The Cochrane Library and the treatment of childhood and adolescent anxiety disorders: an overview of reviews


  • Katharina Manassis,

    Corresponding author
    1. Department of Psychiatry, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada
    • Department of Psychiatry, Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada.
    Search for more papers by this author
  • Kelly Russell,

    1. Alberta Research Centre for Child Health Evidence, Department of Pediatrics, University of Alberta, Edmonton, Alberta, Canada
    Search for more papers by this author
  • Amanda S. Newton

    1. Departments of Pediatrics and Psychiatry, Faculty of Medicine and Dentistry, University of Alberta, Edmonton, Alberta, Canada
    2. Women and Children's Health Research Institute, University of Alberta, Edmonton, Alberta, Canada
    Search for more papers by this author



Anxiety disorders are among the most common psychiatric disorders diagnosed during childhood and adolescence. Reported lifetime prevalence of children or adolescents meeting criteria for at least one anxiety disorder in industrialized countries ranges from 8–27%. Current treatment includes psychotherapy (cognitive and behavioural therapies) as well as medication which is almost always used together with psychotherapy, rather than as a stand-alone treatment.


To synthesize the evidence currently in the Cochrane Database of Systematic Reviews (CDSR) related to the question: ‘In the treatment of childhood and adolescent anxiety disorders, which pharmacologic or nonpharmacologic treatments are known to improve symptom response, response rates, functional capacity, adherence, persistence, and acceptability as well as increase diagnostic remission and decrease adverse events?’.


The CDSR was searched using the term ‘anxiety disorders’ in the title for all systematic reviews examining pharmacologic and nonpharmacologic interventions for the treatment of anxiety disorders in children and adolescents, including pharmacotherapy and psychotherapy. Data were extracted and entered into tables; data were synthesized using qualitative and quantitative methods.

Main Results

Of the studies included in the CDSR, treatment of childhood and adolescent anxiety disorders with cognitive behavioural therapy (CBT) led to significant reductions in anxiety symptoms and increased recovery. Treatment with CBT or behavioural therapy (BT) led to notable reductions in Obsessive–Compulsive Disorder (OCD) severity and a reduced risk of treatment failure. Use of selective serotonin reuptake inhibitors (SSRIs) and the selective norepinephrine reuptake inhibitor (SNRI) venlafaxine were superior to placebo in treating OCD and other anxiety disorders. There was no clear evidence that any particular SSRI or venlafaxine was most efficacious or best tolerated. While few studies were available, CBT combined with a SSRI or SNRI led to significant reductions in both anxiety and OCD symptoms. Psychotherapy (CBT/BT), used alone or in combination with medication, had a mixed impact on reducing risk of treatment failure for OCD.

Author's Conclusions

For childhood and adolescent anxiety disorders, including OCD, the CDSR reviews suggest that psychotherapy treatments are efficacious in reducing symptom severity. Although the CDSR does not include a number of recent research publications on CBT, these newer studies further reinforce CBT efficacy. Pharmacotherapy evidence from the CDSR supports using medication in treating anxiety disorders, and while few studies examined combined pharmacological and psychological treatment, results to date are also favourable for this combination. Clinicians should rely on expert consensus guidelines vis-à-vis this evidence as treatment decision-making should be moderated by the patient's illness severity. Psychotherapy remains the first line treatment for mild to moderate symptoms, whereas pharmacotherapy is used for severe or treatment-resistant disorders. In conclusion, there is a body of literature in the CDSR to support evidence-based treatment decisions for pediatric anxiety disorders; however, as this is a field that is rapidly expanding its knowledge base, efforts must be made to ensure the most recent evidence is consistently incorporated. Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.

Editors' note:Overviews of reviews, compiling evidence from multiple Cochrane reviews into one accessible and usable document, is a regularEditors' note:Overviews of reviews, compiling evidence from multiple Cochrane reviews intoEditors' note:Overviews of reviews, compiling evidence from multiple Cochrane reviews into one accessible and usable document, is a regular feature of this Journal. Our aim for each overview is to focus on the treatment question, ‘which treatment should I use for this condition?’, and to highlight the Cochrane reviews and their results in doing so. It is our hope that the overview will serve as a ‘friendly front end’ to the Cochrane Library, allowing the reader a quick overview (and an exhaustive list) of Cochrane reviews relevant to the clinical decision at hand.


Description of the condition

Anxiety disorders are among the most prevalent of mental, emotional, and behavioural problems to be diagnosed during childhood and adolescence1–3, and are considered one of the earliest forms of psychopathology4. The anxiety disorders include Generalized Anxiety Disorder (GAD; formerly Overanxious Disorder), Separation Anxiety Disorder, Social Phobia (formerly Avoidant Disorder), Specific Phobia, Panic Disorder, and Obsessive-Compulsive Disorder (OCD).1 These disorders share common clinical features including developmentally inappropriate and persistent anxiety with associated physiological symptoms, behavioural disturbances (i.e. avoidance of feared objects, repetitive behaviours) as well as impaired quality of life and daily functioning (at school, during social interactions)4, 5. Differences in clinical presentation exist across the disorders and within diagnostic groups with core diagnostic criteria presenting differently depending on age4, 6, 7. For example, anxious children may present with physical complaints or temper tantrums rather than anxieties and worries. There can also be fluctuation in symptom severity over time, which is predominantly notable in OCD8. Reported lifetime prevalence of children or adolescents meeting criteria for at least one anxiety disorder in industrialized countries ranges from 8–27%2, 9–11 with comparable prevalence between boys and girls12, 13. Early onset anxiety disorders (<13 years of age) tend to follow a chronic course14. The most frequently diagnosed disorders include GAD (over 10% of children and adolescents) with an average age of onset of 8 years15, followed by Specific and Social Phobias (10% and 7%) with ages of onset during early and late adolescence, respectively6, 16 and Separation Anxiety Disorder (4%) with an average age of onset of 7 to 8 years11. OCD and Panic Disorder are lower-prevalence conditions in childhood. OCD has a lifetime prevalence of 2–3% and age of onset is seen across childhood and adolescence17. Panic Disorder occurs in less than 1% of children with a slightly higher prevalence in adolescence (2–3%)6; age of onset is typically in later adolescence3.

