Key Points
Question
How do psychological treatments compare in terms of alleviating pediatric posttraumatic stress disorder (PTSD)?
Findings
In this systematic review and network meta-analysis, trauma-focused cognitive behavior therapies (TF-CBTs), eye movement desensitization and reprocessing (EMDR), non–trauma-focused interventions, and multidisciplinary treatments (MDTs) were all associated with a significant reduction in pediatric PTSD relative to passive control conditions in the short term. TF-CBTs were associated with the largest short- and long-term reductions in pediatric PTSD, but EMDR and MDTs had insufficient long-term data.
Meaning
The findings suggest TF-CBTs should be the first-line treatment recommendation for pediatric PTSD; while data for other treatment approaches are emerging with some promising findings, more data (including long-term data) are needed to draw firmer conclusions.
This systematic review and network meta-analysis evaluates psychiatric interventions in pediatric posttraumatic stress disorder.
Abstract
Importance
Pediatric posttraumatic stress disorder (PTSD) is a common and debilitating mental disorder, yet a comprehensive network meta-analysis examining psychological interventions is lacking.
Objective
To synthesize all available evidence on psychological interventions for pediatric PTSD in a comprehensive systematic review and network meta-analysis.
Data Sources
PsycINFO, MEDLINE, Web of Science, and PTSDpubs were searched from inception to January 2, 2024, and 74 related systematic reviews were screened.
Study Selection
Two independent raters screened publications for eligibility. Inclusion criteria were randomized clinical trial (RCT) with at least 10 patients per arm examining a psychological intervention for pediatric PTSD compared to a control group in children and adolescents (19 years and younger) with full or subthreshold PTSD.
Data Extraction and Synthesis
PRISMA guidelines were followed to synthesize and present evidence. Two independent raters extracted data and assessed risk of bias with Cochrane criteria. Random-effects network meta-analyses were run.
Main Outcome and Measures
Standardized mean differences (Hedges g) in PTSD severity.
Results
In total, 70 RCTs (N = 5528 patients) were included. Most RCTs (n = 52 [74%]) examined trauma-focused cognitive behavior therapies (TF-CBTs). At treatment end point, TF-CBTs (g, 1.06; 95% CI, 0.86-1.26; P < .001), eye movement desensitization and reprocessing (EMDR; g, 0.86; 95% CI, 0.54-1.18; P < .001), multidisciplinary treatments (MDTs) (g, 0.88; 95% CI, 0.53-1.23; P < .001), and non–trauma-focused interventions (g, 0.95; 95% CI, 0.62-1.28; P < .001) were all associated with significantly larger reductions in pediatric PTSD than passive control conditions. TF-CBTs were associated with the largest short-term reductions in pediatric PTSD relative to both passive and active control conditions and across all sensitivity analyses. In a sensitivity analysis including only trials with parent involvement, TF-CBTs were associated with significantly larger reductions in pediatric PTSD than non–trauma-focused interventions (g, 0.35; 95% CI, 0.04-0.66; P = .03). Results for midterm (up to 5 months posttreatment) and long-term data (6-24 months posttreatment) were similar.
Conclusions and Relevance
Results from this systematic review and network meta-analysis indicate that TF-CBTs were associated with significant reductions in pediatric PTSD in the short, mid, and long term. More long-term data are needed for EMDR, MDTs, and non–trauma-focused interventions. Results of TF-CBTs are encouraging, and disseminating these results may help reduce common treatment barriers by counteracting common misconceptions, such as the notion that TF-CBTs are harmful rather than helpful.
Introduction
One- to two-thirds of children and adolescents from the general population report exposure to at least 1 traumatic event.1,2,3,4 While most children and adolescents react resiliently to trauma, about one-fifth develop posttraumatic stress disorder (ie, pediatric PTSD).4,5 Pediatric PTSD is a common, impairing,4 and often chronic6 mental disorder characterized by reexperiencing of trauma, avoidance of trauma-related stimuli, changes in cognition and emotion, and hyperarousal.7 Given the high prevalence and disease burden of pediatric PTSD,8,9,10,11 the examination and implementation of efficacious treatments constitutes a public health priority.
International treatment guidelines recommend trauma-focused cognitive behavior therapies (TF-CBTs; eg, prolonged exposure12) as first-line treatment for pediatric PTSD.13,14,15,16,17,18 Research on other psychological interventions, such as eye movement desensitization and reprocessing (EMDR) or non–trauma-focused interventions, is also emerging. In recent years, the number of randomized clinical trials (RCTs) published on these therapies has increased substantially. To inform clinical practice about the relative reductions in pediatric PTSD of all treatment approaches, a comprehensive systematic review and network meta-analysis (NMA) is required.
NMAs integrate data from both direct (ie, comparison of arms within an RCT) and indirect (ie, comparison of arms across RCTs) comparisons, which enables conclusions about the relative effects of all interventions.19 Three NMAs of psychological interventions for pediatric PTSD have been published.20,21,22 However, 4 omissions in previous work need to be addressed. First, a comprehensive NMA is needed. Caro et al22 only focused on pediatric PTSD relating to sexual abuse. Mavranezouli et al21 analyzed follow-up data up to 4 months posttreatment and cannot discern long-term reductions in pediatric PTSD. Xiang et al20 included data from 56 RCTs published until 2020, compared to 70 RCTs published until 2024 in the present work. Second, no previous NMA included a sensitivity analysis of high-quality evidence. Low-quality evidence may bias results in quantitative synthesis.23 Third, no NMA has performed sensitivity analyses concerning delivery format (eg, individual delivery only or treatments with parent or caregiver involvement only), and reductions in pediatric PTSD may differ by delivery format. Fourth, reductions in pediatric PTSD might also differ by age group (ie, children vs adolescents), and no NMA has yet performed an age-based sensitivity analysis. To enhance our understanding of the relative performance of psychological interventions for pediatric PTSD, the present work addresses these 4 omissions.
Methods
We followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guideline.24 The systematic literature search, data extraction, and risk of bias assessment were carried out independently by at least 2 authors. Disagreements were discussed between at least 3 authors (T.H.H., L.W., A.K., and N.M.). To address missing data, we sent data requests to corresponding authors, with a reminder 1 month later. The objectives and methods of the present NMA were preregistered in the PROSPERO database (CRD42020206290). We defined the main research question (population, intervention, comparison, outcome, and study) as follows. In children and adolescents with full or subthreshold PTSD, how do psychological interventions—compared to passive control conditions, active control conditions, or among different categories of interventions—perform in terms of lowering PTSD symptom severity in RCTs?
Identification and Selection of Studies
Search Strategy
From inception to April 21, 2022, we relied on our previous literature search,25 including 57 eligible RCTs. We conducted a new search wave with identical search strategy on January 2, 2024, covering literature published between April 21, 2022, and January 2, 2024. eAppendix 1 in Supplement 1 outlines the full search string. We performed all-field searches in PsycINFO, MEDLINE, Web of Science, and PTSDpubs with various search terms for PTSD (ptsd OR ptss OR posttraumatic stress OR posttraumatic stress) and treatment (trial* OR treatment* OR intervention*). Consistent with our previous search, no restrictions were applied to languages or publication formats. We also screened 74 recently published related reviews (eAppendix 2 in Supplement 1) and the reference lists of included trials.
Eligibility Criteria
In line with our previous work,25 we included trials that met all of the following inclusion criteria: (1) RCT, (2) investigating a psychological intervention for pediatric PTSD compared to a control condition, (3) all participants had full or subthreshold PTSD, (4) sample mean age younger than 19.0 years, and (5) outcome data reported for at least 10 participants per arm.
Quality Assessment
Risk of bias was independently assessed by 2 authors (T.H.H. and N.M.) based on 8 quality criteria reported by Cuijpers and colleagues.23 These 8 criteria originated from the Cochrane Collaboration criteria26 and authoritative criteria for evidence-based psychological interventions.27 In the present study, RCTs were defined as high-quality trials (ie, with low risk of bias) when fulfilling at least 6 of the 8 quality criteria (for quality criteria and quality ratings per trial see eAppendices 3 and 4 in Supplement 1, respectively). Initial agreement between independent raters was good (91.35%).
