Video Abstract

Video Abstract

Close modal
CONTEXT:

In several studies, authors have reported on cognitive behavioral therapy (CBT) for children and adolescents with autism spectrum disorders (ASDs), but inconsistent treatment effectiveness was revealed from these studies.

OBJECTIVE:

To evaluate the effectiveness of CBT on the symptoms of ASD and social-emotional problems in children or adolescents with ASD by using a meta-analytic approach.

DATA SOURCES:

Data sources included PubMed, Embase, and the Cochrane Library.

STUDY SELECTION:

We selected randomized controlled trials (RCTs) in which authors reported effectiveness of CBT on the symptoms of ASD and social-emotional problems in children or adolescents with ASD from database inception to May 2019.

DATA EXTRACTION:

For each study, 2 authors extracted data on the first author’s surname, publication year, country, sample size, mean age, CBT target, intervention, outcome measurement, follow-up duration, and investigated outcomes.

RESULTS:

Forty-five RCTs and 6 quasi RCTs of 2485 children and adolescents with ASDs were selected for the final meta-analysis. There was no significant difference between CBT and control for symptoms related to ASD based on self-reported outcomes (standard mean difference: −0.09; 95% confidence interval: −0.42 to 0.24; P = .593), whereas CBT significantly improved the symptoms related to ASD based on informant-reported outcomes, clinician-rated outcomes, and task-based outcomes. Moreover, the pooled standard mean differences indicated that CBT has no significant effect on symptoms of social-emotional problems based on self-reported outcomes.

LIMITATIONS:

The quality of included studies was low to modest, significant heterogeneity among the included studies for all investigated outcomes was detected, and publication bias was inevitable.

CONCLUSIONS:

These findings indicate that CBT may significantly improve the symptoms of ASD and social-emotional problems in children or adolescents with ASD.

Autism spectrum disorder (ASD) is the collective term for neurodevelopmental disorders and is characterized by core deficits of social communication and social interactions, atypical repetitive behaviors, and restricted interests.1,2  Currently, the prevalence of ASD in high-income countries has nearly reached 1% because of its association with greater risk of peer conflicts and societal exclusion.35  Moreover, the comorbid condition of childhood anxiety in children with ASD reached up to 80%, and 40% of children met the criteria of anxiety disorder.6,7  Several studies have already indicated that ASD with comorbidity of anxiety contributed a deleterious impact on the lives of children and their families, affecting the negative consequences and difficulties of the disorder.8,9 

Currently, cognitive behavioral therapy (CBT) is widely used in young people with ASD on the basis of their behavioral manifestations of ASD,10  providing an opportunity for patients with ASD to learn the skills to challenge dysfunctional beliefs and replace them with more adaptive and positive thinking.11  CBT could make patients better adapt to environments or contexts and improve their ability to reflect on thoughts and feeling for environments or contexts causing anxiety. Authors of several systematic reviews and meta-analyses have already evaluated the effects of CBT on patients with ASD. Spreckley and Boyd12  recruited 13 studies and revealed that applied behavioral intervention significantly improved the symptoms of children with autism. According to the study conducted by Sukhodolsky et al6  on anxiety in children with high-functioning autism, children with high-functioning ASD who received CBT showed a significant improvement in parent and clinician ratings of anxiety when compared with children on the waiting list or those under treatment-as-usual control conditions. Similarly, Ung et al13  found that CBT demonstrated robust efficacy in reducing anxiety symptoms in youth with high-functioning ASD. However, these meta-analyses studies have been focused on specific symptom of ASD or anxiety disorder in children or adolescents, whereas social-emotional problems (depression alongside anxiety disorders) were not reported in previous studies. Therefore, the current meta-analysis was conducted on the basis of published randomized controlled trials (RCTs) to evaluate the effectiveness of CBT on the symptoms of ASD and social-emotional problems in children or adolescents with ASD.

The current systematic review and meta-analysis study was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis statement issued in 2009.14  Studies designed as RCTs and those that evaluated the effectiveness of CBT on the symptoms of ASD and social-emotional problems in children or adolescents with ASD were considered eligible for inclusion in this study. There were no restrictions placed on the published language and status of the study. Electronic databases such as PubMed, Embase, and the Cochrane Library were systematically searched for studies that were published from their inception until May 2019. The core search terms included (“CBT” OR “Cognitive behavioural therapy”) AND (“autism spectrum disorder” OR “ASD”) AND “randomized controlled trials” AND “child (birth to 18 years).” The details of search strategy are presented in the Supplemental Information. The reference lists obtained from the eligible studies were manually searched for any new eligible study.