Pediatric-onset anxiety disorders do not necessarily remit over time6, 14 and can have long-term implications if left untreated. Lifetime co-morbidity between anxiety and other psychiatric disorders (primarily depression) is also substantial18–21. Access to timely assessment and treatment is essential. For mild anxiety symptoms, psychotherapy can be sufficient in providing symptom relief22–24. For moderate to severe conditions, medication is almost always used together with psychotherapy, rather than as a stand-alone treatment22. Combining medication with psychotherapy often increases disorder acceptability for families and reduces the risk of symptom recurrence after discontinuation.

Description of the Interventions

Drug-based treatments

Many intervention studies (pharma- and psycho-therapeutic) have included children with one or more anxiety disorders as there is a high degree of co-morbidity among these conditions in pre-pubertal children8. The exception is OCD which is a relatively distinct condition and is therefore studied separately25. Early studies focused on tricyclic antidepressants (TCAs) and benzodiazepines, but results were mixed and side effects were often problematic (anticholinergic effects and QT prolongation in the former; disinhibition and tolerance in the latter)22. More consistent evidence of efficacy has been found in recent studies using selective serotonin reuptake inhibitors (SSRIs)26–28. Short-term use of benzodiazepines is still common for specific anxiety-provoking situations (for example, first few days of school in a child with school phobia) or while waiting for an SSRI to take effect. Clomipramine (a TCA that is relatively serotonin-specific) is still used for OCD, particularly in children who do not respond to SSRI treatment27. Augmenting an SSRI with a second medication (most often an atypical antipsychotic) is more common in OCD, as SSRI's often result in only moderate symptom reduction.

Selective serotonin re-uptake inhibitors (SSRI's)

Serotonin has been linked to mood, anxiety, and obsessive–compulsive behaviours29, prompting the use of SSRIs in childhood anxiety disorders, including OCD. Because of parental concerns about use of medication in their children, they are often reserved for children who are severely impaired by anxiety or OCD. Recent multi-site controlled trials suggest these medications are efficacious for reducing anxiety and OCD symptoms, and further symptom reduction may occur when they are combined with cognitive behavioural therapy25, 28, 30. It is not clear, however, whether any one SSRI is superior to the others for either anxiety disorders or OCD26. Despite promising outcomes in clinical trials, many patients remained symptomatic at the end of treatment31.

Although they are generally well-tolerated, SSRI's have been linked to increases in suicidal thinking in some studies, particularly in depressed children and youth13. When examining the risk-benefit ratio of these medications, however, benefits usually outweigh risks in children with anxiety disorders or OCD who are not depressed32. Gastrointestinal upset, restlessness, and sleep disturbance are more common adverse effects33. Frequent monitoring of the child for beneficial and adverse effects is prudent, especially in the first few weeks of treatment22.

Psychotherapy interventions: behavioural and cognitive-behavioural treatments

Behavioural and cognitive-behavioural treatments are the mainstay of treatment for children with mild to moderate anxiety, and are often used together with medication with those who are severely impaired34. Many studies of psychotherapy treatment excluded children on medication and many parents will not consider participation in a medication trial if their child is only mildly anxious, so treatment studies generally include children with less severe anxiety-related impaired than medication-focused trials. Behavioural treatment (BT) is based on the premise that repeated exposure to feared situations results in desensitization to them, reducing anxiety35. Cognitive-behavioural treatment (CBT) combines exposure with attempts to replace anxious, catastrophic thoughts (termed ‘anxious cognitions’) about the situation with more adaptive, encouraging thoughts36. Relaxation strategies are sometimes included as well. Anxious feelings will subside as the child thinks less catastrophically. Given children's cognitive development, these strategies are simpler and more concrete than in adults and often include participation by parents. Generally, the younger the child the more behavioural (relative to cognitive) strategies are used and the greater the need for parental involvement37.

Philip Kendall pioneered the ‘Coping Cat’ CBT program for non-OCD childhood anxiety disorders, and various adaptations of this program (both group and individually based) have been shown efficacious in randomized controlled trials internationally34. This program was recently adapted for the CAMS (Child/Adolescent Anxiety Multimodal Study), a large-scale examination of combined pharmacological and psychotherapy treatments for three childhood and adolescent anxiety disorders (SAD, GAD, social phobia)38. ‘Coping Cat’ emphasizes cognitive restructuring (replacing anxious thoughts with more adaptive thoughts) together with exposure to feared situations and relaxation exercises36. CBT programs for OCD have been combined with SSRI use39, 40 and typically emphasize exposure and response prevention (resisting engaging in compulsive behaviours in situations where they typically occur) to reduce compulsive behaviours. Cognitive and relaxation strategies are usually needed as well, however, to help children tolerate the exposures and to address obsessions.

Studies determining treatment components essential to success, comparing CBT to other interventions, adapting CBT for specific populations (e.g. children with cognitive limitations, with certain co-morbidities, or in certain age groups), and examining effectiveness in community samples are a priority28, 38, 39, 41, 42.