Data Extraction
Trial characteristics (eg, treatment delivery format), sample characteristics (eg, mean age), and PTSD outcome data were extracted independently by at least 2 authors (T.H.H., L.W., A.K., and N.M.). When applicable, intention-to-treat data were prioritized over per-protocol data. When applicable, outcome data from clinician-based measures were prioritized over self-report measures.
Categorization of Psychological Interventions and Control Conditions
The present study compared 4 categories of psychological interventions based on the number of trials available: (1) TF-CBTs (ie, any CBT-based intervention with a trauma focus, such as prolonged exposure12 or cognitive processing therapy28), (2) EMDR, (3) non–trauma-focused interventions (ie, any intervention without a trauma focus), and (4) multidisciplinary treatments (MDTs; ie, treatments that combine techniques from at least 2 of the aforementioned categories, such as the intensive multimodal group program29). Other trauma-focused interventions (ie, interventions with a trauma focus but not based on CBT or EMDR principles, such as expressive supportive groups30) were planned as a fifth category, but lacked evidence (ie, <4 direct comparisons). Control conditions were divided into passive control conditions (eg, wait-list control) and active control conditions (eg, treatment as usual) (for all categorizations see eAppendix 5 in Supplement 1).
Categorization of Assessment Time Points
Consistent with previous research,31 we distinguished between 3 assessment periods: (1) posttreatment (ie, short term); (2) midterm, which we defined as assessments of up to 5 months after treatment end point; and (3) long term, which we defined as assessments longer than 5 months after treatment end point. When several assessments fell into the midterm and long-term categories, the longest assessment was chosen.
Statistical Analysis
The primary outcome of interest was the standardized mean difference (Hedges g)32 in PTSD severity between comparator groups. Random-effects NMAs were conducted given that high heterogeneity in outcomes was expected.32 Level of statistical significance was set to 2-sided P < .05 for all analyses, including the Egger test. Analyses were performed in R version 4.1.1 (R Foundation)33 with the netmeta package.34 Effect sizes (Hedges g) were first calculated at the study level35 and then pooled and compared between all comparison dyads in NMA.32 Following Cohen,36 g was interpreted as small (0.20), medium (0.50), and large effect (0.80). We only included intervention categories with minimally sufficient evidence (ie, ≥4 direct comparisons).37
For transitivity, we examined whether the distribution of various trial and sample characteristics was similar across comparison dyads and performed sensitivity analyses. We analyzed inconsistency between direct and indirect evidence globally38 and locally (ie, per comparison dyad)24,39 with the net splitting method.40 We also inspected net heat plots.41 We performed inconsistency-corrected analyses when applicable. We calculated outlier-adjusted NMAs whenever 1 or more outlier was detected. Outliers were defined as effects 3.3 standard deviations or more above or below the pooled g.42 To examine potential small-study effects, we performed the Egger test43 and inspected comparison-adjusted funnel plots44 (ie, comparisons of interventions to passive and active control conditions only). We calculated the I2 and τ2 statistics as estimates of overall heterogeneity45 and the Q statistic to estimate heterogeneity within and between comparison dyads.39 We also calculated surface under the cumulative ranking (SUCRA; 50 000 resamples), allowing for a ranking by reductions in pediatric PTSD. To visualize distribution of available evidence, we build network graphs. In addition to the main NMAs (across all data), we performed four sensitivity NMAs including (1) only high-quality trials, (2) only trials delivering treatment(s) individually, (3) only trials involving parents or caregivers in treatment, and (4) only trials with a sample mean age younger than 12 years (ie, involving mainly children) as well as only trials with a sample mean age of 12 years and older (ie, involving mainly adolescents).
Results
Study Selection Process
The new search wave covered 10 752 unique electronic records with 13 additional eligible RCTs. Thus, a total of 70 independent RCTs were eligible. Figure 1 details the study synthesis process.
Figure 1. PRISMA Flow Diagram.
PTSD indicates posttraumatic stress disorder; TF-CBTs, trauma-focused cognitive behavior therapies.
aSee eAppendix 7 in Supplement 1.
Study Characteristics
The 70 RCTs reported data of 5528 participants (for trial characteristics and their references, see eAppendices 6 and 7 in Supplement 1, respectively). Apart from 1 dissertation,46 all RCTs were peer reviewed. Only Dorsey et al47 reported more than 1 RCT (ie, 4 RCTs). In total, 41 RCTs (59% of trials) delivered interventions individually and 29 (41% of trials) involved parents or primary caregivers in treatments. Across trials, the mean (SD) number of total sessions was 10 (5) and total mean (SD) duration of treatments (ie, total sessions × length) was 11 (6.5) hours. In total, 40 RCTs (57% of trials) assessed follow-up data (range, 1 month to 24 months posttreatment). Intention-to-treat PTSD data were reported in 44 RCTs (63% of trials). In total, 41 RCTs (59%) were conducted in high-income countries and the remaining 29 (41%) in low- and middle-income countries. However, 2 RCTs conducted in high-income countries exclusively involved refugees originating from low-income countries.48,49 Across 52 trials reporting this information (74%), 3979 of 4413 participants (90%) met diagnostic criteria for full PTSD at baseline. Fifty-seven trials (81%) involved mixed sex samples, whereas 9 (13%) included only female participants and 4 (6%) only male participants. Across all trials, 3334 of 5528 participants (60%) identified as female. The mean (SD) age across trials was 12.21 (3.08) years. In terms of trauma history, 31 RCTs (44%) included a sample with varying trauma histories. In the other trials, only participants with a particular trauma history were included, such as sexual assault (k = 10 [14%]) or parental death (k = 6 [9%]).
Network Meta-Analyses of Short-, Mid-, and Long-Term Outcomes
Assumptions
Assumptions were mostly met. Apart from 2 analyses, no inconsistencies were observed. In the main NMA on midterm outcomes, significant inconsistency (ie, between direct and indirect evidence) was found for MDTs and a corrected analysis without MDTs was performed. In the sensitivity NMA on midterm outcomes for treatments with individual delivery, significant inconsistencies were found for all comparisons (precluding correction) and results are thus not reported. The distribution of sample and methodological characteristics across comparison dyads is presented in eAppendix 8 in Supplement 1.
Network Graphs
Figure 2 shows the network graphs for the NMAs on short-, mid-, and long-term outcomes. Most available trials assessed TF-CBTs. Only TF-CBTs had enough accumulated evidence across all 3 assessment periods.
Figure 2. Network Graphs for the Main Analyses Concerning Short-, Mid-, and Long-Term Outcomes.
Node sizes are proportional to the number of included participants per dyad and the thickness of lines is proportional to the number of direct comparisons (see also number of direct comparisons inserted for each comparison dyad). ACCs indicates active control conditions (eg, treatment as usual); EMDR, eye movement desensitization and reprocessing; MDTs, multidisciplinary treatments (eg, bounce back); non–TF-PIs, non–trauma-focused psychological interventions (eg, non–directive supportive therapy); PCCs, passive control conditions (eg, waitlist); TF-CBTs, trauma-focused cognitive behavior therapies (eg, prolonged exposure).
