Developing the STASH intervention

This chapter describes how the intervention developed from the initial plan set out in the funding proposal, to the intervention as delivered in the feasibility study. It details methods (including a pilot) and explains how we integrated our learning from three development stages – (1) evidence collation and synthesis, (2) co-production and (3) optimisation – into design decisions. The intervention development process is summarised in Figure 1.

FIGURE 1 The STASH study intervention development process (by 6SQuID component). 6SQuID, Six Steps in Quality Intervention Development; PPI, patient and public involvement; YP, young person. (Figure not reproduced here; see the original article.)

Background

Our starting point was the proposal version of the STASH study logic model (Figure 2). This offers a visual representation of the problem (as conceived), inputs (intervention components that were expected to bring about change), process by which change would occur, intermediate factors and expected end outcomes. This was subsequently developed into a full programme theory that, in addition to theorising the mechanisms of change, hypothesised the conditions under which they would operate effectively. The post-pilot version of the programme theory (see Figure 4) guided the evaluation and identified important contextual factors that might shape outcomes and effects.

FIGURE 2 The STASH study logic model: proposal stage. Reproduced from Forsyth et al.62 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original figure. (Figure not reproduced here; see the original article.)

Developing a theory-informed young people’s sexual health intervention

Development work is essential to tailor theoretically grounded programmes to the needs of the target population.63–65 The STASH study aimed to adapt what was known to work in a successful existing public health intervention to a new problem and context. The planned intervention drew on theories and mechanisms that were effective in ASSIST40 (see Chapter 1), while targeting an older age group (14–16 instead of 12–13 years) and focusing on STI prevention instead of smoking. Therefore, although ASSIST offered a theoretical starting point, the appropriateness of intervention components needed to be reviewed and revised accordingly. We anticipated that these differences would require significant adaptations to the ASSIST content and mode of delivery. Crucially, we also set out to explore how we might harness the potential of social media as a mechanism to communicate messages and, in doing so, modernise ASSIST.

Key intervention components brought forward from ASSIST

We began with the overarching mechanisms of ASSIST:

• Recruitment and training of ‘influential’ students (PSs) to spread and sustain positive norms through informal interactions with peers. Grounded in diffusion of innovation theory59 and recruitment of ‘early adopters’,66 ASSIST found that recruiting 15% of the year group was effective in the context of smoking prevention.67
• Use of professional trainers to train PSs to increase intervention efficacy and credibility, reduce burden on schools and ultimately increase potential for scalability and widespread adoption.

We intended to use the following components from ASSIST.

Nomination

Identify the most influential students in secondary 4 (S4) by blinded peer nomination.

Recruitment

From this influential group, recruit a percentage of the entire year group to act as PSs.

Training

A 2-day PS training programme, delivered by a third-sector organisation in school time, at an external venue, covering PS role, skills and responsibilities, and knowledge building.

Peer supporter-delivered activities

A period in which trained PSs carry out intervention activities.

Trainer-led activities

Follow-up sessions delivered during the PS activity period.

Participant acknowledgement

Peer supporters receive formal recognition for their role.

Our development work followed MRC guidance61 on the development and evaluation of complex interventions, which acknowledges that intervention development is often not linear and may have several phases. We focused on exploring the evidence base and identified relevant theory and pilot work to address ‘key uncertainties’.61 In the STASH study, ‘key uncertainties’ were the feasibility of a PS approach in a sexual health context, the older age group (compared with ASSIST) and use of social media as a forum for interaction.

We also followed the Six Steps in Quality Intervention Development (6SQuID) framework, which sets out ‘practical, logical, evidence-based ways to maximise effectiveness’.68 These six steps are:

1. define and understand the problem and its causes
2. identify which causal or contextual factors are malleable and have the greatest scope for change (and who would benefit most)
3. decide on mechanism(s) of change to be utilised
4. clarify how these will be delivered
5. test and refine on a small scale
6. collect sufficient evidence of effectiveness to justify rigorous evaluation/implementation.

We detail the STASH study development stages in line with 6SQuID stages 1 and 2 (grouped here as ‘evidence collation and synthesis’), stages 3 and 4 (grouped here as ‘co-production’) and stage 5 (grouped here as ‘optimising the intervention’). Stage 6 is addressed in Chapters 5–7. We included Michie’s Behaviour Change Wheel/Theoretical Domains Framework in the preliminary logic model69,70 and referred to it in designing the STASH study website and training materials, but found it less applicable to determining macro-level mechanisms of change, as these were, to a large extent, determined by the ASSIST intervention.