Parental/family involvement in treatments

Child CBT programs generally include parental psycho-education about the nature of childhood anxiety and its treatment. Programs vary, however, in the degree and nature of parent or family involvement in sessions43. Some programs involve parents in only a few sessions, while others involve them in all sessions; some treat parents and children separately (i.e., parent training) while others include family or multi-family sessions. Several studies have shown greater impact on treatment outcomes and family dynamics when parents are involved in both child and adolescent care44, 45–47, especially when a parent also suffers from an anxiety disorder48, 49.

Why This Overview is Important

Anxiety disorders are among the most prevalent of psychiatric disorders to be diagnosed during childhood and adolescence, and are chronic in nature. The consequences of poorly managed or untreated anxiety disorders include substantial concurrent and long-term impairment, and place the child/adolescent at risk for additional psychiatric problems such as depression or substance abuse9. Evidence-based treatment offers the hope of ameliorating these adverse outcomes and restoring the child/adolescent to a more normative developmental course. By summarizing relevant information from the Cochrane Library for the practicing clinician, this review aims to contribute to that goal.

Objective of the Overview

The objective of this overview was to synthesize evidence currently in the Cochrane Library Database of Systematic Reviews (CDSR) related to the following question: ‘In the treatment of childhood and adolescent anxiety disorders, which pharmacologic or psychotherapy treatments are known to improve symptom response, response rates, functional capacity, adherence, persistence, and acceptability as well as increase diagnostic remission and decrease adverse events?’


Search strategy

The CDSR was searched for all systematic reviews published prior to 2010 examining the treatment of anxiety disorders in children and adolescents, including pharmacotherapy and psychotherapy treatments. By these means, we indentified five reviews examining treatment options for childhood and adolescent anxiety50, 51–54. Reviews that included both pediatric and adult trials were included provided that pediatric results could be extracted. Reviews that examined anxiety disorders and other psychiatric conditions were included provided the results pertaining to the treatment of anxiety disorder could be extracted. One review that examined anxiety treatment included only one study and this same study and associated results was included in a second review52. For this reason, one review was excluded from this overview of reviews53. A second review was also excluded because the authors could only identify trials that assessed exercise as a preventative measure against developing an anxiety disorder54. Three reviews met the inclusion criteria and were included in this overview of reviews50, 51, 52. One review published an update in 201051 that was not captured in our overview, but reported main findings of the updated review are consistent with the earlier publication. We comment on the updated review's findings in our discussion.

Data extraction and synthesis

One reviewer (KR) extracted the following data from each review: search strategies, inclusion criteria including population, intervention, comparison(s), and outcomes, methodological quality assessment, and results. Both quantitative and qualitative data were entered directly into MS Word® tables. We did not include results from non-parametric data analysis from one review50 and narrative data from another review52. Across reviews, dichotomous data were presented as relative risks (RR) with 95% Confidence Intervals (CI) and continuous data were summarized as standardized (SMD) or weighted (WMD) mean differences with 95% CI. The RR describes the probability of the event in the treatment group compared to the probability of the event in the control group. When the same scale was used to measure the outcome, a WMD was calculated. If the outcome was measured using a variety of scales, a SMD was calculated. When data from only one study were included, a mean difference (MD) was calculated. The 95% CI is an estimate of the true population value and if it did not cross 1, the result was interpreted as statistically significant. In the event that a pooled effect estimate measure had to be calculated, random effects modelling was conducted to provide the most conservative estimate; the accompanying I2 value represents the degree of statistical heterogeneity among the studies.


Description of the Cochrane reviews identified

Criteria for including studies in the reviews

Two reviews included randomized controlled trials50, 52 and a third review included both randomized and quasi-randomized controlled trials51. Reviews were published between 2005 and 2009. The age of the included children was similar: one review included children and youths less than nineteen years of age52, a second review included children 6–18 years of age50, while the third study included those less than nineteen years of age or described as ‘children or adolescents’51. Two reviews used universally standardized diagnostic techniques to assess anxiety50, 52. Ipser et al. included children with a DSM-III/III-R/IV diagnosis of anxiety disorder and excluded children with anxiety symptoms secondary to a primary disorder52. Children with a secondary disorder were included provided they were not receiving treatment for the non-anxiety disorder. James et al. included DSM III/III-R/IV/IV-TR or ICD 9/10 diagnosis of several anxiety disorders; however, excluded children with post-traumatic stress disorders, simple phobias, (s)elective mutism, or OCD50. O'Kearney et al. limited their inclusion criteria to children with OCD, as determined by clinical assessment or standardized diagnostic interview51.

Both the interventions and control groups were variable. The interventions included medications (benzodiazepines, SSRIs, TCAs, and others) and psychotherapy treatments. Two reviews included behavioural therapy or cognitive behavioural therapy treatments51; one of the reviews excluded children taking concomitant medications50. Within each review and between reviews, the control groups were quite diverse and included both active controls, such as diary, support, medications, and non-active controls such as placebo or wait list. The primary outcome in all the reviews was treatment response, as measured by diagnostic remission or reduced symptomatology.

Search strategies

The search strategies used to identify the included studies were comparable. All three reviews searched Medline, PsycINFO, EMBASE, and the CCDANCTR (Cochrane Collaboration Depression, Anxiety and Neurosis Controlled Trials Register). One review performed more exhaustive searches and also searched Clinical trials registries and Dissertation Abstracts52. All reviews checked the reference lists of relevant studies and/or reviews. Two reviews conducted hand-searching50, 51, and one also reviewed pertinent book chapters52. Key researchers were contacted by all three reviews while one review also communicated with pharmacological companies52.

