Implementing a whole-school relationships and sex education intervention to prevent dating and relationship violence: evidence from a pilot trial in English secondary schools

ABSTRACT

Adolescent dating and relationship violence is associated with health harms and is an important topic for sex education. School-based interventions addressing this have been effective in the USA, but schools in England confront pressures that might hinder implementation. We assessed the feasibility of, and contextual enablers/barriers to implementing Project Respect, a whole-school intervention. We conducted a pilot trial with process evaluation in six English secondary schools. Intervention comprised: training; policy-review; mapping and patrolling ‘hotspots’; parent information; help-seeking app; and a curriculum (including student-led campaigns) targeting dating violence. Process evaluation included assessments of fidelity and interviews with the trainer and school staff. Schools delivered training and lessons partially or completely and made parent and app information available. Two schools conducted policy reviews; none patrolled hotspots or implemented campaigns. Implementation was strengthened where staff saw dating violence as a priority. Delivery was undermined where staff were insufficiently involved, lacked time for planning or struggled to timetable lessons, and where new school challenges undermined engagement. School-based health interventions must work to build staff buy-in and ensure they do not overburden schools. Dating and relationship violence might best be addressed in this context as a broader aspect of sex education.

KEYWORDS:

• Dating and relationship violence
• school intervention
• process evaluation
• adolescents
• england

Introduction

Dating and relationship violence refers to intimate-partner violence during adolescence (Mulford and Giordano 2008; Offenhauer and Buchalter 2011), encompassing threats, emotional abuse, controlling behaviours, physical violence, and coerced, non-consensual or abusive sexual activities perpetrated by current or former ‘dating’ or ‘boyfriend’/’girlfriend’ partners (Saltzman et al. 2002). Globally, 30% of ever-partnered women report any lifetime violence from a partner, with similar prevalence among adolescents (World Health Organization 2013). Young people who have experienced dating and relationship violence are more likely to be the victims or perpetrators of relationship-violence in adulthood (Krug et al. 2002; Loh and Gidycz 2006; Exner-Cortens, Eckenrode, and Rothman 2013). Dating and relationship violence have been associated with substance use and anti-social behaviour (Exner-Cortens, Eckenrode, and Rothman 2013; Foshee et al. 2012); STIs and teenage pregnancy (Campbell 2002); eating disorders (Exner-Cortens, Eckenrode, and Rothman 2013); suicidal behaviours (Orpinas, Nahapetyan, and Truszczynski 2017) and mental-health problems (Exner-Cortens, Eckenrode, and Rothman 2013; Temple et al. 2016); physical injuries (Foshee et al. 2001); and low educational attainment (Banyard and Cross 2008). Dating and relationship violence is thus an important topic for relationships and sex education.

Universal prevention is required since dating and relationship violence is widespread and under-reported (Barter, Aghtaie, and Larkins 2014; Barter et al. 2017). Prevention during early and middle adolescence, defined, respectively, as 10–13 and 14–16 years (UNICEF 2006), is important, as this is often the period when dating behaviours begin, behavioural norms start to become established and dating and relationship violence starts to manifest (Furman and Rose 2013). Intervention to prevent dating and relationship violence needs to occur when these transitions are apparent to young people but before behaviours and norms are too established. Schools are key sites of socialisation into gender norms and are settings in which significant amounts of gender-based harassment and dating and relationship violence occur (Jamal et al. 2015). Multi-component interventions, for example, addressing school curricula, policies and environments, are promising because dating and relationship violence arises from individual deficits in communication and anger-management skills (Slaby and Guerra 1988), as well as from sexist norms and pervasive gender-based harassment (Foshee et al. 2001; Stanley et al. 2018).

Recent systematic reviews of school-based dating and relationship violence prevention, largely comprising curriculum-based interventions, have found effects on knowledge and attitudes, but not behaviour (Fellmeth, Heffernan, and Nurse et al. 2013; De La Rue et al. 2014). However, findings from two US randomised controlled trials (RCTs) suggest that multi-component interventions might be promising. In the Shifting Boundaries four-arm school cluster (RCT), schools were allocated to receive: a curriculum-only intervention; a school-environment intervention (staff patrols of hot-spots for gender-based harassment; posters; sanctions for perpetrators); curriculum plus environment components; or usual practice (Taylor et al. 2013). The environment and combined interventions were effective in reducing sexual-violence victimisation and perpetration. In the Safe Dates RCT, a dating and relationship violence prevention curriculum was delivered over ten sessions to students aged 13–15 years and focused on: the consequences of dating and relationship violence; gender roles; conflict-management skills; and student participation in drama and poster activities. A school cluster-RCT reported effects on reduced perpetration and victimisation of physical and sexual dating and relationship violence at 4-year follow-up (Foshee et al. 1998, 2004).

