to the parent and child jointly. The research assistant then provided instructions on completing the st udy measures. Immediately followi ng this assessment, the child, accompanied by the parent, began the first session of CBT. Treatment Treatm ent consisted of up to 14 sessions of CBT for OCD based on the protocol used in POTS (2004), which includes psychoeducation, cognitive training, and exposure with response prevention. Treatment was terminated early if both the therapist a nd the patient agreed that an optimal level of symptom reduction had been ac hieved. In our study, the average duration of treatment was 12 sessions. The first session of CBT for OCD consisted of rapport building, information gathering, and psychoeducation regard ing the nature of OCD and the rationale for treatment. Session 2 consisted of therapeutic tasks that include the creation of a fear hierarchy and cognitive training. Sessions 3 and beyond consis ted of both therapis t-assisted and homebased exposure exercises with response preventi on. An important part of the treatment entailed the assignment of homework, which focused on e ither self-monitoring ex ercises or exposure exercises. Treatment was provided by 9 clinical child psychol ogy postdoctoral fellows ( n = 3) or doctoral candidates ( n = 6) within the UF OCD Clinic. All therapists had previous training and experience in implementing CBT for pediatric OCD. Supervision was held on a daily (for participants in intensive treatment; n = 10) or weekly (for participants in weekly treatment; n = 15) basis to ensure competence and adherence to the treatment protocol Therapists included 3 males and 6 females.
31 Treatment Assessments Treatm ent assessments were conducted followi ng the first and fifth treatment sessions. After completion of the treatment session, a resear ch assistant (not the therapist) met privately with the child and parent and asked them to complete the TASC and WAI, respectively. The child and parent were instructed to insert the completed measur es in a sealed (confidential) envelope and were notified that their therapist would not have access to their responses. After the fifth session, a research assistant also administ ered the CY-BOCS. The therapist completed the therapist versions of the TASC and WAI after the first and fift h sessions. Additionally, after each session, the therapist rated homework compliance. Post-Treatment Assessment At the post-treatm ent assessment, which immediately followed the last treatment session, a research assistant administered the CY-BOCS along with other treatment-related measures. Additionally, the child and parent privately completed the TASC and WAI respectively, and the therapist completed the therapist versions of these measures. Table 2-1 represents a visual representation of the assessment schedule.
32Table 2-1. Assessment schedule Pre-treatment 1 2 3 4 Session 5 6 7 8 9 10 11 12 13 14 Posttreatment CY-BOCS CBCL TASC WAI C C C CY-BOCS TASC WAI C C C C C C C C C C CY-BOCS TASC WAI CY-BOCS Childrens Yale-Brown Obsessive Compulsive Scale, CBCL Child Behavior Checklist, TASC Therapeutic Alliance Scale for Children, WAI Working Alliance Inventory, C Compliance ratings.
33 CHAPTER 3 RESULTS Sample Characteristics Means, stan dard deviations, and ranges for each study measure are shown in Table 3-1. To evaluate the normality of the TA variables, we analyzed the skewness and kurtosis of each alliance variable. Analyses indi cated that all session 1 and most 5 TA variables (child-rated, parent-rated, and therapist-rated) were roughly univariate normal (skewness and kurtosis estimates less than two). Post-treatment TA vari ables were somewhat skewed (statistics between -1.41 to -2.66) and highly kurtotic (statistic s between 1.05 and 8.39), with post-treatment TA variables tending to be rated more uniformly positive. Scores on the CY-BOCS were reduced across treatment, with an overall Cohens d effect size of 2.89. Fifty-nine percent of our studys sample achieved remission status (CY-BOCS < 10) at the end of treatment. All measures demonstrated acceptable internal consistency; alpha coefficients ranged from .70 to .98. Independent group t -tests and one-way analyses of vari ance (ANOVAs) indicated no significant differences on measures of TA and CY-BOCS as a function of gende r, ethnicity, or the presence of a comorbid psychiatric disord er. Correlational analyses indicat ed that age was significantly correlated with session 1 therapist ratings of th e parent-therapist TA. Therefore, in analyses incorporating the session 1 therapist-rated parent -therapist TA as a pred ictor variable, age was entered as a covariate. Child age was not significa ntly correlated with any other ratings of the TA or the CY-BOCS at the first, fifth, or final sessions. Hypothesis 1: TA Predicting Outcome Using hierarchical regression analyses, we evaluated whether the strength of the TA predicted treatm ent outcome, as measured by pos t-treatment CY-BOCS scores, after controlling for pre-treatment symptom severity (i.e., pretreatment CY-BOCS scores). Results of these
34 analyses are shown in Tables 3-2 and 3-3. For an alyses conducted for se ssion 1 ratings of the TA, findings indicated that the therapist-rated child-therapist TA significantly predicted treatment outcome after controlling for pre-tr eatment symptom severity. Additionally, the therapist-rated parent-therapist TA was found to significantly predict treatment outcome after controlling for both age and pre-treatment sympto m severity. In both cases, stronger alliances resulted in greater reductions in OCD symptoms (Table 3-2). Additionally, there was a trend for parent-rated parent-therapist TAs, such that higher ratings of the alliance tended to predict greater reductions in OCD symp toms. For analyses conducted for mid-treatment ratings of the TA, findings indicated that all ra tings of the TA (ch ild-rated TA, therapist-rated child-therapist TA, parent-rated TA, therapist-rated parent-the rapist TA) significantly predicted treatment outcome, and again, the stronger the alliances, the greater the reductions in OCD symptoms (Table 3-3). Overall, the results provide st rong support for hypothesis 1, and findings suggest that both initial (i.e., 1st session) ratings and mid-treatment (i.e., 5th session) ratings of the TA are robust predictors of outcome. Using the criteria proposed by Cohen (1988) fo r effect sizes (ESs) for R2 in which.01 signifies a small effect size, .09 medium, and .25 large, the ESs for significant findings were in the medium to large range (Tab les 3-2 and 3-3). Researchers have raised questions as to whet her the TA is an impor tant causal agent in therapy outcome or whether the alliance is, in fact, a product of symptom improvement (Feeley, DeRubeis, & Gelfand, 1999). Alth ough several correlational studi es have found a significant positive relationship between alliance and treatme nt outcome, few have established a temporal precedence of the alliance, a criteri on necessary to make a causal claim about a process variable (Judd & Kenny, 1981). To establish temporal preceden ce of the process variable, it is necessary to assess the process variable (TA) at a given session and to also assess symptom change that
35 occurs prior to and subsequent to that se ssion (Feeley et al., 1999). The study design, which measured symptom severity as well as the TA at sessions 1, 5, and the final session allows for such an analysis of temporal precedence. Thus, a series of hierarchical regression analyses were conducted in which we evaluated whether the stre ngth of the TA (child-rated, parent-rated, and therapist-rated) at session 5 predicted subsequent change in symptom seve rity after controlling for prior change in symptom severity. The prior change score represente d the difference between the session 1 CY-BOCS score a nd the session 5 CY-BOCS score, and the subsequent change score represented the difference between the final sessions CY-BOC S score and the midtreatment CY-BOCS score. All cha nge scores were in the form of residualized change scores in order to remove the stability in each measure a nd therefore only capture the variance that is explained by factors other than the initial level of the measure. Results indicated that child-rated and therapist-rated mid-treatment ratings of the child-therapist TA, as well as parent-rated midtreatment ratings of the parent-the rapist TA, were significant predic tors of subsequent change in symptom severity after control ling for prior symptom change (T able 3-4). Additionally, there was a trend for therapist-rated mi d-treatment ratings of the parent -therapist TA to be predictive of subsequent change in symptom severity. Hypothesis 2: TA Discrepancy Scores Predicting Outcome Parent and child standardized ratings of the TA were negligibly correlated at session 1 (r = -.12, p =.60), but were moderately correlated at session 5 ( r = .52, p =.01). Standardized therapist ratings of the child-therapist TA a nd the parent-therapist TA were m oderately correlated at session 1 ( r = .54, p =.01) and markedly correlated at session 5 ( r = .64, p < .01). Using hierarchical regression an alyses, we evaluated whether the discrepancies between the child-therapist TA and the parent-therapist TA at session 1 and mi d-treatment predicted treatment outcome after controlli ng for pre-treatment symptom seve rity scores. Of note, childs
36 age was not significantly correla ted with any discrepancy scor es. No discrepancy scores significantly predicted treatme nt outcome (Table 3-5). As an extension of regression analys es, we sought to evaluate whether large discrepancies (defined as a 1.5 standard deviation difference be tween child TA and parent TA) were associated with outcome. Independent samples t -tests indicated that parent-child dyads ( n = 7) with large discrepancies (using parent a nd child ratings) in the TA following the first session had significantly higher posttreatment CY-BOCS scores (mean sc ore = 14.58) compared to parentchild dyads ( n = 15) who did not have large discrepancies (mean score = 5.33; t  = -3.46, p < .01). Similar results were found for mid-treatment ratings by parent and ch ild, such that parentchild dyads ( n = 3) with large discrepancies at midtreatment tended to have higher posttreatment CY-BOCS scores (mean score = 15.00) compared to parent-child dyads ( n = 19) who did not have large discrepa ncies (mean score = 7.21; t  = -1.84, p = .08). Analyses incorporating therapist-ratings of TAs did not yield signifi cant findings regarding large discrepancies. Hypothesis 3: Alliance Shifts Predicting Outcome Session 1 ratings of the T A were highly correlated with mid-treatment ratings of the TA for child-rated child-therapist alliance ( r = .87, p < .01) and therapistrated child-therapist alliance (r = .85, p < .01), were moderately correlated for pa rent-rated parent-t herapist alliance (r = .44, p = .04), and were markedly correlated for th erapist-rated parent-therapist alliance ( r = .71, p < .01). Additionally, mid-treatment ratings of th e TA were highly correlated with final session ratings of the TA for child-rated child-therapist alliance ( r = .93, p < .01), therapist-rated childtherapist alliance ( r = .85, p < .01), parent-rated parent-therapist alliance ( r = .92, p < .01), and therapist-rated parent-therapist alliance ( r = .89, p < .01).
