Programme potential for the prevention of and response to sexual violence among female refugees: a literature review

Abstract Continuing international conflict has resulted in several million people seeking asylum in other countries each year, over half of whom are women. Their reception and security in overburdened camps, combined with limited information and protection, increases their risk and exposure to sexual violence (SV). This literature review explores the opportunities to address SV against female refugees, with a particular focus on low-resource settings. A systematic literature review of articles published between 2000 and 2016 was conducted following PRISMA guidelines. Databases including Medline (Ovid), PubMed, Scopus, PsychINFO, CINAHL and the Cochrane Library. Grey literature from key refugee websites were searched. Studies were reviewed for quality and analysed according to the framework outlined in the UNHCR Guidelines on Prevention and Response of Sexual Violence against Refugees. Twenty-nine studies met the inclusion criteria, of which 7 studies addressed prevention, 14 studies response and 8 addressed both. There are limited numbers of rigorously evaluated SV prevention and response interventions available, especially in the context of displacement. However, emerging evidence shows that placing a stronger emphasis on programmes in the category of engagement/participation and training/education has the potential to target underlying causes of SV. SV against female refugees is caused by factors including lack of information and gender inequality. This review suggests that SV interventions that engage community members in their design and delivery, address harmful gender norms through education and advocacy, and facilitate strong cooperation between stakeholders, could maximise the efficient use of limited resources.


Background
By the end of 2016, continuing conflicts, threats of persecution, violence and human rights violations have led to the displacement of an estimated 65.6 million people globally, of whom 22.5 million are refugees. 1 This is the highest refugee population ever recorded and much of the growing refugee population was driven by the Syrian conflict between 2012 and 2015. However, other conflicts in Iraq, Yemen and Sub-Saharan Africa have also contributed to the global refugee crisis. 1,2 As a result, the main country of origin of refugees in 2016 was Syria, followed by Afghanistan, South Sudan and Somalia. 1,2 Over half of all refugees are being hosted by just 10 countries, of which nine are low-and middle-income countries (LMICs) without sufficient resources to sustain and support these refugee numbers. 1,3 Europe also received an increased influx of refugees with two million arriving between 2015 and 2016, 55% being women and girls. [4][5][6][7][8] Sexual violence (SV) is experienced by many women living in conflict and post-conflict settings, during transit, and within destination countries. [9][10][11][12][13] A recent systematic review estimated that 21% of women in conflict countries have experienced SV either by a stranger or an intimate partner. 14 Another study carried out in Belgium and the Netherlands, which interviewed 223 refugees in 2012, showed that 69.3% of refugee women experienced SV since their arrival in the European Union. 15 However, data available on SV prevalence against female refugees are limited as incidents are often not reported or documented due to lack of SV training of humanitarian staff, SV-related stigma or lack of information about available services. 16,17 This article is focused on SV, which is defined as "any sexual act, attempt to obtain a sexual act, unwanted sexual comments or advances, or acts to traffic (…) against a person's sexuality using coercion (…)". 18 SV is a form of gender-based violence (GBV), which describes violence against women (VAW) perpetrated predominately by men against women and girls as a result of unequal power relations between genders. SV can be perpetrated by strangers, acquaintances, family members and intimate partners, the latter identified as sexual inter-personal violence (IPV). 19,20 Although men can be subjected to SV, women are more vulnerable to SV due to entrenched inequities and discrimination against women. 14,21,22 This is a consequence of sociocultural gender norms and expectations that situate women at a disadvantage compared to men and contribute to their increased risk of violence. Displacement and crisis situations can further aggravate gender inequality and with it the risk of SV due to disruption of social structures and loss of extended family support, economic and physical insecurity, lack of access to information and resources or services as well as an uncertain legal status. 12,15,23 SV against female refugees includes rapes, sexual harassments, forced marriage in order to gain male protection and survival sex in return for documents, food or transport. 12,13,24 Women are targeted by smugglers, human traffickers, male refugees, humanitarian staff or strangers. [11][12][13] Consequences of SV are severe and can lead to many short-and long-term health consequences, such as unintended pregnancies, transmission of HIV and other sexually transmitted infections, as well as mental health disorders including anxiety, post-traumatic stress disorder (PTSD) and depression. In addition, sexually abused women are often stigmatised, which further aggravates their mental health, and can result in nonreporting of abuse, suicide, social rejection or murder of victims by the community or family members. 