Women’s Voice and Agency Research Series 2014 No.10 92713 INTERVENTIONS TO PREVENT OR REDUCE VIOLENCE AGAINST WOMEN AND GIRLS : A SYSTEMATIC REVIEW OF REVIEWS Diana J. Arango Matthew Morton Floriza Gennari Sveinung Kiplesund Mary Ellsberg This paper was commissioned by the World Bank Group to help inform the report Voice & Agency: Empowering Women and Girls for Shared Prosperity. It does not necessarily reflect the views and research of the World Bank Group. Feedback and comments are welcome at: genderandagency@worldbank.org More details about the report are available at: www.worldbank.org/gender/agency 0 Executive Summary Violence against Women and Girls (VAWG) is a pervasive global problem . It is a violation of basic human rights and a drag on development. Much of the research to-date on the topic— including a major recent World Health Organization (WHO) study to produce global prevalence rates—has focused on better understanding the scale and nature of the problem. The present study builds on this body of research while shifting focus to synthesizing global evidence on potential solutions. “Violence against Women and Girls” refers to any act of gender-based violence that results in, or is likely to result in, physical, sexual, or psychological harm or suffering to women or girls, including threats of such acts, coercion, or arbitrary deprivation of liberty, whether occurring in public or in private life. The above mentioned WHO study estimates that 35 percent of women around the world, at some point in their lives, have experienced physical and/or sexual violence by an intimate partner or sexual violence by a non-partner. While both men and women can be victims or perpetrators of violence, women are more likely to be physically assaulted or murdered by someone they know; women are also at a much greater risk of being sexually assaulted or exploited in childhood, adolescence, or adulthood. This paper, a systematic review of reviews, breaks new ground by synthesizing evidence on the effects of VAWG prevention interventions. It examines the diversity of geographical context, the types of violence addressed, and the numerous approaches that have been used to combat VAWG. Additionally the review summarizes the quality of evidence on efficacy and effectiveness in order to highlight strengths and gaps of interventions on a global scale and could serve as a point of reference for those intending to undertake future design, implementation, and evaluation of interventions. Through an extensive search, 3,710 citations were identified and 58 met all of the eligibility criteria. The 58 included reviews focused on synthesizing the effectiveness evidence of interventions aimed at reducing various forms of VAWG and were included in the review. They collectively summarized evidence on 290 tested interventions. Topics covered included child sexual abuse, harmful traditional practices, intimate partner violence, non-partner rape, sexual assault, and harassment. Twenty-one evaluations were identified that had statistically significant positive effects on reducing VAWG. The global evidence base is heavily skewed towards the Global North. Over 70 percent of the impact evaluations were conducted in just seven high income countries comprising six percent of the world's population. This skewed distribution of evidence demonstrates an urgent need for more investment in rigorous evaluations of a range of interventions across different sectors to prevent VAWG in low- and middle-income countries. Although drawn largely from high-income countries, this evidence still offers important lessons that could inform piloting and testing in low-resource settings. For instance, psychosocial support has, in some cases, decreased violence in high-income settings. Various modalities of psychosocial support are being increasingly implemented and tested in low- and middle-income settings and could be usefully applied toward those at risk of experiencing new or repeated exposure to or perpetration of violence. Lessons from the more limited evidence base in low- and middle-income country settings may also be instructive. For example, the focus on primary prevention in low- and middle-income settings is worth noting, and, despite fewer evaluations, several innovative programs with promising results were identified that resulted in a reduction of VAWG. 1 Lessons may also be learned from the included reviews that are likely applicable to most VAWG interventions. In the cases of batterer intervention programs (BIP) and sexual assault education programs, the reviews for each emphasize both poor quality of program implementation and the absence of methodological rigor in the research undertaken. An hour- long video on sexual assault prevention cannot realistically be expected to change youth attitudes or reduce date rape on a university campus. Similarly, failing to adapt a batterer intervention program to the specificities of the diverse perpetrators, even when most drop out, indicates the need for a different approach. While scarcely reported, findings related to triggers of negative effects could inform better design of interventions to prevent and respond to VAWG and to avoid unintended harm. The results underscore the importance of having evaluations that carefully measure and report both positive and negative intervention effects. Several types of interventions suggested as promising by advocacy groups, as well as by the literature, have the potential to prevent VAWG. Yet according to the reviews conducted, many have not been rigorously evaluated. Moreover, some evaluations have observed adverse effects. These include interventions meant to curb child sexual abuse by strangers and interventions that employ police officers as home visitors paired with social workers. In sum, the paper finds that knowledge of intervention impacts on VAWG prevention is growing, but is still highly limited. Nonetheless, a small but growing body of rigorously tested interventions demonstrates that preventing VAWG is possible and can achieve large effect sizes. The interventions with the most positive findings used multiple, well-integrated approaches and engaged with multiple stakeholders over time. They also addressed underlying risk factors for violence, including social norms regarding gender dynamics and the acceptability of violence. These examples point to the imperative of greatly increasing investment both in innovative programming in primary prevention, as well as in high-quality experimental and quasi-experimental evaluations to guide international efforts to end VAWG. 2 Contents Acknowledgements ....................................................................................................................................... 5 Acronyms ...................................................................................................................................................... 6 Background ................................................................................................................................................... 7 Introduction ................................................................................................................................................... 7 Methods......................................................................................................................................................... 9 Eligibility ............................................................................................................................................ 10 Search Strategy ................................................................................................................................... 12 Analysis............................................................................................................................................... 13 Results ......................................................................................................................................................... 14 Characteristics of Reviews .................................................................................................................. 14 Assessing Quality of Reviews............................................................................................................. 15 Characteristics of Impact Evaluations Identified from Reviews ......................................................... 17 Geographic Spread of Evidence .......................................................................................................... 19 Types of Violence Studied .................................................................................................................. 21 Duration, Frequency and Target Population ....................................................................................... 22 Types of Interventions ........................................................................................................................ 23 Summary of Findings by Type of Violence and Intervention Strategy .............................................. 24 Main Findings from Reviews by Type of Violence .................................................................................... 27 Child Sexual Abuse................................................................................................................................. 27 Harmful Traditional Practices ................................................................................................................. 27 Intimate Partner Violence ....................................................................................................................... 29 Primary Prevention Interventions ....................................................................................................... 30 Secondary Prevention Interventions ................................................................................................... 32 Non-Partner Sexual Abuse ...................................................................................................................... 35 Discussion ................................................................................................................................................... 