Recent surveys of English young people with experience of dating or relationships suggest victimisation prevalence of 22-48% for young women and 12-27% for young men aged 14–17 years (Barter, Aghtaie, and Larkins 2014). This suggests a need for prevention targeting those in early and middle adolescence informed by existing evidence. Implementing relationships and sex education and other health interventions in schools is best facilitated by committed school leaders and staff trained and supported to deliver health lessons (Pearson et al. 2015; Tancred et al. 2018). However, delivering health interventions in schools is challenging because of the limited incentives for schools to address students’ health and the lack of training and support available on how to do this (Tancred et al. 2018). Multi-component school-based health interventions depend on multiple school stakeholders (Pearson et al. 2015), and public-health professionals may have little traction to promote implementation in schools (Buchanan et al. 2005; Aarons, Hurlburt, and Horowitz 2011). In England, these challenges may be compounded by pressures on schools increasing as a result of inspections; high-stakes testing and school league tables (Sturgis, Smith, and Hughes 2006; Han and Weiss 2005); and high rates of staff turnover leading to staff trained to lead or deliver a particular intervention moving on. All of these can erode schools’ capacity and commitment to promote health (House of Commons Committee of Public Accounts 2016., Bonell et al. 2014).

With these challenges and processes in mind, we aimed to pilot Project Respect, a new multi-component whole-school relationships and sex education intervention to prevent dating and relationship violence, developed and delivered in partnership with the National Society for the Prevention of Cruelty to Children (NSPCC). Drawing on quantitative data, we examined whether Project Respect was feasible and acceptable for school staff to deliver with fidelity. Drawing on qualitative data, we examined what school contextual factors affected this. Our evaluation was informed by normalisation process theory, which proposes that the implementation of interventions is promoted by an intervention being made sense of as coherent and important by potential deliverers; these individuals ‘cognitively engaging’ with and thereby ‘buying-in’ to an intervention; deliverers engaging in collective action so that implementation is shared and coordinated; and reflexive monitoring where an intervention is formally and informally assessed as being useful and so maintained (May and Finch 2009).

Materials and methods

Design

We conducted a pilot cluster-RCT (four intervention, two control schools) with embedded process evaluation. The study protocol was registered on-line (ISRCTN65324176) and published (Meiksin et al. 2019). State secondary schools within one hour’s journey time from London or Bristol could participate. Of 437 schools invited by email to participate, 25 expressed interest. Three schools in south-east England and three in south-west England were recruited, determined by response time and purposive sampling to ensure variation by neighbourhood disadvantage, as well as school academic attainment. After baseline surveys, schools were randomly allocated 2:1 to intervention/control by the clinical trials unit of the London School of Hygiene and Tropical Medicine, stratified by region (south-east or south-west).

Intervention

In this pilot RCT, Project Respect was implemented in the 2017–2018 school-year. This new, manual-guided, multi-component whole-school universal relationships and sex education intervention was informed by previous studies (Foshee et al. 2004; Taylor et al. 2013), addressing dating and relationship violence perpetrated by girls or boys in heterosexual or same-sex relationships. The intervention was whole-school in that components included but went beyond classroom curricula (Smith et al. 2004), an approach with strong evidence of effectiveness across health outcomes (Langford, Bonell, and Jones et al. 2014).