Intervention development aims and objectives

The overall aims of the development stage were to finalise the intervention design and refine and test the initial logic model. The broad objectives were to:

• fully articulate the problem [i.e. young people’s sexual behaviour and risk of STI transmission (6SQuID stage 1)]
• explore and clarify causal and contextual factors relating to young people’s sexual behaviour and identify those most amenable to change (6SQuID stages 2–4)
• identify contextual factors that might shape implementation (6SQuID stages 2–4)
• establish how the STASH study might bring about change in young people’s sexual behaviour (6SQuID stage 3)
• identify intervention content, including existing high-quality resources (6SQuID stage 3)
• refine programme theory (overall 6SQuID aims).

Intervention development process

Each stage in this iterative process – and its constituent activities, objectives and methods – is outlined, followed by a synthesis of findings (see Synthesising and operationalising development process findings).

Stage 1: evidence collation and synthesis – 6SQuID steps 1–3 (months 0–9)

The objective of stage 1 was to collate and synthesise evidence to inform key intervention content, establish the most effective delivery format and develop the most appropriate theoretical framework to guide this. We explored necessary adaptations to ASSIST, including a shift to social media as an alternative mechanism for delivering content. Key findings and recommendations derived from each activity are outlined in Appendix 1.

Review of the evidence base: ‘scoping’ literature review

This review examined academic literature for evidence to identify drivers of target behaviours, effective components of school- and digital media-based interventions, and behaviour change theories relevant to young people and sexual health.

Following relevant guidance,71 targeted scoping searches of high-level literature were conducted, including systematic reviews of interventions [peer education, social media, school-based sex and relationships education (SRE)]; review, discussion or state-of-evidence articles (key drivers of risk, facilitators of safer practices and well-being, and issues in young people’s sexual behaviour); and behaviour change theories relevant to young people’s sexual health and well-being.

Search inclusion criteria were:

• age range 12–19 years (to include age groups encompassing target group)
• English as the publication language
• a publication date post 1990
• Organisation for Economic Co-operation and Development (OECD)/high-income countries.

We excluded studies that were published before 1990, were not in English, focused primarily on primary school or post-statutory education-aged children or were based in non-OECD countries.

Searches used combinations of controlled vocabulary and free text (e.g. title/abstract = ‘adolescent’ and ‘condom use’). We undertook targeted searches of the following databases: Child Development & Adolescent Studies, EMBASE, POPLINE, PsycINFO, PubMed and Web of Science. We also reviewed pertinent theoretical literature on complex interventions and behaviour change, beginning with diffusion of innovation theory (the basis of ASSIST).59 Targeted manual searches also included reviewing article reference lists, reviews of subsequent publications citing key papers and papers identified through the expertise of the Trial Management Group (TMG).

Reviews of resources and materials: relevant grey literature, sex and relationships education/practitioner resources, digital media sources, potential social/digital media platforms

We collated high-quality SRE resource exemplars (websites, resource packs). We sought to identify key features, examine theoretical grounding, if available, and identify potential content relevant to the STASH study. Grey literature and relevant web resources were identified opportunistically via professional networks of the TMG and by non-systematic and targeted online searches. We identified policy and guidance documents focused on SRE delivery in the UK. Resources were categorised by source, type and quality, and key points summarised. We assessed quality of youth-oriented sexual health web resources [including information sites, YouTube videos (URL: www.youtube.com; YouTube, LLC, San Bruno, CA, USA) and vloggers/influencers], rating them according to relevance to the STASH study, content quality, design quality (appeal to young people) and apparent authority or legitimacy.

Finally, we reviewed candidate social media platforms for user demographics, functionality, appeal and regulatory information. Platforms [e.g. Instagram, WhatsApp (URL: www.whatsapp.com; Facebook, Inc., Menlo Park, CA, USA), Snapchat and Facebook] were appraised for age restrictions, message-sharing functions (including traceability over time, images, audio, video and ability to interface with a website), private group messaging options and whether or not posted content could be visible to the research team to enable evaluation (feasibility).

Consultations (patient and public involvement): young people’s consultation groups

Consultations explored the target group’s attitudes to and use of social media, and views on the proposed intervention format.

