Public Health Research (Feb 2024)
School-based interventions TO Prevent Dating and Relationship Violence and Gender-Based Violence: STOP-DRV-GBV systematic review
Abstract
Background Schools have a duty of care to prevent violence between students but a significant amount of dating and relationship violence and gender-based violence occurs in schools. These are important public health issues with important longitudinal consequences for young people. Objectives To understand functioning and effectiveness of school-based interventions for the prevention of dating and relationship violence and gender-based violence. Review methods We undertook a mixed-methods systematic review to synthesise different types of evidence relating to school-based interventions for the prevention of dating and relationship violence and gender-based violence to understand if, how and in what ways these interventions are effective. We searched 21 databases and 2 trial registers and undertook forwards and backwards citation chasing, author contact and other supplementary search methods. Searches identified all literature published to June 2021. All screening was undertaken in duplicate and independently, and we quality appraised all included studies. Results We included 247 reports (68 outcome evaluations, 137 process evaluations). Synthesis of intervention components produced an intervention typology: single-component, curricular, multicomponent, and multilevel programmes. Synthesis of intervention theories suggested that interventions aiming to increase students’ sense of school belonging and sense of safety in the school building could encourage increased learning of prosocial skills and increased prosocial peer norms, and so potentially reducing dating and relationship violence and gender-based violence. Synthesis of factors affecting delivery highlighted school organisation and leaders who believed in the importance of addressing dating and relationship violence/gender-based violence, along with time and resources to deliver the interventions. The ease with which the intervention could be delivered and modified was also important. Meta-analysis found stronger evidence for intervention effectiveness in reducing dating and relationship violence than for gender-based violence, with significant long-term impacts on dating and relationship violence victimisation and perpetration, and some evidence that interventions in high-income countries could be effective for reducing victimisation and perpetration of gender-based violence in the long-term. Impacts on knowledge and attitudes were primarily short-term. Network meta-analysis did not suggest superiority of any intervention type. Moderation evidence suggested interventions reduced dating and relationship violence perpetration in boys more than girls, but reduced gender-based violence perpetration more in girls. Metaregression by intervention component did not explain heterogeneity in effectiveness, but qualitative comparative analysis suggested that reducing perpetration was important to reducing victimisation, and that perpetration could be reduced via focus on interpersonal skills, guided practice and (for gender-based violence) implementation of social structural components. Limitations Despite an exhaustive search, trials may have been missed and risk of publication bias was high for several analyses. Conclusions This is the most comprehensive systematic review of school-based interventions for dating and relationship violence and gender-based violence to date. It is clear that the prevention of dating and relationship violence and gender-based violence in schools will require longer-term investment to show benefit. Future work Future research is needed to understand why intervention effectiveness appears stronger for dating and relationship violence than gender-based violence. Study registration The study is registered as PROSPERO CRD42020190463. Funding This award was funded by the National Institute for Health and Care Research (NIHR) Public Health Research programme (NIHR award ref: NIHR130144) and is published in full in Public Health Research; Vol. 12, No. 3. See the NIHR Funding and Awards website for further award information. Plain language summary Schools are places where dating and relationship violence and gender-based violence occur. Therefore, interventions conducted within schools are ideally placed to prevent and reduce dating and relationship violence and gender-based violence. We reviewed existing research on these interventions, exploring how they were expected to work, what factors affected their implementation in practice, how they had an impact on dating and relationship violence and gender-based violence, and what specific parts of the interventions were most effective and in what contexts. We defined what sort of evidence to include in the review, carried out a comprehensive search and found 247 reports on school-based interventions to prevent dating and relationship violence or gender-based violence, most of which were conducted in North America. Most interventions aimed to provide students with the knowledge, attitudes and skills needed to prevent perpetration and victimisation. They varied in complexity; some had one activity, others had multiple activities, some were integrated into the existing school curricula and others were complex in that they sought to change how schools as a whole respond to dating and relationship violence or gender-based violence. We theorised that complex interventions would bring about greater and more sustainable change, but this was not supported by our findings. The implementation of interventions was affected by factors such as school organisation and leaders who believed in the importance of addressing dating and relationship violence/gender-based violence, along with time and resources to deliver the interventions. The ease with which the intervention could be delivered and modified