Network Meta-Analysis of Short-Term Outcomes
Table 1 provides all short-term results. At treatment end point, TF-CBTs, EMDR, MDTs, and non–trauma-focused interventions were all associated with significantly larger reductions in pediatric PTSD than passive control conditions (see eAppendix 9 in Supplement 1 for the corresponding forest plot). At treatment end point, TF-CBTs (g, 1.06; 95% CI, 0.86-1.26; P < .001), EMDR (g, 0.86; 95% CI, 0.54-1.18; P < .001), MDTs (g, 0.95; 95% CI, 0.62-1.28; P < .001), and non–trauma-focused interventions (g, 0.88; 95% CI, 0.53-1.23; P < .001) were all associated with significantly larger reductions in pediatric PTSD than passive control conditions. Compared to active control conditions, only TF-CBTs (g, 0.55; 95% CI, 0.36-0.74; P < .001), MDTs (g, 0.43; 95% CI, 0.09-0.77; P = .01), and non–trauma-focused interventions (g, 0.36; 95% CI, 0.01-0.72; P = .04) were associated with significantly larger reductions in pediatric PTSD (eAppendix 10 in Supplement 1). Differences in pooled effect sizes between treatment categories were not significant, with few or no direct comparisons for most comparison dyads. Heterogeneity was large within and between comparison dyads (τ2, 0.14; I2, 68.9%; total Q, 196.06 [df, 61; P < .001]; Q for within‐design heterogeneity, 173.76 [df, 50; P < .001]; Q for between‐design heterogeneity, 21.40 [df, 11; P = .03]). No significant inconsistencies were detected in the net splitting method (eAppendices 11 and 12 in Supplement 1). No evidence for small-study effects was found (eAppendix 13 in Supplement 1). Two outliers50,51 investigating TF-CBTs were detected. Outlier-adjusted analysis produced similar results (eAppendix 14 in Supplement 1).
Table 1. Short-Term Outcomes.
| Reference group | Psychological intervention | No. of direct comparisons (No. of participants) | SMD (95% CI)a | P valueb | I 2 | τ2 |
|---|---|---|---|---|---|---|
| Main analysis (across all data, irrespective of trial quality and trauma history) | ||||||
| Relative to PCCs | TF-CBTs | 18 (1145) | 1.06 (0.86 to 1.26) | <.001 | 68.9 | 0.14c |
| EMDR | 7 (297) | 0.86 (0.54 to 1.18) | <.001 | |||
| Non–TF-CBTs | 1 (40) | 0.88 (0.53 to 1.23) | <.001 | |||
| MDTs | 4 (270) | 0.95 (0.62 to 1.28) | <.001 | |||
| ACCs | 1 (22) | 0.52 (0.24 to 0.79) | <.001 | |||
| Relative to ACCs | TF-CBTs | 21 (2048) | 0.55 (0.36 to 0.74) | <.001 | ||
| EMDR | 0 (0) | 0.35 (−0.04 to 0.73) | .08 | |||
| Non–TF-CBTs | 0 (0) | 0.36 (0.01 to 0.72) | .04 | |||
| MDTs | 4 (146) | 0.43 (0.09 to 0.77) | .01 | |||
| Relative to EMDR | TF-CBTs | 3 (185) | 0.20 (−0.14 to 0.55) | .24 | ||
| Non–TF-CBTs | 0 (0) | 0.02 (−0.43 to 0.47) | .93 | |||
| MDTs | 0 (0) | 0.09 (−0.36 to 0.53) | .70 | |||
| Relative to non–TF-CBTs | TF-CBTs | 9 (631) | 0.19 (−0.11 to 0.48) | .22 | ||
| MDTs | 0 (0) | 0.07 (−0.38 to 0.51) | .77 | |||
| Relative to MDTs | TF-CBTs | 2 (72) | 0.12 (−0.21 to 0.45) | .49 | ||
| Sensitivity analysis: high-quality trials onlyd | ||||||
| Relative to PCCs | TF-CBTs | 10 (882) | 1.05 (0.83 to 1.28) | <.001 | 67.8 | 0.10c |
| EMDR | 3 (145) | 0.95 (0.57 to 1.34) | <.001 | |||
| Non–TF-CBTs | 0 (0) | 0.80 (0.37 to 1.23) | <.001 | |||
| MDTs | 3 (218) | 0.91 (0.54 to 1.28) | <.001 | |||
| ACCs | 0 (0) | 0.53 (0.23 to 0.82) | <.001 | |||
| Relative to ACCs | TF-CBTs | 13 (1792) | 0.53 (0.33 to 0.72) | <.001 | ||
| EMDR | 0 (0) | 0.43 (0.01 to 0.85) | .047 | |||
| Non–TF-CBTs | 0 (0) | 0.28 (−0.14 to 0.69) | .19 | |||
| MDTs | 1 (46) | 0.38 (−0.03 to 0.80) | .07 | |||
| Relative to EMDR | TF-CBTs | 3 (185) | 0.10 (−0.28 to 0.48) | .60 | ||
| Non–TF-CBTs | 0 (0) | −0.15 (−0.67 to 0.37) | .57 | |||
| MDTs | 0 (0) | −0.04 (−0.56 to 0.47) | .87 | |||
| Relative to non–TF-CBTs | TF-CBTs | 5 (322) | 0.25 (−0.11 to 0.61) | .17 | ||
| MDTs | 0 (0) | 0.11 (−0.42 to 0.64) | .69 | |||
| Relative to MDTs | TF-CBTs | 1 (50) | 0.14 (−0.25 to 0.53) | .47 | ||
| Sensitivity analysis: individual treatment delivery only | ||||||
| Relative to PCCs | TF-CBTs | 10 (563) | 1.07 (0.79 to 1.35) | <.001 | 63.0 | 0.12c |
| EMDR | 4 (150) | 1.02 (0.62 to 1.41) | <.001 | |||
| Non–TF-CBTs | 1 (40) | 0.88 (0.49 to 1.27) | <.001 | |||
| ACCs | 0 (0) | 0.42 (0.03 to 0.82) | .03 | |||
| Relative to ACCs | TF-CBTs | 10 (766) | 0.65 (0.37 to 0.92) | <.001 | ||
| EMDR | 0 (0) | 0.59 (0.11 to 1.07) | .02 | |||
| Non–TF-CBTs | 0 (0) | 0.46 (0.06 to 0.86) | .02 | |||
| RELATIVE to EMDR | TF-CBTs | 3 (185) | 0.06 (−0.34 to 0.45) | .78 | ||
| Non–TF-CBTs | 0 (0) | −0.13 (−0.62 to 0.35) | .59 | |||
| Relative to non–TF-CBTs | TF-CBTs | 9 (631) | 0.19 (−0.10 to 0.48) | .20 | ||
| Sensitivity analysis: treatments with parent involvement only | ||||||
| Relative to PCCs | TF-CBTs | 5 (364) | 1.07 (0.76 to 1.38) | <.001 | 55.5 | 0.07c |
| Non–TF-CBTs | 0 (0) | 0.72 (0.28 to 1.16) | .001 | |||
| MDTs | 3 (188) | 0.94 (0.59 to 1.29) | <.001 | |||
| ACCs | 0 (0) | 0.65 (0.28 to 1.02) | <.001 | |||
| Relative to ACCs | TF-CBTs | 9 (852) | 0.42 (0.19 to 0.66) | <.001 | ||
| Non–TF-CBTs | 0 (0) | 0.07 (−0.32 to 0.46) | .72 | |||
| MDTs | 2 (56) | 0.29 (−0.11– 0.69) | .16 | |||
| Relative to non–TF-CBTs | TF-CBTs | 5 (476) | 0.35 (0.04 to 0.66) | .03 | ||
| MDTs | 0 (0) | 0.22 (−0.26 to 0.70) | .38 | |||
| Relative to MDTs | TF-CBTs | 2 (72) | 0.14 (−0.23 to 0.50) | .47 | ||
| Sensitivity analysis: samples with mean age <12 y (ie, mostly children) | ||||||
| Relative to PCCs | TF-CBTs | 4 (252) | 1.08 (0.66 to 1.49) | <.001 | 69.2 | 0.12c |
| EMDR | 3 (89) | 0.86 (0.35 to 1.36) | <.001 | |||
| Non–TF-CBTs | 0 (0) | 0.78 (0.24 to 1.32) | .004 | |||
| ACCs | 0 (0) | 0.52 (0.03 to 1.01) | .04 | |||
| Relative to ACCs | TF-CBTs | 10 (1104) | 0.55 (0.29 to 0.81) | <.001 | ||
| EMDR | 0 (0) | 0.34 (−0.29 to 0.96) | .29 | |||
| Non–TF-CBTs | 0 (0) | 0.26 (−0.17 to 0.69) | .24 | |||
| Relative to EMDR | TF-CBTs | 1 (52) | 0.22 (−0.35 to 0.78) | .45 | ||
| Non–TF-CBTs | 0 (0) | −0.08 (−0.74 to 0.58) | .82 | |||
| Relative to non–TF-CBTs | TF-CBTs | 6 (507) | 0.29 (−0.05 to 0.63) | .10 | ||
| Sensitivity analysis: samples with mean age ≥12 y (ie, mostly adolescents) | ||||||
| Relative to PCCs | TF-CBTs | 13 (854) | 1.09 (0.82 to 1.37) | <.001 | 73.6 | 0.19c |
| EMDR | 3 (147) | 0.93 (0.42 to 1.45) | <.001 | |||
| MDTs | 1 (82) | 1.02 (0.53 to 1.52) | <.001 | |||
| ACCs | 1 (22) | 0.56 (0.18 to 0.95) | .004 | |||
| Relative to ACCs | TF-CBTs | 11 (944) | 0.53 (0.24 to 0.82) | <.001 | ||
| EMDR | 0 (0) | 0.37 (−0.22 to 0.96) | .22 | |||
| MDTs | 4 (146) | 0.46 (0.02 to 0.90) | .04 | |||
| Relative to EMDR | TF-CBTs | 2 (133) | 0.16 (−0.36 to 0.68) | .55 | ||
| MDTs | 0 (0) | 0.09 (−0.59 to 0.77) | .79 | |||
| Relative to MDTs | TF-CBTs | 2 (72) | 0.07 (−0.39 to 0.53) | .77 | ||
Abbreviations: ACCs, active control conditions (eg, treatment as usual); EMDR, eye movement desensitization and reprocessing; MDTs, multidisciplinary treatments; PCCs, passive control conditions (eg, waitlist); SMD, standardized mean difference (ie, Hedges g); TF-CBTs, trauma-focused cognitive behavior therapies.