Pre-pilot consultations with young people built on earlier consultations at the pre-proposal stage (16 participants aged 14–19 years in central Scotland). In the first development stage, we conducted three group interviews (20 participants in total) with young people outside the intervention area. Two group interviews were school based and the third was with a community centre-based youth group (conducted at a community centre). We identified each via the Project Executive Group (PEG) and sought to ensure a range of socioeconomic backgrounds [two schools serving populations from mixed Scottish Index of Multiple Deprivation (SIMD) quintiles and a youth group located in an area of high relative deprivation]. Discussions were facilitated by the research team (KRM, CP and RF) and audio-recorded. A semistructured topic guide (see Report Supplementary Material 1) asked for views on classroom SRE (including gaps) and key perceived influences on young people’s sexual behaviour and attitudes. We also asked about preferences in relation to proposed intervention format, including what would engage and motivate PSs. We also explored views on the use of social media to spread messages. Discussions were summarised and reviewed for key themes, commonalities and differences, including comparison with the pre-proposal young people data.

Other stakeholder consultations

Consultation activities were designed to explore local constraints on and facilitators of implementation (including concerns) and stakeholder perspectives on young people’s sexual health, SRE and social media use.

We conducted semistructured interviews with four teachers at two secondary schools considering participation in the feasibility trial. Participants were members of school senior management teams, with some involvement in guidance and SRE. A topic guide (see Report Supplementary Material 1) explored factors emerging from the reviews, primarily relating to views on young people’s sexual health and behaviour, young people’s social media use, SRE delivery in their school (topics addressed, effectiveness) and the planned intervention (including suggested improvements and practical advice). Interviews were transcribed, summarised and reviewed for key themes and learning.

We also met with two relevant quality improvement officers/child protection leads in participating local authorities and with each participating school’s senior management team to further explore the above issues. An additional meeting was held with one school to address concerns around the intervention model. Notes from meetings were reviewed for key learning and integrated with the teacher interview findings.

Development workshop with relevant professionals (sexual health and health promotion specialists, peer educators, youth workers, academics)

The first of two ‘professional’ workshops explored and established key messages to underpin intervention format and content, and likely mechanisms of change. Participants (n = 16) were identified via the TMG’s and PEG’s professional networks and knowledge of the local health and education services landscape. In the workshop, small group discussions focused on key issues emerging from literature reviews (e.g. barriers to and facilitators of positive sexual health and relationships for young people) and how to address these in the STASH study, potential behaviour change mechanisms, function and format of the intervention, and potential challenges for PSs.

In Appendix 1, we outline key findings from each review and consultation, and recommendations for intervention design. How these were actioned is detailed in Synthesising and operationalising development process findings.

Stage 2: intervention co-production – 6SQuID step 4 (months 1–9)

Stage 2 addressed 6SQuID step 4 and aimed to develop and refine proposed intervention format and content, guided by learning from stage 1.

First phase website co-development with Antbits development consultancy

Co-production with web development consultancy Antbits (Cambridge, UK) aimed to develop an interactive website and content management system to provide mobile-optimised access and desktop functionality, and multilevel access for use by PSs (level 1) and by trainers and the research team (level 2). In multiple teleconferences with Antbits, the team explored design options and potential for presentation of curated sexual health content (sexual health information, web links, videos, etc.).

Drafting intervention content

An iterative drafting process aimed to generate a set of content for both the STASH study website and PS training and follow-up sessions, organised by topic (e.g. around relationships, consent, STIs), which would constitute a comprehensive resource. Drafts were led by KRM and built on a synthesis of evidence gathered in stage 1, guided by relevant behaviour change theories identified in the reviews. Topics and content were discussed and agreed at development ‘away days’. Draft content was reviewed by an expert SRE educator to refine focus, tone and language.

Young people’s development workshops

KRM and CP revisited the two school-based groups to gauge opinions on draft website content. Following a loose topic guide (see Report Supplementary Material 1), participants reviewed drafts (via tablet and smartboard) and gave views on visual appeal and ease of use, topics and messages, language, acceptability of external links, information gaps and general acceptability. During the sessions, five students took part in brief cognitive interviews to test selected questionnaire items. Workshops were conducted in school time, lasted approximately 1.5 hours and were audio-recorded, with notes written up and reviewed for key learning.

Stage 3: intervention optimisation – 6SQuID step 5 (months 10–21)

The overarching aim of stage 3 was to pilot the intervention in one school and to use learning from the pilot to refine the intervention. Guided by 6SQuID step 5, this stage consisted of six phases, described as follows.

