Running head: EXTERNALIZING DISORDERS AND TREATMENT

Recent research has suggested that individuals may respond to differing treatment approaches with varying degrees of success depending upon their personality characteristics. The current paper attempts to examine how a tendency to engage in externalizing behaviors, as measured by the Child Behavior Checklist, may impact on treatment success for anxiety in a sample of 71 children. Results lent preliminary support to the hypothesis that the presence of an externalizing behavior disorder in a child may interfere with treatment success for anxiety disorders. Externalizing Disorders and 3 Externalizing Disorders and the Treatment of Child Anxiety: A Preliminary Study Much of the first century of modern psychology was spent in the development of theoretically derived treatment modalities for psychological problems. Ultimately this has led to a plethora of competing treatment approaches from which a practitioner may choose each with differing techniques and philosophies (Gurman & Messer, 1995). While this is advantageous in that the practitioner may be presented with multiple tools from which her or she may choose, this also may lead to some confusion in regards to the efficacy of vying treatment approaches. As such the last few years have seen an increased interest into studies for treatment outcome for both children and adolescents for internalizing and externalizing disorders. It was hoped that such studies might elucidate once and for all which treatment modalities might prove most effective. Surprisingly, most of these studies suggested that there were, in fact, no leading therapeutic approaches, and that all therapeutic modalities, including behavioral, cognitive, psychodynamic and even psychotropic medications produce roughly equal success rates (Elkin et al., 1989; Robinson, Berman, & Neimeyer, 1990; Silverman et al., 1999). Unfortunately little of this research focuses exclusively on child populations, and continued investigation into the efficacy of therapy and psychotropic medications in children focuses primarily on attentiondeficit hyperactivity disorder (ADHD; Brown & Ievers, 1999) rather than on phobic or mood disorders. Nonetheless it is gradually being accepted that no one therapeutic intervention is generally superior to any other (Imber et al., 1990) for treatment of internalizing disorders. This observation that varying treatment modalities appear to work equally well may be explained in two possible ways. The first of these is that any therapy is better than no therapy, and exactly what takes place in therapy is relatively unimportant so long as therapy happens. Externalizing Disorders and 4 This is, in essence, the “common factors” explanation of psychotherapy (Weinberger, 1995). This explanation suggests that the “common factors” of psychotherapy, namely the caring therapist, optimism for success, a treatment plan, etc. are all that are necessary for treatment success. Indeed this is in keeping with research which suggests that even non-directive therapy (used as a control group) appears to work in the treatment of child phobias (Silverman et al., 1999) or that placebo treatment appear to work nearly as well as psychotherapy or psychotropic medication (Elkin, et al., 1989). However, it must be pointed out that this conclusion makes a potential error that is common to psychology and the social sciences: namely that group tendencies can be blindly applied to individuals. The “common factors” explanation assumes that everyone in such groups are attaining equal levels of success for each treatment modality, regardless of individual characteristics. Unfortunately, group data tend to blur the role of individuals in their own treatment. For instance, it may be true that different individuals are improving in behavioral psychotherapy than those improving in cognitive therapy or by the use of psychotropic medications. Unless looked for, studies of group data would not likely notice such subtle effects, and thus reach erroneous conclusions about group trends. Alternatively, it may be argued that a number of characteristics of the individual client may effect the outcome of psychotherapy (Beutler, 1991). For instance a patient who is higher in intelligence and more comfortable with verbal exchange may do well in a cognitively oriented treatment, but may find a more behaviorally oriented treatment to be rigid and unrewarding. This may be as true for children as it is adults. Although many individuals may wish to think of children as equal in all qualities, and essentially blank slates, research is clear that children are actually quite different from each other in personality and temperament (Caspi, Henry, McGee, Moffitt, & Silva, 1995), intelligence (Bouchard, Lykken, McGue & Segal, 1991) and even Externalizing Disorders and 5 empathy (Hare, 1993). Exactly how children come to have such diverse abilities and temperaments, whether biological, sociological or simply a function of personality is beyond the scope of this paper. However, it may be that individual differences play an important role in treatment success. Similarly, the nature of a child’s symptomatology, and potential combinations of symptoms from multiple disorders may have a moderating and potentially negative effect on treatment success. Research aimed at elucidating possible interacting mechanisms between therapeutic modality and the individual child client may be helpful to practitioners, clinicians and pediatricians as they seek to tailor treatment approaches to individual children. The alternative is a trial-and-error approach by which varying therapies or medications are attempted on the child with relatively little theoretical or empirical guidance to suggest how and why such efforts would work. Treatment for Child Anxiety It has been noted (Brown & Ievers, 1999) that the majority of treatment outcome studies for children have been for children with the diagnosis of ADHD. Relatively little attention has been paid to the treatment of phobic and other internalizing disorders. Studying the relative effectiveness of varying treatment modalities has been compromised somewhat by the tendency for many therapists to combine treatment methods, using an eclectic approach, rather than diligently following a single theoretical