Description of studies

Table 1 describes the specific inclusion criteria of the included reviews. The number of pediatric trials in the reviews ranged from 4 to 22 and the sample sizes of the individual trials ranged from 10 to 322. Treatment effect was measured differently in the three reviews and a variety of scales were used. One review reported treatment response in a dichotomous manner by assessing the presence or absence of ICD or DSM anxiety disorder50. O'Kearney et al. assessed post-treatment scores in OCD clinical symptoms and several clinician-rated or self-reported scales were used51. When possible, the continuous scores were dichotomized to represent responders vs non-responders and diagnosis remitted vs diagnosis not remitted. The third review, conducted by Ipser et al., assessed treatment response by dichotomizing the Clinical Global Impressions scale (CGI-I): children were considered to have responded to pharmacological treatment if they experienced ‘very much improvement’ or ‘much improvement’52. If the CGI-I was not used, other scales were dichotomized in a similar manner. The mean (or median) responses were not provided.

Table 1. Description of included reviews
Title Authors Date of most substantive amendmentNumber of studies Study sample size (range)Clinical population and severity Co-morbiditiesInterventionControlRange of follow-upOutcomes
  • CBT: Cognitive Behavioural Therapy; DSM: Diagnostic and Statistical Manual; ICD: International Classification of Disease; NR: not reported; OCD: Obsessive–Compulsive Disorder.

  • For key therapeutic techniques see: Kendall PC. Coping Cat Workbook, 3nd edn. Workbook Publishing: Philadelphia, PA, 2006.

  • ±

    For key therapeutic techniques see: March J, Mulle K. OCD in Children and Adolescents: A Cognitive-Behavioral Treatment Manual. Guilford: New York, NY, 1998.

Cognitive behavioural therapy for anxiety disorders in children and adolescents13Children aged 6–18 years diagnosed with anxiety disorders according to ICD 9/10 or DSM III/III-R/IV/IV-TR. Those with post-traumatic stress disorder, simple phobias, elective mutism, and OCD were excluded.CBTWait list or active controls who received attention only (diary or support)3–12 monthsPrimary
James ACJ, Soler A, Weatherall RRW809 (12–100)Severity: NR   • Presence or absence of a diagnosis of anxiety disorder as evaluated by reliable and valid structured interviews
October 2005 Co-morbidities: NR   Secondary
      • Reduction in anxiety symptoms
      • Acceptability
Behavioural and cognitive behavioural therapy for obsessive compulsive disorder in children and adolescents4Children aged ≤18 years or described as ‘children and adolescents’ in the study with OCD diagnosed by a clinical assessment or standardize diagnostic interview.CBT and BT±Active drug, pill placebo, or wait list12–43 weeksPrimary
O'Kearney RT, Anstey K, von Sanden C222 (10–112)Severity: NR  (43 weeks of drug alone and 33 weeks of BT/CBT)• OCD clinical symptom frequency, duration and degree of distress of obsessions and compulsions
October 2006 Co-morbidities: NR   • Treatment response
      • OCD remitted
      • General levels of distress and disruption
      • Quality of life
      • Adverse events
      • Dropouts
      • Acceptability of treatment
Pharmacotherapy for anxiety disorders in children and adolescents22Children and/or adolescents diagnosed with anxiety disorders according to DSM-III, DSM-III-R, and DSM-IV.SSRI or TCAPlacebo≤16 weeksPrimary
Ipser JC, Stein DJ, Hawkridge S, Hoppe L2519 (15–322)Severity: Studies concerning anxiety symptoms that were part of the clinical presentation of another primary diagnosis were excluded.   • Treatment response measured by Clinical Global Impressions (CGI) scale. Score of 1 (very much improved) or 2 (much improved)
August 2008 Co-morbidities: Other psychiatric diagnoses were not excluded providing they were secondary to the anxiety disorder being treated and the child was not undergoing treatment for secondary disorder.   • Similar scale if CGI was not used
      • Response of co-morbid symptom of medication
      • Quality of life
      • Functional disability
      • Withdrawal due to adverse event
      • Most common adverse events
      • Long-term treatment response

Methodological quality of the studies included in the reviews

The methods used to evaluate methodological quality varied by review. Ipser et al. used the Risk of Bias tool and the 23-point scale developed by the CCDAN Quality of Research Scale (CCDAN-QRS) to assess various aspects of methodological quality including sample size, duration of intervention, inclusion and exclusion criteria, and adequate power52. O'Kearney et al. also used the CCDAN-QRS; however, the results were not provided. This review also assessed allocation concealment (selection, performance, detection, and attrition bias)51. The final review, James et al., utilized the Jadad Scale and Concealment criteria to assess methodological quality. The James review was the only review that excluded low methodological quality trials; low quality trials were defined as those with inadequate concealment or unclear concealment and a Jadad Score of less than 2; the Jadad Score of included studies, however, was not reported50. The average CCDAN-QRS score was 27/4652. The proportion of trials with adequate allocation of concealment ranged from 1/13 (8%)50 to 2/4 (50%)51.

