A school-based social-marketing intervention to promote sexual health in English secondary schools: the Positive Choices pilot cluster RCT

Intervention funding

Intervention delivery in the feasibility assessment and pilot RCT phases was funded by the NCB, and by schools, which provided staffing for project oversight, the running of SHPCs, curriculum delivery and social-marketing teams.

Outcome measures

The pilot RCT did not aim to assess intervention effects. The pilot primary outcomes were to meet the criteria for progression to Phase III, which were that the intervention was implemented with fidelity in three or four out of four intervention schools; process evaluation indicated that the intervention was acceptable to a majority of students and staff involved in implementation; randomisation occurred and five or six out of six schools accepted randomisation and continued in the study; student questionnaire follow-up rates were ≥ 80% in five or six out of six schools; and linkage of self-report and routine administrative data on pregnancies was feasible. Secondary outcomes addressed other research questions, including the feasibility of economic evaluation.

We did, however, pilot the primary outcome and other measures that we anticipated would be used in a future Phase III RCT. The primary outcome in such a RCT was anticipated to be routine data on births and abortions at 48 months (age 16/17 years) with secondary outcomes measured via self-reports at 24 months (age 14/15 years). Routine data on abortions minimise information bias and clearly indicate unintentional pregnancy; however, some unintended pregnancies will not result in abortion, and changes in abortion rates may also reflect variations in access;41 therefore, our indicative primary outcome measure piloted in this study also encompassed routine data on live births. Although it is recognised that around half of teenage pregnancies will, to some extent, be intended,42 this outcome measure was anticipated, nonetheless, to provide a better indication of the overall impact of the intervention.

We aimed to pilot indicative secondary outcome measures to examine broader intervention effects via survey self-reports:

• pregnancy and unintended pregnancy for girls and initiation of pregnancy for boys using adapted versions of the Ripple measures41

• diagnosed sexually transmitted infections (STIs), which focused on diagnosis with common infections, using an adapted version of the Ripple measure41

• age at sexual debut, which focused on age at vaginal sexual debut with someone of the opposite sex using an adapted version of the Ripple measure41

• number of sexual partners, which focused on vaginal sex partners using an adapted Ripple measure41

• use of contraception at first and last sex, which focused on which common contraceptives, if any, were used in first and last vaginal sex, using an adapted Ripple measure41

• non-volitional sex, which examined experience within the last 12 months of forced sex using an adapted version of the Conflicts in Adolescent Dating Relationships Inventory (CADRI) measure43

• educational attainment, which is a plausible intervention effect and, for scale-up, a critical outcome of our intervention,44 but which was not possible to pilot in this study because the pilot RCT involved a shorter period of follow-up than would be the case in a Phase III RCT and so did not include the time when students took public examinations.

Mediators and covariates

Informed by our theory of change, we also piloted the following mediators using existing self-report measures:

• school-level social norms supportive of positive relationships and sexual health, using an adapted measure from the Safer Choices study22

• perceived behavioural norms about early sexual experience and use of condoms and contraception, using an adapted measure from the Safer Choices study22

• sexual health knowledge, which used an adapted version of a measure from the Share trial27

• sexual health and contraceptive skills and access, using a measure adapted from the Share study27

• sexual health service access, using a measure adapted from the Share study27

• sexual communication self-efficacy, using selected items appropriate for young people from the Sexual Communication Self-Efficacy scale45

• sexual competence, using the Natsal measure46

• communication with parents using a measure from the Ripple study41

• school engagement using the Beyond Blue School Climate Questionnaire (BBSCQ) measure47

• career/educational aspirations, using an adapted version of the measure from the Ripple study. 41

We did not pilot a measure of intentional self-regulation as originally planned because we determined that, given the intervention theory of change, it was sufficient to measure self-efficacy of sexual communication.

All of the above measures were assessed for reliability in our pilot, by reporting intracluster correlation coefficients (ICCs) for repeat measures over time and Cronbach’s alpha statistics at baseline and follow-up for scaled outcomes.

Potential confounders/moderators were SES, as measured using the Family Affluence Scale (FAS),48,49 sex and ethnicity, using standard Office for National Statistics (ONS) categories. 50 The FAS score was calculated by scoring item responses numerically, with the least affluent options being scored 0, and the item scores being summed to give a total scale score. We originally planned to pilot moderator analyses, but this was not done as these analyses would have been severely underpowered.

A detailed table describing key measures is provided in Report Supplementary Material 3.

Economic evaluation outcomes

Given the lack of previous economic evaluations in this area, the aim of the economic analysis was to examine whether or not it would be feasible to assess the cost-effectiveness of the Positive Choice intervention using a cost–consequences analysis within a Phase III trial. The objectives of the economic analysis were to estimate the costs of delivering this new intervention; collect data on health-related quality of life among students and examine response rates and data quality, and to undertake a pilot analysis of intervention impact; and make recommendations about the feasibility of undertaking a cost–consequences analysis alongside a Phase III RCT.

Within the pilot, study methods to measure the incremental cost of the intervention in a Phase III RCT study were developed and piloted. With use of a broad public and third-sector perspective, resources to be measured included those used by SEF, schools and the NHS. Within this, key interventional resources included SEF and school staff time, training events/workshops and consumables. Measures included standardised sessional checklists to monitor and document attendance, preparation and delivery time for key training events and SHPCs; logbooks e-mailed to school staff charged with intervention delivery, assessing time spent on tasks relating to intervention; and all intervention staff travel and other expenses relating to the intervention charged to a specific project grant code.

The Child Health Utility-9D (CHU9D) measure was used to assess students’ health-related quality of life. 51 The CHU9D is a validated, age-appropriate measure that was explicitly developed using children’s input and has been suggested to be more appropriate and function better than other health utility measures for children and adolescents. Student utility values were converted into utility scores using a UK valuation set. 52 CHU9D utility values are measured on a scale from 0 to 1, anchored at 1 for full health and 0 for dead. We calculated the completion rates for the CHU9D based on the number of students for whom it was possible to compute CHU9D utility scores at baseline and follow-up. We then observed the distribution of CHU9D scores at both time points, calculating summary statistics and drawing histograms showing the percentage of values at each CHU9D value. We assessed the inter-item reliability of the nine dimensions of the CHU9D using Cronbach’s alpha and the ordinal alpha.

Finally, we undertook a pilot analysis of intervention impact on CHU9D scores using unadjusted and adjusted regression analyses. We regressed (using ordinary least squares) student-level CHU9D utility scores at follow-up against whether the student was in an intervention school or not (1 = yes, 0 otherwise). We ran unadjusted and adjusted analyses. For the latter, we ran two models: the first controlled for baseline age, sex, ethnicity and SES as measured by the FAS, and the second controlled for these variables plus baseline CHU9D utility scores. Results were presented as marginal effects [mean difference (MD) in utility scores between students in intervention and those in control schools] and corresponding 95% confidence intervals (CIs).

Assessment of harms

It was considered unlikely that any harms would arise because of the intervention or the research. The pilot study was not powered to examine intervention effects (beneficial or harmful), but qualitative data were collected as part of the process evaluation to explore any potentially harmful mechanisms.

