Indashyikirwa Women’s Safe Spaces: Informal Response for Survivors of IPV within a Rwandan Prevention Programme

: Within intimate partner violence (IPV) prevention programmes that raise awareness of women’s rights and the forms and consequences of IPV, there is a need to ensure response mechanisms for IPV survivors. Indashyikirwa is a Rwandan IPV prevention programme, which established 14 women’s safe spaces, whereby men and women could access support for IPV, be referred or accompanied to other services. This paper draws on qualitative interviews with safe space facilitators, attendees, staff and observations of activities at various points across the programme. Thematic analysis was conducted to assess the process and impact of the spaces. Attendees generally preferred the women’s safe spaces over formal services for IPV disclosure and support, and the spaces also enhanced the quality of and linkage to formal IPV response services. The safe spaces further supported well-being and economic empowerment of attendees. Lessons learned from implementing this model are offered, including how to ensure safe, inclusive and integrated sources of support within broader IPV prevention efforts.


Background
With the encouraging growth of global programming to prevent intimate partner violence (IPV), it is imperative to understand effective integration of response mechanisms for survivors of IPV. This is especially important among programmes that raise awareness of women's rights and the forms and consequences of IPV, as this can create a greater demand on available IPV response services, including social, health and criminal justice services. It is also critical to understand the context of IPV prevention programme implementation, as it is often more difficult to ensure the safety of IPV survivors, or can generate a burden in areas with limited available response services. There have also been recent programmatic efforts to support informal responses to IPV survivors, given that these sources are more commonly used. For instance, globally, across 24 nationally representative Demographic and Health Surveys (DHS), 40% percent of women experiencing gender based violence (GBV) previously disclosed to someone, yet only 7% reported to a formal source or person in authority (Palermo et al. 2014). Women's safe spaces or support groups can be important and effective sources of informal support (Rodriguez 1999;Sullivan 2012). Yet there is limited understanding of best practices for incorporating IPV prevention programmes with informal and formal response components. This paper will contribute to this gap by reviewing qualitative process and evaluation data of dedicated IPV response services implemented as a critical component of Indashyikirwa, an IPV prevention programme in Rwanda. In particular, we assess the successes and challenges of piloting and implementing the women's safe spaces model, which aimed to be gender transformative and

Rwandan Context
The Rwandan government has made significant strides to improve gender equality over the last few decades through the implementation of policies, laws and programmes promoting women's rights. In 1999, the Law on Matrimonial Regimes, Liberalities, and Successions established women's right to inherit land for the first time in Rwanda, including divorced women, to ensure they do not become landless (Debusscher and Ansoms 2013). In 2008, the Rwandan government adopted the Prevention and Punishment of GBVLaw, which includes all forms of violence with the minimum penalty being 6 months in prison, while sexual abuse or rape leading to terminal illness or death can generate a life imprisonment sentence (Umubyeyi et al. 2014). The Rwandan government has supported the development of improved responses to GBV, through initiatives such as the establishment of GBV desks at the Ministry of Defense and National Police, and GBV prevention committees at the village level (Slegh and Augustin 2010). The committees consist of stakeholders from various institutions dealing with GBV (i.e., justice officials, national women's council, civil society actors, religious leaders), are responsible for advocating for services, referring and assisting GBV survivors, and offer valuable social support (Mannell et al. 2018). In 2009, the Rwanda National Police and Ministry of Health launched the Isange One Stop Centers, which provide free medical care, psychosocial support, and legal services to victims of IPV and child abuse, as well as emergency accommodation for a few days (Umubyeyi et al. 2016).
Despite these efforts, there continues to be a lack of access to comprehensive medical and psychological care for Rwandan survivors of GBV including IPV (Russell et al. 2016;Zraly et al. 2011). According to the WHO (2005), Rwanda only spends 1% of its health budget on mental health services (Zraly et al. 2011). Although mental health has been highlighted as a national priority for the Rwandan government, it has been significantly underfunded by international donors (Mannell et al. 2018). Mannell et al. (2018) found that common responses of the Rwandan GBV committees include encouraging women to marry or remain with abusive partners, and prioritizing physical over emotional forms of violence, indicating limited understanding of the mental health consequences of IPV. Access to psychological care is critical for GBV survivors, as those who receive positive and supportive responses are less likely to experience post-traumatic stress (Edwards et al. 2015;Mannell et al. 2018), or remain at heightened risk for ongoing emotional, physical and economic suffering (Mahr and Campbell 2015). There is a particular gap in understanding Rwandan men's access to mental health support and trauma care, which has implications for both GBV prevention and response. For instance, Verduin et al. (2013) study found that common mental health disorders and post-traumatic stress disorder (PTSD) among Rwandan men could lead to an increase in perpetration of IPV.
Moreover, IPV in Rwanda is a persistent phenomenon. According to the 2014-2015 Rwandan DHS (NISR 2016), an estimated 20.7% of women aged 15 to 49 reported experiencing physical or sexual violence by their spouse in the past 12 months, and 7% of men reported experiencing physical or sexual violence by their spouse in the past 12 months. There is also very high acceptability of IPV, and the recent DHS found that 41% of women aged 15-49 believe a husband is justified in hitting or beating his wife for at least one of six reasons common within Rwandan culture, including burning the food, refusing to have sex, or arguing with him (NISR 2016).
Women's tolerance of IPV can be a significant barrier to women disclosing IPV (Mannell et al. 2016). Indeed, the Rwanda 2014-2015 DHS (NISR 2016) found that 23% of women and 20% of men aged 15-49 who had experienced physical or sexual violence by anyone never sought help or told anyone. Other identified barriers for Rwandan women to report IPV include fear of retaliation from their partner, humiliation, shame, financial dependency on their partner, and fear of worsening their overall life situation (Kubai and Ahlberg 2013;Mannell et al. 2016;Umubyeyi et al. 2016). A recent study conducted by the University of Rwanda with Oxfam (2017) found that 78.8% of women reported "fear of stigma" as the primary barrier to reporting violence, followed by because an "arrangement between families was made instead" (73.5%), or "feeling that it will change nothing" (65.2%). Given that Rwandan women's poverty and financial dependence on their spouses have been identified as significant barriers to women seeking care, the need to provide integrated health and psychosocial services that support women economically has been strongly emphasized (Russell et al. 2016;Umubyeyi et al. 2016;Mannell and Dadswell 2017). Supportive community responses to GBV survivors are especially important in Rwanda, given the evidence that men and women are overwhelmingly more likely to ask for help from their neighbors or family members when experiencing physical or sexual violence (NISR 2016).

