Preventing anxiety and low mood with school-based CBT

Anxiety and depressive disorders in children are common. The British Mental Health survey estimated that 3.7% of 5-15 year-olds met the diagnostic criteria for an anxiety disorder and 1% had a depressive disorder. Co-morbidity between anxiety and depression is common and around 15-18% of children will have experienced an impairing anxiety or depression by the age of 16-17 years old.

The prevalence of emotional problems, which often can persist into adulthood and the subsequent impact on individuals’ lives, is a major public health issue. The difficulty is that many children who experience anxiety and depression will not access evidence based treatment or specialist services. Very few will receive any form of help.

A recently published NIHR-funded study attempts to address this problem by focusing on the delivery of a prevention programme delivered in schools.

Around 1 in 6 young people will have experienced an impairing anxiety or depression before age 18.

Methods

The study employed a pragmatic, cluster three arm RCT (randomised controlled trial) design.

The three arms were:

FRIENDS delivered by health staff
FRIENDS delivered by school staff
The usual school curriculum (PSHE).

Break down of study arm and delivery

40 schools in the south of England (within a 50 mile radius of Bath) were recruited and 1,448 children were eligible for the study, of which 1,392 consented to take part.
The only exclusion criteria was for children 9-10 years old who were not attending school or did not participate in PSHE for religious or other reasons.
The FRIENDS intervention consisted of a manualised CBT programme delivered over 9 one-hour sessions. The programme focused on increasing emotional awareness, emotional regulation, problem solving and coping skills to reduce avoidance.
PHSE sessions were provided by the school and followed the UK national curriculum programme.
Outcome data was collected at 6, 12 and 24 months. The primary outcome was the Revised Child Anxiety and Depression Scale (RCADS)
Questionnaires were administered by researchers who were blind to the children’s treatment allocation.
The data analysis was assessed using an intent-to-treat analysis without the use of imputation. Pre-defined analysis was conducted with corrections for multiple analysis applied.
The authors also conducted an economic analysis, which compared each of the interventions and focused on the service usage (e.g. visits to a GP, outpatient appointment and A&E visits).

Results

All 9 FRIENDS sessions were delivered in both treatment arms in the allocated schools. Fidelity raters who were independent of the study rated them as good. However, this was based on a sample of just one single session.
In the Health-led FRIENDS sessions, 100% of the core tasks and homework were delivered. In the School-led FRIENDS, only 60% of the core task and homework were delivered.
Outcomes at 12 months showed a difference in adjusted mean RCADS scores between Health-led and School-led FRIENDS (Adjusted Difference -3.91, 95% Confidence Interval -6.48 to -1.35; p=0.0004)

There was also a difference between Health-led and School-led FRIENDS and usual school provision (Adjusted Difference -2.66, 95% Confidence Interval -5.22 to -0.09; p=0.043) at 12 months.
There were 12-month differences in the RCADs scores for generalised and social anxiety, but not for depression.
A pre-defined sub group analysis of high- and low-risk groups was also conducted at 12 months. Authors reported significant reductions in the high-risk group, but no difference between the Health-led and School-led FRIENDS interventions. There was a small treatment effect for both Health-led FRIENDS (Cohens d=0.22, 95% CI 0.38 to 0.07) and School-led FRIENDS (Cohens d= 0.25, 95% CI 0.40 to 0.11) in the low risk group.
Outcomes at 24 months were based on 43.6% of the original sample. The differences observed between interventions had diminished across all three groups and there were no between group differences in total anxiety. There was no difference between high- and low-risk groups at 24 months.

Children’s anxiety levels improved irrespective of the intervention that they received.

Conclusions

The authors concluded that the FRIENDS programme was acceptable to both children and school staff. It was also possible to integrate the FRIENDS programme successfully into the academic teaching timetable.
In the short-term, the effectiveness of the programme depended on who was delivering it with Heatlh-Led FRIENDS achieving greater reductions in anxiety at 12 months.
However at 24 months the between-group effects had reduced and there was no between-group effects present.
The authors reported no evidence of cost effectiveness for the universal delivery of the FRIENDS program over a 6-month period.
Economic analysis did not detect change in service usage of a 6-month period post-intervention, however this was based on a small subset of the overall sample and it could be argued that service usage is a low frequency event, which may mean a longer time period was needed to capture potential benefits.

This RCT does not support the universal provision of specific anxiety prevention programmes in UK primary schools.

Strengths and limitations

This study has several strengths; firstly the authors provided a clear and transparent report of the methodology, analysis strategy and findings and in addition pre-registered both the 12 and 24 months parts of the study. The authors also provided a power calculation to allow the reader to determine if the study was appropriately powered. This is a really great example of the authors showing their working. Others publishing intervention trials should follow this example.
In terms of addressing potential sources of bias, the authors address this in multiple ways. Independent assessors were used in the randomisation, data collection and rating of fidelity. It would have been good to know if the authors had been blind to the allocation until the analysis was complete, as this would have added another layer of rigour to reducing possible sources of bias.
Regarding limitations in the sample, this only included schools in a 50-mile radius of a city in southern England (Bath). It is not clear if the findings would be generalisable to other areas of the UK or a more urban environment (e.g. London). Also there was no assessment or control for the relative levels of social economic deprivation as these have often been cited as leading to poorer outcomes across health (Wilkinson & Picket, 2009). Including such an analysis may have identified who the intervention was most or least effective for.
There was also substantial attrition at 24 months and as a consequence effects may have been masked by the reduction in the data pool and the relative reduction in power to detect if an effect was present. Also the economic analysis was based on a small subtest of the sample over a short period of time and as such the conclusions of effectiveness are limited, although the authors freely point this out in the paper.

Discussion

This large and well-conducted trial throws up a number of questions. The recent government announcement to increase funding in mental health (HM Government, 2016) emphasised the need to invest in provision to young people. The key question is, how can this be well targeted in a cost-effective manner. The concept of universal provision, with the aim of preventing anxiety in children, appears at least from this study to have limited support, as the interventions performed no better than each other at 24 months. This may speak to the existing strengths of PHSE education and also little need for health professional involvement in delivery. Currently it would seem this approach would be outside of the financial abilities of most schools. Also it is unclear how well any intervention would generalise to more diverse schools and in highly economically deprived areas.

Most interesting is that it appears that the delivery of the programme had the same effect even when only 60% of the core material and between session tasks were assigned. It may therefore be useful for future research to establish the active ingredients of the interventions, which may lead to a reduced cost. The alternative is to research to identify those who are at risk of developing anxiety and depression at a young age and target interventions at this group, rather than pursing a course of universal delivery.

The rates of anxiety and depression in children is still a major public health concern, but this study suggests blanket delivery of prevention programmes in schools may not be the answer to nudging children away from a trajectory of developing anxiety or depressive disorder.