A positive SMD indicates superior reductions in pediatric posttraumatic stress disorder of the given psychological intervention relative to the given reference group; a negative SMD, inferior reductions.
P values correspond to the respective Q statistic as a measure of heterogeneity in outcomes.
P < .001.
Meeting at least 6 of 8 trial quality criteria (Cuijpers et al23).
Sensitivity Analyses for Short-Term Outcomes
In high-quality trials only, the results for comparisons to passive control conditions were similar to those of the main analysis. Only TF-CBTs (g, 0.53; 95% CI, 0.33-0.72; P < .001) and EMDR (g, 0.43; 95% CI, 0.01-0.85; P = .047) were associated with larger short-term reductions in pediatric PTSD than active control conditions. In the sensitivity analysis concerning trials with individual treatment delivery, results were similar to the main analysis with the most favorable outcomes for TF-CBTs. In the sensitivity analysis concerning only trials with parent or caregiver involvement, results were similar for the comparison to passive control conditions, with TF-CBT, MDTs, and non–trauma-focused interventions being associated with significantly larger reductions in PTSD. Compared to active control conditions, however, only TF-CBTs (g, 0.42; 95% CI, 0.19-0.66; P < .001) were associated with significant reductions in PTSD. Moreover, TF-CBTs with parent or caregiver involvement were associated with larger reductions in pediatric PTSD than non–trauma-focused interventions with caregiver involvement (g, 0.35; 95% CI, 0.04-0.66; P = .03). In the sensitivity analysis of samples with mean age younger than 12 years (mostly children), results were similar to those of the main analysis. Yet, only TF-CBTs were associated with significantly larger reductions compared to active controls (g, 0.55; 95% CI, 0.04-0.66; P < .001). In the sensitivity analysis of samples with mean age 12 years and older (mostly adolescents), results were also similar to those of the main analysis. Only TF-CBTs (g, 0.53; 95% CI, 0.24-0.82; P < .001) and MDTs (g, 0.46; 95% CI, 0.02-0.90; P = .04) were associated with significantly larger reductions in PTSD than active controls.
Network Meta-Analyses of Mid- and Long-Term Outcomes
Table 2 provides all results for mid- and long-term results. For non–trauma-focused interventions, too few direct comparisons were available. At midterm (up to 5 months posttreatment), TF-CBTs, EMDR, and MDTs were associated with significantly larger reductions in pediatric PTSD than passive control conditions, with g values being moderate to large and ranging from 0.59 (95% CI, 0.03-1.15; P = .04) for MDTs to 0.87 (95% CI, 0.58-1.17; P < .001) for TF-CBTs to 0.95 (95% CI, 0.48-1.41; P < .001) for EMDR (eAppendix 15 in Supplement 1). Compared to active control conditions, only EMDR (g, 0.52; 95% CI, 0.04-1.00; P = .03) and TF-CBTs (g, 0.45; 95% CI, 0.17-0.73; P = .002) were associated with significant pooled g values (eAppendix 16 in Supplement 1). Heterogeneity was large within and between comparison dyads (τ2, 0.15; I2, 66.4%; total Q, 68.49 [df, 23; P < .001]; Q for within‐design heterogeneity, 49.40 [df, 16; P < .001]; Q for between‐design heterogeneity, 19.85 [df, 7; P = .006]). There was no evidence for small-study effects (eAppendix 17 in Supplement 1). Significant inconsistency was detected for MDTs (eAppendices 18 and 19 in Supplement 1). Results remained similar to those of the main analysis in a reanalysis excluding MDTs (eAppendix 20 in Supplement 1).
Table 2. Midterm and Long-Term Outcomes.
| Reference group | Psychological intervention | No. of direct comparisons (No. of participants | SMD (95% CI)a | P valueb | I 2 | τ2 |
|---|---|---|---|---|---|---|
| Midterm data (ie, ≤5 mo follow-up): main analysis | ||||||
| Relative to PCCs | TF-CBTs | 9 (389) | 0.87 (0.58 to 1.17) | <.001 | 66.4 | 0.15c |
| EMDR | 2 (86) | 0.95 (0.48 to 1.41) | <.001 | |||
| MDTs | 1 (52) | 0.59 (0.03 to 1.15) | .04 | |||
| ACCs | 2 (74) | 0.42 (0.06 to 0.79) | .02 | |||
| Relative to ACCs | TF-CBTs | 9 (813) | 0.45 (0.17 to 0.73) | .002 | ||
| EMDR | 1 (74) | 0.52 (0.04 to 1.00) | .03 | |||
| MDTs | 3 (79) | 0.17 (−0.36 to 0.69) | .54 | |||
| Relative to EMDR | TF-CBTs | 2 (125) | −0.07 (−0.51 to 0.37) | .75 | ||
| MDTs | 0 (0) | −0.35 (−1.02 to 0.31) | .30d | |||
| Relative to MDTs | TF-CBTs | 1 (20) | 0.28 (−0.26 to 0.83) | .31 | ||
| Sensitivity analysis: high-quality trials d only | ||||||
| Relative to PCCs | TF-CBTs | 4 (223) | 1.06 (0.64 to 1.49) | <.001 | 71.9 | 0.14c |
| EMDR | 1 (23) | 1.15 (0.56 to 1.75) | <.001 | |||
| ACCs | 1 (52) | 0.73 (0.24 to 1.21) | .004 | |||
| Relative to ACCs | TF-CBTs | 7 (781) | 0.33 (0.03 to 0.63) | .03 | ||
| EMDR | 1 (74) | 0.43 (−0.08 to 0.94) | .10 | |||
| Relative to EMDR | TF-CBTs | 2 (125) | −0.09 (−0.58 to 0.39) | .70 | ||
| Sensitivity analysis: treatments with parent involvement only | ||||||
| Relative to PCCs | TF-CBTs | 3 (128) | 1.09 (0.56 to 1.63) | <.001 | 71.1 | 0.14c |
| MDTs | 1 (52) | 0.43 (−0.25 to 1.10) | .22 | |||
| ACCs | 0 (0) | 0.73 (0.11 to 1.35) | .02 | |||
| Relative to ACCs | TF-CBTs | 5 (619) | 0.37 (−0.02 to 0.75) | .06 | ||
| MDTs | 2 (43) | −0.30 (−0.95 to 0.34) | .36 | |||
| Relative to MDTs | TF-CBTs | 1 (20) | 0.67 (0.02 to 1.32) | .04 | ||
| Sensitivity analysis: samples with mean age ≥12 y (ie, mostly adolescents) | ||||||
| Relative to PCCs | TF-CBTs | 6 (252) | 0.76 (0.43 to 1.09) | <.001 | 44.9 | 0.07e |
| MDTs | 0 (0) | 0.70 (0.05 to 1.34) | .03 | |||
| ACCs | 2 (74) | 0.22 (−0.18 to 0.61) | .28 | |||
| Relative to ACCs | TF-CBTs | 7 (492) | 0.54 (0.26 to 0.83) | <.001 | ||
| MDTs | 3 (79) | 0.48 (−0.05 to 1.01) | .08 | |||
| Relative to MDTs | TF-CBTs | 1 (20) | 0.06 (−0.52 to 0.64) | .83 | ||
| Long-term data (ie, 6-24–mo follow-up): main analysis | ||||||