37 Interestingly, change score analyses revealed an overall slight positive early alliance shift, such that, on average, between sessions 1 and 5, the strength of the child-rated TA increased by 1.82 points (SD = 4.53), the strength of the therap ist-rated child TA increased by 2.05 points (SD = 6.20), the strength of the parent-rated TA increased by 8.41 points (SD = 16.24), and the strength of the therap ist rated parent TA increased by 5.27 points (SD = 16.74). For the late alliance shift, analyses revealed mixed findings. On average, between session 5 and the final session, the strength of the ch ild-rated TA increased by 0.82 point s (SD = 3.20), the strength of the therapist-rated child TA increased by 2.91 points (SD = 10.67), the strength of the parentrated TA decreased by 11.77 points (SD = 8.90), and th e strength of the therapist rated parent TA increased by 6.32 points (SD = 12.94). Using paired sample t -tests, the only significant alliance shift was the decrease in strength of the parent-r ated TA across mid-treatment to the final session ( t  = 6.20, p < .01). Using hierarchical regression analyses, we eval uated whether the change s in the strength of TAs across treatment sessions predicted treatmen t outcome after controlling for pre-treatment symptom severity scores. The early alliance shift change sc ore represented the difference between the session 1 TA score and the session 5 TA score and the late alliance shift change score represented the difference between the final sessions TA score and the session 5 TA score. All change scores were in the form of residuali zed change scores. Early alliance shifts based on child-rated ratings of the TA significantly predic ted treatment outcome, such that the larger and more positive the change in the strength of th e alliance between sessions 1 and 5, the greater the reduction in OCD symptoms (Table 3-6). Additionally, there were tre nds for early alliance shifts regarding the therapist-rated child-therapist TA and the parent-rated parent-therapist TA, such that the larger and more positive the change in the strength of alliance over the first 5 sessions,
38 the greater the reduction in OCD symptoms. No late alliance shifts significantly predicted treatment outcome. Hypothesis 4: TA Predicting Homework Compliance The m ean for homework compliance ratings ac ross treatment for the studys sample was 3.91 (SD = 1.14), with a minimum of 1.00 and a maximum of 5.80. Using regression analyses, we evaluated whether the strength of the TA measured after session 1 predicted homework compliance across treatment. Average homework co mpliance for each participant was the sum of all homework compliance scores for all sessions divided by the sum of the number of sessions attended. The session 1 therapistrated child TA significantly pr edicted homework compliance ( r = .48, p < .05), such that the stronger the alliance, the greater the homework compliance ratings. Additionally, the session 1 therapist-rated parent TA si gnificantly predicted homework compliance ( r = .44, p < .05), and again, the stronger the alliance, the greater the homework compliance ratings. Child and parent ratings of the TA at session 1 were not significantly predictive of average homework compliance acro ss treatment. We repeated the analyses to examine whether the strength of the TA rated af ter session 5 predicted homework compliance in all sessions occurring after session 5. Results in dicated that no mid-treatment ratings of the TA significantly predicted homework compliance. Overal l, the results suggest that initial therapist ratings of both child-therapist a nd parent-therapist TAs are predic tive of the childs homework compliance across treatment. Hypothesis 5: TA and Homework Compliance Predicting Outcome A series of hierarchical linear regression an alyses were computed separately for childrated, parent-rated, and therapist-rated alliance, in which TA ratings (averaged across therapy) and compliance ratings (averaged across therapy) were entered simulta neously as predictor variables after controlling for pr e-treatment symptom severity scores. All overall models were
39 significant (Table 3-7). Results fr om analyses incorporating the child-rated TA indicated that both the average child-rated TA and averag e compliance emerged as significant unique predictors of treatment outcome. Results from an alyses incorporating th e therapist-rated childtherapist TA revealed that no individual variab les emerged as significant unique predictors of outcome, although there was a trend for the TA to be uniquely predictive. Regarding analyses incorporating the parent-rated TA, both the average parent-rated TA and average compliance were unique predictors of treatment outcome. Fi nally, results from analyses incorporating the therapist-rated parent-therapist TA indicated that no individual variables emerged as significant unique predictors of outcome, although again, th ere was a trend for the TA to be uniquely predictive of outcome (Table 3-7). Hypothesis 6: Associations among TA and Pre-Treatment Characteristics Zero-order correlations between session 1 child-rated, parent-rated, and therapist-rated TAs and various pre-treatment characteristics (i .e., child age, child social problems, child oppositionality, pre-treatment symptom severity) ar e shown in Table 3-8. Results indicated that child age was significantly correlated with therap ist-rated parent-therapist TA, such that older child age was associated with hi gher ratings of the TA. No othe r of these characteristics was significantly associated with alliance scores. Fu rthermore, chi-square analyses indicated no significant relation between session 1 TA ratings and gender or ethnicity. We repeated these analyses to examine whether pr e-treatment characteristics were related to session 5 TAs, and none of the pre-treatment characteristics was sign ificantly associated with alliance scores. Post Hoc Analyses In addition to research investigating the influence of the TA on treatment outcome, studies have also focused on the influence of the TA on attrition (Garcia & Weisz, 2002; Hawley & Weisz, 2005; Kazdin et al., 1997). Therefore, post hoc analyses were conducted to examine
40 the potential impact of the TA on treatment dropout. In our study, 3 participants dropped out of treatment and 22 completed treatment. A series of logistic regression an alyses were conducted separately for the child-rated, parent-rated, and th erapist-rated alliances, with therapist alliance ratings as the predictor variable and attrition status (completer or drop out) as the criterion variable. Additionally, we used t-tests to examine whether TA discrepancies or alliance shifts predicted attrition status. Findings indicated none of these variables si gnificantly predicted attrition status.
41 Table 3-1. Means (standard deviati ons) and ranges of study variables Variable M ( SD ) Range CY-BOCS Pre-treatment Session 5 Post-treatment 25.73 (4.59) 17.59 (5.78) 8.27 (7.19) 16.00 to 35.00 10.00 to 27.00 0.00 to 22.00 Child-rated child-therapist alliance Session 1 Session 5 Post-treatment 37.86 (8.59) 39.68 (8.85) 40.50 (8.49) 12.00 to 48.00 12.00 to 48.00 12.00 to 48.00 Therapist-rated child-therapist alliance Session 1 Session 5 Post-treatment 34.86 (11.57) 36.91 (10.43) 39.82 (9.13) 14.00 to 47.00 12.00 to 48.00 18.00 to 48.00 Parent-rated parent-therapist alliance Session 1 Session 5 Post-treatment 227.41 (15.98) 235.82 (14.70) 224.05 (20.29) 194.00 to 250.00 194.00 to 252.00 150.00 to 239.00 Therapist-rated parent-therapist alliance Session 1 Session 5 Post-treatment 200.45 (17.35) 205.73 (23.76) 212.05 (28.57) 169.00 to 230.00 150.00 to 239.00 140.00 to 249.00 Oppositionality 3.45 (2.50) 0.00 to 8.00 Social Problems 5.82 (4.00) 0.00 to 13.00 CY-BOCS Child Yale-Brown Obsessive Compulsive Scale
42 Table 3-2. Session 1 TA predicting outcome: Hier archical regression analyses after controlling for pre-treatment symptom severity Step Variable(s) R2 R2 F Child-rated child-therapist alliance 1 .03 .03 .67 Pre-treatment Symptom Severity .18 2 .14 .11 2.35 Pre-treatment Symptom Severity .29 Session 1 Alliance -.34 Therapist-rated child-therapist alliance 1 .03 .03 .67 Pre-treatment Symptom Severity .18 2 .30 .27 7.34* Pre-treatment Symptom Severity .32 Session 1 Alliance -.54* Parent-rated parent-therapist alliance 1 .03 .03 .67 Pre-treatment Symptom Severity .18 2 .19 .16 3.82+ Pre-treatment Symptom Severity .07 Session 1 Alliance -.42+ Therapist-rated parent-therapist alliance 1 .08 .08 .80 Pre-treatment Symptom Severity .16 Child Age -.21 2 .27 .19 4.75* Pre-treatment Symptom Severity .18 Child Age .13 Session 1 Alliance -.56* + p < .10, *p < .05, **p < .01
43 Table 3-3. Session 5 TA predicting outcome: Hier archical regression analyses after controlling for pre-treatment symptom severity Step Variable(s) R2 R2 F Child-rated child-therapist alliance 1 .03 .03 .67 Pre-treatment Symptom Severity .18 2 .42 .39 12.87* Pre-treatment Symptom Severity .32 Session 5 Alliance -.64** Therapist-rated child-therapist alliance 1 .03 .03 .67 Pre-treatment Symptom Severity .18 2 .43 .40 13.12** Pre-treatment Symptom Severity .23 Session 5 Alliance -.64** Parent-rated parent-therapist alliance 1 .03 .03 .67 Pre-treatment Symptom Severity .18 2 .30 .27 7.43* Pre-treatment Symptom Severity .07 Session 5 Alliance -.53* Therapist-rated parent-therapist alliance 1 .03 .03 .67 Pre-treatment Symptom Severity .18 2 .26 .23 5.77* Pre-treatment Symptom Severity .18 Session 5 Alliance -.48* + p < .10, *p < .05, **p < .01