Effects of interventions

Drug-based treatments

One review examined the effectiveness of pharmacological treatment compared to placebo (Table 2)52. When assessed using Clinical Global Impression (CGI-I scale), anxiety symptoms significantly improved among children and adolescents prescribed a SSRI (RR = 0.42, 95% CI:0.35,0.50) or SNRI (RR = 0.68, 95% CI:0.58,0.88) compared to placebo. Measures from the Diagnostic and Statistical Manual of Mental Disorders (DSM) did not detect any significant difference in anxiety severity among those treated with SSRI or placebo. Compared to placebo, treatment with a SSRI significantly reduced OCD symptoms (RR = 0.61, 95% CI:0.49,0.76) and severity (DSM-based measures: SMD = − 0.48, 95% CI:− 0.71,− 0.25; Children's Yale-Brown Obsessive Compulsive Scale [CY-BOCS]: WMD = − 3.78, 95% CI:− 5.01,− 2.55). Children and adolescents who were treated with a SNRI experienced an even greater reduction in OCD severity (DSM-based measures: MD = − 1.15, 95% CI:− 1.70, − 0.60; CY-BOCS: MD = − 8.90, 95% CI:− 12.73, − 5.07) (Table 2). Quality of life was significantly higher among those with anxiety who were treated with SSRI vs placebo (SMD = 0.57, 95% CI:0.35, 0.79); however, there was no significant difference in quality of life for children and adolescents suffering from OCD. Among children and adolescents with OCD, there was no significant difference in relapse among those treated with SSRI or placebo (RR = 0.78, 95% CI:0.55,1.11). Compared to placebo, significantly more children and adolescents randomized to medication experienced a drug-related adverse event (RR:1.91, 95% CI:1.2,3.05). Ipser et al. found little evidence of publication bias for treatment response or reduction in symptom severity52.

Table 2. Effect of drug-based treatments in treating childhood and adolescent anxiety disorders
OutcomeComparisonAdditional detailsNumber of studies (subjects)Intervention (n/N)Control (n/N)Measure of treatment effect (95% CI)Interpretation
  • n/N = number who had the event/total number who had the treatment

  • Excluding Obsessive–Compulsive Disorder; CY-BOCS: Children's Yale-Brown Obsessive Compulsive Scale; OCD: Obsessive-Compulsive Disorder; CGI-I: Clinical Global Impression-Improvement Scale; DSM: Diagnostic and Statistical Manual; SSRI: Selective Serotonin Reuptake Inhibitor; SNRI: Selective Norepinephrine Reuptake Inhibitor.

Improvement of anxiety symptomsSSRI vs placeboCGI-I7 (842)106/389295/453RR = 0.42 (0.35,0.50)Significantly favours SSRI
      I2 = 0% 
 SNRI vs placeboCGI-I2 (606)131/311183/295RR = 0.68 (0.8,0.88)Significantly favours SNRI
      I2 = 0% 
Anxiety severity reductionSSRI vs placeboDSM-based measures4 (770)N = 241N = 529SMD = − 0.58 (−1.54,0.39)No significant difference
      I2 = 95.6% 
Improvement of OCD symptomsSSRI vs placeboCGI-I5 (654)83/317152/337RR = 0.61 (0.49,0.76)Significantly favours SSRI
      I2 = 0% 
OCD severity reductionSSRI vs placeboDSM-based measures4 (322)N = 177N = 145SMD = − 0.48 (−0.71,− 0.25)Significantly favours SSRI
      I2 = 0% 
  CY-BOCS6 (705)N = 363N = 342WMD = − 3.78 (−5.01,− 2.55) 
      I2 = 0% 
 SNRI vs placeboDSM-based measures1(60)N = 31N = 29MD = − 1.15 (−1.70,− 0.60)Significantly favours SNRI
  CY-BOCS1 (60)N = 31N = 29MD = − 8.90 (−12.73,− 5.07) 

Behavioural and cognitive behavioural treatments

Two reviews examined the effects of individual- and/or group-based CBT/BT on symptomatology, remission, and treatment failure (Table 3)50, 51.

Table 3. Effect of BT/CBT alone in the treatment of childhood and adolescent anxiety disorders
OutcomeComparisonAdditional detailsNumber of studies (subjects)Intervention (n/N)Control (n/N)Measure of treatment effect (95% CI)Interpretation
  1. n/N = number who had the event/total number who had the treatment

  2. CBT: Cognitive Behavioural Therapy; BT: Behavioural Therapy; CY-BOCS: Children's Yale-Brown Obsessive-Compulsive Scale; NIMH-GOCS: National Institute of Mental Health Global Obsessive–Compulsive Score; OCD: Obsessive-Compulsive Disorder; RCMAS: Revised Children's Manifest Anxiety Scale; FSSC-R: Revised Fear Survey Schedule for Children; SSRI: selective serotonin reuptake inhibitor; TCA: Tricyclic Antidepressant.

Remission of anxietyCBT vs wait list or active controlAll formats of CBT12 (765)209/475208/290RR = 0.58 (0.50,0.67)Significantly favours CBT
      I2 = 32% 
  Individual CBT5 (290)78/170100/120RR = 0.56 (0.47,0.67) 
      I2 = 0% 
  Group CBT7 (314)70/162107/152RR = 0.61 (0.50,0.75) 
      I2 = 10% 
  Family CBT4 (187)40/12158/66RR = 0.38 (0.29,0.51) 
      I2 = 9% 
  All formats of CBT12 (668)147/415201/253RR = 0.43 (0.38,0.50) 
  Completer analysis   I2 = 0% 
Reduction in anxiety symptomsCBT vs wait list or active controlRCMAS10 (579)N = 375N = 204SMD = − 0.58 (−0.76,− 0.40)Significantly favours CBT
      I2 = 0% 
  RCMAS4 (260)N = 155N = 105SMD = − 0.20 (−0.46, 0.05)No significant difference
      I2 = 62% 
  FSSC-R10 (579)N = 375N = 204SMD = − 0.55 (−0.74,− 0.36)Significantly favours CBT
      I2 = 6% 
OCD treatment failureIndividual BT/CBT vs wait list or placebo 2 (104)20/5251/52RR = 0.31 (0.05,1.95)No significant difference
      I2 = 92.2% 
 Group BT/CBT vs wait list or placebo 1 (53)7/2924/24RR = 0.26 (0.14,0.48)Significantly favours group BT/CBT
 BT/CBT vs medication (SSRI or TCA) 2 (78)22/4028/38RR = 0.75 (0.54,1.05)No significant difference
      I2 = 0% 
Change in OCD severityIndividual BT/CBT vs wait list or placeboCY-BOCS2 (102)N = 50N = 52WMD = − 11.67 (−19.69,− 3.66)Significantly favours individual BT/CBT
      I2 = 6% 
  NIMH-GOCS1 (48)N = 24N = 24MD = − 5.50 (−6.72,− 4.28) 
 Group BT/CBT vs wait list or placeboCY-BOCS1 (53)N = 29N = 24MD = − 15.76 (−18.90,− 12.62)Significantly favours group BT/CBT
  NIMH-GOCS1 (48)N = 24N = 24MD = − 5.69 (−6.87,− 4.51) 
 BT/CBT vs medication (SSRI or TCA)CY-BOCS2 (78)N = 40N = 38WMD = − 3.87 (−8.15,0.41)No significant difference
      I2 = 25% 