| Relative to PCCs | TF-CBTs | 3 (118) | 0.76 (0.27 to 1.26) | .002 | 67.6 | 0.11c |
| Non–TF | 1 (40) | 0.71 (0.15 to 1.27) | .01 | |||
| ACCs | 1 (51) | 0.21 (−0.31 to 0.73) | .43 | |||
| Relative to ACCs | TF-CBTs | 9 (920) | 0.55 (0.30 to 0.81) | <.001 | ||
| Non–TF-CBTs | 1 (45) | 0.50 (0.09 to 0.93) | .02 | |||
| Relative to non–TF-CBTs | TF-CBTs | 5 (343) | 0.06 (−0.29 to 0.40) | .75 | ||
| Sensitivity analysis: high-quality trials onlyd | ||||||
| Relative to ACCs | TF-CBTs | 8 (887) | 0.53 (0.27 to 0.80) | <.001 | 70.7 | 0.11c |
| Non–TF-CBTs | 1 (45) | 0.46 (−0.03 to 0.95) | .07 | |||
| Relative to non–TF-CBTs | TF-CBTs | 3 (151) | 0.07 (−0.37 to 0.52) | .74 | ||
| Sensitivity analysis: individual treatment delivery only | ||||||
| Relative to PCCs | TF-CBT | 3 (118) | 0.78 (0.34 to 1.22) | <.001 | 46.2 | 0.07 |
| Non–TF-CBTs | 1 (40) | 0.64 (0.13 to 1.15) | .01 | |||
| ACCs | 1 (51) | 0.17 (−0.33 to 0.67) | .50 | |||
| Relative to ACCs | TF-CBT | 5 (280) | 0.61 (0.28 to 0.94) | <.001 | ||
| Non–TF-CBTs | 0 (0) | 0.47 (0.01 to 0.92) | .04 | |||
| Relative to non–TF-CBTs | TF-CBT | 5 (343) | 0.14 (−0.18 to 0.46) | .38 | ||
| Sensitivity analysis: samples with mean age <12 y (ie, mostly children) | ||||||
| Relative to PCCs | TF-CBTs | 1 (26) | 1.00 (−0.11 to 2.11) | .08 | 78.5 | 0.15c |
| Non–TF-CBTs | 0 (0) | 0.83 (−0.37 to 2.03) | .18 | |||
| ACCs | 0 (0) | 0.46 (−0.71 to 1.63) | .44 | |||
| Relative to ACCs | TF-CBTs | 5 (673) | 0.54 (0.18 to 0.90) | .003 | ||
| Non–TF-CBTs | 1 (45) | 0.37 (−0.16 to 0.91) | .17 | |||
| Relative to non–TF-CBTs | TF-CBTs | 3 (266) | 0.17 (−0.29 to 0.63) | .47 | ||
Abbreviations: ACCs, active control conditions (eg, treatment as usual); EMDR, eye movement desensitization and reprocessing; MDTs, multidisciplinary treatments; PCCs, passive control conditions (eg, waitlist); SMD, standardized mean difference (ie, Hedges g); TF-CBTs, trauma-focused cognitive behavior therapies.
A positive SMD indicates superior reductions in pediatric pediatric posttraumatic stress disorder of the given psychological intervention relative to the given reference group; a negative SMD, inferior reductions.
P values correspond to the respective Q statistic as a measure of heterogeneity in outcomes.
P < .001.
Meeting at least 6 of 8 trial quality criteria (Cuijpers et al23).
P < .05.
In the long term (6 to 24 months posttreatment), only TF-CBTs and non–trauma-focused interventions had sufficient evidence. Compared to passive control conditions, both TF-CBTs (g, 0.76; 95% CI, 0.27-1.26; P = .002) and non–trauma-focused interventions (g, 0.71; 95% CI, 0.15-1.27; P = .01) were associated with significantly larger reductions in pediatric PTSD (eAppendix 21 in Supplement 1). Both TF-CBTs (g, 0.55; 95% CI, 0.30-0.81; P < .001) and non–trauma-focused interventions (g, 0.50; 95% CI, 0.09-0.93; P = .02) were associated with significantly larger reductions than active control conditions (eAppendix 22 in Supplement 1). Heterogeneity was large within and between comparison dyads (τ2, 0.11; I2, 67.6%; total Q, 46.30 [df, 15; P < .001]; Q for within‐design heterogeneity, 32.33 [df, 10; P < .001]; Q for between‐design heterogeneity, 14.02 [df, 5; P = .02]). No inconsistencies (eAppendices 23 and 24 in Supplement 1) and no evidence for small-study effects (eAppendix 25 in Supplement 1) were found.
Sensitivity Analyses for Mid- and Long-Term Outcomes
At midterm, sensitivity analysis on high-quality trials could be conducted with TF-CBTs and EMDR only. EMDR (g, 1.15; 95% CI, 0.56-1.75; P < .001) and TF-CBTs (g, 1.06; 95% CI, 0.64-1.49; P < .001) were associated with large effect sizes compared to passive control conditions. Compared to active control conditions, however, only TF-CBTs (g, 0.33; 95% CI, 0.03-0.63; P = .03) were associated with significantly larger reductions in PTSD. The sensitivity analysis concerning trials with individual treatment delivery only was infeasible given detected inconsistency for all intervention categories. The sensitivity analysis of trials with parent/caregiver involvement produced similar results to the main analysis. Yet, TF-CBTs were associated with significantly larger reductions in pediatric PTSD relative to MDTs (g, 0.67; 95% CI, 0.02-1.32; P = .04). The sensitivity analysis of trials with mean age younger than 12 years was infeasible due to lacking evidence. The sensitivity analysis of trials with mean age 12 years and older produced similar results to the main analysis. TF-CBT (g, 0.76; 95% CI, 0.43-1.09; P < .001) and MDTs (g, 0.70; 95% CI, 0.05-1.34; P = .03) were both associated with significantly larger reductions in pediatric PTSD than passive controls. Yet, only TF-CBTs (g, 0.54; 95% CI, 0.26-0.83; P < .001) were associated with significantly larger reductions when compared to active controls.
At long term, only TF-CBTs and non–trauma-focused interventions had sufficient data. Passive control conditions were also lacking. Compared to active control conditions in high-quality trials, only TF-CBTs were associated with significantly larger reductions in PTSD (g, 0.53; 95% CI, 0.27-0.80; P < .001). Sensitivity analysis concerning trials with individual treatment delivery only produced similar results to the main analysis. Sensitivity analysis concerning trials with parent/caregiver involvement was infeasible (<4 direct comparisons). The sensitivity analysis of trials with mean age 12 years and older was infeasible due to lacking evidence. The sensitivity analysis of trials with mean age younger than 12 produced similar results to the main analysis, with only TF-CBT (but not non–trauma-focused interventions) being associated with significantly larger reductions in pediatric PTSD compared to active controls (g, 0.54; 95% CI, 0.18-0.90; P = .003).