framework (Hill, 1989). Most psychotherapeutic approaches designed to treat phobic disorders have as a principle component exposure to the feared object (Silverman, et al., 1999). However, treatment approaches may differ in the extent to which the child’s parents or the child he or herself may control the exposure process. Contingency management approaches for instance focus on parental initiated rewards for exposure behavior by the child. By contrast, self-control therapy is more cognitive, in that Externalizing Disorders and 6 greater emphasis is placed on the child’s ability to self regulate his or her own exposure and approach behavior. Although, both contingency management and self-control therapy have received attention in the literature, empirical studies are sparse and often combine strategies, or rely on case reports, thus presenting an unclear or incomplete picture of treatment efficacy. Kendall and Braswell (1982) attempted to discern if the cognitive component was an improvement over the behavioral contingency management. They studied 27 children (aged 812) with low self-control randomly assigned to a cognitive-behavioral, a behavioral, or an attention-control group. In this study self-control was measured by the Self-Control Rating Scale, and referred to an ability to refrain from impulsive behavior. Both the cognitivebehavioral and the behavioral therapy approaches were found to lower teachers' ratings of hyperactivity. However, only the cognitive-behavioral approach was found to increase teacher's ratings of self-control, or the child's self-concept of their ability to control their actions. The results of this study suggest that the cognitive components to cognitive-behavior therapy produce additive and positive effects. While this is an interesting and positive finding, it still is not clear that cognitive approaches are beneficial to all children. For example, it was not clear from this study if the intellectual or verbal capabilities of the children played a role in their treatment success. Silverman et al. (1999) conducted a treatment comparison study for anxious children examining the treatment success of contingency management, self-control and educational support treatment. In this study, contingency management referred to an operant behavioral treatment in which parents of the child provided pre-determined reinforcers for a target behavior. Self-control treatment referred to a cognitive-behavioral approach that attempted to teach the child how to identify his or her symptoms of anxiety and develop the ability to control their Externalizing Disorders and 7 anxious symptoms through cognitive restructuring. Education support referred to a comparison treatment approach in which anxiety disorders were discussed, but no specific interventions were conducted with the child. As educational support was a non-directive treatment and did not specifically encourage exposure, the group of children who received this treatment approach were used as a control group. It was hypothesized that children in exposure based groups would demonstrate greater treatment success than those in the educational support treatment group. Similarly it was expected that self-control therapy would demonstrate greater success than contingency management, as a factor of the added cognitive component of that approach. Surprisingly, all three treatment groups demonstrated treatment success, with little difference between the three. The success of the educational support control group may be possibly explained in two fashions. The first factor which may have facilitated treatment success in that group, is that the therapy itself may have raised optimism and self esteem in the children and their parents, and this may have fostered self-initiated exposure. The second factor is simply that maturation may have naturally lessened some of the phobic symptoms regardless of therapy. Once again, however, it remains unclear which treatment approaches may be valuable to which clients. That not all clients in any of the treatment groups were successful leads to the question of how those clients differed from those who did succeed. Furthermore, it may be that client drop out from therapy may have been related to frustration with treatment modality. For example, a child with relatively poor self-control abilities may find self-control therapy to be frustrating and thus drop out. Ultimately it can taken from this study that therapy for child phobic disorders works, but it is not clear for whom it works. Barrett (1998) examined the addition of a family management component to cognitive behavioral treatment for child phobic disorders. Family management referred to fostering a Externalizing Disorders and 8 family’s ability to identify problem behaviors, engage in problem solving as well as mutual support. The rationale for such an addition was that treatment of the child singularly may be less effective if the family unit continues to foster an atmosphere that promotes the child’s phobic behavior. Compared to both a wait-list group and to cognitive behavioral therapy without family management, those children who received both cognitive behavioral therapy and family management demonstrated a greater degree of treatment success. This study is important in demonstrating an important interaction between family, child and therapy. As such, this study represents an increased degree of sophistication over many of the other studies available. Yet while involvement of the family into the therapeutic process is an important step, the specific role of the child has still been largely ignored. A question that may be important for researchers and therapists to consider may be not only which treatment modality is more effective, but under what circumstances is one treatment modality more effective than another? Specifically, given the heterogeneity of the population of phobic children, which children would benefit from one treatment strategy compared to another, and what individual characteristics may be prognostic of treatment success or failure with the use of certain strategies? To date little research has been conducted on variables that may moderate the effectiveness of treatments for childhood anxiety disorders. It may be possible that many of the children who present to practitioners