37 Lessons Learned...................................................................................................................................... 37 Strengths and Limitations ....................................................................................................................... 38 Implications for Research ....................................................................................................................... 39 Annexes....................................................................................................................................................... 41 Annex A - Full Search Strategy Search Terms ....................................................................................... 41 Annex B - Websites Consulted for Systematic or Comprehensive Rreviews......................................... 41 3 Annex C - Data Extraction ...................................................................................................................... 42 Annex D – All Included Reviews ........................................................................................................... 44 Annex E – Lessons from World Bank Group Impact Evaluations ......................................................... 51 Bibliography ............................................................................................................................................... 53 4 Acknowledgements The team would like to acknowledge the advice and support of World Bank Group colleagues who peer reviewed this document: Markus Goldstein, Jennifer McCleary-Sills, Elizaveta Perova, and Alys Wilman. The team appreciates the work of Manuel Contreras, Amber Hill, Isabella Mansilla, and Stephanie Perlson, from the Global Women’s Institute (GWI) at George Washington University, for their help in data cleaning and report editing. GWI would also like to acknowledge the support of the Australian Department of Foreign Affairs and Trade. We are also very grateful to Sarah Jackson-Han and GSDTI’s English Editing service for their assistance in making this paper more accessible and user friendly for a broad audience. 5 Acronyms AFR Africa AMSTARA Measurement Tool to Assess Systematic Reviews ASSIA Applied Social Sciences Index and Abstracts BIP Batterer Intervention Program CBO Community-Based Organization CBT Cognitive Behavioral Therapy CINAHL Cumulative Index to Nursing and Allied Health Literature EAP East Asia and the Pacific ECA Eastern Europe and Central Asia ERIC Education Resources Information Center FBO Faith-Based Organization FGM/C Female Genital Mutilation/Cutting GBV Gender-Based Violence GWI Global Women’s Institute HARK Humiliation, Afraid, Rape, Kick HITS Hurt, Insult, Threaten, Scream HTP Harmful Traditional Practices HV Home Visiting IE Impact Evaluation IPV Intimate Partner Violence LAC Latin America and the Caribbean LSHTM London School of Hygiene and Tropical Medicine MCH Maternal and Child Health MENA Middle East North Africa NGO Non-Governmental Organization OECD Organization for Economic Co-operation and Development OVAT Ongoing Violence Assessment Tool PTSD Post-Traumatic Stress Disorder PVS Partner Violence Screen RCT Randomized Control Trial SAR South Asia Region SASA! Start Awareness Support Action UNICEF United Nations Children’s Fund USPSTF United States Preventive Services Task Force VAWG Violence Against Women and Girls WAST Women Abuse Screening Tool WBG The World Bank Group WHO The World Health Organization WDR World Development Report 6 Background The 2012 World Development Report (WDR) on Gender Equality and Development (World Bank, 2011) identified women’s voice, agency, and participation as key dimensions of gender equality, alongside endowments and opportunities. The WDR 2012 goes on to highlight gender equality as a major policy priority for the World Bank Group. The report recognizes freedom from the risk of violence among the key aspects of ensuring that women and girls have the ability to make meaningful choices in their lives and to act on those choices (World Bank, 2011, p. 150). It further acknowledges that gains in women’s agency improve not only their lives, but also their children’s future and welfare, as well as offering broader development objectives. As a background paper to the World Bank Group’s report, Women’s Voice & Agency: Empowering Women and Girls for Shared Prosperity (Klugman et al., 2014), the Global Women’s Institute (GWI) at the George Washington University, jointly with the World Bank Group (WBG), undertook a systematic review of reviews to gather evidence on what is known about the impact of interventions to reduce and prevent violence against women and girls (VAWG). Introduction Violence against women and girls (VAWG)—also referred to as violence against women, gender- based violence, or sexual and gender-based violence—is a widespread and pervasive infringement on human rights and well-being that has no social or economic boundaries. According to the United Nations Declaration on the Elimination of Violence Against Women (1993), VAWG refers to any act of gender-based violence that results in, or is likely to result in, physical, sexual, or psychological harm or suffering to women (including threats of such acts), or coercion or arbitrary deprivation of liberty, whether occurring in public or in private life. Violence against women and girls includes, but is not limited to, physical violence, such as slapping, kicking, hitting, or the use weapons; emotional abuse, such as systematic humiliation, controlling behavior, degrading treatment, insults, and threats; sexual violence, which includes any form of non-consensual sexual contact—female genital mutilation/cutting (FGM/C) is an act of violence that impacts sexual organs and as such is included under this category of violence; forced marriage,1 which is the marriage of an individual against her or his will; and denial of resources, services, and opportunities, also known as economic abuse, such as restricting access to financial, health, educational, or other resources with the purpose of controlling or subjugating a person.2 1 Early/child marriage is defined by the age of the survivor at the time the incident of forced marriage took place. Any forced marriage that occurred before the age of 18 is considered an early/child marriage. The definition of who is a child is taken from the UN Human Rights: The Convention on the Rights of the Child. 2 For the purposes of this review and analysis, the types of VAWG were classified into 5 categories: (i) Intimate Partner Violence; (ii) Non-partner Sexual Abuse; (iii) Harmful Traditional Practices; (iv) Human Trafficking; and (v) Child Sexual Abuse(see diagram Annex C). 7 A World Health Organization (WHO) report on global and regional prevalence of violence against women, released in June 2013, estimates that 35 percent of women around the world have experienced physical or sexual violence at the hands of an intimate partner, or sexual violence perpetrated by a non-partner, at some point in their lives (World Health Organization, 2013). The report, which aggregates data on the victimization or perpetration of VAWG from over 80 countries around the world, calls for a multisectoral response to eliminate tolerance for violence, increased investment in prevention efforts, and strengthened services for survivors. Epidemiological research has demonstrated that VAWG is a major cause of ill health among women and girls (Garcia-Moreno, Jansen, Ellsberg, Heise, & Watts, 2006; Ellsberg, Bradley, Egan, & Haddad, 2008). Its impact can be seen in death and disability caused by injuries, increased vulnerability to contracting sexually transmitted infections, increased physical and mental illness, and increased alcohol use. VAWG may also result in unwanted pregnancy and abortions and low birth weight among infants (World Health Organization, 2013). As the WDR 2012 underscored, violence and the fear of violence severely limits women’s agency and affects their potential contributions to social and economic development. Both men and women can be victims or perpetrators of violence, but the characteristics of violence commonly committed against women and men differ. Women are more likely to be physically assaulted or murdered by someone they know. WHO reports that 38 percent of all murders of women globally are reportedly committed by a previous or current intimate partner (Stockl et al., 2013). Women are also at a much greater risk of being sexually assaulted or exploited in childhood, adolescence, or adulthood. In contrast, men are more likely to be assaulted or murdered by unknown assailants. Men are the most common perpetrators of violence against men as well as against women (Fulu, Jewkes, Roselli, & Garcia-Moreno, 2013). During the last two decades, there has been increased focus on developing and implementing interventions to address VAWG around the world. Drawing on evidence regarding risk and protective factors contributing to violence or to the lack thereof, existing interventions have used approaches ranging from community mobilization efforts aimed at changing norms that support VAWG to improving the economic opportunities available to women through micro-credit programs. Through these programmatic efforts, many “promising approaches” for violence prevention have been identified. Yet knowledge of what works to prevent violence has been limited by several factors: a poor overall understanding of which contributing factors are amenable to change and can lead to significant reductions in violence; an overemphasis on single- factor solutions; limited consistency, rigor, and quality of evaluation approaches, measures, and methodologies; and a lack of experimental and quasi-experimental evaluations in research, monitoring, and evaluation efforts (Bott, Morrison, & Ellsberg, 2005). 