Ranking of Intervention Categories
Table 3 shows SUCRA rankings. TF-CBTs were the highest-ranking category of interventions at all time points and all analyses, except at midterm when it was second to EMDR.
Table 3. Rankings of Psychological Interventions for Pediatric Posttraumatic Stress Disorder.
| Psychological therapy | SUCRA | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Short term | Midterm | Long term | ||||||||||||||
| All dataa | Outlier adjusteda | High qualitya,b | Individual deliverya | Caregiver involvementa | Mean sample age <12 ya | Mean sample age ≥12 y a | All datac | MDTs deletedc | High qualityc,b | Caregiver involvementc | Mean sample age ≥12 yc | All datab | High quality b,d | Individual deliveryb | Mean sample age <12 yb,e | |
| TF-CBTs | 0.90 | 0.89 | 0.87 | 0.88 | 0.94 | 0.87 | 0.83 | 0.81 | 0.80 | 0.77 | 0.98 | 0.86 | 0.87 | 0.81 | 0.93 | 0.91 |
| EMDR | 0.58 | 0.61 | 0.71 | 0.77 | NA | 0.69 | 0.64 | 0.87 | 0.86 | 0.87 | NA | NA | NA | NA | NA | NA |
| MDTs | 0.70 | 0.72 | 0.66 | NA | 0.74 | NA | 0.74 | 0.50 | NAe | NA | 0.36 | 0.78 | NA | NA | NA | NA |
| Non–TF-CBTs | 0.60 | 0.57 | 0.52 | 0.60 | 0.46 | 0.60 | NA | NA | NA | NA | NA | NA | 0.79 | 0.67 | 0.73 | 0.68 |
| ACCs | 0.21 | 0.21 | 0.23 | 0.25 | 0.36 | 0.34 | 0.28 | 0.32 | 0.34 | 0.36 | 0.62 | 0.31 | 0.26 | 0.02f | 0.25 | 0.29 |
| PCCs | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 | 0.01 | 0.00 | 0.01 | 0.00 | 0.00 | 0.04 | 0.05 | 0.07 | NA | 0.09 | 0.12 |
Abbreviations: ACCs, active control conditions (eg, treatment as usual); EMDR, eye movement desensitization and reprocessing; MDTs, multidisciplinary treatments; NA, not applicable (given insufficient accumulated evidence; ie, <4 direct comparisons); PCCs, passive control conditions (ie, waitlist); SUCRA, surface under the cumulative ranking curve (ie, reductions in pediatric posttraumatic stress disorder ranked by means of surface under the cumulative ranking with 50 000 resamples); TF-CBTs, trauma-focused cognitive behavior therapies.
Short term defined as assessments at treatment end point.
Long term defined as assessments 6 to 24 months after treatment end point.
Midterm defined as assessments 5 months or fewer after treatment end point.
High quality indicates fulfilling at least 6 of 8 quality criteria (Cuijpers et al23).
Excluded from this analysis given that significant inconsistency was detected.
Reference group for this analysis given insufficient number of direct comparisons (k <4) for PCC.
Discussion
This systematic review and network meta-analysis synthesized data from 70 RCTs. TF-CBTs were the most evaluated treatment category for PTSD in children and adolescents. TF-CBTs were associated with the highest reductions in pediatric PTSD relative to control conditions in the short and long term, followed by (in this order) EMDR, MDTs, and non–trauma-focused interventions. This supports recommendations of international treatment guidelines for pediatric PTSD, including the International Society for Traumatic Stress Studies17 and the National Institute of Clinical Excellence.16 Our review confirms and extends previous NMAs20,21,22 and pairwise meta-analyses.25,52,53 TF-CBTs were associated with significant reductions in pediatric PTSD relative to passive and active comparators, across assessment periods, and when restricting analyses to trials with high quality, trials examining individually delivered interventions, trials examining interventions with parent or caregiver involvement, and trials examining samples mainly involving children or adolescents, respectively. These results are important for the training of therapists and implementation in clinical practice and might help in reducing treatment barriers.
Our review also reveals remaining gaps in the literature. While short- and midterm data for EMDR showed that EMDR was associated with significant reductions in pediatric PTSD, our review highlights the lack of long-term follow-up periods. The present results therefore support some international treatment guidelines16,18 that list EMDR as second-line treatment recommendation. Data for MDTs and non–trauma-focused interventions are emerging. However, more data (including long-term data) are needed to robustly investigate the (relative) reductions in pediatric PTSD of these 2 categories. For the time being, about three-fourths of the available data concern TF-CBTs, which means that treatment effects of TF-CBTs could be estimated most robustly. There was some evidence for TF-CBTs being associated with significantly larger reductions in pediatric PTSD than non–trauma-focused interventions, but more data are needed to draw firmer conclusions.
Limitations
This study has limitations. First, the categories of non–trauma-focused interventions, MDTs, and other trauma-focused interventions are heterogenous with regards to theoretical foundations. However, there is no solid ground for further subcategorization given the low number of available RCTs. As more RCTs accumulate, more homogenous categorizations will become feasible. Second, we found evidence for inconsistency in the NMA regarding midterm outcomes. However, results remained similar in a consistency-corrected reanalysis. Third, the age group sensitivity analyses provide approximations of the reduction in PTSD for children and adolescents, as the categorization was based on sample mean age. An individual patient data meta-analysis would allow for a solid differentiation of children and adolescents, which was beyond the scope of the present work. Fourth, while the distinction between passive and active control conditions is a strength of the present work, active control conditions, such as treatment-as-usual, can comprise different elements, depending on the context. Future research might be able to disentangle this heterogenous comparator group. Fifth, relatively low rates of reported intention-to-treat data are concerning and trialists are encouraged to report intention-to-treat data.
Conclusions
This study presents robust evidence indicating that psychological treatments, and in particular TF-CBTs, are associated with significant reductions in pediatric PTSD. A large evidence base for TF-CBTs supports reductions in PTSD relative to both passive and active controls at short, mid, and long term. A comparably thin evidence base indicates that EMDR was associated with significant reductions in PTSD relative to both passive and active controls in the short and midterm. Data for MDTs and non–trauma-focused interventions are emerging. More high-quality data (including long-term data) are needed to draw firmer conclusions regarding the relative performance of psychological treatments for pediatric PTSD.
