This paper describes the protocol of a study funded by the Wolfson Foundation. The study was designed to test the effectiveness of an adapted version of CwS, called Skills for adolescent WELLbeing (SWELL). The goal of the SWELL intervention is to prevent the onset/recurrence of MDD in YP at high-familial risk of depression. When designing the study, several adaptations were made with input from the trial team, the trial management group (TMG), the original intervention developers, clinicians and YP, and the potential impact of these was considered.21 22 The main adaptation involved including treatment optimisation for parents in a depressive episode at study entry before randomisation of the YP. This adaptation was intended to maximise the effectiveness of the intervention for YP. Although the addition of an intergenerational component might add to the burden of delivery, we judged that the potential benefits of adding parent treatment optimisation (PTO) outweighed the risk of increasing complexity of delivery.

We also made two main adaptations to the delivery of CwS. First, we reduced the number of continuation sessions from six to three. This was done with future implementation in mind, given that YP depressive symptomatology is common and access to psychological interventions is limited.23 As SWELL is intended to be a preventive intervention, we decided the content could be delivered effectively over a shorter period, reducing the burden of participation for YP. Indeed, a secondary analysis of CwS showed that efficacy remained with partial completion of intervention sessions.24 Second, we adapted the intervention for online delivery, due to scalability considerations about possible future implementation in healthcare settings. Evidence indicates that online psychological interventions are as effective as face-to-face formats,25 especially when this is facilitated by a therapist.26

In terms of presentation, we made minor updates to the materials included in the original CwS workbook. This included the use of more contemporary comic strips to illustrate key learning points. We also translated the YP workbook into Welsh to create a bilingual resource (although all intervention groups will be delivered in English during the planned trial). The adaptations described were judged necessary for a UK context with a view to future implementation in NHS settings.21 22

Given the amendments made to the intervention, we will carry out interviews with therapists following completion of the first two YP intervention groups and analyse with rapid content analysis.27 This will inform if minor procedural updates are required to ensure feasible and accessible ongoing delivery for the remainder of the trial.

The primary aim is to test the effectiveness of SWELL in increasing the time-to-onset of YP MDD by 9-month follow-up. We will also examine the effects of the intervention on secondary outcomes including depression-free days, depression risk score, mental health problem symptoms, functioning and quality of life. A nested process evaluation will assess fidelity, adherence, the impact of adaptations on effectiveness and implementation and possible intervention mechanisms. We will undertake quantitative analyses of potential intervention mediators (eg, dysfunctional attitudes, self-efficacy) and exploratory analysis of factors that might modify effectiveness (eg, parent depression status).

This is a multisite, two-arm, parallel-group effectiveness RCT with assessors blinded to allocation. YP will be randomised 1:1 to the intervention condition (the SWELL programme plus usual care) or the usual care only condition. Given that most YP do not receive preventive interventions, usual care is the appropriate comparator. The method of randomisation will be simple-permuted-block randomisation stratified by site, with the unit of randomisation being the YP (online supplemental material 2).

These are outlined in box 1 and involve identifying a group of YP aged 13–19 years at high risk of depression (depression in a parent/carer plus YP current elevated symptoms/a past depressive episode). Other eligibility criteria will ensure YP can participate fully and that the parent/carer will consider engaging with PTO. We will also ensure parents do not have additional mental health diagnoses likely to impact on the appropriateness and effectiveness of PTO (eg, bipolar disorder/psychosis).28 29

At study enrolment, if the parent is currently depressed at the screening stage, they will be offered PTO. Baseline assessments will follow completion of PTO (or immediately if the parent refuses the PTO offer), and the YP will be randomised.

The trial will be conducted by Cardiff University and supported by University Health Boards (UHBs) in Wales and Clinical Research Networks (CRNs) in England. UHBs and CRNs will facilitate recruitment via general practices (GPs) and mental health services. Secondary schools will also be approached to support recruitment.

We will recruit via: (1) existing cohorts (https://www.ncmh.info/, https://gladstudy.org.uk/), (2) primary care and mental health services, (3) schools and (4) advertising (eg, posters/leaflets, videos) in health/community settings. Standard emails/letters/text messages with options to respond via multiple methods (study website/email/telephone) will be used to distribute study information. To help ensure a diverse sample, we intend to minimise inclusion criteria and recruit in socioeconomically diverse areas. Interested individuals will be asked to complete an online form to register interest and provide contact details. They will be contacted by telephone/videocall to complete screening with a researcher (and ask any questions). Once confirmed as eligible, the parent and YP will complete the appropriate informed consent/assent forms via phone/videocall with a researcher (online supplemental material 3). Of parents who register an interest, we estimate 25% will have a child who fulfils the eligibility criteria (ie, they have a history of depression/currently elevated symptoms) based on data from this population.30–32 Recruitment progress will be monitored regularly against expected targets.