8 Despite the scarcity of empirical evidence, some interventions evaluated using experimental and quasi-experimental study designs have emerged showing significant positive effects in reducing or preventing violence against women and girls.3 The current paper presents the results of a systematic review of reviews of evidence on reducing the victimization or perpetration of VAWG. The review examines the geographical and topical distribution of evidence, as well as the quality of evidence on efficacy and effectiveness, in order to highlight strengths and gaps at the global scale. Secondarily, the paper reviews findings on impacts of interventions on changing norms and attitudes that underlie VAWG when available from eligible reviews. The study also incorporates lessons from a small number of World Bank Group impact evaluations that have measured VAWG outcomes (see Appendix E). One of the strengths of the review is that it covers a wide range of types of VAWG, from violence occurring in the context of conflict and intimate partner relationships, to female genital mutilation and child marriage. A systematic review of reviews synthesizing evidence from all reviews focusing on the reported effects of prevention interventions aimed at reducing violence against women has not been previously conducted. This study may serve as a point of reference for those intending to move forward with the development, implementation, and evaluation of interventions, as well as with more specific systematic reviews, to fill gaps on the subject matter. This review is especially timely because it complements the recent WHO prevalence study by providing an analysis of the evidence base on various intervention approaches designed to prevent this global public health concern. To the extent feasible, this review seeks to present operational recommendations from the available international evidence in order to enable the World Bank Group and other multilateral, bilateral, government, and non-governmental institutions to inform their decision making when it comes to investing in interventions to prevent and reduce VAWG. At the same time, this review of reviews is intentionally broad and synthetic, seeking to distill common lessons from the wider evidence base. More specific analyses of programmatic approaches and evidence for specific interventions can be found in the individual reviews and evaluations summarized by this study. This systematic review complements a paper published in The Lancet on interventions to reduce the prevalence and incidence of violence against women and girls (Ellsberg et al., 2014). Methods This study was conducted in accordance with the guidelines provided in the Cochrane Handbook for Systematic Reviews of Interventions (Green & Higgins, 2009), to the extent that it applied to a review of reviews. The Cochrane methodology was developed for the purpose of conducting preplanned and transparent reviews of the evidence on a particular intervention question. When conducted properly, the approach provides an exhaustive and unbiased synthesis of the evidence base with respect to given interventions for pre-specified populations, outcomes, and research 3 For the remainder of the document, the word “positive” refers to the intended direction of the outco me, rather than the statistical direction. 9 designs. When adequate data are identified, systematic reviews also conduct statistical meta- analyses according to a pre-specified analysis plan. Increasingly, however, as multiple evidence reviews of specific intervention questions emerge for topic areas of interest, the systematic search, selection, and—to some extent—analytical methodology is increasingly applied to reviews of reviews. Accordingly, the present review applies a systematic review of reviews approach to the evidence base on any and all interventions to prevent or reduce VAWG (for example health sector, community-based, school, public awareness, infrastructural, criminal justice, or economic empowerment interventions) for which reviews exist. The intent is to summarize the evidence of intervention effectiveness across a broad range of sector entry points. Key characteristics of reviews, and the experimental and quasi-experimental impact evaluations they contain, were extracted. The quality of the reviews themselves were assessed using A Measurement Tool to Assess Systematic Reviews (AMSTAR), a check list tool for appraising systematic review quality which has been used as well in other reviews of reviews, (for example, Butchart & Mikton, 2009; Shea, Andersson, & Henry, 2009). AMSTAR asks questions, for example, related to the reporting of a predetermined review protocol, the thoroughness and transparency of the search strategy, the assessment and use of study quality in analyzing and summarizing results, and reporting any potential conflicts of interest. The review protocol was submitted and approved for registry in PROSPERO, the international prospective register of systematic reviews at the Centre for Reviews and Dissemination, University of York, UK.4 The authors chose the systematic review of reviews approach, but recognized its limitations. For example, a pre-specified review process can result in greater selectivity and more restrictions in terms of what can be included and analyzed. Although the methodology limits the kind of information that will be gathered, it also enables the creation of an authoritative state of the evidence on a given research question. There are various systematic reviews summarizing individually evaluated interventions. In aggregate, these comprise a wide- ranging mesh of information that can leave readers and decision makers with disparate and sometimes contradictory information. A review of reviews helps to synthesize and distill that information in order to provide readers with a more unified and thorough understanding of the evidence on various interventions for preventing and reducing VAWG. Eligibility To be eligible for inclusion in this study, reviews had to synthesize evidence on the impacts of interventions that aimed to reduce the victimization or perpetration of selected forms of VAWG. Reviews had to have been either systematic or comprehensive and to have been completed between January 2000 and April 30, 2013. This time period was chosen in view of previous evidence summaries, such as the In-depth Study on all forms of violence against women: a report to the Secretary General of the UN 2006, and the Population Reports: Ending Violence Against Women written by Lori Heise, Mary Ellsberg (co-author of this review) and Megan Gottemoeller in 1999. 4 Registry No. CRD42013004422. Available at http://www.crd.york.ac.uk/PROSPERO/ 10 As defined by the Cochrane Handbook (Green & Higgins, 2009), a systematic review attempts to identify, appraise, and synthesize all the empirical evidence that meets pre-specified eligibility criteria to answer a given research question. Systematic reviews that identified no includable evidence (that is “empty reviews”) were eligible for inclusion. This is because the finding of a lack of rigorous evidence on a particular intervention question can itself help prompt future research priorities, and in some cases, can challenge commonly used approaches that have little grounding in reliable evidence (Schlosser & Sigafoos, 2009; Yaffe et al., 2012).5 Reviews that did not meet the aforementioned tenets of a systematic review, but which did seek to review and describe evidence on the impacts of interventions to reduce VAWG, are described in this study as “comprehensive reviews.” To be eligible, comprehensive reviews must have indicated that a primary objective of its study is to review the evidence of the impacts of interventions designed to prevent or reduce VAWG, and also must have included empirical results from two or more impact evaluations. Impact evaluations from reviews were eligible if they included experimental designs or quasi-experimental designs with well-defined comparison groups. All reviews must have included this review of review’s primary outcome–victimization or perpetration of VAWG–for eligibility. As a secondary outcome, however, this study also reviewed results described in eligible reviews that dealt with changes in attitudes and social norms that regulate the acceptability of VAWG. Box 1: Outcomes of Interventions Reviewed Primary outcome(s) Secondary outcome(s) Victimization or perpetration of violence Change in attitudes and social norms that against women and girls. All eligible reviews regulate the acceptability of violence against must specifically aim to synthesize evidence women and girls. This secondary outcome of the impacts of an intervention type or category can include, for example, measures multiple interventions types on this outcome. of attitudes that condone violence against The outcome can include both victimization women and girls in general or under specific and perpetration of VAWG, and the circumstances, perceptions of fault for certain interventions may focus on either primary or types of violence or attitudes about secondary prevention. bystanders intervening in violence against women and girls. Eligible reviews summarize the evidence for one or more of the following types of violence: intimate partner violence; rape or sexual assault, including sexual violence in conflict settings; child sexual abuse; sexual harassment; female genital mutilation/cutting; forced/child marriage and other harmful traditional practices; psychological/emotional abuse; physical assault; trafficking; or other similar activities. For analytical purposes, after categorizing all reviews in detail, the authors then reclassified the VAWG types into six broad categories: Intimate Partner 5 Comprehensive reviews, which did not include eligible trials, were not included in this review of reviews because unlike systematic reviews, comprehensive reviews do not include a systematic and prospective design that can sufficiently summarize the state of the evidence on a particular research question— including when the state of the evidence is that there is none. 11 Violence (IPV), Non-Partner Sexual Abuse (NPSA), Harmful Traditional Practices (HTP), Human Trafficking (HT), and Child Sexual Abuse (CSA). Figure 1. Classification of VAWG Categories6 Non- Intimate Harmful Child Partner Human Partner Traditional Sexual Sexual Trafficking Violence Practices Abuse Abuse Any incident of violence Violence perpetrated in the experienced by Sexual violence Any type of name of social, someone who is committed by a Rape or sexual violence cultural or recruited or non-intimate assault where committed by a religious values. harbored by partners survivor is less current or former Can include another through (irrespective