44 Table 3-4. Session 5 TA predicting subsequent change in symptom severity: Hierarchical regression analyses after controlling fo r prior change in symptom severity Step Variable(s) R2 R2 F Child-rated child-therapist alliance 1 .00 .00 .03 Prior Change in Symptom Severity .04 2 .35 .35 10.41** Prior Change in Symptom Severity -.12 Session 5 Alliance -.62** Therapist-rated child-therapist alliance 1 .00 .00 .03 Prior Change in Symptom Severity .04 2 .32 .32 8.95** Prior Change in Symptom Severity -.31 Session 5 Alliance -.66** Parent-rated parent-therapist alliance 1 .00 .00 .03 Prior Change in Symptom Severity .04 2 .24 .24 5.90* Prior Change in Symptom Severity -.04 Session 5 Alliance -.49* Therapist-rated parent-therapist alliance 1 .00 .00 .03 Prior Change in Symptom Severity .04 2 .17 .17 3.85+ Prior Change in Symptom Severity -.11 Session 5 Alliance -.44+ + p < .10, *p < .05, **p < .01
45 Table 3-5. TA discrepancy scor es predicting outcome: Hierarchical regression analyses after controlling for pre-treatme nt symptom severity Step Variable(s) R2 R2 F Session 1 Childand parent-rated alliances 1 .03 .03 .67 Pre-treatment Symptom Severity .18 2 .04 .01 .07 Pre-treatment Symptom Severity .16 Discrepancy Score -.06 Session 5 Childand parent-rated alliances 1 .03 .03 .67 Pre-treatment Symptom Severity .18 2 .05 .02 .25 Pre-treatment Symptom Severity .23 Discrepancy Score .13 Session 1 Therapist-rated alliances 1 .03 .03 .67 Pre-treatment Symptom Severity .18 2 .03 .00 .01 Pre-treatment Symptom Severity .19 Discrepancy Score .03 Session 5 Therapist-rated alliances 1 .03 .03 .67 Pre-treatment Symptom Severity .18 2 .06 .03 .62 Pre-treatment Symptom Severity .21 Discrepancy .18
46 Table 3-6. TA early alliance shift change scor es predicting outcome: Hierarchical regression analyses after controlling for pr e-treatment symptom severity Step Variable(s) R2 R2 F Child-rated child-therapist alliance 1 .03 .03 .67 Pre-treatment Symptom Severity .18 2 .50 .47 18.04** Pre-treatment Symptom Severity .10 Early Alliance Shift -.69** Therapist-rated child-therapist alliance 1 .03 .03 .67 Pre-treatment Symptom Severity .18 2 .17 .14 3.21+ Pre-treatment Symptom Severity .13 Early Alliance Shift -.38+ Parent-rated parent-therapist alliance 1 .03 .03 .67 Pre-treatment Symptom Severity .18 2 .18 .15 3.32+ Pre-treatment Symptom Severity .14 Early Alliance Shift -.38+ Therapist-rated parent-therapist alliance 1 .03 .03 .67 Pre-treatment Symptom Severity .18 2 .07 .04 .82 Pre-treatment Symptom Severity .17 Early Alliance Shift -.20 + p < .10, *p < .05, **p < .01
47 Table 3-7. TA and homework compliance predicti ng outcome: Hierarchical regression analyses after controlling for pre-tr eatment symptom severity Step Variable(s) R2 R2 F Child-rated child-therapist alliance 1 .03 .03 .67 Pre-treatment Symptom Severity .18 2 .63 .60 14.41** Pre-treatment Symptom Severity .38* Alliance -.53** Homework Compliance -.46** Therapist-rated child-therapist alliance 1 .03 .03 .67 Pre-treatment Symptom Severity .18 2 .53 .50 9.66** Pre-treatment Symptom Severity .32+ Alliance -.39+ Homework Compliance -.40 Parent-rated parent-therapist alliance 1 .03 .03 .67 Pre-treatment Symptom Severity .18 2 .56 .53 10.43** Pre-treatment Symptom Severity .12 Alliance -.44* Homework Compliance -.42* Therapist-rated parent-therapist alliance 1 .03 .03 .67 Pre-treatment Symptom Severity .18 2 .46 .43 7.13** Pre-treatment Symptom Severity .25 Alliance -.50+ Homework Compliance -.19 + p < .10, *p < .05, **p < .01
48Table 3-8. Correlations for session 1 TA and pre-treatment characteristics 1 2 3 4 5 6 7 8 1. Child-rated childtherapist alliance --.68** -.12 .36 .23 -.17 -.22 .32 2. Therapist-rated childtherapist alliance --.05 .54** .26 -.06 -.16 .27 3. Parent-rated parenttherapist alliance --.32 .23 .16 -.16 -.26 4. Therapist-rated parenttherapist alliance --.61** .03 -.11 -.02 5. Age --.26 -.04 -.07 6. CBCL OPP --.55** .10 7. CBCL SOC --.16 8. CY-BOCS --Mean (SD) CY-BOCS Childrens Yale-Brown Obsessive Compulsive Scale, CBCL SOC Child Be havior Checklist Social Problems, CBCL OPP Child Behavior Checklist Oppositionality. p < .05, ** p < .01
49 CHAPTER 4 DISCUSSION The prim ary aim of our study was to examine the role of the TA in CBT for pediatric OCD. The studys main findings were that (1) stronger child-rated, parent-rated, and therapistrated TAs were predictive of bette r treatment outcome, (2) large discrepancies in child-therapist versus parent-therapist TAs were associated w ith poorer treatment outcome, (3) larger and more positive early alliance shifts (as rated by changes in child-rated TA between sessions 1 and 5) were predictive of bette r treatment outcome, (4) therapist-rate d child-therapist TA and therapistrated parent-therapist TA at se ssion 1 predicted homework compliance across treatment, (5) both TA and compliance ratings were uniquely predictive of treatment outcome, and (6) with the exception of child age being correlated with the therapist-rated parent-the rapist TA, no measured pre-treatment child characteristics were predictive of TA ratings. Hypothesis 1: TA Predicting Outcome Our study found that both child-rated and th erapist-rated child-therapist T As were significantly related to treatment outcome, such that stronge r ratings of the alliance at the beginning of treatment and at mid-treatment pred icted a greater reduction in OCD symptoms at post-treatment. Additionally, results indicated th at both the parent-rated and the therapist-rated parent-therapist TA was also a significant predictor of treatment outcome. Furthermore, analyses examining the temporal precedence of these fi ndings provided support for the causal hypothesis that the TA predicts outcome, as results indica ted that the TA was a predictor of subsequent change in OCD symptoms after control ling for prior change of symptoms. Findings from our study correspond with alliance-outcome research among youth being treated with evidence based th erapies for disruptive behavior disorders (Kazdin et al., 2005, Kazdin & Whitley, 2006), substance abuse (Hogue et al., 2006; Tetzlaff et al., 2005), and non-
50 OCD anxiety disorders (Hughes & Kendall, 2007 ), and they lend further support for the importance of investigating the TA in evidence -based treatments for children and adolescents (Green, 2006). Additionally, results from our st udy are congruent with re search regarding the influence of the parent-therapist TA on treatment outcome, with most studies reporting a significant alliance-outcome association (Kazdin & Whitley, 2006; Kazdin, Whitley, & Marciano, 2006; McLeod & Wiesz, 2005). The effect sizes in our study were larger than ones found in previous research with youth ( r = .24; Shirk & Karver, 2003) and adults ( r = .22; Martin, Garske, & Davis, 2000), suggesting that the TA may be a particularly important predictor of treatment outcome within the cognitive-behavioral treatment of pediatric OCD. Researchers have conceptualized the TA as a catalyst for treatment involvement, such that the TA generates motivation to work on problem s and promotes collaboration with treatment tasks (Shirk & Karver, 2005). Within an anxi ety-provoking treatment, in which youth experience intense emotional arousal as a result of therap eutic tasks (i.e., exposure to feared thoughts and situations), it is likely that a strong TA provides an empathic, supportive, and motivating environment that facilitates involvement in such tasks designed to reduc e OCD symptoms. It is noteworthy that findings from our study indicate that both early and mid-treatment ratings of the TA were predictive of outcome, as they suggest that both before and after the initiation of exposure exercises, the alliance be tween child and therapist is an important predictor of outcome. Alliance ratings following the first session are likely reflective of how well the therapist was able to establish rapport with the child a nd deliver the treatment rationale in a way that matched therapeutic tasks with the childs goal s. Alliance ratings following mid-treatment are likely to be influenced by the introduction of ex posure exercises into therapy. Exposure exercises facilitate reduction in symptoms through helping the patient to approach and attend to anxiety-
51 provoking stimuli so as to activat e the fear network, tolerate th e emotional arousal associated with such activation, and change the cognitions involved in the maladaptive fear network (or schema) to allow for a more adaptive understanding (Foa & Kozak, 1986). Helping the patient through this difficult proce ss requires substantial clinical skill (Foa & Franklin, 2000), with the clinician needing to adequately respond to the patients emotions in an empathic way that communicates to the patient that they are safe and supported through the process (Greenberg & Pascual-Leone, 2006). Thus, findings from the study suggest that within an exposure-based treatment for pediatric OCD, a strong TA may pr ovide a validating and encouraging environment that helps the patient to understand the importance of difficult th erapeutic tasks and supports the patient through completing these tasks. Greenbe rg and Pascual-Leone (2006) have likened the therapeutic relationship to a thermostat for the fire of emotional arousal, such that the through a positive relationship, the therapist can empathically evoke emotional arousal (i.e., heat) (to facilitate activation of the fear ne twork) and can validate and facilitate adaptive regulation of overly hot emotions (to prevent es cape and avoidance within the exposure task). The strong parent-therapist alliance-outcome relation in our study is not surprising, given that parental involvement within CBT for pediatric OCD is considered an essential aspect of successful treatment. Within family-based CBT for OCD, parental involvement serves many different functions, as pare nts learn to reduce criticism decrease accommodation and reinforcement of OCD symptoms, limit modeling of anxious responding, increase reinforcement of adaptive coping skills, and provide support during in-session and out -of-session exposures (Barrett, Dadds, & Rapee, 1996; Barrett et al., 2004; Freeman et al., 2003). Our studys findings suggest that the parent-therapi st TA may promote compliance w ith and support of the treatment within and outside of the treatment sessions.