Student surveys

Baseline surveys were conducted before randomisation as students neared the end of year 8 (age 12/13 years) in June 2018, collecting data on outcome measures and potential confounders and moderators. Around 1 week before completing the surveys, students were given an information sheet about the study and the survey. Immediately prior to completing the surveys, students were given an oral description of the study and had the chance to ask questions. Students were then invited to assent to participate in data collection. As is conventional with UK RCTs in secondary schools (including RCTs of sexual health interventions),41,53,54 parents/carers were sent a letter and a detailed information sheet around 1 week before data collection and asked to contact the school or research team should they wish for their child not to participate in the RCT. Paper questionnaires were completed confidentially in classrooms, supervised by fieldworkers, with teachers remaining at the front of the class to maintain quiet and order, but unable to see student responses. Absent students were surveyed by leaving questionnaires and stamped, addressed envelopes with schools for the students to complete.

Students were resurveyed 12 months later in June 2019 as they neared the end of year 9 (age 13/14 years), collecting data on outcomes and mediators using the same consent and fieldwork procedures. Fieldworkers were blind to allocation. Based on past experience, we expected 95% baseline survey participation and 90% at follow-up. 28,53

Data linkage

In the pilot, we sought to link our self-report data with administrative data on abortions and births up to 18 months after baseline surveys, in collaboration, respectively, with the Department of Health and Social Care (DHSC) and the ONS. This was to occur by linking data on female RCT participants to routine data on statutory abortion notifications and registration of births, by staff blind to allocation. Linkage of such data had been previously conducted for observational studies26 and initial discussion with ONS had established that data linkage was feasible despite the limited identifiers attached to abortion records, and was consistent with DHSC and ONS guidance and data protection law.

In the case of obtaining birth data from ONS, the research team aimed to implement the following procedures to ensure the anonymity and confidentiality of data. The fieldwork team would securely transfer a password-protected data set of female participants’ names, postcodes and dates of birth (DOBs), to which the CTU did not have access, linked to a unique identifier code for each participant (with the data set not including any self-report survey data). ONS would then prepare a data set containing unique identifier codes (but not other identifiers) linked to any births among trial participants. Having been accredited by the ONS, Elizabeth Allen would then attend the ONS secure data centre to access and carry out analysis of birth data with these having been imported into the main anonymised trial data set. A similar process was to occur to access abortion data, but, in this case, using only DOB and postcode information and not participant names (because routine records of abortion data do not contain patient names), and would involve the secure sending of each data set, rather than in-person visits. Although ONS initially committed to providing data on births up to 18 months after baseline surveys, in the course of the project it reported that it could provide such data up to only 11 months after baseline surveys (from July 2018 to June 2019) because of the time needed to undertake internal quality checks before releasing data.

Process evaluation

Ethics review

Ethics approval for the study was obtained from the LSHTM Ethics Committee on 21 March 2017.

Policies and checks

Any member of the research/fieldwork team visiting a school was required to have a full Disclosure and Barring Services check. All work was carried out in accordance with guidelines laid down by the Economic and Social Research Council, the Data Protection Act 1998, the latest Directive on Good Clinical Practice (2005/28/EC) and the General Data Protection Regulation 2018.

Informed consent/assent

Informed, written consent was sought from head teachers for random allocation and intervention. Informed written consent (for adults) and assent (for young people) was sought for participation in research activities from all research participants judged competent to provide this. We sought advice from Professor Richard Ashcroft, Professor of Medical Ethics at Queen Mary University of London, who is an expert on informed consent. He advised that in the case of secondary-school students, unless individuals were deemed not competent by teachers to provide it, assent for participation should be sought from the young people themselves.

For surveys, interviews and focus groups, observations and audio-recordings, participants were given an information sheet around 1 week before data collection. Just before data collection, participants also received an oral description of the study and had the chance to ask questions. Participants were advised that participation was voluntary and that they could withdraw at any point. All participants were advised that they were free to withhold consent and this matter would not be communicated to teachers or, in the case of staff participants, their managers. Students opting not to participate in surveys were offered alternative activities in the classroom. Those opting out of other data collection were free to continue with their normal activities.

The research also involved the piloting of the linkage of student survey data to administrative data on births and abortions with, respectively, the ONS and DHSC. Female survey participants were informed of this process as part of consent procedures for follow-up surveys, and their specific assent for this was sought.

In addition, students’ parents were contacted by letter 1 week prior to research activities, informing them about informed consent and providing them with the option of withdrawing their child by contacting the school or the research team. Parental consent for intervention was not required because this was a universal intervention delivered in schools as part of broader educational activities.

Confidentiality, safeguarding and serious adverse events

All participants, including students, were informed in the consent materials of the confidentiality with which the information they provided would be treated, as well as the circumstances in which we would need to breach confidentiality to ensure student safeguarding. In collaboration with a qualified social worker specialising in child protection, we developed a priori categories of abuse that, if reported through the research, would necessitate our breaching confidentiality to ensure that individuals were offered care and protection: if they reported sex at age ≤ 12 years or if they wrote anything on the questionnaire that would lead us to believe that they were at risk of ongoing serious harm. These criteria were established so that we could balance our ethical duty of promoting participant autonomy by respecting confidentiality and our ethical duty of safeguarding participant well-being when we determined that we needed to breach confidentiality to address abuse that appeared to be serious and ongoing. It was planned that where such abuse was reported through a questionnaire, we would contact the safeguarding lead at the school. Where such abuse was reported directly to research staff, we planned to first discuss the need for a response with the research participant prior to contacting the school safeguarding lead (see Appendix 3).

Qualitative research (i.e. interviews, focus groups and observations) did not ask staff or students about their experience of sex; however, if participants nonetheless described any sexual abuse, or otherwise became upset in any way, our researchers were trained in how to respond. In the case of focus groups, our researchers were trained to ensure that discussions did not move in the direction of personal disclosures of sexual behaviour since this was not the purpose of the groups and it would have been very difficult to ensure that all focus group participants did not talk about such disclosures outside the group. Our staff were trained to identify the potential for such disclosures, working to avoid them but then approaching participants immediately after the focus group to offer support and to assess whether or not any other response was needed, using the same procedures as described above.

In addition, we requested information on SAEs from participating schools and assessed the plausibility that any of the SAEs were related to intervention or research activities in collaboration with school safeguarding leads.

It was planned that the SSC and LSHTM Ethics Committee would be provided with anonymised reports of all disclosures of serious abuse and SAEs, categorised by type, circumstances and the extent of any possible connection with intervention or research activities.

In each school and within NCB, a senior member of staff was identified who was not directly involved with the intervention and whom staff or students could go to if they had complaints about any elements of the research study. This was communicated to students outside the research process to increase trust that this was truly independent.

Data anonymity and security

Quantitative and qualitative data were managed by project staff using secure data management systems and stored anonymously using participant identification numbers. Quantitative data were managed by LSHTM’s accredited CTU. Files containing participant identification numbers and corresponding participant names were held in password-protected files separate to files containing self-report survey data, which were attached only to student identification numbers. No single researcher had access to both sets of data. The names used in qualitative data were replaced with pseudonyms in interview/focus group transcripts. In reporting the results of the process evaluation, care was taken to use quotations that did not reveal the identity of respondents. The methods by which anonymity was maintained in relation to the linkage of survey data and routine data on births and abortions are described in Data linkage.