of than 18 years old partner female genital use of force for age of survivor) mutilation/cutting the purposes of and child exploitation. marriage. The study excluded reviews that focus on child maltreatment by a caregiver (meaning in this review a parent or guardian in the household), including incidents of violence by such caregivers against girls. Child maltreatment can include some forms of gender-based violence against girls and is also an important risk factor for future exposure to and perpetration of VAWG (Renner & Slack, 2006). Child maltreatment by a caregiver, however, involves a unique set of programmatic considerations (for example, related to parenting interventions), and the topic is adequately and recently addressed by other reviews, and reviews of reviews, leaving little value that could be added by including it in the present study (for example Knerr, Gardner, & Cluver, 2011; Butchart & Mikton, 2009). Search Strategy The literature for the review was identified by implementing a preplanned search strategy in relevant databases and supplemental sources, including outreach to over 60 experts in the field of VAWG. Key characteristics of reviews and experimental and quasi-experimental impact evaluations were extracted. The search was conducted in English using Psychinfo, Embase, Medline, Applied Social Sciences Index and Abstracts (ASSIA), Social Services Abstracts, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Education Resources Information Center (ERIC), and the Cochrane 6 The classification tool used is based on the incident classification system created by the Gender-based Violence Information Management System team: www.gbvims.org 12 Database of Systematic Reviews. A full list of search terms and parameters is provided in Appendix A. Additionally, the team compiled reference lists of review articles and consulted with a group of experts in the field of VAWG. Grey literature was identified by searching key Websites (also outlined in Appendix A), and conducting outreach to a wide range of organizations and individuals known for producing or disseminating relevant research on this topic. The process followed for data extraction can be found in Annex C. If the team was unable to find the relevant characteristics and results of an evaluation from the systematic reviews, the authors then reviewed any available original articles to obtain the missing information. For the evaluations that showed statistically significant positive results, two researchers were responsible for verifying these results by reading each study and obtaining effect sizes and any additional information on its intervention and evaluation design. During this process, articles were removed either because the design did not fit the inclusion criteria; because the results were not appropriately disaggregated by sex or age (where in the case of child marriage, for example, the outcome could not be attributable to reduction in child marriage because the age of marriage was not clear); or because the findings on the primary outcome were not evaluated using rigorous statistical methods. Analysis Because of the extent of intervention and outcome-measure heterogeneity that is inherent in a systematic review of reviews, the authors determined that statistical meta-analysis (pooling of data) would be inappropriate and would lack credibility. Descriptive statistics are provided to help summarize the nature and scope of the evidence base, both at the level of the reviews themselves and at the level of evaluations they contain. This exercise involved identifying the numbers and characteristics of included evaluations that reported statistically significant positive and negative—as well as null—results on the outcomes of interest. This is to present a general description of findings from the individual evaluations. However, these descriptive statistics should be very cautiously used, if at all, to draw inferences about the effectiveness or ineffectiveness of any particular intervention. The fragility of statistically significant results of trials showing only modest changes to participant experiences is well-established (Walsh et al., 2014), indicating that a wider array of statistics and information should be used to form strong conclusions about the effectiveness of any particular intervention. Furthermore, simply “counting” statistically positive or negative effects by intervention types for the purpose of summarizing them as effective or not would necessarily ignore important details about interventions and trial designs, sample sizes, follow-up periods, effect sizes, and cultural context, among others, and therefore could lead to misleading conclusions. Because specific evidence reviews themselves are better positioned for characterization of the evidence based on specific interventions, the approach of this review is to provide a narrative analysis that summarizes and critically appraises findings from the range of reviews. 13 Results Characteristics of Reviews A total of 3,687 citations were retrieved from electronic databases, and 23 more were retrieved from institutional, Web-based databases and expert outreach, resulting in a total of 3,710 citations. One hundred and forty-seven reviews were retrieved for full-text inspection by two reviewers using a predetermined screening guide. The team identified 58 eligible reviews7 focused on synthesizing the evidence of interventions aimed at reducing various forms of VAWG. Of the 58 included reviews, 23 were designated as systematic and 35 as comprehensive reviews as per the above outlined definitions. Figure 2 includes a flow diagram of the review’s search and selection process. Figure 2. Flow Diagram for Selecting Systematic and Comprehensive Reviews 3,687 reviews identified 3,540 reviews were excluded after abstract review 147 full-text articles read to further assess for inclusion 109 reviews were excluded as they did not meet the inclusion criteria 23 eligible reviews were identified through gray literature 3 reviews were removed searches, backward references as they were only and contact with experts in the summarized or identical field to studies already identified 58 reviews included  23 Systematic Reviews  35 Comprehensive Reviews 7 See Annex D for a complete list of all included systematic and comprehensive reviews with complete citation. 14 Figure 3. Number of Reviews by Type of Violence and Type of Review 40 35 Number of Reviews 30 15 25 20 15 3 10 19 5 12 2 1 1 3 0 1 1 0 Type of Violence Reviewed CR SR Figure 3 presents the distribution of reviews according to the types of VAWG addressed. A majority (n=34) of the reviews assessed interventions aimed at reducing IPV. Fifteen of the eligible reviews focused on evidence related to reduction of non-partner sexual abuse. Five reviews summarized findings related to several types of violence faced by women and girls. Two reviews related to harmful traditional practices (HTP), one on FGM/C and the other on child marriage. Only one review dealing with child sexual abuse met the inclusion criteria. One review focused on trafficking. Assessing Quality of Reviews Table 1 describes the reviews according to the assessed quality (based on AMSTAR ratings), the type of violence, and the number of impact evaluations that were extracted for the review of reviews. Because the AMSTAR tool was designed to assess the methodological quality of systematic reviews, it was not applied to the 35 reviews designated in this review as comprehensive reviews. Lacking a predetermined systematic methodology, comprehensive reviews should generally be viewed as inherently more susceptible to bias in both the identification and analysis of evidence than would systematic reviews. However, the decision to conduct a comprehensive rather than a systematic review may arguably be justified in some cases—for example, when the authors of a particular review determine that gains in breadth or flexibility outweigh potential costs in terms of bias, in view of the desired purpose or audience of a particular review. Only one review was considered to be of “low quality” (that is an AMSTAR score between 0-4), whereas 10 were of “moderate quality” (5-8) and 12 were of “high quality” 15 (9-11).8 The mean AMSTAR score for the 23 systematic reviews included in this study was 8 (median=9), with an average of 12 (median=7) eligible impact evaluations per review. Overall, the AMSTAR scores illustrate substantial heterogeneity in the quality of adherence to established systematic review standards. While most fall into at least the moderate-quality range or above, the results imply that additional caution is needed in drawing and interpreting conclusions from many of these reviews because there may have been greater scope for bias in their research. Moderate- and especially low-quality ratings imply that these reviews diverged from several common methodological standards expected in high-quality systematic reviews and/or inadequately reported their analysis in reference to these standards. Because the objective of a systematic review is to provide a minimally biased synthesis of the state of the evidence on a particular research question, lower AMSTAR ratings raise additional questions regarding a particular systematic review’s strength in fulfilling this objective. Table 1. Quality of Reviews on VAWG Interventions found through Systematic Reviews of Reviews Systematic Reviews AMSTAR° Type of Number of IEs Score Violence eligible for extraction Anderson 2005 3 Non-Partner 69 Sexual Abuse Ashman 2004 11 Multiple Types 0 Berg 2012 10 HTP 8 Bilukha 2005 5 IPV 1 Coulthard 2010 9 IPV 0 Davis 2008 7 IPV 9 Feder 2008 10 IPV 10 Jahanfar 2013 10 IPV 9 Kataoaka 2004 6 IPV 6 Morrison 2004 7 Non-Partner 57 Sexual Abuse Nelson 2012 6 IPV 7 O’Reilly 2010 5 IPV 4 Ramsay 2005 9 IPV 13 Ramsay 2002 7 IPV 2 Ramsay 2009 11 IPV 9 8 The authors systematically deviated from the AMSTAR tool’s guidance in one respect. In the event of “empty reviews” (that is systematic reviews in which no studies met the inclusion criteria), the current AMSTAR phrasing of guidance for multiple items suggests that the review should be rated as “not applicable” on these items, which would automatically trigger a point reduction in scoring for each item (only “yes” answers award a point). In turn, this would automatically demote “empty reviews” to the moderate-quality range. Because the authors of this review believe that perfectly well-conducted systematic reviews can result in no includable trials, and therefore “empty” and “quality” are not necessarily correlated, the team opted to rate “yes” on any item for which the only reason for rating a particular review as “not applicable” was due to lack of included studies. AM STAR authors were notified of this deviation. 