52 Hypothesis 2: TA Discrepancy Scores Predicting Outcome A unique aspect of our study, as com pared to other alliance-outcome studies in youth psychotherapy, is that both parent and child partic ipated in each therapy session together with the same therapist. As a result, we were able to include an assessment of the parent-therapist TA within this family-based CBT treatment approach. Given our studys design, we were able to examine whether discrepancies between child-the rapist TAs and parent-therapist TAs were related to treatment outcome. Results from regression analyses revealed that discrepancies were not predictive of treatment outcome. One possible expl anation for the lack of significant results may be related to the studys operational definiti on of a discrepancy (i.e., difference between a measure of childtherapist TA and parent-therapist TA). Subtrac ting TA scores may not adequately reflect the degree to which a child and his or her parent disagree about the strength of the TA. Study findings may have been different had we been able to assess discrepancies using observational data. For example, the System for Observing Family Therapy Alliances (SOFTA; Friedlander, Escudero, & Heatherington, 2006) incorporates an assessment of individual family members sense of comfort with each othe r in therapy and their agreem ent about attending therapy and working together towards common goals. However, analyses comparing those parent -child dyads with particularly large discrepancies in TAs (discrepancies of 1.5 standard deviations or higher between child-therapist and parent-therapist TAs) with those dyads withou t such a large discrepanc y indicated that larger discrepancies were significantly associated w ith poorer treatment outcome. Interestingly, these results suggest the need to monitor and assess fo r a noticeably large imbalance in the degree to which each family member involved in treatment believes that the therapist is working hard to respect them, understand their goals, and help them overcome problems. Within any family-
53 based treatment, the therapist has a challenging ta sk of aligning simultaneously with both parents and children, a task that may be particularly de manding if there are significant differences in the perception of the problem or disagreements regarding the goals of treatment (Friedlander, Lambert, & Muniz de la Pena 2008; Robbins et al., 2003). Shirk and Saiz (1992) have highlighted that im portant factors to cons ider when developing alliances with youth in therapy include beliefs about the need to change, causal locus of problems, and the contingency of the problem so lution. If there are disagreements between child and parent regarding these issues (e.g., parent does not understand role of family accommodation in maintenance of symptoms, child experi ences ego-syntonic OCD symptoms), such disagreements could limit the development of the alliance and/or interfere with treatment progress. Thus, findings highlight the need for th e therapist to be mindful of discrepancies in TAs and reflect the importance of adequately responding to both parents and childrens needs and concerns and to attempt to incorporate each family members goals in to the overall treatment plan. If sensing that a considerable imbalance exists, findings fr om our study suggest that it may be especially critical to attend to possible family conflict or disagreements and to work to equally engage and connect with each family member in each session. Given the studys finding that TA does pr edict outcome, these discrepancy findings highlight the importance of utilizing intervention strategies aimed at transforming the childs initial resistance or negativity about treatment into a sense of colla boration with th e therapist and their parent on treatment tasks (Diamond et al., 1999). For example, in a study investigating alliance building behaviors within family therapy, Diamond et al (1999) found that improvement in adolescent-therapist alliances was related to therapist behaviors of attending to the adolescents experience, presenting themselves as the adolescents ally, and helping the
54 adolescent formulate personally meaningful goals Thus, it appears that a therapists willingness to advocate for youth in therapy and to spend time early in th erapy developing goals that are mutually agreed upon by all treatm ent participants may assist in the development of strong alliances and consequently faci litate treatment progress (DiGiu seppe, Linscott, & Jilton, 1996). Hypothesis 3: Alliance Shifts Predicting Outcome Findings fro m analyses investigating the influence of alliance shifts on outcome found that early alliance shifts based on child-rated ratings of the TA si gnificantly predicted treatment outcome, such that the larger and more positive th e change in the strength of the alliance over the first 5 sessions, the greater the reduction in OCD symptoms. Additionally, results revealed trends for early alliance shifts based on the therapistrated child-therapist TA and the parent-rated parent-therapist TA to be pred ictive of treatment outcome, and again, the larger and the more positive the change in the strength of the allia nce over the first 5 sessions tended to predict a greater reduction in OCD sy mptoms. Findings from our study are congruent with a study examining the impact of parent-therapist and ch ild-therapist alliances in family therapy for adolescent substance use, in which results indi cated that adolescents whose alliances improved from early to mid-treatment demonstrated im provement in externalizing symptoms whereas adolescents with worsening allia nces between these time points demonstrated a worsening of externalizing symptoms (Hogue et al., 2006). In other words, in bot h studies, it appears that a positive shift in early alliances is related to symptom improvement where as the reverse (a negative shift) is related to symptom worsening. These results are encouraging, as they suggest that despite in itial difficulty in establishing a TA, it is possible for the child an d/or parent to percei ve that he or she and the therapist can become more collaborative across time. Findings from this set of analyses are especially interesting as they indicate an average positive shift in alliances across sessions 1 to 5, which
55 reflects an improving alliance in therapy desp ite the initiation of exposure and response prevention exercises that have the potential to result in considerable distress. These findings have important implications for early tasks in therapy. It may be more important to devote extra time with patients and/or parents who have difficulty connecting with the therap ist in order to promote a collaborative relationship. Analyses of TA change scores, which indicated fluctuations in ratings of the TA over time, provide data suggesting that the TA is a modifi able construct, and ther efore, support the notion that the TA in and of itself can be a target for intervention. For example, therapists can be trained to increase certain behaviors associated with strong alliances (e.g., therapist responsiveness; Russell, Shirk, & Jungbluth, 2008) as well as decr ease other behaviors associated with weak alliances (e.g., pushing the child to talk; Cree d & Kendall, 2005; Karver et al., 2008). Thus, findings suggest that it remains crucial for therap ists to continue to fo cus on the development of the TA while simultaneously attending to specif ic therapeutic tasks, such as exposure and response prevention. Hypothesis 4: TA Predicting Homework Compliance Analyses ex amining the potential predictive value of the TA on homework compliance provided partial support for hypothesis 4, with resu lts indicating that therapist ratings, but not child or parent ratings, of the session 1 TA significantly predicted homework compliance across treatment. Researchers have conceptualized the TA as a catalyst for treatment involvement, such that the alliance is thought to generate motivation to work on problems and promote collaboration with treatment tasks (Shirk & Ka rver, 2005). Given our studys findings, it appears that the therapists perception of the initial alliance with the patient and the patients parent may be an indicator of the degree to which the patient successfully completes out-of-session homework assignments across treatment. It may be that therapists who establish strong initial
56 alliances with their patients are able to use the collaborative relationship to garner motivation for involvement in therapy and/or highlight the impo rtance of participati on in treatment tasks. Alternatively, however, it may be that therapists ra tings of the initial TA are influenced by their perception of their patients motivation for involve ment in therapy, which may be manifested in subsequent sessions by good compliance with homewo rk tasks. It is also important to note the potential influence of shared me thod variance on these findings, given that only therapist-rated TAs were associated with homework compliance, which is also ultimately rated by therapists (with input from the child and parent). Interestingly, child-rated and parent-rated TAs were not related to homework compliance. Perhaps these non-significant findi ngs are a result of how our study assessed compliance, as it focused exclusively on homework compliance. Abra mowitz et al (2002) examined the role of compliance within several different aspects of CBT for adult OCD, including psychoeducation, in-session exposure, homework exposure, and ritual prevention. Perhaps examining patient compliance in this more comprehensive manner ma y have resulted in significant findings for parent and child ratings as well as for mid-treat ment ratings of the TA. Consistent with this proposition, other studies examini ng child involvement in therapy, a variable that includes an array of behaviors associated with participating in therapy (e.g., discussi ng thoughts and feelings, responding to therapist requests, completing homework), have found that this more comprehensive variable has been related to strong TAs (Chu & Kendall, 2004; Karver et al., 2006; Karver et al., 2008). Other potential reasons for the studys nonsignificant findings may be related to methodological issues associated with how homework compliance was assessed. One issue relates to the studys reliance on a 1-item therap ist rating of homework compliance based on a
57 review of child and parent repor t of compliance at the beginning of each treatment session. This type of assessment may increase risk of inaccu racy as well as reporting bias. Additionally, we did not control for the amount of homework given to each patient nor were we able to account for quality of homework assigned (e.g., difficulty level; Woods, Chambless, & Steketee, 2002). Future investigations of the influence of the TA on homework compliance should include a more comprehensive assessment of patient compliance, and should consider ways of assessing for both quantity and quality of therapeu tic tasks (Woods et al., 2002). Hypothesis 5: TA and Homework Compliance Predicting Outcome Findings from analyses examining the relativ e predictive value of the TA and homework compliance on outcome suggest that both the TA and homework compliance are robust and unique predictors of OCD symp tom reduction in CBT for pediat ric OCD. These findings are partially congruent with a previous study examining the relative influence of the therapeutic relationship and homework compliance on non-OCD anxiety disorder symptoms, in which the therapeutic relationship, but not homework complia nce, was a significant predictor of whether participants had a post-treatment anxiety diso rder diagnosis (Hughes & Kendall, 2007). Results from our study highlight the importance of both non-specific, relationshipbased factors as well as specific cognitive-behavioral techniques in the treatment of pediatric OCD. The finding that homework compliance was a unique predictor of OCD symptom reduction after accounting for pre-treatment symptom severity and the TA is congruent with past research examining the compliance-outcome asso ciation in adult OCD patients (Abramowitz et al., 2002; De Araujo et al., 1996; Tolin et al., 20 04; Whittal et al., 2005). This finding suggests that exposure and other skills-based homework exercises practiced outside of the treatment session (i.e., in the child s natural environment) aid in th e reduction of OCD symptoms, likely through increased opportunities for habituation to f eared stimuli as well as through providing