25,26 SV survivors are also at increased risk of being subjected to subsequent abuse, further aggravating long-term health outcomes. 15 Despite attempts by refugee host countries, UN and humanitarian organisations to implement protective measures, basic security measures and protection guidelines are often not adequately implemented, as shown by recent situation assessments in refugee-receiving countries. 10,16,17,27 Many international NGOs and UN agencies have published prevention and response guidelines with the most relevant ones formulated by the United Nations High Commissioner for Refugees (UNHCR), the Inter-Agency Standing Committee (IASC) and the United Nations Population Fund (UNFPA) and include the following: In response, the UNHCR, which has the primary UN mandate of protecting and supporting refugees, has formulated comprehensive guidelines that are specific for the prevention of and response to SV against refugees. 28 The guidelines have an overarching human-rights based approach, which aims to address the root causes of SV and ensure that international human rights standards are protected and promoted. UNHCR recognises the significance of harmful gender norms and power dynamics in regard to SV and recommends human rights-and community-based approaches including: initiatives that empower women, through targeted livelihood interventions or creation of women's groups; ensuring confidentiality for SV survivors and legal action for perpetrators; safety and security measures including secure accommodation; and increased education regarding SV. [28][29][30] Despite these guidelines, there is a gap in understanding how to implement SV prevention and response programmes across diverse humanitarian settings and in particular within transit or in temporary refugee settlements. 33 This review addresses this "know-do" gap, synthesising published literature on the effectiveness of interventions to prevent and respond to SV against female refugees. The review focuses on promising approaches aligned with UNHCR guidelines that have evidence of impact on SV prevalence in humanitarian settings with limited resources, as the majority of refugees are hosted by LMICs. 1 While the focus is on refugees protected by international law, displaced people in refugee-like situations, who are also in need of protection but might not meet the definition of the 1951 UN Convention, are also included. 1,2

Methods
A literature search was conducted in July-August 2016. The review focused on peer-reviewed articles accessed from the following databases: Medline (Ovid), PubMed, Scopus, PsychINFO, CINAHL and the Cochrane Library. Grey literature was reviewed from websites of the World Health Organisation (WHO), UNHCR, Women's Refugee Commission, United Nation Population Fund and the World Bank. In addition, reference lists of relevant articles were manually searched. The review was limited to articles and reports in English and German, which have been published between 1 January 2000 and 31 August 2016. Search terms encompassing variants of SV, humanitarian crisis settings and programme evaluation were incorporated. A summary of search terms can be found in Table 1.
The title and abstract of all articles were screened for eligibility according to specific inclusion and exclusion criteria as described in Table 2. Articles and reports, which matched the inclusion criteria, were then obtained as full text and screened for further eligibility according to content. Summary data from all included articles were extracted into an Excel spreadsheet for analysis by both authors.

Quality assessment
Included studies were evaluated for the quality of their methodological rigour, and rated as either "weak", "moderate" or "strong" and assessed with a different appraisal tool according to the type of study: quantitative studies were evaluated with the quality assessment tool for quantitative studies of the Effective Public Health Practice Project (EPHPP). 34 For qualitative studies, the Critical Appraisal Skills Programme (CASP) quality appraisal tool 35 was chosen and for systematic reviews, the AMSTAR measurement tool was used. 36,37 The Mixed Methods Appraisal Tool was used to review included mixed methods studies. 38

Analysis of studies
Included studies were reviewed and evaluated against a framework based on the UNHCR Sexual Violence against Refugees: Guidelines on Prevention and Response, including subsequent updates, due to its specificity to the topic of review. [28][29][30] The framework is categorised into prevention and response strategies and sub-categories, which shared common protective elements for either prevention or response. Prevention was sub-categorised into four sections: (a) participation and engagement, (b) safety, (c) legal protection and (d) education and training. Response was divided into the following four sections: (a) treatment and counselling, (b) protection from repeat abuse, (c) legal protection and (d) education and training. Articles were analysed according to this framework to evaluate the focus and orientation of available evidence and to identify research gaps and intervention opportunities. Table 3 provides an overview of the guidelines and the applied framework.

