Exploring opportunities for coordinated responses to intimate partner violence and child maltreatment in low and middle income countries: a scoping review

Abstract

Intimate partner violence (IPV) and child maltreatment (CM) by a parent or caregiver are prevalent and overlapping issues with damaging consequences for those affected. This scoping review aimed to identify opportunities for greater coordination between IPV and CM programmes in low- and middle-income countries (LMIC). Nine bibliographic databases were searched and grey literature was identified through the scoping review team. Eligible studies were published in English; described primary prevention programmes in LMIC that addressed IPV and CM, or addressed one form of violence, but reported outcomes for the other; reported IPV and CM outcomes; and evaluated with any study design. Six studies were identified published between 2013 and 2016 (four randomised controlled trials, one pre-post non-randomised study and one qualitative study). Programmes were based in South Africa (2), Uganda, (2), Liberia (1) and Thailand (1). All except one were delivered within parenting programmes. The emphasis on gender norms varied between programmes. Some parenting programmes addressed gender inequity indirectly by promoting joint decision-making and open communication between caregivers. Conclusions are tentative due to the small evidence base and methodological weaknesses. More robust evaluations are needed. Improved coherence between IPV and CM programmes requires equal attention to the needs of women and children, and the involvement of fathers when it is safe to do so.

Keywords:

• Intimate partner violence
• child maltreatment
• low middle income countries
• interventions
• scoping review

Introduction

Violence against women (VAW) and violence against children (VAC) are recognised as a serious global health concern and a violation of human rights (World Health Organization, 2000). Yet historical developments in the VAW and VAC fields have led to these phenomena being considered distinct from one another. Whilst this has been essential for the development of adequate laws, advocacy and programmes, there have been calls for prevention strategies that address both forms of violence. This review aimed to scope the current evidence to identify synergistic intervention opportunities for a more coordinated approach.

Intimate partner violence (IPV) is one of the most common forms of VAW and includes acts of physical, sexual, and emotional abuse and controlling behaviours by an intimate partner (World Health Organization, 2000). Global estimates show that 30% of women aged 15 and over have ever experienced physical and/or sexual violence by an intimate partner (Devries et al., 2013) and 38.6% of all female homicides are perpetrated by intimate partners (Stöckl et al., 2013). Women who are physically or sexually abused by their partners are more likely to have an abortion, suffer from depression, and in some regions are more likely to acquire HIV compared with women who have not experienced IPV (World Health Organization, 2013).

Child maltreatment (CM) is a common form of violence against children (VAC) perpetrated by a parent or caregiver and includes acts of commission, such as physical, sexual and psychological abuse, as well as acts of omission such as neglect or exposure to violent environments (Leeb, Paulozzi, Melanson, Simon, & Arias, 2008). However, in the context of this review which focuses on CM that occurs in the family context (i.e. by a parent or caregiver), we limit our definition of ‘acts of omission’ solely to children’s exposure to IPV in the home, which is associated with impairment similar to other types of maltreatment (MacMillan & Wathen, 2014). The United Nations Children’s Fund (UNICEF) reports that around 6 in 10 children worldwide aged 2–14 experience regular physical punishment and 7 in 10 experienced psychological aggression by caregivers (United Nations, 2014). The immediate consequences of CM include physical injury, cognitive impairment, impaired attachment, and symptoms consistent with depression and post-traumatic stress disorder (Norman, Byambaa, Butchart, Scott, & Vos, 2012) although the damage to health and social functioning can last into adolescence and adulthood (Felitti et al., 1998; Oladeji, Makanjuola, & Gureje, 2010; Ramiro, Madrid, & Brown, 2010; Releva, Peshevska, & Sethi, 2013; Tran, Dunne, Van Vo, & Luu, 2015).

The rationale for this review stems from increasing evidence that IPV and CM intersect on a number of levels. Guedes, Bott, Garcia-Moreno, and Colombini (2016) define four aspects of this intersection: (i) overlapping risk factors (e.g. unemployment, poverty and social isolation); (ii) the presence of social norms that condone violence; (iii) co-occurrence of IPV and CM in the same family, which has implications for the intergenerational transmission of violence; and (iv) similar health outcomes. IPV and CM share a number of commonly associated underlying risk factors, which include unemployment, poverty, high levels of community violence and social isolation, as well as individual level factors such as poor mental health and substance abuse (Alhusen, Ho, Smith, & Campbell, 2014). IPV and CM are also associated with some of the same social norms that condone violence and reinforce gender inequality. These norms include victim blaming attitudes that reinforce male sexual entitlement and support men’s right to control women, as well as norms that prioritise family privacy and the belief that corporal punishment of children is necessary (Alhusen et al., 2014; Guedes et al., 2016).

The presence of IPV in the home is a risk factor for CM (Hamby, Finkelhor, Turner, & Ormrod, 2010) and the high co-occurrence of IPV and CM has been reported in LMIC (Dalal, Lawoko, & Jansson, 2010; Gage & Silvestre, 2010; Rada, 2014). Childhood exposure to IPV is associated with multiple health problems including internalising behaviour problems (e.g. anxiety and depression), externalising behaviour problems (e.g. aggression, delinquency) and trauma symptoms (Evans, Davies, & DiLillo, 2008). Exposure to certain forms of IPV and CM, for example sexual abuse, have been shown to have similar consequences with regards to mental health outcomes and social functioning (Wilkins, Tsao, Hertz, Davis, & Klevens, 2014). Furthermore, the risk of adult victimisation and/or perpetration of IPV is greater amongst children who have been abused, highlighting the intergenerational transmission of violence (Abramsky et al., 2011; Fry, McCoy, & Swales, 2012; Spatz Widom, Czaja, & Dutton, 2014). These intersections are important to consider as this paper examines promising interventions to prevent and respond to IPV and CM.

The dominant programmatic efforts for addressing IPV in high income countries have been response-driven and focused on providing services to survivors (Ellsberg et al., 2014). In comparison, research and programmes in LMIC have prioritized primary prevention. These use multiple approaches such as media campaigns and community mobilisation, economic empowerment and group education which aim to change attitudes and norms that reinforce violence against women and girls and promote gender-equitable behaviours (Abramsky et al., 2014; Jewkes, Nduna, Levin, et al., 2008; Kim, Watts, Hargreaves, et al., 2007; Wagman, Gray, Campbell, et al., 2015). According to recent reviews, the most effective programs in reducing IPV in LMIC are those that involve community mobilisation and/or economic empowerment paired with gender equality training (Ellsberg et al., 2014; Heise, 2011). Some of these programmes also seek to address issues that are concomitant with violence against women and girls, such as HIV, poverty, low education and women’s economic dependence on men.

Prevention strategies to reduce violence against children outside of the family have predominantly been implemented in schools in high-income countries, mainly in the form of group education and training to address either teen dating violence or childhood sexual abuse (De La Rue, Polanin, Espelage, & Pigott, 2014; MacMillan et al., 2009; Walsh, Zwi, Woolfenden, & Shlonsky, 2015). Prevention programmes for violence in the family have generally been embedded within home visiting programmes, group or individual based parenting programmes, and paediatric care. Some perinatal home visiting programmes, such as the Nurse Family Partnership, and early childhood parenting programmes have been shown to prevent or reduce certain forms of CM such as physical abuse and neglect, whilst others have been found to prevent reoccurrence of CM (Barlow, Simkiss, & Stewart-Brown, 2006; MacMillan et al., 2009). However, the evidence base is relatively modest and sometimes of poor quality. With regards to children exposed to IPV, a systematic review found that parent skills training, delivered in combination with practical support for non-abusing mothers and group based psycho-education delivered to mothers and children may be effective for improving children’s behavioural outcomes, although this is a tentative conclusion based on a small number of studies (Howarth et al., 2016). The evidence for CM prevention programs in LMIC is scant (Mikton & Butchart, 2009). A systematic review of parenting programmes in LMIC to reduce harsh and abusive parenting identified 12 randomised trials, of which two high quality trials reported positive effects of the intervention in reducing dysfunctional or harsh parenting (Knerr, Gardner, & Cluver, 2013).

The pervasive link between IPV and CM has resulted in calls for the provision of comprehensive and complementary services to families affected by these forms of violence (Guedes & Mikton, 2013; Herrenkohl, Higgins, Merrick, & Leeb, 2015; Lessard & Alvarez-Lizotte, 2015). Amongst the various intervention approaches in high-income countries, only the health care sector has recognised the potential for addressing IPV and CM as co-occurring issues, for example in home visiting and paediatrics (Dubowitz, Feigelman, Lane, & Kim, 2009; Dubowitz, Lane, Semiatin, & Magder, 2012; Prossman, Lo Fo Wong, van der Vouden, & Lagro-Janssen, 2015). In the United States, where there has been a substantial investment in perinatal home visiting programmes, guidelines for joint IPV and CM interventions have been developed for policymakers (Family Violence Prevention Fund, 2010). In LMIC where resources are low, there is a need to maximise prevention efforts particularly in view of the shared risk factors and health consequences.

