and/or their combination are the treatment of choice for anxiety conditions in youth.

We also know that, as is the case with adults, some patients do not respond to CBT and pharmacotherapy and many remain somewhat symptomatic. Progress will come from further understanding of biological and psychological mechanisms underlying the disorders and the ways in which our treatment ameliorates anxiety symptoms. Anther avenue worthy of pursuit is the study of predictors for treatment response and failure.

For most child and adolescent anxiety disorders we rely on only a few randomized control trials (RCTs) to guide our decisions about initial treatments, a situation that is in stark contrast to the adult literature, where dozens of studies have been conducted within almost every disorder. Also, with the exception of OCD, there are no completed RCTs that examine the relative and combined efficacy of psychosocial and pharmacological treatments for children and adolescents with anxiety disorders. The adult literature informs us that simultaneous administration of CBT and pharmacotherapy has generally not enhanced outcome of monotherapies. This may be because the effects of many medications are delayed in anxiety disorders, producing symptom reduction after about three months, when most CBT protocols are completed. To maximize the utility of combining treatments, it may be an option to initiate medication first, and begin CBT after medication has reduced anxiety sufficiently to enhance tolerability and to promote information processing from CBT exercises. A recently completed study with adult PTSD sufferers indicates that the addition of exposure therapy to sertraline enhanced medication effects, especially for partial responders (Cahill et al., 2004). No such investigations have been made in youth as yet. In contrast, an initial successful trial with medication may undermine the potency of the exposure task and preclude or reduce its effectiveness.

There is little to no research regarding the management of partial response and nonre sponse in the treatment of anxiety disorders in youth, and the effect of prior treatment with one modality on response to the other has also not been explored. Moreover, although it is unknown at this time, findings from the adult literature suggest that perhaps some children and adolescents will require medication treatment for years. The efficacy and safety of continued pharmacotherapy have not been studied sufficiently. Given recent concerns about growth suppression in youth taking SSRIs and possible increased risk for suicidal ideation in youth receiving paroxetine, the lack of studies of long-term effects of medication in youth constitute a critical gap in our knowledge. With CBT, most studies with adults and youth report long-term naturalistic follow-ups. However, no studies (in adults and youth) available examined the short and long differential responding to different doses of treatment (e.g., 10 vs. 20 sessions). Studies of the effectiveness of different lengths of CBT, both long and short term, are very much needed.

The treatment outcome literature with youth thus far consists of studies that have included wide age ranges, with the majority of participants being younger children. The samples of adolescents in most studies are too small to examine whether children respond differently than adolescents. Given the special developmental stage of adolescence, it is imperative that treatments effective for adults and younger children not be applied blindly to adolescents. Thus, there is a large and significant gap in our knowledge regarding how applicable the available treatments are to adolescents with an anxiety disorder.

Given the available and impressive evidence with adults and the emerging evidence with young children that the various anxiety disorders can be treated effectively, and given the evidence that adolescence is a stage in the lifespan that is fundamentally different from both childhood and adulthood, it is essential that treatments specifically designed for adolescents be developed and that large-scale RCTs be con

ducted to evaluate their acute and long-term safety and efficacy. After a series of RCTs, assuming favorable outcomes from more than a single site and across dependent variables and evaluators, examination of the specific contribution of treatment components would then be needed. With psychopharmacology, issues of tolerance and dosage need to be studied; with regard to psychological treatments, issues related to the role of parents and the influence of peers should be emphasized. It is also extremely important to examine whether the effects of treatment provided during adolescence persist (maintenance of gain), preventing the negative sequelae of anxiety so often seen in adults with these disorders, such as major depression, substance abuse, and underemployment. Studies of samples that span adolescence may also be needed, however, if findings suggest that adolescence itself may moderate outcomes and comparisons across ages would be deemed necessary and appropriate.

Medications tested thus far with youth have generally been well tolerated in the acute treatment, but more information is needed to see whether any gains are maintained as the adolescent faces new challenges. One set of common SSRI side effects that may be especially relevant for adolescents are the sexual side effects, as it is often during this period of life that young people begin to explore their sexuality. Specifically, difficulties with anorgasmia and retarded ejaculation may lead to reduced medication compliance. Much more information is needed to examine this important clinical management issue, although this is but one example of an untoward medication effect that needs to be studied in youth. It is also crucial to explore the interaction between pharmacotherapy and experimentation with alcohol and other drugs, as here again adolescence is a time in life when such experimenting often begins.

Most studies of anxious youth combine patients with various anxiety disorders. Future research should use homogenous samples to evaluate the specificity and generality of the interventions. Such studies will tell us whether medications and/or psychosocial treatments (e.g., for OCD, PTSD, GAD) are differentially effective for adolescents with varying principal diagnoses, or are applicable to more than one di agnostic presentation. We also need to examine the possible complicating role of psychiatric comorbidity, as perhaps we may learn that combined treatments are most needed for patients with other significant and impairing problems beyond the primary diagnosis. To achieve this goal, studies will need to include patients with a wide range of comorbidity.

More information is also needed about prior treatment history and its effect on the evaluation of a current intervention, and there is need for studies of the preferred treatment for treatment-resistant patients. Studies examining the prediction of partial response are also needed, as perhaps a triage approach could be employed if we learned up front which patient and disorder characteristics predict partial response or relapse following delivery of acute treatments.

Fundamental issues facing the field regarding psychosocial treatments for anxiety disorders are as follows:

Recognition of the social, biological, cognitive, and emotional changes that emerge during adolescent development should be used to design treatments specific for adolescents with anxiety disorders. Treatment development and treatment evaluation should follow a logical course, with empirical data on the nature of adolescence used to provide a basis for the intervention. Initially, treatment development

should be guided by empirical data on the nature of adolescence. That is, as treatments are developed, the biological changes and the cognitive and emotional processing features of adolescents should inform them. Given pubertal development and the emerging autonomy and independence of adolescents, researchers need to consider how best to provide treatments (e.g., role of parental involvement). Once developed, these empirically based treatments would then need to be evaluated in RCTs.