was also important. There was stronger evidence for intervention effectiveness in reducing dating and relationship violence than for gender-based violence, with significant long-term impacts on dating and relationship violence victimisation and perpetration. There was some evidence that interventions in high-income countries could be effective for reducing victimisation and perpetration of gender-based violence in the long-term. More complicated interventions were not more effective, and interventions tended to reduce dating and relationship violence perpetration in boys more than in girls. We have also been able to identify where there are gaps in available evidence, which may provide avenues for future research. Scientific summary Background and rationale This systematic review focused on dating and relationship violence (DRV) and gender-based violence (GBV). The long-term impacts of both for young people are numerous and, consequently, addressing them is a public health priority. Adolescence is a crucial stage for focusing on the prevention of DRV and GBV and schools are an ideal location for this, as this is where young people are socialised into gender norms and where significant amounts of DRV and GBV occurs. Schools also offer an opportunity to reach many young people who may not otherwise be accessible for intervention. Previous reviews have evaluated the effectiveness of interventions for DRV or GBV for adolescents but do not specifically focus on interventions in schools or analyse outcomes jointly. Aim and review questions We systematically searched for and synthesised the evidence for the following research questions (RQs): What are the theories of change and components of evaluated interventions? What factors affect the implementation of evaluated interventions? Are interventions effective and cost-effective in preventing DRV and GBV and reducing social inequalities in these outcomes? Based on the findings of RQs 1–3, what factors are important for joint effectiveness on DRV and GBV outcomes? What is the comparative effectiveness of different approaches to DRV and GBV prevention? What do the different sources of evidence suggest about intervention mechanisms and how these are contingent on context? Methods Inclusion criteria We included randomised controlled trials and process evaluations of school-based interventions for DRV and/or GBV for children between 5 and 18 years. DRV was defined as physical, sexual and emotional violence (including coercive control) in relationships between young people. GBV was defined as violence rooted in gender equality and sexuality such as harassment or bullying on the basis of gender or sexuality, sexual violence, coercion and assault including rape, within or outside dating relationships. Searching information sources The search strategy included both free-text terms and subject headings (e.g. MeSH in MEDLINE) for the school setting and DRV/GBV outcomes. In order to identify outcome, process and economic evaluations, we did not apply publication type or study design limitations. In July 2020, we searched 21 bibliographic databases from inception and without limitation on date or language: MEDLINE, EMBASE, PsycINFO, Social Policy and Practice (Ovid); Cumulative Index to Nursing and Allied Health Literature, Education Resources Information Center, British Education Index, Education Research Complete, EconLit, Criminal Justice Abstracts (EBSCOhost); Cochrane Database of Systematic Reviews and the Cochrane Central Register of Controlled Trials; NHS Economic Evaluation Database (via the Centre for Reviews and Dissemination); Social Science Citation Index and Conference Proceedings Citation Index (Web of Science, Clarivate Analytics); Australian Education Index, ProQuest Dissertations and Theses Global, Sociological Abstracts including Social Services Abstracts, Applied Social Sciences Index and Abstracts (ProQuest); Trials Register of Promoting Health Interventions and Bibliomap [Evidence for Policy and Practice Information (EPPI)-Centre]; Campbell Systematic Reviews (Campbell Collaboration). The bibliographic database searches were updated in June 2021 with a revised strategy developed to improve precision, and added further search terms for named interventions. We completed forwards and backwards citation chasing on included studies in Scopus (Elsevier), Web of Science (Clarivate Analytics) and Google Scholar, and reviewed the reference lists of relevant systematic reviews and reports. To identify linked studies and further grey literature, we conducted targeted searches in Web of Science and Scopus and searched Google Scholar for specific intervention names. We also searched publication lists on key websites. Information management and study selection Studies were screened by the title and abstract by six reviewers. The reviewers screened a random sample of 100 records and then discussed disagreements before proceeding to screening by title and abstract, and then full-text, independently and in duplicate. Data extraction and assessments of quality Data were extracted into a data extraction tool that was developed and piloted a priori. Data extracted included details about the study design, study sample, intervention characteristics, analysis methods and outcome data. Studies were appraised using the Cochrane risk of bias tool or the EPPI-Centre tool. Synthesis of theories and components We synthesised theories of change for the interventions via a staged approach: (1) using line-by-line coding and developing a coding template; (2) synthesising programme theories of change for the interventions and (3) using a meta-ethnographic approach to develop a line-of-argument for an overarching theory of change. We undertook an intervention components analysis to analyse intervention descriptions. We used an inductive approach (open coding initially and then axial coding) to comprehensively describe and categorise intervention