Intervention pilot

The pilot school was selected from among those expressing interest in participation. It was chosen as typical of schools in the locality (regarding size and relative deprivation), and because the senior management team were supportive of the STASH study. The intervention was implemented over 9 weeks, comprising peer nomination, recruitment meeting, PS training and five follow-up sessions. Of 163 students in the year group, 31 (19%) were invited to recruitment, having received the highest number of nominations. Of these students, 19 (61% of those invited, 12% of year group) were trained and 14–17 attended follow-up sessions (attendance varying weekly). No students explicitly withdrew from the PS role, so all 19 students were considered ‘completers’ and awarded a certificate.

The purpose of the pilot was to test and refine the intervention on a small scale. We undertook a process evaluation primarily to identify what did or did not work (and why), and to identify potential improvements in anticipation of the feasibility study. The eight data collection strands were:

1. semistructured interviews with a convenience sample of intervention participants (14 PSs and three S4 friends of PSs who had some engagement with the STASH study)
2. semistructured observation of the PS training
3. PS evaluation of training via a one-page evaluation form (n = 17)
4. trainer feedback sessions (at the close of training days) and in-depth interviews with all three trainers (at the end of pilot)
5. PS online questionnaire (n = 14)
6. semistructured interviews with three teachers (including the STASH study contact teacher and two other teachers with guidance responsibilities and some awareness of the STASH study)
7. Facebook monitoring data (group membership, number of messages sent)
8. project monitoring log (participation numbers, key contextual information, e.g. school interactions, contemporaneous events).

The pilot enabled us to test the data collection tools and analytical approach. The latter was guided by an analytic framework designed to explore fidelity, acceptability, reach, recruitment and retention, context and perceived impact, and to capture any emergent and unanticipated issues. All pilot data were assessed for common themes using an approach informed by the framework method.72,73 Data from across the data collection strands were collated by the above categories and reviewed in a matrix for commonalities, differences and key learning for the next stage of the study. We also piloted the baseline and follow-up questionnaires.

Second professional workshop

Those attending the first professional workshop were invited back to a second workshop to discuss key findings from stages 1 and 2, particularly the pilot results. Small group discussions focused on how to improve the STASH study website and message-sharing process, on intervention reach and impact, and on PS training and follow-ups. Meeting notes were reviewed for key learning points.

Intervention development meetings

Intervention development meetings consisted of four ‘away days’ with the researchers and delivery partners (West Lothian Drug and Alcohol Service and Fast Forward). Sessions focused primarily on optimising website messages and training content, including the trainer manual, slides and materials (such as games and activities). Detailed notes were taken by the research team and content drafts revised in the course of discussion.

Second phase website co-development

Following handover of the content management system from Antbits, the second phase of website development was led by the research team. Supported by Antbits and the MRC/Chief Scientist Office (CSO) Social and Public Health Sciences Unit in-house graphic designer, the research team worked to streamline and simplify content and structure, and improve visual appeal, primarily via the creation of bespoke infographics.

Website testing with young people

We conducted group interviews with two youth groups: one face to face with an established young people’s action research group (Young Edinburgh Action) and one online using Facebook Messenger (Facebook, Inc., Menlo Park, CA, USA) with a pre-existing friendship group identified through the professional networks of the research team. Groups comprised six (all female) and seven (four female, three male) participants, respectively. Participants were slightly older than the target group (aged 16–17 years), to enable reflection on what information and format they would have preferred at age 14–15 years. Groups discussed relevance, credibility and relatability of newly created content. The face-to-face group was audio-recorded and notes were taken from the audio, whereas the online group generated a transcript of the online discussion. Each was reviewed for key revisions.

Ad hoc consultation meetings (across months 0–21)

A series of meetings were held with school contact teachers, head teachers (HTs) and deputy head teachers (DHTs) at schools considering participation in the trial (two or three per school), academic colleagues, intervention development partners, professional panel participants and local NHS young people’s sexual health specialists. In addition, discussions held in regular PEG, TMG and Trial Steering Committee (TSC) meetings contributed significantly to intervention optimisation.

Synthesising and operationalising development process findings

The following section synthesises findings from across all development activities, showing how these shaped the final intervention. Findings are organised by intervention component (see Key intervention components brought forward from ASSIST). Appendix 1 details specific recommendations from each activity. The synthesis begins with a brief summary of higher-level decisions made regarding the theory underpinning the intervention.