We know that the ecology of adolescent development and culture involves an expanded network of peer, school, and community affiliations and that this expanded network increases adolescent risk for exposure to events and circumstances that have been empirically linked with the development of anxiety disorders. The events and circumstances associated with the development of anxiety disorders include possible high-risk sexual and self-injurious behaviors, smoking and drug use, and exposure to traumatic events (e.g., interpersonal and community violence). Some of these factors appear to put young people (i.e., children and adolescents) at risk for developing anxiety disorders as adults. Identifying which factors place young people at risk is a beginning step toward developing selective prevention intervention programs.

Specifically, we know that there are distinct temperamental vulnerabilities to anxiety disorders and anxiety symptoms in some individuals. Moreover, anxiety disorders tend to aggregate in families, and such familial aggregation is due to genetic contributions as well as family environment. A family environment that limits youth independence may particularly put young people at risk. Furthermore, the interactions of children and parents, in which either the child or parent has an anxiety disorder, maintain anxiety and its disorders in young people. Another individual characteristic, anxiety sensitivity, may also serve as a potential risk factor for anxiety disorders, especially panic attacks and panic disorder. A third factor, the presence of anxiety symptoms, at a subthreshold level, may also place young people at risk for developing full-blown DSM-IV anxiety disorders.

We know that not all vulnerable children develop anxiety disorders—in fact, relatively few do. Thus, there are factors that protect young people (i.e., children and adolescents) from developing anxiety disorders. In addition to knowledge about risk factors, knowledge about protective factors constitutes the building block for developing selective preventive interventions programs that will foster protection from pathological anxiety.

Specifically, individual characteristics such as child perceived competence, child coping skills and behavior, level of intelligence, and general “resourcefulness” (e.g., knowing how to solve problems and whom to seek out to help solve problems) serve protective function. Certain environmental resources, such as adequate social support systems (and knowing how to reach out to these systems), and parents who are relatively free of psychopathology and serve as model of coping, also may serve a protective function.

Although research has informed us about potential risk and protective factors, we do not know which factors are linked specifically to anxiety disorders, rather than being general risk and protective factors that play a role in the development of psychopathological conditions (either internalizing, externalizing, or both). Moreover, we do not know how the risk and protective factors differentially affect specific anxiety disorders. The research conducted to date has included diagnostically heterogeneous samples of youths with anxiety disorders (e.g., SAD, GAD, etc.), and participants often are comorbid with other disorders. This limitation is more pronounced in research of protective factors than of risk factors. Further, we know much less about protective and risk factors of anxiety disorders in ethnic and racial groups and in socially and economically disadvantaged groups.

Studies of risk and protective factors conducted thus far have primarily used “main-effects” models, rather than “interactive models.” Consequently, we do not how risk and protective factors interact with one another or whether they serve to either mediate or moderate (or both) anxiety or other psychopathology. We also do not know when in the developmental trajectory of the child potential risk and protective factors have particular influence. Most of the research on risk factors has included either mixed samples of younger and older children (including adolescents) or younger children only. Studies using samples of adolescents only are rare. We should not presume that potential risk and protective factors operate in a similar fashion across developmental stages.

From a selective preventive intervention perspective, we do not know whether targeting any of identified risk factors, protective factors, or some combination (or “packages”) of risk and protective factors in certain subsamples of adolescents would in fact lead to a prevention of anxiety disorders. Relatedly, we do not know whether targeting either particular risk factor(s) (i.e., reducing risk factors), protective factors (i.e., enhancing protective factors), or some combination thereof (i.e., reducing and enhancing risk and protective factors, respectively) would lead to “resilience-building” in adolescence. Indeed, no prevention or resilience building research in the context of anxiety and its disorders has been conducted among adolescents.

We also do not know whether prevention programs aimed at a universal level, which target particular facets of the ecology of adolescent development and culture, reduce anxiety disorders in adolescents. For example, do teen smoking cessation programs reduce or prevent anxiety disorders in teens? If they do, what might be the mediational processes that are operating and might they be moderated by certain adolescent characteristics? We are not even close to knowing the answers to such questions.

It is critical that more research be conducted on obtaining basic knowledge about specific risk and protective factors of anxiety disorders in adolescents. This research must carefully consider the context of adolescent development and culture in trying to discern the particular factors and the manner in which these factors interact. Most risk studies to date combine patients with various disorders, anxiety and otherwise. Future research should use homogenous samples to evaluate the specificity of these factors to anxiety. In addition, research needs to be focused exclusively on specificity of protective factors of adolescents. This research needs to consider carefully not only the role of these factors in the “mainstream” population but also in diverse samples that represent the adolescent population of the United States.

A particularly high research priority is to develop and evaluate selective intervention programs (i.e., programs that target particular risk and protective factors) to learn about their efficacy in preventing anxiety disorders, building resilience, or both in adolescence. If preliminary data indicate that such prevention programs have some positive outcome, the field can move on to discerning more specific questions about enhancing program effectiveness, as well as why these program work and for whom.

As with the development of treatment interventions, recognition of the social, biological, cognitive, and emotional changes that emerge during adolescent development should be used to design prevention programs specific for adolescents; thus far this important work has not been done. Given pubertal development and the critical role of the social network for teenagers, programs developed for adolescents should carefully consider how best to provide such interventions in the school setting and at the same time take into account the possible issues that arise by implementing interventions in schools (e.g., problems with confidentiality).