Results
The literature search resulted in 1422 articles. After removing 54 duplicates and a screening of title and abstract of the remaining articles, 110 articles were included for further review. The full text of seven articles could not be located and these were excluded. After full-text screening, 74 articles were excluded due to incomplete information on interventions described, uncertainty on whether or how SV was targeted, or when implemented in a high-resource setting. At the end of this process, 29 articles were included for the synthesis (see Figure 1). Included articles consisted of nine qualitative studies, six quantitative studies, two mixed method studies, seven reviews and five reports. Seven publications addressed prevention, 14 focused on response and 8 publications addressed both. Finally, 12 studies were of "weak", 14 of "moderate" and 3 of "strong" quality. Table A1 summarises the key characteristics of the included articles and Table A2 summarises the studies against the UNHCR Framework.

Prevention
Identified articles discussed SV prevention with respect to male involvement in protection measures, community mobilisation as well as skills and livelihood enhancing strategies similar to the recommendations of the UNHCR guidelines. Several articles addressed more than one guideline category.
Participation and engagement Nine articles described interventions which engaged refugee communities or men in SV prevention efforts.
One trial, of weak quality, conducted in conflictaffected Côte d'Ivoire, engaged men in reflective discussion groups and showed that even a short but intensive intervention period of four months was successful in changing negative gender norms and expectations, increased conflict management skills and respect for women's rights. The intervention recruited 174 men and their female partners and conducted a survey about the incidence of IPV, including sexual IPV, preand one-year post-intervention. According to the surveys one year after the intervention, women reported experiencing less sexual IPV and more men agreed with the statement that women have  . Ensure legal protection and prosecution of SV perpetrators with the help of women's organisations, for example, legal protection . Provide personal documentation to all refugee women . Ensure protection and enjoyment of one's rights as refugee (rights-based approach) Legal protection . Report abuse to police . Interview victim to obtain information for legal prosecution . Ensure legal action (if victim agrees to it) and punishment of perpetrator(s) . Provide essential information about rights and available options to victim Education and training . Run information campaigns and training courses about SV and related issues for staff and refugees . Activities that promote gender equality as this is a cause of SV (help women to obtain capacity and skills to control over their own livelihood)

Education and training
. Training courses about SV and related issues for staff and refugees In addition the following general principles apply to all programmes: . Protection measures need to be gender-sensitive and responsive . Measures need to include immediate, intermediate and long-term solutions for effective sustainable protection . Include LGBT women as a vulnerable target group . Include women with disabilities as a vulnerable target group . Empowerment and engagement are key components to majority of protective measures . Prevention and response measures need to be multi-sectoral (health care, psychosocial and legal support, etc.) . Improve data collection methods and interagency information sharing to design evidence-informed and effective programmes a right to refuse sex regardless of the circumstances. The strength of this study was the inclusion of female partner reports in the analysis, which reduced the possible bias inherent with relying on reports given by participating men. 39 According to Gurman, 40 SV reduction strategies also have to consider the influence of social networks and the local community on the actions of the individual. Community-based interventions appear most suitable as people can be engaged as agents for social change. 40 One promising strategy is the use of short films and video projects in order to educate the local community and to initiate a public discussion about GBV and SV. [41][42][43] One multi-year participatory video project rated as being of weak quality, which took place in Thailand along the Thai-Myanmar border and in Rwanda, Liberia, South Sudan and Uganda in Africa, actively engaged the community and included different stakeholders such as ministries, religious leaders and youth and women's groups in the creation of several targeted videos, which educated about GBV, including SV. 40 Each public screening was then followed by reflective discussions. A qualitative evaluation of 18 focus group discussions showed that participants credited the project with an increased awareness and respect for women's rights, more equitable gender dynamics within relationships and families, a decrease in forced early marriage and increased reporting of GBV including SV and use of related services. Nevertheless, it was also recognised that continuous activism is still necessary to change gender norms long-term as some abused women still deal with stigma and hence, may choose to remain silent. 40 In a qualitative study in Rwanda, refugees also expressed the need for collaboration between camp leaders, local communities and NGOs to avoid a top-down intervention approach and thereby increase accountability, ownership and acceptability of protective services. 43 One intervention of weak quality in a refugee camp in Darfur showed that actively consulting and engaging female refugees and camp leaders in developing solutions to safety concerns was more effective and sustainable than endeavouring to enforce existing guidelines and policies. 44 Another article by Spangaro et al 41 found that SV interventions designed and implemented with active involvement of the community were associated with greater reductions in SV exposure and incidence, concluding that engagement of targeted communities may be fundamental to the success and sustainability of SV prevention efforts.