Despite growing evidence of the intersection between IPV and CM, there is a paucity of research regarding effective strategies for addressing both forms of violence. The goal of this scoping review was to identify interventions that have measured outcomes for both IPV and CM and programme components that may have contributed to positive outcomes. Due to the fact that current evidence focuses largely on high-income countries, this paper focuses on interventions in LMIC to build the knowledge base in less developed settings.

Methods

Our scoping approach was informed by Arksey and O’Malley’s (2005) methodological framework which comprises six phases that do not necessarily occur in a linear manner: (i) identifying the research question; (ii) searching for relevant studies; (iii) selecting relevant studies; (iv) charting the data; (v) collating, summarising and reporting the results; (vi) and consulting with stakeholders. The research question for this paper was developed through stakeholder consensus at the Know Violence in Childhood Learning Initiative meeting on intersections between violence against children and violence and women (22–24 April 2015). The workshop brought together leading experts in the fields of VAW and VAC to explore the potential for core principles and key skills involved in developing a shared approach to preventing these forms of violence, as well as identify where priorities differ. A scoping study team, comprised of the co-authors of this paper, was formed to contribute expertise (i.e. programmatic, policy, research and advocacy) at various stages of the scoping review (Levac, Colquhoun, and O’Brien (2010).

Search procedure

The scoping team agreed on the inclusion and exclusion criteria for the review (Figure 1).

Figure 1. Inclusion and exclusion criteria.

Nine bibliographic databases (MEDLINE, EMBASE, Global Health, Health Management Information Consortium, Cumulative Index to Nursing and Allied Health Literature [CINAHL], Africa Wide, Latin American and Caribbean Health Sciences [LILACS], Index Medicus for South-East Asia [IMSEAR], Index Medicus for the Eastern Mediterranean Region [IMEMR] were searched from 2010 to 2015 using controlled vocabularies for each database and text words (Appendix 1). To identify studies prior to 2010, we drew upon systematic reviews of interventions to prevent or reduce violence against women and girls (Arango, Morton, Gennari, Kiplesund, & Ellsberg, 2014; Ellsberg et al., 2014) and of interventions to prevent child maltreatment (Knerr et al., 2013; MacMillan et al., 2009; McCloskey, 2011; Mikton & Butchart, 2009). Study selection was an iterative process. First, a list of potentially relevant studies resulting from this search was sent to the scoping team requesting that they check for missing studies, particularly unpublished data and grey literature. Bacchus contacted implementers of three parenting programmes in process that address IPV and CM in order to enquire about interim findings. One team member who was involved in a large review of interventions for violence against women and girls checked the list against their own database (Contreras Urbina). Another team member (Gardner) with expertise in parenting programmes provided a list of 20 contacts for parenting programmes in LMIC, to whom an email and one follow-up was sent to request unpublished findings regarding IPV related outcomes. Following this, an amended list of studies was sent to the scoping team for review.

Screening and data extraction

One author (Bacchus) screened all abstracts and full texts of potentially eligible studies and extracted the data. The scoping team agreed on what information should be extracted which included: country; programme name; programme aims; key components (i.e. topics in the curriculum); setting in which the intervention was delivered; target groups; evaluation methods (study design, measures used to assess IPV and CM, data collection activities, sample size, follow-up period); and IPV/CM findings. Key components were extracted directly from the training manuals where these were available, from publications and by contacting authors and programme implementers. We looked specifically for programme content which engaged participants in exploring gender norms and values in the context of parenting, decision-making and violence against women and children. This stems from our feminist epistemological position which seeks to understand violence against women within the interaction of gender norms, power and patriarchy which create inequalities and disadvantage women in many spheres of life (Yllo, 2005). Furthermore, programmes that address gender inequitable norms have been shown to reduce violence and improve health outcomes (Dworkin, Treves-Kagan, & Lippman, 2013; Jewkes, Flood & Lang, 2015).

The data was tabularised in Word and reviewed by the scoping team to determine whether additional data needed to be extracted and agree on information to be presented in the synthesis. The findings and abridged tables were also presented to leading experts at the second Know Violence in Childhood Learning Initiative meeting on intersections between VAW and VAC (10–11 March 2016). This provided another opportunity to discuss the findings and key recommendations arising from the review.

Synthesis

A narrative approach was used which provides descriptive information about the overall number of studies included, years of publication, the countries in which they were based, and whether they addressed IPV only, CM only or both. A formal quality appraisal was not undertaken due to the variation in the study designs used to evaluate programmes. Therefore, the findings from individual programmes are presented by levels of evidence starting with randomised controlled trials. With regards to mixed methods studies (i.e. trials with a nested qualitative study) the trial evidence is presented first. Findings from the individual programmes include a brief description of the content (including whether or not it addressed gender norms and values) and the target groups. Statistical tests from pre and post measures are reported for quantitative studies and key themes are presented from qualitative data, supported by quotes. The limitations of the study designs are elaborated in the discussion.

Results

The search strategies retrieved 1387 studies published between 2013 and 2016, of which 6 were directly relevant to the aims of the review. Two of the studies were based in South Africa, two in Uganda, one in Liberia and one in Thailand. Of the six studies, two described programmes that were designed to address IPV and CM jointly (Ashburn, Kerner, Ojamuge, & Lundgren, 2016; Hatcher, Colvin, Ndlovu, & Dworkin, 2014; Van den Berg et al., 2013); one was designed to address IPV, but reported unintended outcomes for CM (Abramsky et al., 2014, 2016; Kyegombe et al., 2015); and three were designed to address CM, but reported unintended outcomes for IPV (Cluver, Lachman et al., 2016; Cluver, Meinck et al., 2016; Sim, Puffer et al., 2014; Sim, Annan, Puffer, Salhi, & Betancourt, 2014).

Randomised controlled trials

Table 1 presents four studies which used a randomised controlled trial with a nested qualitative component: REAL Fathers in Uganda (Ashburn et al., 2015); SASA! in Uganda (Abramsky et al., 2014, 2016); Parents Make the Difference in Liberia (Sim, Puffer et al., 2014); and Building Happy Families in Thailand (Sim, Annan, et al., 2014).

Table 1. Evidence from randomised controlled trials.