Components comprised training by an NSPCC trainer for school senior leadership team members and other key staff to enable them to plan and deliver the intervention in their schools; training by these school staff of other school staff to prevent and respond to gender-based harassment and dating and relationship violence; senior leadership team staff reviewing school rules and policies so that these aimed to prevent and respond to gender-based harassment and dating and relationship violence; staff and students mapping ‘hotspots’ (i.e. geographical sites in the school where dating and relationship violence and gender-based harassment tended to occur); senior leadership team planning a rota of staff patrols targeting these hotspots whereby staff visit these sites to prevent or intervene in such behaviours; information for parents on preventing and responding to dating and relationship violence; distributing to students the existing, freely available ‘Circle of 6ʹ app (www.circleof6app.com), which helps individuals discreetly request help from their pre-identified contacts for support if threatened by/experiencing dating and relationship violence; and a classroom curriculum delivered by teachers in tutor group, ‘personal, social and health education’ or other sessions to students in years 9 (6 lessons) and 10 (2 lessons) aged 13–15. Lessons which were newly developed by NSPCC and informed by input from the research team focused on challenging gender norms; defining healthy relationships; inter-personal boundaries, consent, and mapping ‘hotspots’ for gender-based harassment and dating and relationship violence at school; helping friends at risk of dating and relationship violence and planning campaigns challenging gender-based harassment and dating and relationship violence; communication and anger management skills for relationships; and accessing local services relating to dating and relationship violence. Learning activities included information giving; discussion; videos; quizzes; role plays and exercises; and cooperative planning and review of student-led campaigns.

The intervention was underpinned by the theory of planned behaviour (Ajzen 2012) and the social development model (Catalano and Hawkins 1996), supported by reviews which suggest that interventions should promote the development of skills and control over behaviour, as well as challenge attitudes and perceived norms concerning gender stereotypes and violence (De La Rue et al. 2014; Fellmeth, Heffernan, and Nurse et al. 2013). The comparator condition was schools allocated to the control group, which did not implement Project Respect and continued with existing gender, violence or sexual-health-related provision.

Outcomes and data collection

In the pilot RCT, the primary outcome was whether progression to a full trial (i.e. a ‘phase-III’ RCT which aims to assess effectiveness) was justified in terms of the pre-specified criteria which included the intervention being implemented with fidelity in at least three of the four intervention schools. Fidelity is commonly measured for public-health interventions with some evidence that strong fidelity is necessary for effectiveness (Mihalic 2004). Data on implementation were collected via the audio-recording of all training; log-books completed by teachers delivering lessons which recorded what they actually delivered; structured observations of a randomly selected lesson per school; two interviews with the NSPCC trainer; and interviews with four staff per school, purposively sampled by seniority/role in implementing the intervention. Students were also interviewed; results from these interviews are reported in a forthcoming publication and are outside the scope of this paper.

Researchers arranged interviews with the NSPCC trainer directly and staff interviews were arranged by intervention schools. Log-books and observation guides monitored actual elements delivered against planned elements for the training sessions and curriculum lessons, listing planned topics and activities for each lesson with tick boxes for completion. Fidelity was defined as 100% delivery of essential elements for the NSPCC-delivered training and 75% delivery of essential elements for school-delivered components. Trained researchers conducted interviews in private rooms in schools or by telephone, using semi-structured guides. Interviews were audio-recorded and transcribed in full.

Analysis

Descriptive statistics on fidelity drew on audio-recordings, log-books and observations comparing actual to planned delivery and assessing whether this reached the threshold of 100% for NSPCC-delivered training and 75% for school-delivered training and lessons. Descriptive statistics on acceptability drew on staff interviews to give a summary indication of whether this was positive or not. Assessment drew on log-books with data from observations acting as a check on the accuracy of log-books. Qualitative data were subject to thematic content analysis using in vivo/axial codes and constant comparison to explore factors affecting feasibility (Green and Thorogood 2004). Our analysis was sensitised by normalisation process theory (May and Finch 2009) concepts of intervention sense-making and coherence, and participant cognitive engagement, collective action and reflexive monitoring. Analyses were conducted by two researchers working in parallel on different transcripts but meeting to discuss emerging themes and sub-themes and agree their overall structure.

Ethics

Ethical approval for the study was obtained from the LSHTM and NSPCC Ethics Committees. Individuals were given an information sheet one week in advance of data collection. Researchers orally described the study and individuals were given the opportunity to ask questions before deciding whether to participate. Participants were provided with information about sources of support for those experiencing dating and relationship violence or other abuse. We then sought written consent. Interviewees were informed that our safeguarding policy would require researchers to report to school safeguarding leads if interviewees suggested that a young person was at risk of serious harm.

