With the exception of the study in Indonesia [
4], the randomized trials on CBI did not show a significant treatment effect for depression, PTSD, or anxiety. The simple conclusion that can be drawn from these findings is that the widespread use of CBI is not justified, at least not for the treatment of trauma-related disorders. The group was very careful to analyze differential effects for several subgroups. However, these effects were inconsistent across studies. In Indonesia, positive effects on PTSD symptoms were found for girls, yet not for boys and no effects were found on symptoms of depression and anxiety. In Sri Lanka [
7], boys and children with less ongoing trauma-exposure benefited with regard to PTSD and anxiety symptoms, yet no effects were found on symptoms of depression. In Burundi [
6], children living with both parents benefited from CBI with regard to symptoms of PTSD and depression and symptoms of depression were reduced for those living in larger households. This pattern of inconsistent results has several implications. First of all, there are contextual and individual moderators that determine whether a specific child will benefit from treatment or not. These results question the use of CBI as a universal (one size fits all) intervention/prevention approach for all children. Secondly, the moderators are inconsistent across studies and all of the findings are
post-hoc. It is, therefore, not possible for differential recommendations for the use of the intervention to be made, that is, a prediction about who will benefit from the treatment, in which context and, therefore, who should receive CBI and who should not. Thirdly, it is important to note that most trials had an overall zero effect on symptoms. The mere presence of significant moderators where there was an average zero-change is a strong indicator of the so-called deterioration effect [
9]. The deterioration effect relates to the fact that the average trajectory of study participants in a trial does not necessarily inform about potential harmful effects of the intervention. It is possible, and not improbable, that some subjects benefit from a treatment while others do get worse, which may still result in an overall positive (or null) effect of the intervention. However,
primum non nocere (first do no harm) must be the major ethical principle of any treatment, and all interventions, especially those with small overall-benefits, must show that they are not harmful for some individuals. Unfortunately, there is no information on clinically significant worsening or impaired recovery in the CBI trials, but Tol
et al. [
7] report that Sri Lankan girls with CBI were doing worse than girls in the waiting list condition in terms of their change in PTSD symptoms. Moreover, children dropping out of CBI in Burundi had significantly higher levels of PTSD symptoms [
6]. The negative effect of CBI is probably not dramatic given the small or absent overall effect. However, it appears that those who suffer the most in terms of symptoms of mental health disorders benefit the least, or may even deteriorate through CBI, which questions the use of this intervention as a treatment for trauma-related disorders.