Anxiety and depression are the most common mental health difficulties experienced by children and young people (CYP) worldwide, yet many struggle to access timely treatment (Westin, 2014). This is a significant issue in the UK, with demand for child and adolescent mental health services surpassing service capacity. This is further exacerbated by limited funding and sufficiently qualified staff (Holmes et al., 2018), which can make it difficult to imagine how this situation will improve anytime soon.
Internet-delivered interventions could provide a solution, as they are designed to encourage positive behaviour change with limited clinical support and are delivered via websites, apps, wearable technology, and more (Hollis et al., 2017) – the possibilities are endless! Interventions delivered via the internet have been shown to be as effective as face-to-face treatments for addressing anxiety and depression symptoms in adults (Richards et al., 2020), and are considered scalable and personalisable for CYP (Hollis et al., 2017). Plus, we all know how CYP typically consume technology at a rate of knots (see our blog by Sarah Hetrick on digital technology and youth mental health to learn more). However, far less is known about the effectiveness of internet-delivered interventions when designed to address anxiety and depression symptoms in CYP.
There is therefore a gap in our knowledge, which the current blog is all about! Nora Eilert and colleagues conducted a meta-analysis, with two aims:
- What existing evidence looks at this topic, and is it good?
- How effective are internet-delivered interventions for treating symptoms of anxiety and depression in children and young people?
The authors searched 3 electronic databases in November 2020 for studies of low intensity interventions for depression or anxiety which were delivered remotely, either directly to CYP or via their parents/carers.
Initial searches found 1,014 articles. After one author excluded any “entirely off-topic studies”, two authors independently double-screened the remaining studies and extracted relevant data. Studies were assessed for quality using the Checklist to Evaluate a Report of a Nonpharmacological Trial (CLEAR NPT).
Intervention effectiveness was represented by effect sizes (Hedges’ g). Random effects models were used in the meta-analysis, meaning that the authors combined the estimates of treatment effects from each study to produce an estimate of overall treatment effect.
What existing evidence looks at this topic, and is it good?
23 studies were included in the systematic review: 16 randomised controlled trials (RCTs) and 7 non-RCTs. Most studies (n = 21) were conducted in the West, with one study based in Iran and 1 in China. Studies included a total of 6,981 participants, aged between 3 and 21 years old (sample means < 18).
Internet-delivered Cognitive Behavioural Therapy (iCBT) was the most popular intervention approach used (n = 12), but cognitive/attentional bias modification (n = 3), problem-solving therapy (n = 1), psychodynamic therapy (n = 1) and a spiritually-informed intervention (n = 1) were also used. Control groups differed in nature and included waitlist (n = 8), placebo (n = 2), face-to-face (n = 2) and other non-internet delivered interventions (n = 4). Study quality was adequate or satisfactory for most articles.
How effective are internet-delivered interventions for treating symptoms of anxiety and depression in children and young people?
- Anxiety: Findings from 15 RCTs (8 anxiety-focused interventions, 3 depression-focused, and 4 transdiagnostic), suggested a small, significant improvement in anxiety symptoms for those receiving internet-delivered interventions (g = -0.25, 95% CI [-0.38 to -o.12], p <0.001), with moderate heterogeneity (I2 = 41.4%).
- Depression: Findings from 10 RCTs (4 depression-focused interventions, 2 anxiety-focused, and 4 transdiagnostic) suggested a small, but non-significant improvement in depression symptoms for those receiving internet-delivered interventions (g = -0.27, 95% CI [-0.55 to 0.01], p = 0.06) with moderate (bordering on high) heterogeneity (I2 = 71.4%).
- Impaired functioning: Analysis of 7 RCTs showed a moderate effect on impaired functioning after treatment (g = 0.52, 95% CI [0.24 to 0.80], p <0.001), demonstrating improved scores on clinical assessments of psychological and social functioning.
- Quality of life: Analysis of 5 RCTs showed no significant effect of treatment on quality of life scores (g = -0.01, 95% CI [-0.23 to 0.21], p = 0.94), but this should be interpreted with caution as fewer than half of the included studies measured this.
- Follow-up outcomes: No significant long-term effects were found for anxiety (g = -0.17, 95% CI [-0.58 to 0.24], p = .42) or depression (g = -0.18, 95% CI [-0.39 to 0.03], p = 0.09) symptoms.
In a nice shiny nutshell, the authors found:
- 16 studies that addressed their research question and met the criteria for inclusion in a meta-analysis.
- A small, significant effect for internet-delivered interventions in treating anxiety in children and young people when compared to control groups.
- A small but non-significant effect for internet-delivered interventions in treating depression in children and young people when compared to control groups.
Overall, the authors conclude that the results are “in line, if a little pessimistic” compared to previous meta-analyses of similar topics, and that “the results of the study do not encourage the effectiveness of digitally delivered interventions for treating the symptoms of anxiety and depression in children and young people” due to small effect sizes and no evidence of longer-term effects. Internet-delivered interventions for this population could be beneficial, but further research is needed.
