Promoting wellbeing in children and young people is a key priority for governments and health services across the world (e.g. NHSE, 2019). Particularly after two years of disrupted schooling, and possibly more to come in the future.
However, this is a daunting challenge. In 2017/18, approximately 30% of children and young people with a mental health condition received NHS care (NHSE, 2019). This is by no means just a UK problem. Sufficient capacity to support children and young people with mental health conditions is a challenge internationally.
Therefore, it is important that we prioritise the delivery of efficacious and acceptable interventions for children and adolescents with mental disorders. Correll et al. (2021) conducted an umbrella review to summarise this evidence. This is a monumental task, they aimed to assess 48 medications, 20 different psychosocial interventions, and four brain stimulation interventions. They summarised evidence for these interventions across 52 mental disorders!
How to cover such a huge amount of ground? For the last 20 years systematic reviews (comprehensive and transparent summaries of evidence) have become a necessity in making sense of the growth in evidence based mental health research. But keeping up with the ever-increasing number of systematic reviews has itself become unmanageable. Hence the rise of umbrella reviews; systematic reviews of systematic reviews. In other words, comprehensive and transparent summaries of systematic reviews.
The authors used standard methods for conducting umbrella reviews. They searched for network meta-analyses (NMAs) and meta-analyses (MAs) which mathematically combine data from different studies. The authors searched well known databases (Pubmed, PsycINFO, Cochrane Library).
The inclusion criteria were NMAs or MAs of RCTs of a priori defined 48 psychotropic medications, 20 psychosocial interventions and four brain stimulation interventions (see below) in children / adolescents with any of 52 a priori defined mental disorders (see below). Studies were excluded if they: i) did not include a meta-analysis ii) pooled studies other than RCTs or iii) focussed on interventions for other than the pre-defined disorders / outcomes.
52 mental disorders were eligible for inclusion: including neurodevelopmental disorders (e.g. autism, attention-deficit/hyperactivity disorder (ADHD), schizophrenia and other primary psychotic disorders (e.g. schizophrenia), mood disorders (e.g. depression) and anxiety disorders.
A comprehensive range of medications such as antidepressants, antipsychotics and anti-ADHD medications were included. As were a broad range of psychosocial interventions like cognitive behavioural therapy, problem solving, family-based therapy and psychodynamic psychotherapy. The brain stimulation techniques included were transcranial magnetic stimulation, transcranial direct current stimulation, electroconvulsive therapy, and neurofeedback.
14 network meta-analyses (NMAs) and 90 meta-analyses (MAs) were included in the umbrella review. These reviews provided data on intervention effectiveness and acceptability for 15 mental disorders.
With a review this size and breadth it’s always challenging to summarise the results. Some of the key findings are summarised below.
For depression, a number of psychological interventions were more effective than control (based on 70 RCTs and 8,906 participants) including:
- Interpersonal therapy (SMD -1.37, 95% CI -2.04 to -0.70)
- Problem solving (SMD -1.26, 95% CI -2.48 to -0.03)
- Family therapy (SMD -1.03, 95% CI -1.66 to -0.4)
- Cognitive behavioural therapy (SMD -0.94, 95% CI -1.40 to -0.48).
Similar patterns of data were found across a variety of conditions. For example, CBT or behavioural interventions were effective for children and adolescents in several other conditions like:
- Anxiety disorder
- Social anxiety disorder
- Post-traumatic stress disorder (PTSD)
- Obsessive compulsive disorder (OCD)
Of all the antidepressants (also based on 70 RCTs and 8,906 participants):
- Fluoxetine was the only medication with evidence of effectiveness compared with placebo for reducing depression symptoms (SMD -0.51, 95% CI -0.84 to -0.18)
- People receiving nefazodone, duloxetine, or fluoxetine were more likely to respond to depression treatment compared with placebo
- Selective serotonin reuptake inhibitors (SSRIs) were also effective in other conditions including OCD, and anxiety disorders.
For antipsychotic medication in children and adolescents with bipolar disorder:
- There was limited support for the use of quetiapine compared with placebo for reducing depression symptoms in children and adolescents with bipolar disorder (SMD -0.10, 95% CI -0.32 to 0.13; 2 RCTs and 224 participants)
- Aripiprazole was associated with an improvement in symptoms during a manic episode (SMD -1.08, 95% CI -1.32 to -0.85; 2 RCTs and 339 participants)
There was also evidence for combination interventions to reduce symptoms of depression:
- Fluoxetine and CBT were more effective than placebo (SMD -0.73, 95% CI -1.39 to -0.07; 70 RCTs and 8,906 participants)
The authors drew extensive conclusions. We cannot summarise these comprehensively here, but the paper is open access if you want to see the detail. They identified the most convincing evidence for nine conditions including:
- The combination of fluoxetine and CBT (Cognitive behavioural therapy)
- Interpersonal therapy
- Behavioural therapy with exposure and response prevention
Strengths and limitations
A clear strength of the review is the synthesis of a very broad scope of mental disorders. This is a great place to start if you are unfamiliar with the evidence in this area.
Another strength is that the methods used are logical and appropriate. An umbrella review is the most practical method for synthesising such a large topic area.
But the key limitation of umbrella reviews is that they depend on published systematic reviews providing an up-to-date picture of the evidence. In some topic areas, the systematic reviews may not reflect current knowledge.
For example, our recent Cochrane review on new generation antidepressants in children and adolescents (Hetrick et al, 2021) suggests more recent evidence no longer supports the view that fluoxetine is the only effective pharmacological intervention in this population.
Evidence for psychological interventions in children and adolescents with PTSD was based on a single review published in 2013. Similarly, data on treatment of tic disorders, are largely based on a review published in 2009. This raises questions on whether more recent systematic reviews would come to different conclusions.
This isn’t a criticism of the how the umbrella review was conducted. Only that where there is an absence of recent systematic reviews, this may impact on the validity of conclusions.
A further limitation of the review was that authors did not make available a pre-existing protocol and search strategy. However, further information is available through contacting the authors directly.
This level of transparency is usually an expectation of high-quality reviews, all the more important in a review of this complexity and breadth where subtle decisions could impact on conclusions.
Implications for practice
This umbrella review provides a nice summary of a large body of evidence. The review is informative for those who are unfamiliar with the children and adolescent mental disorders literature. But it is also informative for those who have detailed knowledge of the data in some mental disorders who may have gaps in knowledge for other disorders. Or those who just want to check their knowledge of the evidence is up to date.
Statement of interests
Nick Meader is a co-author of a recent network meta-analysis on the effectiveness of new generation antidepressants in children and young people (Hetrick et al 2021). He is also the Deputy Coordinating Editor for the Cochrane Common Mental Disorders group who were responsible for editorial decisions on the publication of various systematic reviews included in the umbrella review.
Correll CU, Cortese S, Croatto G. et al (2021) Efficacy and acceptability of pharmacological, psychosocial, and brain stimulation interventions in children and adolescents with mental disorders: an umbrella review. World Psychiatry 2021 20 244-275.
Hetrick SE, McKenzie JE, Bailey AP. et al (2021) New generation antidepressants for depression in children and adolescents: a network meta-analysis. Cochrane Database of Systematic Reviews 2021 5(5) CD013674.
NHS England. The NHS Long Term Plan (PDF).