There was significant recovery when CBT was administered in an individual, in a group, or in a family environment50. Among children and adolescents with an anxiety disorder (excluding OCD), treatment with CBT compared to wait list or active control (receipt of support or diary treatments) resulted in a significant reduction of fears as measured by the Revised Fear Survey Schedule for Children (FSSC-R) (SMD = − 0.55, 95% CI:− 0.74,− 0.36), but report of reduced anxiety symptomatology was mixed when measured with the Revised Children's Manifest Anxiety Scale (RCMAS) (Table 3). There was no significant difference in the loss to follow-up between the two groups (RR = 1.02, 95% CI:0.54,1.94). There was a significant increase in the remission in anxiety diagnosis for all CBT formats (individual-, group-based) (RR = 0.58; 95% CI:0.50,0.67). For every three children/adolescents (95% CI:2.5,4.5) who received CBT, one child/adolescent would recover.

A second review assessed BT/CBT among those diagnosed with OCD51. Compared to medication, there were no differences in treatment failures or symptom severity among those children and adolescents who received CBT/BT. There was also no significant difference in treatment failures among those who received individual CBT/BT versus wait list or placebo; however, there were notable clinical changes in symptomatology. Children and adolescents who received individual- or group-based CBT/BT had a significant improvement in symptom severity irrespective of the clinical measure used. Group-based CBT/BT, however, resulted in slightly more significant reductions compared to individual-based treatment: group-based (CY-BOCS: MD = − 15.76, 95% CI:− 18.90,− 12.62; National Institute of Mental Health Global Obsessive Compulsive Rating Scale [NIMH-GOCS]: MD = − 5.69, 95%CI:− 6.87,− 4.51), individual-based (CY-BOCS: WMD = − 11.67, 95% CI:− 19.69,− 3.66; NIMH-GOCS: MD = − 5.50, 95%CI:− 6.72,− 4.28). Children who received group CBT/BT had lower self-reported depression than those who were placed on a wait list (MD = − 0.75; 95% CI:− 1.34,− 0.17), but no similar finding was observed for those who were randomized to individual CBT/BT. Mixed findings were reported for treatment failures when comparing CBT/BT to wait list or placebo treatment (Table 3).

Combination treatments

Among children and adolescents with OCD, the use of CBT/BT alongside medication had mixed effect on treatment outcomes depending on the comparison group (Table 4). Compared to placebo and medication alone, use of CBT/BT with medication (SSRI) resulted in a significant reduced risk of OCD treatment failure (placebo comparison: RR = 0.48, 95% CI:0.32,0.78; medication alone comparison: RR = 0.59, 95% CI:0.38,0.92). There was no difference risk of treatment failure of symptom severity when comparing combined treatment with CBT/BT alone. Compared to placebo or medication alone, use of CBT/BT in conjunction with medication (SSRI) significantly reduced OCD symptom severity across multiple clinical measurement scales including the CY-BOCS (placebo comparison: MD = − 10.30, 95% CI:− 14.06,− 6.54; medication alone comparison: WMD = − 4.55, 95% CI:− 7.40,− 1.70).

Table 4. Effect of BT/CBT in combination with drug-based treatments in treating childhood and adolescent anxiety disorders
OutcomeComparisonAdditional detailsNumber of studies (subjects)Intervention (n/N)Control (n/N)Measure of treatment effect (95% CI)Interpretation
  1. BT: Behavioural Therapy; CBT: Cognitive Behavioural Therapy; CY-BOCS: Children's Yale-Brown Obsessive Compulsive Scale; CGI-I: Clinical Global Impression-Improvement Scale; CGI-S: Clinical Global Impression-Severity Scale; OCD: Obsessive-Compulsive Disorder; NIMH-GOCS: National Institute of Mental Health Global Obsessive-Compulsive Score; SSRI: selective serotonin reuptake inhibitor.

OCD treatment failureBT/CBT and medication (SSRI) vs placebo 1 (56)13/2827/28RR = 0.48 (0.32,0.72)Significantly favours BT/CBT and medication
 BT/CBT and medication (SSRI) vs medication 2 (66)18/3327/33RR = 0.59 (0.38,0.92) 
      I2 = 0% 
 BT/CBT and medication (SSRI) vs BT/CBT 1 (56)13/2817/28RR = 0.76 (0.47,1.2)No significant difference
Change in OCD severityBT/CBT and medication (SSRI) vs placeboCY-BOCS1 (54)N = 28N = 28MD = − 10.30 (−14.06,− 6.54)Significantly favours BT/CBT and medication
 BT/CBT and medication (SSRI) vs medication (SSRI)CY-BOCS2 (76)N = 38N = 38WMD = − 4.55 (−7.40,− 1.70)Significantly favours BT/CBT and medication
      I2 = 0% 
  CGI-I1 (20)N = 10N = 10MD = − 0.70 (−1.34,− 0.06) 
  CGI-S1 (20)N = 10N = 10MD = − 0.70 (−1.15,− 0.25) 
  NIMH-GOCS1 (20)N = 10N = 10MD = − 0.20 (−2.31, 1.91)No significant difference
 BT/CBT and medication (SSRI) vs BT/CBTCY-BOCS1 (54)N = 28N = 28MD = − 2.80 (−7.55,1.95)No significant difference