58 additional opportunities to alter distorted cognitions and to enhan ce the childs sense of mastery (Hudson & Kendall, 2002; Huppert, Ledley, & Foa, 2006). Given these findings, efforts should be made to use strategies to promote homework compliance and to problem-solve issues of noncompliance in a quick and effective manner. Hudson and Kendall (2002) discuss various techniques to improve homework compliance with in CBT for children with anxiety disorders, such as using positive reinforcement strategies (e.g., praise, rewards) for compliance and involving parents in the process to help remind the child of their homework tasks. Hypothesis 6: Associations among TA and Pre-Treatment Characteristics Researchers have highlighted the importance of investigating the precu rsors of the TA, as identifying such characteristics increases our unde rstanding of the nature of the alliance and may help to reveal possible factors that impede th e development of a strong alliance (Kazdin & Whitley, 2006). Interestingly, our study found that no measured pre-treatment child characteristic (i.e., age, gender, ethnicity, OCD symptom severity, oppositionality, and social problems) was associated with the child-rated child-therapist TA the therapist-rated child-therapist TA, or the parent-rated parent-thera pist TA. This finding may speak to the therapists ab ility to respond appropriately to the childs pres entation and to modify or adap t therapeutic behaviors based on the presence of behavior or social problems (e.g., employ beha vioral modification principles when handling oppositional behavior). Researchers have highlighted the im portance of therapist responsiveness to patient factors that may interfere with the de velopment of the TA (Newman & Stiles, 2006). For example, they suggest spending more time on the explanation of the treatment rationale for a patient with low treatment expecta tions and incorporating so cial interventions to build positive support for patients with negativ e or inadequate social support networks. Unfortunately, because we did not use observationa l methods, we are unable to examine whether therapist responsiveness wa s a factor in assuaging potential barriers to the development of the
59 TA. These null findings are also encouraging in that they suggest largel y that the degree of a childs oppositionality, social problems, or OCD symptom severity does not necessarily have to negatively impact the development of a strong th erapeutic alliance with in CBT for pediatric OCD. Our study found that only one measured pre-treatment child characteristic, age, was related to the TA, such that the older the child, the stro nger the initial therapist-rated parent-therapist TA. Perhaps therapists sense a strong connectio n from parents of adolescents struggling with OCD because these parents may express or exhibit a particular appreciation fo r and/or interest in having a supportive and understand ing person help their child a nd family cope with OCD. Future research should continue to inves tigate other potential pr edictors of the TA, including child (e.g., comorbidit y, level of peer support), pa rent (e.g., degree of family accommodation, parental psychopathology), and therap ist (e.g., level of experience, degree of empathy) characteristics. Furthermore, given this studys findings on al liance shifts and their importance in predicting outcome, it would also be interesting to investigate possible predictors of such shifts. Limitations Despite the many strengths of the study, incl uding its multi-informant and multiple timepoint design within an evidenced based treatm ent protocol, there are several limitations worth acknowledging. First, the design of the present study does not allow for a causal examination of the alliance-outcome relationship, given that there was no experime ntal manipulation of the TA. Thus, although findings suggest a significant re lationship between the TA and outcome, other factors, such as maturation or the passage of time, confound the interp retation of results. On balance, we were able to establish temporal precedence of the TA as it relates to symptom change, which provides some support for the causal hypothesis. Furthermore, therapists were not
60 systematically randomized to patients, and therefor e we cannot rule out the influence of selection bias. Therapists were assigned cases based on availability and schedul ing, and there was no attempt to match a particular patient with a th erapist. Chi-square analyses were conducted to examine the possible influence of therapist on homework compliance and outcome, and findings revealed no significant differences on these variables as a function of therapist. The number of variables in our analyses was limited by the studys small sample size. As a result, potential moderators of the alliance-outco me relationship were not examined in our study. Within the TA literature for youth, few clinical m oderators of the alliance-outcome relationships have been identified, with the most robust finding indicating that the type of presenting problem moderates the relationship (Shirk & Karver, 2003) Such research is needed to understand for whom and under what conditions the TA is im portant with regards to outcome (Kendall & Choudhury, 2003). Future studies of larger magnitude should focus on examining moderators within pediatric OCD treatment to identify disord er-specific variables th at impact the allianceoutcome relationship. Although our study did investigate the influence of the TA on both homework compliance and OCD symptom reduction, it did not examine the alliances effect on other potential outcomes, including depressive symptoms, degree of family accommodation, and levels of functional impairment. It may be that the TA impacts these variable s as well, and if so, positive findings would likely strengthen the fields focus on the TA and its importance within CBT. Additionally, the study did not incorporate an objective m easure of the TA. Given the variability in ratings of alliance between child a nd therapist as well as parent and therapist, it appears as though perceptions of the alliance vary depending on the rating source. Despite this, it appears as though all perceptions ma tter in terms of outcome, and ther efore, it is worthwhile to
61 gain subjective ratings. However, this study ma y have been enhanced by comparing subjective versus objective ratings and e xploring whether discrepancies between these ratings signify meaningful data. Another limitation pertains to th e studys inability to examine all sources of variability in the TA. DeRubeis, Brotman, & Gib bons (2005) have identified four sources of variability in the TA, including the patient, therapist, patient -therapist interaction, and outcome. Although our study did evaluate pre-treatment characteristics of the patient that may contribute to the development of the alliance (e.g., age, gender, level of oppositionality), it did not examine therapist characteristics (e.g., leve l of experience, ability to be empathic) that may contribute to the variability in the alliance. Furthermore, our study did not utilize observational methods to assess the moment-to-moment dyadic interaction be tween patient and therap ist in session. Such investigation into these others sources of variab ility will further our unders tanding of factors that lead to the development of strong TAs and may pr ovide us with information pertaining to what specific types of patients work well with particular types of therap ists. Future research utilizing multilevel modeling methods will allow researchers to further understand the degree to which these different sources of variability influe nce outcome (Baldwin, Wampold, & Imel, 2007). On balance, the study did control for symptom severity when examini ng the influence of the alliance on outcome, thereby accounting for the s ource of outcome in our analyses. Study Implications Given the studys finding that the T A is indeed an important factor in outcome within CBT for pediatric OCD, research should begin to i nvestigate specific therapist attributes and techniques that contribute to the development of strong TAs. Within the child anxiety literature, Creed and Kendall (2005) have identified specific therapist behaviors that predicted strong TAs. However, they only examined these behaviors during CBT sessions focusing on general CBT
62 strategies (e.g., learning emotion management, cognitive restructuring skills) rather than during sessions involving exposure and response prev ention exercises, which are arguably more anxiety-provoking and difficult for children but wh ich comprise the majority of the treatment protocol for pediatric OCD. The field will benefit from research examining specific therapist attributes (e.g., empathic, confident, flexible) and actions (e.g., provides encouragement, engages in exposure with child, reviews past exposure su ccesses) that promote strong TAs. Furthermore, given the numerous roles of the therapist within CBT for OCD (including educator, consultant, support, modeler, motivator, and accountabi lity agent; Tolin & Hannan, 2005), it will be important to identify specific therapeutic behavi ors within each role that create an optimal TA. Results indicated that therapist-rated al liances significantly pr edicted both symptom reduction and homework compliance. Implications for this finding suggest the potential benefit of therapists regular evaluation of the alliance, especially in the instan ces in which weak TAs may be hindering therapeutic progr ess (Baldwin et al., 2007). Within such an evaluation, it will be important to address the different aspects of the TA, including problems with the therapeutic bond as well as disagreements on tasks or goals Furthermore, findings that child-rated and parent-rated TAs significantly predicted outcome highlight the importance of obtaining patient and parent feedback regarding the strength of the TA at multiple time points during therapy. Research with poorly responding adult patients has highlighted the advantage of obtaining patient feedback about TAs w ith regard to improved outcomes (Harmon, Hawkins, Lambert, Slade, & Whipple, 2005), and Nafisi and Stanley (2007) have proposed several different strategies for gathering such feedback, includ ing subjective rating forms as well as regular inquiry at the end of all sessions.
63 Another implication of the studys findings re lates to reports of therapist reluctance to conduct E/RP with patients given that such trea tment entails anxiety-provoking exercises that are distressing to the patient. Ind eed, researchers have hypothesized that one reason for the underutilization of E/RP (despite its proven effectiveness) is that clinicians are concerned that such treatment approaches will damage the therapeu tic relationship (Rosqvist 2005). Findings from our study indicated that, overall, patients and their parents genera lly rated their TA with their therapists as positive, with an average child-rated TA of 38.56 (on a scale from 12 to 48) and an average parent-rated TA of 228.59 (on a scale fr om 36 to 252) These results suggest that patients and their parents generally perceive thei r therapist positively. A strength of our study is that the TA was assessed at multiple time point s during therapy (both before and after E/RP exercises were implemented). Findings revealed that, at both time points, as well as at the end of therapy, ratings of the TA were relatively hi gh. Such findings will hopefully assuage clinicians concerns regarding E/RP as a therapeutic techni que and will encourage apprehensive clinicians to learn and utilize the effectiv e approach to better the outcom es among children and adolescents with OCD. Conclusions Despite subs tantial evidence demonstrating th at manualized treatments, which describe specific procedures and techniques for specific problems, work, and work very well, there are concerns that such treatment protocols neglect to focus on the practitioners que stion, How do I get my patients to do these procedures? (Shirk & Karver, 2005). In fact, although most evidence-based treatment manuals mention the importance of the TA, there is often a dearth of information regarding how to specifically develop the alliance or what specific therapeutic actions will enhance the patients motivation a nd engagement in therapeutic tasks. Our study evaluated the TA within a manualized treatmen t, and findings revealed the importance of
64 examining the construct with regards to its relati onship to compliance with therapeutic tasks as well as treatment outcome. Results suggest that a certain degree of variability in outcome for pediatric OCD patients is related to the nature of the TA between therapist and patient as well as between therapist and parent. Given these findings, it is essential to continue to investigate this variable as the field moves towards enhanci ng outcomes for OCD patients. We should move from simply acknowledging the importance of th e TA within our manuals to explicitly citing those therapist attributes and actions that yield strong alli ances and better outcomes.