eAppendix 1. Search string used for systematic literature search
eAppendix 2. References of screened reviews as part of the systematic literature search
eAppendix 3. Quality criteria for risk of bias assessment
eAppendix 4. Quality coding of included trials
eAppendix 5. Categorization of interventions and control conditions
eAppendix 6. Trial characteristics of included trials
eAppendix 7. References of studies included in the present network meta-analysis
eAppendix 8. Trial and sample characteristics across comparison dyads
eAppendix 9. Short-term outcomes: Forest plot compared to passive control conditions
eAppendix 10. Short-term outcomes: Forest plot compared to active control conditions
eAppendix 11. Short-term outcomes: Net splitting results
eAppendix 12. Short-term outcomes: Net heat plot
eAppendix 13. Short-term outcomes: Funnel plot
eAppendix 14. Short-term outcomes: Outlier-adjusted results
eAppendix 15. Mid-term outcomes: Forest plot compared to passive control conditions
eAppendix 16. Mid-term outcomes: Forest plot compared to active control conditions
eAppendix 17. Mid-term outcomes: Funnel plot
eAppendix 18. Mid-term outcomes: Net splitting results
eAppendix 19. Mid-term outcomes: Net heat plot before (top) and after (bottom) exclusion of MDTs
eAppendix 20. Mid-term outcomes: MDTs excluded due to detected inconsistency
eAppendix 21. Long-term outcomes: Forest plot compared to passive control conditions
eAppendix 22. Long-term outcomes: Forest plot compared to active control conditions
eAppendix 23. Long-term outcomes: Net splitting results
eAppendix 24. Long-term outcomes: Net heat plot
eAppendix 25. Long-term outcomes: Funnel plot
Data sharing statement
References
- 1.Copeland WE, Keeler G, Angold A, Costello EJ. Traumatic events and posttraumatic stress in childhood. Arch Gen Psychiatry. 2007;64(5):577-584. doi: 10.1001/archpsyc.64.5.577 [DOI] [PubMed] [Google Scholar]
- 2.Saunders BE, Adams ZW. Epidemiology of traumatic experiences in childhood. Child Adolesc Psychiatr Clin N Am. 2014;23(2):167-184, vii. doi: 10.1016/j.chc.2013.12.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Danese A, Smith P, Chitsabesan P, Dubicka B. Child and adolescent mental health amidst emergencies and disasters. Br J Psychiatry. 2020;216(3):159-162. doi: 10.1192/bjp.2019.244 [DOI] [PubMed] [Google Scholar]
- 4.Lewis SJ, Arseneault L, Caspi A, et al. The epidemiology of trauma and post-traumatic stress disorder in a representative cohort of young people in England and Wales. Lancet Psychiatry. 2019;6(3):247-256. doi: 10.1016/S2215-0366(19)30031-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Alisic E, Zalta AK, van Wesel F, et al. Rates of post-traumatic stress disorder in trauma-exposed children and adolescents: meta-analysis. Br J Psychiatry. 2014;204(5):335-340. doi: 10.1192/bjp.bp.113.131227 [DOI] [PubMed] [Google Scholar]
- 6.Widom CS. Posttraumatic stress disorder in abused and neglected children grown up. Am J Psychiatry. 1999;156(8):1223-1229. doi: 10.1176/ajp.156.8.1223 [DOI] [PubMed] [Google Scholar]
- 7.American Psychiatric Association . Diagnostic and Statistical Manual of Mental Disorders: DSM-5. American Psychiatric Publishing; 2013. [Google Scholar]
- 8.Gilbert R, Widom CS, Browne K, Fergusson D, Webb E, Janson S. Burden and consequences of child maltreatment in high-income countries. Lancet. 2009;373(9657):68-81. doi: 10.1016/S0140-6736(08)61706-7 [DOI] [PubMed] [Google Scholar]
- 9.Lansford JE, Dodge KA, Pettit GS, Bates JE, Crozier J, Kaplow J. A 12-year prospective study of the long-term effects of early child physical maltreatment on psychological, behavioral, and academic problems in adolescence. Arch Pediatr Adolesc Med. 2002;156(8):824-830. doi: 10.1001/archpedi.156.8.824 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Guiney H, Caspi A, Ambler A, et al. Childhood sexual abuse and pervasive problems across multiple life domains: Findings from a five-decade study. Dev Psychopathol. 2024;36(1):219-235. doi: 10.1017/S0954579422001146 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Cohen JA, Deblinger E, Mannarino AP, Steer RA. A multisite, randomized controlled trial for children with sexual abuse-related PTSD symptoms. J Am Acad Child Adolesc Psychiatry. 2004;43(4):393-402. doi: 10.1097/00004583-200404000-00005 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Foa EB, McLean CP, Capaldi S, Rosenfield D. Prolonged exposure vs supportive counseling for sexual abuse-related PTSD in adolescent girls: a randomized clinical trial. JAMA. 2013;310(24):2650-2657. doi: 10.1001/jama.2013.282829 [DOI] [PubMed] [Google Scholar]
- 13.Bandelow B, Allgulander C, Baldwin DS, et al. World Federation of Societies of Biological Psychiatry (WFSBP) guidelines for treatment of anxiety, obsessive-compulsive and posttraumatic stress disorders—version 3. part II: OCD and PTSD. World J Biol Psychiatry. 2023;24(2):118-134. doi: 10.1080/15622975.2022.2086296 [DOI] [PubMed] [Google Scholar]
- 14.World Health Organization . Guidelines on Mental Health Promotive and Preventive Interventions for Adolescents: Helping Adolescents Thrive. World Health Organization; 2020. [PubMed] [Google Scholar]
- 15.Cohen JA, Bukstein O, Walter H, et al. ; AACAP Work Group on Quality Issues . Practice parameter for the assessment and treatment of children and adolescents with posttraumatic stress disorder. J Am Acad Child Adolesc Psychiatry. 2010;49(4):414-430. [PubMed] [Google Scholar]
- 16.National Institute for Health and Care Excellence . Post-traumatic stress disorder (NICE guideline NG116): [B] Evidence reviews for psychological, psychosocial and other non-pharmacological interventions for the treatment of PTSD in children and young people. Accessed May 14, 2024. https://www.nice.org.uk/guidance/ng116/resources/posttraumatic-stress-disorder-pdf-66141601777861 [PubMed]
- 17.Bisson JI, Berliner L, Cloitre M, et al. The international society for traumatic stress studies new guidelines for the prevention and treatment of posttraumatic stress disorder: Methodology and development process. J Trauma Stress. 2019;32(4):475-483. doi: 10.1002/jts.22421 [DOI] [PubMed] [Google Scholar]
- 18.Phelps AJ, Lethbridge R, Brennan S, et al. Australian guidelines for the prevention and treatment of posttraumatic stress disorder: updates in the third edition. Aust N Z J Psychiatry. 2022;56(3):230-247. doi: 10.1177/00048674211041917 [DOI] [PubMed] [Google Scholar]
- 19.Cipriani A, Higgins JPT, Geddes JR, Salanti G. Conceptual and technical challenges in network meta-analysis. Ann Intern Med. 2013;159(2):130-137. doi: 10.7326/0003-4819-159-2-201307160-00008 [DOI] [PubMed] [Google Scholar]
- 20.Xiang Y, Cipriani A, Teng T, et al. Comparative efficacy and acceptability of psychotherapies for post-traumatic stress disorder in children and adolescents: a systematic review and network meta-analysis. Evid Based Ment Health. 2021;24(4):153-160. doi: 10.1136/ebmental-2021-300346 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Mavranezouli I, Megnin-Viggars O, Daly C, et al. Research Review: Psychological and psychosocial treatments for children and young people with post-traumatic stress disorder: a network meta-analysis. J Child Psychol Psychiatry. 2020;61(1):18-29. doi: 10.1111/jcpp.13094 [DOI] [PubMed] [Google Scholar]
- 22.Caro P, Turner W, Caldwell DM, Macdonald G. Comparative effectiveness of psychological interventions for treating the psychological consequences of sexual abuse in children and adolescents: a network meta-analysis. Cochrane Database Syst Rev. 2023;6(6):CD013361. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Cuijpers P, van Straten A, Bohlmeijer E, Hollon SD, Andersson G. The effects of psychotherapy for adult depression are overestimated: a meta-analysis of study quality and effect size. Psychol Med. 2010;40(2):211-223. doi: 10.1017/S0033291709006114 [DOI] [PubMed] [Google Scholar]
- 24.Hutton B, Salanti G, Caldwell DM, et al. The PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care interventions: checklist and explanations. Ann Intern Med. 2015;162(11):777-784. doi: 10.7326/M14-2385 [DOI] [PubMed] [Google Scholar]
- 25.Hoppen TH, Meiser-Stedman R, Jensen TK, Birkeland MS, Morina N. Efficacy of psychological interventions for post-traumatic stress disorder in children and adolescents exposed to single versus multiple traumas: meta-analysis of randomised controlled trials. Br J Psychiatry. 2023;222(5):196-203. doi: 10.1192/bjp.2023.24 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Higgins JPT, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions. Version 5.0.2. Cochrane; 2009. [Google Scholar]
- 27.Chambless DL, Hollon SD. Defining empirically supported therapies. J Consult Clin Psychol. 1998;66(1):7-18. doi: 10.1037/0022-006X.66.1.7 [DOI] [PubMed] [Google Scholar]
- 28.Rosner R, Rimane E, Frick U, et al. Effect of developmentally adapted cognitive processing therapy for youth with symptoms of posttraumatic stress disorder after childhood sexual and physical abuse: a randomized clinical trial. JAMA Psychiatry. 2019;76(5):484-491. doi: 10.1001/jamapsychiatry.2018.4349 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Roque-Lopez S, Llanez-Anaya E, Álvarez-López MJ, et al. Mental health benefits of a 1-week intensive multimodal group program for adolescents with multiple adverse childhood experiences. Child Abuse Negl. 2021;122:105349. doi: 10.1016/j.chiabu.2021.105349 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Shechtman Z, Mor M. Groups for children and adolescents with trauma-related symptoms: outcomes and processes. Int J Group Psychother. 2010;60(2):221-244. doi: 10.1521/ijgp.2010.60.2.221 [DOI] [PubMed] [Google Scholar]
- 31.Hoppen TH, Jehn M, Holling H, Mutz J, Kip A, Morina N. The efficacy and acceptability of psychological interventions for adult PTSD: a network and pairwise meta-analysis of randomized controlled trials. J Consult Clin Psychol. 2023;91(8):445-461. doi: 10.1037/ccp0000809 [DOI] [PubMed] [Google Scholar]
- 32.Borenstein M, Hedges LV, Higgins JPT, Rothstein HR. Introduction to Meta-Analysis. John Wiley & Sons; 2011. [Google Scholar]
- 33.R Core Team . R: a language and environment for statistical computing. http://www.R-project.org/.