Indashyikirwa Women's Safe Spaces
The Rwandan context of limited availability of formal (medical, legal, social, etc.) services for IPV survivors, especially psychological support, and the significant barriers for survivors to disclose IPV to formal services, informed the design of the Indashyikirwa women's safe spaces. Fourteen women's safe spaces (one per intervention sector) were established to ensure dedicated support for IPV survivors as part of the programme model (Stern et al. 2018). These draw on RWN's experience implementing the Polyclinic of Hope spaces since 1997, which were designed to address the health, psychosocial, shelter and socio-economic needs of GBV survivors. The spaces were identified as 'women's safe spaces' given the lack of women-led spaces in the context, and since Rwandan women are more likely to experience IPV than men. The programme conceptualized 'safety' of the spaces in terms of striving to be confidential, inclusive, and promoting a participatory approach where attendees should never be judged for their views or held accountable for their experiences of violence. Safety also entailed supporting attendees to feel comfortable sharing personal and/or difficult experiences. The spaces thus tried to carefully balance women's active engagement, while being inclusive of and catered to men, as both potential perpetrators and survivors of IPV. At each women's safe space, 22 female facilitators were recruited from the intervention communities based on their community leadership and/or related experience, and through a community election process. The women's safe space facilitators (WSFs) completed a training over two weeks at the beginning of the programme, facilitated by RWN staff, to offer dedicated support to women and men that report IPV, educate women about their rights, and refer or accompany individuals who wish to report IPV to local authorities or criminal justice services, and/or seek health or social services. Drawing on the SASA! programme established by Raising Voices in Uganda, the curriculum with WSFs emphasized positive types and uses of power ('power within', 'power to', 'power with') alongside negative types and uses of power ('power over'). The curriculum moved from knowledge, attitudes, and skills to actions to prevent and more appropriately respond to IPV. There were also dedicated modules on the roles of WSFs, participatory facilitation, communication skills for providing support, and the foundations of advocacy and reporting skills. The WSFs received ongoing support and refresher trainings from RWN staff throughout the programme, including to facilitate participatory dialogues with the support of SASA! adapted communication materials. Before implementation, the curriculum was pre-tested in one location from August-September 2015. One to two external researchers, including the first author, observed several sessions. After each session, focus groups were conducted with participants, and interviews were conducted with each RWN facilitator, to assess their impressions and understanding of the sessions. Daily notes were compiled documenting insights around participant engagement, comprehension and quality of activities, which informed key recommendations for refining the approach and content of the WSFs training.
Each safe space has a sector level referral list of healthcare, justice or social services and support mechanisms for survivors of violence, which are regularly updated. Three mornings per week, the women's safe spaces are open to provide dedicated, private spaces for men or women to disclose IPV, whereby WSFs can offer support, referral and/or accompaniment to other needed services. In the afternoons, WSFs conduct activism dialogues primarily around uses of power, gender equality, IPV, women's rights, and healthy relationship skills. SASA! communication materials including power posters and community conversations were adapted for use by WSFs. Income generating activities (IGAs), most commonly handicraft skill development and production, also take place in the afternoons. The WSFs also undertake advocacy to improve formal services for survivors of IPV. To inform this advocacy, attendees of the safe spaces were encouraged to provide feedback to the WSFs on the quality of services they were referred or accompanied to. The programme ensured linkages between the various components of Indashyikirwa, including through hosting quarterly meetings among WSFs, community activists and opinion leaders to facilitate joint planning and learning. The safe spaces were intended to be a critical referral point for the engaged community activists and opinion leaders.

