Exercise for Depression in Young People

Aims

1. Compare mental health outcomes in depressed young people receiving a tailored exercise intervention with mental health outcomes for depressed young people in a waiting list control group.
2. Explore young people's experiences of exercise in the treatment of their depression.
3. Investigate the attrition rate for the tailored exercise intervention group.
4. Investigate longevity of changes in exercise habit with a six month follow up.
5. Conduct an economic evaluation of the tailored exercise intervention.

Research Questions

1. Does an exercise intervention programme tailored to young people's exercise levels significantly improve mental health outcomes and exercise attritions rates in young people with depression?
2. Are the changes in exercise habit and outcomes brought about by participation in the tailored exercise programme still evident at six months post intervention?
3. How cost-effective is a tailored exercise intervention in young people with depression?
4. How acceptable is a tailored exercise intervention to young people with depression?

Design: Pragmatic Randomised Controlled Trial. An experimental arm of a tailored exercise intervention and a waiting list control group.

Sample: 14-17yr olds attending Tier 2 and Tier 3 Child and Adolescent Mental Health outpatient services in Nottingham City and Nottinghamshire County, presenting with depression. For 80% power to detect a difference in mean change between the two groups of 8 points on the Children's Depression Inventory (the primary outcome measure), and assuming equal standard deviations of 16 points, significance level of p<0.05 we need 126 participants. Assuming equal attrition of 20% we will recruit 158 participants. Random allocation to the experimental or control arm.

Inclusion/exclusion criteria: Presenting symptoms defined by their assigned clinician in the CAMHS team, as opposed to specific diagnosis. Eligible participants: those presenting with depression. Exclusions: Unable to participate because of injury or physical health problem that precludes participation, and regular exercisers as indicated by their responses to the SCQ.

Intervention: Intervention consists of 12x 60 minute aerobic exercise sessions, in groups of 10, tailored to participants' exercise levels and preferences as assessed using the Stages of Change for Exercise Questionnaire (SCQ). Psychosocial and motivational support throughout each session to facilitate ongoing participation and enduring change consisting of behaviour modification: goal setting, self-monitoring, social support, enhancing self-efficacy and shaping. A variety of lifestyle activities will be used as these are associated with improved compliance rates.

Control group: Waiting list receiving usual care. The same facilitator will be used for all experimental interventions. Outcome primary measures: Depression, captured using the Child Depression Inventory (CDI) - 5 scales measuring negative mood, interpersonal difficulties, negative self-esteem, ineffectiveness and anhedonia, designed for 7-17 year olds, sensitive to changes over the proposed timescale, quick to administer and yields an aggregate score indicating depressive illness and clinically significant depression. We will measure the number of exercise sessions completed, physical fitness using maximum energy expended, health and social care costs using the Client Receipt of Service Inventory (CRSI), Quality of life using the EQ-5D and total number of incidents of self-harm and violence, treatment received and compliance with this, during the 12 week intervention period. Outcome measures will be collected at baseline, post-intervention and six months follow-up.

Data Analysis: Two groups, described in terms of their baseline characteristics. Primary outcome will be the CDI post intervention. A regression model will be used to compare the two groups in terms of the CDI score post intervention conditional on baseline score. No other confounders will be adjusted for unless there's an imbalance at baseline between the two groups on pre-specified covariates that are considered to influence outcome. For secondary outcomes, similar models will be built all conditional on the relevant baseline score in addition to baseline CDI score. Analyses will be intention-to-treat and conducted on a dataset where codes for the two intervention groups are unlabeled. All estimates of effect sizes and numbers needed to treat will be reported with 95% confidence intervals. The statistical analyses will be conducted independently of the research team with assistance from the Clinical Trials Support Unit, University of Nottingham Medical School.

Economic evaluation: We will collect resource use data using the CRSI. Resource use will be valued using published and as appropriate patient reported, unit costs. The primary measure for the cost effectiveness analysis will be the SPPA. Assessing health related quality of life using the EQ-5D will enable a cost utility analysis to be undertaken for the trial period. Using the information on costs and benefits an incremental economic analysis comparing the intervention to usual treatment will be undertaken to estimate mean cost-effectiveness. If one group is clearly dominant (less costly and more effective) a recommendation will be made. If benefits are not significantly different then a cost minimisation approach will be adopted to recommend the cheapest option. If non-dominance occurs (that is if costs are greater and the intervention is more effective or if the intervention is cheaper and less effective), an incremental cost-effectiveness ratio will be produced and a value judgement about value for money will need to be made. To test the robustness of results in the face or any uncertainties, and to improve the generalisability of results, sensitivity analysis will be conducted. In addition, uncertainty surrounding the cost-effectiveness of the tailored exercise intervention will be presented graphically using Cost-Effectiveness Acceptability Curves. Should the data be skewed then non-parametric bootstrap analysis using the percentile method confidence interval will be done and incremental net benefits estimated.

The qualitative study: The MRC Complex intervention Framework, shows that service delivery often involved therapeutic elements that are difficult to define but essential, and also the use of qualitative methodology alongside traditional trial design to elicit barriers and drivers to service delivery of interventions. Therefore, we will conduct 3 focus groups of 45 minutes, each with 6-7 participants selected from the intervention group for their diverse opinions, and facilitated by two service users. We will explore acceptability, pros and cons of the tailored approach, what worked and what was less successful. This will help explain why/how the intervention was effective, if this is the case, or why it failed. We will record the focus groups and transcribe them. Transcripts will be coded and analysed for patterns, themes and sub-categories, paying particular attention to outliers and rival explanations given by participants. We will supplement these data with one minute 'box pops' video recordings, of two participants per week, to further explore the young peoples' experience of the intervention. To ensure validity, the results will be fed back to the focus group leaders and other interested participants, whose interpretations will be integrated into our conclusions.

Implementation Analysis: The analysis of implementation and diffusion of innovations is increasingly being undertaken in order to understand knowledge transfer from research to clinical practice in the health service. We propose to use the Steering Group to do an implementation analysis of the project's findings. Such an approach is in line with MRC guidance for developing and evaluating complex interventions. In the steering group we propose to elicit from members the issues relating to the uptake of the findings from the proposed study and the barriers we may encounter in doing this. We will ask each member of the group to complete a proforma eliciting a Social Network Analysis of their interactions with other service providers to better understand the concerns of providers not immediately associated with the project.