Two authors independently searched the literature, and a standardized approach was used for study selection. Any conflicts between these authors were resolved by group discussion. A study was included if it met the following inclusion criteria: (1) patients: patients diagnosed with ASD and <18.0 years of age; (2) intervention: CBT (contained both cognitive and behavioral components irrespective of whether they were clearly described); (3) outcomes: symptoms of ASD based on self-reported outcomes, informant-reported outcomes, clinician-rated outcomes, and task-based outcomes and social-emotional problems based on self-reported outcomes, informant-reported outcomes, and clinician-rated outcomes; and (4) study design: RCT or quasi RCT. Observational studies were excluded because of uncontrolled confounding factors.

The data abstraction and quality assessment were conducted by 2 authors independently, and any disagreements were settled by an additional author referring to the original article. The collected items included the first authors’ surname, publication year, country, sample size, mean age, CBT target, intervention, outcome measurement, follow-up duration, and investigated outcomes. The effect estimates were maximally adjusted and were abstracted if the study reported multiple outcome measures. Study quality was assessed by using the JADAD scale, which is based on randomization, blinding, allocation concealment, withdrawals and dropouts, and the use of an intention-to-treat analysis, with scores ranging from 0 to 5.15  In this study, a study that scored 4 or 5 was regarded as high quality.

Treatment effectiveness between CBT and control groups regarding the symptoms of ASD and social-emotional problems in children or adolescents with ASD was calculated on the basis of mean, SD, and sample size in each group in each individual trial. Moreover, the standard mean difference (SMD), with corresponding 95% confidence intervals (CIs), was calculated by using the random-effects model.16,17  Heterogeneity among the included studies was assessed by using I2 and P values for the Q statistic, and I2 > 50.0% or P < .10 was considered to be significant.18,19  The stability of pooled results was assessed by using sensitivity analyses through sequential exclusion of each study.20  Subgroup analyses for investigated outcomes were conducted on the basis of mean age, intervention, CBT target, and study quality, and P values between subgroups were also calculated.21  Publication biases were assessed by using funnel plots, Egger’s test,22  and Begg’s test,23  and the results, adjusted by the trim-and-fill method, were provided if any potential publication bias was observed.24  All reported P values are 2 sided, and P values <.05 were considered to be statistically significant for all the included studies. Statistical analyses were performed by using Stata software (version 10.0; Stata Corp, College Station, TX).

The electronic searches from PubMed, Embase, and the Cochrane Library yielded 2347 records. Of these, 546 were excluded because of duplications, and 1722 studies were excluded because of irrelevant topics. The remaining 79 studies were selected for further detailed evaluations; 45 RCTs and 6 quasi RCTs were selected for the final quantitative meta-analysis. Manual search of the reference lists yielded no new studies. A detailed study selection process is presented in Fig 1, and the baseline characteristics of the included studies are summarized in Supplemental Table 2.

FIGURE 1

Flow diagram of literature search and study selection process.

FIGURE 1

Flow diagram of literature search and study selection process.

Close modal

Forty-five RCTs and 6 quasi RCTs involving a total of 2485 children and adolescents with ASD were included in this meta-analysis. The mean follow-up duration for patients was 2.0 weeks to 6.0 months, and 9 to 196 patients were included in each trial. CBT for anxiety was reported in 18 trials, whereas other CBTs were reported in 33 trials. Forty-six trials were conducted in Western countries, and the remaining 5 trials were conducted in Eastern countries. Study quality was assessed by using the JADAD scale, on which 24 trials scored 4, 16 trials scored 3, 6 trials scored 2, and the remaining 5 trials scored 1.

Data regarding the effect of CBT on symptoms related to ASD based on self-reported outcomes were available in 10 trials. There was no significant difference between CBT and control groups regarding symptoms related to ASD based on self-reported outcomes (SMD: −0.09; 95% CI: −0.42 to 0.24; P = .593; Fig 2), and the included trials showed significant heterogeneity. This pooled conclusion was unchanged by sequential exclusion of any individual trial (Supplemental Fig 9). Subgroup analysis indicated no significant differences between CBT and control groups regarding the symptoms related to ASD based on self-reported outcomes in all subsets (Table 1). No significant publication bias for symptoms related to ASD based on self-reported outcomes was detected (P value for Egger’s test = .810; P value for Begg’s test = .996; Supplemental Fig 16).

FIGURE 2

Forest plot for the effectiveness of CBT on symptoms related to ASD based on self-reported outcomes.

FIGURE 2

Forest plot for the effectiveness of CBT on symptoms related to ASD based on self-reported outcomes.