with phobic disorders may have other comorbid conditions, or even personality-related disturbances that may interfere with therapy. Thus even the child’s temperament or personality may impact on treatment success. For example, although anxiety disorders have been found to have a significant degree of comorbidity with externalizing disorders (Clark, Smith, Neighbors, & Skerlec, 1994; Russo & Beidel, 1994), few studies have considered the impact which externalizing symptoms may have on treatment outcome. Indeed, Externalizing Disorders and 9 there is very little literature in regards to what child variables may moderate treatment success, either for treatment in general, or as a means for comparing the relative success of varying treatment modalities. With the goal in mind of examining whether some children may respond to therapeutic interventions differently as a factor of their own behavioral characteristics, the purpose of the present study was to examine the role of children's externalizing behavior problems as a moderating variable for the treatment of phobia in children. Specifically, the role of externalizing behavior problems was considered in comparing the relative effectiveness of selfcontrol therapy and contingency management as well as educational support in the treatment of specific phobias. Given that self-control therapy is more highly dependent upon internal regulatory systems that may or may not be possessed by specific children, it may be that this modality will be particularly vulnerable to impulsive or oppositional behavior on the part of the child. A child who demonstrates externalizing behavior problems may not be able to manage the cognitive skills necessary to organize and direct their own thoughts and behavior consistently. As such, cognitive self-control therapy may ultimately prove counter-productive for such children. In the current study it was hypothesized that externalizing behavior problems have a negative impact on treatment success across all treatment conditions. It also was hypothesized that externalizing behavior problems will predict significantly treatment failure in self-control therapy, relative to the contingency management and control conditions. This may be due to a block in the transfer-of-control from parents to children brought on by the inability of children to assume rational control of their own behavior. Transfer-of-control refers to the treatment goal of transferring the responsibility for improving target behaviors from the parent to the child. For instance, parents may initially control a target behavior through reinforcement of that behavior, Externalizing Disorders and 10 though it may be desirable for the child to ultimately demonstrate the ability to exercise selfcontrol (thus transferring control from parent to child) over that behavior. For the purposes of this study, externalizing behavior problems were operationally defined by behaviors as measured by the externalizing sub-scale of the Childhood Behavior Checklist (Achenbach & Edelbrock, 1983). This study is viewed as important in the determination of specific utility of these cognitive-behavioral strategies for children with different levels of internal self-control. The ultimate goal of this study is to provide empirical data by which specific treatments may be matched to a child's unique requirements. Methods Participants Participants of the study were 71 children and adolescents between the ages of 6 and 17 (mean age=10.37) who presented at a university sponsored child phobia program with a primary diagnosis of specific or social phobia. Thirty-seven of the subjects were boys and thirty-four were girls. Participants were referred to the clinic most often either by clinicians, school counselors and psychologists, or by self-referral. Although children referred to the clinic presented with a variety of symptoms and disorders, only those with a primary diagnosis of specific phobia or social phobia were included in this study. Measures Child Behavior Checklist (CBCL). The CBCL (Achenbach & Edelbrock, 1991) is designed to measure a variety of behavioral problems and social factors in children. The behavior problems portion of the CBCL consists of 118 items rated on a scale from 0-2, with 0 indicating not true for the child, 1 indicating somewhat true, and 2 indicating very true within past 6 months. Profiles have been revised and standardized in 1991 for boys and girls of age Externalizing Disorders and 11 ranges 4-11 and 12-18. Two global factors, Internalizing and Externalizing have been identified for the CBCL, with reported reliabilities of .87. Validity of the CBCL has been supported by its ability discriminate between clinically referred and non-referred children, as well as convergent validation with other behavioral measures. (Achenbach & Edelbrock, 1991). The Externalizing factor contains a number of items related to impulsivity and hyperactivity, and thus will be considered as a part of this study. Anxiety Disorders Interview Schedule for Children (ADIS-C and ADIS-P). The ADISC/P (Silverman & Nelles, 1988) is a semi-structured interview schedule designed for the assessment of anxiety disorders in children and adolescents. Interviews are conducted with both children and their parents, from which a composite diagnosis is attained, as well as an indication of distress and interference. The ADIS has been found to have satisfactory reliability, both interrater (Silverman & Nelles, 1988) and test-retest (Silverman & Rabian, 1995), as well as good validity (Rabian, Ginsburg & Silverman, 1994). Composite ratings of the severity of diagnosis for phobic disorders, based upon child and parent interviews will be used as a primary outcome measure. Revised Children's Manifest Anxiety Scale (RCMAS, RCMAS-P). The RCMAS (Reynolds & Richmond, 1978), consisting of child and parent versions, is a 37-item scale designed to measure anxious symptomatology. Responses are answered 'yes' or 'no'. The scale consists of a total score, an 11-item lie scale, and three factor scales, 'Physiological Anxiety', 'Worry/Oversensitivity' and 'Social Concerns/Concentration'. Reliability, internal consistency, and validity were measured extensively for total scale score, factor scales and the lie scale with all results within the satisfactory range (Reynolds &