The empowerment of women through vocational skills training and education has also been found to be promising as an attempt to increase women's livelihoods and decrease risk of SV. 43,[45][46][47] In Sierra Leone, this approach was reported to increase women's choices and capability to earn a sustainable income and thereby decreased women's vulnerability to sexual exploitation. 46 However, projects supporting women's economic opportunities can also increase their risk to IPV, which may include sexual IPV. 47 Ray et al 47 strongly recommend engaging with men in any livelihood programmes. Implementing these interventions in isolation is discouraged given the complexity of GBV in displacement situations. Despite the recognition that livelihood interventions might be crucial for protecting refugee women, there was limited evidence among the included studies with regard to the effectiveness of economic opportunities in displacement settings for reducing SV. In many hosting countries, governments apply significant restrictions on the right of refugees to work. 47 None of the identified studies evaluated efforts to conserve the original community structures from the country of origin within refugee camps or on how to engage refugee women in the design, management or leadership of SV protection measures as recommended by UNHCR. Safety Eleven studies evaluated the impact of SV prevention on increasing personal and environmental safety.
All studies within the safety domain addressed the risk of SV that women faced while collecting firewood. [41][42][43][44][45][48][49][50][51][52][53] Intervention approaches included firewood provision or training in fuel-efficient stove production. [41][42][43] One comprehensive evaluation of a firewood project in Kenya, rated as being of moderate quality, provided firewood in an attempt to reduce rapes. While the incidence of rape was 10% lower during wood collection, it increased in other situations within the same camp. 53 In addition, the monetary value of wood increased during the project and some women collected additional firewood for income generation, augmenting their risk, despite their own fuel needs being supplied. 52,53 The project was assessed as neither cost-effective nor sustainable, and was found to increase tensions between refugee and host communities as the firewood provision was resented by host communities who did not receive the same assistance. 53 Tappas and colleagues reviewed the provision of fuel-efficient stoves to reduce the trips necessary to collect firewood and reduce SV exposure of women and found the need to purchase fuel can in turn increase the economic burden and SV vulnerability. 43 Despite its questionable impact, stove or firewood provision remains a major approach by humanitarian organisations to increase safety and decrease exposure to SV. 51 However, one project, in a Sudanese displaced persons camp, showed a promising approach, which included the formation of a women-led firewood patrol committee. Regular meetings of these women allowed them to discuss any safety concerns and it increased trust, problem-solving skills and ultimately women's safety. 42,44,52 Legal protection Five articles, among them three systematic reviews, discussed legal interventions, such as protective laws, policies, guidelines and training programmes as a preventative measure against SV.
The limited literature on protective legal interventions in the prevention of SV against refugees indicates a severe lack of evidence available. Moreover, included literature discussed interventions with limited effectiveness as evaluations showed that refugees are insufficiently educated about their rights or available legal mechanisms and may regard SV as something shameful but inevitable, thereby somewhat "normalising" SV. Therefore, many SV cases may not be reported despite protective legal mechanisms in place. [41][42][43]54 Legal interventions are, however, considered important components for SV prevention to deter perpetrators through comprehensive laws and policies. 42 Nevertheless, a systematic review of moderate quality by Spangaro et al 42 found that even if legal policies and mechanisms are in place and SV charges are reported, many cases have not been prosecuted, leaving many SV perpetrators unpunished and SV survivors without justice. Consequently, protective policies and laws are limited in their impact of deterring future or repeat SV perpetrators. [41][42][43] Education and training Seven studies evaluated education and training interventions for SV prevention.