Strengths and limitations
Given the current state of CYP’s access to mental health support, it is clear that this review was timely, necessary and had many strengths, including:
- The use of a narrow search criteria, which allowed the authors to address knowledge gaps left by previous meta-analyses which were broader and included CD-ROMs, SMS messaging and video-conferencing.
- Registering the review on PROSPERO and reporting any deviations from the published protocol to ensure reliability and minimise bias.
- Using random effects models, which allow for between-study heterogeneity to be tested. This is important given the variation in individual study aspects (e.g., populations, interventions, theoretical approach, length of time to follow up).
However, there were also some limitations:
- Differences in aspects such as control groups, age ranges, and the amount of support offered within interventions introduced a “considerable amount of heterogeneity” to the study. Did the meta-analysis measure like for like, or was this a case of comparing apples and oranges? Some argue that the I2 is not a sensitive enough measure to reliably account for all of these differences (Borenstein et al., 2017).
- Most included studies were from Western countries. Could there be cultural effects on how effective or feasible internet-delivered interventions are (e.g., for less affluent, more remote populations)? This requires further investigation.
- Some studies had very high drop-out rates (e.g., in one study, 22% of participants didn’t even complete the first of 10 modules), introducing issues such as attrition bias. Who were the participants who dropped out, and does this impact intervention effectiveness?
Implications for practice
Intuitively, it makes sense for researchers to be investigating how effective digital mental health support is for a technology-savvy generation. Long waiting lists show that the need for more accessible interventions is there, but these findings indicate that current internet-delivered interventions may not be effective enough to meet this need.
Given that a small but significant effect was found for anxiety, it may be that future research could focus on iCBT to identify what the effective components of this intervention are and what is needed to implement these techniques in practice (e.g., amount of parental support required, number of modules, etc.). Qualitative research and co-design methods (where stakeholders themselves help to create what the intervention looks like – read more about co-production on The Mental Elf website) could explore and build on service user’s views around how to make interventions effective and engaging for CYP.
Interestingly, new NICE guidelines for clinicians have been published emphasising the usefulness of various internet-delivered interventions for CYP. This may seem in contradiction to the findings of this meta-analysis, but it could suggest that there is a middle ground where face-to-face interventions are the ‘gold standard’ for high-intensity needs and internet-delivered interventions could support those on waiting lists for low-intensity needs. From a service perspective, it would be good to see clinicians armed with different solutions to help them work through their waiting lists to provide timely access to mental health support for all.
Statement of interests
I am in the fortunate position of currently working on a large scale randomised controlled trial looking at ways of identifying and minimising anxiety in primary school children and the intervention we use is an internet-delivered intervention to parents using CBT. This intervention did not form part of the meta-analysis.
Eilert, N., Wogan, R., Leen, A., and Richards, D. (2022). Internet-Delivered Interventions for Depression and Anxiety Symptoms in Children and Young People: Systematic Review and Meta-analysis. JMIR Pediatrics and Parenting, 5(2) 1-20.
Borenstein, M., Higgins, J., Hedges, L.V. and Rothstein, H.R. (2017). Basics of meta-analysis: 12 is not an absolute measure of heterogeneity. Research Synthesis Methods, 8(1) 5-18.
Hetrick, S. Digital technology and youth mental health: recommendations from the Royal College of Psychiatrists. The Mental Elf, 16 April 2020.
Hollis, C., Falconer, C.J., Martin, J.L., Whittington, C., Stockton, S., Glazebrook, C., et al. (2017). Annual Research Review: Digital health interventions for children and young people with mental health problems – a systematic and meta-review. The Journal of Child Psychology and Psychiatry, 58(4) 474-503.
Holmes, E.A., Ghaderi, A., Harmer, C.J., Ramchandani, P.G., Cuijpers, P., Morrison, A.P., et al. (2018). The Lancet Psychiatry Commission on psychological treatments research in tomorrow’s science. Lancet Psychiatry, 5(3) 237-286.
National Institute for Health and Care Excellence. Digital mental health tech for children and young people recommended by NICE in first rapid healthtech guidance. 8 February 2023.
Richards, D., Enrique, A., Eilert, N., Franklin, M., Palacios, J., Duffy, D., et al. (2020) A pragmatic randomized waitlist-controlled effectiveness and cost-effectiveness trial of digital interventions for depression and anxiety. NPJ Digital Medicine, 3(1) 1-10.
Westin, A.M., Barksdale, C.L. and Stephan, S. H. (2014) The effect of waiting time on youth engagement to evidence based treatments. Community Mental Health Journal, 50(2) 221-228.
- Photo by Annie Spratt on Unsplash
- Photo by Reba Spike on Unsplash
- Photo by Mimi Thian on Unsplash
- Photo by Brands&People on Unsplash
- Photo by Kristina Tripkovic on Unsplash
- Photo by Windows on Unsplash
Thanks for writing an interesting and accessible summary of the key issues and findings, it was an enjoyable read. I definitely do believe there is a place for internet-based interventions but they will not suit everyone, just as face-to-face is not acceptable or accessible for others. Hopefully there will be more research into the possibilities moving forward, taking into account young people’s views and preferences!