Developing the STASH study programme theory

We conceptualised and used relevant theories at three levels (Figure 3): (1) the social system level (macro theory), (2) the individual level (mid-level theory) and (3) the behaviour level (micro theory). We identified, via literature review, mid-level theories relevant to young people’s sexual health and behaviour. This review, together with stakeholder consultations, informed the choice of behaviour change techniques and training, and website content. We developed our programme theory in line with what we already knew about mechanisms of change in ASSIST and adapted these for the sexual health context.

FIGURE 3 Use of theory in the STASH study. (Figure not reproduced here; see the original article.)

A detailed description of the theories and how we used them is given in Appendix 1 (see Table 19). Following the literature review, we excluded the least relevant of any overlapping theories and applied those most relevant at three levels (see Figure 3): (1) the social system of the intervention, (2) individual behaviour change and (3) specifics of intervention content. Together, these theories describe potential mechanisms for change that might be expected to occur as a result of the intervention.74 Having explored their relevance and utility in the pilot, we integrated these into the revised programme theory (Figure 4) and feasibility trial.

FIGURE 4 The STASH study programme theory: pre pilot. (Figure not reproduced here; see the original article.)

Guided by these theories and our consultations (see Synthesising and operationalising development process findings), we established that it would be crucial to prioritise skills development (particularly communication), encourage and support active participation and responsibility-taking (within the intervention and with regard to sexual behaviour), boost intrinsic motivation and self-efficacy, provide SRE information relevant to young people and tailor the intervention to the needs of the target group.

Synthesis and actions relating to key intervention components

Nomination and recruitment

Motivating schools

Our development activities and literature30,65,75 reviews established that school buy-in would be key to successful implementation. ASSIST highlighted that use of external trainers is a key means of limiting school burden and enhancing intervention workability. Our grey literature review and expert consultations identified that helping schools meet national policy requirements would be a strong motivator. This related particularly to the Scottish Government’s Curriculum for Excellence,19 Getting it Right for Every Child20 policy and the National Health and Wellbeing Outcomes.76 Our schools brochure, therefore, explicitly outlined how the STASH study could help schools meet relevant indicators, particularly around enhancing students’ confidence and sense of responsibility, and addressing stigmatising attitudes.

Peer nomination process

Our consultations with young people and professionals affirmed learning from ASSIST, suggesting that peer nomination (as opposed to teacher or self-selection) would identify the most ‘credible’ or ‘trustworthy’ students40 and that schools should not be allowed to shape the list of peer-nominated students. This was stipulated in the guidance provided to schools.

The ASSIST study also found that to obtain the 15% ‘critical mass’ required for effective diffusion of innovation throughout the year group, 17.5% of nominated students should be invited to recruitment. However, our consultations suggested, and the pilot confirmed, that the sensitive topic and older cohort would require a tougher sell. In the pilot, inviting 31 nominees (19% of year group) resulted in only 19 (12% of year) being trained. To enhance the likelihood of reaching the 15% ‘critical mass’, we increased the proportion of nominees invited to recruitment to 25%. The pilot and young people consultations identified that emphasising potential curriculum vitae (CV) benefits, time with friends, and the sense of responsibility and satisfaction at being nominated, might also enhance uptake.

Discussions on peer recruitment at professional workshops and TMG meetings mooted whether or not the older cohort and sensitive topic might necessitate alternative recruitment questions to those used in ASSIST. The TMG therefore designed four new nomination questions (see The STASH study intervention design for the feasibility trial) and piloted those alongside existing ASSIST questions. Learning from the pilot also highlighted the importance of practical considerations – such as allowing plenty of time for circulation and return of parent and carer consent forms – to avoid losing some potential recruits.

Peer supporter training

Intervention content

Most development activities fed into decisions about training and website content. The literature reviewed suggested that young people’s sexual health knowledge tends to converge around risk, STIs and contraception, and that current gaps in SRE include communication, consent, pornography, sexual pleasure and online literacy.6,77 Stakeholder consultation highlighted that current SRE packages are outdated and lack information on contemporary issues, such as consent and coercion, a view echoed in the young people consultations. Young people also expressed a desire for relatable experiential information and, echoing the literature, suggested that the use of humour78 and interactive components would enhance training and online content. In collaboration with the trainers, we integrated each of these considerations into content design.