Close modal
TABLE 1

Subgroup Analyses for Investigated Outcomes

Outcomes and SubgroupsNo. Included StudiesSMD (95% CI)PHeterogeneity, %P for HeterogeneityP Between Subgroups
Symptoms related to ASD based on self-reported outcomes       
 Mean age, y      .479 
  ≥10.0 −0.14 (−0.38 to 0.10) .264 27.6 .208 — 
  <10.0 0.17 (−1.51 to 1.86) .840 94.8 <.001 — 
 Intervention      .696 
  Group based −0.10 (−0.46 to 0.26) .581 72.5 <.001 — 
  Individual based 0.03 (−0.64 to 0.69) .940 — — — 
 Target       
  Anxiety — — — — — 
  Other 10 −0.09 (−0.42 to 0.24) .593 69.3 .001 — 
 Study quality      .479 
  High 0.17 (−1.51 to 1.86) .840 94.8 <.001 — 
  Low −0.14 (−0.38 to 0.10) .264 27.6 .208 — 
Symptoms related to ASD based on informant-reported outcomes       
 Mean age, y      .007 
  ≥10.0 11 −0.76 (−1.30 to -0.22) .006 92.0 <.001 — 
  <10.0 12 −0.40 (−0.79 to −0.02) .039 77.1 <.001 — 
 Intervention      .815 
  Group based 21 −0.56 (−0.91 to −0.21) .002 88.2 <.001 — 
  Individual based −0.67 (−2.19 to 0.84) .384 90.8 .001 — 
 Target       
  Anxiety — — — — — 
  Other 23 −0.57 (−0.90 to −0.24) .001 87.8 <.001 — 
 Study quality      .001 
  High 10 −0.69 (−1.30 to −0.09) .024 93.9 <.001 — 
  Low 13 −0.45 (−0.70 to −0.20) <.001 46.6 .032 — 
Symptoms related to ASD based on clinician-rated outcomes       
 Mean age, y      .136 
  ≥10.0 0.44 (−0.02 to 0.90) .059 0.0 .343 — 
  <10.0 1.26 (−0.07 to 2.59) .064 65.5 .055 — 
 Intervention      .610 
  Group based 0.87 (−0.17 to 1.91) .101 63.6 .064 — 
  Individual based 0.73 (−0.57 to 2.02) .270 68.4 .075 — 
 Target       
  Anxiety — — — — — 
  Other 0.75 (0.10 to 1.41) .024 55.1 .063 — 
 Study quality      .052 
  High 0.15 (−0.47 to 0.76) .643 0.0 .899 — 
  Low 1.31 (0.27 to 2.35) .014 61.0 .077 — 
Symptoms related to ASD based on task-based outcomes       
 Mean age, y      .237 
  ≥10.0 −0.48 (−1.01 to 0.05) .077 79.8 <.001 — 
  <10.0 −0.36 (−0.65 to −0.07) .014 9.3 .347 — 
 Intervention      .110 
  Group based 10 −0.45 (−0.81 to −0.10) .012 71.7 <.001 — 
  Individual based −0.07 (−0.63 to 0.49) .808 — — — 
 Target       
  Anxiety — — — — — 
  Other 11 −0.41 (−0.75 to −0.08) .014 70.9 <.001 — 
 Study quality      <.001 
  High −0.69 (−1.09 to −0.29) .001 69.4 .011 — 
  Low −0.01 (−0.30 to 0.27) .932 0.0 .442 — 
Symptoms of social-emotional problems based on self-reported outcomes       
 Mean age, y      .255 
  ≥10.0 −0.48 (−1.02 to 0.07) .086 80.6 <.001 — 
  <10.0 0.03 (−0.64 to 0.70) .935 — — — 
 Intervention      .069 
  Group based −0.56 (−1.17 to 0.05) .071 82.4 <.001 — 
  Individual based 0.06 (−0.42 to 0.54) .816 0.0 .902 — 
 Target       
  Anxiety −0.42 (−0.90 to 0.07) .093 78.6 <.001 — 
  Other — — — — — 
 Study quality      <.001 
  High −0.10 (−0.35 to 0.15) .445 0.0 .961 — 
  Low −1.16 (−2.67 to 0.36) .136 90.6 <.001 — 
Symptoms of social-emotional problems on informant-reported outcomes       
 Mean age, y      .391 
  ≥10.0 13 −0.72 (−1.16 to −0.29) .001 80.0 <.001 — 
  <10.0 −0.71 (−1.24 to −0.17) .009 73.3 .002 — 
 Intervention      .481 
  Group based 12 −0.83 (−1.32 to −0.34) .001 82.4 <.001 — 
  Individual based −0.54 (−0.95 to −0.14) .008 63.5 .012 — 
 Target      .461 
  Anxiety 17 −0.71 (−1.09 to −0.34) <.001 79.1 <.001 — 
  Other −0.74 (−1.34 to −0.13) .017 57.5 .125 — 
 Study quality      .895 
  High 10 −0.62 (−0.96 to −0.27) <.001 64.5 .003 — 
  Low −0.86 (−1.50 to −0.21) .009 85.2 <.001 — 
Symptoms of social-emotional problems on clinician-rated outcomes       
 Mean age, y      .018 
  ≥10.0 0.80 (0.40 to 1.19) <.001 49.4 .065 — 
  <10.0 1.64 (0.10 to 3.19) .037 87.5 .005 — 
 Intervention      .335 
  Group based 0.94 (0.30 to 1.57) .004 60.7 .054 — 
  Individual based 1.07 (0.40 to 1.73) .002 76.4 .002 — 
 Target       
  Anxiety 1.02 (0.58 to 1.46) <.001 68.6 .001 — 
  Other — — — — — 
 Study quality      .802 
  High 0.97 (0.44 to 1.50) <.001 73.4 .002 — 
