Comprehensive Sex Education Addressing Gender and Power: A Systematic Review to Investigate Implementation and Mechanisms of Impact

Searches for this review were conducted in six scientific databases: Medline, EMBASE, PsycINFO, Web of Science, ERIC, and the Cochrane Library of Systematic Reviews. The search strategy was developed iteratively based on repeated scoping searches and employed the following four concepts: programmes and interventions; sexuality education/schools; gender; power and rights (full search strategy available in online supplementary material). Synonyms and proximity operators were used to enable identification of studies that were not explicitly labelled as addressing gender and power or as evaluation studies. Additionally, we screened articles referencing the seminal Haberland review (2015) and its sibling publication (Haberland & Rogow, 2015).

The following inclusion criteria were applied for screening:

Previous work identified ‘gender and power’ content as an important working component in sex education (Haberland, 2015). Thus, even if our ultimate aim was to inform CSE programming, we included studies about sex education programmes that were not explicitly labelled as ‘CSE’, as well as other school-based interventions, as long as they included gender and power content meeting the above definition. As it has been demonstrated that a wider range of interventions in the school environment may affect (sexual) health (Shackleton et al., 2016), we expected to improve our understanding of the wider social context of the intervention by including a broader set of studies.

The following exclusion criteria were applied:

The systematic review software EPPI-reviewer was used for screening (University College London, 2017). After piloting and refinement of the screening criteria including double screening of a subset of studies, the first author conducted title and abstract and then full text screening. Items coded as ambiguous were discussed with a second author to reach a consensus.

A cluster-search was performed for evaluation studies of five programmes that were referred to multiple times in screened full-texts but were not represented among included records. Additionally, reference lists of articles included in our review were cluster-searched for sibling publications reporting on the same programmes and a Google and Google Scholar search for the programmes and lead authors of all included articles was performed to identify further relevant process-focused articles (Booth et al., 2013).

For assessment of study quality, the Critical Appraisal Skills Programme (CASP) checklist for qualitative research was used (Critical Appraisal Skills Programme, 2018). It comprises 10 questions that prompt the user to consider potential for bias, along with methodological and ethical issues. We derived scores from these questions to indicate study quality, with a 10 out of 10 indicating high quality. As we were primarily interested in qualitative results, mixed-methods implementation studies were assessed with the CASP checklist for qualitative research as well.

Data from included studies was extracted into a Microsoft Excel–based data extraction sheet. Study and intervention details, qualitative outcomes, and large sections of text covering process-related aspects (mechanisms of change, context, and implementation (MRC, 2015)) were extracted comprehensively. Data on process was extracted from the introduction, methods, results, and discussion sections of included studies and it was noted which of these sections the respective data originated from.

Based on the respective research question asked, the study design, and methods, the included studies were categorised into two mutually exclusive types. One group of studies examined programme implementation, employing quantitative and/or qualitative methods. These studies were reporting data explicitly about intervention implementation, which was defined as ‘the structures, resources and processes through which delivery is achieved, and the quantity and quality of what is delivered’ (MRC, 2015, p. 10), including contextual factors (Pfadenhauer et al., 2017). The second set of studies investigated the impacts on social and behavioural outcomes and underlying processes, employing qualitative methods. These studies had a focus on exploring the links between programme outcomes, programme characteristics, and/or mechanisms of impact. We subsequently refer to the first group of studies as ‘implementation studies’ and to the second group of studies as ‘studies exploring mechanisms of impact’. We conducted two distinct syntheses, one of each of these study types.

The synthesis of implementation studies was informed by the Context and Implementation of Complex interventions (CICI) framework (Pfadenhauer et al., 2017). Categories within this framework comprise the implementation agents (individuals concerned with running or receiving an intervention), implementation process, implementation strategies, and context (Pfadenhauer et al., 2017). Results from implementation studies were organised into the distinct implementation categories and summarised narratively.