The online resource review identified a vast array of training activities and websites aimed at young people, of varying relevance to the STASH study. Selection from among these was based on the extent to which resources linked to our key content-guiding theories (e.g. sought to change perception of social norms, boost self-efficacy and support new behaviours) (see Appendix 1), as well as their likely appeal to young people and their potential for interactive learning. Several shortlisted resources were tested with young people in the development phase. Among these were two key videos that conveyed experiential information and behaviour modelling. [‘Ryan and Natalie’ uses a gently comedic approach to illustrate a young couple navigating their first sexual encounter [URL: www.truetube.co.uk/film/screwball (accessed 4 August 2020)], whereas ‘Tea and Consent’ [URL: https://www.youtube.com/watch?v=pZwvrxVavnQ (accessed 4 August 2020)], developed for Thames Valley Police, also uses humour and a tea metaphor to explain sexual consent.] We identified Hannah Witton as an important ‘influencer’ [URL: https://hannahwitton.com (accessed 4 May 2020)] and that the BISH (best in sexual health) website had strong appeal to young people [URL: www.bishuk.com (accessed 4 May 2020)].

Some activity types were excluded or reduced post pilot. Role-plays, in particular, were used more selectively in the feasibility study, following observation of PS discomfort in performing these during the pilot training.

Values

Our theoretical grounding suggested that a clear set of values should be conveyed to participants, so that they could judge congruence with their own priorities and understandings.60 We therefore included an overview of ‘STASH study values’ within the PS training. These values drew on the literature, policy and stakeholder consultations. Foremost among these were aims to address negative gender norms,23,79 to take a broadly ‘sex-positive’ rather than risk-focused approach and to draw on a rights-based discourse of sexual health.80 A rights-based approach was familiar to students and teachers from the current Scottish curriculum.

Peer supporter role

Theory-guiding development, expert consultations and findings from the pilot similarly highlighted the necessity of clearly articulated role expectations for PSs. From pilot to feasibility trial, the role was refined and distilled to four key components: (1) give information, (2) influence attitudes, (3) promote the STASH study website and (4) signpost friends to trusted adults or sexual health services. These components were revisited repeatedly throughout training. Beyond these, the other key feature of the PS role was to establish and post in a secret Facebook group.

During the development stage, concerns were raised by professional workshop participants and education stakeholders about whether or not the role required too much of students. Fears coalesced around whether PSs would have the requisite maturity to handle potentially inappropriate responses from peers (including online ridicule) and whether or not they could appropriately manage potentially sensitive disclosures, including those raising safeguarding or child protection issues. To signal the responsibility inherent in the role, PSs were asked to sign a ‘STASH study charter’ on completion of the training. This reiterated role expectations and the STASH study values. The training included a session on recognising and handling sensitive disclosures, with emphasis on swift referral to an appropriate adult if anything gave cause for concern. We emphasised that PSs should not offer any sort of counselling beyond what would typically happen in their everyday friendships (see Chapter 3, Assessment of harms).

Discussion of the PS role, supported by the literature, foregrounded the importance of building generic skills of communication and listening, essential to the PS role.81 We therefore kept the ASSIST format of building generic PS skills on training day 2, including several activities adapted from ASSIST.

Training activity format

Close consideration was given to the most effective format for all training activities, whether newly designed or drawn from ASSIST. Adaptations and revisions drew primarily on observational data from the pilot training and intervention delivery partner meetings, in which the training format was finalised. It was agreed that activities should prioritise smaller group work – which process evaluation data suggested was more effective – and that fewer topics should be covered, but in more depth (with remaining topics moved to the follow-up sessions). Based on these same sources, it was decided that use of the website and Facebook groups should be more fully integrated into the training (to improve PSs’ familiarity and confidence with content and process). It was also decided that information and skills training should be integrated across both days, rather than allocated 1 day each.

Peer supporter-led activities

Using social media

The use of social media in the STASH study was novel and, therefore, a key focus of the development stage. Our review of platforms, stakeholder consultations (patient and public involvement) and design decisions, suggested the following priorities for functionality: appeal to young people; provide options to post web links, images and/or text; provide an option to create private, invite-only groups; ease of viewing messages and message stability; potential for monitoring; and potential for interface with a website. This ruled out popular platforms such as Snapchat (messages disappear, primarily images) and Twitter (messages likely to be lost in traffic, primarily appeals to older users) (URL: www.twitter.com; Twitter, Inc., San Francisco, CA, USA). Instagram (primarily visual) at the time had no option for private groups or for a bespoke application programming interface (API) to link to the STASH study website. We ruled out WhatsApp, on the basis that this would necessitate participants sharing mobile telephone numbers with fellow group members. We also ruled out school-linked platforms [such as Glow (URL: https://glowconnect.org.uk/)], which might limit credibility and perceived student ownership of the project. Ultimately, Facebook was identified as the only platform to meet all priorities. Although not the most popular choice for the target group, consultations did suggest that the platform was widely used and that young people would be willing to utilise it for the STASH study. A priority emphasised by young people throughout the consultations was privacy and control over who would see their posts, and the ‘secret’ group functionality of Facebook addressed this particularly well.