One weak quality study implemented by UNI-FEM in 11 refugee camps in Darfur between 2008 and 2011 intended to increase the internal security for female refugees by training camp leaders, police and humanitarian staff in gender sensitivity and GBV/SV protection protocols. 44 The training resulted in increased awareness and understanding of VAW and improved capabilities and problem-solving skills of the community to protect refugee women against SV more efficiently. 44 Thompson et al 44 stressed the importance of the involvement of camp leaders in interventions and a participatory training, with regular continuous training, as a long-term strategy to change social norms and harmful attitudes that result in SV. Moreover, in this study, the involvement of religious leaders was shown to play an important role to reinforce important messages against SV. 44 Another training programme implemented in Guinea further showed success in SV prevention through the deployment of "community trainers" in refugee settlements, who facilitated and educated refugees about SV and involved the community in a participatory problem-solving process to establish SV preventative systems. The programme also succeeded in recruiting refugees, including women, to be trained as paraprofessional social workers to support awareness-raising activities such as through the organisation of women's groups, aiming to inform and sensitise women about their rights in cases of abuse, and has led to an increased uptake of SV-related services. 41,42 Four studies also focus on encouraging further education and skills training as promising approaches for the prevention of sexual exploitation of refugees. [45][46][47] However, in some circumstances, economic opportunities for displaced women also increased the exposure to sexual harassment or attacks by employers, especially if working illegally, as women may fear deportation or detention if they seek help from the authorities or police. 47 Limited access to employment opportunities for displaced girls and women without family or spousal support is of particular concern as food rations provided by humanitarian agencies are often insufficient, which may force girls and women to seek precarious means to earn an income through prostitution or illegal domestic work, where they risk sexual harassment. Ray et al 47 strongly recommend that governments need to grant refugee women legal status as quickly as possible and allow them to work legally in the country of refuge to prevent the cycle of vulnerability that favours sexual exploitation.
Two systematic reviews on SV in humanitarian settings worldwide, both of moderate quality, determined that training programmes for volunteers, staff, community leaders and refugees have the greatest impact on reducing the risk of SV, when combined with multi-component interventions, which may include community discussions and awareness-raising activities. 41,42 According to Spangaro et al, 41,42 this is due to the fact that this intervention approach better addresses the broad spectrum of issues, which initially led to the exposure and risk of SV.

Response
The identified literature predominately focused on treatment and counselling as a response to SV. Several articles addressed more than one guideline category.

Treatment and counselling
Eighteen studies addressed the treatment of SV and counselling of SV survivors with only one study of strong quality.
The majority of the identified literature predominately focused on alleviating PTSD through individual or group therapy and support groups with skills training; these have had some success in alleviating mental distress and increasing perceived social support. 41,42,[44][45][46][55][56][57][58][59][60][61] However, studies often lacked rigorous methodology and reporting, especially if conducted in conflict areas. 56,57 In a weak quality trial conducted in the Democratic Republic of Congo (DRC), female SV survivors with high levels of PTSD, anxiety and depression were randomly assigned to either individual support or cognitive processing therapy (CPT). Evaluations showed that women completing the CPT intervention recorded greater reduction in mental health distress than women receiving individual support only. Additionally, functional impairment was also significantly decreased in the CPT intervention group. 55 Another intervention, implemented in a conflict setting of Congo, showed that psychological support provided by trained psychologists can successfully reduce SV trauma-related impairment by around 60% and its benefit was maintained up to two years after the intervention. 58 Five of these 16 studies also suggest that the inclusion of traditional rituals helped SV survivors to deal with trauma and build resilience. 41,42,59,60,62 According to a moderate quality study by Stark,60 traditional healing ceremonies held in communities for former girl soldiers who survived SV during the conflict in Sierra Leone supported their personal healing and reintegration process into society. Another study from Sierra Leone, which interviewed local social workers and women who had survived war-related SV, showed that a personal care approach, which includes directive advice, self-disclosure and personal involvement of social workers, who were often SV survivors themselves, was greatly appreciated by help-seeking women and seen as a source of knowledge and courage. 59 These findings are reinforced by the analytical analysis of a failed GBV/SV intervention in a displaced people camp in Eritrea, which attempted to provide health services and individual counselling to SV survivors of the Saho ethnic group after the second Ethiopia-Eritrea war. Most women did not attend those services and later evaluations concluded that the implemented intervention might have imposed models of Western psychotherapy, which were insufficient to attend the needs of SV survivors in Eritrea, as Saho women are more accustomed to group support with women in their community facing similar issues. Therefore, traditional or community support systems were not fully considered and might have affected the impact of the intervention. 