Method
Qualitative research with the women's safe spaces was conducted as part of the external evaluation of Indashyikirwa, embedded within the DFID-UK funded What Works to Prevent Violence against Women and Girls Programme. The broader evaluation also includes longitudinal qualitative research with all of the Indashyikirwa stakeholders (couples, community activists, opinion leaders, staff), as well as a randomized control trial with couples and communities exposed and not exposed to the programme. The qualitative research was conducted in three Indashyikirwa intervention sectors (Rurembo Sector, Western Province; Gishari Sector, Eastern Province; and Gacaca Sector, Northern Province), which were purposefully selected to represent a diversity of environments including rural, urban and peri-urban locations. For the recruitment process, the first author informed RWN staff members of recruitment criteria, aims, benefits, and risks of the study, so that these could be disseminated to potential participants. Attendees were purposefully selected for having attended the safe spaces at least a few times, and for attending both the morning disclosures and afternoon group discussions, to assess their experiences with different aspects of the safe spaces. After obtaining the WSFs and attendees consent to do so, RWN provided the qualitative researchers with their contact details. The qualitative researchers than contacted the WSFs and attendees, and set up a time and location for conducting these interviews.
Data was collected at various times throughout the programme to unpack the implementation evolution of the women's safe spaces. In May 2016 (Round One), three WSFs (one per sector) were interviewed to assess their motivations as facilitators and their impressions of the training they received. They were also asked about typical motivations for safe space attendees, attendees' awareness of rights including the GBV law, and their perceived impact of the spaces. In September 2016, six safe space attendees (two per sector) were interviewed to assess their reasons for visiting the safe spaces, and the difference the spaces make in their lives (if any). In June 2017 (Round Two), a different subset of three WSFs (one per sector) were interviewed to assess their perceived impact of the safe spaces and the ongoing support they receive as facilitators. Six female attendees (two per sector), one male attendee in the Northern Province and one male attendee in the Western Province were also interviewed to assess their impressions of and experiences with the women's safe spaces. The RWN programmatic team wanted to better understand men's impressions of and engagement with the women's safe spaces, which is why men were included in the sample. In May 2018 (Round Three), three WSFs (one per sector) were interviewed to assess their perceived impact of the safe spaces, and what they had been doing at the safe spaces recently. Four female attendees (two in Ruhimbi sector, one in Rurembo sector and one in Gacaca sector) were also interviewed to assess their impressions of and perceived impact of the safe spaces.
In May 2016, six in-depth interviews were conducted with RWN field officers and supervisors across all intervention sectors, which assessed their perspectives of successes and lessons learned from facilitating the WSFs training. Seven RWN staff were also interviewed in May 2017, and asked to describe key successes and challenges of the women's safe spaces, assess motivations and abilities of the WSFs, and their general perceived impact of the women's safe spaces. Interviews with WSFs, attendees, and RWN staff lasted approximately one hour. All interviews were conducted at preferred locations deemed appropriate and private for participants. Two female Rwandan qualitative researchers external to the programme conducted the interviews with WSFs and attendees. All these interviews were conducted in Kinyarwanda and audio recorded. The first author conducted the interviews with RWN staff in English, and took detailed notes of these interviews.
Self-reports of programming may be biased, due to social desirability (James Bell Associates 2009). Observations can often provide a more objective assessment of programme implementation, including whether facilitators use appropriate delivery methods, or actively engage participants (James Bell Associates 2009). There are further benefits of observations, drawing on anthropological and ethnographic perspectives, whereby observations alongside interviews equips researchers to better understand the depth of participants' experiences (Najafi et al. 2016). The Rwandan female qualitative researchers observed two women's safe space activities per research sector (six in total) in September 2016, two women's safe space activities per sector (six in total) in June 2017, and one women's safe space activity per sector (three in total) in May 2018. For the observations, the researchers took notes on participation engagement, comprehension, and facilitator skills, using a guidance document.

Ethics
Ethical approval to undertake the study was obtained from the Rwandan National Ethics Committee (RNEC) (REF: 340/RNEC/2015) and the National Institute of Statistics Rwanda (REF:0738/2015/10/NISR). Secondary approval was also obtained from the South Africa Medical Research Council (REF: EC033-10/2015) and from the London School of Hygiene and Tropical Medicine, and ethical procedures were followed in accordance with the review board guidelines. All identifying information of participants have been removed for the presentation of the findings.

Analysis
The first author debriefed with the qualitative researchers after data collection to capture their initial impressions, non-verbal and contextual insights. These research summaries were used to inform the analysis. Using the audio files, the data from WSFs and attendees were transcribed and translated verbatim into English by a language specialist and professional translator. After carefully reading the transcripts from WSFs, attendees, staff interview and observation notes, the first author established a preliminary coding structure to analyse the data. Thematic analysis was conducted to uncover predominant themes in order to provide a rich, detailed and holistic account of the data (Braun and Clarke 2006). All of the transcripts were then analysed by the first author using this thematic coding framework with the assistance of NVIVO 11 software. An additional transcriber coded a small subset of the baseline transcripts using NVIVO 10 and inter-coder agreement was found to be 95%. The first author regularly presented the emerging findings to the Indashyikirwa senior programme staff at participatory workshops. Staff members offered their insights of the data at these workshops, including around social norms and cultural values, to validate programmatic insights and consider lessons learned.