In line with Medical Research Council guidance on personal information in medical research, we will retain all research data for 20 years after the end of the study. This is to allow secondary analyses and further research to take place, and to allow any queries or concerns about the conduct of the study to be addressed. In order to maintain the accessibility of the data, the files will be refreshed annually and upgraded, if required.

Needs survey to tailor intervention components

The original protocol had specified that a student needs survey of year 8 students (drawing on the baseline RCT survey) would be used to tailor Positive Choices to local priorities in each school. There was, however, a need for some elaboration of the content of the needs survey and how it would be used to inform the intervention. Regarding the curriculum, the protocol originally specified that, informed by the needs assessment data, SHPCs would select the order in which to deliver modules; whether to deliver them within personal, social, health and economic education (PSHE), tutor groups or by integrating them into other lessons; and whether to use materials from Positive Choices or existing materials, if these conformed to our curriculum. However, SEF advised that, in practice, not all of these things would be likely to be open to influence by the needs data or the SHPC. In which lessons the curriculum would be taught, for example, would probably already be determined by school leaders. Whether to use existing or Positive Choices materials would be determined through an assessment by SEF of the resources to which the school already had access. SEF also felt that there should be a logical order to lessons (e.g. understanding reproduction before contraception), which precluded SHPCs determining this.

It was agreed that the most useful way to use the student needs assessment was to inform the selection of curriculum lessons for each school. Students would be asked in the needs survey what RSE topics they had covered previously in school, how well they had been covered and what they wanted to cover in year 9. To ensure alignment with the intervention theory of change and that all basic SEF-recommended material was covered in a logical order, the curriculum was initially designed to include a set of five ‘essential’ lessons to be covered by all schools and a bank of 10 ‘add-on’ lessons, five of which could be selected by the SHPC based on the student needs data. This was a deviation from the protocol.

Questions on how well RSE topics had been covered in school, knowledge of conception, contraception and STIs, knowledge of services, parental communication, experiences of sending/receiving sexual imagery and sexual harassment at school were included in the needs survey (see Report Supplementary Material 5) to lend more general support for implementation of Positive Choices in schools and to inform the content of parent newsletters, the focus of student-led social-marketing activities and the review of sexual health services.

School health promotion council activities

Although the original protocol stated that the role of the SHPC would be to review local needs data and to use these to tailor each intervention component, and co-ordinate intervention delivery, the exact tasks of the SHPC (other than those set in relation to the student curriculum) were not specified. Based on a review of the Safer Choices materials and discussion between researchers and SEF, it was agreed that the main activities of the SHPC would involve reviewing the student needs data and identifying how the data could inform decisions about the implementation of each intervention component of Positive Choices; launching the intervention in schools and to parents; selecting the ‘add-on’ curriculum topics informed by the needs data; taking a role in recruiting the social-marketing team and monitoring campaign activities; preparing and distributing parent newsletters; and carrying out the school sexual health services review. These elaborations did not constitute a deviation from the protocol.

Topics covered by curriculum

Although the curriculum areas identified in the protocol reflected those recommended by SEF for inclusion in a comprehensive year 9 curriculum, they did not map exactly onto the SEF’s suggested curriculum structure and there were some additional topics such as ‘pornography and the law’, ‘female genital mutilation’ (FGM) and ‘body image’ that SEF wanted to ensure were covered for year 9. Inclusion of topics such as ‘the male/female body and reproduction’ were also considered to be essential, as these provided the building blocks for later learning, which many students may not have received in primary school or in years 7 or 8. It was agreed that the curriculum format and lesson titles suggested by SEF would be adopted, but that all topics referred to in the protocol would be embedded within the lessons to ensure alignment between the lessons and intervention theory of change. This did not constitute a deviation from the protocol.

The curriculum was initially designed within the optimisation and feasibility-testing phase to include the following five essential lessons: (1) the female/male body and reproductive organs; (2) reproduction, pregnancy and contraception; (3) STIs and safer sexual practices; (4) sexual response and pleasure; and (5) building blocks to good relationships to cover the core topics specified in the protocol. Five of the following 10 ‘add-on’ lessons were also to be selected, informed by the student needs data: (6) readiness for intimacy; (7) unsafe relationships; (8) love; (9) sexual identity, gender and orientation; (10) understanding consent; (11) pregnancy options; (12) FGM; (13) pornography and the law; (14) body image and the digital world; and (15) values in our community. However, the number of lessons and titles, and the balance between ‘essential’ and ‘add-on’ lessons were modified again when moving from the feasibility-testing to the pilot RCT phase of the study.

Student-led social marketing

The initial protocol specified the number of students that should be involved in developing student-led social-marketing campaigns (n = 12–18); that students from a range of backgrounds should be recruited to take part; the topics that campaigns should cover; and the kind of activities that the social marketing might involve. There was, however, a need to elaborate how student-led social marketing would be implemented in schools. Drawing on Andreasen’s six characteristics of social marketing,18 SEF developed the student-led social-marketing component of Positive Choices to include the following steps: selecting a campaign focus; carrying out new consumer research and/or reviewing student needs data to inform campaign messages; deciding on targeting of subgroups; selecting the marketing mix; considering the ‘exchange’/benefit to students; assessing barriers to achieving the behaviour change goal (the ‘competition’); pitching ideas to the SHPC and SLT; finalising a campaign plan; working on campaign development/implementation; and debriefing by reviewing the campaigns implemented in the school. Guidance materials were designed for each of these core steps. These elaborations did not constitute a deviation from the protocol.

The initial protocol had also stated that student social marketers would use data from the student needs survey to segment consumers based, for example, on cultural styles; however, SEF staff were not comfortable with targeting in this way as they judged that it might be offensive and stigmatising to some students. This was confirmed through consultation with Karin Coyle and with the ALPHA group (see Findings from consultation with the ALPHA group). Karin Coyle also advised that younger students might be less able to contribute to campaign activities, especially where independent work outside structured lessons was required.

Following discussions between SEF and researchers, it was agreed that segmentation and targeting would be considered only on the basis of year group and that the student-led social-marketing team should focus on including students from years 8 to 11. These decisions were reflected in the content of social-marketing materials, and questions on student ‘cultural styles’ were ultimately not included in the student needs survey. This was a deviation from the protocol.

Review of sexual health services

The protocol originally specified that SEF would review school sexual health services. However, SEF advised that the most efficient and sustainable approach to reviewing and improving local sexual health services was for SEF to develop an audit tool that schools could use to carry out their own reviews and make recommendations to improve and promote services on an ongoing basis. This was a deviation from the protocol.