16 Ricardo 2011 6 Multiple Types 24 Smedslund 2007 10 IPV 6 Spangaro 2013 9 Non-Partner 3 Sexual Abuse Taft 2013 11 IPV 2 Van Der Laan 2011 10 Trafficking 0 Wathen 2003 5 IPV 7 Zwi 2007 11 Child Sexual 15 Abuse Whitaker 2006 5 IPV 11 Mean (SD) 8 (2.47) 12 (17.18) ° AMSTAR is a recognized tool for the assessment of systematic reviews (Shea et al. 2007). The maximum score on AMSTAR is 11, and scores of 0-4 indicate that the review is of low quality; 5-8, of moderate quality; and 9-11, of high quality Characteristics of Impact Evaluations Identified from Reviews From the 58 included reviews, a subset of data was extracted for individual evaluations that were focused on preventing VAWG. Two-hundred ninety non-duplicated citations of tested interventions were identified. Among these 290 intervention evaluations, 149 identified reduction of VAWG as a primary outcome. Among these intervention evaluations, 98 evaluations used experimental or quasi-experimental designs to test the intervention’s effectiveness, of which 84 evaluations provided information on the effectiveness of the intervention. Among these, 21 evaluations were identified that had significant positive results on reducing VAWG. Two researchers reviewed each of the evaluations and agreed upon the final classification. Figure 4 presents the flow diagram explaining the individual evaluation extraction process. The analysis presented below includes a summary of the characteristics and results of the 84 experimental and quasi-experimental evaluations for which data are available. 17 Figure 4. Flow Diagram Extracting Intervention Studies Derived from Included Systematic Reviews 58 reviews included: - 23 systematic reviews - 35 comprehensive reviews 290 individual impact evaluations identified from reviews that addressed the primary or secondary outcome 141 only reported on secondary outcomes of changing attitudes and norms around VAWG 149 measured the primary outcome of reduction of VAWG (change in behavior) 51 impact evaluations did not use experimental or quasi-experimental design 98 impact evaluations using experimental or quasi-experimental design 14 impact evaluations found in systematic reviews did not provide enough information on the effect of the tested intervention 84 impact evaluations with outcome data - 2 impact evaluations with null and available negative results - 59 with null results - 2 with significantly negative results 21 with one or more statistically significant outcomes of reducing VAWG 18 Geographic Spread of Evidence The review of reviews found a significant amount of geographic concentration in the evidence base for VAWG interventions. As shown in Figure 5, the vast majority (77 percent) of experimental and quasi-experimental evaluations identified were conducted in North America. Eleven percent of evaluations were conducted in Africa, and 7 percent were conducted in the South Asia region. Two percent or less of all extracted evaluations were conducted in the Latin America and Caribbean region, the Middle East and North Africa region, or the East Asia and Pacific region. No intervention evaluations meeting the inclusion criteria were identified for the Eastern Europe and the Central Asia or East Asia region. The vast majority of the evaluations identified through this systematic review were conducted in high-income countries. In other words, over 70 percent of the evidence found using the review’s methodology on what does and does not work to prevent or reduce VAWG is derived from seven high income countries9 (Australia, Canada, Denmark, Hong Kong, New Zealand, United Kingdom and the United States) comprising 6 percent of the world's population. Sixty-six percent of experimental or quasi-experimental evaluations found were carried out in the United States alone. Notably, the three regions with the highest reported rates of IPV (South Asia, Middle East and North Africa, and Sub-Saharan Africa) according to recent WHO estimates (WHO, 2013) were settings for less than one-fifth of the experimental or quasi-experimental trials measuring IPV victimization or perpetration as an outcome. In other words, while experimental and quasi-experimental interventions to address IPV are needed everywhere, the current state of evidence on the topic is not only inadequate overall, but it is especially lacking in the regions that need it the most. 9 Countries classified according to the World Bank Group’s Country and Lending grouping. http://data.worldbank.org/about/country-and-lending-groups accessed July 7, 2014 19 Figure 5. Number of Eligible Impact Evaluations According to Region — -   - —  —  —   — —     - 20 Types of Violence Studied Table 2 presents a summary of the characteristics of the impact evaluations according to the type of violence addressed. IPV was the most commonly studied form of violence, with 69 percent (n=58) of the experimental and quasi-experimental evaluations identified in the review of reviews having evaluated programs meant to reduce or prevent IPV. Non-partner sexual abuse, which includes rape or sexual assault perpetrated by a non-intimate partner, comprises 20percent (n=17) of included intervention evaluations. Evaluations reporting effects related to change in harmful traditional practices (including FGM/C and child marriage) account for 11 percent (n=9) of all evaluations. No impact evaluations meeting the inclusion criteria were identified for child sexual abuse or trafficking. Table 2. Describing the Evidence Base: Characteristics of Experimental and Quasi-Experimental Impact Evaluations Experimental or Quasi- Intimate Non- Experimental evaluations Harmful Partner Partner of interventions to reduce Traditional Violence Abuse Total or prevent VAWG Practices n=58 n= 17 n =9 (11%) (69%) (20%) N= 84 (100%) Impact Evaluation (IE) Design Randomized Control 3 43 13 59 (70%) Trials (RCTs) Quasi-Experimental 6 15 4 25 (30%) Sample Size (Average size = 600, Median 342) Less than 100 0 11 4 15 (18%) 101-299 0 15 4 19 (23%) 300 or more 2 30 7 39 (46%) No Data 7 2 2 11 (13%) Age During Infancy, Childhood, And Early Adolescence 1 1 1 3 (4%) During Adolescence And Early Adulthood 4 6 13 23 (27%) During Adulthood 0 33 0 33 (39%) All Life Stages 2 18 3 23 (27%) 21 No Data 2 0 0 2 (2%) Target Population Male focused 0 20 4 24 (29%) Female focused 3 30 10 43 (51%) Both men and women 6 8 3 17 (20%) Type of Intervention Primary Prevention 9 10 14 33 (39%) Secondary Prevention 0 48 3 51 (61%) Duration One event 0 14 10 24 (28%) Less than 1 month 0 2 1 3 (4%) 1-6 months 0 24 0 24 (28%) More than 6 months 3 11 0 14 (17%) No Data 6 7 6 19 (23%) Geographic Location Low and Middle Income 9 6 2 17 (20%) High Income 0 52 15 67 (80%) Duration, Frequency and Target Population Nearly one-third of interventions studied (n=27) had a duration of less than one month, and most of these were single events. Twenty-four interventions lasted between 1-6 months, and 14 interventions lasted more than six months. The mean frequency or dosage of the interventions was high, at least 10 hours long, occurring over several weeks or month. The longest intervention was a home visitation program that occurred over a three-year period (Duggan et al., 1999). Most evaluations (n=33) targeted adults. Twenty-three interventions targeted adolescents and young adults, or individuals in any age group. 22 Types of Interventions Box 2: Typology of Intervention Advocacy Activities that improve general awareness among communities on issues related to VAWG. Interventions inform survivors and the general public of their rights, and the services available to them, and improve knowledge of the different forms, risk factors, and consequences of violence Group training Programs that use training to improve awareness, knowledge, and/or skills related to VAWG among a group of individuals (for example students, women, men, adolescents, and so forth) Livelihood Programs that include activities to help generate income, such as skills trainings, business development, micro- financing, apprenticeship programs, and/or programs related to food, agriculture, and livestock (Women’s Refugee Commission, 2014) Psychosocial support Group or individual counseling that provide survivors of violence with emotional, psychological, and social support Batterer interventions Programs that focus reducing recidivism among perpetrators of violence by using various techniques, such as cognitive behavioral therapy, the Duluth model, and anger management sessions Home visitation Visits by nurses, community health workers, advocates, or other individuals to households. Visitation sessions can include training components in addition to monitoring of progress on desirable behavioral outcomes. Cash transfers Financial incentive programming where participants receive cash payments after meeting predetermined requirements Community mobilization Programs that work with communities as a whole through educational activities that focus on a broader spectrum of issues, such as health, literacy, and human rights, allowing community members to identify key actions that can be taken to shift norms and behavior Figure 6 below presents the intervention components that were most tested in the regions according to the reviews by type of violence. The most common type of intervention was group training, with 22 evaluations, followed by batterers’ intervention (n=18) and psychosocial s upport (n=12). Advocacy interventions were used in 10 evaluations, and police interventions or home visitations accounted for eight evaluations. The other interventions included home visitations by nurses or trained peers (n=5), livelihood interventions (n=5) and community mobilization (n=3). A cash transfer program was also tested. Some interventions used a combination of strategies (for example, psychosocial support plus home visitation, victim advocacy or training, livelihood plus group training, and so forth). To simplify the analysis, the team classified the evaluations by the primary approach described. 