We conducted a thematic synthesis of qualitative studies exploring programme outcomes, characteristics, and mechanisms of impact (Thomas & Harden, 2008). We conceptualised mechanisms of impact as the link between intervention activities and outcomes including ‘[p]articipant responses to, and interactions with, the intervention’ as well as mediators (MRC, 2015, p. 24).

Data analysis was undertaken at the level of the extracted data: the sections of the data extraction sheet containing data on qualitative outcomes and process from the results and discussion sections of included qualitative studies were analysed thematically. Data excerpts served as the unit of analysis for coding. Codes were developed inductively and subsequently compared across studies and grouped and regrouped together in an iterative process to develop themes, resulting in the development of an initial mindmap. This process was informed by key findings from the preceding synthesis of implementation studies and by the four programme characteristics previously identified by Haberland (2015, textbox 1), which shaped our initial understanding of relevant programme aspects of sex education with gender and power content.

These four previously identified characteristics were compared, contrasted, and linked with the newly developed themes to facilitate differentiation of the new themes as programme characteristics or potential mechanisms of impact. The themes and respective links are visualised in Fig. 2.

Where we encountered data that was not sufficiently rich to describe mechanisms of impact, we made inferences about potential mechanisms, which are identified as hypothesised mechanisms in the results section.

Eleven studies were process evaluations that focused on programme implementation and employed qualitative or mixed methods. Eight studies were qualitative studies exploring qualitative intervention outcomes, programme characteristics, and mechanisms of impact. Only four of these eight studies were explicitly referred to as evaluation studies. For the majority of included studies (n = 17), there was very little concern with study quality (Table 1).

The 19 included primary studies were conducted in 15 countries (Table 1). Most evidence was from Europe (n = 6 countries, 3 studies) and Africa (n = 6 studies from six countries), followed by North America (n = 4 studies) and Australia (n = 4). Most articles reported evaluations of locally implemented programmes targeting boys and girls.

Only five programmes were labelled by the authors as CSE or holistic sex education programmes (Boonmongkon et al., 2019; Browes, 2015; Chandra-Mouli et al., 2018; Rijsdijk et al., 2014; Wood et al., 2015). Other programmes exhibited key CSE characteristics but were not labelled as such: five articles reported on school-based interventions labelled violence prevention programmes (including dating, domestic, and gender-based violence) (Jaime et al., 2016; Joyce et al., 2019; Kearney et al., 2016; Ollis, 2017; Williams & Neville, 2017). Four programmes were explicitly called gender-transformative (Jaime et al., 2016; Sánchez-Hernández et al.,, 2018) or ‘healthy’ or ‘positive masculinities’ programmes (Claussen, 2019; Namy et al., 2015). Three programmes were sports- or PE-based (Jaime et al., 2016; Merrill et al., 2018; Sánchez-Hernández et al., 2018) and three were focused on critical media literacy or critical thinking related to gender (Berman & White, 2013; Jacobs, 2016; Jearey-Graham & Macleod, 2017). In three of these programmes, the gender and power content constituted the only sex education component of the programme (Berman & White, 2013; Jacobs, 2016; Sánchez-Hernández et al., 2018).

Programmers incorporated creative, non-conventional, and innovative teaching methods: participatory methods like role-plays and discussions were utilised in most included studies. Further methods included the following: artwork, dance, drama, film and media production (Berman & White, 2013; Jacobs, 2016; Jearey-Graham & Macleod, 2017; Namy et al., 2015). Beyond classroom-based intervention components, six studies included activities at the school (Joyce et al., 2019; Kearney et al., 2016; Namy et al., 2015; Williams & Neville, 2017) and/or community level (Chandra-Mouli et al., 2018; Robertson-James et al., 2017).