Participant characteristics

Five girls and four boys from year 8, and eight staff holding various roles (including one member of the school SLT, and the school’s marketing and communications manager), participated in an initial optimisation session held at the school in June 2017. Although we had intended to hold more optimisation sessions at the school to gather feedback on draft intervention materials, this proved impossible because of significant problems with the school’s capacity, which emerged in the course of the year and are in Qualitative findings on implementation. In this section, we describe the feedback received from the initial consultation, which focused on the overall acceptability of the proposed intervention aims, components and delivery models (particularly the tasks of the SHPC); teacher preference for the format and content of the manual guiding the intervention and presentation of the needs data; students’ views on the curriculum content; and implementation of the SHPC.

Participant characteristics

The first two consultations with the ALPHA group informed optimisation. Five boys and four girls took part in the first session in July 2017. Three boys and four girls took part in the second session in April 2018. Table 2 gives a breakdown of participants by age and gender.

Views on intervention content

In the first ALPHA session in July 2017, participants proposed possible activities for the launch of Positive Choices in schools that could be incorporated as suggestions into guidance materials. These included events for parents, presentations in assemblies or in tutor time, use of social media, communications to parents, updating the school website, running RSE-themed interactive performances, using promotional screens in schools and creating a graffiti wall for questions about sexual health.

In the second session in April 2018, participants reported that they were supportive of the parent component of the Positive Choices intervention, recognising the value of informing parents about the RSE being taught in school and involving them in supporting their children’s learning at home. Consultation with ALPHA on the most appropriate models of engaging parents revealed that schools had a variety of ways of communicating with parents. These included sending letters home with children; communicating via e-mail, text messages and, in fewer cases, via social media; posting information on school websites; and holding parent events.

Some participants were sceptical about whether or not homework assignments could prompt conversations at home, suggesting that many students would avoid such homework activities because of the risk of embarrassment or breaching existing child/parent boundaries. In line with SEF’s intended plan, homework activities were selected with these considerations in mind and aimed to ease into parent/child discussion at home, focusing initially on the less sensitive topic of ‘rites of passage’ before progressing to ‘abusive and healthy relationships’ in a later assignment. Participants also suggested that there was a need for the content of parent materials to be sensitive to different home cultures. Templates for parent newsletters were therefore included in intervention materials, but the precise content and mode of delivery was left to schools.

The ALPHA members were also generally supportive of the student-led social-marketing element of Positive Choices, on the grounds that student-led campaigns could ensure messaging was relevant to young people, complementing teacher-led lessons. Resonating with SEF concerns outlined above, ALPHA participants suggested that younger students might be less competent to participate fully in student-led social-marketing campaigns. They also agreed that targeting students on the basis of knowledge, cultural style or peer group affiliation could be stigmatising. Participants further raised the importance of ensuring flexibility in the programme so that campaigns could fit the context and resources available in each school. Campaigns being genuinely student led and messaging being consistent with programme aims were also highlighted as important issues. The latter informed our decision to ensure student-led social-marketing groups were accountable to SHPCs.

Table 3 outlines how feedback from the ALPHA meetings were incorporated during optimisation. A full summary of each of the ALPHA sessions is provided as Report Supplementary Material 6.

Student needs survey

The survey was completed by 127 (80%) year 8 students. Twenty-two students did not consent to participate, six were absent and three were considered not competent to give consent because of language difficulties or special educational needs and disabilities (SEND). No students were withdrawn by parents.

Researcher observations and teacher logbooks

All three SEF-delivered training sessions were audio-recorded and observed by a researcher. Six SHPC meetings and five student-led social-marketing meetings were observed. Owing to implementation challenges with the curriculum (see Quantitative findings on implementation, Fidelity of classroom curriculum), only one curriculum lesson was observed. Staff logbooks for curriculum delivery and SHPC meetings were not completed, but was completed for the student-led social-marketing meetings. The fidelity checklist for the sexual health services review was completed by researchers in consultation with students from the SHPC.

Table 5 outlines the data collection activities that were carried out.

Staff, student and Sex Education Forum interviews

Five school staff involved in the implementation were interviewed (three as individuals and two in a paired interview). Eight year 9 pupils (five girls and three boys) who received the Positive Choices curriculum, two of whom were also involved in the SHPC, participated in a focus group discussion about their experiences. Interviews were carried out with three SEF staff members: one involved in the initial development of the SHPC and curriculum components, and training students and staff on the SHPC; one leading on the development of the curriculum, parent and student-led social-marketing components, and training school staff on these; and one contributing to the social marketing and review of sexual health services. All interviews took place following the end of the optimisation and feasibility-testing phase.

Fidelity of staff training

All SEF trainings were delivered with > 70% fidelity overall (Table 6). Time limitations affecting the curriculum training reduced the fidelity of this element.

Fidelity of school health promotion council meetings

Six SHPC meetings were held (two in term 1 and four in term 3). Although no teacher logbooks were received, according to researcher observations of all six meetings, implementation of the SHPC met fidelity across all measures, with overall fidelity of 72% (Table 7).

Fidelity of classroom curriculum

Although Positive Choices included 10 hours of classroom curriculum for year 9 students, the collaborating school experienced a number of challenges to implementation (outlined in more detail in Curriculum delivery) and was only able to deliver the five lessons specified as ‘essential’ in the initial intervention design. These were the female/male body and functions of reproductive organs; fertility and contraception; STIs and safer sex; building blocks to good relationships; and consent.

Overall fidelity of the lesson that was observed by a researcher was poor, at only 54% (Table 8). Possible reasons for this are explored in more detail using qualitative data in Qualitative findings on implementation.

Fidelity of student-led social marketing

Fidelity of the student-led social-marketing meetings was high at 82% overall (Table 9). However, despite their investment in planning a campaign, students were unable to deliver their final production in school. The reasons for this are explored in more detail using qualitative data in Qualitative findings on implementation.

Fidelity of school sexual health services review

The SHPC completed the review of school-based and other local sexual health services. According to the logbook completed with students, the overall fidelity for implementation of this component of the programme was 83% (Table 10).

Overall fidelity of implementation

Overall, the participating school implemented Positive Choices with 80% fidelity.

Acceptability

Staff and student acceptability of the optimised Positive Choices intervention was quantitatively assessed via two questions included in staff interviews and the focus group with year 9 pupils. These were:

1. Do you think your school should deliver the Positive Choices programme again?

2. Do you think other schools should?

In response to these questions, staff and students unanimously (100%) agreed that their school should deliver Positive Choices again and that it should be delivered in other schools.

Staff views on implementation and feasibility

Views of Sex Education Forum staff on implementation and feasibility

Staff from the SEF were similarly concerned about the fact that teaching support staff rather than trained teachers were being selected by the school to be trained to deliver the curriculum:

It shouldn’t be taught by non-teaching staff, it’s got to insist on teachers coming on the training and not support staff, otherwise you’ll get support staff teaching, and if they don’t do that for other subjects they shouldn’t do it for Positive Choices.

SEF trainer

The trainer felt that this exemplified the low priority Positive Choices and RSE more generally were afforded in the school under new management. SEF staff felt that the change in management and crisis situation in the school had rendered it a poor site to test feasibility of the intervention, notwithstanding the fidelity results being good enough for progress to the pilot RCT.