23 Figure 6. Experimental and Quasi-Experimental Evaluations – Effects of Intervention on Victimization or Perpetration of VAWG 1 Non-Partner advocacy 1 Abuse group training 1 12 livelihood intervention 1 psychosocial support 1 advocacy 8 batterer Intervention 18 group training 7 IPV home visitation 1 4 Secondary livelihood intervention 1 primary police advocacy/home visitation 8 psychosocial support 1 10 cash transfers 1 community mobilization HTP 3 group training 2 livelihood intervention 3 0 5 10 15 20 Summary of Findings by Type of Violence and Intervention Strategy Of all the 84 interventions that used an experimental or quasi-experimental study design and that were summarized in included reviews, 21 significantly reduced the victimization or perpetration of VAWG according to at least one measure. These 21 interventions were distributed across three types of violence in four regions and used a variety of interventions or a combination of intervention components (see Table 3). The greatest numbers of impact evaluations with significantly positive or mixed effects were those measuring change in intimate partner violence. “Mixed effects” is categorized as results obtained when one or more, but not all, outcomes used to measure reduction of violence against women, or decrease in perpetration of violence, were significantly attained. This review team also included Table 4, which summarizes the evidence base for various intervention strategies and which was first published in Ellsberg et al., 2014. The specific evidence is discussed in greater detail in the section “Main findings from reviews by type of violence” below. Some additional intervention types have been added based on wider research more recently presented. A few of the intervention categories have also been re-named in this table. Table 3. Reported Effects of Experimental and Quasi-Experimental Evaluations to Reduce VAWG VAWG Sig. Mixed- Null Mixed Sig. positive effects w/negative negative Total HTP 2 2 5 9 IPV 7 8 40 1 2 58 Non-Partner Abuse 2 14 1 17 Total 9 12 59 2 2 84 24 Table 4. Effectiveness of Intervention Strategies to Reduce VAWG, According to Current Evidence Base Evidence level Type of Intervention strategy a Example High-income Low- and middle- violence countries income countries Response to Violence against Women Women-centered programs for Psychosocial counseling, post-exposure prophylaxis and emergency IPV, NPSA Conflicting Insufficient evidence survivors* contraception as needed, risk assessment, referrals, safety planning Perpetrators programs* Interventions for men who assault their female partners IPV Conflicting Insufficient evidence One-stop crisis centers Multidisciplinary crisis centers (community or hospital based) IPV, NPSA N/A or no evidence Insufficient evidence Safe accommodations that provide short-term refuge and other Insufficient Shelters IPV evidence Insufficient evidence services Specialized police services for survivors of VAW, can include Women’s police stations IPV, NPSA N/A or no evidence Insufficient evidence psychosocial counseling and referrals Victim Advocacy* Case management, connection to legal services and information IPV Promising Insufficient evidence Insufficient ICT services National emergency hotlines or mobile applications IPV, NPSA evidence Insufficient evidence Population-based Prevention Participatory projects, community-driven development engaging IPV, NPSA, Community mobilization* N/A or no evidence Promising multiple stakeholders and addressing gender norms FGM/C, CM One-off information or media efforts, billboards, radio programs, IPV, NPSA, Awareness-raising campaigns* Ineffective Ineffective posters, television advertisements FGM/C, CM Social marketing campaigns or Long-term programs engaging social media, mobile applications, IPV, NPSA, Insufficient edutainment plus group thematic television series, posters, together with interpersonal evidence Insufficient evidence FGM/C, CM education* communication activities Group-based Training or Workshops for Prevention of Violence against Women and Girls School or community programs to improve women’s agency. Can Empowerment training for IPV, NPSA, Insufficient include other components such as safe spaces, mentoring, life skills evidence Promising women and girls* FGM/C, CM or self-defense training School programs and community workshops to promote changes in Men and boys norms Insufficient social norms and behavior that encourage VAWG and gender IPV, NPSA evidence Conflicting programming* inequality School or community workshops to promote changes in norms and Insufficient Women and men* IPV, NPSA evidence Promising behavior that encourage VAWG and gender inequality Training for girls in life skills culminating in a ceremony without Alternative rites of passage FGM/C FGM/C N/A or no evidence Insufficient evidence Economic and Livelihoods Economic empowerment and Microfinance; vocational training or job placement; cash or asset IPV, NPSA, N/A or no evidence Conflicting income supplements* transfers (for example, land reform) FGM/C, CM 25 Economic empowerment and Microfinance; vocational training or job placement; cash or asset IPV, NPSA, income supplements plus transfers (for example, land reform); plus gender equality/violence N/A or no evidence Promising FGM/C, CM gender-equality training * prevention training Retraining for traditional Microfinance or vocational training FGM/C N/A or no evidence Ineffective excisors System-wide Approaches Screening* Universal IPV screening among nurses and doctors at all visits IPV, NPSV Ineffective N/A or no evidence Home visitation and health Visits by community health workers or nurses to households IPV Promising Insufficient evidence worker outreach* Justice and law-enforcement Mobile courts, increased enforcement, second response IPV, NPSV Ineffective N/A or no evidence interventions Sensitization, identification or response training with institutional IPV, NPSA, Personnel training* personnel (for example teachers, police officers, first responders, Ineffective Ineffective FGM/C, CM health professionals) Insufficient Infrastructure and transport Improving the safety of public transport, street lighting NPSA evidence Insufficient evidence a Programs will often incorporate multiple components and overlaps reflecting more than one intervention type. * Classification based on trials including randomized controlled trials (RCTs) or quasi-experimental trials with comparison groups. Evidence classification adapted from: WHO (2010). Preventing intimate partner and sexual violence against women: Taking action and generating evidence. Geneva, Switzerland: World Health Organization (WHO). Table taken from (Ellsberg et al., 2014) 26 Main Findings from Reviews by Type of Violence Child Sexual Abuse Overview of the Evidence One systematic review (Zwi et al., 2007) on child sexual abuse, that was rated of high quality, met the inclusion criteria for the systematic review of reviews. This review had been previously included in at least one other systematic review of reviews by Mikton (2009) and was also rated of high quality by the reviewers. Interventions Group training for boys and girls: The review summarized the evidence for school-based interventions. The interventions that were included used a variety of methods, such as role-play, and videos and other multimedia tools. The review found no evidence that any of the interventions reduced the victimization or perpetration of child sexual abuse or led to greater access to services for children who had been sexually assaulted (Zwi et al., 2007). The behavioral outcome that was measured was children’s self-protective behavior when faced with a situation that could lead to abuse. This outcome was measured in two of the evaluations reviewed by Zwi (2007). Both used a simulated abduction situation. Of the 13 individual impact evaluations that the team extracted from the Zwi (2007) review, nine reported positive effects in changing knowledge and attitudes around child sexual assault. Although the review found significant improvements in knowledge of protective behaviors on the part of children, the authors strongly caution against the assumption that increased knowledge by children on what constitutes abuse will lead to changes in behavior when they are confronted with a potential situation of abuse. Three evaluations reviewed by Zwi (2007) showed some negative effects resulting from interventions, such as nightmares, increased aggressive behavior towards peers, increased dependency, fearfulness of strangers, bed-wetting, reluctance to go to school, and so forth. While these evaluations represented only a minority of evaluations, the findings reinforce the need to assess potentially harmful impacts of interventions. Zwi and her fellow authors recommend further research on the optimal age for children to receive interventions and on the best format for school-based interventions of this nature. Harmful Traditional Practices Overview of the Evidence Studying harmful traditional practices, including female genital mutilation/cutting (FGM/C) and early or forced child marriages, requires a careful understanding of the cultural context within which these practices take place. Impact evaluations from two identified reviews indicate that transforming such strong norms and practices requires careful communication and the use of community dialogue and participation, as well as the involvement of multiple sectors and community stakeholders. Broaching this topic within other initiatives, such as health literacy programs or economic empowerment activities, may heighten community involvement and openness to community dialogue about FGM/C. The reviews highlighted a paucity of evidence, especially for the prevention of FGM/C, and called attention to the need to safeguard against negative outcomes resulting from HTP interventions. 