  Low 1.24 (0.18 to 2.29) .022 69.8 .036 — 
Outcomes and SubgroupsNo. Included StudiesSMD (95% CI)PHeterogeneity, %P for HeterogeneityP Between Subgroups
Symptoms related to ASD based on self-reported outcomes       
 Mean age, y      .479 
  ≥10.0 −0.14 (−0.38 to 0.10) .264 27.6 .208 — 
  <10.0 0.17 (−1.51 to 1.86) .840 94.8 <.001 — 
 Intervention      .696 
  Group based −0.10 (−0.46 to 0.26) .581 72.5 <.001 — 
  Individual based 0.03 (−0.64 to 0.69) .940 — — — 
 Target       
  Anxiety — — — — — 
  Other 10 −0.09 (−0.42 to 0.24) .593 69.3 .001 — 
 Study quality      .479 
  High 0.17 (−1.51 to 1.86) .840 94.8 <.001 — 
  Low −0.14 (−0.38 to 0.10) .264 27.6 .208 — 
Symptoms related to ASD based on informant-reported outcomes       
 Mean age, y      .007 
  ≥10.0 11 −0.76 (−1.30 to -0.22) .006 92.0 <.001 — 
  <10.0 12 −0.40 (−0.79 to −0.02) .039 77.1 <.001 — 
 Intervention      .815 
  Group based 21 −0.56 (−0.91 to −0.21) .002 88.2 <.001 — 
  Individual based −0.67 (−2.19 to 0.84) .384 90.8 .001 — 
 Target       
  Anxiety — — — — — 
  Other 23 −0.57 (−0.90 to −0.24) .001 87.8 <.001 — 
 Study quality      .001 
  High 10 −0.69 (−1.30 to −0.09) .024 93.9 <.001 — 
  Low 13 −0.45 (−0.70 to −0.20) <.001 46.6 .032 — 
Symptoms related to ASD based on clinician-rated outcomes       
 Mean age, y      .136 
  ≥10.0 0.44 (−0.02 to 0.90) .059 0.0 .343 — 
  <10.0 1.26 (−0.07 to 2.59) .064 65.5 .055 — 
 Intervention      .610 
  Group based 0.87 (−0.17 to 1.91) .101 63.6 .064 — 
  Individual based 0.73 (−0.57 to 2.02) .270 68.4 .075 — 
 Target       
  Anxiety — — — — — 
  Other 0.75 (0.10 to 1.41) .024 55.1 .063 — 
 Study quality      .052 
  High 0.15 (−0.47 to 0.76) .643 0.0 .899 — 
  Low 1.31 (0.27 to 2.35) .014 61.0 .077 — 
Symptoms related to ASD based on task-based outcomes       
 Mean age, y      .237 
  ≥10.0 −0.48 (−1.01 to 0.05) .077 79.8 <.001 — 
  <10.0 −0.36 (−0.65 to −0.07) .014 9.3 .347 — 
 Intervention      .110 
  Group based 10 −0.45 (−0.81 to −0.10) .012 71.7 <.001 — 
  Individual based −0.07 (−0.63 to 0.49) .808 — — — 
 Target       
  Anxiety — — — — — 
  Other 11 −0.41 (−0.75 to −0.08) .014 70.9 <.001 — 
 Study quality      <.001 
  High −0.69 (−1.09 to −0.29) .001 69.4 .011 — 
  Low −0.01 (−0.30 to 0.27) .932 0.0 .442 — 
Symptoms of social-emotional problems based on self-reported outcomes       
 Mean age, y      .255 
  ≥10.0 −0.48 (−1.02 to 0.07) .086 80.6 <.001 — 
  <10.0 0.03 (−0.64 to 0.70) .935 — — — 
 Intervention      .069 
  Group based −0.56 (−1.17 to 0.05) .071 82.4 <.001 — 
  Individual based 0.06 (−0.42 to 0.54) .816 0.0 .902 — 
 Target       
  Anxiety −0.42 (−0.90 to 0.07) .093 78.6 <.001 — 
  Other — — — — — 
 Study quality      <.001 
  High −0.10 (−0.35 to 0.15) .445 0.0 .961 — 
  Low −1.16 (−2.67 to 0.36) .136 90.6 <.001 — 
Symptoms of social-emotional problems on informant-reported outcomes       
 Mean age, y      .391 
  ≥10.0 13 −0.72 (−1.16 to −0.29) .001 80.0 <.001 — 
  <10.0 −0.71 (−1.24 to −0.17) .009 73.3 .002 — 
 Intervention      .481 
  Group based 12 −0.83 (−1.32 to −0.34) .001 82.4 <.001 — 
  Individual based −0.54 (−0.95 to −0.14) .008 63.5 .012 — 
 Target      .461 
  Anxiety 17 −0.71 (−1.09 to −0.34) <.001 79.1 <.001 — 
  Other −0.74 (−1.34 to −0.13) .017 57.5 .125 — 
 Study quality      .895 
  High 10 −0.62 (−0.96 to −0.27) <.001 64.5 .003 — 
  Low −0.86 (−1.50 to −0.21) .009 85.2 <.001 — 
Symptoms of social-emotional problems on clinician-rated outcomes       
 Mean age, y      .018 
  ≥10.0 0.80 (0.40 to 1.19) <.001 49.4 .065 — 
  <10.0 1.64 (0.10 to 3.19) .037 87.5 .005 — 
 Intervention      .335 
  Group based 0.94 (0.30 to 1.57) .004 60.7 .054 — 
  Individual based 1.07 (0.40 to 1.73) .002 76.4 .002 — 
 Target       
  Anxiety 1.02 (0.58 to 1.46) <.001 68.6 .001 — 
  Other — — — — — 
 Study quality      .802 
  High 0.97 (0.44 to 1.50) <.001 73.4 .002 — 
  Low 1.24 (0.18 to 2.29) .022 69.8 .036 — 