Author/year	Country	Intervention name	Programme aims/key components	Setting and target group	Methods	IPV and CM outcomes/themes
Ashburn et al. (2016)	Uganda	Responsible, Engaged and Loving (REAL) Fathers	Programme aims: A 6-month father-centred mentoring programme plus a community awareness campaign to improve knowledge and skills in positive parenting and conflict resolution, and reflection on gender roles or parents in childcare	Setting: delivered in community settings. Amuru district, Northern region of Uganda	Randomised controlled trial with a nested qualitative study. N = 250 men assigned to the mentoring programme and community poster series and N = 250 men assigned to the community poster series only in two cohorts. Fathers interviewed at baseline, end line (4 months post intervention) and follow-up (12 months for cohort 1 and 8 months for cohort 2, after completing the intervention).	CM quantitative findings: significant reduction in physical punishment in the exposed versus the unexposed group at follow-up, but not at end line (62.0 vs. 42.1; p < .000). Parents attitudes rejecting use of physical punishment were significantly higher at end line (53.6 vs. 64.5; p = .023) and follow-up (42.2 vs. 65.7; p < .000). Parents confidence in dealing with a child without using threats, shouting or beating significantly increased at end line (15.1 vs. 28.9; p < .001) and follow-up (22.9 vs. 45.1; p < .000)
			Key components: Six large scale posters were sequenced over time in central meeting locations to spark community dialogue and reinforce mentoring sessions. 64 mentors, chosen by the young men from within their community, were trained and supported up to 4 fathers each. Mentoring included 6 individual (2 of which included wives) and 6 group (1 of which included wives) mentoring sessions. Topic areas: understanding gender values and norms; parenting (respecting the child and the child’s mother, spending time with children, disciplining with love, being a role model and teacher, talking and listening to children, showing love); effective communication in the home, stronger couples through communication (including how to resolve differences within couples without violence); dealing with stress and managing emotions	Target group: young fathers aged 16–25 years of age who have a child aged 1 to 3 years and are cohabiting with an intimate partner	CM outcomes taken from parent-child Conflict Tactics Scale: six items measuring types of violence during child discipline in the past month; attitudes toward the use of physical punishment. IPV perpetration in the past three month measured with an adapted Conflict Tactics Scale including verbal, psychological, physical; justification of IPV. Qualitative interviews with 20 men and 10 women in the intervention group.	CM qualitative themes: some men reported spending more time with children in play and positive interaction; that their children were less afraid of them and that they had greater awareness that the use of physical punishment made their children more aggressive
Ashburn et al. (2016)						IPV quantitative findings: IPV scores (any form) were significantly lower in the exposed versus unexposed group at end line (68.7 vs. 53.1; p < .001) and follow-up (47.6 vs. 28.8; p < .000). Significant reduction in verbal IPV at end line (58.1 vs. 42.6; p < .001) and follow-up (38.6 vs. 23.6; p < .001)
						IPV qualitative themes: some men reported a reduction or cessation of alcohol use which they linked to less use of violence towards their wives and being more cooperative. Some women still felt a strain in the relationship with reports of men resuming their use of violence after alcohol use.
						Significant reduction in psychological IPV at end line (36.9 vs. 25.0; p = .008) and follow-up 25.3 vs. 12.0; p < .001). The reduction in physical IPV was significant at endline (36.8 vs 27.7; p=0.043), but not at long term follow up
						Significant reduction in attitudes related to justification of IPV at end line (39.1 vs. 28.2; p = .019) and follow-up (35.4 vs. 21.2; p = .002)
Kyegombe et al. (2015)	Uganda	SASA!	Programme aims: Community mobilisation approach that seeks to prevent violence against women and HIV. SASA! is unique in its focus on power (positive and negative uses) and shifts away from a traditional focus on gender norms. Includes a range of activities that focuses on changing attitudes, norms and behaviours that underpin power imbalances between men and women and reinforce HIV risk behaviours	Setting: delivered in community settings. Eight communities in Kampala (Rubaga and Makindye Divisions)	Mixed methods design using baseline and follow-up survey data from the SASA! randomised controlled trial and interview data from a nested qualitative study. Quantitative outcomes based on women with at least one biological/step child living in the household	IPV quantitative findings: women in the intervention communities were less likely to report past year physical or sexual IPV than controls (aRR = .68; 95% CI .16, 1.39)
				Target group: community members	1538 baseline and 2532 4 year follow-up surveys completed	CM quantitative findings: of women who did report past year IPV, fewer reported that a child was present or overhead physical or sexual IPV (aRR = .58; 95% CI .19, 1.74). Authors estimate that reductions in IPV combined with reduced witnessing by children when IPV did occur, led to a 64% reduction in the prevalence of children witnessing IPV in their home (aRR = .36; 95% CI .06, 2.20)
			Key components: The programme includes four phases: (i) START: staff from implementing organisation learn about the community (ii) AWARENESS: building activists’ confidence as they conduct activities in the community (iii) SUPPORT: skills and connections between community members are strengthened to encourage people who are trying to foster change and (iv) ACTION: individuals try out new behaviours and celebrate change within their community. Intervention staff work with four groups: community activists selected from progressive men and women in the community; community leaders (i.e. religious, cultural); professionals (i.e. police, health etc. who provide direct services); and institutional leaders who have power to implement policy change. Strategies used included: local activism, media and advocacy, communication materials and training. One-on-one and group based activities are used		Past year experience IPV and attitudes towards IPV measured with the WHO Multi-Country study on Women’s Health and Domestic Violence Survey	CM qualitative themes: being a positive role model for children (i.e. showing that violence is not normal or acceptable); men spending more time with their children, change in attitude about being solely a provider of money; improved communication with children; improved anger management strategies; less use or rejection of harsh discipline and violence towards children. In reducing IPV, children’s exposure to IPV is also reduced
					For the qualitative evaluation, 82 men and women who participated in SASA! activities were interviewed including community members, community activists and community leaders
Abramsky et al. (2014, 2016)	Uganda	SASA!	Same as Kyegombe et al. (2015)	Same as Kyegombe et al. (2015)	Pair matched cluster randomised controlled trial in eight communities	IPV quantitative findings: the intervention was associated with significantly lower social acceptance of IPV among women (aRR = .54; 95% CI .38, .79) and men (aRR = .13; 95% CI .01, 1.15); significantly greater acceptance that a woman can refuse sex among women (aRR = 1.28; 95% CI 1.07, 1.52) and men (aRR = 1.31; 95% CI 1.00, 1.70)
					Past year experience IPV and attitudes towards IPV measured with the WHO Multi-Country study on Women’s Health and Domestic Violence Survey	52% lower past year experience of physical IPV among women (aRR = .48; 95% CI .16, 1.39); and lower levels of past year sexual IPV (aRR = .76; 95% CI .433, 1.72) although not statistically significant
					Baseline survey: (Intervention communities: n = 374 women and n = 419 men; 97%) and (Control communities: n = 343 women and n = 447 men; 98%). At four year follow up: (Intervention communities: n = 600 women and n = 768 men, 99%) and (Control communities: n = 530 women and n = 634 men; 98%)	Intervention was also associated with lower continuation of prior abuse. Statistically significant effects were observed for continued physical IPV (aRR = .42; 95% CI .18 to .96); continued sexual IPV (aRR = .68; 95% CI .53 to .87); continued emotional aggression (aRR = .68; 95% CI .52 to .89); continued fear of partner (aRR = .67; 95% CI .51 to .89); and new onset of controlling behaviours (aRR = .38; 95% CI .23 to .62)
Sim, Puffer et al. (2014)	Liberia	Parents make the difference	Programme aims: To decrease the use of physical and psychological punishment; increase the use of positive parenting strategies; and increase malaria prevention	Setting: delivered in community settings plus one home visit. Rural Liberia not specified further	Randomised waitlist controlled design with a nested qualitative study and observations of caregiver and child in an unstructured play activity	CM quantitative findings: average decrease of 56% in the use of physical and psychological punishment (effect size −.61; p < .001). Percentage of caregivers who reported beating, whipping and spanking their child in the last month decreased by 64% (effect size −.67; p < .001); 62% (effect size −.62; p < .001) and 56% (effect size −.42; p < .001) respectively. The use of psychological punishment (e.g. yelling at child) decreased by 29% (effect size −.65; p < .001). 9% of caregivers in the treatment group reported beating their child the last time they misbehaved compared to 45% in the control group
				Target group: caregivers and children aged 3 to 7 years. 57% of caregiver participants were women	135 families in the intervention and 135 in the control group, randomised by lottery and stratified by community	CM qualitative themes: some caregivers reported using non-violent discipline techniques, increased awareness that beating was not effective and would harm their child’s development and adopting a more nurturing parenting approach. Some respondents reported that learning about the negative effects of domestic violence on children was motivation for reducing family conflict and being a positive role model
			Key components: 10 weekly group sessions facilitated by two trained Liberian staff from IRC, plus one individual home visit. The home visit was used for individualised support and to discuss previous sessions. Parent support groups provided a forum for sharing		Baseline and one month follow-up surveys measured parenting attitudes, beliefs and practices including use of violent and non-violent discipline using validated measures: Discipline Module of Multiple Indicator Cluster Survey (MICS); Parental Acceptance and Rejection Questionnaire (PARQ); Adult-Adolescent Parenting Inventory (AAPI-2); Dyadic Parent-Child Interaction Coding System (DPICS)
			Underpinned by behavioural theory, highly skills-based, provides opportunities for discussion and skills practice. Content includes: negative effects of physical and psychological punishment; use of non-violent discipline; positive parenting interaction and communication; strategies for stimulating children’s cognitive development including communication and activity to promote children’s numeracy, vocabulary and critical thinking skills; basics of child development and importance of active involvement in children’s education. The programme also included one session on malaria prevention		Purposive sample of 30 caregivers participated in a semi-structured interview
						IPV qualitative themes: the intervention had a positive effect on some caregiver’s relationship with their partner. More open communication, collaborative problem solving and understanding of one another appeared to reduce conflict and violence in the home. Some men and women reported less anger towards each other. Some men also reported decreased use of drugs and alcohol and spending more time at home, which may also have helped to reduce conflict and violence in their relationships with their spouses
Sim, Annan et al. (2014)	Thailand	Building Happy Families	Programme aims: Increase the use of positive parenting skills; decrease harsh punishment; improve family functioning and child psychosocial wellbeing	Setting: delivered in community settings. 479 households from 20 urban and rural communities in the Tak province	Waitlist randomised controlled design	CM quantitative findings: caregivers reported an average decrease of 13% in the use of harsh discipline overall as measured by Discipline Interview (effect size −.40; p < .001). 90% decrease in scaring their child into behaving well, 18% decrease in beating their child and 17% decrease in swearing at their child
				Target group: caregivers and children aged 8 to 12 years. Burmese migrant and displaced families living on Thai-Burmese border. 83% of parent/caregiver participants were women	Intervention group: n = 256 caregivers and n = 240 children. Control group: n = 257 caregivers and n = 239 children	Children reported a small, non-significant decrease in their caregivers’ use of harsh punishment overall on the Discipline Interview (effect size −.12) and a 15% reduction in spanking and slapping (effect size −.33; p < .001). Using the second measure (MICS), caregiver reports found a small and non-significant decrease in harsh discipline overall (effect size −.10). Analysing individual items on MICS, only using a hard object to beat significantly decreased by 16% (effect size −.22; p < .01). Reductions in harsh parenting overall were maintained at 6 months follow-up in the intervention group
			Key components: Delivered by IRC programme staff and community based facilitators. Adapted from the Strengthening Families Programme. A 12 week group-based parenting family skills intervention for children aged 8–12 years and their caregivers. Caregivers and children participate in parallel group sessions each week followed by joint activities in which skills can be practice under supervision. The programme also included structured opportunities for positive interactions (e.g. family meal ending with games)		Baseline, 1 month end line, and six-month follow-up. measures included: Discipline module of Multiple Indicator Cluster Survey, MICS); Parental Acceptance and Rejection Questionnaire, Burmese Family Functioning Scale; Child Behaviour Checklist/Youth self-report; Burmese Child Resilience Scale; and Alcohol Use Disorders Identification Test (AUDIT) administered to caregivers. End line survey was conducted with all participants at 1 month after the intervention. Only participants in intervention group completed a 6-momth follow-up
			Caregiver content included: setting appropriate developmental expectations; understanding negative consequences of harsh punishment; non-violent discipline strategies; positive communication and problem solving skills; behaviour management; effects of alcohol and drugs on families; managing stress; and maintaining change. Children’s content included: speaking and listening to others; rewarding good behaviour; peer resistance; communication skills; coping and resilience skills; problems solving; recognising feelings; dealing with criticism and anger management
					Semi-structured interviews conducted with a purposive sample of 25 families in the intervention group	CM qualitative themes: some caregivers described decreased use of or cessation of harsh physical punishment and were no longer swearing or shouting at their children, or using hurtful language towards them. They also reported increased empathy for children in relation to how harsh punishment can negatively affect their development
						IPV qualitative themes: some caregivers reported improvements in their relationships with their partners. For example, reduced conflict, fewer fights, improved communication, more discussion regarding household finances and problem solving. The relaxation techniques they were taught to reduce stress (e.g. breathing and relaxation exercises) may have contributed to improved relationships and interactions with children, partners and community members. Some caregivers reported reducing their use of alcohol