Results

Participants

The NSPCC-delivered training was audio-recorded in all four intervention schools and school-delivered training was audio-recorded in three intervention schools; the school-delivered training did not take place in the fourth intervention school. Staff from all four intervention schools returned log-books, with the number per school ranging from four to 13. One lesson was observed in each of the three intervention schools. Two interviews were conducted with the NSPCC trainer, one mid-way through and one after implementation. Staff interviews were conducted near the end of implementation and included four staff-members in each intervention school plus one additional staff-member in one school.

Implementation fidelity and acceptability

NSPCC trained key staff in all four schools, with fidelity of 76–86% (Table 1). School-delivered training occurred in three schools with fidelity of 71–93%. Policy-review occurred in two schools. Hotspot-mapping was undertaken by staff in four schools and by students in three schools. No school modified staff patrols. Parent information was distributed and details of the Circle of 6 app provided to students in four schools. All schools delivered lessons for year-9 and year-10 students. In three schools, lessons were delivered during personal, social and health education lessons. In the fourth, lessons were delivered in tutor-group time with each lesson split into two 20-minute sessions. In two schools, the number of year-9 lessons was reduced from six to four or five. Overall, the student curriculum was delivered with fidelity of 52–98%, and with fidelity over 75% in three schools. Staff interviews suggest student-led campaigns were planned in two schools but not implemented. Observations confirmed the accuracy of data from log-books, with an agreement at school level ranging 73–100%.

Table 1. Overall fidelity of intervention in pilot.

Intervention component		Intervention schools*				Total/4 implementing with fidelity
		1	2	3	4
Training by NSPCC (100% fidelity target)	Attendance, n (sheet)	4	3	19	7	N/A
	% coverage of essential topics	86%	86%	76%	86%	0
In-school training for all staff (75% fidelity threshold)	% coverage of essential topics	93%	93%	0%	71%	2
School policies reviewed to ensure address dating and relationship violence, y/n		Y	N	Y	N	2
Potential hotspots for dating and relationship violence mapped – staff, y/n		Y	Y	Y	Y	4
Potential hotspots for dating and relationship violence mapped – student, y/n		Y	N	Y	Y	3
Reorientation of school patrol to potential hotspots, y/n		N	N	N	N	0
Parent/guardian information on dating and relationship violence disseminated, y/n		Y	Y	Y	Y	4
Student information on Circle of 6 app, y/n		Y	Y	Y	Y	4
Student curriculum, % coverage of essential topics across classes (75% fidelity target)	Year 9 Lesson 1	100%	57%	73%	88%	2
	Year 9 Lesson 2	100%	50%	89%	79%	3
	Year 9 Lesson 3	100%	36%	77%	93%	3
	Year 9 Lesson 4	88%	54%	73%	83%	3
	Year 9 Lesson 5	0%	39%	84%	86%	2
	Year 9 Lesson 6	0%	33%	55%	93%	1
	Year 10 Lesson 1	100%	79%	97%	93%	4
	Year 10 Lesson 2	100%	57%	91%	100%	3
	Overall across all lessons	98%	52%	83%	90%	3
School-delivered components delivered with fidelity, # (75% fidelity target)		7	4	4	5	1
Delivered with overall fidelity (100% NSPCC training fidelity target; 75% school-delivered fidelity target), y/n		N	N	N	N	0

*Shading indicates fidelity below target.

According to interviews with staff in intervention schools, the intervention was acceptable (described positively) to ten (59%) staff and unacceptable to two (12%) staff with three (17%) staff having mixed feelings and two (12%) being insufficiently aware of the intervention to have an opinion.

Factors affecting feasibility and acceptability

Staff engagement with the topic

This theme was informed by the concepts within the normalisation process theory of coherence and cognitive engagement. Staff were consistently interested in the topic of dating and relationship violence. However, sub-themes indicated a variable understanding of the scope and range of dating and relationship violence. Among staff from all schools, there was broad support for preventing and addressing gender- or relationship-based abuse or harassment among students. Some staff referred to specific examples where such abuse had come to the school’s attention:

I think that’s probably something that I see more, is more of a controlling aspect rather than let’s say physical violence or … controlling behaviours in general. I think that’s probably something that we see a lot more. We’ve had other pupils as well come to speak to us worried about people who’re in relationships as well that might not be considered healthy. (Assistant head of year, school 1)