Summary of main results

Overall completeness and applicability of evidence

This Overview provides a synthesis of current evidence available from the CDSR for a range of pharmaco- and psychotherapeutic treatments for childhood and adolescent anxiety disorders, but highlights that (1) the most recent trial evidence in the CDSR exists for medication management, and (2) more contemporary studies examining the efficacy of psychotherapeutic programs and their combination with pharmacotherapeutics still need to be synthesized. These latter areas of clinical research are ones that are rapidly expanding and efforts must be made to ensure uptake into Cochrane reviews.

Despite the need for ongoing contributions to the CDSR, the findings in this Overview point to promising treatments to guide current clinical decision-making. Of the studies included in the CDSR50, 51, treatment with individual- or group-based CBT has been shown to lead to significant reductions in of anxiety symptoms and increased recovery. CBT and BT have also been shown to reduce OCD severity and risk of treatment failure. A recent update to the O'Kearney review51 maintained this conclusion. Recently published studies further reinforce these findings, but also examine specific active treatment techniques used in different clinical populations see e.g. 31, 48, 55–57 while other multi-site studies are emerging in this field38, 41.

The pharmacotherapy review included in this Overview41 demonstrated that use of SSRIs and the SNRI venlafaxine are superior to placebo in treating OCD and other anxiety disorders. Ipser's52 review pointed out there was no clear evidence that any particular SSRI or venlafaxine was most efficacious or best tolerated. This conclusion, while important to circumscribing the evidence base, does not help to optimize clinician decision-making26. Targeted investigations are needed to develop an understanding of the utility of specific pharmacologic agents in this population. These investigations should include combined treatment arms (medication + psychotherapy). While few studies were available in the CDSR, CBT combined with medication was shown to lead to significant reductions in both anxiety and OCD symptoms, and have mixed impact on reducing risk of treatment failure. O'Kearney's updated review51 recommended that further studies be conducted to clarify questions about the relative efficacy of BT/CBT against medication and when combined with medication.

On a final note, report of outcomes important for re-establishing a normative course of emotional and cognitive development including functional capacity, adherence/persistence, and acceptability were either absent from studies in the CDSR (functional capacity, adherence/persistence) or varied in measurement (acceptability defined as ‘lost to follow-up’, quality of life, drug-related adverse events). The role of therapeutic relationship factors was also absent in studies in the CDSR, but recently discussed in the literature47. This absence and the heterogeneity across individual trials in definition and measurement reduces the ability of this overview to draw definitive conclusions on treatment effect. The trend of significance across many outcome measures, however, provides confidence in the utility of current treatments. One notable exception is combined treatments (medication plus psychotherapy). The small number of available trials and limited sample sizes leaves this approach to treatment in need of further investigation to provide high-quality evidence, particularly as this approach is most common in clinical practice.

Quality of the evidence

The reviews in this overview reported modest quality for the studies evaluating interventions for children and adolescents with anxiety disorders. While one review excluded low quality trials50, methodological quality of studies in the other reviews was not reported. These reviews reported concerns with studies' methodological rigor including high risk of bias from a lack of adequate allocation of concealment. Further, variability in sampling, nature and duration of interventions, comparison groups, and outcome measures across studies clearly tempers the conclusions that can be drawn. Several reviews did report significant statistical heterogeneity using I2 values indicating that the studies may have been too different to pool results (see Tables 2 and 3). On a final note, this Overview's synthesis should be interpreted with the understanding that any methodologic (including analytic) limitations across the three reviews cannot be resolved. To reduce the risk of perpetuating conclusions from flawed research, we have therefore limited our definitive conclusions to broad findings that have emerged in a preponderance of studies across time and across various sites. The use of individual patient data in subsequent Cochrane review updates would be ideal to compare contemporary pooled results to past analyses.


Implications for research

The methodological variability in sampling, nature and duration of intervention, comparison groups, statistical heterogeneity between the studies, and outcome measures used and how they were defined in the reviewed literature clearly limits the definitive conclusions that can be drawn from this Overview. This variability does, however, point to methodological areas of focus for both individual trials and subsequent Cochrane reviews. Further, this Overview highlights the need to incorporate more recent clinical trial data in the CDSR and to replicate recently conducted studies to mature the knowledge base51. Nevertheless, the cumulative evidence for BT, CBT, SSRI's and venlafaxine is substantial, suggesting that further studies of their benefits (alone or in combination) for specific populations within childhood anxiety disorders would be worthwhile. Such studies would be enhanced by the use of the same ‘gold standard’ outcome measures, for ease of interpretation and to allow comparisons across sites. Most OCD studies, for example, have used the CY-BOCS but outcome measures used for non-OCD anxiety studies are quite variable.

One approach to ensuring large samples and methodological consistency in such studies is exemplified by the Research Units of Psychopharmacology (RUPP) group in the United States30. This group has a network of multiple, geographically diverse sites that all have investigators experienced in psychopharmacology trials and have agreed to work together using common protocols across sites. For large trials of well-established treatments or for studies combining psychological and pharmacological treatment (that require several treatment arms and therefore large samples), the RUPP approach appears ideal.

