Findings in Relation to What Is Known

In regards to the CBT, the current analysis is consistent with previous ones, such as a Cochrane systematic review, concluding that CBT is an effective treatment for childhood and adolescent anxiety disorders. However, the current review extends the empirical support for CBT, by finding moderate support for the superiority of CBT over treatment as usual or attention control. The current analyses also contribute additional information to the understanding of the necessary and sufficient components of CBT. Specifically, the data suggest that relaxation and cognitive-restructuring do not increase effectiveness above exposure. In fact the presence of cognitive-restructuring was associated with worse outcomes in terms of functioning. Moreover, the current analyses found no differences or fewer dropout rates between CBT and pill placebo, waitlisting, or active control therapies. This finding refutes the belief that patients find CBT (particularly exposure) aversive and unacceptable. Finally, the current report highlights factors that may reduce the effectiveness of CBT including younger age. As such, the current review bolsters the empirical support for CBT by supporting its incremental effectiveness over common therapeutic factors, elucidating some of its active ingredients, and demonstrating its acceptability.

In regards to medication, the current analyses are consistent with previous systematic reviews of psychopharmacologic interventions suggesting that SSRIs and SNRIs have demonstrated effectiveness in the reduction of anxiety symptoms. However, the evidence support for SSRIs is somewhat lessened by the fact that superiority over pill placebo was not found with child report. This issue of inconsistent report is more concerning for SNRIs where support for effectiveness was only found through clinician report, and not through parent or child report of symptoms, or on measures of functioning. In terms of adverse events (AEs), the current review provides the most comprehensive evaluation to date and suggests that short-term AEs tended to be not serious and generally did not lead to discontinuation. Studies were generally of small sizes and short duration and did not report the incidence of serious AEs. There was no evidence of suicidal behavior or ideation associated with the use of SSRIs in children with anxiety, although one trial showed a nonsignificant increase in suicidal ideation with venlafaxine. This contrasted with the well reported two fold increase in suicidal behavior and/or suicidal thoughts associated with SSRIs used for the treatment of depression in children and adolescents, a finding which led to the black box warning. Besides the small sample sizes and short duration, this discrepancy could also be due to the lack of a standardized mechanism for coding and assessing akathisia, aggression, hostility, and suicidal events in pediatric trials and the resulting underreporting of harm events ¹⁶⁵. Evidence on the effectiveness of benzodiazepines and tricyclic antidepressants remains minimal and insufficient to recommend their routine use.^{15, 166}

The current analysis also contributes to the understanding of effectiveness in terms of patient centered outcomes. Whereas previous analyses typically focused exclusively on reduction of anxiety symptoms, the current analyses also examined effects on functioning, anxiety related constructs (such as coping skills), social functioning, and AEs. The results suggested that both CBT and SSRIs improve functioning. The current analyses are also the first to our knowledge to examine the symptom improvement from the perspectives of each stakeholder. Specifically, rather than selecting a single outcome measure from each study, we examined the parent, child, and clinician/evaluator reported outcomes.

Overall, our findings are consistent with existing evidence synthesis reports in terms of demonstrating effectiveness of CBT, SSRIs and SNRIs and pointing out to the need for comparative effectiveness evidence and concerns about long term safety of medications. In terms of existing guidelines, the World Health Organization (WHO), National Institute for Health and Care Excellence (NICE), and British Columbia Medical Services Commission^10–12 are congruent with the current findings in that they recommend CBT as the first-line treatment with medication treatment as a reasonable alternative if preferred by the patient or if CBT was unavailable. In contrast, the American Academy of Child and Adolescent Psychiatry (AACAP) guideline recommends that treatment be multimodal (including a variety of education, psychotherapy interventions, and medications) and informed by the severity of the symptoms and level of impairment.¹⁹ The current findings provide some data to address the need acknowledged in the AACAP guidelines for comparative effectiveness by supporting the superiority of CBT over treatment as usual.

Limitations

Despite anxiety being a common disorder in children, the body of evidence was relatively small and had short followup. A large number of scales were used across studies in overlapping domains, which created a challenge for evidence synthesis, interpretation and translation. Components of interventions and description of participants comorbidities, demographics and social support was either lacking or was provided without stratification per intervention. This rendered numerous subgroup analyses unfeasible. Results of such subgroup analyses would have been most helpful to guideline developers, practitioners and patients because it could have led to nuanced and personalized recommendations. We found indications of potential publication bias when CBT was compared to waitlisting on primary anxiety symptoms. We were unable to statistically evaluate publication bias for most of the comparisons due to small numbers of studies (n<20). The synthesis of data on AEs in particular is limited by the fact that the vast majority of CBT studies do not evaluate AEs and by the lack of a structured consistent approach to measurement in medication studies.