REAL Fathers, a father-centred mentoring programme in Uganda which targets young fathers aged 16–25, is designed to address IPV and CM. The programme aims to improve knowledge and skills in positive parenting, communication and conflict resolution, encourage reflection on the gender roles of parents in childcare, and improve acceptance of non-traditional gender roles. Mentors are recruited from the study communities and chosen by the young men as people whom they respect and can take advice from.

In the REAL Fathers trial, young fathers aged 16–25 who were cohabiting with their partner were eligible to participate, with 250 assigned to the intervention group and 250 to the control group. Men who received the intervention engaged in mentor facilitated discussion groups with other fathers, and individual and couple mentoring. Topics include: understanding gender values and norms; parenting (including talking and listening to children and showing love); effective communication in the home and between couples (including resolving conflict without violence); and dealing with stress and managing emotions. In the final group session on parenting, wives are invited to participate. In addition, a poster series representing topics from the curriculum is implemented in locations frequented by young fathers in order to stimulate discussion. Fathers in the control group were only exposed to the posters series. Men completed outcomes measures at end line (4 months post intervention) and at longer term follow-up (12 and 8 months respectively for cohorts 1 and 2). Women were not followed-up so as not to compromise their safety and due to the lack of local IPV resources (Ashburn et al., 2015). Analysis compared men exposed to the intervention (defined as at least one individual and one group mentoring session) versus men not exposed at endline and longer term follow up. Unique identification codes were not used during data collection because of concerns about confidentiality. Therefore the survey data were analysed as two cross sectional surveys post intervention rather than panel data. According to men’s reports, there was a significant reduction in the use of physical punishment to discipline children at longer term follow up (aOR = .52, 95% CI: 32 to .82, p < .001) and in IPV at end line (aOR .48, 95% CI: .31 to .76, p < .01) and at longer term follow up (aOR = .48, 95% CI: .31 to .77, p < .01). Men also reported significantly higher levels of confidence in dealing with their child’s behaviour without resorting to violence or verbal threats at end line (aOR = 2.5, 95% CI: 1.50 to 4.28; p < .001) and over the longer term (aOR = 2.4; 95% CI: 1.55 to 3.98, p < .001).

The qualitative component of REAL Fathers involved interviews with 20 men and 10 women in the intervention group Ashburn et al., (2015). Some men reported an increased awareness that using physical punishment to discipline only made children more aggressive. In addition, some of the wives and partners of the young fathers commented that children appeared to be less afraid of their fathers. The positive impact of the programme on men’s relationships with their partners may, in part, be related to a reduction in their use of alcohol which some of them linked to less use of violence and being more cooperative in the home regarding household chores and child care.

Before this mentorship, I was a drunkard and violent [fought] my wife a lot. I had the wrong peer company who only know drinking alcohol as a way of life, but after the REAL Fathers mentorship, I could see and understand clearly. I had to dump my friends and become a real friend to my family. I stopped drinking alcohol, my violence vanished, we started communicating and working well … [Young Father]

According to the interviews with women, this behaviour change was not always sustained and some reported that their partner had reverted to using violence, often accompanied by alcohol use.

At the beginning of the program things were working well, he would understand me and I also understand him. Later came a time when we went to attend the training, upon my returning home my husband started a terrible quarrel, accusing me that I took long there [at the training] … he threatened to slap me if I answered him … Adding to that he could return home very late … he resorted to too much drinking. [Wife of a REAL participant]

SASA! in Uganda was designed to address IPV, but reported unintended outcomes for CM. Drawing on the ecological model in its programme design, SASA! challenges social norms and beliefs about gender that contribute to violence using a community mobilisation approach which actively engages stakeholders within the community including activists, local government, cultural leaders, religious leaders, and professionals such as the police and health care providers. The language used in the programme focuses on how power can produce positive and negative outcomes, and encourages participants to consider this in the context of relationships between men and women in different spheres of life. The programme has four phases described further in Table 1 and engages men and women of all ages in a range of one-to-one and group activities to discuss and engage on issues of gender inequality, violence and HIV. SASA! encourages critical reflection on violence against women and the development of communication and relationship skills. It also encourages activism against violence at the community level. Trained community activists conduct informal activities within their own social networks, among their families, friends, colleagues and neighbours. Consequently, community members are exposed to SASA! ideas repeatedly and in diverse ways within the course of their daily lives, from people they know and trust as well as from more formal sources within their communities (Abramsky et al., 2014).

The IPV outcomes and children’s exposure to IPV are derived from a pair-matched cluster randomised controlled trial in eight communities in Kampala, with 1538 men and women (aged 18–49 years) completing a baseline survey and 2532 completing the four-year follow-up. The evaluation also included a qualitative component from which the CM outcomes are derived. The intervention was associated with significantly lower social acceptance of IPV among women (aRR = .54, 95% CI .38, .79) and men (aRR = .13, 95% CI .01, 1.15); and significantly greater acceptance that a woman can refuse unwanted sex among women (aRR = 1.28, 95% CI 1.07, 1.52) and men (aRR = 1.31, 95% CI 1.00, 1.70). There was a 52% lower past year experience of physical IPV among women and lower levels of past year sexual IPV, although this was not statistically significant (Abramsky et al., 2014). However, SASA! did have a significant impact on stopping violence from continuing, where it occurred previously (Abramsky et al. 2016). Statistically significant effects were observed for continued physical IPV (aRR = .42; 95% CI .18 to .96); continued sexual IPV (aRR = .68; 95% CI .53 to .87); continued emotional aggression (aRR = .68; 95% CI .52 to .89); continued fear of partner (aRR = .67; 95% CI .51 to .89); and new onset of controlling behaviours (aRR = .38; 95% CI .23 to .62).

