Comparison the effects of EMDR and trauma-focused art therapy practices on post-traumatic stress levels in children and adolescents: Systematic review and meta-analysis

This research aimed to compare the effects of Eye Movement Desensitization and Reprocessing (EMDR) and Trauma-Focused Art Therapy (TF-ST) applications on children and adolescents showing symptoms of post-traumatic stress or diagnosed with Post-Traumatic Stress Disorder (PTSD). In addition, in examining the effects of both therapy methods, the differences in effect sizes created by the characteristics of the studies or participants (age group, gender, duration and number of sessions of therapy, involvement of family members, country where the study was conducted, number of traumatic experiences that the participants were exposed to, dropout rates in the study, selection of the control group, whether the study was conducted as individual or group therapy, and the time elapsed since the traumatic event) were conducted using Metaregression analyses. Effect size calculation was performed using the standardized mean difference (Hedges' g); analysis of heterogeneous distributions was conducted using the Random Effects Model. For publication bias, Trim and Fill, Funnel Plot, Begg rank correlation test, Egger regression test, and Rosenthal and Orwin fail-safe numbers were examined. The meta packages in R and R Studio v.4.3.3 were used. First, 1516 records were reached in the electronic database review. Within the scope of the determined inclusion criteria, 17 studies (NEMDR = 11, NTO-ST = 6) were reached at the end of the research. When the general effect size was examined over the mean differences before and after treatment, it was observed that both methods were marginally significant in the mentioned clinical population; however, the effect size was in favor of Art Therapy (for EMDR Hedges' g = .68, 95% CI [-.03; 1.41], for Art Therapy p = .06; Hedges' g = 1.19, 95% CI [-.20; 2.58], p = .09). Potential risks of publication bias were discussed using fail-safe numbers (for EMDR Rosenthal fail-safe N = 177, Orwin fail-safe N = 161; for Art Therapy Rosenthal fail-safe N = 222, Orwin fail-safe N = 150), Begg rank order correlation test (for EMDR Kendall's tau = .27, p = .28; for Art Therapy Kendall's tau = .33, p = .46), and Egger regression test (for EMDR t(9) = 4.04, p<.001; for Art Therapy t(4) = 4.16, p<.001). Finally, in the moderator analyses, the number of traumas had a significant moderator effect for both therapy methods (for EMDR B = -.29, p = .08; for Art Therapy B = .05, p<.001), while the dropout rate was found to be significant for Art Therapy (B = .02, p<.001); however, no significant moderator effect was found for other variables. The obtained data were compared with other studies in the literature, and possible reasons for the greater effect of Art Therapy compared to EMDR were discussed. Factors such as neurobiological effect differences, youths' dominance in artistic creations, framing Art Therapy in an appropriate way for children's developmental periods, and the presence of concrete artistic productions explain the greater effect size of Art Therapy. It is observed that this study has a pioneering position since EMDR and Art Therapy were compared firstly in the trauma literature through randomized controlled studies.