Staff view on intervention acceptability

Staff were very positive about the intervention and saw the content as important and necessary for their students. Staff were particularly enthusiastic about the student participation element of the intervention, which they suggested had given students a sense of empowerment and ownership over the programme and made sure it was relevant to their experience:

I think it’s a good way of getting the students involved in it and making them feel like they are creating something for themselves. It’s all about empowering students, getting them to be in a leadership position, I think they feel that.

SLT lead

The social-marketing facilitator commented on the value of providing students who were usually less vocal or engaged with school with an alternative space to express themselves and flourish:

So, I think like for the boy who was very quiet but also quite astute . . . he came religiously to every meeting and would ask me about when the next one was. And this is a boy that’s very, you know, that was a shining example of someone who doesn’t get tapped up to be part of these groups and really flourished . . . And even one of the girls . . . who was reluctant to separate from her friend during the workshop but then eventually came to me afterwards and said how much she’d enjoyed it. So, I think it’s giving students who don’t always get tapped up to take part in things opportunity to express themselves and I thought they did really well overall.

Social-marketing facilitator

Staff reported that they thought Positive Choices had had an impact on student knowledge, with the lesson on sexual consent in particular increasing awareness and reporting of sexual harassment in school:

‘Yes, I would [link the increase in reporting] to Positive Choices because actually, the consent stuff was delivered in the April . . . And from that point the girls were actually making disclosures to counsellors . . . So I think it was giving them a little bit more self-confidence to say, well actually . . . you shouldn’t be doing that to me and I’m going to tell someone about that . . . So a positive impact on the girls and in theory, a positive impact on the boys because hopefully they now know what they can and can’t do’.

SLT lead

Student views on intervention acceptability

Students were enthusiastic about Positive Choices, reporting that the curriculum covered topics more comprehensively than most RSE delivered in schools:

What they [in other schools] learnt is basically just condoms, condoms, condoms, condoms . . . We’ve had conversations like with my friends from different schools, and then we was talking about like just sex ed[ucation] in general, and how it can be uncomfortable and it’s like, ‘The only thing we’ve learnt about is like condoms,’ I was like, ‘You haven’t learnt about sexual beliefs, none of that?’

Year 9 girl

Students generally appreciated the focus on relationships, not just sex. The content on different types of abuse in relationships was particularly powerful for some female students:

I learnt like that, like, I don’t know what the word’s called again, but like controlling, how you’re being controlling, that can count as a mental abuse. Because always wanting to know where someone is and stuff like that, I didn’t realise that before, that that could count as abuse, or any different things, so I learnt that there’s different . . . it was more of an eye-opener.

Year 9 girl

Most students felt that the curriculum content was appropriate and important to enable them to protect themselves and develop their own sense of what was right for them now or in the future:

I feel like the key messages was to prepare us for the future and like if you decide, OK you want to be in a relationship now, that’s your choice, but make sure you’re doing what’s right for you and your body and like know how, where the law fits in within it, and what’s legal for you to do and what’s illegal for you to do.

Year 9 girl

Students also appreciated the opportunity to discuss with their peers in lessons, exploring other students’ views. This represented a different pedagogical approach to that used in their other lessons. Some, nevertheless, felt that they wanted more time to explore issues with each other in class:

Compared to maths, it wasn’t as fast, but I feel like there still wasn’t enough time, to become in depth and understand the situation and anything that people go through, or what they can go through.

Year 9 boy

Underscoring the value of mixed-sex teaching, one student also reported that she valued the space for discussing and for challenging the opinions of male students in the class:

When the boys were there, there was a lot of . . . controversy, so it was a lot of arguing and disagreements, and it was good arguing . . .

Year 9 girl

Births

Name, postcode and DOB data were matched to ONS births registration data for the period 1 July 2018–1 June 2019. There were no births among the trial cohort based on exact matching of name, DOB and postcode. Focusing on the subset of the births data where the mother’s year of birth was equivalent to that in the survey data (2004–5) produced 81 matches. Focusing separately on given/first and family/second names and running searches for matching values across both data sets resulted in 44 records that did not match to either a forename or a surname; 26 that matched to a forename only; 10 that matched to a surname only; and one to a surname and forename. For the one record that matched on surname and forename, neither the DOB nor the postcode matched the participant data. It was determined, therefore, that there were no probable matches in our cohort.

Abortions

Postcode and DOB information was matched by DHSC with abortions data in the Abortion Notification System (ANS) for the period from 1 June 2018 to 9 October 2019. There were no abortions among the trial cohort based on exact matching of DOB and postcode. There were 166 cases in our data set with an exact match on postcode only. Some of these matched more than once because there was more than one record with the same postcode in the ANS, giving a total of 235 individual matches on postcode only (Table 16).

However, none of these matches in the ANS had a birth year of 2004 or 2005, as was the case for all female students in our cohort, and none of the DOBs contained in our data set was considered close enough to those in the ANS to constitute a possible match. There were no matches on DOB alone.

A search of partial matches based on postcode sector, combining a postcode area, a postcode district and a single character indicating the location’s inward code (e.g. YO23 2), revealed 428 cases that matched at least one case in the ANS. There were a total of 9784 individual abortion records in the ANS database that matched on postcode sector. Thirty-eight cases (or 47 matches as a result of the same postcode sector appearing more than once in the ANS data) had a year of birth of either 2004 or 2005, for the female students in our cohort. Nineteen of these 47 matched our data on year of birth, but none matched on day or month. From this, we concluded that there were also no probable partial matches in our cohort.

Completion rates and utility scores

Completion rate for the CHU9D was 1058 (81%) at baseline and 1056 (81%) at follow-up (Table 21). CHU9D data were available for 863 students at follow-up who had also completed the CHU9D at baseline. Mean [standard deviation (SD)] and median [interquartile range (IQR)] utility scores at baseline were 0.864 (SD 0.105) and 0.884 (IQR 0.808–0.952), respectively. At follow-up, these scores were 0.835 (SD 0.115) and 0.856 (IQR 0.768–0.921), respectively. The distributions of utility scores at baseline and follow-up were broadly similar (Figures 3 and 4, respectively). Minimum and maximum values were similar at both time points, as were the proportion of cases reporting maximum utility scores. The distributions of utility scores at baseline and follow-up were broadly similar, with few students reporting utility scores under 0.5, a right-skewed distribution with modal values of around 0.95, and 5–10% students reporting maximum utility scores (see Figures 3 and 4, respectively).

FIGURE 3.

Distribution of baseline CHU9D utility scores among students: 1058 observations, all respondents.

FIGURE 4.

Distribution of follow-up CHU9D utility scores among students: 1056 observations, all respondents.

Reliability

Reliability of the nine CHU9D items was assessed separately at baseline and at follow-up using Cronbach’s alpha and the ordinal alpha. In every case, it was over 0.8; at baseline, it was 0.805 (Cronbach’s alpha) and 0.874 (ordinal alpha). At follow-up, it was 0.826 (Cronbach’s alpha) and 0.876 (ordinal alpha).