65 LIST OF REFERENCES Abramowitz, J. S., & Foa, E. B. (2000). Does comorbid major depressive disorder influence outcome of exposure and response prevention for OCD? Behavior Therapy, 31, 795-800. Abramowitz, J. S., Franklin, M. E., Schwartz, S.A., Furr, J. M. (2003). Symptom presentation and outcome of cognitive-behavioral therapy for obsessive-compulsive disorder. Journal of Consulting and Clinical Psychology, 71, 1049-1057. Abramowitz, J. S., Franklin, M. E., Street, G., Kozak, M., & Foa, E. B. (2000). Effects of comorbid depression on response to treatme nt for obsessive-compulsive disorder. Behavior Therapy, 31, 517-528. Abramowitz, J. S., Franklin, M. E., Zoellner, L. A., & DiBernardo, C. L. (2002). Treatment compliance and outcome in obsessive-compulsive disorder. Behavior Modification, 26, 447-463. Abramowitz, J. S., Whiteside, S. P., & Deacon, B. J. (2005). The effectiveness of treatment for pediatric obsessive-compulsive disorder: A meta-analysis. Behavior Therapy, 36, 55-63. Achenbach, T. M., & Rescorla, L. A. (2001). Manual for the ASEBA School-Age Forms & Profiles. Burlington, VT: University of Vermont, Re search Center for Children, Youth, & Families. Alonso, P., Menchon, J. M., Pifarre, J., Mataix-Col s, D., Torres, L., Salgado, P., & Vallejo, J. (2001). Long-term follow-up and predictors of clinical outcome in obsessive-compulsive patients treated with serotonin reuptake inhibitors and beha vioral therapy. Journal of Clinical Psychiatry, 62, 535-540. American Academy of Child and Adolescent Psyc hiatry. (1998). Practice parameters for the assessment and treatment of children and adolescents with obsessive-compulsive disorder. Journal of the American Academy of Child and Adolescent Psychiatry, 37, 27S45S. American Psychiatric Association (2000). Diagnostic and statistica l manual of mental disorders fourth edition text revi sion. Washington, DC: Author. Baldwin, S. A., Wampold, B. E., & Imel, Z. E. (2007). Untangling the alliance-outcome correlation: Exploring the rela tive importance of therapist and patient variability in the alliance. Journal of Consulting and Clinical Psychology, 75, 842-852. Barrett, P. M., Dadds, M. R., & Rapee, R. M. (1996). Family treatment of childhood anxiety: A controlled trial. Journal of Consulting and Clinical Psychology, 64, 333-342. Barrett, P., Farrell, L., Dadds, M., & Boulter, N. (2005). Cognitive-behavioral family treatment of childhood obsessive-compulsive disorder: Long-term follow-up and predictors of outcome. Journal of the American Academy of Child and Adolescent Psychiatry, 44, 1005-1014.
66 Barrett, P., Healy-Farrell, L., & March, J. S. (2 004). Cognitive-behavioral family treatment of childhood obsessive-compulsive diso rder: A controlled trial. Journal of the American Academy of Child and Adolescent Psychiatry, 43, 46-62. Barrett, P. Shortt, A., & Healy, L. (2002). Do pa rent and child behaviours differentiate families whose children have obsessive-compulsive di sorder from other clinic and non-clinic families? Journal of Child Psychology and Psychiatry, 43 597-607. Besiroglu, L., Uguz, F., Ozbebit, O., Guler, O ., Cilli, A. S., & Askin, R. (2006). Longitudinal assessment of symptom and subtype categories in obse ssive-compulsive disorder. Depression and Anxiety, 0, 1-6. Calvocoressi, L., Lewis, B., Harri s, M., Trufan, S. J., Goodman, W. K., McDougle, C. J. et al. (1995). Family accommodation in obsessive-compulsive disorder. American Journal of Psychiatry, 152, 441-443. Calvocoressi, L., Mazure, C. M., Kasl, S. V., S kolnick, J., Fisk, D., Vegso, S. J., van Oppen, B. L., & Price, L. H. (1999). Family accomm odation of obsessive-compulsive symptoms: Instrument development and assessment of family behavior. Journal of Nervous and Mental Disease, 187, 636-642. Carter, A. S., & Pollock, R. A. (2000). Obsession and compulsions: The developmental and familial context. In A. J. Sameroff & M. Lewis (Eds.), Handbook of developmental psychopathology (2nd ed., pp. 549-566). New York: Kluwer Academic/Plenum. Chambless, D. L., & Steketee, G. (1999). Expressed emotion and behavior therapy outcome: A prospective study with obsessive-com pulsive and agoraphobic patients. Journal of Consulting and Clinical Psychology, 67, 658-665. Chu, B. C., & Kendall, P. C. (2004). Positive as sociation of child involvement and treatment outcome within a manual-based cognitive-behavi oral treatment for children with anxiety. Journal of Consulting and Clinical Psychology, 7 2, 821-829. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed). Hillsdale, NJ: Erlbaum. Creed, T. A., & Kendall, P. C. (2005). Therapist alliance-building behavior within a cognitivebehavioral treatment for anxiety in youth. Journal of Consulting and Clinical Psychology, 73, 498-505. De Araujo, L. A., Ito, L. M., & Marks, I. M. (1996). Early complia nce and other factors predicting outcome of exposure fo r obsessive-compulsive disorder. British Journal of Psychiatry, 169, 747-752. de Haan, E. (2006). Effective treatment of OCD? Journal of the American Academy of Child and Adolescent Psychiatry, 45, 383.
67 de Haan, E., Hoogduin, K.A., Buitelaar, J., & Keijs ers, S. (1998). Behavior therapy versus clomipramine for the treatment of obsessive-compulsive disorder. Journal of the American Academy of Child & Adolescent Psychiatry, 37 1022. DeRubeis, R. J., Brotman, M. A., & Gibbons, C. J. (2005). A conceptual and methodological analysis of the nonspecifics argument. Clinical Psychology: Science and Practice, 12, 174-183. De Vet, K. A., Kim, Y. J., Charlot-Swilley, D., & Ireys, H. T. (2003). The therapeutic relationship in child therapy: Pers pectives of children and mothers. Journal of Clinical Child and Adolescent Psychology, 32, 277-283. Diamond, G. M., Diamond, G. S., & Liddle, H. A. (2000). The therapist-parent alliance in family-based therapy for adolescents. Journal of Clinical Psychology, 56, 1037-1050. DiGiuseppe, R., Linscott, J., & Jilton, R. (1996) Developing the therapeu tic alliance in childadolescent psychotherapy. Applied and Preventive Psychology, 5, 85-100. Douglass, H. M., Moffit, T. E., Dar, R., McGee, R., & Silva, P. (1995) Obsessive compulsive disorder in a birth cohort of 18 y ear olds: Prevalence and predictors. Journal of the American Academy of Child and Adolescent Psychiatry, 34, 1424-1431. Eltz, M. J., Shirk, S. R., & Sarlin, N. (1995) Alliance formation and treatment outcome among maltreated adolescents. Child Abuse and Neglect, 419-431. Emmelkamp, P. M., G., Hoekstra, R. J., & Vi sser, A. (1985). The behavioral treatment of obsessive-compulsive disorder: Prediction of outcome at 3, 5 years follow-up. In P. Pichot, P. Berner, R. Wolf, & K. Thau (Eds.) Psychiatry: The state of the art (pp. 265270). New York: Plenum. Feeley, M., DeRubeis, R. J., & Gelfand, L .A. (1999). The temporal relation of adherence and alliance to symptom change in cognitive therapy for depression. Journal of Consulting and Clinical Psychology, 67, 578-582. Flament, M.F., Koby, E., Rapoport, J.L., & Ber g, C.J. (1990). Childhood ob sessive compulsive disorder: A prospective follow-up study. Journal of Child Psychology and Psychiatry, 31, 363-380. Flament, M.F., Whitacker, A., Rapoport, J.L., Da vies, M., Berg, C., Kalikow, K., et al. (1988). Obsessive compulsive disorder in ad olescence: An epidemiological study. Journal of American Academy Children and Adolescent Psychiatry, 27, 764-771. Florsheim, P., Shotorbani, S., & Guest-Warnick, G. (2000). Role of the working alliance in the treatment of delinquent boys in community-based programs. Journal of Clinical Child Psychology, 29, 94-107. Foa, E. B. (1979). Failure in treat ing obsessive-compulsive disorder. Behaviour Research and Therapy, 17, 169-176.