The reviewed literature suggested that the best use of a digital platform in this context would be one that drew on dynamic and interactive components. We therefore created sharable content for the STASH study website, which participants were able to post and comment on in Facebook groups, and adapt if desired.

The STASH study website

The first website iteration required PSs to log in with individual passwords. However, pilot data suggested that this created an unnecessary barrier, should technical issues arise (e.g. forgotten password). We also subsequently realised that opening the website to all students could allow PSs to direct friends to visit for themselves. The final version, therefore, required only a school-wide password (circulated by PSs). An additional login for PSs, via a bespoke API (created by Antbits), allowed PSs to post messages directly to their Facebook groups. Website accessibility was thus increased, while maintaining a degree of security that limited contamination beyond S4 at participating schools. We reiterated at training that PSs should give the STASH study password to friends in S4 at their school only, and they were provided with STASH study cards indicating the web address to facilitate this.

Based on the pilot and professional workshop findings, the website was significantly streamlined for the feasibility trial stage. Key revisions included removal of subsections for different activity types, replacing textual messages with more visually appealing bespoke infographics and memes, reducing external links (participants tended not to click on these) and dropping conversation prompts (which were minimally engaged with). The TMG proposed ‘gamifying’ the online component of the STASH study as a means of boosting interactivity, engagement and motivation, but ultimately dropped this idea, prioritising simplicity over desirable ‘add-ons’. Instead, it was agreed that PSs could be motivated in a ‘low-tech’ way by offering a ‘PS of the week’ prize at follow-up. We also considered a search function and ‘escape’ button (allowing students to quickly close the site); however, given the significant work required to refine basic functionality, these additions were not possible within our resources.

Participants in website testing groups were positive about the revised website, particularly with the inclusion of humorous memes and brightly coloured visuals.

The final version of the STASH study website [URL: www.stashtrial.org.uk (accessed 4 May 2020)] was organised under eight topics:

1. What’s STASH all about?
2. Is my relationship good?
3. What’s normal?
4. Am I ready for sex?
5. What’s consent?
6. How do I avoid risk?
7. Are pics and porn OK?
8. How do I feel good about myself?

Each topic had a website page, comprising an overview and a series of easily sharable messages (text, memes, infographics, embedded videos, external links; see Appendix 2 for examples) and a list of support sources.

School support and intervention profile

The pilot process evaluation echoed the existing literature in highlighting that school leadership support and contact teacher amenability to the intervention were key preconditions for successful implementation. Learning from ASSIST suggested that the contact teacher should attend PS training. Although invited, the pilot contact teacher (a DHT) was able to attend for only a brief portion of the training and instead sent a pupil support staff member. This led the TMG to reflect on whether or not a less senior member of staff would be more suited to the contact teacher role. The pilot findings also generated TMG discussions about how we might increase awareness of the STASH study across the entire staff body and led to suggestions such as highlighting the STASH study at staff meetings. Post-pilot student and teacher interviews indicated that the profile of the STASH study across the school was low. The TMG discussed and agreed that at the training PSs (and the teacher present) should agree how best to ‘announce’ the STASH study to their wider year group. Following the suggestion of one PS, STASH study PS badges were provided to PSs to identify them as such.

Trainer-led activities

Adaptations from ASSIST

The ASSIST study used pro forma diaries to track PS activity and, although these helped prompt PSs, the process evaluation found that entries were sometimes fabricated or omitted and did not offer an accurate activity record.82 In the STASH study, the trainers and research team were able to monitor Facebook activity directly and a decision was made to capture face-to-face conversations via reports to trainers at weekly follow-ups, rather than via the more cumbersome and potentially less reliable diary method (although the conversations remained self-reported).