In consultation with the study TMG and trial steering committee (TSC), minor changes were made to broaden inclusion criteria to better reflect how the intervention would be delivered in practice (online supplemental material 4; Analyses).

We aim to recruit 400 YP and parents/carers, based on a previous study17 and sample size calculations. We allow for 10% loss to follow-up based on previous studies of this population and depressed YP.16 30–32 The effect size observed in the study by Garber et al16 was a HR of 0.63. We expect the effect of the intervention in this RCT to be slightly greater than this due to the addition of PTO before randomisation. We therefore define the minimum clinically important difference for this study as an HR of at least 0.60 (equivalent to an 11% difference in MDD between study arms). With a sample size of up to 400 randomised participants, we estimate statistical power as 79.9% to detect a difference of at least HR=0.60 between study arms, with an assumed 10% attrition rate at 9 months, expected event rate probabilities at follow-up of 0.327 (control) and 0.211 (intervention) based on17 a one-sided alpha of 0.05. A HR of 0.56 (equivalent to a 12% difference between study arms) will attain 87.0% power. A STATA V.17 programme was used for this estimation.33

Blinding therapists or participants is not possible for the intervention under study. Assessors at baseline and follow-ups will be blind to the intervention arm. Participants will be reminded of the importance of assessors remaining blind to allocation as recommended,34 and assessors will report any instances of potential unblinding. The trial statistician will remain blinded to group allocation during analysis. It will not be possible to blind the qualitative analyst(s). It is not possible to blind YP, which could lead to reporting bias. We will mitigate this by using an interviewer-led criterion-based clinical interview to define the primary outcome and standardised prompts for questionnaires. We will also evaluate the consistency of effects for outcomes reported by different raters (YP and parents/carers).

Adherence to the intervention for YP will be defined as attending 6 out of 8 acute sessions and 2 out of 3 continuation sessions. Adherence will be rewarded by providing certificates following completion of the programme. Therapists will also offer sessions scheduled outside the school day and one-to-one catch-up sessions if YP are unable to attend a session. Attendance for each participant will be monitored. For PTO, the opportunity to be assessed and treated by a specialist (psychiatrist) may be attractive, and we will offer flexibility in how parents complete appointments (in person/videocall/phonecall). Adherence to PTO will be defined as completion of the baseline assessment and four follow-up calls.

Participants (parents and YP) will each receive £60 in gift vouchers for participation (£10 at baseline, £20 at 3 months, £30 at 9 months) and study completion certificates at 3 months and 9 months. Those who also participate in qualitative interviews and/or focus groups will receive an additional £20 voucher.

Adaptations to the intervention were made following discussions with the TMG, the CwS intervention developers, professionals and our study advisory group of 12 YP with lived experience of depression/anxiety (YPWLE). The latter advisory group contributed feedback on protocol development during three meetings (two online, one hybrid) and a smaller group contributed on one additional occasion. YPWLE advised on: the study name, assessments, timings of intervention/assessments and adaptation of the intervention (workbook content including language, visuals, online delivery methods). One young person and a parent with experience of depression are members of the independent TSC.

Parents and YP will complete assessments with a researcher via videocall at initial screening, at baseline and 9-month follow-up post randomisation. Self-report questionnaires will be completed at baseline, 3 months and 9 months (figure 1). We selected 9-month post randomisation to assess the primary outcome for consistency with prior studies16 and because this aligns with completion of the full intervention. The 3-month postrandomisation follow-up was selected to allow for measurement of potential mediators and secondary outcomes following completion of the acute intervention sessions. Table 2 shows the schedule of measures.

A nested process evaluation (figure 2) will assess fidelity, adherence, feasibility and acceptability of the intervention and possible intervention mechanisms. It will also assess how key adaptations made to SWELL (ie, the addition of PTO) influence effectiveness and implementation (table 3). We aim to conduct focus groups with YP, parents and professionals involved in the youth-focused CBT intervention and PTO. We will additionally ask for feedback on trial processes which will include those allocated to both study arms (table 3). The sample size will be reviewed at each data collection phase to ensure data saturation has been reached and that no new themes emerge. We will complete a focus group/interviews with eight parents in PTO, the PTO professional team and the four CBT therapists. Participants will be purposively sampled to ensure maximum variation in terms of sex, age, ethnicity, educational level and outcome. Interviews and/or focus groups will be audio recorded, transcribed and de-identified.

