Internet‐delivered cognitive behavior therapy with minimal therapist support for anxious children and adolescents: predictors of response.

Matt Kempen
Marketing Manager for ACAMH

Posted on

Internet‐delivered cognitive behavior therapy with minimal therapist support for anxious children and adolescents: predictors of response. A video abstract from Professor Susan Spence on her JCPP paper.

Authors; Susan H. Spence, Samantha J. Prosser, Sonja March, Caroline L. Donovan
First published:19 May 2020


In general, Internet‐delivered cognitive behavior therapy (iCBT) produces significant reductions in child and adolescent anxiety, but a proportion of participants continue to show clinical levels of anxiety after treatment. It is important to identify demographic, clinical, and family factors that predict who is most likely to benefit from iCBT in order to better tailor treatment to individual needs.

Read The Bridge research digest on this paper

Professor Sue Spence
Professor Sue Spence

Emeritus Professor Sue Spence has a BSc (Hons – 1st class) and a PhD in Clinical Psychology. She is currently based in the Australian Institute of Suicide Research and Prevention, and the School of Applied Psychology at Griffith University and also holds the position of Honorary Professor at the University of Queensland. Her research focusses on the etiology, assessment, treatment and prevention of anxiety and depression in young people. She has received over $10.5 million in research grants during her career and has managed numerous randomised controlled trials to successful completion. She is the author of 8 books, 37 book chapters and over 130 research articles in refereed journals. Her research is highly influential, with over 13,000 citations, with publications in top international journals in her field. She places strong emphasis on translation of her research into clinical practice, being the author of the internationally used Spence Children’s Anxiety Scale and one of the founders of the BraveOnline treatment program. Professor Spence is a member of several editorial boards and is regularly invited to present keynote addresses at international conferences. Her work has been recognised through the awards of Officer of the Order of Australia, Fellowship of the Australian Psychological Society, British Psychological Society; the Academy of the Social Sciences of Australia and the Academy of Cognitive Therapy. (bio via Griffith University)


This is Dr. Sue Spence, Professor Emeritus at Griffith University, Brisbane, Australia. And I’m going to be presenting a summary of an article recently published in the Journal of Child Psychology and Psychiatry entitled, ‘Internet-delivered cognitive behaviour therapy with minimal therapist support for anxious children and adolescents. Predictors of response’. And I’d like to acknowledge my co-authors, Samantha Prosser, Sonja March and Caroline Donovan.

In terms of background, delivering CBT using the internet rather than face-to-face offers a way of increasing access to treatment for many young people with mental health problems who would otherwise not receive the help that they need. This is particularly topical at the moment, with the Covid-19 virus not only impacting upon the mental well-being of young people but also restricting their opportunities for receiving clinic-based treatment. The option of effective internet-delivered CBT, or what I’m going to call ICBT, for youth mental health is therefore more important than ever.

There are now several control trials showing that ICBT with minimal therapist support can produce significant improvements in anxiety for children and adolescents presenting with anxiety disorders. This includes four RCTs conducted along with my colleagues from the BRAVE-Online programme. However, we can’t assume that ICBT is suitable for all clinically anxious children. And it’s important that we identify the characteristics of those who respond well versus those who do not. This study aims to address this issue.

We followed up 175 clinically anxious children and adolescents aged seven to 18 years for 12 months after they completed BRAVE-Online programme. They all had a diagnosis of either separation, social, generalised anxiety disorder or a specific phobia. On average they had around three anxiety disorder diagnoses, with a clinician severity rating in the moderate-to-severe range. In general they came from middle- to high-income families with relatively well-educated parents, tended to speak English at home and had been born in Australia. All children were required to have access to a computer and the internet at home.

The BRAVE-Online programme involves ten online sessions once per week, each lasting about 40 minutes. There were two versions of the programme. One for children aged seven to 12 years and one for teenagers, so that the graphics and wording is age-appropriate. Their parents also completed sessions. There’s six for the parents of children or five for the parents of the teens.

The children had no direct contact with the therapist, other than short weekly e-mails providing feedback and encouragement about their progress. Plus a 30-minute phone call midway through treatment to help develop the exposure hierarchy.

The CBT content included psychoeducation about anxiety and skills for relaxation, problem-solving, helpful thinking and graduated exposure to feared situations.

We examine predictors of response to treatment, measured using child and parent report of their anxiety symptoms, using the Spence children’s anxiety scale. Anxiety was assessed at pre- and post-treatment and six and 12-month follow-ups. The data analysis used multi-level modelling to examine factors that predicted the degree of change in anxiety over time. All analyses controlled for baseline clinician severity ratings.

The predictors that we examined related firstly to demographic characteristics, such as the child age and gender and family characteristics, such as their income, parental age and education. Then we looked at clinical characteristics of the presenting anxiety problem, such as the type of disorder, comorbidity with depression, number of anxiety disorders and anxiety severity. And for family characteristics, we looked at parental mental health, parenting style, family functioning and the couple relationship quality.

When it came to the results, basically we found very few factors that predicted response to treatment. The majority of the children responded well to the programme. They generally completed the sessions and showed strong reductions in anxiety. In terms of demographic factors, only having an older age of mother predicted poorer change in anxiety. Child age, gender, father age, family income and parent education did not.

For clinical factors, comorbidity with depression, greater number of anxiety disorders, greater severity of anxiety, did not predict a worse outcome, contrary to our predictions. For family factors, again contrary to predictions, parental mental health, parenting style and family functioning did not predict poorer treatment response. But the children whose parents experienced poor couple relationship quality tended to show lower reductions in anxiety over time.

Turning now to the discussion. While we note that the majority of children did respond well to ICBT for child anxiety, irrespective of most of the child, family and clinical characteristics that we examined. And there are some implications then for treatment in the future. For example, for children whose parents are experiencing couple conflict, then perhaps additional parent-focused intervention may be needed or a face-to-face family-focused approach. In terms of older mothers, we don’t know why older mothers tended to have children who experienced weaker response to treatment. We need to explore this further. Perhaps these older mothers were working more hours and therefore had less time to spend with their children on the programme. Or perhaps they don’t have such strong skills for use of the online programme.

It’s important that we consider the limitations of the study. We need to note that the sample that we included here tended to be families with relatively well-educated parents with above average incomes. So in future research we need to examine whether these results would apply with other groups of children, for example, those from disadvantaged backgrounds or from other cultural groups.

And it’s still important that we continue to investigate which children do respond best to ICBT. And there may be other factors that we didn’t examine in this study that will turn out to be important predictors. Because it’s important that we identify who those children are for whom other treatment methods might be more appropriate rather than ICBT.

Thank you for listening to my presentation.

Add a comment

Your email address will not be published. Required fields are marked *