Training and Motivations of WSFs
There was general consensus from RWN staff and WSFs that the participatory curriculum was both informative and engaging. A few staff members noted how pre-testing the WSF curriculum aided participants' comprehension of the training: "After the pre-test, we revised the module and we put in some content that was missing. Their participation was active in the pre-test and it helped us revise the module to use more relevant examples. Like the session of what is GBV, where there are stories that helped them to understand what is happening in their communities." (Field Officer Western Province Round 1) WSFs generally reported having benefited from the training, including to be better equipped to combat GBV, and thus demonstrated high levels of dedication as volunteers. As one WSF said: 'It made me happy to be able to be useful and help women in my neighborhood, to help reduce conflicts in families. It helped me to contribute so that a Rwandan woman can have peace." (WSF Northern Province Round 2) Both the status of the role and the monthly allowance they received to cover transportation costs sustained their motivation as facilitators. Observations of safe space activities found that WSFs had good facilitation skills, but ongoing support and training were essential. Regular monitoring by RWN staff and refresher trainings were critical for strengthening the participatory facilitation, listening skills, reporting mechanisms, and confidence of the WSFs. The extent to which WSFs support each other, after having developed close relationships through the initial training, was also given as an achievement by some of the staff and WSFs.

Motivations to Attend the Women's Safe Spaces
WSFs, attendees, and staff concurred that the majority of women attended the safe spaces due to their own experiences of regular conflict or IPV. For instance, one attendee expressed the appeal of the safe spaces, as she had no one else to talk to about her experiences of IPV. A few attendees shared their initial reluctance to attend the safe spaces, due to concerns of being stigmatized as GBV survivors. They overcame this through witnessing the benefits the safe spaces had for fellow community members, and since the spaces were framed to develop healthy relationships, families, support solidarity and skills development, rather than only providing GBV response and prevention. Another common motivation to attend the safe spaces was to connect with and speak openly with other women, including about the taboo issue of IPV, to the extent that the safe spaces were commonly identified as 'tuganire', meaning 'let us have discussions.' Although not a formal programme component, across all safe spaces, women initiated income-generating activities (IGAs) through encouraging savings and teaching each other skills such as basket weaving. In a context with high poverty levels, this was an additional incentive for many women to attend the safe spaces. The fact that the safe spaces promoted women's empowerment, and raised awareness of their national rights (i.e., to own property, to live free of GBV) was another crucial motivating factor for women to attend the safe spaces: "We give them information through sensitizing them about their rights and most especially that is what draws them to the Space." (WSF Eastern Province Round 2) Many attendees reported their desires to regularly attend the spaces, and the majority of RWN staff and WSFs noted women are repeatedly using the safe spaces. Some women shared that they would prefer to attend the safe spaces more regularly, but were prevented due to transport costs, long distances they had to travel, and/or domestic and agricultural duties. One RWN staff member reported how limited resources could constrain women's ability to attend the safe spaces and/or join the IGAs: Sometimes they are looking for other support, because most of the cases, women are depending on men. They may want more income. Even if we encourage them to come to do solidarity groups, sometimes they don't have money or capacity to contribute to those income-generating activities, or even money for transport. (RWN Field Officer Eastern Province Round 2) This indicates the challenge of only having one WS per sector, which was regularly identified by staff, attendees and WSFs. One staff member (RWN Field Supervisor Eastern Province Round 2) emphasized that this challenge is being mitigated through the WSFs conducting outreach activities. While the majority of safe space attendees were women, men also attended the safe spaces. Indicative of the value of a benefits-based approach, one male attendee noted he was motivated to attend the safe spaces for promoting the development of healthy families. Another male attendee noted his motivation to attend to support survivors of GBV in his community: I could see some of my neighbours being the victims of violence against women and children. I decided to study those lessons so that one day I could be any help because sometimes the victims prefer to keep silent, because they don't care about it or simply because they don't have enough knowledge about it. I felt like I should not fall into that trap of ignorance. (Attendee Northern Province Male Round 2)