Data will be kept on a REDCap database held on a secure server at Cardiff University. Participant data will be encrypted, time-stamped and pseudoanonymised using an identification number. Only the researchers will have the information needed to translate the identification number to the participant identity.

The flow of participants through the trial will be presented as a Consolidated Standards of Reporting Trials diagram. Baseline demographic data will be presented for all consented and randomised participants, and patterns of attrition will be examined.

The primary analysis will be time-to-event analysis of the primary outcome (time to the first depressive episode collected at 9-month follow-up) and will use the complete case dataset. However, multiple imputation will be considered as a sensitivity analysis to examine possible dropout bias in the treatment effect estimate. Consistent with clinical practice, depression in the YP will be defined as YP depression reported by either parent/YP with a planned sensitivity check for YP report only. A Cox regression model will be used to determine the difference between trial arms, the HR and 95% CI. We will investigate whether a random effect to account for the group delivery clustering in the intervention arm needs to be included in the primary model. If the effect of the clustering is negligible, this will be excluded. Assumptions of the model (eg, constant hazard) will be checked to ensure model validity. Prespecified subgroups to be compared descriptively include those defined by baseline characteristics of the YP (eg, comorbidity) and parents (eg, depression status at randomisation). External treatment use will also be compared descriptively across study arms, and its effect will be evaluated by including this as a time-varying covariate in the Cox model.

Possible effect modification (eg, by PTO response, baseline demographic factors) will be investigated by including an arm*modifier interaction term. Mediation analyses will examine the effects of potential mediators (eg, improved problem-solving skills, improved parent-child relationship) on the primary outcome via inclusion of the mediator predictor in the primary model.

A sensitivity check will explore the effect of changes to inclusion criteria on the effectiveness of the intervention (online supplemental material 4). Only 20 participants were consented prior to the implementation of these changes, meaning their inclusion is unlikely to influence the main study findings in a substantial way.

Repeated measures analysis will be used for secondary outcomes collected at 3 months and 9 months. Linear mixed effects regression will be used to account for clustering by group delivery in the intervention arm (partial effect model) and will include baseline values as covariates where appropriate. Where the assumptions of the linear regression model are not met, transformation or categorisation will be considered. Logistic regression models will be used for binary outcomes. Main group effects and 95% CIs will be presented for all outcomes.

Qualitative data analysis will use a thematic analysis approach, a process of identifying, analysing, reporting and interpreting patterns or themes.48 Codes will be applied to broad themes and transcripts will be examined to identify key themes and subthemes. To ensure validity, a proportion of transcripts (approximately 20%) will be double-coded by a second qualitative researcher, and any discrepancies resolved by discussion. Thematic analysis will be supported by NVivo software.

A TMG and an independent TSC have been assembled to monitor progress and provide advice. The TSC has agreed to play the role of a data monitoring committee. The TMG is composed of the PI, co-investigators and the SWELL research team at the Wolfson Centre and the Centre for Trials Research. The TSC comprises external experts on trials involving YP’s mental health, an independent statistician, a young person and a parent with lived experience of depression.

In the YP eligible for this study, there is an inherent risk of self-harm and suicide. We have therefore defined expected adverse events a priori as suicidal thoughts with a plan, self-harm or suicide attempts. These adverse events are not expected to be related to the intervention, but the additional contact with the study team for YP in the intervention arm may create greater opportunity for reporting adverse events. Adverse events reporting procedures will be followed by all researchers working on the trial. A standard study procedure for documenting and dealing appropriately with these events will be followed. This includes risk assessment, signposting to services and liaison with clinical services as needed (see online supplemental material 5 for procedure).

Ethical approval was granted by Wales NHS Research Ethics Committee 5 ([email protected]) before the trial commenced (IRAS 305331; REC 22/WA/0254). This manuscript is based on V.5.7 of the protocol dated 17 January 2025. Significant protocol modifications will be communicated by submitting an amendment for ethical review and updating trial registry details. We will share deidentified individual participant-level data following publication of the main trial results on Cardiff University’s research data repository. We will follow BMJ regulations on authorship. Trial results will be published in peer-reviewed journals.

Recruitment started on 21 August 2023, the first participant was consented on 19 September 2023 and 267 participants had been recruited by 02 June 2025. Recruitment is planned to be completed by August 2025 with analysis and dissemination between September 2025 and February 2027. Cardiff University is the study sponsor (ref: SPON1902-22) and the study contact for enquiries is [email protected].

Not applicable.