Many staff commented that students often used sexist terms of abuse targeting female students and that incidents of sexual harassment were also common:

Sort of boys being heavy-handed I suppose with girls and not realising that that’s a problem. Sort of comments, snarky little comments and comments that then they don’t know are necessarily harmful, I’m trying to think of examples … Yeah, I mean the word, slag, gets, like bounced around a lot. (Teacher, school 4)

Staff commented that one reason for their strong commitment to the intervention was because it concerned safeguarding students from harm, for which schools have legal responsibility. According to one personal, social and health education coordinator, ‘The first thing that will close the school is safeguarding, not their English results.’

Staff in one school reported that while their school had systems for responding to dating and relationship violence, the school now wanted to move towards prevention. The intervention was attractive because of its universal primary prevention rather than responsive approach:

If a female, or even a male student come up to, you know, head-teacher or whoever and said, ‘You know, this, this has happened.’ You know, we would deal with it …, because we understand that that could be some form of like harassment, sexual harassment or relationship of course. But we never had sort of this Project Respect kind of make that message more widespread throughout the year-groups … Not sending out a general message of, you know, this is right, this is wrong, you know, what is consent, what is not consent and I think that’s why this has been quite good for the school because it’s sort of made kids more aware, so hopefully the number of times we have to step in reduces.” (Head of house, school 2)

However, schools took time to engage with the concept of dating and relationship violence as it was presented in intervention materials. The NSPCC trainer commented that the term was not previously used in these schools. He advised that the term ‘violence’ could cause confusion because some associated this only with physical violence, thereby eroding the coherence of the term for some other staff. He suggested that ‘abuse’ might be a better term. He also commented that the extent to which staff initially recognised whether dating and relationship violence was a problem varied with school location and staff gender:

There’s one school in the south-west where you know, there was almost a divide between the male and female staff about their views on it. And the training had a bit of a, there was a clear distinction between who got that it’s an issue and who didn’t, as in like the males, sort of didn’t as much. And I was actually pulled aside by the leader saying that they, that they struggle, they feel that they struggle with the male staff in the school.

Another sub-theme was that some staff perceived the ‘dose’ of the intervention as too large considering that dating and relationship violence was just one among many health topics that schools needed to address:

I don’t think we can commit that amount of curriculum time to it, particularly in year 9 … I would say whoever’s organising the package, they need to remember that everybody, so drugs awareness, smoking, tobacco-awareness you know, all the resources you can get are about five, six weeks. (Assistant head, school 3)

Insufficient lead-in time

Schools were informed in July whether they had been allocated to deliver the intervention the following September, and a theme apparent across interviews was that this timescale was too short. This could erode schools’ abilities, as described within normalisation process theory, to ensure broad staff buy-in to the intervention, and time to organise delivery. The short lead-in time did not give staff sufficient time to schedule times and arranging cover for training, meetings or lessons. Staff also reported that the results of hotspot-mapping could not be used to modify patrols because staffing for patrols had already been negotiated and could not be changed. As the assistant head of school 3 described it,

“The duty rota is huge. The documentation about who’s going where and what their actual duties are. And to change that massively means you’re, you can’t take somebody off one area without it affecting … So it’s difficult.”

In terms of lessons, the intervention leads in each school scrambled to work out where in year-9 and year-10 timetables, the lessons could occur. They also had to identify and secure staff agreement to deliver these lessons, often in a context of high turnover and low morale:

I think the things that really made me nervous … was the lessons. Because that team did not know that was coming their way so their planning had not been able to consider how and when they would fit in. And they became a bit of an add-on, rather than being properly incorporated to complement other lessons that they might have been delivering at the same time. So I was then in this position where I was having to get other people to do things that they didn’t know about … But we managed it. (Deputy head, school 1)

Insufficient whole-school buy-in

The NSPCC trainer perceived that in some schools, the decision to participate was taken by one individual, with insufficient buy-in from other staff. This could cause problems when, for example, other stakeholders in the schools, such as staff coordinating personal, social and health education, were not consulted, or when the lead person left the school without a plan for who would take over responsibility. Staff in one school, in particular, described poor communication at the start of the intervention:

Project Respect fell into a series of problems from the very beginning in that the member of senior leadership team who commissioned it didn’t speak to me about it and yet it was going to be taught in my curriculum. So I had no idea until September that it was happening … The person who set it up left the school and handed it over to someone who was pushing it through without actually considering whether it, you know, what needed to work on it. (personal, social and health education coordinator, school 1)

Another member of staff inherited the intervention at the start of the implementation period, who was not briefed by their predecessor on what the intervention involved, resulting in a delay to intervention activities:

We were kind of all a bit in the dark really. So [name] had left … I had no idea that it was happening. So then [name] left and then I guess [name] just kind of picked it up and was like, ‘Oh, okay, so this is happening, like I had no idea’.” (Assistant head of year, school 1)

A sub-theme across schools was that senior leadership team members were insufficiently involved in intervention activities, such as the training by NSPCC. This adversely affected the implementation of other components, such as the policy review and staff training. The NSPCC trainer commented:

I think the problem when the SLT members aren’t attending the training, standard staff wouldn’t have the responsibility of editing the policies. So that’s again making sure that the person who’s responsible for policy review is involved … I think for the success of the project I think there needs to be a commitment from the senior leadership team at the training as well because without that the implementation of the whole staff training can be a bit problematic.

Another staff-member described how the review of policies had been hampered by the senior leadership team not being sufficiently involved in the intervention:

It hasn’t got anywhere if I’m honest. I think trying to, no, I’ll rephrase that, finding the time I think to discuss with the senior leadership team has been quite difficult if I’m honest, I probably haven’t pushed it as much as I need to. (personal, social and health education coordinator, school 4)

In some schools, there was poor communication between the staff-member leading the intervention and those attending the training, so they could come to the training with little understanding of why they were there. The NSPCC trainer commented:

Yeah. I mean staff buy-in to be honest … And it’s about that communication. Because the schools where we’ve had trouble are the ones where there’s been a lack of communication from the senior leadership team down to the staff-members. So …, if we take [school] for instance, when we sat there and there’s just clearly someone massively disengaged, you know, and it’s awkward. And, you know, and then at the end it’s like well we don’t know why we’re here … So yeah, it’s that relying on schools to communicate it down to their staff.

Teacher ability to teach the curriculum

Another theme was that there was variation in the extent to which teachers who were to deliver the classroom curriculum were committed and prepared for this. In one school, lessons were delivered by teachers not specialising in relationships and sex education or personal, social and health education and lacking experience in health education. Intervention leads were candid that some teachers lacked the skills. These staff’s commitment to the intervention could also vary, with some seeing this as marginal to their particular role:

That’s an issue with all staff teaching personal, social and health education. I think that’s a whole-school issue than kind of Project Respect issue. It’s a timetabled lesson. Staff have time to teach it and time to plan for it. It was quite evident to see, as I was [observing lessons], staff that had clearly gone through and looked at the resources and were clear about what they were teaching beforehand and staff that hadn’t. (personal, social and health education coordinator, school 4)

Teachers varied in how comfortable they were delivering lessons. Intervention leads and classroom teachers acknowledged that some teachers were not comfortable addressing challenging topics or lacked the skills to facilitate participative learning methods:

“I think that there are some staff that are absolutely fabulous at delivering stuff like that. And then some others who should not be allowed anywhere near it. Because it can … be quite damaging if it’s not done the right way” (Deputy head, school 1).

Context of schools under strain

A major theme concerned school context and how this could influence schools’ commitment and capacity to implement the intervention. Staff described that most participating schools were experiencing high levels of staff turnover:

“There’s like a crazy amount of staff leaving and coming and going, yeah, it’s mad” (Assistant head of year, school 1)

Some schools were reported to be undergoing staff restructuring and redundancy programmes as a result of budget cuts, which were eroding relationships among staff:

The relationship between senior leadership team and staff at the moment is a little bit frayed and that is purely down to the fact that there is a complete restructure going on in place and I think, you know, people are very sensitive at the moment, so things like huge new initiatives we’ve actually said ‘no’. That’s not nice to be doing at the moment because staff are concerned, they’re having to be re-interviewed for jobs that they’ve been doing for years. (Assistant head, school 3)

This had resulted in staff being less willing to take on additional work such as that arising from Project Respect:

When I was thinking about getting involved with Project Respect I had no idea that we were going to this year have so much disruption, so this year we’ve gone through a … restructuring process which means that teaching and learning responsibility is extra responsibility and the money for that position, and a lot of those have been stripped out for next year, and the process off sorting that out, who’s losing what and who’s going to therefore have to do more in order to get all the jobs done that need doing has been very painful for the staff, resulting in low morale and I would say a reluctance to take any more on than they have to. (Assistant head, school 3)

The NSPCC trainer reported that implementation could be impeded when schools faced challenges such as those described above. Two schools received a downgraded inspection rating during the course of implementing the intervention. This led to a shift in management priorities to increasing educational attainment and a need to scale back their involvement in the intervention to a core package.

Staff discontinuities and low morale could also undermine teachers’ commitment to delivering the curriculum. In his second interview, following intervention delivery, the NSPCC trainer reflected on how staffing problems had meant that driving implementation was challenging across all four schools:

I didn’t anticipate it being quite as difficult to get answers … off the schools. I think that, you know, as I’ve said previously that the schools that were involved have … they seem to have all had staffing issues as the project’s gone on … When I first went into teaching, you know, you’d go into teaching and it would be a job for life, whereas now, you know, people do go through redundancy processes … So you know, the four schools that … I think three out of the four, or four … to have, you know, go through that sort of stuff in crisis and … I would say, is quite unusual.

Discussion

Project Respect, a multi-component, school-based relationships and sex education intervention focused specifically on preventing dating and relationship violence, was delivered with variable fidelity. Some components, such as policy-review and changes to school patrols, were implemented patchily or not at all. Lesson delivery for three-quarters of participating year-groups began late in the school year. The intervention was judged acceptable by just under two-thirds of the staff interviewed.

There was broad support among school leaders and teaching staff for addressing dating and relationship violence in schools and for an emphasis on prevention. This was grounded in a recognition that abusive relationships were prevalent among students and that it was the school’s role to address. However, some participants noted that this recognition was not evenly distributed among staff, which existing studies suggest might undermine implementation (Pearson et al. 2015). Uncertainty as to what was meant by the term dating and relationship violence was also a barrier to staff’s initial ‘cognitive engagement’ with the intervention. Although delivery of lessons by teachers rather than external specialists offers the most promise for embedding dating and relationship violence interventions in school curricula (Stanley et al. 2015; Ollis 2014), it was however also clear that some teachers lacked the skills or interest to deliver high-quality lessons and facilitate participative discussions on challenging topics, consistent with some previous research findings (Pound et al. 2017).

Implementation was also undermined by the short lead-in time for the intervention, which did not give schools enough time to build support and collectively plan intervention activities, and by insufficient buy-in from some school staff and a lack of involvement from some senior leadership teams. In some cases, the training for staff was not attended by senior leadership team and other intended participants. Furthermore, some participants were unsure why they had been asked to attend, suggesting communication problems within schools and between schools and the training provider. As suggested in previous studies, schools found it difficult to find space in timetables for lessons focused solely on a single health topic (Bonell, Allen, and Warren et al. 2018).

In a context of budget cuts, inspections, high-stakes testing and school league tables (Sturgis, Smith, and Hughes 2006; Han and Weiss 2005), there was evidence that stressed schools struggled to prioritise this work. Schools’ commitment could be undermined by new priorities (e.g., responding to worsened inspection ratings or exam results); and reduced staff morale (e.g., because of staff-restructuring programmes). Staff turnover was extremely high and hindered the extent to which implementation could be initiated in the autumn term and proceed incrementally over the school year. These attributes are likely to remain a feature of the English education policy landscape for some time and as previous studies have suggested, it is difficult for such challenges to be mitigated when the agency coordinating the intervention has little power to drive delivery.