In order to examine changes in children’s exposure to IPV, Kyegombe et al. (2015) used baseline and follow-up survey data from a subset of women in the trial who also reported past year IPV at baseline, and qualitative interviews with a sub-sample of men and women who participated in the intervention. Women in the intervention communities were less likely to report past year physical or sexual IPV than those in the control communities (aRR = .68; 95% CI .16, 1.39). Amongst the women who experienced past year IPV, fewer reported that a child was present or overhead physical or sexual IPV. The reduction in past year IPV combined with reduced witnessing by children when IPV did occur, led to a 64% reduction in the prevalence of children witnessing IPV in their home (aRR = .36; 95% CI .06, 2.20). Amongst couples that experienced a reduction in IPV, qualitative data suggested that this had a positive effect on parent-child relationships through improved parenting and discipline practices. Some participants also reported being less tolerant of violence against children in their community and more willing to intervene when necessary.

Before I joined SASA! I used to think that as a man I used to have all the power in the home so whenever a child made a mistake I would, without understanding, punish the child badly. But from when I joined SASA! whenever a child makes a mistake, I have to first understand the cause of the mistake. [Male Community Member 18]

For us who have been to those sessions [SASA! activities] we are like attorneys for such children or we are like watchmen for abused people in the community. On many occasions I have confronted parents and rebuked their actions. [Male Community Member 10]

Parents Make the Difference in Liberia (Sim, Puffer, et al., 2014) and Building Happy Families in Thailand (Sim, Annan, et al., 2014) were designed to address CM, but reported unintended outcomes for IPV. In both studies, the CM outcomes were assessed quantitatively within the trial, whilst the IPV outcomes were reported in the qualitative component. Parents Make the Difference in rural Liberia is adapted from existing evidence-based parenting programmes and targets parents and caregivers with children aged 3–7 years of age. It uses behavioural theory and is highly skills-based, providing caregivers with specific techniques to promote positive caregiver-child interactions and positive discipline strategies. Group sessions are designed to be interactive, with a focus on discussion, modelling and practicing of skills. This is supplemented with a single home visit for individualised feedback and ongoing support is provided via parenting groups. The programme does not explore parenting within the context of traditional gender norms and values, or how these perpetuate violence in the home. In a randomised controlled waitlist trial with a one month follow-up, 135 families were randomly assigned to the intervention group and 135 families to the control group. An observational assessment was conducted of each caregiver and child pair comprising a brief unstructured play activity that was audio-recorded. Additionally, semi-structured interviews were conducted with 30 caregivers who participated in the intervention. In relation to the quantitatively measured CM outcomes, the study reported a statistically significant decrease of 56% (effect size −.61; p < .001) in caregiver’s use of physical and psychological punishment. Qualitative interviews also revealed reports of some parents no longer beating their children or using harsh discipline such as shouting or denying their children food, and having an increased recognition of the harmful effects of aggressive discipline on children.

First I used to beat on them because they were not understanding me at all, but right after this training the people taught me how to counsel your children, how to talk to them so that they can change and be somebody better, which I did. My children now don't hestitate to do things I ask them which is the change I saw in them. It's beause the way I used to treat them, I'm not treating them like that again. [31 year old father]

With regards to the IPV findings, in the qualitative interviews some caregivers reported that the intervention had an unintended positive impact on their relationship with their partner in terms of improved communication, problem solving and understanding of each other. Some men also reported decreased use of drugs and alcohol and spending more time with their families, which they perceived to contribute to reduced conflict in the home and less use of violence with spouses (Sim, Puffer et al., 2014).

One of the main changes in my woman and I are not making confusion again like the way we used to make palaver [arguments] every time, and the people [facilitators] are even telling us not to be making palaver and abusing our woman because if we have confusion, our children will practice that from us. [47 year old father]

I used to drink and smoke, but thank God I'm dropping all those things now, because the money I'm taking to buy cigarettes and liquor I can use that as recess for my children. Since the people came and started advising us how to take care of our children I looked into it and I left all of things. [39 year old father]

Happy Families is a parenting and family skills intervention for Burmese migrant and displaced families living on the Thai-Burmese border, and is adapted from the evidence-based US programme Strengthening Families (Kumpfer, Pinyuchon, Teixeira de Melo, & Whiteside, 2008). Sim, Annan, et al. (2014) contend that many of the stressors associated with forced migration, such as economic hardship, psychological distress, discrimination, abuse and weakening of social support structures, are known to have a negative impact on parenting. These factors can also compromise the protective capabilities of parents and may increase the risk of child maltreatment and abuse. The 12-week course provides parallel group sessions for caregivers and their children, in addition to joint sessions in which they can practice the skills learned. The programme content focuses on helping caregivers to understand their children’s development and teaches caregivers and children communication and problem solving skills. For example, parents are taught how to reward good behaviour, set goals and objectives with children, problem solve, manage behaviour and maintain changes. Child sessions focus on speaking and listening to others, problem solving, recognising feelings, and dealing with criticism and anger. Both parents and child sessions include content on the effects of alcohol and drugs. The curriculum does not include content which encourages participants to explore traditional social norms and beliefs about gender in the context of parenting or in relation to gender based violence.

A waitlist randomised controlled design with a one month follow-up was used to measure the impact of the intervention on CM outcomes. The programme was implemented in 20 urban and rural communities in the Tak province. Parents or caregivers of Burmese origin with children aged 8–12 years of age were eligible to participate in the programme. A further 6-month follow-up was undertaken with the intervention, but not the control group. A purposive sample of 25 families from the intervention group participated in a semi-structured interview in which themes relating to IPV were identified. With regards to CM, the intervention had a significant medium effect on reducing harsh discipline overall as measured by the Discipline Interview (effect size −.40, p < .001). Children reported a small, but non-significant decrease in their caregiver’s use of harsh discipline overall on the child report version of the Discipline Interview and a 15% reduction specifically in spanking and slapping which was significant (effect size −.33; p < .01). Using the second measure, the discipline module of the Multiple Indicator Cluster Survey, there was a small, non-significant decrease in harsh discipline practices overall as reported by caregivers only (effect size -0.10). When looking at individual items on harsh discipline, the only significant result was a 16% decrease in using a hard object to beat their child (-0.22, p<0.01). Reductions in harsh or negative parenting overall were maintained at six-month follow up with the intervention group.

In the semi-structured interviews, some caregivers reported using less harsh physical punishment to discipline their children and described an increased awareness of the negative impact that harsh discipline had on their child’s development.

If my children don’t listen to me, I do meditation. I don’t let myself have a hot temper. [Before] I threw everything, cooking pots and plates. My children didn’t dare to stay with me when I was angry. Now I try to control my temper … They told us to calm down by using breathing exercises, controlling our mind. [56 year old mother]

Although IPV was not assessed quantitatively within the trial, the author reports that the qualitative interview data highlighted that the communication and problem solving skills component had a positive impact on some caregiver’s relationships, evidenced through fewer fights with spouses. Furthermore, that the intervention may have had an unintended and positive impact on parents’ wellbeing, for example, through the relaxation and breathing techniques that they were taught which helped them to regulate negative emotions. More open communication and shared decision-making between caregivers was perceived to reduce the levels of conflict in the home. Men’s reduced alcohol use may also have contributed to the improved relationships with children and reduced conflict with spouses. However, no illustrative quotes were provided for these latter findings (Sim, Annan, et al., 2014).

Pre and post non-randomised pilot evaluation

Table 2 presents one study which used a pre-post non-randomised design. Sinovuyo Caring Families in South Africa is distinct from other parenting programmes described, as it targets older children who have been identified as having behavioural problems or a suspected history of abuse (Cluver, Lachman et al. 2016; Cluver, Meinck et al. 2016). Designed specifically to address CM, it also reported outcomes for IPV, both measured quantitatively. The programme uses group-based parent workshops, adolescent and joint parent-adolescent sessions so that skills can be practiced together. A buddy system consisting of peer support provides help to participants between sessions. The curriculum draws on evidence-based parenting programmes and includes collaborative problem solving, home practice and discussion (Cluver, Lachman et al. 2016). Session content includes trust building, talking about emotions, dealing with stress and anger, joint problem solving, non-violent discipline techniques, rules and routines, responding to a crisis, and keeping adolescents safe in the community. The curriculum does not explore parenting within the context of traditional gender norms and values, or how gender norms reinforce violence in the home.

Table 2. Evidence from pre-post non-randomised studies.