Reporting and Disclosure of IPV at the Safe Spaces
There was strong consensus on behalf of staff, attendees and WSFs of attendees' preferences to report experiences of IPV to the women's safe spaces, over other available options. Reasons given for this were having time, confidential and non-judgmental spaces, being offered solutions, and not fearing consequences of reporting IPV, such as victim blaming, their experience being shared publicly, having to pay a fine, or their partner being arrested. As one attendee emphasized: The facilitators keep a secret, and they interact softly with people in a way that you cannot hide anything from them because of the way they are. They are calm and they don't even tell you things such as: 'but I have heard that you get drunk!' so you tell something to a facilitator and she listens to you then you tell her: 'please don't tell this to anyone' then she tells you: 'among things which we have studied, we have also studied to keep the information confidential. You can tell us if you would like that we visit you at home with your husband being present.' (Attendee 2 Northern Province Round 3) One male safe space attendee (Round 2) emphasized the unique contribution of such dedicated support: Another woman will never stop and listen to that problem, in some case she might even laugh at those having problems. On contrary, the WSFs will stop, listen, understand and provide solutions.
One attendee noted the preference for herself and husband to seek help for IPV from the safe spaces, as the WSFs were not from their same close-knit neighborhood: The reason he has even accepted to come to see the WSFs is because they are not from our close neighborhood. He is afraid of the local authorities and he can't accept to go to the local authorities! (Attendee 2 Eastern Province Round 3) One attendee (Attendee 1 Western Province Round 1) shared her appreciation that the WSFs do not ask for money when reporting, and noted how this contrasted to having once been asked for money by community elders upon reporting IPV. She expressed her appreciation that WSFs will accompany women to local authorities, which she felt would reduce the likelihood of being asked for money. Another attendee (WSA02 Western Province Round 1) previously asked local authorities for help with IPV, but they were unavailable after several attempts. This contrasted with the dedicated support attendees were said to receive at the safe spaces: "Women like to come here; there are some who can talk from morning to evening, and they come the next day and ask more questions. They feel that you have time to talk to them." (RWN Field Officer 1 Round 2). One male safe space attendee noted how women prefer reporting to the safe spaces given their general trust of the WSFs: When a given wife has family disputes, she seeks support from one of those facilitators instead of deciding to go back to her parents' place or neighbours to seek support. I have noticed that WSFs play an important role in mediating families that have problems. They trust facilitators. (Attendee 1 Northern Province Male Round 2) The majority of attendees expressed appreciation that WSFs are from their communities, which was said to be critical for establishing such trust and rapport. Several staff and WSFs emphasized how many cases of IPV are resolved through the women's safe spaces, which highlights the value of informal responses: Most of the cases have been addressed at the women's space. I see a small number of cases referred to other institutions for support. This is a big achievement; seeing a person come when their home is almost broken, and their home starts afresh, it's a great achievement. (RWN Field Supervisor Eastern Province Round 2) Yet, the spaces critically supported attendees' access to formal responses where needed, including health, justice and social services, through raising awareness of these services, as well as inviting service providers to the safe spaces. RWN staff and WSFs reported that many cases were resolved at the women's spaces, which commonly included supporting conflict resolution among spouses, providing encouragement, advice or relevant information to attendees. For cases not resolved at the safe spaces, common referrals offered and accessed were to medical services, police and local authorities. Several WSFs and attendees asserted the value of accompaniment to access services. As one WSF said: "when they see that we are sending them there and that we are with them accompanying them, it shows them that it is serious and they immediately receive them, but when they go alone it is hard." (WSF Northern Province Round 3) Several staff members and WSFs described the importance of the safe space referral forms, and the linkages between safe spaces and engaged opinion leaders, which supported programme advocacy for improved GBV services. For instance, a key advocacy success highlighted by RWN staff was to ensure health care providers provide services to GBV survivors free of charge, as part of national health insurance rather than fining them for physical violence, due to an application of a policy to fine individuals who report physical violence, intended as a way to discourage such violence. The linkages fostered between WSFs and opinion leaders supported their information being relayed around improving formal services for IPV services. One of the most significant impacts of the spaces given was that IPV survivors know where to seek support from, and receive better care when doing so: Before one had experienced violence and she reported it to the village level-one would be confused about where to report her case. But now that there are WSFs, when one has a problem she goes to see them and even if it is necessary to seek help from the local leadership, it passes through the WSFs. Before, no one could report her problem to the village level leadership twice. On contrary, thanks to the WSFs, when women have been victims of violence and they seek help from the WSFs, they do advocacy for them. The village leaders didn't used to pay attention to those problems. (WSF Eastern Province Round 3)