68 Foa, E. B., & Franklin, M. E. (2000). Psychotherapies for obsessive-c ompulsive disorder: A review. In M. Maj, N. Sartorius, A. Okas ha, & J. Zohar (Eds.), Obsessive-compulsive disorder (pp. 93 115). New York: Wiley. Foa, E. B., & Kozak, M. J. (1986). Emotiona l processing of fear: Exposure to corrective information. Psychological Bulletin, 99, 20-35. Franklin, M. E., Abramowitz, J. S., Kozak, M. J., Levitt, J. T., & Foa, E. B. (2000). Effectiveness of exposure and ritual prevention for obs essive-compulsive disorder: Randomized compared with nonrandomized samples. Journal of Consulting and Clinical Psychology, 68, 594-602. Franklin, M. E., Kozak, M. J., Cashman, L. A., Co les, M. E., Rheingold, A. A., & Foa, E. B. (1998). Cognitive-behavioral treatment of pediatric obsessive-compulsive disorder: An open trial. Journal of the American Academy of Child and Adolescent Psychiatry, 37, 412-419. Freeman, J. B., Garcia, A. M., Fucci, C., Kar itani, M., Miller, L., & Leonard, H. L. (2003), Family-based treatment of early-onset obsessive-compulsive disorder. Journal of Child and Adolescent Psychophamacology, 13, S1: S71-S80. Friedlander, M. L., Escuder o, V., & Heatherington, L. (2006). TAs with couples and families. Washington, DC: American Psychological Association. Friedlander, M. L., Lambert, J. E., & Muniz de la Pena, C. (2008). A step toward disentangling the alliance/improvement cy cle in family therapy. Journal of Counseling Psychology, 55, 118-124. Fricke, S., Moritz, S., Andresen, B., Jacobson, D., Kloss, M., Rufer, M., & Hand, I. (2006). Do personality disorders predict negative tr eatment outcome in obsessive-compulsive disorders? A prospective 6-month follow-up study. European Psychiatry, 21, 319-324. Garcia, J. A., & Weisz, J. R. (2002). When youth mental health car stops: Therapeutic relationship problems and other reasons for ending youth outpatient treatment. Journal of Consulting and Clinical Psychology, 70, 439-443. Geffken, G. R., Storch, E. A., Duke, D., Le win, A., Monaco, L., & Goodman, W. K. (2006). Hope and coping in family members of pa tients with obsessive-compulsive disorder. Journal of Anxiety Disorders 20, 614-629. Geller, D., Biederman, J., Jones, J., Park, K., Schwartz, S., Shapiro, S., & Coffey, B. (1998). Is juvenile obsessive-compulsive disorder a developmental subtype of the disorder? A review of the pediatric literature. Journal of the American Academy of Children and Adolescent Psychiatry, 37, 420-427. Green, J. (2006). Annotation: The therapeutic alli ance a significant but neglected variable in child mental health treatment studies. Journal of Child Psychol ogy and Psychiatry, 47, 425-435.
69 Greenberg, L. S., Pascual-Leone, A. (2006). Em otion in psychotherapy: A practice-friendly research review. Journal of Clinical Psychology: In Session, 62, 611-630. Harmon, C., Hawkins, E. J., Lambert, M. j., Slade, K., & Whipple, J. L. (2005). Improving outcomes for poorly responding clients: The use of clinical support tools and feedback to clients. Journal of Clinical Psychology, 61, 175-185. Hawley, K. M., & Weisz, J. R. (2005). Youth vers us parent working alliance in usual clinical care: Distinctive associations with rete ntion, satisfaction, and treatment outcome. Journal of Clinical Child and A dolescent Psychology, 34, 117-128. Hogue, A., Dauber, S., Stambaugh, L. F., Cecero, J. J., & Liddle, H. A. (2006). Early therapeutic alliance and treatment outcome in individual a nd family therapy for adolescent behavior problems. Journal of Consulting and Clinical Psychology, 74, 121-129. Hoogduin, C. A. L., de Haan, E., & Schaap, C. ( 1989). The significance of the patient-therapist relationship in the treatment of obsessive-compulsive neurosis. British Journal of Clinical Psychology, 28, 185-186. Horvath, A. O., & Bedi, R. P. (2002). The alliance. In J. C. Norcross (Ed.), Psychotherapy relationships that work: Therapist cont ributions and responsiveness to patients (pp. 2769). New York: Oxford University Press. Hudson, J. L., & Kendall, P. C. (2002). Showing you can do it: Homework in therapy for children and adolescents w ith anxiety disorders. Journal of Clinical Psychology, 58, 525534. Hughes, A. A., & Kendall, P. C. (2007). Predic tion of cognitive behavior treatment outcome for children with anxiety disorders: Therapeu tic relationship and homework compliance. Behavioural and Cognitive Psychotherapy, 35, 487-494. Huppert, J. D., Ledley, D. R., & Foa, E. B. (2006) The use of homework in behavior therapy for anxiety disorders. Journal of Psychotherapy Integration, 16, 128-139. Judd, C. M., & Kenny, D. A. (1981). Process an alysis: Estimating mediation in treatment evaluations. Evaluation Review, 5, 602-619. Karver, M. S., Handelsman, J. B., Fields, S., & Bickman, L. (2006). Meta-analysis of therapeutic relationship variables in youth and family ther apy: The evidence for different relationship variables in the child and adolescent treatment outcome literature. Clinical Psychology Review, 26, 50-65. Karver, M. S., Shirk, S., Handelsman, J. B., Fiel ds, S., Crisp, H., Gudmundsen, G., & McMakin, D. (2008). Relationship proce sses in youth psychotherapy. Journal of Emotional and Behavioral Disorders, 16, 15-28. Kazdin, A. E. (2002). The state of child and adolescent psychotherapy research. Adolescent Mental Health, 7, 53-59.
70 Kazdin, A. E., Holland, L., & Crowley, M. (1997) Family experience of barriers to treatment and premature termination from child therapy. Journal of Consulting and Clinical Psychology, 65, 453-463. Kazdin, A. E., & Nock, M. K. (2003). Delineating m echanisms of change in child and adolescent therapy: Methodological issues and research recommendations. Journal of Child Psychology and Psychiatry, 44 1116-1129. Kazdin, A. E., Whitley, M. K. (2006). Pre-treatmen t social relations, therapeutic alliance, and improvements in parenting practices in parent management training. Journal of Consulting and Clinical Psychology, 74, 346-355. Kazdin, A. E., Marciano, P. L., & Whitley, M. K. (2005). The therapeutic alliance in cognitivebehavioral treatment of children referred for oppositional, aggressive, and antisocial behavior. Journal of Consulting and Clinical Psychology, 73, 726-730. Kazdin, A. E., Whitley, M. K., & Marciano, P. L. (2006). Child-therapist and parent-therapist alliance and therapeutic change in the trea tment of children referred for oppositional, aggressive, and antisocial behavior. Journal of Child Psychology and Psychiatry, 47, 436-445. Keijsers, G. P., Hoogduin, C. A., & Schaap, C. P. (1994). Predictors of treatment outcome in the behavioural treatment of obs essive-compulsive disorder. British Journal of Psychiatry, 165, 781-786. Kendall, P. C. (1994). Treating anxiety disorders in children: Results of a randomized clinical trial. Journal of Consulting and Clinical Psychology, 62 100-110. Kendall, P. C., & Choudhury, M. S. (2003). Childre n and adolescents in cognitive-behavioral therapy: Some past efforts and current a dvances, and the challenges in our future. Cognitive Therapy and Research, 27, 89-104. Knox, L. S., Albano, A. M., & Barlow, D. H. (1996). Parental involvement in the treatment of childhood obsessive-compulsive disorder: A multiple baseline examination incorporating parents. Behavior Therapy, 27, 93-114. Leonard, H., Swedo, S.E., Lenane, M., Rettew, D.C., Hamburger, S.D., & Bartko, J.J. (1993). A two-to-seven year follow-up study of 54 obsessive-compulsive children and adolescents. Archives of General Psychiatry, 50,429-439. Lewin, A. B., Storch, E. A., Adkins, J. W., Mu rphy, T. K., & Geffken, G. R. (2005). Update and review of pediatric obses sive-compulsive disorder. Psychiatric Annals, 35 745-751. March, J.S., & Mulle, K. (1998). OCD in child ren and adolescents: A cognitive behavioral treatment manual. New York: Guilford. March, J. S., Franklin, M., Nelson, A., & Foa, E. (2001). Cognitive-behavioral psychotherapy for pediatric obsessive-c ompulsive disorder. Journal of Clinical Child Psychology, 30, 8-18.
71 March, J. S., Mulle, K., & Herbel, B. (1994). Behavioral psychotherapy for children and adolescents with obsessive-compulsive disord er: An open trial of a new protocol-drive treatment package. Journal of the American Academy of Child and Adolescent Psychiatry, 33, 333-341. Martin, D. J., Garske, J. P., & Davis, M. K. (2000). Relation of the therapeutic alliance with outcome and other variables: A meta-analytic review. Journal of Consulting and Clinical Psychology, 68, 438-450. Masi, G., Millepiedi, S., Mucci, M., Bertini, N., Pfanner, C., & Arcangeli, F. (2006). Comorbidity of obsessive-compulsive diso rder and attention-deficit/hyperactivity disorder in referred children and adolescents. Comprehensive Psychiatry, 47 42-47. Mataix-Cols, D., Marks, I. M., Greist, J. H ., Kobak, K. A., & Baer, L. (2002). Obsessivecompulsive symptom dimensions as predictors of compliance with and response to behaviour therapy: Results from a controlled trial. Psychotherapy and Psychosomatics, 71, 255-262. McLeod, B. D., & Weisz, J. R. (2005). The Th erapy Process Observational Coding SystemAlliance Scale: Measure characteristics and prediction of outcome in usual clinical practice. Journal of Consulting and Clinical Psychology, 73, 323-333. Meyer, V. (1966). Modification of expecta tions in cases with obsessive rituals. Behavioral Research and Therapy, 4 270280. Mowrer, O. H. (1939). A stimulus-response analysis of anxiety and its role as a reinforcing agent. Psychology Review, 46 553-65. Mowrer, O. H. (1960). Learning theory and behavior. New York: Wiley. Minichiello, W., Baer, L., & Jenike, M. A. (1987). Schizotypal person ality disorder: A poor prognostic indicator for behavi or therapy in the treatment of obsessive-compulsive disorder. Journal of Anxiety Disorders, 1, 273-276. Moritz, S., Fricke, S., Jacobsen, D., Kloss, M ., Wein, C., Rufer, M., Katenkamp, B., Farhumand, R., & Hand, I. (2004). Positive schizotypal symptoms predict treatment outcome in obsessive-compulsive disorder. Behavior Research and Therapy, 42, 217-227. Nafisi, N., & Stanley, B. (2007). Developing and maintaining the TA with self-injuring patients. Journal of Clinical Psychology, 63, 1069-1079). Newman, M. G., & Stiles, W. B. (2006). Therapeutic factors in treatment anxiety disorders. Journal of Clinical Psychology, 62, 649-659. Pauls, D. L., Alsobrook, J. P., Goodman, W, Rasmussen, S., & Leckman, D. L. (1995). A family study of obsessive-compulsive disorder. American Journal of Psychiatry, 152, 76-84.