63 Leskes et al 61 further evaluated two psychosocial programmes in Liberia, which were implemented for three months each by two different local NGOs. The evaluation compared their respective interventions, which included one intervention with individual counselling and another one with support group and skills training for income generation. The aim of the evaluation was to measure the impact on stress and trauma reduction experienced by SV survivors. Both interventions had positive effects, although for many participants their stress was not due to SV experience alone, but was compounded by ongoing socio-economic deprivation. 61 Only four articles evaluated healthcare services in regard to SV response. 48,49,63,64 One weak quality study, situated in Colombia, stressed the importance of a culturally sensitive integrated response to SV that allowed vulnerable women to access comprehensive sexual and reproductive health services. The project closely collaborated with different local governmental and civil society organisations, such as health councils, health committees, pastors, women's groups and youth networks, to create strong links with other relevant services such as legal aid, and to challenge local attitudes and social mechanisms that condone SV through awareness-raising campaigns. This approach was seen as essential to the achievement of the project to increase the visibility of available services and to increase access to appropriate care. 46 In addition, Henttonen et al 64 further found that appropriate health services require regular training in SV of all health staff as well as a stronger orientation of services towards adolescents to maintain proper care. Particularly adolescent women were found to be vulnerable to SV, which required health services to offer youthfriendly services with special hours to better meet their need of care in case of SV. 64 A strong quality study by Wirtz et al 49 and a weaker one by Vu et al 48 described the development and testing of the ASIST-GBV screening tool, which was specifically designed for conflict areas of Colombia and Ethiopia to increase timely identification and referral of SV survivors for treatment and appropriate services. The tool has been shown to be accurate, practical and easy to use, adaptable to context, specific to women's and girl's needs and suitable for resource-constrained humanitarian settings. Collected patient data can be linked with reporting and monitoring systems, such as GBV-IMS, to avoid duplication of information and re-traumatisation of abused women. The ASIST-GBV screening tool is to date the only tool being tested in a humanitarian context, which does not only target IPV. However, the acceptability of the tool by end-users, such as refugees and service providers, has not been extensively studied and is currently being evaluated. 48,49 Despite benefits of early SV detection, screening is not advised in the absence of a functioning and efficient referral system in place. 49 Protection from repeat abuse Eight included studies addressed this domain of the UNHCR framework. One study of strong methodological rigour by Lattu et al 54 explored various mechanisms in place in different refugee settings in case of sexual exploitation and abuse, in particular if committed by humanitarian staff. This qualitative research 54 showed that camp residents were often hesitant and fearful to complain and were insufficiently informed about how to protect themselves from abuse or how to safely report sexual misconduct. Therefore, refugees suggested firing humanitarian staff found guilty of perpetrating SV, and that displaced communities needed to be better informed about their rights, including how to access freely available humanitarian services. Awareness-raising activities, such as workshops and youth drama clubs, further proved to be effective in initiating discussions within displaced communities about SV and safe reporting. Moreover, collaboration with gatekeepers, such as camp leaders, was found to be essential for successful implementation of related interventions. 54 Another study in Kakuma refugee camp in Kenya further explored community mechanisms in response to IPV, including sexual IPV, and discovered a hierarchy of responses, which showed that community-based responses, which often involved negotiations with community leaders and relatives, were preferred to services provided by humanitarian staff. This was also due to the fact that many refugees were mistrusting of services implemented by international agencies as they believed them to act with a top-down approach and to act only in favour of women and against the interest of the involved families. However, the study also showed that community measures do not necessarily result in the protection of abused women and can even support re-victimisation, as many women were blamed for their own abuse at the hands of their partners. 43,65 In addition, included literature also pointed towards the importance of changing harmful gender norms, raising public awareness about SV and providing economic opportunities to female refugees to decrease their vulnerability to repeated sexual exploitation as well as to address the underlying social mechanisms that lead to repeat abuse of female refugees. [40][41][42]45,50 Legal protection Five studies discussed legal interventions such as legal aid services and prosecution of SV offences but there was limited evidence regarding their effectiveness.