27 The two reviews included in the study were comprised of a high quality systematic review (Berg & Denison, 2012) and a comprehensive review (Lee-Rife, Malhotra, Warner, & Glinski, 2012). The first review focused on summarizing interventions designed to reduce the victimization or perpetration of FGM/C, while the second one reviewed interventions or policies that had documented measurement of change in behavior, knowledge, or attitudes related to child marriage among relevant stakeholders. In the review of FGM/C conducted by Berg and Denison (2012), the authors also highlight the overall weakness of evidence on this topic. The majority of evaluations focused on changes in attitudes towards the practice itself, or on the intention of mothers to have their daughters undergo the practice in the future. Very few evaluations measured changes in victimization or perpetration of FGM/C, or presented the data in a way that allows the estimation of effect sizes. This review specifically reiterates the need to measure the negative effects of interventions, reporting that one reviewed evaluation found that after the program was completed, fewer health personnel wished to play a role in educating clients on FGM/C (Berg & Denison, 2012, p.142). Another evaluation reviewed by Berg and Denison involving health practitioners may have unintentionally resulted in the misconception that FGM/C is acceptable as a medically safe procedure (Berg & Denison, 2012, p. 142). Given how embedded FGM/C is within cultural and traditional practice, the authors recommend taking time to understand this context and to then design culturally relevant and appropriate interventions as a critical step in creating new interventions. They suggest that impact evaluations providing inconsistent and non-significant findings could largely reflect imperfect responses to the populations’ needs. In other words, achieving community involvement and ownership can be essential for interventions that aim to change community attitudes and norms related to marriage and purity. Interventions Community mobilization: One of the most featured models for reducing FGM/C that was reviewed by Berg & Denison (2012) is the TOSTAN model. This intervention, developed in Senegal by the Non- Governmental Organization (NGO) TOSTAN, has been replicated in several countries in Sub-Saharan Africa, and utilizes community-based education programs that address a variety of issues, including health, literacy, and human rights. Through these programs, villagers identify priority issues for community action, and both FGM/C and IPV have emerged as key issues. In many cases, villagers have taken pledges to renounce FGM/C and to encourage people in neighboring villages to do the same. A quasi-experimental evaluation (n=1332) of the program in Thies, Senegal found that women in the 20 intervention villages reported significantly less violence in the last 12 months than women in the comparison villages (p<0.001), according to a post-intervention survey. Mothers of girls aged 0-10 also reported significantly less FGM/C in the intervention villages (p<0.05). It is particularly noteworthy that women in the intervention villages who were not directly involved in the TOSTAN education program also reported lower levels of violence and FGM/C, indicating successful diffusion of program impact. Group training for women and girls and cash transfers: The review conducted by Lee-Rife et al. (2012) on effective measures to prevent child marriage, concluded that the programs that work to end early marriage are designed to acknowledge and address the multiple drivers of the phenomena. Three interventions showed a statistically significant positive result according to the criteria of this review. One program in Maharashtra, India (Pande, Kurz, Walia, MacQuarrie, & Jain, 2006) and the Berhane 28 Hewan program in rural Amhara, Ethiopia(Erulkar & Muthengi, 2007; Erulkar & Muthengi, 2009) used a comprehensive set of activities, including intensive “life skills” training for unmarried girls, “community conversations,” and mentorship, and community service activities to encourage parents to keep girls in school and to delay marriage. The Berhane Hewan program also provided support to obtain basic school supplies, and an economic incentive (for example, a goat) for families whose daughters were still unmarried by the end of the program. Another program in Western Kenya used a vertical approach, including teacher training, some in-class activities, and distribution of school uniforms (Duflo, Dupas, Kremer, & Sinei, 2006). While all three programs showed some success in delaying the age of marriage by one or more years, the first two interventions showed additional benefits by addressing multiple drivers of early marriage, such as by providing increased knowledge and skills among the girls, and also by fostering changes in attitudes towards child marriage. The Kenya program shows how a relatively modest financial incentive can achieve benefits on a large scale. Conversely, it was found that stand-alone interventions, such as awareness-raising and national advocacy campaigns, combined with legislative measures, did not achieve statistically significant results in delaying child marriage (Lee-Rife et al., 2012). Intimate Partner Violence Overview of the Evidence Prevention of IPV was, by far, associated with the greatest number of both evidence reviews and included impact evaluations. Hence, this subsection includes more extensive material than is included in the material describing other types of Box 3: Defining Primary and Secondary Prevention VAWG prevention, which are Interventions structured around different common Primary prevention refers to reducing the number of new instances intervention approaches. Overall, the of violence by intervening before violence takes place. This bulk of trials in this area that took place involves fostering societies, communities, organizations, and relationships in which violence is less likely to occur in high-income countries have focused (for example by challenging attitudes, behaviors, and on two types of secondary prevention practices which justify, excuse, or condone violence). approaches—batterer interventions and While violence may in practice have occurred among some survivor services. The reviews generally of the population served, the intervention does not target find batterer interventions to lack individuals on the basis of violence that has already occurred. positive effects on VAWG repeat Secondary prevention refers to both mitigating the immediate perpetration. Women-centered survivor consequences of abuse by providing already-abused services have achieved more mixed women and girls with services and supports (for example results, but some models—especially emergency contraception, post-exposure prophylactic-PEP, psychosocial support, and counseling), and more those including intensive advocacy pertinently for this review’s primary outcome, also services and psychosocial support— preventing recurrence or repeat abuse (for example have shown positive effects in reducing through timely protection and safety for domestic violence revictimization. Furthermore, while survivors, removal of perpetrators from the household, and primary prevention approaches have orders of protection) (Fergus, L., 2012). been much less frequently studied, there are encouraging results emerging from interventions in middle and low income settings—for instance, from multisectoral community mobilization interventions. 29 A total of 34 reviews (19 comprehensive reviews and 15 systematic reviews) were identified that examined the evidence on interventions aimed at reducing or preventing IPV. The systematic reviews averaged a score of 8 on the AMSTAR scale, meaning the reviews for this topic are generally scored as of moderate quality. Altogether, 58 distinct impact evaluations meeting the study criteria were identified. Of these, 15 evaluations reported statistically significant reductions in the occurrence of IPV. In high-income countries, two main approaches were used: interventions for male perpetrators (batterer intervention programs) and women-centered services for survivors of violence. Both approaches target individuals who have either experienced or used violence in the past, with the aim of reducing either revictimization or recidivism. General findings coming from the reviews specific to IPV response interventions suggest that the intervals at which effectiveness is measured should be lengthened. Evaluations that measure results solely at the conclusion of an intervention may fail to capture fade-out effects, or results that could emerge 6 or 12 months later (Feder et al., 2008; Ramsay, Rivas, & Feder, 2005). The authors of the reviews also encourage researchers to test interventions with positive findings in different modalities; this could facilitate scale-up, particularly in resource-limited settings. It is not known, for instance, whether findings from individual psychosocial counseling trials are transferable to group therapy or couples therapy, whether the effectiveness would continue to be maintained if there were fewer or shorter sessions, or whether there would be differences between using trained non-professional staff or professionals, and so forth The reviews also highlighted the lack of evaluations containing robust designs and quality implementation: sample sizes tended to be small; sampling strategies were unclear (Feder et al., 2008; Jahanfar et al., 2013; Kataoka et al., 2004; Ramsay et al., 2005); and there is a lack of data on the cost-effectiveness of interventions (Ramsay et al., 2005). Despite the knowledge that the response to IPV should be multisectoral, there is a dearth of evaluations that measure the impact of working in a multisectoral manner (Ramsay et al., 2005). Primary Prevention Interventions High-Income Countries Primary prevention of IPV has received much less attention overall than the health and justice sector interventions described above. Only four evaluations with positive findings, all from high-income countries, were identified, including the Hawaii Health Start Program (n=643) (Bair-Merritt et al., 2010; Duggan et al., 1999), and a reproductive coercion evaluation10 in California (n=906) (Miller et al., 2011). The other two evaluations assessed group training programs on “Healthy Relationships” to reduce dating violence among adolescents in Canada (n=158 & n=1722) (Wolfe, Crooks, Lee, McIntyre-Smith, & Jaffe, 2003; Wolfe et al., 2009). Both programs, one conducted with male and female high school students, and the other, a community-based intervention with male and female at- risk youth, found significant reductions in perpetration of dating violence in the intervention arm compared to the control groups. 