Pilot analysis of intervention impact on Child Health Utility-9D scores

In the unadjusted model, utility scores were significantly lower at follow-up among students in the intervention schools, although the difference was small [marginal effect –0.020 (95% CI –0.037 to –0.004); Table 22]. After adjustment, this effect was non-significant. We reran the unadjusted model on the smaller sample that was included in the second adjusted model (841 observations) and the difference in utility scores remained statistically significant [marginal effect –0.023 (95% CI –0.041 to –0.004)]; this suggests that the non-significance in the adjusted models was due to the adjustment, rather than the different sample.

Fidelity

Intervention reach and acceptability: students

Students in intervention and control schools were asked about their awareness of new RSE provision in their school. Those in control schools were asked about this as a check on the rate of over-reporting of awareness where no new programme had been implemented. Students in intervention schools were more likely to report that their school had recently been running a new programme (Table 33). But of those reporting a new programme, large majorities in both arms reported that this was a good thing and that it included a new curriculum. Of students reporting a new intervention, those in the intervention group were more likely to report that it included parental information and help for students seeking sexual health services, but were actually less likely to report that it involved student involvement in decisions. Much higher proportions of students in intervention schools than in control schools reported coverage of various topics in RSE at school. A majority of students in intervention schools reported coverage of all topics except ‘masturbation’ and ‘love’.

Among the intervention group, there were minimal differences in reported awareness of the intervention by sex or SES as measured by FAS, but significant differences by ethnicity, with a reduction in reported awareness when comparing white, black African/black Caribbean/black British, Asian/Asian British and mixed ethnicity subgroups (Table 34).

Overall intervention acceptability: staff

Of the staff we interviewed who had been involved in implementing Positive Choices, almost all of them indicated positive views about the acceptability of the intervention in their school. Of the staff who expressed views on this (18 out of 19), 100% stated that Positive Choices was feasible and acceptable in their school. All four lead staff stated that they would use the Positive Choices teaching materials again. In all four schools, other aspects of the programme, such as the student-led social marketing and/or SHPC, were planned to continue the following year, albeit in adapted forms in three schools. Seventeen (94%) of the 18 classroom teachers interviewed stated that they enjoyed teaching most of the materials, but two (11%) expressed concerns that they were personally not best placed to have done so. Four (22%) also queried the age-appropriateness of the materials.

School context: staff and student views

Implementation, feasibility and acceptability

Control condition and potential contamination

Both control schools had recently revised their RSE programmes. The deputy head teacher of control school 1 reported that the school had over the last couple of years developed a curriculum for students from years 7 to 11, revisiting topics across years to provide age-appropriate learning.

According to the PSHE lead, control school 2 had minimal RSE teaching until recently, which covered biological aspects in science and forced marriage in religious education:

I do know that it wasn’t taught I believe because it was just lack of knowledge, lack of experience for staff. Could be taboo as well being a [faith] school, I think that’s probably the biggest worry, worrying what teachers think, worrying what the school views are, worrying what parents, so I know for a long time there wasn’t any RSE taught at the school . . .

Control school 2, PSHE lead

However, control school 2 had recently developed and piloted a new RSE programme for students in years 7–11 led by the PSHE lead with, according to the school’s PSHE lead and the SLT member leading on RSE, strong support from the head teacher and governors. The PSHE lead reported that she had attended a course at an external charity, reviewed externally produced resources and then led the development of the new programme.

In terms of consultation with parents and students about RSE provision, the deputy head teacher at control school 1 reported that student councils were consulted on students’ RSE needs, parental representatives on the governing body were informed about RSE provision and there were details on the school website. However, both the deputy head teacher and head of year 8 reported that broader parental views were not solicited. Students reported that there was little consultation with students about RSE. In control school 2, the PSHE lead reported that parents were consulted on the programme and were able to review some of the teaching materials:

Before we launched, it we spoke to parents. We invited parents in. We said to parents, ‘Look, we are going to teach this’. To be fair, all the parents are happy, there’s no concerns.

Control school 2, SLT lead

The SLT lead in control school 2 reported that implementation of the new RSE programme had been delayed to ensure that there was adequate time for consultation. The SLT lead and year 9 students agreed that student consultation had occurred through the student council, which offered strong support.

In terms of timetabling, staff in control school 1 reported that RSE lessons were delivered in 30-minute tutor-time sessions by form tutors, most of whom were teachers but some of whom, according to the head of year 8, were teaching assistants. RSE was provided via one or two sessions per fortnight. In control school 2, RSE was also taught in tutor time, comprising six sessions per year each of 25 minutes, according to the PSHE lead.

In control school 1, the deputy head teacher said that RSE was taught in tutor sessions because of the difficulty fitting PSHE into the school timetable. The head of year 8 also suggested that form tutors were more likely to engage students and be able to provide follow-up care for students when needed:

Generally, I would think yes, I suppose from the perspective of our mentors, the people the kids see every day, they tend to be people the kids trust.

Control school 1, head of year 8

Staff in control school 1 reported that teaching RSE was challenging and staff varied in their comfort teaching it. The deputy head teacher and head of year 8 both said that the quality of delivering RSE varied among staff, reflecting differences in values and aptitude. The deputy head teacher commented:

Any school that is having its PSHE programme delivered by form tutors or mentors has to acknowledge the fact that there is an inconsistency, there is a difference between your strongest and your weakest members of staff. Our responsibility as leaders is to narrow that gap through the quality of the resources that are in place, through the training the staff get and through exactly how you deliver it.

Control school 1, deputy head teacher

There was some training (according to the head of year 8, there was not very much), for example on handling disclosures of safeguarding issues and challenging inappropriate language. The deputy head teacher acknowledged that training for tutors was currently suboptimal:

And what sort of training is provided to those who are doing the RSE teaching?

At the moment, not enough, is the honest answer. A lot of the training is through their [line manager] and they meet weekly . . . And a lot of that will be as a year group training session. We do some work on our inset days that cover various bits and pieces, but we’ve identified this year that, actually . . . training needs to be, you know, a higher priority for us, especially with the SRE stuff coming in. Yeah, so it’s, it’s not bad, but it isn’t anywhere near what it could be.

Control school 1, deputy head teacher

Staff reported that, where form tutors did not feel comfortable delivering a specific lesson, more senior staff would do so. The head of year 8 reported that some teaching, for example about contraception, was taught in science and some, for example about relationships, was taught in religious education.

Control school 2 also delivered RSE primarily via form tutors. According to the PSHE lead, tutors were regarded as the best providers because of their existing relationships with students and their pastoral role:

They’ve already got that relationship with the students anyway and form tutors, you know, when you come in year 7 it’s always the first person they approach isn’t it when they’ve got a problem? The big problems of friendships and oh someone’s picking on me. And, and every time, you know, there’s messing around in lessons, the form tutors are informed about it so it’s always the central person that they come to.

Control school 2, SLT lead

Most tutors volunteered to teach RSE and received 2–3 hours of training from an external charity. The PSHE lead commented:

We invited staff and said ‘Look, if you want to become specialist teachers you are just going to teach those topics whatever time in the year, so probably once a year for 6 weeks, you’ll be teaching those topics and you’ll be trained and become a specialist RSE team.’ So, we’ve gone with that and it’s gone down quite well.