Two studies focused on public awareness-raising campaigns, which inform about women's rights and legal services available in case of SV, and these were found to be crucial to signal zero tolerance towards SV and to encourage SV survivors to seek justice. 43,46 However, two included reviews of moderate quality stated that without adequate protection for abused women against stigma, negative attitudes of the public or avoidance of re-victimisation and traumatisation, legal interventions will be ineffective in reaching out to women and enabling them to receive justice as women will fear to speak up. 41,42 Nevertheless, legal services have been found to be essential to counteract the prevailing impunity for SV against refugees and therefore this needs to be addressed in response measures. 41,42,50 Education and training Nine identified studies evaluated effective SV education and training interventions.
A majority of studies either focused on training refugees about SV through awareness-raising campaigns, workshops or through teaching vocational skills in order to decrease SV vulnerability and increase well-being. [40][41][42][44][45][46]50 One study of moderate quality described a multi-component programme in the DRC, which targeted women who survived SV during the conflict. The intervention included advocacy and support for local women's groups in the organisation of educational opportunities for women and girls and was found to have a positive impact on participants' well-being and functioning in daily and social tasks. 50 A similar intervention of weak quality, implemented in Sierra Leone, targeted women affected by SV during the civil war who had been forced into commercial sex. The intervention aimed to provide knowledge and skills, which would enable these women to find alternative income activities and protect themselves from HIV. Applied strategies to educate these women included music, dance and drama and peer education and were found to be effective in delivering sensitive and important messages. Impact evaluations also showed that the programme was successful in encouraging women to learn new skills and to find alternative livelihoods. HIV infections and mortality decreased among female programme participants. 46 Another intervention in Colombia, which aimed to protect and educate women affected by violence including SV, used radio and flyers and trained mobile teams in order to teach women about GBV and SV, to increase the visibility of available services and to confront the cultural acceptance of GBV. Moreover, the cooperation with youth, women's and social organisations increased the reach of the project. 46 Another determining factor for training programmes was their need to be empowering and participatory in order to enable refugees to gain practical skills to solve SV-related safety concerns and to change harmful attitudes that condone SV. A successful intervention, which applied this approach, was conducted by Thompson et al,44 training female refugees in women's rights and leadership skills, who then passed on their skills to other women through meetings at women's centres and committees. This allowed women to voice their concerns and issues to decision-makers and enabled them to have an active role in the planning of safety measures within their own camps. 44 Three studies also evaluated training on SV response for service providers and humanitarian staff. 44,64,66 One moderate quality study, which trialled a multi-media training for health providers in humanitarian settings of four different countries, further succeeded in increasing respect for patients' rights, case management and treatment. 66 However, negative attitudes towards SV survivors were not affected by the training, posing a substantial barrier to the reporting of SV abuse to service providers. 66 Another evaluation of health services provided to SV survivors in post-conflict Uganda further showed health personnel should be regularly trained on SV case management as part of their continuous medical education, as one-off trainings in SV response protocols through international agencies were deemed insufficient to ensure appropriate care. Moreover, a greater focus on dealing with adolescent SV survivors was recommended as predominately young girls were found to be sexually abused. 64 Within this category only one study, of weak quality, 44 pointed out the importance and impact on SV reduction by training both refugees and service providers, in order to increase the understanding and communication between refugees and the professionals working with them, as well as to build an enabling and empowering environment for refugees to find practical and tangible solutions for SV protection.
However, while this review identified studies with the objective of transforming harmful gender norms within displaced communities, none of them addressed attitudes and prejudices held by service providers and other professionals, who directly engage with SV survivors.