Not every type of childhood anxiety disorder or anxiety-related condition has well-established treatments. For example, literature on treating children with selective mutism (a condition often associated with social phobia) consists largely of case reports58. Children with comorbid anxiety and depression or comorbid anxiety and externalizing disorders may require treatment tailored to their complex presentations. Such treatments are being developed59, but await evaluation. Treatments (especially psychological treatments) evaluated in North America may also require adaptation for other countries and cultures. Thus, there will continue to be a role for treatment development and for preliminary studies of efficacy, in addition to large, multi-site trials. Studies to determine the most crucial components of treatment (whether related to treatment content or to process/therapeutic relationship elements) are also needed to improve outcomes and reduce rates of treatment failure.

Implications for practice

The treatments reviewed have demonstrated efficacy for the treatment of childhood anxiety disorders, including OCD. Efficacy in research centers does not always translate into effectiveness in community practice, however. In contrast to academic practice, community practice is often characterized by more complex presentations, less training and supervision of therapists, and organizational barriers to successful treatment37. The fewer of these problems that exist in a given community case, the greater the chances of treatment success. Careful assessment of children and adolescents with anxiety disorders is warranted, not only to confirm diagnosis but to assess the likelihood that the child or teen can benefit from the evidence-based treatment proposed. Community practitioners must also be wary of claims by pharmaceutical firms that their product is the best or the only helpful treatment for childhood anxiety or OCD. As reviewed in Table 2, there is evidence of efficacy relative to placebo for a number of SSRI and SNRI medications. When evaluating the risks and benefits of a given treatment for a child, it is also prudent to consult expert consensus guidelines. For example, our review suggests that many children with anxiety or OCD can benefit from pharmacologic treatments (SSRI's, SNRI's), psychological treatments (BT or CBT), or a combination of these. Consensus guidelines, however, generally advocate offering psychological treatment first to mildly affected children, and reserving pharmacological treatment (which has a higher potential for adverse events) for non-responsive or severely impaired children (cite CPA anxiety guidelines).

Even when a child or youth is well-suited to a given treatment, the treatment may be difficult to obtain in community settings. This is a particularly common problem with psychological treatments such as CBT. In addition to advocating for more therapists (see below), it is worth placing suitable children on wait lists for psychotherapy and encouraging families to take some simple steps to begin to address the child's anxiety. For example, parents can encourage the child with phobias to face some feared situations, or encourage exposure and response prevention for situations where the child already has limited control of his or her compulsions. Books that can guide this process include Helping Your Anxious Child: A Step-By-Step Guide for Parents60, Keys to Parenting Your Anxious Child, 2nd edition61 and OCD in Children and Adolescents: A Cognitive-Behavioral Treatment Manual62.

Finally, the treatments reviewed above were shown to benefit children and adolescents with anxiety disorders, including OCD, but the benefits were somewhat variable. For example, some studies found BT or CBT alleviated OCD severity and reduced rates of treatment failure relative to control groups, but others did not (see Table 3). Therefore, the clinician should caution children, youths, and families to not expect complete resolution of the anxiety- or OCD-related problems with treatment particularly for OCD where treatment response may consist of a modest decline in symptoms rather than complete remission. It is important to encourage families to persevere with treatment despite limited gains, and to focus on evidence of success, even if it occurs in small steps.

Implications for policy-makers

Given the high prevalence and burden of suffering associated with childhood and adolescent anxiety disorders, access to evidence-based treatments is essential for all children and youths who could benefit. Access to medical practitioners increases the chances of obtaining an SSRI or SNRI for childhood anxiety, and speaks to the need for universal health insurance. Access to non-medical personnel trained in BT or CBT, however, is often difficult or expensive to obtain even in jurisdictions where medical care is covered. Advocacy for training more therapists in these evidence-based treatments, particularly in publicly funded mental health centers, is clearly needed.

A further barrier to obtaining anxiety-focused medications for children and adolescents is the public perception that SSRIs or SNRIs are dangerous, especially in children. The proposed link between SSRI/SNRI use and suicide in young people, spurred by a US FDA review that suggested an increased risk of suicidal events (defined as emergent/worsening suicidal thoughts and/or behaviours) associated with SSRI/SNRI use63, 64, remains an area of debate65, 66. While the debate continues, recent findings suggest that the benefits of these antidepressants can outweigh the risks32, especially for children with non-OCD anxiety disorders. Education of both physicians and the general public around the risks and benefits of these medications in childhood anxiety disorders would be helpful in ensuring that they are provided to children who could benefit. Further studies of safety and efficacy specific to children would also contribute to this goal.


1 Post-Traumatic Stress Disorder (PTSD), Acute Stress Disorder, (S)elective Mutism, and Adjustment Disorder with anxiety are also categorized in this diagnostic group, but will not be covered in this overview.


Dr Newton is salary supported by a four-year career award from the Canadian Child Health Clinician Scientist Program, in partnership with SickKids Foundation, Child & Family Research Institute (British Columbia), Women & Children's Health Research Institute (Alberta), Manitoba Institute of Child Health. Kelly Russell holds an Alberta Heritage Foundation for Medical Research doctoral student award.

Contributions of Authors

All authors contributed to this review. KR extracted the data and contributed to writing the Methods and Results sections. AN contributed to writing the Introduction and Discussion sections. KM contributed to writing the Introduction and Conclusions sections. KM is the primary author of this report. All authors contributed to editing all sections of the overview and take responsibility for the manuscript.

Declarations of Interest

Dr Manassis receives author's royalties from the books ‘Keys to Parenting Your Anxious Child, 2nd Edition’ (Woodbine House, 2008) and ‘Cognitive Behavioral Therapy with Children: A Guide for the Community Practitioner’ (Routledge, 2009).