Control school 2, PSHE lead

Those staff who had opted out did so primarily because of religious objections or embarrassment. In these cases, their lessons were delivered by SLT members or science teachers who had previous experience teaching some aspects of RSE. The SLT lead commented that it was important for SLT to be involved to signal the priority given to RSE:

I think it’s good that the SLT are involved almost on a point of the students understand . . . it’s really, it’s actually really important maybe that they’re delivering it. But we want to show staff as well that it is important. So, it’s not just to the kids, it’s to staff as well to say that we should be involved with this. We’re not just telling you to do it, we see it as a significant part of students’ or young people’s lives, so you should be doing it too.

Control school 2, SLT lead

This individual added that some of the staff opting out sat in on the new lessons and some of these planned to opt in to teach RSE in future years:

There were some, well since then though, which has been a positive thing, some form tutors who were originally a little bit worried about it [but] have seen the lessons. Because they’ve observed lessons that have been going on and they’ve said ‘Actually, this next training session you’re having, can we be involved?’ Which is a good thing because we want teachers to feel comfortable about what they’re doing.

Control school 2, SLT lead

In terms of classroom activities and materials, most lessons in control school 1, according to the head of year 8, were structured around slides with some use of class surveys to assess knowledge and needs. Boys and girls were generally taught together, according to both the deputy head teacher and head of year 8, but there were some single-sex lessons, for example on consent. The deputy head teacher said that the curriculum aimed to be inclusive, for example in terms of LGBTQ identities. The deputy head teacher and head of year 8 both reported that the school used some externally produced RSE materials (the head of year 8, for example, referred to the ‘cup of tea’ video about consent), but, according to the deputy head teacher, these were of variable quality and could sometimes encourage staff to underprepare for lessons. According to the head of year 8, form tutors each led on developing materials for different PSHE topics, which were then collectively reviewed to ensure their quality.

In control school 2, RSE lessons typically involved a mix of information, whole-class and small-group discussion, videos (including the ubiquitous cup of tea/consent video) and student presentations with little written work. All RSE teaching was in mixed-sex groups, primarily for logistical reasons, according to the SLT lead. The PSHE lead produced an overview of the curriculum and the team of specialist teachers then planned lessons, with the PSHE lead providing some externally produced materials. According to the SLT lead, these were adapted for the school so that they met teachers’ individual needs and preferences. Each teacher planned and delivered lessons as they wished, depending on their preferences and what was most appropriate to each topic.

In terms of topics, year 7 students in control school 1 learnt about relationships (initially mostly focused on family and friendship relationships), consent, sexting, internet safety and grooming, self-esteem, and assertiveness, for example in challenging sexual harassment. Older students learnt about sexual health in more detail, such as safer sex, STIs, deeper learning about intimate relationships and consent (building on year 7 provision), pornography (in terms of exploitation in its production, its consequences for realistic understanding of sex and the legal status of revenge pornography). There was some RSE for year 11 students but, according to the deputy head teacher, this was limited by GCSE preparation. The head of year 8 and the school nurse reported that RSE tended to focus on ‘the basics’, for example safer sex, and did not engage more deeply with the emotional aspects of sexual relationships. The deputy head teacher reported that control school 1’s approach to RSE emphasised that sexuality was normal and healthy but did not address sexual pleasure. The RSE curriculum involved no homework.

In control school 2, year 7 students learnt about romantic relationships, family life, puberty, managing online and off-line relationships, unwanted contact and FGM. Year 8 covered sexuality, consent and contraception. Year 9 focused on healthy and unhealthy relationships, STIs, sexting and pornography. Year 10 covered different types of families, parenting skills, risk of STIs, sexting, pornography, tackling transphobia and homophobia, and managing romantic relationships. Year 11 students learnt about domestic abuse and forced marriage.

In both schools, there was also some RSE provided outside the classroom. Staff in control school 1 reported that some RSE learning was delivered via assemblies and off-curriculum days. Assemblies could involve external speakers, according to the school nurse. However, students in control school 1 reported that no external speakers came to assemblies for RSE. In control school 2, the SLT lead and year 9 students reported that an external charity came in to lead some assemblies, for example on gender identity and legal issues in sexual relationships.

In terms of student views, a focus group with students in control school 1 suggested that most sex education was actually taught in religious education (marriage, relationships, cheating, same-sex relationships) or science (anatomy, fertility and contraception, STIs). Students commented that provision varied greatly between teachers:

We haven’t really done much PSHE.

I don’t think it’s every week. I think it’s like every, once a term.

Yeah. Like we barely do it really. Like we did one the other day about like puberty and stuff but we haven’t really done, we’ve done about like three lessons on it.

I think we learn about growing up. We don’t learn about like sex education.

Control school 1, year 9 students in focus group

Students reported that some tutors appeared ill-informed and underprepared, and awkward teaching RSE and went rapidly through their slides asking students to copy down the information:

Like they go through slides and stuff but they go over like puberty and things like that but they don’t go through like sex education. And even if they do get like a slide show about it they just kind of go through the slides really quick.

Control school 1, year 9 student in focus group

Another commented:

They just, like they don’t really know about the topic before they teach it so they kind of just look at when the slides are up, they just look at it and then just say what’s on there. Like they need to know more about the situations and stuff.

Control school 1, year 9 student in focus group

Students complained about one teacher who told a student off for asking too many questions in RSE. However, students also reported that some teachers were good RSE educators. Such teachers were more confident and comfortable, and more prepared to engage in discussion and practical demonstrations (for example, of condom use). Students said of one particularly well-regarded teacher:

He like told us like everything.

He used to just tell us straight, like he didn’t sugar coat it or anything like.

He kept it real. He kept it like as a teacher like in the right boundaries but he told us straight up like.

Control school 1, year 9 student in focus group

Students in control school 2 were much more positive about RSE, and about their school and teachers in general. They reported that they had learnt about contraception, pornography, relationships, consent, STIs and LGBTQ rights, largely in weekly PSHE lessons in tutor time, but also with some learning in science (e.g. conception) and religious education lessons (e.g. different faiths’ teachings on sex and relationships). Students were particularly enthusiastic that those teaching them RSE had been trained, which they felt enhanced both their expertise and communication skills:

It’s trained.

There’s trained teachers that teach.

Trained teachers particularly like just teach us about these different things.

Control school 2, year 9 student in focus group

Some students preferred RSE to be taught by staff other than their form tutors because this enabled students to know there was an additional teacher from whom they could get support if needed. Other students preferred non-form tutors because they wanted RSE to be taught by someone not responsible for behaviour and other aspects of students’ lives at the school:

I feel like form tutors deal with our behaviour. So, say if someone’s messing about in lessons, they will talk about behaviour, they will talk about what you’re doing in school, your lessons. And I feel like if you’re going to talk about them, they’re still very open. Like my form tutor she’s like really, like she’ll talk to us about anything really. But something about like SRE, I just feel a bit awkward to talk to her about that kind of stuff, because we talk to our form tutors about different things. And then because SRE just is like a little bit of a different kind of subject, it’s a bit awkward talking to a teacher that you talk about with educational, other educational purposes, yeah.

Control school 2, year 9 student in focus group

However, some students were happy with RSE being provided by form tutors as long as the tutors were trained and comfortable with this role.

