Journal of Childhood & Developmental Disorders

Introduction

Pediatric Anxiety Disorders (ADs) affect up to one-third of all children, cause significant impairment in day-to-day functioning, and predict psychopathological difficulties in adulthood. A range of trials has shown that Cognitive-Behavioral Therapy (CBT) is an efficacious treatment of pediatric ADs and CBT is considered a first-line treatment in most guidelines.

Few children and adolescents with ADs seek care and among those who do, few are offered evidence-based interventions. This treatment gap has led to calls for widespread dissemination of CBT and other evidence-based interventions for mental health symptoms among youth. Traditionally, CBT has been conducted using face-to-face formats, either individually or in group settings, but during the last decade, computer- and Internet-based adaptations of CBT (tCBT) have been developed in which the full treatment (or portions of it) is delivered using technological devices such as computers or cellphones. Given that CBT is a highly structured treatment, it may lend itself well to module-based interventions carried out online or by means of technological devices. Proposed advantages of tCBT compared with face-to-face CBT are that tCBT is not bound by geographical distance and can make evidence-based treatment more available and scalable. Further, children and adolescents are often highly active online and may prefer tCBT over face-to-face interventions. tCBT also requires less therapist time than faceto- face treatment (contact with patients are often conducted via chat or email) and may therefore be more cost-effective than traditional face-to-face CBT, which can facilitate dissemination. tCBT has also been proposed as a potential first phase in stepped-care models of pediatric ADs.

Understanding Cognitive-Behavioral Therapy (CBT)

Several tCBT protocols have been developed and evaluated for pediatric ADs, with a seminal study published as early as 2006. To date, two meta-analyses have synthesized some of this evidence. Zhou et al., conducted a network meta-analysis and compared the efficacy of different types of delivery formats of CBT in relation to change in self- and caregiver/teacher-reported anxiety. Eleven trials of Internet-assisted CBT were grouped and results indicated that Internet-assisted CBT was more effective than waitlist and no treatment, but less effective than groupbased CBT. Of importance, remission rates were not analyzed and some trials included participants with obsessive-compulsive disorder (OCD). Another meta-analysis found that the efficacy of Internet-assisted CBT may be similar to face-to-face CBT in relation to change in self-reported anxiety but nonrandomized trials and trials that did not use structured diagnostic interviews were included. Other reviews have included trials with participants over 18 years of age and have pooled treatment and prevention studies. Thus, current meta-analyses have several limitations that prohibit the generalization of effects into the treatment of anxiety-disordered youth. A major limitation is that no meta-analysis has synthesized trials that used structured diagnostic interviews at pre- and posttreatment, leaving the efficacy of tCBT in relation to diagnostic remission unknown. Further, inclusion has often been based on elevated scores on broad self-reported anxiety measures, which is not the same as meeting criteria for an AD. Therefore, it is currently unclear whether tCBT is efficacious for pediatric ADs and whether it has an effect on remission. These are important gaps in the literature as information on remission and effects for young individuals that meet criteria for ADs is crucial for whether tCBT should be recommended in treatment guidelines for ADs and offered in child and adolescent mental health settings.

To address these gaps, we conducted a systematic review of randomized controlled trials (RCTs) of tCBT that included patients <18 years of age with a primary AD according to a structured diagnostic interview. The primary outcome was remission for the principal/primary AD. Secondary outcomes were remission for all ADs, posttreatment self- and caregiverreported anxiety, and posttreatment clinician-rated functioning.

Symptom Reduction

A majority of studies included participants with different types of ADs and type of AD was not accounted for in the metaanalyses. Detailed accounts of therapist involvement were seldom provided; however, as mentioned above, some trials included face-to-face sessions and these trials were contrasted with trials without face-to-face sessions in meta-regression. Three types of control conditions were used: waitlist, TAU, and psychological placebo. Because of the small number of studies, control condition was dichotomized and coded as either waitlist or active intervention (i.e., TAU, psychological placebo). Adverse events were only reported in one study and for the five studies that were preregistered, the assessment of adverse events was only mentioned in the Jolstedt et al., trial registration protocol. Study characteristics are presented.

Limitations of the present report merit mentioning. First, few trials fulfilled PICO criteria, which reduces precision in estimates and results in low power for meta-regression and all metaregression results should be considered preliminary and interpreted with caution. Second, because of few trials, we pooled different control conditions. It may be argued, and rightly so, that the effect of tCBT versus waitlist is a different effect than tCBT versus psychological placebo or TAU, but because of low power we could not conduct separate metaanalyses or use proper moderator meta-analysis. Last, few studies reported separate child and caregiver information about proportion of completed sessions and we analyzed a pooled estimate that included both child and caregiver data.

In sum, tCBT has a moderate effect on remission for pediatric ADs and overall functioning compared with control conditions. With respect to youth-reported anxiety, tCBT may not result in a better effect than control conditions, and it may only produce a small effect in relation to caregiver-reported anxiety. Remission rates across studies vary greatly and tCBT may be less efficacious than face-to-face CBT, but noninferiority trials are needed. Future trials of tCBT should follow preregistered protocols, use structured diagnostic interviews, assess and report adverse events, and consider contrasting ‘pure’ tCBT with partial tCBT.