Author/year	Country	Intervention name	Programme aims and key components	Setting and target group	Methods	IPV and CM outcomes/themes
Cluver, Lachman et al., (2016)	South Africa	Sinovuyo Caring Families Teen Programme	Programme aims: Increase parenting skills and confidence; decrease harsh discipline; help teenagers manage their behaviour; improve mental health and social support; improve problem solving skills; help families respond better to crisis situations; improve knowledge of services for violence, illness and arrest; reduce stress that families feel about money	Setting: delivered in community settings	Pre-post non-randomised pilot test of the programme. 60 participants (30 caregiver-adolescent dyads) from two high-poverty, rural communities of South Africa’s Eastern Cape province.	CM quantitative findings: significant reductions in the use of violent and abusive discipline following the intervention, as reported by caregivers (pre-test  = 7.94, SD = 7.72; post-test  = 1.63, SD = 2.83; t = 4.18, df = 15, p = .001; effect size −1.09; 95% CI −1.80, −.32) and adolescents (pre-test  = 25.53, SD = 4.52; post-test = 21.87, SD = 2.11; t = 2.39, df = 29, p = .024; effect size −.47; 95% CI −.99, .06). Improvements in positive parenting following the intervention as reported by caregivers (pre-test = 117.75, SD = 14.27; post-test = 132.13, SD = 13.20; t = −4.49, df = 23, p = .000; effect size 1.05; 95% CI .43, 1.63) and adolescent (pre-test = 118.24, SD = 12.99; post-test = 127.38, SD = 13.98; t = −3.85, df = 23, p = .001; effect size .68; 95% CI .08, 1.25)
				Target group: families were identified by the NGO as having expressed challenges with their adolescents, or with adolescent behavior, or families where the NGO or community suspected violence. There were no exclusions for severity of circumstances nor for mental or physical health problems or any other cause. Adolescents were aged 10 to 17 years. 97% of parents/caregivers participants were women	Interviews at baseline and two weeks post-intervention. Violent and abusive discipline measured using the International Society for Prevention of Child Abuse and Neglect (IPSCAN) child and parent version of the International Child Abuse Screening Tool (ICAST-C and ICAST-P). Positive parenting assessed using relevant child and parent subscales of the Alabama Parenting Questionnaire (APQ).	IPV quantitative findings: significant reduction in acceptance of gender and sexual violence on the GEM scale following the intervention as reported by amongst caregivers (pre-test = 24.23, SD = 3.72; post-test = 22.37, SD = 3.88; t = 3.39, df = 29 p = .002) and adolescents (pre-test = 27.90, SD = 5.16; post-test = 25.62, SD = 3.88; t = 2.18, df = 29, p = .038) - personal communication.
			Key components: Comprised 12 weekly sessions, each lasting 2–2.5 h. For eight sessions caregivers and adolescents attended jointly, and in four sessions they attended in separate caregiver and adolescent groups. The programme followed core principles of evidence-based parenting programmes, including collaborative problem-solving, home practice and discussion, and skills-based active participation. Content included: building trust and spending time; praising each other; identifying and talking about emotions; dealing with stress, fear, shame and anger; joint problem-solving; non-violent discipline; rules and routines; keeping adolescents safe in the community; and responding to crisis. A peer-support system of ‘Sinovuyo buddies’ was introduced to help participants between sessions, as low literacy levels limited the use of written materials. Also included mindfulness-based physical exercises for stress reduction		Attitudes to gender and sexual violence (parents and adolescents) were measured using the Gender Equitable Men scale (GEM).
Cluver, Meinck et al. (2016)	South Africa	Sinovuyo Caring Families Programmes	Same as Cluver, Lachman et al. (2016)	Setting: delivered in community settings	Larger pre-post study. 230 participants (115 adolescent-caregiver dyads), living in six deprived rural and peri-urban communities	CM quantitative findings: abuse of adolescents within the home (physical, emotional, neglect) significantly decreased following the intervention (p < .001 adolescent and caregiver reports), dropping from an average score of 4.33 (SE.57) to 1.33 (SE.27) for adolescents and an average score of 11.32 (SE.84) to 1.68 (SE.36) for caregivers. Proportions of adolescents reporting within-home abuse were 63.0% at pre-test, and 29.5% at post-test, and proportions of caregivers reporting within-home abuse were 75.5% at pre-test and 36.5% at post-test
				Target group: Two thirds of families were referred by NGOs, schools, clinics, chieftains, and social workers, based on family conflict or challenges in dealing with adolescents. The remainder of families were approached door-to-door. No eligibility criteria exclusions were mde regarding factors such as parental or adolescent literacy, mental or physical health, or domestic violence. Adolescents were aged 10 to 17 years. 94% of parents/caregivers participants were women	Interviews at baseline and two to six weeks post-intervention. Violent and abusive discipline measured using the International Society for Prevention of Child Abuse and Neglect (IPSCAN) child and parent version of the International Child Abuse Screening Tool (ICAST-C and ICAST-P). Positive parenting assessed using relevant child and parent subscales of the Alabama Parenting Questionnaire (APQ). Attitudes to gender and sexual violence (parents and adolescents) were measured using the Gender Equitable Men scale (GEM)	Positive and involved parenting showed improvements following the intervention as reported by adolescents (pre-test  = 48.71, SE = 1.07; post-test = 51.62, SE = .91; p = .01) and caregivers (pre-test  = 49.23, SE = .98; post-test = 53.83, SE = .81; p = .01) df = 23, p < .001). Poor monitoring and inconsistent discipline decreased following the intervention, as reported by adolescents (pre-test  = 19.64, SE = 1.01; post-test = 15.52, SE = .90; p < .001) and caregivers (pre-test  = 24.36, SE = 1.21; post-test = 16.87, SE = 1.06; p < .001)

In the first pre-post pilot study 30 adolescents and their caregivers were referred to the programme by a local NGOs. At the two week follow-up there were significant reductions in the use of violent and abusive discipline on parent (pre-test  = 7.94, SD = 7.72; post-test  = 1.63, SD = 2.83; t = 4.18, df = 15, p = .001) and adolescent measures (pre-test  = 25.53, SD = 4.52; post-test  = 21.87, SD = 2.11; t = 2.39, df = 29, p = .024). Positive parenting also significantly improved according to parent (pre-test  = 117.75, SD = 14.27; post-test  = 132.13, SD = 13.20; t = −4.49, df = 23, p = .000) and adolescent reports (pre-test  = 118.24, SD = 12.99; post-test  = 127.38, SD = 13.98; t = −3.85, df = 23, p = .001).

IPV amongst caregivers was measured using the Conflict Tactics Scale (Straus, Hamby, Boney-McCoy, & Sugerman, 1996) and children were also asked two separate questions regarding how many days there were arguments with adults shouting at each other and hitting each other in the home. There were no significant differences between the pre and post test scores in relation to the parent or children measures. Attitudes towards gender based and sexual violence was assessed using the Gender Equitable Men Scale (Pulerwitz & Barker, 2008). The pre-post pilot found significant reductions amongst parents (pre-test  = 24.23, SD = 3.72; post-test  = 22.37, SD = 3.88; t = 3.39, df = 29, p = .002) and adolescents (t = 2.18, df = 29, p = .38) in their acceptance of gender and sexual violence post intervention (Cluver – personal communication).

A second, larger pre-post study of Sinovuyo was conducted with 115 adolescents and their caregivers (Cluver, Meinck et al., 2016). At the two to six week follow-up physical, emotional abuse and neglect of adolescents within the home significantly decreased according to adolescent and caregiver reports (p < .001) dropping from an average score of 4.33 (SE.57) to 1.33 (SE.27) for adolescents and 11.32 (SE.84) to 1.68 (SE.36) for caregivers. Positive and involved parenting showed significant improvement following the intervention, as reported by both adolescents (pre-test  = 48.71, SE 1.07; post-test  = 51.62, SE .91; p = .01) and caregivers (pre-test  = 49.23, SE .98; post-test  = 53.83, SE .81; p < .001). There was also a reduction in physical abuse, emotional abuse and neglect following the intervention according to adolescent (pre-test  = 4.33, SE = .57; post-test = 1.33, SE = .27; p = .001) and caregiver reports (pre-test  = 11.22, SE = .84; post-test = 1.68, SE = .36; p = .001). A large cluster randomised trial is currently underway.

Qualitative evaluation

One Man Can (OMC) Fatherhood Programme in South Africa was evaluated qualitatively and presented in Table 3 (Hatcher et al., 2014; Van den Berg et al., 2013). OMC is a gender-transformative, masculinities and rights-based programme that aims to reduce violence against women, decrease levels of unsafe sex and promote more gender equitable relations. It is underpinned by the notion that fatherhood is an opportune time for challenging harmful masculine norms and explores with men the disadvantages that accompany male privilege such as reduced intimacy with partners and children. The curriculum includes specific sessions that address gender violence, gender, sex and HIV/AIDS, developing healthy relationships, and content on fatherhood that explores non-traditional gender roles in caregiving, non-violence and the needs and rights of children.