Inclusivity at the Women's Safe Spaces
While the safe spaces target women, they are also open to and engage men. This was said to be important for transparency, to avoid backlash from men and support women's access to the spaces. For instance, several participants noted how men could prevent their wives from accessing the safe spaces: "Her husband said 'you are a crazy woman! What are you going to say about us?' He was worried that she would tell everything related to the relationship and family." (WSA01 Northern Province) A few WSFs and attendees discussed the value of home visits that WSFs regularly conducted to engage men and obtain their support. Some attendees and WSFs reported changes from their male partners through their own involvement at the safe spaces: They told me to tell my husband: 'nowadays, everyone has the right to property. You should rather share ideas when you want to sell something.' I told it to my husband and he accepted it. Now there is no problem and when he wants to sell something he first asks me: 'I would like to sell this thing. Do you agree?' then I also agree. (Attendee 1 Western Province Round 3) Yet, many attendees and WSFs shared the difficulties they experienced with engaging their male partners, and requested that men be more actively engaged at the safe spaces: They [the WSFs] said, if I understand that I have power within but my husband can't allow me to do everything I want, he is still controlling me. They were saying it is better if you could be with our partners here so they can have the same understanding. With my partner, it is not easy to tell him I have power! They have insisted please invite our partners. Even to one or two sessions. (Field Officer Eastern Province Round 2) One attendee suggested that only targeting women might suggest women are responsible for IPV: 'I see that women are the ones who attend the discussions more. So it would be better if men could also attend the discussions because violence is not done by only one person: it is committed by both a man and a woman.' (Attendee 2 Eastern Province Round 3) RWN staff responded to this request by inviting WSFs' spouses (if married) to one of the ongoing refresher trainings with WSFs. The safe spaces also invited men (including attendees' spouses) to attend the safe spaces on specific days, which appeared to have a positive impact as anticipated. For instance, a male safe space attendee discussed learning the value of sharing household roles: I learnt an inspiring lesson thanks to an image showing a man and his wife -on that picture the wife was carrying many things while her husband was walking empty handed. The lesson, I learnt from it was that the man was misusing his power. Ever since I decided that I will not do that again because from that image I thought it was describing what I used to do and I decided to change. (Attendee Western Province Male Round 2) Yet, in some observed safe space activity sessions, men appeared to dominate the discussions. One attendee shared her experience of men victim blaming women for IPV at the safe spaces, although she also noted that the safe spaces support men's understanding of IPV: When men are here, they respond and they really show that they do understand, they also admit that the violence is really done. But when they are here, they sometimes blame women to be responsible for what happens. (Attendee 2 Eastern Province Round 3) One staff member (RWN Field Officer Eastern Province, Round 2) indicated the importance of continuing to have dedicated days for women only, to ensure their safety and active involvement. A male attendee suggested reframing the name of the spaces and having male facilitators to make the spaces more inclusive for men, and since men are more likely to listen to other men: Because it is called women's space, they may think it is only reserved for women, but, because the topics covered here are designed for both men and women, I would suggest having male facilitators, because there are some men who would not accept having female facilitators leading them-you would hear them saying, 'what would she tell me?' But if it was a man teaching, he would definitely understand more. (Attendee Western Province Male Round 2) Overall, it was a difficult balance for the programme team to strike between having safe, dedicated spaces for women and promoting men's engagement, which required ongoing critical reflection and adaptive programming.
A few RWN staff members lamented that few adolescent girls attended the spaces, and emphasized the need for more targeted engagement of adolescence. In response to this identified gap, during the last year of the programme, 20 girl champions at each space received training around power and GBV, with the support of the National Youth Council. Girl champions were connected to WSFs for continuous mentorship, and supported to conduct community discussions and encourage female youth to attend the safe spaces. One WSF reflected on the heightened barriers for women with disabilities to attend or regularly attend the safe spaces, and noted the value of WSFs conducting home visits to engage people with disabilities, and offer dedicated support: We visit people with disabilities and discuss power and GBV, because people with disabilities don't attend the discussions we conduct. There is one person I visited and she was like: 'I also have power just like someone who doesn't have any physical disability? I use to feel so small given the way I walk! I didn't even wish to go where others are gathering. I won't feel shy anymore to go where others are. It is so good to know that there are people who think about us. (Women's Safe Space Facilitator, Eastern Province, Round 3) In collaboration with the National Council of Persons with Disabilities, all programme staff and WSFs were trained on disability inclusion and mainstreaming during the last year of the programme. RWN staff then delivered a version of this training to all Indashyikirwa community activists, WSFs and 280 community members living with disabilities. Such efforts were in response to encouragement from the donor, DFID Rwanda, to ensure more inclusive programming for people living with disabilities.

Impact of the Women's Safe Spaces
Staff, WSFs and attendees shared numerous positive impacts of the women's safe spaces. A few staff members related the success of the solidarity groups and IGAs initiated by the spaces, whereby women contribute some money each month and use the cumulative amount to meet each other's needs, such as purchasing materials or food products. In doing so, a critical impact of the safe spaces given was to reduce immediate household poverty, related family conflict and stress. The majority of WSFs and attendees related how their self-confidence improved after having reflected on their 'power within'. This helped challenge inequitable gender norms, such as feeling unable to achieve economic growth or speak in public, with many women feeling better equipped to take on these roles: 'Now I openly speak out and I use the power that I have in me and I feel there is something that I can do to make my family developed. That is a very big thing." (Female Attendee Western Province Round 2) For many attendees, this translated into a commitment to ensure their sons and daughters received equal opportunities, including prioritizing education for both, related to greater recognition that boys and girls should have equal power: We had lessons about treating our children the same way, training them to do the same type of work and to teach them that they both have the same power. This will help our children in the future. I used to think that my daughter should not do the same work as my son. I used to think that a girl should not continue her studies because I was wondering, what will she do with her studies? But now I have come to understand that a girl has the same right as her brother and that she also has the right to go to school and she has the power to do something. (Attendee 1 Eastern Province Round 3) Many women noted that the solidarity at the safe spaces helped reduce their sense of loneliness and anxiety, and meant they could draw on others for emotional and sometimes financial support. Several attendees noted that the quality of the support and care received at the spaces developed their sense of self-worth and strength, which helps to counter the emotional legacy of IPV: In fact, when one has been victim of violence, he/she feels useless. When you have been a victim of violence you even lose hope. So that is exactly the way I was feeling when I came here, I had lost hope and I was feeling that I was unable to do anything but they have helped me out. I have regained more strength gradually. (Attendee 1 Eastern Province Round 3) Several staff members, attendees and WSFs emphasized the value of the safe spaces for raising awareness of women's rights according to Rwandan laws and policies: People know there is a place they can ask for information and advice, which is different from the community. I think it is important to have a physical place where people can discuss, have knowledge of laws. We had a community debate on the family law recently, and some were not aware of that law. (RWN Field Officer 1, Eastern Province Round 2) In addition, several attendees and WSFs reported greater awareness of different types of IPV (emotional, economic, physical and sexual), and openness to discuss and take action against IPV: "Before, conflict between men and women was treated normally in the community but now they know the consequences, they know that GBV can affect the entire family, they want to respond to and prevent it." (RWN Field Officer Eastern Province Round 2) The majority of WSFs and attendees also appreciated the skills-building focus of the safe spaces, which included an emphasis on improving relationship skills through constructive communication and non-violent conflict resolution, as well as access to income-generating skills. Sessions that focused on how to identify and manage triggers of IPV (i.e., alcohol abuse, economic stress/disagreements) were also highly valued: I have learnt how one should behave towards her husband, even when he is drunk, and they have taught us how we should behave when there is someone who is committing violence against us and how we should avoid arguing with such a person. Briefly, the women's safe spaces have given me inner peace. Thanks to them I came to know how to have a good relationship and I learnt about having common agreement with my husband before we take any action regarding the household property. (Attendee 2 Eastern Province Round 3) Many of the WSFs and attendees interviewed reported improved family relationships and reduced personal experiences of IPV, indicative of how the safe spaces can play a role in both response AND prevention of IPV: WSFs keep telling us the importance of the programme in their own lives. They are women who live in the communities. They were also victims of GBV. Their success is not just helping other women but also helping their relationships with their husbands, their children, other colleagues in the community. (RWN Field Supervisor Western Province Round 2)