Discussion
The review demonstrates that SV against refugee women is a complex public health concern requiring a comprehensive, multi-component and culturally sensitive solution. The interventions included were generally not well designed, lacked adequate monitoring and evaluation and were often shortterm, preventing a robust impact assessment of interventions on SV long-term or a reliable estimate of the impact interventions have on SV. [41][42][43]45 Few programmers followed the UNHCR guidelines, with the exception of the limited intervention of firewood collection within the safety (prevention) domain and numerous programmes providing counselling within the treatment and counselling (response) domain. Notably, no study evaluated the methods of data collection or interagency information sharing to improve SV prevention and response measures. [28][29][30]43 Humanitarian actors all endorse the need to mainstream SV prevention and response in all programmes to achieve a true impact on the protection of refugee women and girls. 28 However, none of the identified studies evaluated the approaches to mainstream SV prevention within broader humanitarian efforts.
Nevertheless, the review identified a number of promising approaches for SV prevention and response, deserving greater attention from humanitarian actors and researchers. These include the active involvement of female refugees in the design, planning and implementation of preventative measures. The involvement of refugees increased empowerment and ownership of implemented programmes and built capacity reducing SV across all included studies. [41][42][43][44]46 Community engagement strategies also helped to raise awareness about SV, transform harmful gender norms and strengthen accountability within communities to reduce and prevent SV. 39,40,41,[43][44][45][46] Training and education interventions, which engage and teach refugees about SV and the value of gender-equitable norms, are also effective in the prevention and response to SV, according to our review. [40][41][42]44,46,50 However, few studies targeted service providers directly in contact with survivors. 64,66 In addition, very few included studies addressed effective methods of collaboration between organisations, services or even between humanitarian staff and refugees in order to better address safety concerns or to better integrate and improve services for SV survivors. 44,46 Although the UNHCR guidelines are important for SV protection, the recommendations have a tendency to position refugees in isolation and disconnected from context, thereby disregarding the often complex dynamics and factors contributing to SV exposure of female refugees. The guidelines also focus on SV perpetrated by strangers rather than intimate partners, even though sexual IPV cases are not uncommon during times of crisis and displacement. 14 Risks posed by the host community, including smuggling and trafficking networks, are barely acknowledged and discussed. 15 Moreover, abused women face many barriers due to harmful gender biases, stigma, xenophobia and social marginalisation despite protection measures in place. [66][67][68] The existing UNHCR guidelines are more applicable for refugee camp settings. [28][29][30] Refugee populations in transit face constantly changing circumstances and situations on their way to their destination, a logistically challenging research setting, reflected in the paucity of studies in moving populations.
A number of interventions commonly implemented by NGOs or UN agencies in humanitarian settings were excluded from our review due to the limited evidence of effectiveness. These include women's safe spaces or youth tents, where displaced women and girls can socialise, retreat, build skills and receive information about GBV. 23 Other approaches trialled include the establishment of mobile health teams that reach remote displaced populations which would otherwise be marginalised. 23 Finally, while UNHCR acknowledges the importance of multi-sectoral and coordinated SV/GBV responses, none of the identified literature in this review examined the potential benefit of effective coordination between humanitarian agencies in a given setting. The UNHCR's Refugee Coordination Model provides guidance on harmonisation of humanitarian action in regard to GBV protection, which includes the establishment of a Refugee Protection Working Group for the coordination of humanitarian services and the mainstreaming of GBV protection in all sectors. 69 The absence of any evaluations of coordinating mechanisms is a significant research gap.

Limitations
This review has some limitations. The focus on English and German publications may have excluded relevant papers only available in other languages. The screening of articles was conducted by a single reviewer which may have resulted in reviewer bias. In addition, the complex context that is the hallmark of conflict and humanitarian crises means that rigorous programme evaluation is particularly challenging. It is noteworthy that only three of the included papers were of strong methodological rigour.

Conclusion
There is limited evidence available for the effectiveness of SV interventions which protect refugees and prevent SV. This review suggests that empowerment, engagement, gender-transformative, culturally sensitive and comprehensive approaches, which build the capacities of the refugee community, provide the most promising strategies to prevent SV. Refugee women cannot rely on short-term SV responses but require a long-term commitment across government and humanitarian sectors to ensure protection for all women. We recommend a greater focus on evaluating and implementing effective models for legal protection, male-engaging strategies, livelihood interventions and training for service providers. Our review provides examples of promising interventions within the domains of engagement and participation as well as training and education, and highlights the significant evidence gaps, which need to be addressed to ensure female refugees enjoy their full rights and can live valuable and healthy lives.