Patient and Public Involvement

This project was supported by two Young People’s Advisory Groups: the NeurOX YPAG, a group of young people aged 14–18 years who support research in ethics and youth mental health [31], and the Uplift YPAG, set up specifically for this project. Uplift YPAG members were recruited from a larger group of young people aged 16–18 years trained in peer support in a previous project [30], on the basis of suitability and readiness to deliver peer support (assessed by professional judgement from Youth Era) and availability. Priority was given to trained young people from ethnic minority backgrounds and low family affluence. The YPAGs provided extensive input into the course design, including overall structure, content and delivery methods, and Uplift YPAG members additionally acted as peer leaders. Both groups provided input into trial design (e.g., recruitment strategies, data collection methods, outcome measures) and results interpretation (see Additional file 1 for further details on young people’s contributions). All YPAG members were UK residents from a range of socioeconomic and ethnic backgrounds.

RecruitmentParticipants and recruitment context

The study protocol was approved by the University of Oxford Interdivisional Medical Sciences Ethics Committee (R69810/RE001). Participants were recruited through social media adverts and schools across the UK between the 14th of November 2020 and 2nd of December 2020. The poster/advert invited young people to take part in a ‘Coping during COVID’ online course while contributing to research. The aim was to reach at nation-wide sample of adolescents experiencing common emotional difficulties during this period. Eligibility criteria for the study included being aged 16–18, UK resident, a sufficient level of English, consenting to random assignment to one of two iterations of the Coping during COVID course, access to a device with camera, sound and microphone. Those interested were directed to an “Expression of Interest” form via an online survey platform. Potentially eligible participants were invited for a call with Youth Era staff where eligibility and suitability were confirmed. Once confirmed, written informed consent and baseline measures were obtained via the Qualtrics online survey platform. Participants were incentivised to complete the surveys at three assessment points with a £20 voucher reimbursement at the end of their participation.

During the time of data collection, England was operating a three-tier system, with stricter restrictions for areas of high incidence; however, schools, shops and personal care centres were open across three tiers. Social distancing was enforced, and people were encouraged to stay in their local areas as far as possible. To reach young people for whom the intervention would be most valuable and relevant, social media adverts targeted areas classified as Tier 2 and Tier 3, as well as areas of high incidence across Northern Ireland, Wales and Scotland. Once these places were filled, we also set adverts and contacted schools in other areas in the UK to reach our recruitment target.

Procedure

After completion of the baseline measures, participants were automatically randomly allocated in a 1:1 ratio to intervention or wait-list control using the Qualtrics randomisation tool. We chose to use this randomisation procedure because feedback from participants in a previous trial indicated that young people wished to know their course allocation immediately, to allow them to save the dates and prepare for the intervention. The automated tool ensured that the research team could not affect randomisation. Participants randomised to the intervention arm completed the course from the 2nd-6th of December 2020 (course 1) whilst the wait-list control were offered to complete the course after the final assessment from the 16–20th of December 2020 (course 2). Both arms were assessed via Qualtrics surveys immediately post-course (approximately 1 week post-randomisation) and at a follow-up point 1 week post-course (approximately 2 weeks post-randomisation). Participants completed all the assessments independently online; surveys were distributed via email following standard wording, by a researcher blind to condition. Survey answers were identified by a randomly assigned ID and participants were made aware that neither Youth Era nor their peer leaders had access to their answers.

Trial registration and deviation from protocol

The trial was registered under ISRCTN77941736 (https://doi.org/https://doi.org/10.1186/ISRCTN77941736). Because of the rapid planning and delivery of this project in the context of the COVID-19 pandemic, the trial registration was only submitted shortly before recruitment started (12-Nov-2020, with recruitment commencing 14-Nov-2020) and therefore registration was published during the recruitment period as retrospectively registered (23-Nov-2020). Participant recruitment was slower than anticipated and as a result we made two changes to the original protocol. Firstly, we changed the second follow-up from 2 weeks post-course to 1 week post-course. This allowed us to extend our recruitment period without extending the overall trial duration and ensured follow-up assessments were complete before the Christmas holiday period. Secondly, we reviewed our sample size calculation and reduced the target sample size from 120 to 100 participants (see Statistical analysis).

Coping during COVID intervention and peer leader training

Coping during COVID was co-delivered to participants via Zoom by a selected group of young people from the Uplift YPAG group (peer leaders) and a team of specialist Youth Era staff, over 5 consecutive days (2 h/weekday; 4 h/weekend day). The course was made up of educational lectures delivered by Youth Era staff to the whole group, and small group activities and supportive and sharing sessions, led by the peer leaders. Small groups consisted of 5–7 participants, with mixed gender and age, and small group activities were delivered using both Zoom and WhatsApp. Participants were unacquainted to their peer leaders prior to the course.

The course focused on self-care, coping with emotions, identifying strengths, building resilience, improving coping skills, developing purpose, and forming community. Activities included, for instance, generating coping strategies for a variety of scenarios; identifying and challenging self-limiting beliefs; sharing letters of appreciation among group members; and drafting a self-care plan. More details of the course content are provided in Additional file 2. Prioritisation of course content was influenced by data from the literature and YPAG input on what young people were struggling with during that current period of the COVID-19 pandemic. Even though the socioemotional skills targeted (e.g., coping) were broadly relevant, COVID-specific cases and examples were provided where possible. For example, icebreaker questions included “What have you learned about yourself during COVID?”.