Discussion
The findings emphasize the value of the women's safe spaces for navigating complex social dynamics to provide inclusive and safe support for survivors of IPV, which were key components of the Indashyikirwa model. The safe spaces offered dedicated and unique informal GBV response, which also enhanced the quality of and linkage to formal GBV response services. Attendees generally preferred the women's safe spaces over formal services for IPV disclosure and support, which emphasizes the importance of programmes providing informal sources of support, and questions the current emphasis in the field on primarily strengthening formal/institutional responses to GBV. The fact that WSFs were elected by community members and resided in intervention communities supported attendees' rapport with and trust of WSFs. WSFs received intensive, strategically-designed training and ongoing support from RWN, which also equipped their quality of care. According to staff and WSFs, regular monitoring, constructive feedback, support by RWN staff and the refresher trainings were critical for strengthening participatory facilitation, listening skills, reporting mechanisms, and the confidence of the WSFs. Although access was limited due to having only one women's safe space per sector, the fact that WSFs were not from the same village as attendees helped to support a sense of confidentiality. As Mannell et al. (2018) note, closely integrated communities can deter women from disclosing violence for fear this will increase the severity of the violence from their partners. The safe spaces allowed women to publicly discuss experiences of IPV, which challenges the salient social norm in Rwanda that IPV should be kept hidden in the private domain (Uwineza and Pearson 2009). Individual support and home visits were also important for those who faced barriers to attend the safe spaces or publicly share their experiences. This openness supported attendees' well-being, which confirms other research in Rwanda indicating that GBV survivors wished to openly discuss their experiences to move forward with their lives (Russell et al. 2016). The solidarity developed at the safe spaces and the concept of 'power within' also supported attendees' confidence and well-being, and was said to reduce anxiety and loneliness. Extensive research conducted on the impact of women's safe spaces or solidarity groups reveals similar benefits through being with others who are facing similar issues including improvements in psychological well-being, self-esteem and self-efficacy, reductions in depression, self-blame and isolation (Sullivan 2012;Santos et al. 2017;Koegler et al. 2018). Furthermore, the positive impact of this type of support for IPV survivors has been shown to be sustainable over longer periods of time (Santos et al. 2017). Another major strength of the women's safe spaces was the opportunity for livelihoods skills training and income generation, which is consistent with other research in Rwanda emphasizing the value of safe spaces combined with income generation activities (Mahr and Campbell 2015). Focus groups with Rwandan GBV survivors that were members of a support group found that these groups strengthened safety, stability, trust and interpersonal connections. Yet members struggled with poverty, hunger and an inability to support their families and requested the groups to begin IGAs (Walstrom et al. 2013).
The safe spaces also supported prevention of GBV through informing attendees about different forms, triggers and consequences of IPV, rights to live free from violence, and supporting healthy relationship skills. Through critically reflecting on gender norms and inequalities underlying IPV, the spaces encouraged healthier and more equitable gender norms, including among boys and girls. This is similar to other studies, which found that women's safe spaces can help challenge the norms that encourage and justify IPV against women (Rodriguez 1999). Yet, a key challenge of comprehensive IPV prevention was the limited engagement of men. Many attendees desired more active male engagement at the spaces, although there were concerns about, and experiences of, women's less active engagement with men present. This is not unusual in the Rwandan context, where cultural norms dictate that men should speak first when present with women, and in many cases women do not speak at all with men present, as men are considered to be more adept leaders (Carlson and Randell 2013). This challenge reflects one of the primary critiques of safe spaces as being exclusionary (i.e., women only) for the sake of unity, empowerment or safety, rather than being inclusive in order to challenge oppression and gender binary categories (Noterman and Rosenfeld 2014). In order to cultivate spaces that do not replicate the inequalities they are meant to dismantle, they should be critically reflexive of what they are promoting safety from and for, and regularly evaluate inclusions and exclusions (Noterman and Rosenfeld 2014). The programme team importantly adjusted their programming in response to identified needs, and tried to manage these dynamics by having some days for women only, and established days inviting men and women together. It is also important to recognize that a physical space can be considered safe for some people, but unsafe for others of a different gender, race, class, age or other identity (Koskela and Pain 2000). Indeed, the programme team reflected on the need for targeted recruitment to ensure safety and engagement of women living with disabilities and youth.
This study is not without limitations. There is a risk of social desirability bias whereby programme participants report overly favourable impacts. This risk was mitigated through staff external to the programme conducting the research, and emphasizing respondents' confidentiality. One of the framed research objectives of the interviews was to learn how to improve the safe spaces, and to emphasize the value of openness including limitations or challenges of the safe spaces. Generally, RWN staff were more open about implementation challenges than beneficiaries, which may be related to such biases. This also speaks to the value of including staff perspectives as part of a comprehensive process evaluation. The authors aimed to be critically reflexive of their identities as Canadian, white, middle-class women. Both authors have lived and worked in Rwanda in the area of GBV prevention and gender equality for a number of years, and the analysis was enriched through the input from Rwandan Indashyikirwa programme partners.

