BESST is a cluster randomised controlled trial (cRCT) (ISRCTN registry ISRCTN90912799, registered with ISRCTN 28 May 2020). Schools from 15 geographical areas across four regions of England were invited to participate in the study. The primary and secondary outcome measures were assessed at baseline, 3-month, and 6-month post randomisation. The researchers collecting the post-baseline measures, chief investigator and senior analyst were blinded to the trial arm throughout the trial.

School recruitment took place in four regions in England, the Northwest, West Midlands, London, and the Southwest. Recruitment was over two school years (2021/2022 and 2022/2023), aligning with the start of the school academic years, with intake after the summer holidays in order to follow up participants at 6 months prior to the end of the year exam period in May. School recruitment began in September 2021 and closed in December 2021 for year 1 and took place between September 2022 and December 2022 for year 2.

Schools were approached in regions served by the local MHSTs and were identified via two methods: an introduction to a relevant school staff member was made via the MHST, after which members of the research team would meet to discuss the trial with the school staff. Alternatively, and less frequently, the research team would identify and contact the sixth-form staff in appropriate schools directly and follow the same process if interest was indicated.

One hundred and eleven state-funded schools containing a sixth form (students aged 16–18 studying A levels) were first identified through either MHST referral or searching the local records. These schools were first screened for eligibility prior to contact being made; they were then formally screened once they had expressed interest in participating in the trial. In order to be eligible for the study schools must be (i) secondary schools with sixth form or dedicated sixth form college, (ii) state-funded and have (iii) sufficient resources available to host the trial. Exclusion criteria were (i) further education college, (ii) privately funded school/college and (iii) sixth form student population < 70.

One hundred and four eligible sixth forms were approached to take part in the study. They were asked to express written interest in the study, after which a final screening for eligibility was conducted, and an information session was organised to explain the study to the relevant staff members at each school and 57 consented into the trial. The necessary school level summary characteristics were sent to the statistician to carry out the randomisation of the block of schools for each area. Following this, dates were organised with the school for the site trial Research Assistant (RA) to attend a sixth-form assembly and present the study to the students. Any interested students were invited to attend a lunchtime information session where they were able to find out more information about participation and receive a written information sheet and consent form. Students were asked to review the information and inform a designated member of school staff by a set date if they would like to participate.

From the 57 consented schools, 1407 students attended the project information meetings, of which 991 went on to consent to the study and were screened for eligibility. Once consented, the list of students was sent to an independent statistician to randomly generate a list to determine which 19 students from each school would be invited to take part in the study and be screened for eligibility. This was limited to 19 students per school for practical reasons in implementing the DISCOVER workshop. Participant inclusion criteria were (i) aged between 16 and 18 years, (ii) attending school or college, (iii) sufficient English to provide valid informed consent and complete assessments in the BESST study, (iv) seeking psychological help for stress, (v) able to attend the DISCOVER workshop on school premises and (vi) able to provide informed written consent to participate. Exclusion criteria were (i) identified as actively suicidal through risk assessment, (ii) current involvement in psychological therapy for anxiety or depression with CAMHS, or (iii) severe learning difficulties. The 926 eligible participants were then invited to take part in the study and complete baseline assessments. Once all baseline assessments had been completed, the allocation for that school was released and the school was informed.

Full flow of schools and participants recruitment into the study can be seen in Fig. 1.

Randomisation was done at school level in a 1:1 ratio. The sequence was generated by a statistician not within the research team using a randomisation algorithm for cluster randomised trials developed by Carter [22] and was stratified by site with balancing covariates of school size and index of multiple deprivation (IMD) [23]

Primary, secondary, and economic measures shown in Table 1 were collected at the participant level at baseline, 3-month, and 6-month follow-up. Full information about the outcome measures can be found in the protocol [21]. The primary outcome measure for this study is the Mood and Feelings Questionnaire (MFQ). The MFQ is a 33 item self-report depression measure with scores ranging from 0 to 66 with a higher score indicating a higher level of depression. Baseline and follow-up assessments took place in schools and were facilitated by local/site RAs. The RA arranged a 1–2–1 appointment with each participant, via relevant school staff. The appointment took place in a private room at the participants’ school, where the RA explained each measure to the participant and then allowed the participant to complete the measures whilst remaining present for any questions.

Prior to 3- and 6-month follow-up appointments, the participants were briefed by school staff (and by the RA at the start of the appointment) not to reveal the workshop allocation of the school in order to keep the RA blinded. To improve follow-up rates, vouchers were offered to participants for completion of assessments at post randomisation timepoints. If participants were absent on the day the assessments took place, the RA attempted to return to see them when they were available or sent an assessment pack out by post for completion and return.

The sample size required for this study was 900 participants recruited from 60 schools. This number accounts for an assumed loss to follow-up of 12.5% of students and 4% of schools. As this is a cluster randomised trial, the sample size estimation is inflated based off an intra-class correlation of 0.02 to account for similarities within cluster. A sample of this size would provide 90% power to detect a clinically meaningful effect size of 0.28 on the primary outcome the Mood and Feelings Questionnaire [24].

The baseline characteristics of the schools and participants were summarised using appropriate summary statistics. Baseline measures of the primary and secondary outcomes were scored and summarised to characterise the sample. All data manipulation and summaries were performed using Stata version 18. The trial statistical analysis plan (SAP) can be found as supplementary material to the protocol [21]. The primary outcome was analysed using a multi-level multivariable regression fitting schools with a random intercept, and adjusting for pre-specified covariates and presenting the adjusted mean difference with 95% CI, p-value and intra-cluster correlation.

Adolescent PPI groups from the Anna Freud Centre in London were consulted to inform effective recruitment strategies. PPI members advised on the content and delivery of participant recruitment presentations to provide optimal clarity of trial information, ensure appropriateness for the target population, and maximise engagement of the presentations. Focus groups were also run with groups of adolescent and young adult males from two of the trial regions to understand improved strategies for engagement of boys [25]. The findings of these focus groups were used to inform recruitment approaches, in particular transparency and honesty around participation and presence of a male ‘role model’ during recruitment events.