Table 3. Evidence from qualitative studies.

Author/year	Country	Intervention name	Programme aims and key components	Setting and target group	Methods	IPV and CM outcomes/themes
Van den Berg et al. (2013)	South Africa	One Man Can Fatherhood Project	Programme aims: Gender transformative programme to: (i) increase men’s involvement in the lives of their own children, and ensure that children in general have access to essential social services and have their psychosocial and educational needs met (ii) develop men’s capacity to be activists in efforts to eliminate violence against women and children and prevent the spread of HIV/AIDS (iii) give voice to vulnerable children and raise awareness of their needs within the community	Setting: delivered in community settings	Qualitative evaluation consisting of 90 in-depth interviews with men during the 6 months after they completed the programme	CM qualitative themes: shifts in parenting style from disciplinary and authoritarian to a more caring, nurturing and protective role; men’s recognition that their role extended beyond that of being a financial provider for the family; less use of violence and corporal punishment; and improved communication with children and spending more time with them.
			Key components: Workshops of 15–20 men are facilitated by men. Activities focus on the costs of masculinity and encourages men to reflect on how masculinities is practiced in relationships with women, men and the broader community. Workshop content: gender power and health; gender and violence; gender, sex and HIV/AIDS; and healthy relationships. Fatherhood content: father’s impact; pregnancy; birth; family planning; caregiving; gender roles; non-violence; needs and rights of children; division of caregiving	Target group: Black South African men aged 18 and over in communities affected by high AIDS mortality and where children are left vulnerable by the deaths of one or both parents
Hatcher et al. (2014)	South Africa	One Man Can Fatherhood Project	Same as Van den Berg et al. (2013)	Same as Van den Berg et al. (2013)	Qualitative evaluation. In-depth interviews with 53 men during the 6 months after they completed the programme	IPV and CM qualitative themes: Reduced alcohol use by men potentially linked to less risky sexual behaviour and reduced relationship conflict. More open and emotional communication with partners and children. Communication changes including shifts towards more gender equality in decision-making, including sexual decision-making and more respectful handling of volatile emotional states to prevent escalation to violence

The programme targets black South African men aged 18 years and over living in communities with high AIDS morality where children are left vulnerable by the deaths of one or both parents. Since one of the aims is to increase men’s involvement in the lives of their own children and children in general, it is not restricted to men who are fathers. The programme is delivered as a series of workshops in community settings. Van den Berg et al. (2013) conducted interviews with 90 men within six months of completing the programme. One of the key themes identified focused on a shift in parenting style from disciplinary and authoritarian, to a more caring and nurturing role, with men reporting less use of violence and corporal punishment and improved communication with their children. Hatcher et al. (2014) also conducted interviews with 53 men within six months of completing the programme. Three key themes were identified which included reduced alcohol intake which was linked to shifting ideals of manhood; improved communication with more equal and shared decision-making between men and their partners; and shifting views around sexual entitlement and more shared sexual decision-making. It appeared that the programme helped some men to learn new ways of communicating respectfully with their partners and children in order to avoid escalation of emotions and the use of violence. Within the specific sub-theme of ‘reduced violence’, a few men also attributed this to their reduced alcohol and marijuana use.

In one of my frequent drunken states, I would go and look for my girlfriend and when I wanted her to come along with me there would be no compromise. My word was the final word and I would not take any input from her. Attending the OMC workshops, I got to understand the wrongs of my past behaviour and I started understanding that men should listen to women’s inputs. [Khuzani, 33 years, 9 sessions]

… I used to drink every day and go home drunk and shouting to my children. I really have changed, I have completely stopped drinking. [Sizwe, 62 years, 1 session]

It was one of my fascinations to hear men defining power that people have within the communities that included sexual power carried by men over women. When I looked at the topic deeply, I then had to search inside me and compare what I do to women as well to influence their decision … But things have changed. Even women can protect and provide for the family. [Makondelela, 42 years, 2 sessions]

Discussion

This review identified six programmes in LMIC that demonstrate promise for developing a coordinated response to IPV and CM. It also identified key programme components that may have contributed to positive outcomes for both forms of violence. With the exception of SASA!, all were parenting education programmes targeting caregivers and children. Only two (One Man Can Fatherhood Programme in South Africa and Real Fathers in Uganda) were specifically designed to address both forms of violence. All programmes were delivered in community settings and engaged participants in group education and discussion sessions, although REAL Fathers in Uganda supplemented this with individual and couple mentoring sessions and Parents Make the Difference in Liberia included a home visit.

The interventions were complex and it was difficult to determine precisely which components were directly responsible for producing the promising outcomes for IPV and CM. The emphasis on gender and gender norms varied between programmes. Parenting programmes focussed primarily on improving parent-child relationships and reducing harsh and abusive parenting. However, they also enabled couples to develop better communication skills with each other, which encouraged joint decision-making in relation to caregiving, household and financial issues and collaborative problem solving. This indirect mechanism for addressing gender equity appeared to reduce conflict in caregiver relationships, and improve overall family functioning and cohesion. Caregivers appeared better equipped to resolve emotionally heightened situations with each other before they escalated to violence. This highlights the potential for caregivers to be positive role models for their children and prevent future IPV perpetration and victimisation.

There were other unintended positive outcomes in some programmes. Building Happy Families and Parents Make the Difference documented reports of reduced alcohol use by male partners in the qualitative interviews, although this was only a specific programme component in Building Happy Families. Reduced alcohol use by male partners appeared to contribute to improved relationships and some men reported spending more time at home with their families as opposed to drinking with peers, which helped to reduce partner conflict. Whilst some lower level forms of aggression may be borne out of family stress, there is a danger in conveying to programme participants that IPV is solely a communication issue or due to family stress and programmes must ensure that men take responsibility for their use of violence.

The focus on female caregivers is a noted commonality in the general literature on parenting programmes and the subject of considerable discussion. Three parenting programmes targeted male and female caregivers (Sinovuyo, Parents Make the Difference and Building Happy Families). However, with the exception of Parents Make the Difference in Liberia (in which 57% of attendees were female), the majority of caregiver participants were female (94–97% in Sinovuyo and 92% in Happy Families). There are concerns that interventions that target or are primarily attended by female caregivers fail to address structural and other contextual factors that impact children, families and communities. Therefore, such programmes may inadvertently reinforce traditional gender roles and ideologies that can increase women’s risk of gender based violence (Daly et al., 2015). However, SASA! demonstrated that it is possible to address gender based violence without focussing explicitly on traditional gender norms that can sometimes discourage community participation. Discussions about the positive and negative aspects of power can be an indirect way of addressing the imbalances between men and women and how this manifests in different spheres of life (Abramsky et al., 2014; Kyegombe et al., 2015).

In recent years, fatherhood programmes have been identified as a key environment in which to transform harmful masculine norms that underpin gender based violence (Levtov, van der Gaag, Greene, Kaufman, & Barker, 2015; McAllister, Burgess, Kato, & Barker, 2012). Although the two fatherhood programmes in this review included components that address traditional gender norms in relation to caregiving, female caregiver involvement was minimal or absent (Ashburn et al., 2015; Hatcher et al., 2014; Van den Berg et al., 2013). There is growing recognition that fatherhood programmes need to work alongside efforts to support and protect women and children exposed to family violence. Another gap within parenting programmes is the lack of provision for adolescents as many are designed for young children (Daly et al., 2015). Although interventions for parents of adolescents are rare in LMIC countries, our review identified two which targeted older children (Building Happy Families in Thailand and Sinovuyo Caring Families in South Africa). Future programmes and services need to be tailored to the needs of both younger children and adolescents, where the latter are at higher risk of exposure to and perpetration of IPV.

Although the evidence in this review was concentrated in parenting programmes, there are other settings in which greater coherence between CM and IPV programming can be achieved. The SASA! community mobilisation programme in Uganda did not include a specific parenting component, yet the study found that after the intervention fewer children witnessed IPV and that men were spending more time with their children, using less or rejecting harsh discipline and violence towards their children. The mechanisms through which the intervention may have impacted on children witnessing and experiencing violence are multiple. The programme encouraged participants to reflect on the consequences of violence for their relationships with partners, children and other community members. It taught communication, joint decision-making and conflict resolution skills between couples, which parents may have adopted in their parenting practices. It also encouraged more connected and intimate relationships, which may have impacted on parents spending more time with children and listening to them. Beyond the relationship level, SASA! also played a role in reducing the acceptability of violence and fostering a sense of responsibility to act to prevent violence and communities had more supportive structures (e.g. community activists) to do this. Both SASA! and REAL Fathers, which identified and trained local men to act as mentors to deliver the programme, illustrate the importance of broader social networks, influential people and community engagement in supporting and sustaining positive changes in behaviour.