—, not applicable.

Data regarding the effect of CBT on symptoms related to ASD based on informant-reported outcomes were available in 23 trials. The results revealed that individuals who received CBT showed improvement in symptoms related to ASD based on informant-reported outcomes (SMD: −0.57; 95% CI: −0.90 to −0.24; P = .001; Fig 3), and substantial heterogeneity was detected across the included trials. The results of sensitivity analysis indicated that this conclusion was stable and unaltered by sequential exclusion of each trial (Supplemental Fig 10). Although significant differences between CBT and control groups for symptoms related to ASD based on informant-reported outcomes were observed in most of the subsets, there was no significant difference when patients received an individual-based intervention (Table 1). There was no significant publication bias detected for symptoms related to ASD based on informant-reported outcomes (P value for Egger’s test = .350; P value for Begg’s test = .635; Supplemental Fig 17).

FIGURE 3

Forest plot for the effectiveness of CBT on symptoms related to ASD based on informant-reported outcomes.

FIGURE 3

Forest plot for the effectiveness of CBT on symptoms related to ASD based on informant-reported outcomes.

Close modal

Data regarding the effect of CBT on symptoms related to ASD based on clinician-rated outcomes were available in 5 trials. These studies revealed that CBT was associated with high symptoms of ASD based on clinician-rated outcomes (SMD: 0.75; 95% CI: 0.10 to 1.41; P = .024; Fig 4), and significant heterogeneity was observed among the included trials. This conclusion varied because of a marginal 95% CI and a smaller number of included trials (Supplemental Fig 11). A subgroup analysis indicated that CBT could improve symptoms related to ASD based on clinician-rated outcomes when the CBT target was other than anxiety and when studies were of low quality (Table 1). No significant publication bias for symptoms related to ASD based on clinician-rated outcomes was detected (P value for Egger’s test = .272; P value for Begg’s test = .221; Supplemental Fig 18).

FIGURE 4

Forest plot for the effectiveness of CBT on symptoms related to ASD based on clinician-rated outcomes.

FIGURE 4

Forest plot for the effectiveness of CBT on symptoms related to ASD based on clinician-rated outcomes.

Close modal

Data regarding the effect of CBT on symptoms related to ASD based on task-based outcomes were available in 11 trials. The summary SMD indicated that CBT was associated with lower symptoms of ASD based on task-based outcomes (SMD: −0.41; 95% CI: −0.75 to −0.08; P = .014; Fig 5), and significant heterogeneity was observed across the included trials. A sensitivity analysis indicated that the pooled conclusion varied because of a marginal 95% CI (Supplemental Fig 12). A subgroup analysis indicated significant differences in the mean age of patients <10.0 years, in patients who received a group-based intervention, when the CBT target was other than anxiety, and in studies of high quality (Table 1). There was no significant publication bias for symptoms related to ASD based on task-based outcomes (P value for Egger’s test = .427; P value for Begg’s test = .876; Supplemental Fig 19).

FIGURE 5

Forest plot for the effectiveness of CBT on symptoms related to ASD based on task-based outcomes.

FIGURE 5

Forest plot for the effectiveness of CBT on symptoms related to ASD based on task-based outcomes.

Close modal

Data regarding the effect of CBT on symptoms of social-emotional problems based on self-reported outcomes were available in 9 trials. There was no significant difference between CBT and control groups for symptoms of social-emotional problems based on self-reported outcomes (SMD: −0.42; 95% CI: −0.90 to 0.07; P = .093; Fig 6), and significant heterogeneity was observed across the included trials. This conclusion was stable and unaltered by exclusion of any particular trial (Supplemental Fig 13). The results of subgroup analyses were consistent with the overall analysis in all subsets (Table 1). No significant publication bias for symptoms of social-emotional problems based on self-reported outcomes was observed (P value for Egger’s test = .215; P value for Begg’s test = .175; Supplemental Fig 20).