The gender and power content was delivered across a range of different school subjects, i.e. social studies, PE, home economics, health, science, language, and religious education classes (Boonmongkon et al., 2019; Sánchez-Hernández et al., 2018; Williams & Neville, 2017; Wood et al., 2015). In addition to teacher or facilitator-led programmes, some included peer mentors and student-led initiatives outside of the classroom (Berman & White, 2013; Namy et al., 2015; Williams & Neville, 2017).

Included implementation studies stressed the critical role of the implementation agent and their skill set in delivering sex education. Programmes reported in the implementation studies were delivered by teachers (Boonmongkon et al., 2019; Browes, 2015; Rijsdijk et al., 2014; Wood et al., 2015), sports coaches (Jaime et al., 2016; Merrill et al., 2018), and external facilitators (Claussen, 2019). Whole-school approaches were further supported by an external project implementer (Joyce et al., 2019; Kearney et al., 2016; Robertson-James et al., 2017). Whilst reports suggest that teacher-delivered CSE was implemented as intended when teachers participated in high-quality training focused on gender and human rights (Wood et al., 2015), teachers who were unprepared to deliver CSE were found to omit relevant programme topics and frame adolescent sexuality as a risk or problem (Boonmongkon et al., 2019), reflecting teachers’ values (Browes, 2015; Rijsdijk et al., 2014). Teacher training for CSE was thus recommended to address both teachers’ knowledge and teachers’ gender attitudes (Browes, 2015; Wood et al., 2015).

Implementation support from an external change agent was described as instrumental in mainstreaming programme content beyond the classroom, e.g. by addressing gender in school policies, providing gender training to teachers, and undertaking a gender-focused audit and staff surveys (Joyce et al., 2019; Kearney et al., 2016; Robertson-James et al., 2017). The latter were fed back to schools as part of the intervention in one study, thus serving as feedback loops enhancing an overall change process (Kearney et al., 2016).

Implementation strategies: In programmes delivered by sports coaches and other external facilitators, non-hierarchical participatory teaching strategies promoted student engagement with intervention content and supported implementation. Engagement was reportedly fostered by creating a safe space and building student-facilitator relationships, with facilitators acting in a non-authoritative, non-judgemental, approachable manner and sharing personal experiences whilst addressing real-life issues (Claussen, 2019; Jaime et al., 2016; Merrill et al., 2018). Programmes facilitated by these ‘adult allies’ (Jaime et al., 2016) were often delivered in same-sex groups by same-sex facilitators acting as role-models that encouraged student engagement (Claussen, 2019; Jaime et al., 2016; Merrill et al., 2018). Other implementation strategies included a strong focus on interaction, reflection, and discussion (Claussen, 2019; Jaime et al., 2016; Merrill et al., 2018; Wood et al., 2015) and allowing for curricular flexibility to adapt programmes to students’ needs and knowledge (Claussen, 2019; Rijsdijk et al., 2014). One report suggested that the ‘dose delivered’ in process evaluations of these programmes should consider the degree of student engagement and their relating of programme content to their experiences (Jaime et al., 2016).

In terms of implementation context, included studies suggest that CSE programming is likely to be met with contradictory messages from schools, families, and communities (Browes, 2015) and sometimes with resistance from diverse actors in these settings (Chandra-Mouli et al., 2018), especially in conservative contexts, which could impact on programme implementation. While this may restrain programme effectiveness or may lead to programme adaptations (Browes, 2015; Chandra-Mouli et al., 2018; Wood et al., 2015), studies suggested that the implementation process can be tailored to build support for these programmes: successful approaches included framing the programmes around healthy skills instead of sexuality (Chandra-Mouli et al., 2018; Wood et al., 2015), getting stakeholder and community buy-in during the programme development phase (Chandra-Mouli et al., 2018), and building support networks or enhancing pre-existing networks for the programmes in schools and communities (Joyce et al., 2019; Kearney et al., 2016; Rijsdijk et al., 2014; Robertson-James et al., 2017), including with other initiatives that promote gender equality, such as non-governmental organisations providing teacher training (Wood et al., 2015).