The push towards more coordinated responses to IPV and CM has given rise to discussions about the potential risks of a combined agenda. Advocates in the fields of VAW and VAC have highlighted that integration may not always be the best approach, or in the best interests of women and children, and that separate interventions are sometimes necessary (Guedes et al., 2016). Specific concerns relate to the historical protection of children being prioritised over the safety of women, if it is determined that their children are being exposed to IPV. Typically, mothers are held solely responsible for the health, safety and wellbeing of their children, regardless of whether or not they are responsible for their abuse. Failing to address the needs of maltreating fathers creates problems for all family members and increases the risk of violence to women and children. As the fields of IPV and CM start to converge there has been growing awareness that both parents play an important role in ensuring child safety and wellbeing and that interventions are needed for fathers that address both CM and IPV. However, interventions must be underpinned by accountability principles that prioritise the safety and wellbeing needs of children and mothers (Peled, 2000; Scott & Crooks, 2007). Furthermore, research and discussion is needed regarding how to support the parenting practices of men who are known to be abusive to their partner and in which circumstances this should be promoted or restricted. McMahon and Pence (1995) maintain that a considered and critical perspective is needed to ensure that the policies and procedures of such programmes do not perpetuate gender inequality and the damaging consequences for women and children.

Conflicting priorities, policies and the differential allocation of resources across sectors and organisations may impede a coordinated response. More work is needed to encourage collaborative working across the VAW and VAC sectors without either feeling undermined. This includes agreed safety measures to reduce possible increased risk of violence to women and children when IPV and CM are identified as occurring within the same household, with equal attention to women and children. There must also be consideration of the level of resources available in countries to develop such programmes and the scarcity of qualified professionals in some settings. Three of the programmes (SASA!, REAL Fathers and Building Happy Families) used trained community members to help deliver the intervention, an approach which may provide a solution to the lack of professional staff. With the exception of SASA! which was community based, the interventions were all targeted, focussing on young children within a specific age group (Parents Make the Difference, Building Happy Families), older children with identified behavioural problems (Sinovuyo), young fathers of toddler aged children (REAL Fathers) or men living in communities affected by high AIDS mortality (One Man Can). Whilst targeted approaches are lower cost than universal programmes, there are disadvantages that should be considered. Not all eligible families will want to enrol in targetted programmes because of stigma (e.g. children with behavioural problems or where there is a known history of abuse or alcohol problems in the home), or due to negative attitudes from the community which result in segregation. Furthermore, strict eligibility criteria may exclude families that could benefit from the programme. Women and children affected by violence in the home, may move in and out of the eligibility criteria for targeted programmes as their family circumstances change. Participation may be higher and with greater community integration in universal programmes, for examples in schools or health care settings. Due to scarcity of resources in some LMIC, a hybrid approach might be preferable. This would ensure that all families receive some services, with more intensive services provided if additional needs are identified.

Limitations

Caution is warranted when interpreting the findings. The evidence is limited due to the small number of studies identified, limitations in the study designs and the exclusion of articles written in languages other than English. With regards to the strength of the evidence, not all outcomes were measured quantitatively in the studies that used a trial design, but relied on qualitative data from interviews (Kyegombe et al., 2015; Sim, Puffer et al., 2014). The IPV finding for Sinovuyo Caring Families was partially based on a quantitative measure of caregiver and adolescent attitudes towards sexual and gender based violence and traditional gender roles (Cluver, Lachman et al. 2016). However, the use of scales that measure attitudes and knowledge are not reliable indicators of behaviour change and direct measures of IPV are needed. One of the studies that included a qualitative component presented the author’s interpretation of the data without including illustrative quotes to support the findings. In comparison, the quantitative data was presented in detail (Sim, Annan, et al., 2014). In the case of REAL Fathers, ethical concerns regarding use of unique identifier codes during data collection limited the ability to maintain group assignment in the RCT design. 

Some evaluations relied on men’s reports of IPV with no corroborating evidence from women (Ashburn et al., 2015; Hatcher et al., 2014; Van den Berg et al., 2013). Similarly, reports of changes in relation to harsh and abusive parenting practices were based on caregiver accounts (Ashburn et al., 2015; Kyegombe et al., 2015; Sim, Puffer et al., 2014) and only two studies included follow-up interviews with children (Cluver, Lachman et al., 2016; Cluver, Meinck et al., 2016; Sim, Annan, et al., 2014). Two of the trials and the pre-post pilot study had short follow-up periods (Cluver, Lachman et al., 2016; Cluver, Meinck et al., 2016; Sim, Puffer et al., 2014; Sim, Annan, et al., 2014). Longitudinal research using mixed methods designs is needed to strengthen the evidence in LMIC regarding sustainability of programme outcomes. Our findings and recommendations are partially informed by studies that used qualitative interviews with participants to explore their perceptions of how the programmes changed their behaviour in regard to IPV and CM. Given the highly sensitive and stigmatised nature of these issues, consideration must be given to the presence of social desirability bias which may have influenced participant disclosures of change following their involvement in the programmes.

Future evaluations of programmes that are likely to impact on IPV and CM should include the use of validated measures with men and women which can be triangulated with data from qualitative data sources. Furthermore, this should be corroborated with children’s reports of the nature, frequency and severity of IPV and CM whenever possible. Evidence suggests there are low to moderate associations between parent and child reports of the child’s exposure to IPV (Kolko, Kazdin, & Day, 1996; Litrownik, Newton, Hunter, English, & Everson, 2003). Children may be aware of abuse which parents believe they are shielded from, and parents and children may be aware of, attend to and remember different aspects of hostile interactions leading to differences in their accounts (Grych, Jouriles, Swank, McDonald, & Norwood, 2000; Jouriles et al., 2001). Ethical dilemmas involving children in research relate to conflicting constructions of children as both competent and vulnerable. This necessitates further discussion of the methods that may be used to capture children’s exposure to and descriptions of violence (Cater & Overlien, 2013).

Although the scoping team and other experts were involved at various stages of the review process, an important limitation is that screening of abstracts and data extraction was conducted by one reviewer. It is recommended that at least two reviewers independently review abstracts and extract data to temper biases related to interpretation of the findings (Levac et al., 2010). However, we aimed to limit potential bias by ensuring input from the scoping review team during drafting of the paper, in addition to an independent review of the paper by four international experts, and presentation of the findings at the Know Violence in Childhood expert meeting.

Conclusion

Our review has highlighted that opportunities do exist for greater coherence between IPV and CM programmes, especially in community-based programmes targeting parents. Researchers and programmers should be mindful of these opportunities whilst working towards shared goals, so that violence against women and violence against children are not addressed in isolation.

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

This paper was prepared under the Know Violence in Childhood: Global Learning Initiative (http://www.knowviolenceinchildhood.org/).

Acknowledgements

The author/s acknowledge the support and internal review provided by the Initiative and its funders.

Appendix 1. Example search strategy from medline

(1)	Domestic violence/or battered woman/or family violence/or partner violence/
(2)	gender based violence/
(3)	(spous* abuse or wife abuse or abuse of wives or abuse of women or wife battering or battering of wives or partner abuse or partner violence or family violence or battered wom*n or dating violence or violence against women or gender based violence). mp. [mp = title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier]
(4)	dating violence/
(5)	child neglect/or child abuse/or child sexual abuse/
(6)	(child* maltreatment or child* neglect or child* sex* abuse or violence against children or violence against girls). mp. [mp = title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier]
(7)	child health care/or prenatal care/or community health nursing/
(8)	postnatal care/
(9)	secondary prevention/or prevention study/or primary prevention/
(10)	early childhood intervention/or intervention study/or early intervention/
(11)	((gender adj 3 education) or gender power or prevention or primary prevention or secondary prevention or intervention* or program* or what works or response or approach or approaches or advocacy or perinatal home visit* or parenting program* or home visit* or health visiting or nurse family partnership or family nurse partnership or health sector intervention* or health service intervention or school program* or community mobilisation). mp. [mp = title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier]
(12)	1 or 2 or 3 or 4
(13)	5 or 6
(14)	7 or 8 or 9 or 10 or 11
(15)	12 and 14
(16)	13 and 14
(17)	12 and 13 and 14