FIGURE 6

Forest plot for the effectiveness of CBT on symptoms of social-emotional problems based on self-reported outcomes.

FIGURE 6

Forest plot for the effectiveness of CBT on symptoms of social-emotional problems based on self-reported outcomes.

Close modal

Data regarding the effect of CBT on symptoms of social-emotional problems based on informant-reported outcomes were available in 19 trials. The results revealed that CBT was associated with lower symptoms of social-emotional problems based on informant-reported outcomes (SMD: −0.71; 95% CI: −1.04 to −0.38; P < .001; Fig 7), and significant heterogeneity was observed among the included trials. The results of sensitivity analyses indicated that the pooled conclusion was stable and unchanged by sequential exclusion of each trial (Supplemental Fig 14). Subgroup analyses indicated significant differences in all subsets (Table 1). No significant publication bias was detected for the symptoms of social-emotional problems based on informant-reported outcomes (P value for Egger’s test = .095; P value for Begg’s test = .234; Supplemental Fig 21).

FIGURE 7

Forest plot for the effectiveness of CBT on symptoms of social-emotional problems based on informant-reported outcomes.

FIGURE 7

Forest plot for the effectiveness of CBT on symptoms of social-emotional problems based on informant-reported outcomes.

Close modal

Data regarding the effect of CBT on symptoms of social-emotional problems based on clinician-rated outcomes were available in 9 trials. The patients who received CBT showed a significant improvement in symptoms of social-emotional problems based on clinician-rated outcomes (SMD: 1.02; 95% CI: 0.58 to 1.46; P < .001; Fig 8), and significant heterogeneity was observed among the included trials. The results of sensitivity analyses indicated that the conclusions regarding the symptoms of social-emotional problems based on clinician-rated outcomes remained unchanged by exclusion of any particular trial (Supplemental Fig 15). Subgroup analyses revealed significant differences in all subsets (Table 1). A significant publication bias was detected for symptoms of social-emotional problems based on clinician-rated outcomes (P value for Egger’s test = .014; P value for Begg’s test = .048; Supplemental Fig 22). The conclusions were unchanged after adjustment for publication bias by using the trim-and-fill method.

FIGURE 8

Forest plot for the effectiveness of CBT on symptoms of social-emotional problems based on clinician-rated outcomes.

FIGURE 8

Forest plot for the effectiveness of CBT on symptoms of social-emotional problems based on clinician-rated outcomes.

Close modal

In our current quantitative meta-analysis of RCTs, we evaluated the effectiveness of CBT on symptoms of ASD and social-emotional problems in children or adolescents with ASDs. This meta-analysis included 2485 children and adolescents with ASD with a wide range of individual characteristics from 45 RCTs and 6 quasi RCTs. The results revealed that CBT improved the symptoms of ASD and social-emotional problems based on informant-reported and clinician-rated outcomes, whereas CBT had no significant effect on the above outcomes based on self-reported outcomes. Moreover, the symptoms of ASD based on task-based outcomes were significantly improved in children or adolescents with ASD who received CBT. Furthermore, the treatment effectiveness of CBT might differ when stratified by mean age and study quality.

A previous comprehensive systematic review and meta-analysis conducted by Weston et al,25  with 48 RCTs, revealed that CBT had a small to medium effect size for comorbid social-emotional problems related to ASD, whereas the effect size varied among the outcomes data from self-reported, informant-reported, clinician-reported, and task-based measures. However, this study incorporated both children and adults, and the stratified analyses were conducted only by targeting CBT. Moreover, numerous studies have already been completed, requiring further reevaluation to update the treatment effectiveness of CBT in children or adolescents with ASD. Therefore, the current meta-analysis was conducted on the basis of published RCTs to evaluate the effectiveness of CBT on ASD and social-emotional problems in children or adolescents with ASD.

The summary results regarding the symptoms of ASD based on self-reported, informant-reported, clinician-reported, and task-based outcomes in the current meta-analysis were consistent with those reported in previous meta-analysis studies. Although most of the included studies reported no significant differences between CBT and control groups regarding the symptoms of ASD based on self-reported measures, several studies reported inconsistent results. In their study, Frankel et al26  found that children who received Children’s Friendship Training demonstrated reliable change after 3.0 months’ follow-up. Laugeson et al27  included 28 adolescents with ASD and found that adolescents with peers intervention showed significant improvement in social communication, social cognition, social awareness, social motivation, assertion, cooperation, responsibility, autistic mannerisms, and the frequency of peer interactions. The interpretation of the above outcomes could be due to these studies having employed a parent-assisted manualized intervention design for improving the symptoms of ASD, which in turn promotes continued generalization of newly learned skills. However, there was an inconsistent diagnostic assessment scale, which induced uncontrolled measurement biases.