Our findings identify a number of key barriers and enablers of whole-school health interventions, which resonate with normalisation process theory and previous research on the delivery of such interventions. Implementation was stimulated when staff made sense of the intervention, accepting the importance of addressing dating and relationship violence and understanding both how the intervention was intended to work and their role within its implementation. Fidelity was poorest for the policy review and reorientation of staff patrols, intervention components that schools could not align with their existing procedures and timetables, as suggested in previous research (Domitrovich et al. 2008). Implementation of this whole-school intervention was sometimes significantly undermined by some staff with a critical role in delivery who were not fully bought-in to the intervention; these included senior leadership team members who were intended to lead some components and some classroom teachers with a role in delivering the curriculum. Previous research has frequently referred to the importance of senior ‘champions’ for interventions with the commitment and authority to get things done (Pearson et al. 2015). Implementation was also undermined when schools lacked the time or leadership to develop a collective plan for intervention delivery, such that only one or two individuals were involved in leading delivery, a problem noted in some previous reports of whole-school interventions, linked to over-demanding research time-tables (Bonell, Fletcher, and Fitzgerald-Yau et al. 2013).

In terms of limitations, most elements of the process evaluation had very good response rates but the completion of log-books by staff delivering the curriculum was inconsistent. This meant that an assessment of the fidelity of delivery of this intervention component is somewhat uncertain. We assessed intervention acceptability to staff using interview-based data because our questionnaire survey of staff had a low response rate. Although we found no evidence of staff being upset by any intervention contents, we did not explore whether any staff avoided participating in the project because of the sensitivity of the subject matter.

Conclusion

In conclusion, our study suggests that there should not be an immediate proposal for a phase-III trial of this intervention. While staff showed broad support for school-based prevention of gender-based harassment and dating and relationship violence, interviews suggested that it was not feasible in some schools to implement a relationships and sex education intervention that required considerable space in the school timetable but only addressed one topic, among the multiple topics that should be addressed within comprehensive relationships and sex education. This is particularly relevant in England where new statutory relationships, sex and health education guidance for all secondary schools will be implemented from 2020 (Department for Education 2017). This suggests that if a future phase-III trial is warranted, it should focus on a broader intervention focused on comprehensive relationships and sex education and including dating and relationship violence within this. Placing dating and relationship violence in the broader context of healthy relationships, gender norms and communication skills might also make for a more powerful intervention.

Furthermore, a refined intervention should have a longer preparatory phase and a process for ensuring stronger buy-in including from senior leadership team members and classroom teachers, as well as training to ensure teachers have the skills to deliver the curriculum, with the option of some challenging topics being addressed by external specialists (Foshee et al. 2012). More generally, our study provides evidence that school-based health interventions must ensure they do not overburden schools, particularly in contexts where school systems are stressed by budgetary or staffing problems and performance pressures (Sturgis, Smith, and Hughes 2006; Han and Weiss 2005). In line with previous research, this study raises concerns that schools struggle to deliver separate interventions for each health issue affecting their students (Tancred et al. 2018), and so may be unwilling or unable to deliver an intervention focused solely on dating and relationship violence. Dating and relationship violence might best be addressed as an aspect of comprehensive relationships and sex education as suggested in recent studies (Santelli, Grilo, and Choo et al. 2018; Wolfe et al. 2009), with there being no clear evidence of a certain minimum dose for intervention effectiveness in reducing dating and relationship violence (Fellmeth, Heffernan, and Nurse et al. 2013). There is also increasing evidence that whole-school interventions aiming to ensure healthier school environments can benefit a range of outcomes simultaneously without adding large amounts of lesson time (Bonell, Allen, and Warren et al. 2018; Langford, Bonell, and Jones et al. 2014).

Data availability

Data are available on reasonable request.

Acknowledgments

We thank the NSPCC for their work developing the intervention and for their guidance on developing the child safeguarding policy for this study. We also thank David Humphreys, David Gadd, Amanda Mason-Jones and Mona Kanaan for their feedback on the study design and methods via the Study Steering Committee. We thank members of the ALPHA young researchers group based in DECIPHer, Cardiff University, for their feedback on the intervention and methods. Finally, we thank the students and staff at the schools taking part in optimising the intervention for their contributions and support.

Disclosure statement

The authors declare that they have no competing interests.

Data availability statement

The data described in this article are openly available in the Open Science Framework at DOI: 10.17605/OSF.IO/TPA6U.

Additional information

Funding

This work was funded by the National Institute for Health Research Public Health Research Programme (PHR 15/03/09). This report presents independent research commissioned by the NIHR. The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, MRC, CCF, NETSCC, the Public Health Research programme or the Department of Health. The funder had no role in study design, in the collection, analysis or interpretation of data, or in writing this manuscript.