Six themes emerged in our analysis. Three constitute key programme characteristics: (i) student-facilitator relationship-building, (ii) student participation, and (iii) open discussions or ‘dialogues’ integrating student reflection and experience-sharing with critical content on gender and power. Three additional themes represented potential mechanisms of impact: empowerment, meaningful contextualisation of students’ experiences, and transformation of gender norms. Figure 2 depicts the themes and crosslinks, including further relevant programme characteristics identified in a previous review (Haberland, 2015).

Included studies also highlighted that sex education with gender and power content may leave entrenched norms unchanged, in particular when student exposure to the programme was limited to only a few classroom sessions (Jearey-Graham & Macleod, 2017; Namy et al., 2015), and that teachers at times reinforced gender stereotypes (Ngabaza et al., 2016), resonating with similar findings from the implementation studies.

Our narrative synthesis of findings from implementation studies largely corresponds with other summaries of aspects that facilitate intervention implementation and engagement with CSE, e.g. regarding the importance of educator training and skill, support from external facilitators, non-hierarchical participatory teaching methods to engage students in intervention activities, and a supportive school and community context (Kirby et al., 2007; UNESCO, 2018a; Vanwesenbeeck, 2020). These other reviews have further emphasised the critical role of an enabling school environment and multicomponent approaches for CSE implementation (UNESCO, 2018a; Vanwesenbeeck, 2020).

Whether teachers or external facilitators are best placed to deliver CSE is an area of active debate. For example, in a UK-focused overview of best-practices in sex and relationships education (SRE), students deemed teachers unsuitable to deliver SRE whilst teachers and SRE professionals considered teacher-led SRE to be the most sustainable model long-term (Pound et al., 2017). This is reflected in our review where studies involving outside facilitators appeared to be pilot or one-off projects, or required substantial resources (e.g. Jaime et al., 2016; Joyce et al., 2019; Kearney et al., 2016; Williams & Neville, 2017). However, our results suggest that programmes were generally more successful in empowering students and engaging them in a meaningful way when implemented by an outside facilitator. Pound et al. argue that one of the challenges of teachers implementing SRE is the breeching of boundaries between teachers and students, which may be less of a problem when outside facilitators are involved (Pound et al., 2016). Given that the overall importance of the skill level of the teacher or facilitator for achieving positive programme effects has been strongly emphasised, it remains somewhat unclear to which extent programme ‘success’ can be attributed to the programme content as opposed to the skill level of the implementing agent, in particular with respect to addressing gender and power and facilitating participatory sessions. Whilst this may present an avenue for further research work, it is promising that authors in one implementation study included in our review argued that high-quality teacher training would enable teachers with previously limited CSE teaching skills to implement progressive sex education programmes as intended (Wood et al., 2015).

In our thematic synthesis, empowerment of students emerged as a likely mechanism of impact on SRH outcomes. This is not unexpected, as empowerment is central to many CSE programmes (UNFPA, 2015; Vanwesenbeeck, 2020) and constitutes a key strategy in health promotion more generally (Laverack, 2004). In sex education, student empowerment is theorised to expand the range of ‘sexual or gendered subject positions’ (Jearey-Graham & Macleod, 2017), thus enabling health-promoting attitudes, practices, and behaviours, including sexual agency (Fields, 2008; Jearey-Graham & Macleod, 2017). Our findings suggest that fostering student engagement and egalitarian relationships in the classroom, as well as open discussions allowing students to share experiences and feel validated, led to a shift of power in classrooms and contributed to this mechanism. However, empowerment may be easier envisioned than enacted. Jessica Fields argues that even staunch CSE advocates tend to fall short of embracing the transformative and empowering potential of CSE. By resorting to narratives of danger, they eschew positive messaging that builds on students’ existing sexual knowledge, encourages sexual agency, and equips them to deal with the social challenges that are intertwined with sexuality (Fields, 2008). This is echoed by other authors who observe that even in the most progressive contexts sex education teachers fail to achieve a shift in power hierarchies that would enable student empowerment (Naezer et al., 2017; Sanjakdar, 2019). Thus, while our findings suggest student empowerment is a mechanism of impact that may ultimately affect adolescent SRH, this mechanism likely requires a very facilitative context and skilled implementer.