Of the group of Uplift YPAG members who co-designed the course (21 members), 9 co-delivered the course for the intervention arm based on their availability. Two additional youth acted as background support, providing one-to-one peer support when necessary and assisting with large group activities. The remaining members of the Uplift YPAG acted as peer leaders to the waitlist control arm.

Peer leaders had all received previous training in peer support through the Uplift Peer Support Training Programme. This training course equipped adolescents with the skills to support the mental health and wellbeing of friends and peers during the COVID-19 pandemic. Young people who received this training reported greater ability to help others, compassion and civic engagement, compared to controls [30]. Peer leaders also received an additional 2-day training provided by Youth Era before delivering the Coping During Covid course in the current trial. Peer leaders’ age range matched that of adolescents who took the Coping during COVID intervention (i.e. trial participants): 16–18 years old.

Peer leaders received close support and mentorship from Youth Era staff throughout. This included daily meetings before and after the course to provide guidance and discuss any difficulties. Youth Era staff were also on call during the small group activities in case peer leaders had an emergency or needed support.

Outcome measuresPrimary and secondary outcomes

All quantitative outcomes were assessed in the intervention and control group using self-report measures at baseline, 1 week and 2 weeks post-randomisation. Our primary outcome was wellbeing at 1 week, and secondary outcomes included connectedness, perceived coping skills, sense of purpose, self-esteem and self-compassion at 1 week and 2 weeks, and wellbeing at 2 weeks.

Wellbeing was measured using the Warwick-Edinburgh Mental Wellbeing Scale (WEMWS; [32]), which includes 14 items (e.g., “I’ve been feeling relaxed”; total score: 14–70).

Connectedness was measured using the Social Connectedness Scale [33], including 20 items, adapted to refer to peer relationships (e.g., “I feel close to my peers”; total score: 20–120).

We measured perceived coping skills using items adapted from the COVID-19 Adolescent Symptom & Psychological Experience Questionnaire [34]. The questionnaire included three items about participants’ confidence in their ability to manage negative emotions arising from the COVID-19 pandemic and lockdown measures (total score: 3–15).

Sense of purpose was measured using the 12-item Claremont Purpose Scale [35], which assesses three dimensions of purpose: meaningfulness, goal orientation, and a dimension called “beyond the self” which was modified to assess participants belief in their ability (rather than motivation) to make a meaningful contribution to the world (total score: 12–60).

To measure self-esteem, we used Rosenberg’s 10-item Self-esteem Scale [36] (e.g., “On the whole, I have been satisfied with myself”; total score: 10–40).

Self-compassion was measured using the Self-Compassion Scale [37], including the 6-item Compassionate Engagement Subscale (e.g. “I have been accepting, non-critical and non-judgemental of my feelings of distress”) and the 4-item Compassionate Action Subscale (“I have directed my attention to what is likely to be helpful to me”) (total score: 10–100).

Qualitative outcomes

Open-ended questions immediately post-intervention (1 week post-randomisation) captured participants’ experience of the course, including perceived impact, and plans to implement course content. Participants allocated to the intervention arm were asked: i) how the course impacted their life, either positively or negatively (e.g., how they see themselves, their relationships, day-to-day life), and ii) their plans to use what they have learnt from the course in their life moving forward. Additional questions about views and experience of the course itself (e.g., suggested improvements) and responses to situations when participants felt distressed or upset were also included. Results from these additional questions fall outside the scope of this paper and will not be reported here.

Statistical analysisSample size

Our original target sample size was 120 participants and was determined on the basis that with a retention rate of 90% this would be provide > 85% power at the 2 sided 5% significance level to detect a difference between the intervention and wait-list group on the primary outcome of 0.6 standard deviations (Cohen’s d medium effect size). Participant recruitment was slower than anticipated so during the recruitment period we took the pragmatic decision to adjust our recruitment target to 100 participants on the basis that with a 90% retention rate this sample size would still provide 80% power to detect the planned effect size (0.6). G power was used for sample size calculations.

Analysis

Descriptive statistics (mean, SD; n, %) were used to summarise baseline demographic characteristics for each group. We used linear mixed effects models to compare the two groups over time for each outcome variable. Each model included fixed effects of group (intervention, control), timepoint (baseline, 1 week, 2 week), group by timepoint interaction, gender, age, and corresponding baseline score. To account for the fact that participants completed measures over multiple time points, a random participant effect (random intercept) was included in each model. Primary and secondary outcomes were analysed using intention-to-treat principles, without imputation. For each outcome, we present the difference in estimated means between groups for each time point, together with 95% CI for the difference, p-value, and effect size (Cohen’s d).

Qualitative analysis

We adopted a directed content-analysis approach for qualitative analysis of responses to open-ended questions related to impact of the course and intentions to use skills. The analysis was guided by literature on potential positive outcomes associated with peer-led interventions. New codes emerging from the data were also identified and included as appropriate. The initial coding frameworks was developed by GM based on responses related to course impact, and further iterated through meetings with GP and TR. The same scheme was then applied to intentions to use what they have learned, with additions when new content was identified. The final coding frameworks were applied to the full dataset, where we coded the presence/absence of each code for each response, regardless of response length. A second coder additionally analysed 25% of the data, reaching substantial inter-rater agreement (κ =  > 0.704) [38].