components. Synthesis of process data Process evaluations reported qualitative data and were synthesised qualitatively using thematic synthesis methods. Synthesis of effectiveness Pairwise meta-analysis tested effectiveness on victimisation and perpetration outcomes, knowledge and attitudes, using robust variance estimation meta-analyses with random effects. Synthesis of mediation and moderator data We used harvest plots to examine how interventions impact health inequalities, focusing on ethnicity, socioeconomic position, gender, sexuality and age. We narratively synthesised findings relating to mediation. Effectiveness of different approaches to dating and relationship violence and gender-based violence prevention We used metaregression to test if components explained heterogeneity, and qualitative comparative analysis (QCA) to consider pathways to effectiveness in victimisation and perpetration. We used network meta-analysis to understand the comparative effectiveness of intervention types on DRV and GBV victimisation and perpetration. Synthesis of evidence on intervention mechanisms contingent on context To understand the links between contexts and mechanisms in generating outcomes, we followed a realist synthesis approach and used findings from the syntheses of the theories of change and process evaluations as a framework to infer and induce mechanisms from studies. This allowed previously ‘untheorised’ findings to emerge as relevant from our synthesis. Results Included studies Searches identified 40,160 records after deduplication, of which 793 were screened in full text. Of these, 247 reports were identified as eligible for inclusion, and these were coded into 68 outcome evaluations and 137 process evaluations. No economic evaluations were identified, but we examined seven cost and resource use studies. What are the theories of change and components of evaluated interventions? Most interventions focused on preventing victimisation and perpetration in DRV or GBV and fewer addressed bystander-intervention skills. Interventions had student components (e.g. group discussions, individual reflection), staff components (e.g. training, lesson plans), parent/family components (e.g. involvement in intervention) and school structures or physical environment (e.g. changes to school policies). Interventions were categorised as single-component interventions (usually short and focused on a novel technology or activity) and generally focused on a single change mechanism; curricular programmes integrated into the wider school curriculum, which generally focused on one or more change mechanisms at the student level; multicomponent programmes involving multiple modes of intervention and focusing on multiple change mechanisms operating at the student or staff level; and multilevel programmes, which were complex interventions involving multiple modes focused on change mechanisms within schools at multiple levels including the individual, classroom and school structural/environment context. The interventions were theorised as aiming to trigger a complex set of mechanisms to promote students’ school belonging, engagement with pro-social behaviours and avoidance of DRV/GBV behaviours. This was theorised to occur through interventions that ‘weakened classification’ and ‘reframing’. ‘Weakened classification’ involved strengthening relationships between and among staff and students, between the classroom and the wider school, and between schools and their communities. ‘Reframing’ aimed to increase student involvement in decisions at the level of the classroom and the school. Both ‘weakening classification’ and ‘reframing’ were then theorised as increasing student belonging and a sense of safety in the school building which could encourage increased learning of prosocial skills and interactions. Not all interventions addressed all of these mechanisms; some multilevel interventions could trigger mechanisms at multiple levels of the school system; classroom-level interventions could trigger mechanisms at the level of staff–student relationships or could focus on mechanisms that sought, for example, to promote specific skills for preventing or reducing DRV/GBV. It was theorised that multilevel interventions triggering a complex set of mechanisms across multiple levels would achieve larger effects (e.g. at the school level) and be more sustainable than mechanisms at the individual and group level. What factors affect the implementation of evaluated interventions? Key factors influencing the implementation of interventions were school resources and infrastructure, space and supplies, school organisation and leadership, perceived importance of addressing DRV/GBV, intervention interactivity and development of positive relations among students, facilitator content knowledge and availability of external support, ease of delivery and modification of interventions to suit the particularities of settings. Significant barriers to implementation included time constraints and competing priorities. However, strong staff commitment to prevention of DRV/GBV could offset time and resource limitations. Interventions may be best delivered when they align with school organisational readiness which includes a receptive school climate, staff buy-in and/or strong school leadership. Are interventions effective in preventing dating and relationship violence and gender-based violence and reducing social inequalities in these outcomes? Significant long-term, but not short-term, impacts on DRV victimisation [odds ratio (OR) = 0.82, 95% confidence interval (CI) (0.68 to 0.99)] and DRV perpetration [OR = 0.78, 95% CI (0.64 to 0.94)] were found. Although there was no overall effect for GBV victimisation [long-term OR = 0.93, 95% CI (0.80 to 1.08)] or perpetration [long-term OR = 0.90, 95% CI (0.73 to 1.12)], there was some evidence that interventions in high-income countries could be effective for reducing victimisation