The second potential mechanism of impact we identified was meaningful contextualisation of students’ experiences, facilitated by open discussions that provided a forum for critical thinking on gender and power and personal reflection. Open discussions as interactive learner-centred approaches are emphasised in CSE guidance (UNESCO, 2018a), desired by students (Pound et al., 2017), and theorised to make CSE relevant for the diverse and heterogeneous SRH needs of adolescents (Engel et al., 2019). Similarly, our analyses highlight their central role in CSE programming with gender and power content. Beyond ensuring that programmes are relevant for individual students, open discussions that include critical content on gender and power also address the societal dimensions of sex and relationships, enabling meaningful contextualisation of students’ experiences in particular in relation to gender inequality, and harmful gender norms. As Jessica Fields powerfully states, ‘Sex education offers students an opportunity to grasp sexuality’s place in the context of gender, racial and class inequalities […]’ (Fields, 2008). Ensuring that programmes are linked to the social environment of participants’ lived realities is understood to make SRH interventions more effective (Wingood & DiClemente, 2000).

The power of the interactive discussions led one group of authors in our review to conclude that students needed these open ‘sexuality dialogues’ more than what’s conventionally understood as sexuality education (Jearey-Graham & Macleod, 2017). The term ‘dialogues’ originates from critical pedagogy (Sanjakdar et al., 2015). In this framework, schools are understood as sites that reinforce existing social systems and power structures—which critical pedagogy seeks to counter (Sanjakdar et al., 2015). This approach is a democratic, joint process of knowledge creation that drives on student voice and curiosity, with teachers encouraging questioning and critical thinking via ‘dialogic teaching’ (Sanjakdar, 2019; Sanjakdar et al., 2015). The critical pedagogy framework thus incorporates the programme characteristics and mechanisms we identified, including the disruption of traditional power dynamics which was linked to our empowerment mechanism. A critical pedagogy approach in CSE may therefore facilitate a better understanding of the operationalisation of these mechanisms in educational systems and improve programme implementation.

Systematic review evidence suggests that gender-transformative programmes seeking to improve diverse SRH outcomes impact behaviour more effectively than programmes without this approach (Barker et al., 2010). Interventions addressing gender norms were identified as most promising in addressing a multitude of risk factors to reduce violence against women and girls (VAWG) (Jewkes et al., 2015). In our review, whilst transformation of gender norms emerged as a potential mechanism in our thematic analysis, only few included studies reported using an explicit gender transformative approach, defined as approaches including ‘strategies to foster progressive changes in power relationships between women and men’ by WHO (2011). This reflects findings from a systematic review of reviews on engaging boys and men in SRH programming that reported an overall dearth of gender transformative programmes (Ruane-McAteer et al., 2019). Programmes to prevent VAWG and those set in low- and middle-income countries were most likely to be gender transformative compared to programmes targeting other SRH outcomes (Ruane-McAteer et al., 2019), suggesting that the potential of a gender transformative approach has not yet been harnessed across the educational programming seeking to improve adolescent SRH. Our findings also highlight a barrier to implementation of gender transformative programming: programme implementation is highly dependent on implementation agents whose values and skills influence implementation and may reproduce gender stereotypes, thus maintaining gender relations (Boonmongkon et al., 2019; Browes, 2015; Ngabaza et al., 2016; Rijsdijk et al., 2014).