Lessons Learned
Several implementation lessons around dedicated response mechanisms within IPV prevention programmes have been identified from this research, which we offer for practitioners and/or researchers applying or evaluating a safe space model. These include the importance to frame safe spaces as benefits and skills-based (i.e., to develop healthy families), which helped counter the stigma for GBV survivors to initially attend the safe spaces, and encouraged motivation on behalf of men and women. The data indicate how producing safe spaces entails critical reflection on inclusivity; the programme had to carefully consider how to engage men in a gender transformative way but also ensure safe and dedicated spaces for women, while additionally acknowledging that women cannot be treated as one homogenous unit. In future, the safe spaces should consider more strategies for inclusion such as to ensure selection and training of male facilitators and facilitators living with disabilities, and organize clubs to draw young people in. Safe space programming should consider how to use spaces for different activities, given the value of both receiving individual cases, and offering group activities. The data further indicates the critical importance of integrating economic empowerment activities in settings where GBV survivors require economic support. The IGA component of the safe spaces supported attendees' regular engagement with the safe spaces, and could in turn further challenge harmful gender norms and relieve economic stress. This economic component may be critical for sustainability of the spaces, especially to maintain the unique physical safe spaces. The dedication of the WSFs and their quality of care indicates the strength of the WSFs recruitment process and initial training, which was critically pre-tested and drew on evidence-based programming including SASA! It also speaks to the value of the ongoing support provided to WSFs through mentoring and ongoing trainings by RWN staff. With such programmatic support, community members (WSFs) can provide valuable informal GBV response and prevention, and act as a key liaison to formalized responses. Finally, the ongoing evaluation findings were workshopped with RWN senior staff to strategize solutions for improving the spaces. This included establishing dedicated days where men as individuals or couples are invited to the spaces, and identifying capacity needs of WSFs. This speaks to the importance of research informing programming on an ongoing basis, with an emphasis on adaptive programming.
A critical lesson learned related to the difficult balance of offering safe spaces for women while actively engaging men, is to have more comprehensively worked with couples. The safe spaces could have offered the Indashyikirwa couples curriculum; a 21-session training to support couples to identify and manage triggers of IPV, and foster equitable, healthy relationships (Stern and Nyiratunga 2017). The pathways to reduced IPV among safe space attendees interviewed suggests that women learned to more effectively mitigate the risks of violence, such as avoiding husbands when drunk, more effectively resolving conflict, or raising awareness of their rights. However, they requested more active engagement of their spouses, to facilitate more transformative and comprehensive relationship level changes. Safe space facilitators and attendees would have been aware of the Indashyikirwa couples curriculum, which may explain why they requested similar programmatic emphasis on working with couples. Working with couples more explicitly may have equipped the safe spaces to be more gender transformative, for more adequately shifting inequitable power relations, and ensuring the onus of violence mitigation and prevention is not placed solely on women.

Conclusions
This study uniquely explored the implementation process and assessment of women's safe spaces as part of a comprehensive IPV prevention programme. The findings indicate the preference for women to disclose IPV and seek support from the safe spaces over formalized services, and how the safe spaces could importantly bridge access to formalized services. The findings also suggest that the safe spaces supported attendees' economic empowerment, well-being, and development of skills. Although safe spaces are limited and have challenges as documented, they can be a critical component of GBV prevention programming. Indeed, the local government and community members have recognized the value of the safe spaces, and the majority have continued to be sustained, even after the formal programme and associated funding came to a close. The findings indicate the importance of programmatic conceptualization of safety to entail dedicated, confidential, and empowering mechanisms of support, while critically reflecting on and adapting to ensure inclusivity. We recommend that prevention programming considers such dedicated integration of response mechanisms, and gives more attention to informal sources of GBV response.