- 34.Balduzzi S, Rücker G, Nikolakopoulou A, et al. netmeta: An r package for network meta-analysis using frequentist methods. J Stat Softw. 2023;106(2):1-40. doi: 10.18637/jss.v106.i0237138589 [DOI] [Google Scholar]
- 35.Lipsey MW, Wilson DB. Practical Meta-Analysis. Sage Publications; 2009. [Google Scholar]
- 36.Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Taylor and Francis; 2013. doi: 10.4324/9780203771587 [DOI] [Google Scholar]
- 37.Hoppen TH, Meiser-Stedman R, Kip A, Birkeland MS, Morina N. The efficacy of psychological interventions for adult post-traumatic stress disorder following exposure to single versus multiple traumatic events: a meta-analysis of randomised controlled trials. Lancet Psychiatry. 2024;11(2):112-122. doi: 10.1016/S2215-0366(23)00373-5 [DOI] [PubMed] [Google Scholar]
- 38.Donegan S, Williamson P, D’Alessandro U, Tudur Smith C. Assessing key assumptions of network meta-analysis: a review of methods. Res Synth Methods. 2013;4(4):291-323. doi: 10.1002/jrsm.1085 [DOI] [PubMed] [Google Scholar]
- 39.Balduzzi S, Rücker G, Nikolakopoulou A, et al. netmeta: An R package for network meta-analysis using frequentist methods. J Stat Softw. 2023;106(2):1-40. doi: 10.18637/jss.v106.i0237138589 [DOI] [Google Scholar]
- 40.Dias S, Welton NJ, Caldwell DM, Ades AE. Checking consistency in mixed treatment comparison meta-analysis. Stat Med. 2010;29(7-8):932-944. doi: 10.1002/sim.3767 [DOI] [PubMed] [Google Scholar]
- 41.Krahn U, Binder H, König J. A graphical tool for locating inconsistency in network meta-analyses. BMC Med Res Methodol. 2013;13(1):35. doi: 10.1186/1471-2288-13-35 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Tabachnick BG, Fidell LS. Using Multivariate Statistics. 6th edition. Pearson; 2013. [Google Scholar]
- 43.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. doi: 10.1136/bmj.315.7109.629 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Chaimani A, Salanti G. Using network meta-analysis to evaluate the existence of small-study effects in a network of interventions. Res Synth Methods. 2012;3(2):161-176. doi: 10.1002/jrsm.57 [DOI] [PubMed] [Google Scholar]
- 45.DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7(3):177-188. doi: 10.1016/0197-2456(86)90046-2 [DOI] [PubMed] [Google Scholar]
- 46.Schauer E. Trauma Treatment for Children in War: Build-up of an Evidence-Based Large-Scale Mental Health Intervention in North-Eastern Sri Lanka. Dissertation. University of Konstanz; 2008. http://nbn-resolving.de/urn:nbn:de:bsz:352-opus-54249 [Google Scholar]
- 47.Dorsey S, Lucid L, Martin P, et al. Effectiveness of task-shifted trauma-focused cognitive behavioral therapy for children who experienced parental death and posttraumatic stress in Kenya and Tanzania: a randomized clinical trial. JAMA Psychiatry. 2020;77(5):464-473. doi: 10.1001/jamapsychiatry.2019.4475 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Schottelkorb AA, Doumas DM, Garcia R. Treatment for childhood refugee trauma: a randomized, controlled trial. Int J Play Ther. 2012;21(2):57-73. doi: 10.1037/a0027430 [DOI] [Google Scholar]
- 49.Ruf M, Schauer M, Neuner F, Catani C, Schauer E, Elbert T. Narrative exposure therapy for 7- to 16-year-olds: a randomized controlled trial with traumatized refugee children. J Trauma Stress. 2010;23(4):437-445. doi: 10.1002/jts.20548 [DOI] [PubMed] [Google Scholar]
- 50.McMullen J, O’Callaghan P, Shannon C, Black A, Eakin J. Group trauma-focused cognitive-behavioural therapy with former child soldiers and other war-affected boys in the DR Congo: a randomised controlled trial. J Child Psychol Psychiatry. 2013;54(11):1231-1241. doi: 10.1111/jcpp.12094 [DOI] [PubMed] [Google Scholar]
- 51.Barron I, Freitas F, Bosch CA. Pilot randomized control trial: Efficacy of a group-based psychosocial program for youth with PTSD in the Brazilian favelas. J Child Adolesc Trauma. 2020;14(3):335-345. doi: 10.1007/s40653-020-00328-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.John-Baptiste Bastien R, Jongsma HE, Kabadayi M, Billings J. The effectiveness of psychological interventions for post-traumatic stress disorder in children, adolescents and young adults: a systematic review and meta-analysis. Psychol Med. 2020;50(10):1598-1612. doi: 10.1017/S0033291720002007 [DOI] [PubMed] [Google Scholar]
- 53.Hoppen TH, Morina N. Is high-quality of trials associated with lower treatment efficacy? a meta-analysis on the association between study quality and effect sizes of psychological interventions for pediatric PTSD. Clin Psychol Rev. 2020;78:101855. doi: 10.1016/j.cpr.2020.101855 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eAppendix 1. Search string used for systematic literature search
eAppendix 2. References of screened reviews as part of the systematic literature search
eAppendix 3. Quality criteria for risk of bias assessment
eAppendix 4. Quality coding of included trials
eAppendix 5. Categorization of interventions and control conditions
eAppendix 6. Trial characteristics of included trials
eAppendix 7. References of studies included in the present network meta-analysis
eAppendix 8. Trial and sample characteristics across comparison dyads
eAppendix 9. Short-term outcomes: Forest plot compared to passive control conditions
eAppendix 10. Short-term outcomes: Forest plot compared to active control conditions
eAppendix 11. Short-term outcomes: Net splitting results
eAppendix 12. Short-term outcomes: Net heat plot
eAppendix 13. Short-term outcomes: Funnel plot
eAppendix 14. Short-term outcomes: Outlier-adjusted results
eAppendix 15. Mid-term outcomes: Forest plot compared to passive control conditions
eAppendix 16. Mid-term outcomes: Forest plot compared to active control conditions
eAppendix 17. Mid-term outcomes: Funnel plot
eAppendix 18. Mid-term outcomes: Net splitting results
eAppendix 19. Mid-term outcomes: Net heat plot before (top) and after (bottom) exclusion of MDTs
eAppendix 20. Mid-term outcomes: MDTs excluded due to detected inconsistency
eAppendix 21. Long-term outcomes: Forest plot compared to passive control conditions
eAppendix 22. Long-term outcomes: Forest plot compared to active control conditions
eAppendix 23. Long-term outcomes: Net splitting results
eAppendix 24. Long-term outcomes: Net heat plot
eAppendix 25. Long-term outcomes: Funnel plot
Data sharing statement