This review employed a broad and comprehensive search strategy across six databases and supplemental searches to include a wide range of studies. Inclusion of studies from 15 countries across diverse world regions may support transferability of our findings across contexts. Whilst we were interested in understanding how gender and power content would work in CSE, which is considered best practice in sexuality education, we included other sex education and school-based programmes with relevant gender and power components to broaden our understanding, in particular on potential mechanisms of impact. Since all included studies still contained relevant CSE content, we are confident that our conclusions are applicable for CSE programming. Similarly, we included studies that reported relevant data on the intervention process but were not strictly process evaluations in order to draw on a larger body of evidence. These decisions led to some ambiguity at the screening stage and heterogeneity among included studies but ultimately enhanced the findings in this work, especially the thematic synthesis.

This review also has limitations. Additional cluster-searching, repeated iterative searches, snowballing, inclusion of grey literature (Booth et al., 2013), and inclusion of studies published in languages other than English may have identified additional eligible studies but were beyond the scope of this review. Relevant studies may have been excluded based on programme descriptions, which were often limited in screened reports (Ruane-McAteer et al., 2019). Nevertheless, included studies reported a wide range of observations and reinforced key findings across studies, suggesting that included studies provide reliable insights to support implementation of CSE programmes with gender and power content among adolescents.

Whilst our synthesis does not allow for causal inference on mechanisms of impact that describe how school-based CSE interventions with a gender and power component ultimately impact adolescents’ SRH outcomes, the evidence suggesting empowerment, meaningful contextualisation of students’ experiences, and transformation of gender norms as relevant mechanisms correspond with the theoretical literature and existing empirical evidence. These mechanisms are facilitated by student participation, open discussions, and student-facilitator relationship-building and rely on skills of facilitators and a supportive context for effective intervention delivery.

This research can thus contribute to the growing body of literature to inform programme design, adaptation, transferability, and the evaluation of interventions to improve adolescent SRH. The mechanisms identified in this review can inform future research in which they are empirically tested and refined, for example through linked, rigorous outcome and process evaluations investigating the pathways to change of how sex education programme content impacts on the multiplicity of SRH outcomes, which constitutes a gap in the current literature. Furthermore, as it is considered best practice to guide intervention design, implementation, and the evaluation of complex interventions such as CSE by a relevant theory of change (De Silva et al., 2014; Moore & Evans, 2017), our results can inform the evaluation of ongoing programmes and inform the theory of change of future programmes—which is currently not often made explicit in interventions to improve SRH (Ruane-McAteer et al., 2019). Our review identified only one study which incorporated an explicit focus on complexity into the evaluation of school-based CSE (Joyce et al., 2019; Kearney et al., 2016). Their results suggest that feedback loops and the evaluation itself may have an important effect on programme implementation and outcomes and should be considered in future intervention planning and theories of change (Kearney et al., 2016).

In the field of knowledge co-production in public health research, co-produced (research) knowledge has long been argued to be more relevant to research users, empower communities, increase the chance of research uptake, and to ultimately affect health outcomes, but evidence supporting this has been scarce (Oliver et al., 2019). Whilst preliminary, our findings elucidate how the non-hierarchical, discussion-based co-production of knowledge on gender, sex, and relationships in sex education interventions makes this knowledge more relevant to participants, which may inform other interventions employing co-production approaches as part of an intervention in SRH and other public health fields.

Whilst not discussed in depth, our results also show that sex education with gender and power content does leave some entrenched norms unchanged (Namy et al., 2015; Ngabaza et al., 2016), calling for wider efforts targeting these norms. This should include interventions starting at a much younger age, that is before gender norms and roles become ingrained, continuing through childhood and adolescence (Goldfarb & Lieberman, 2021), and reaching beyond the classroom, including components targeting the broader school environment (Denford et al., 2017; Vanwesenbeeck, 2020). Further evidence suggests that school-based interventions should be coupled with interventions targeting social contexts outside of school, where both policies and community-wide interventions are needed to improve access to youth-friendly SRH services, address discriminatory practices, and support equitable gender norms at scale (Denford et al., 2017; DFID PPA Learning Partnership Gender Group, 2015; Starrs et al., 2018).