**5. Discussion**

We have presented the factors that helped former child brides to leave the violent marriages. The aim was to highlight potential protective factors that can be built, harnessed or reinforced to empower and encourage girls to swiftly leave violent intimate partner relationships.

The study identified four factors that were key in helping child marriage survivors to leave violent unions: having a secure base to return to, where they could be supported emotionally and materially; reaching a tipping point when the violence escalates or they realise that it will only escalate more and that their partners will not change; intervention of significant others who provide assistance that is crucial for leaving or physically picking up the survivor; and financial independence that enables an independent living away from the marriage, though this may only be feasible for those women who are old enough to work independently. The significance of informal support, financial independence and escalation of violence in women's decisions to leave violent intimate relationships has been discussed elsewhere (Anderson and Saunders 2003; Kim and Gray 2008; Lacey 2010; Lacey et al. 2011; Baholo et al. 2015; Khoury and Wehbi 2016). However, this study shows that for adolescents, having a secure place with loving and supportive parents/guardians to return to is particularly critical if they are to sustainably leave violent marriages. It not only encourages them to leave, but also prevents them from returning to violent unions. As shown, abused adolescents typically returned to their parents/guardians. Those who did not find adequate support from their parents/guardians returned to their violent partners because they had neither the resources nor the capacity to su fficiently support themselves without assistance. Therefore, parental support can be seen as a critical factor in helping young child marriage survivors in a low resource setting to escape violence in the relationships. This is because family remains the main fallback for most people experiencing crises or need in these contexts. This pattern supports the findings of Willan and colleagues (2019), who show that young women from impoverished informal settlements of South Africa are more likely to leave violent intimate relationships if they have strong emotional and economic support from their families.

The significance of parental support in child marriage survivors' decisions to leave and stay away from violent relationships underlines the need to equip parents with positive parenting skills. Described as an approach to parenting that is child-centred and which emphasises care, support, encouragement, empathy and positive disciplining (UNICEF 2018), positive parenting will contribute to making homes a safe haven where children and young people facing IPV can seek refuge. Child marriage survivors who anticipated love, empathy and support from their parents/guardians typically thought of returning home to escape the violence. Moreover, positive parenting can address drivers related to the quality of parenting, such as domestic child abuse and neglect (Bantebya et al. 2014), and contribute to preventing child marriages in the long run. Since several child marriages are preceded by elopement and unwanted teenage pregnancies that strain relationships between girls and their parents/guardians, it is important that parents/guardians are helped to understand that sometimes children only learn when they face uncomfortable consequences from their actions (UNICEF 2018). This will enable them to proactively forgive and provide girls who wish to leave violent relationships a second chance to learn and grow from their mistakes. As our data show, girls were only able to sustainably leave violent marriages where their parents/guardians displayed positive parenting attributes such as forgiving and taking them in, listening to them and understanding their situations, encouraging them to leave and providing them both material and emotional support during the process of and after leaving violent unions.

However, positive parenting may be di fficult to realise in a context of poverty and socio-cultural norms that condone violence against women and girls and harmful traditional practices such as child marriage. Therefore, optimising the protective benefits of positive parenting interventions requires that they be proactively complemented with economic strengthening programmes to improve the capacity of parents/guardians to provide for their children. In addition, continuous community engagemen<sup>t</sup> is necessary to change negative social norms and customs that promote the tolerance of violence against women and girls, as these often prevent parents/guardians from helping their children to leave abusive marriages. These include beliefs that normalise IPV in marriage and the custom of paying bride price. Our data show that several parents/guardians encouraged girls to stay in violent relationships because they had been socialised to believe that IPV is a normal day-to-day experience of married women. An analysis of 2016 demographic and health survey data for Uganda also found high levels of tolerance towards IPV against women in the country (Ghose and Yaya 2019). This normalisation of violence could be attributed in part to the forces of patriarchy and male dominance which prevail in most African societies, a demonstration of the preponderance of structure over the human agency of the girls (Archer 2000, 2003). Other parents/guardians may encourage girls to stay in violent relationships when bride price has been paid. It is not uncommon for African women who wish to leave violent marriages to meet resistance from their parents, who are often afraid of being asked to refund the bride price (Wendo 2004), again buttressed by the power of structure (Archer 2003).

Our data show that several girls and women tend to see returning to their natal home as the only viable way out of violent marriages, even where family members are hostile towards them. This pattern not only suggests a paucity of formal protection services for survivors of GBV such as temporary shelters and rehabilitation centres, but also limited awareness of the few available ones. This observation is corroborated by country-specific reports which highlight the limited funding for GBV-related services and activities as a key impediment to its elimination in Uganda (MGLSD 2015b). It is for instance indicated that the country has only 16 shelters for GBV survivors, in spite of the high incidence of violence against women and girls. Moreover, survivors' access to the few GBV shelters and related services is not only constrained by inadequate funding but also by the limited awareness about their availability (MGLSD 2018; Mugerwa and Wesaka 2020). Therefore, there is a need to invest in protection services to increase safety nets for survivors of GBV whose informal support networks may be feeble or absent. In addition, a massive awareness of available services should be created to increase demand for them and encourage women and girls who would otherwise be entrapped in violent relationships due to lack of viable alternatives to leave. Awareness of a shelter has been identified as a key facilitator to women's departure from abusive intimate relationships in other parts of Africa (Baholo et al. 2015).

While the support of significant others such as family and friends is often critical in helping abused women leave violent relationships (Khoury and Wehbi 2016; Willan et al. 2019), our study shows that the intervention of survivors' significant others can be both a facilitator and a barrier in decisions to leave. Several of our participants accessed resources that were crucial in enabling them to leave with the support of their relations, while others left at their encouragement. However, some of the social relations including kin, neighbours and friends encouraged several women and girls to stay in violent relationships by normalising the abuse. This suggests that community members are a potentially powerful resource in fighting IPV and GBV in general and providing support to survivors, if equipped with the correct information. In this regard, harnessing the protective role of the community requires interventions to challenge and change beliefs and social norms that condone violence against women and girls. It is important, furthermore, to increase community awareness of GBV, its forms, dangers, existing interventions, the referral pathway and the potential role of individuals and families in preventing, reporting and helping survivors to escape the abuse. Community dialogues drawing participants from di fferent sections of the community can be a useful strategy for providing information on IPV and GBV and getting them to reflect on the drivers and dangers of the two practices and on possible solutions to addressing them. Mechanisms to engage children and young people in age-appropriate conversations on the meaning and manifestations of IPV and GBV, its consequences, how to protect themselves and others from abuse and where to seek support may help to enhance their capacity to recognise and take action. Some of these topics could be incorporated in the sexual and reproductive health education curriculum for adolescents and young people and tackled during discussions of intimate relationships.

It is evident that in the context of child marriage, the significance of specific factors in decisions to leave may depend on the age of the survivor. For instance, data show that having a secure base to return to was more critical for adolescents who were generally dependent and too young to support themselves and their o ffspring on their own, while financial independence applied more to survivors who left as adults and sought to live independently. This suggests that programmes targeting survivors of IPV involved in child marriage should consider being flexible enough to address the likely variations in the needs of di fferent age groups.

Drawing on structuration theory and the agency-structure debate (Sewell 1992; Wendt 1987; Giddens 1984), we recognise both the importance of human agency and the importance of rules, regulations, norms and value systems in influencing the young women's decisions and ability to leave violent unions. This implies that interventions to encourage young women to leave abusive relationships ought to address both individual and structural constraints for success.