and perpetration of GBV in the long-term, and that the proportion of girls in the trial sample moderated the effect of the interventions for DRV and GBV victimisation, but not DRV or GBV perpetration. This may be where a critical mass of female students might have encouraged greater overall student engagement with the intervention. There was evidence for gender moderating programme effects on DRV perpetration with greater benefits for boys, particularly for emotional and physical DRV perpetration. Interventions improved DRV violence acceptance, knowledge and attitudes to personal help-seeking in the short term only. Interventions improved GBV violence acceptance, knowledge and individual self-efficacy in the short term, and violence acceptance in the long-term. What factors are important for joint effectiveness on dating and relationship violence and gender-based violence outcomes? Metaregression of intervention components did not explain heterogeneity in intervention effectiveness. We were able to estimate QCA models for short-term and long-term DRV victimisation, long-term DRV perpetration and short-term GBV victimisation and perpetration. A key finding from the QCA models for victimisation is that a central causal condition for reduction of victimisation is reduction of perpetration. There were a number of other pathways to the reduction of victimisation, namely the inclusion of single-gender components or a critical mass of girls. A critical mass of girls was especially important where interventions went beyond single components. There was also some evidence that components that were absent, for example, the absence of parental involvement was central to achieving effectiveness for long-term DRV victimisation. Perhaps the absence of such a component reduces opportunities to minimise GBV and DRV, or to receive conflicting messages about their importance. For long-term DRV perpetration and short-term GBV perpetration, interventions that were most effective incorporated a range of opportunities for guided practice of skills and attitudes, and interpersonal components focusing on student relationships. The implementation of social structural components was central to effectiveness for short-term GBV perpetration, but not for DRV perpetration. What is the comparative effectiveness of different approaches to dating and relationship violence and gender-based violence prevention? Network meta-analysis was consistent for DRV but not for GBV outcomes. No one intervention type was clearly more effective than any other, but single-component interventions may have been useful for reducing short-term and long-term DRV victimisation and perpetration, and short-term GBV victimisation. Multilevel interventions showed some effectiveness for DRV outcomes compared to other intervention formulations. For GBV outcomes, there was strong evidence for the role of curriculum interventions which were more successful than any other types at short-term follow-up victimisation outcomes, and short-term and long-term perpetration. What do the different sources of evidence suggest about intervention mechanisms and how these are contingent on context? There is evidence that the interventions in this review worked not by complex mechanisms (e.g. increasing school commitment), as initially hypothesised, but by the ‘basic safety’ mechanism which aimed to disrupt violent behaviours by communicating to students the unacceptability of violence. This simpler mechanism may have involved reductions in DRV perpetration among males. That simpler interventions if delivered well may be effective in reducing DRV would be a significant finding, especially for resource-poor settings. We argue that destabilising harmful practices is foundational whereas establishing prosocial behaviours will take more time and investment, so schools may focus on foundational mechanisms initially. Simpler interventions were easier to implement and receive so it may be more important to deliver simpler programmes achieving basic levels of school safety with fidelity than attempt more complex, multicomponent or multilevel programmes. It may be particularly important to prioritise such programmes in resource-poor settings. We also theorise that the individual-level basic safety mechanisms are more likely to effect change in DRV than in GBV perpetration because of the more dyadic, private nature of DRV. The more public nature of GBV means that it might be influenced by social norms that programmes do not seem to successfully address. We found that programmes could be effective in preventing GBV but that this was only likely in high-income settings. It may be that the transformation of school organisation and culture required to reduce GBV is beyond the reach of many schools in low- and middle-income settings because of low school organisational capacity and higher levels of inequality and GBV. Conclusions The evidence suggests that a well-delivered single-component intervention may be as effective as a complex, multicomponent or multilevel intervention. Such an intervention may be more acceptable within the school environment and receive greater staff buy-in. Preventing GBV may require normative and social structural change which may not be sufficiently triggered by current interventions. Overall, the evidence is more conclusive for DRV than GBV. It is possible that differential mechanisms will need to be activated for DRV than for GBV and that existing theories of change do not adequately account for differences between DRV and GBV in terms of intervention functioning. Study registration This study is registered as PROSPERO CRD42020190463. Funding This award was funded by the National Institute for Health and Care Research (NIHR) Public Health Research programme (NIHR award ref: NIHR130144) and is published in full in Public Health Research; Vol. 12, No. 3. See the NIHR Funding and Awards website for further award information.
Keywords