The summary results indicated that CBT significantly improved the symptoms of social-emotional problems based on informant-reported and clinician-rated outcomes, whereas no significant difference was observed regarding the symptoms of social-emotional problems based on self-reported outcomes. Several reasons could be interpreted for these outcomes: (1) parents and teachers of children with appropriate training could implement behavioral management strategies to improve disruptive and repetitive behaviors, and these play an important role in anxiety reduction and skill generalization28 ; (2) the children’s complex presentations could reflect the value of maintaining focused goals, affecting the anxiety and ASD-related barriers; and (3) treatment effectiveness of CBT could be affected by baseline characteristics of enrolled children. Finally, subgroup analyses indicated that treatment effectiveness of CBT could be affected by mean age and study quality. The reasons for this could be due to learning ability and differences in interventions between children and adolescents, which in turn could affect the treatment effectiveness of CBT. Moreover, the quality of included studies could determine the reliability of conclusions of individual trials.

However, our meta-analysis has some limitations that should be acknowledged: (1) the quality of included studies was low to modest, and the conclusion of this study might be restricted by uncontrolled biases; (2) the outcomes investigated by using a various outcome measurement scale could affect the net effect estimates between CBT and control groups, and therefore the pooled analyses in this study applied SMD as an effect estimate; (3) the various CBT strategies and targets across included trials might have contributed an important role on significant heterogeneity for all investigated outcomes; (4) publication bias was inevitable because the current study was based on published RCTs, and unpublished data were not available; and (5) the analysis based on pooled data restricted us to conduct a more detailed analysis.

The symptoms of ASD and social-emotional problems based on informant-reported and clinician-rated outcomes were significantly improved in children or adolescents who received CBT, whereas these outcomes based on self-reported outcomes between CBT and control groups had no significant association. Future large-scale RCTs using a consistent outcome measurement scale for evaluating treatment effectiveness of CBT in children or adolescents should be conducted.

Dr Hao conceptualized and designed the study, drafted the initial manuscript, and reviewed and revised the manuscript; Dr Luo designed the data collection instruments and critically reviewed the manuscript; Dr Zhao and Ms Wang collected data, conducted the initial analyses, drafted the initial manuscript, and reviewed and revised the manuscript; Dr Huang and Ms Chen coordinated and supervised data collection; Mr Zhou and Mr Li extracted the data and assessed their quality; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

FUNDING: Supported by the Hubei Province Natural Fund Project (2016CFB428), the Huazhong University of Science and Technology Key Projects of Independent Innovation Research Fund (2017 kfyxjj100), the Southeast University in Child Development and Learning Science Ministry of Education Key Laboratory Open Fund (CDLS-2018-01), Huazhong University of Science and Technology Emergency Technology Research Project Response to COVID-19 (2020kfyXGYJ020), and Tongji Hospital After Startup Funds.