Previous research suggested that ‘booster’ sessions enhance intervention effectiveness.65 The pilot confirmed that, as with ASSIST, regular follow-ups provided an acceptable means of ‘checking in’, troubleshooting and encouraging PSs. We extended this model by adding interactive learning activities to these sessions. This was done to supplement the training (as it was not possible to cover all relevant topics therein) and to maintain interest and momentum. Given the complexity of the issues and behaviours the STASH study sought to address, we added one further follow-up session for the feasibility stage, giving a total of five, plus a sixth session to conduct evaluation activities (PS questionnaire, group interviews).

To meet the concerns of teachers and child protection officers consulted regarding online bullying and sensitive online disclosures, we stipulated that a trainer (via a generic STASH study trainer account) should be a member of all PS groups. Pilot data suggested that this was acceptable to PSs, offering additional reassurance and opportunity to contact trainers with queries between follow-ups.

Peer supporter acknowledgement

During all intervention development stages, there was much discussion about how best to motivate PSs, as this was known to be essential.60,81,83 Informed by ASSIST, and by feedback from students and professionals, we opted to provide formal academic recognition with a University of Glasgow certificate, in addition to a £10 voucher on completion of the PS questionnaire. PSs were also given the opportunity to bank time spent on the STASH study towards attainment of a Saltire Award (Scottish Volunteer Award scheme for 12- to 25-year-olds).84 On a week-to-week basis, commitment was rewarded by a ‘PS of the week’ prize (a STASH study highlighter).

Revisiting the STASH study programme theory

As a result of the development work, we were able to more clearly articulate the critical assumptions underlying the intervention, the potential mediating factors and key contextual factors. The pre-pilot programme theory is shown in Figure 4.

The STASH study intervention design for the feasibility trial

Following development, refinement and optimisation, and ahead of the feasibility study, the final design of the STASH trial was determined as follows.

Peer nomination

All students in S4 (aged 14–16 years) are asked to complete a peer nomination questionnaire. Each school is given a different combination of nomination questions. Three are the same as those used in ASSIST:

1. Who do you respect in S4 at your school?
2. Who are good leaders in sports or other group activities in S4 at your school?
3. Who do you look up to in S4 at your school?’

Four questions are newly designed for the STASH study:

1. With whom in S4 would you feel comfortable to talk about something personal or sensitive?
2. Who in S4 is good at encouraging and persuading others to do things?
3. Whose opinion do you trust and value most in S4 at your school?
4. Who in S4 is confident at talking to people outside their friendship group?

The 25% of young people receiving most nominations, stratified by gender, are invited to a recruitment meeting.

Peer supporter recruitment

A meeting is held with nominees in each school, in which trainers explain the intervention purpose, the PS role and answer questions. The aim is to recruit 15% of the year group to participate in PS training. If recruitment attendance is poor or role uptake low (or skewed significantly towards one gender), a second recruitment meeting is held.

Two-day peer supporter training

A 2-day PS training session is held in school time, at an external venue, facilitated by Fast Forward and West Lothian Drug and Alcohol Service. The training is intended to equip PSs with the knowledge, skills and confidence required for the role; build motivation and enthusiasm for the role; generate trust and rapport within the PS group and among PSs and trainers; build sexual health knowledge and skills; improve understanding of risks and consequences; build self-esteem and self-efficacy; reinforce social support for healthy sexual norms; and boost intrinsic motivation and autonomy. PSs are required to sign a code of conduct during training and agree a plan to ‘announce’ the project to their year group.

Peer support work

The period in which PSs are active varies between schools: 5 weeks in three schools and 10 weeks in three schools.

Activities

Peer supporters are to establish ‘secret’ Facebook groups (invitation only groups with the highest privacy setting), comprising friends and the STASH study trainer. PSs are encouraged to post messages from the STASH study website to this group and to initiate face-to-face conversations centred on the STASH study messages. They are asked to alert friends to the STASH study website and to local support sources. To ensure maximum reach, PSs use STASH study cards to advertise a non-sharing version of the STASH study website, particularly to students who do not use Facebook. PSs are supported by a trainer for the duration of this period, as well as by an appointed contact teacher. As far as possible, PSs are encouraged to engage with intervention resources flexibly; for instance, they can choose which messages and links to share, and have the option of editing messages into their own words.

Trainer-led activities include moderation of group discussions and monitoring posts, supporting the PSs and facilitating follow-up meetings (weekly or fortnightly) with all PSs for the intervention duration.

Acknowledgement

Peer supporters who fulfil the requirements of the role would be provided with University of Glasgow certificates and, if they complete the online questionnaire, a £10 gift voucher. Schools may also support PSs towards attainment of an accredited Saltire Award.