There were also some whole-school activities addressing sexual health in control schools. Control school 1 had posters and displays, for example about safeguarding and LGBTQ inclusion. The deputy head teacher said that there were some student-led campaigns that touched on some aspects of sexual health, such as transgender inclusion, but the deputy head teacher as well as the head of year 8 said that there were not, generally, student-led campaigns centrally focused on sexual health. In control school 2, the SLT and PSHE leads reported that their RSE programme included a survey of students to assess need and inform future provision:

There’s definitely elements, for example FGM, where lots of students said they don’t know about that. And that’s obviously something that we need to then obviously look at and incorporate. Well, we’re going to incorporate anyway but it’s interesting to know what they do and they don’t know.

Control school 2, SLT lead

Questions asked about personal hygiene, FGM, consent, contraception, LGBTQ identities, sexting, pornography and managing romantic relationships. Control school 2 had also engaged in work with teachers across the school to normalise the use of inclusive language, for example to recognise same-sex relationships, which was confirmed by students. The SLT lead also referred to the RSE programme enabling students to design leaflets and posters, and to lead assemblies on sexual health, including drama skits.

Sexual health services in both control schools were patchy. Control school 1 had a well-being centre staffed by a school nurse who also had safeguarding responsibilities. The nurse reported that students were generally referred to external services rather than providing STI testing and contraception on site. She reported that she initiated conversations with students who were in relationships at school:

I mean a lot of the students you know, we’re aware of them being in relationships with each other. We will often talk to them if we hear kids talk, you know, about that and make sure that they are practising safe sex, etc. and then pointing them in that direction.

Control school 1, school nurse

Students, however, reported that they were uncomfortable discussing sexual health or pregnancy with school staff, and would not access contraception or pregnancy tests via the well-being centre. They were concerned about the maintenance of confidentiality in the school and there were stories of students being visibly removed from classrooms as a result of safeguarding responses after students had disclosed sexual activity. Students commented:

I feel like if you were to go to them, it’s not really that confidential. So, like if you told them like ‘Oh, someone like did this’ then obviously they tell the other safeguarding, who tell like the well-being, who tell your parents. And then, before you know it, like everyone knows, which is like the last thing you want. And they pull you out of lesson in front of everyone. It’s just like . . . Like they just do it in front of everyone and then people like listen in on conversations. And it’s just like really like not confidential at all.

Control school 1, year 9 student in focus group

In control school 2, there was no sexual heath provision on site. Pregnancy tests, and routine or emergency contraception were not available at the school. There were sexual health services relatively nearby and these were advertised to students via assemblies led by the clinic, which year 9 students did recall. Staff reported that any disclosures of sex by students < 16 years would be shared with parents.

Staff training

For the SEF training of school staff, data were obtained from the trainers on time spent at each individual session on curriculum training and SHPC training. At two schools, curriculum and SHPC training were delivered on the same day. Using the data from the schools where it was delivered separately, the mean total trainer time per school to deliver curriculum training and SHPC training including preparation, travel and delivery time was 12 hours 52 minutes and 12 hours 15 minutes, respectively (Table 35). The daily rate per trainer was £610 plus VAT (Lucy Emerson, NCB, 2019, personal communication). Assuming an 8.5-hour working day, the hourly cost per trainer was £86.12, and the mean cost per school was £1108 for the curriculum training and £1055 for the SHPC training. In addition, on average. 4.5 members of school staff attended the curriculum training for an average of 3 hours 45 minutes and three attended SHPC training for an average of 1 hour 45 minutes. We did not collect data on the grade of each staff member attending training; assuming each staff attendee was paid at the maximum value of the upper pay range for classroom teachers (£39,406 per year)62 and worked 1265 hours per year of directed time,62 the hourly rate was £31.15. The mean cost per school for staff attendance for curriculum training was therefore £526; for the SHPC training, it was £164. Combining trainer and school staff costs, the mean cost per school was therefore £1634 for curriculum training and £1219 for SHPC training. The social-marketing training comprised a teleconference lasting 2 hours with the SEF trainer and one staff member from each school. Assuming 2 hours’ preparation time for the trainer, the trainer costs were £344.47; the school staff costs were £249.20, giving a total cost across the four schools of £593.67 and a mean cost per school of £148.42. The mean cost per school for all three types of training combined was therefore £3001.

School health promotion council

Costs were also included for each school implementing the SHPCs, using logbooks completed by teachers. Data were obtained from three intervention schools including staff preparation time (assuming preparation time involved a single staff member). Each school held three implementation meetings, with durations and numbers of staff present as shown in Table 36. Assuming an hourly cost for school staff of £31.15 (see Staff training), the costs of staff time in schools 1, 3 and 4 were £253, £132 and £548, respectively, giving a mean cost per school of £311.

Curriculum delivery

The Positive Choices intervention also included 10 hours of lessons (eight ‘essential’ lessons and two ‘add-on’ lessons chosen by each school from a selection of topics). We calculated the cost of this component based on the amount of classroom teacher time required to prepare and deliver the lessons, as recorded in logbooks completed by teachers. Data were collected from three intervention schools on the number of times each session was run by a school, the planning time needed for each lesson by the classroom teacher and the mean duration of each session (Table 37). We assumed that the planning time for each lesson was only incurred once per individual teacher, irrespective of the number of times the lesson was taught by each teacher, and calculated the total teacher time per school for each of the essential lessons. The mean total teacher time per school was 596 minutes. Assuming an hourly cost for school staff of £31.15 (see Staff training), and that the two add-on lessons had the same mean duration and planning as the essential lessons, the mean total cost per school was £3096.

Student-led social marketing

Costs were also incurred by each school through implementing the student-led social marketing, measured using logbooks completed by teachers, because it involved school staff time. Data were obtained from three intervention schools, including staff preparation time. Each school held three sessions, with preparation and delivery time and numbers of staff present as shown in Table 38. Assuming an hourly cost for school staff of £31.15 (see Staff training), the costs of staff time in schools 1, 3 and 4 were £93, £187 and £179, giving a mean cost per school of £153.

Parent information

At each intervention school, three newsletters were sent home to parents to explain project progress. Data were not collected in the present study on the resources incurred by these activities. We assume that these were sent electronically at negligible cost, and that each newsletter took 1 hour to produce by a teacher. Assuming an hourly cost for school staff of £31.15 (see Staff training), the mean cost per school was estimated to be £93.

Sexual health services review

Using logbooks completed by teachers, schools were asked to record the amount of staff time used to facilitate each step of the sexual health services review, which was included in the intervention cost. Three schools returned data and the amount of staff time for each school was 2 hours 30 minutes, 2 hours 40 minutes and 7 hours 30 minutes, respectively. The mean staff time per school for the review was, therefore, 4 hours 13 minutes. Assuming an hourly cost for school staff of £31.15 (see Staff training), the mean cost per school was estimated to be £132.

Total intervention cost

Combining the costs of each intervention component, the mean intervention cost per school was estimated to be £6786 (Table 39); the two intervention components contributing the most to this total were the staff training (44%) and the curriculum delivery (46%).