72 Pediatric OCD Treatment Study (POTS) Team. ( 2004). Cognitive-behavior therapy, sertraline, and their combination for children and adolescents with obsessive-compulsive disorder. Journal of the American Medical Association, 292, 1969-1976. Piacentini, J. (1999). Cognitive behavioral therapy of childhood OCD. Child and Adolescent Psychiatric Clinics of North America, 8, 599-616. Piacentini, J., Bergman, R. L., Jacobs, C., McCrack en, J. T., & Kretchman, J. (2002). Open trial of cognitive behavior therapy for ch ildhood obsessive-compulsive disorder. Journal of Anxiety Disorders, 16, 207-219. Piacentini, J., Bergman, R.L., Keller, M., & Mc Cracken, J. (2003). Functional impairment in children and adolescents with obsessive-compulsive disorder. Journal of Child and Adolescent Psychopharmacology, 13S-1, S61-S69. Riddle, M. A. (1998). Obsessive-compulsive disorder in children and adolescents British Journal of Psychiatry, 173 (S135), 91-96. Robbins, M. S., Turner, C. W., & Alexander, J. F. (2003). Alliance and dropout in family therapy for adolescents with behavior proble ms: Individual and systemic effects. Journal of Family Psychology, 17, 534-544. Rosqvist, J. (2005). Exposure treatments for anxiety disorders: A pr actitioners guide to concepts, methods, and evidence-based practice. New York: Routledge. Rufer, M., Fricke, S., Moritz, S., Kloss, M ., & Hand, I. (2006). Symptom dimensions in obsessive-compulsive disorder: Prediction of cognitive-behavior therapy outcome. Acta Psychiatrica Scandanavia, 113, 440-446. Russell, R., Shirk, S., & Jungbluth, N. (2008). First-session pathways to the working alliance in cognitive-behavioral therapy for adolescent depression. Psychotherapy Research, 18, 1527. Salkovskis, P. M. (1999). Understanding a nd treating obsessive-compulsive disorder. Behaviour Research and Therapy, 37 (Suppl 1), S29-S52. Scahill, L., Riddle, M. A., Mc Swiggen-Hardin, M., & Ort, S. I. (1997). Childrens Yale-Brown Obsessive Compulsive Scale: Reliability and validity. Journal of the American Academy of Child and Adolescent Psychiatry, 36, 844-852. Shelef, K., Diamond, G. M., & Guy, S. (2005). Adolescent and parent alliance and treatment outcome in multidimensional family therapy. Journal of Consulting and Clinical Psychology, 73, 689-698. Shirk, S. R., & Karver, M. (2003). Prediction of treatment outcome from relationship in child and adolescent therapy: A meta-analytic review. Journal of Consulting and Clinical Psychology, 71, 452-464.
73 Shirk, S., & Karver, M. (2005). Process issues in cognitive-behavioral therapy for youth. In P. C. Kendall (Ed.), Child and adolescent therapy: Cogniti ve-behavioral procedures (3rd Edition). New York: Guildford Press. Shirk, S. R., & Saiz, C. C. (1992). Clinical, em pirical, and developmental perspectives on the therapeutic relationship in child psychotherapy. Development and Psychopathology, 4, 713-728. Sookman, D., Abramowitz, J. S., Calamari, J. E., Wilhelm, S., & McKay, D (2005). Subtypes of obsessive-compulsive disorder: Implications for specialized cognitive-behavior therapy. Behavior Therapy, 36, 393-400. Steketee, G. (1993). Social support and treatment outcome of obsessive-com pulsive disorder at 9-month follow-up. Behavioral Psychotherapy, 21, 81-95. Steketee, G. S., Chambless, D. L., & Tran, G. (2001). Effects of Axis I and II comorbidity on behavior therapy outcome for obsessive-c ompulsive disorder and agoraphobia. Comprehensive Psychiatry, 42, 76-86. Stewart, S.E., Geller, D.A., Jenike, M., et al. (2004). Long-term outcome of pediatric obsessivecompulsive disorder: A meta-analysis and quantitative review of the literature. Acta Psychiatrica Scandanavia, 110 4-13. Storch, E. A., & Merlo, L. J. (2006). Obsessive-compulsive disord er: Strategies for using CBT and pharmacotherapy. Journal of Family Practice, 55, 329-333. Storch, E. A., Geffken, G. R., Merlo, L. J., Mann, G., Duke, D., Munson, M., Adkins, J., Grabill, K., Murphy, T. K., & Goodman, W. K. (2007a). Cognitive-Behavioral Therapy for Pediatric Obsessive-Compulsive Disord er: Comparison of Intensive and Weekly Approaches. Journal of the American Academy of Child and Adolescent Psychiatry, 46, 469-478. Storch, E. A., Merlo, L. J., Larson, M. J., Ferna ndez, M., Jacob, M. L., Geffken, G. R., Grabill, K., Murphy, T. K., & Goodman, W. K. ( 2007b). Family accommodation in pediatric obsessive-compulsive disorder. Journal of Clinical Ch ild and Adolescent Psychology 36, 207-216. Storch, E. A., Merlo, L. J., Larson, M. J., Ge ffken, G. R., Lehmkuhl, H. D., Jacob, M. L., Murphy, T. K., & Goodman, W. K. (2008a). Impact of comorbidity on cognitivebehavioral therapy response in pedi atric obsessive-compulsive disorder. Journal of the American Academy of Child and Adolescent Psychiatry Storch, E. A., Merlo, L. J., Larson, M. J., Ma rien, W. E., Geffken, G. R., Jacob, M. L., Goodman, W. K., & Murphy, T. K. (2008b). Clinic al features associated with treatmentresistant pediatric obsessi ve-compulsive disorder. Comprehensive Psychiatry, 49, 35-42.
74 Storch, E. A., Murphy, T. K., Geffken, G.R., Bagner, D. M., Soto, O., Sajid, M., Allen, P., Killiany, E. M., & Goodman, W. K. (2005). Factor analytic study of the Childrens YaleBrown Obsessive-Compulsive Scale. Journal of Clinical Child and Adolescent Psychology, 34, 312-319. Storch, E. A., Murphy, T. K., Geffken, G. R., Ma nn, G., Adkins, J., Merlo, L. J., et al. (2006). Cognitive-behavioral therapy for PANDAS-related obsessive-compulsive disorder: Findings from a preliminary waitlist controlled open trial. Journal of the American Academy of Child and Adolescent Psychiatry, 45, 1171-1178. Tetzlaff, B. T., Kahn, J. H., Godley, S. H., Godley, M. D., Diamond, G. S., & Funk, R. R. (2005). Working alliance, treatment satisfac tion, and patterns of post-treatment use among adolescent substance users. Psychology of Addictive Behaviors, 19, 199-207. Thienemann, M., Martin, J., Cregger, B., Th ompson, H. B., & Dyer-Friedman, J. (2001). Manual-driven group cognitive-behavioral therapy for adolescents with obsessivecompulsive disorder: A pilot study. Journal of the American Academy of Child and Adolescent Psychiatry, 40, 1254-1260. Thomsen, P.H., & Mikkelsen, H.U. (1995). Course of obsessive-compulsive disorder in children and adolescents: A prospective follow-up study of 23 Danish cases. Journal of the American Academy of Children and Adolescent Psychiatry, 34, 1432-1440. Tolin, D. F., Abramowitz, J. S., Przeworski, A ., & Foa, E. B. (2002). Thought suppression in obsessive-compulsive disorder. Behaviour Research and Therapy, 40, 1255-1274. Tolin, D. F., & Hannan, S. (2005). The role of the th erapist in behavior therapy for OCD. In J. S. Abramowitz and A. C. Houts (Eds.), Handbook of obsessive-compulsive spectrum disorders. New York: Kluwer Academic Publishers. Tolin, D. F., Maltby, N., Diefenbach, G. J., Hannan, S. E., & Worhunsky, P. (2004). Cognitivebehavioral therapy for medication nonresponders with obsessive-compulsive disorder: A wait-list-controlled open trial. Journal of Clinical Psychiatry, 65, 922-931. Vogel, P. A., Hansen, B., Stiles, T. C., & Gotestam, K. G. (2006). Treatment motivation, treatment expectancy, and helping alliance as predictors of outcome in cognitive behavioral treatment of OCD. Journal of Behavior Therapy and Experimental Psychiatry, 37, 247-255. Waters, T. L., & Barrett, P. M. (2000). The role of the family in childhood obsessive compulsive disorder. Clinical Child and Family Psychology Review, 3, 173-184. Whittal, M.L., Thordarson, D. S., & McLean, P.D. (2005). Treatment of obsessive-compulsive disorder: Cognitive behavior therapy vs. exposure and response prevention. Behavior Research and Therapy, 43, 1559-1576.
75 Woods, C. M., Chambless, D. L., & Steketee, G. (2002). Homework compliance and behavior therapy outcome for panic with agora phobia and obsessive-compulsive disorder. Cognitive Behaviour Therapy, 31, 88-95. Zohar, A. H. (1999). The epidemiology of obsessive-compulsive disorder in children and adolescents. Child and Adolescent Psychiatric Clinics of North America, 8 445-460.
76 BIOGRAPHICAL SKETCH Mary Lorraine Keeley w as born in Atlanta, Georgia in 1982. She graduated from Marist High School in 2000 and received her Bachelor of Arts degree in psychology from the University of North Carolina (UNC) at Chap el Hill in May 2004. Mary worked in the Developmental Risk and Resiliency Lab at UNC from 1998 to 2000, conducting research on how negative affect and maladaptive coping strategies influence adol escent substance use and other risky behaviors during the high sc hool transition. She entered the Clinical and Health Psychology doctoral program at the University of Florida in 2004 with a concentrat ion in clinical child psychology. While at UF, she worked in the Un iversity of Florida Obsessive-Compulsive Disorder Clinic under the mentorship of Gary R. Geffken, Ph.D. and Er ic A. Storch, Ph.D.