Applicability

The results of this review are likely widely applicable to a heterogeneous population of children with separation anxiety disorder, generalized anxiety disorder, social anxiety disorder, panic disorder, and specific phobia; with minimal psychiatric comorbidities, who are on average 8–18 years old and have ready access to mental health professionals who can provide CBT or have access to psychiatrists or pediatricians who are willing to prescribe SSRIs and SNRIs. Studies published in foreign languages (Spanish and German) demonstrated similar or larger effect size (the effect was in the same direction), compared to studies published in English.

Children of younger ages (3–6) were less presented in the current literature. The majority of the studies were conducted with populations that were predominately Caucasian with limited comorbidity. As such, it is unclear how the results would apply to more diverse populations, patients with comorbidity (especially disruptive behavior), or families with significant additional psychosocial stressors. Most studies also studied treatment naïve children. Thus it is unclear how the results apply to practitioners working with children that have received previous ineffective treatments.

A Guide To Aid in Applicability

To facilitate analysis, data had to be standardized (i.e., expressed in multiples of standard deviations and presented as a standardized mean difference called SMD) or combined using a relative association measure (e.g., relative risk). Such measures may be challenging to interpret by guideline developers or practicing clinicians and can be translated to become more clinically meaningful.¹⁶⁷

One way to make SMD more clinically relevant is to translate it back to scales with which clinicians have familiarity. In Table 13, we provide the average standard deviations for commonly used scales that can be multiplied by SMD for conversion.

Table 13

Average standard deviations for commonly used scales that can be multiplied by SMD for conversion.

As an example of this conversion; compared with pill placebo, fluoxetine reduced primary anxiety symptoms by SMD=−0.40. Multiplying this SMD by the average standard deviation of ADIS scale (1.73) results in −0.69 (which is the expected improvement in anxiety symptoms using ADIS scale). Another approach to aid in the interpretation of the SMD is to consider the magnitude of the effect. SMD cutoffs of 0.20, 0.50, and 0.80 are considered to represent small, moderate, and large effect, respectively.

For binary outcomes presented using a relative effect measure, the effect can be multiplied by the baseline risk to produce an absolute effect and number needed to treat. Using the same example, fluoxetine improved remission by RR= 1.75. Using the average risk in the placebo arms of the included studies (36%), we obtain the absolute effect of 270 patients per 1000 achieving remission (number needed to treat =4). Such conversion can be done in each local setting differently (using a baseline risk appropriate for the setting of the stakeholder) and facilitates the applicability of the findings of this review.

Future Research Needs

Interventions for anxiety in children are complex interventions with multiple components and effect modifiers. However, few studies provided sufficient information to determine the relative effectiveness of such components (e.g. relaxation, exposure, and cognitive-restructuring) or to explore contextual factors that can modify the effectiveness of these complex interventions. Research needs to move away from the simple question of ‘does this work?” to “under what circumstances do medications and psychotherapy work best for children with anxiety?” Therefore, studies need to explore the most beneficial components of CBT and the impact of comorbidities, family demographics and stressors as effect modifiers that can change the effectiveness of treatment. Knowing these effect modifiers would help in providing more individualized treatment. Further research is also needed on long term safety of drugs, treatment of refractory anxiety symptoms, and needs to be more inclusive of underserved populations and minorities. Since anxiety outcomes are measured using a variety of scales without established minimally clinically important differences, studies that establish such differences are needed to better enable clinicians and patients gauge the effectiveness of interventions and balance benefits and harms during a shared decision making process. A large number of analyses, when stratified by an individual intervention, had a small number of included patients leading to imprecise estimates. Considering that anxiety in children is a fairly common condition, larger trials (> 400 participants) with follow up that exceeds 2–3 years are likely feasible and are needed to advance patient care.

Conclusion

CBT is effective in reducing anxiety symptoms and improving function. Medications, primarily those targeting serotonin, are also effective but were associated with various short-term AEs, which were mostly not serious, but studies were too small or too short to assess suicidality with SSRI or SNRI. One trial showed a statistically nonsignificant increase in suicidal ideation with venlafaxine. The combination of medications and CBT is likely more effective than either treatment alone. Comparative effectiveness evidence between various medications and comparing CBT versus medications, or the combination, is limited and represents a need for research in this field. Future research is needed to evaluate components of CBT, effect modifiers of treatment, and long-term safety of drugs, and needs to be more inclusive of underserved populations and minorities.