ASD

autism spectrum disorder

CBT

cognitive behavioral therapy

CI

confidence interval

RCT

randomized controlled trial

SMD

standard mean difference

1
American Psychiatric Association
.
Diagnostic and Statistical Manual of Mental Disorders
, 5th ed.
Washington, DC
:
American Psychiatric Association
;
2013
2
Ingersoll
B
.
Recent advances in early identification and treatment of autism
.
Curr Dir Psychol Sci
.
2011
;
20
(
5
):
335
339
3
Elsabbagh
M
,
Divan
G
,
Koh
YJ
, et al
.
Global prevalence of autism and other pervasive developmental disorders
.
Autism Res
.
2012
;
5
(
3
):
160
179
4
Zeedyk
SM
,
Rodriguez
G
,
Tipton
LA
,
Baker
BL
,
Blacher
J
.
Bullying of youth with autism spectrum disorder, intellectual disability, or typical development: victim and parent perspectives
.
Res Autism Spectr Disord
.
2014
;
8
(
9
):
1173
1183
5
Whitehouse
AJO
,
Durkin
K
,
Jaquet
E
,
Ziatas
K
.
Friendship, loneliness and depression in adolescents with Asperger’s syndrome
.
J Adolesc
.
2009
;
32
(
2
):
309
322
6
Sukhodolsky
DG
,
Bloch
MH
,
Panza
KE
,
Reichow
B
.
Cognitive-behavioral therapy for anxiety in children with high-functioning autism: a meta-analysis
.
Pediatrics
.
2013
;
132
(
5
). Available at: www.pediatrics.org/cgi/content/full/132/5/e1341
7
van Steensel
FJA
,
Bögels
SM
,
Perrin
S
.
Anxiety disorders in children and adolescents with autistic spectrum disorders: a meta-analysis
.
Clin Child Fam Psychol Rev
.
2011
;
14
(
3
):
302
317
8
Sukhodolsky
DG
,
Scahill
L
,
Gadow
KD
, et al
.
Parent-rated anxiety symptoms in children with pervasive developmental disorders: frequency and association with core autism symptoms and cognitive functioning
.
J Abnorm Child Psychol
.
2008
;
36
(
1
):
117
128
9
Ghaziuddin
M
.
Mental Health Aspects of Autism and Asperger Syndrome
.
London, United Kingdom
:
Jessica Kingsley Publishers
;
2005
10
Rotheram-Fuller
E
,
MacMullen
L
.
Cognitive-behavioral therapy for children with autism spectrum disorders
.
Psychol Sch
.
2011
;
48
(
3
):
263
271
11
Beck
AT
.
Cognitive therapy: past, present, and future
.
J Consult Clin Psychol
.
1993
;
61
(
2
):
194
198
12
Spreckley
M
,
Boyd
R
.
Efficacy of applied behavioral intervention in preschool children with autism for improving cognitive, language, and adaptive behavior: a systematic review and meta-analysis
.
J Pediatr
.
2009
;
154
(
3
):
338
344
13
Ung
D
,
Selles
R
,
Small
BJ
,
Storch
EA
.
A systematic review and meta-analysis of cognitive-behavioral therapy for anxiety in youth with high-functioning autism spectrum disorders
.
Child Psychiatry Hum Dev
.
2015
;
46
(
4
):
533
547
14
Moher
D
,
Liberati
A
,
Tetzlaff
J
,
Altman
DG
;
PRISMA Group
.
Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement
.
PLoS Med
.
2009
;
6
(
7
):
e1000097
15
Jadad
AR
,
Moore
RA
,
Carroll
D
, et al
.
Assessing the quality of reports of randomized clinical trials: is blinding necessary?
Control Clin Trials
.
1996
;
17
(
1
):
1
12
16
DerSimonian
R
,
Laird
N
.
Meta-analysis in clinical trials
.
Control Clin Trials
.
1986
;
7
(
3
):
177
188
17
Ades
AE
,
Lu
G
,
Higgins
JPT
.
The interpretation of random-effects meta-analysis in decision models
.
Med Decis Making
.
2005
;
25
(
6
):
646
654
18
Deeks
JJ
,
Higgins
JPT
,
Altman
DG
. Analyzing Data and Undertaking Meta-Analyses. In:
Higgins
JPT
,
Green
S
, eds.
Cochrane Handbook for Systematic Reviews of Interventions. Version 5.0.1
.
London, United Kingdom
:
The Cochrane Collaboration
;
2008
:
243
296
19
Higgins
JPT
,
Thompson
SG
,
Deeks
JJ
,
Altman
DG
.
Measuring inconsistency in meta-analyses
.
BMJ
.
2003
;
327
(
7414
):
557
560
20
Tobias
A
.
Assessing the influence of a single study in the meta-analysis estimate
.
Stata Technical Bulletin
.
1999
;
8
(
41
):
7526
7529
21
Altman
DG
,
Bland
JM
.
Interaction revisited: the difference between two estimates
.
BMJ
.
2003
;
326
(
7382
):
219
22
Egger
M
,
Davey Smith
G
,
Schneider
M
,
Minder
C
.
Bias in meta-analysis detected by a simple, graphical test
.
BMJ
.
1997
;
315
(
7109
):
629
634
23
Begg
CB
,
Mazumdar
M
.
Operating characteristics of a rank correlation test for publication bias
.
Biometrics
.
1994
;
50
(
4
):
1088
1101
24
Duval
S
,
Tweedie
R
.
A nonparametric “trim and fill” method of accounting for publication bias in meta-analysis
.
J Am Stat Assoc
.
2000
;
95
(
449
):
89
98
25
Weston
L
,
Hodgekins
J
,
Langdon
PE
.
Effectiveness of cognitive behavioural therapy with people who have autistic spectrum disorders: a systematic review and meta-analysis
.
Clin Psychol Rev
.
2016
;
49
:
41
54
26
Frankel
F
,
Myatt
R
,
Sugar
C
,
Whitham
C
,
Gorospe
CM
,
Laugeson
E
.
A randomized controlled study of parent-assisted Children’s Friendship Training with children having autism spectrum disorders
.
J Autism Dev Disord
.
2010
;
40
(
7
):
827
842
27
Laugeson
EA
,
Frankel
F
,
Gantman
A
,
Dillon
AR
,
Mogil
C
.
Evidence-based social skills training for adolescents with autism spectrum disorders: the UCLA PEERS program
.
J Autism Dev Disord
.
2012
;
42
(
6
):
1025
1036
28
Reaven
J
,
Hepburn
S
.
The parent’s role in the treatment of anxiety symptoms in children with high-functioning autism spectrum disorders
.
Mental Health Aspects Developmental Disabilities
.
2006
;
9
(
3
):
73
80

Competing Interests

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.

Supplementary data