10 Miller et al. define reproductive coercion as spanning “both pregnancy coercion (for example male partners' verbal pressure to get women pregnant) and birth control sabotage (for example condom manipulation and other active interference with contraceptive methods) and results in women's compromised decision making regarding, or limited ability to enact, condom and other contraceptive use” (Miller et al., 2011) 30 Home visitation: Perinatal home visiting (HV) interventions have traditionally been used to reduce risks of adverse pregnancy outcomes and to improve parenting skills and infant development. Given their close contact with women, HV programs also have the potential to reduce IPV (Sharps, Campbell, Baty, Walker, & Bair-Merritt, 2008). Nonetheless, three comprehensive reviews assessed the effectiveness of perinatal and maternal and child health (MCH) home visitations and found limited evidence supporting the use of this type of intervention to prevent or reduce IPV (Bilukha et al., 2005; Evanson, 2006; Sharps et al., 2008). Despite the scarcity of evidence for the effectiveness of home visitations, home-visiting nurses have an important role to play in IPV prevention because they may be able to identify and intervene before survivors are seen in emergency rooms or clinics for the health consequences of the violence. Therefore, authors urge further research to determine and improve the effectiveness of home visitations in IPV prevention (Evanson, 2006). One home visitation intervention with positive findings is Hawaii’s Healthy Start Program (HSP) (Duggan et al., 1999), primarily designed to prevent child abuse and neglect among families considered at risk. A Randomized Control Trial (RCT) with 643 families was conducted to determine the program’s effectiveness in improving outcomes related to the mother and child’s wellbeing, including IPV during the three years of the program’s implementation and three years of follow-up. During the program implementation, mothers in the intervention group reported significantly lower rates of IPV victimization compared with mothers in the control group (incidence rate ratio [IRR], 0.86; 95% CI, 0.73-1.01) Advocacy: The evaluation of a “reproductive coercion” intervention conducted by Miller, et al. (2011), is noteworthy because it addresses a form of IPV that has not previously been studied. Reproductive coercion includes pregnancy coercion (for example male partners’ verbal pressure to get women pregnant) and active interference with conceptive methods (birth control sabotage). It has been associated with an increased risk of unintended pregnancy, HIV infection, and other sexual and reproductive health concerns. An intervention conducted in four family-planning clinics with 906 women aged 16-29 in Urban Northern California found a 71 percent decrease in the odds of pregnancy coercion among women in the intervention group who reported IPV in the past 3 months compared to participants in the control clinics. Women in the intervention arm were also more likely to report ending a relationship because it was unhealthy, or because they felt unsafe, regardless of IPV status. Low- and Middle-Income Countries In low- and middle-income countries, there is a much greater focus on primary prevention of violence. The interventions focusing on primary prevention of IPV use a wide range of approaches, including group training; social communication, such as radio and television spots, billboards, theater, and so forth; community mobilization; and livelihood strategies. Group training for women and men: Many of the interventions emerged out of HIV programming, with the growing recognition of gender inequality and IPV as a driver of HIV infection. For example, Stepping Stones, a widely adapted program, uses participatory learning approaches to build knowledge, risk awareness, and communication and relationship skills relating to gender, violence and HIV. This program, a 70-village cluster-randomized trial conducted in South Africa with young men and women aged 15-26, found that after two years following an intervention, men’s self-reported perpetration of physical and/or sexual IPV was significantly lower compared to men in the control villages (p=0.05). 31 However, no differences were found in women’s reports of IPV victimization between the intervention and control villages during the same period (Jewkes et al., 2008). Group training for men: An intervention targeting men, Yaari Dosti, was carried out in Mumbai and Gorakhpur, India (Verma et al., 2008). The program, based on “Program H,” an intervention developed in Brazil (Barker, Nascimento, Pulerwitz, & Segundo, 2006), aimed to reduce male-perpetrated VAWG by transforming gender-inequitable norms through group training and “social lifestyle marketing.” The participants (n=1015) included both married and unmarried young men between the ages of 15-29. Men in the intervention arms in Mumbai and Gorakhpur were five times and two times less likely, respectively, to report having used physical or sexual violence against a partner during the last 3 months, than participants in the comparison group (p<0.005). Livelihood programs: Another innovative program, IMAGE, used livelihood strategies to address gender, HIV and violence among rural women in South Africa (Kim, 2007). The program combined microfinance with training and skills-building sessions on preventing HIV infection, gender norms, cultural beliefs, communication, and intimate partner violence. A cluster-randomized trial found a reduction at 24 months of over 50 percent in women’s reports of physical or sexual violence from a partner in the intervention group compared to the control group (n=430). Secondary Prevention Interventions Batterer intervention programs: Two systematic reviews (Feder et al., 2008; Smedslund, Dalsbø, Steiro, Winsvold, & Clench-Aas, 2007) and one comprehensive review (Babcock, Green, & Robie, 2004) assessed the evidence on the effects of court-mandated batterer intervention programs (BIP) on IPV. The team extracted 18 experimental or quasi-experimental BIPs from these reviews, all of which took place in high-income settings. Although the authors of reviews on this type of intervention acknowledge the need for additional research, the meta-analysis conducted by Feder et al. (2008) does not provide strong support for the effectiveness of BIPs in reducing violence recidivism among perpetrators. Smedslund et al. (2007) concluded that insufficient RCTs exist to draw evidence regarding the effectiveness of cognitive behavioral therapy for batterers. The comprehensive review carried out by Babcock et al. (2004) found that effect sizes were in the small range, indicating that current interventions have a minimal impact on reducing violence recidivism. Furthermore, Feder et al. argue that more positive findings in this review are the result of the authors’ inclusion of additional quasi - experimental designs prone to selection bias. Various evaluations analyzed also reported high attrition rates for both batterers and survivors, which presents an obstacle for ascertaining either positive or negative effects from batterer intervention programs. BIPs typically involve some type of group education lasting from 8 to 24 weeks. One of the most well- known approaches is the “Duluth Model,” a feminist approach that engages men in discussions about power and control. Other approaches commonly used in the impact evaluations were cognitive behavioral therapy (CBT), and anger management, both of which seek to change violent behavior using established behavioral strategies, as well as discussing thought patterns and beliefs (Smedslund et al., 2007). A few programs tested new approaches, such as combining batterers’ treatment with substance abuse programs, or using racially and culturally adapted programs for specific groups, such as African 32 Americans. Although the literature on BIPs indicates a general decrease in recidivism among men who complete the full program, there are methodological weaknesses in the evidence base, such as the lack of a comparable control group, that make it challenging to ascertain the effectiveness of completing a BIP versus another program or no treatment at all. Most of the evaluations reviewed the histories of men who were court-mandated to participate in BIPs as a result of a domestic violence arrest and then compared recidivism (measured as either new arrests or spousal reports of violence) among men who completed the intervention with men who dropped out of the program or never attended it at all. In such evaluations, the failure to include a true comparison group weakens the study design and its ability to provide strong support for BIPs. Overall, batterers’ programs have very high dropout rates, and there are few consequences for not completing the program. An evaluation of a BIP that showed positive results (Davis, Taylor, & Maxwell, 2000) randomly assigned 376 court-mandated batterers to batterer treatment or to a treatment irrelevant to the battering problem, such as community service in New York. All men assigned to batterer treatment were mandated to attend 39 hours of class time, although some were assigned to complete the treatment in 26 weeks and others in 8 weeks. Defendants assigned to the 26-week group showed significantly lower recidivism at 6- and 12-months post-sentencing compared to defendants assigned to the control condition. However, the groups did not differ significantly at either 6 or 12 months in terms of new incidents reported by victims, suggesting that the violence may have decreased in severity, but not necessarily in quantity. Another set of evaluations involved police follow-up and support to victims, including home visitation. None of the six evaluations included in the review reported any positive findings. Despite small effect sizes, Babcock et al. (2004) urge policy makers to not dismiss the potential for BIPs to have an impact on IPV. To put these results into perspective, the authors compare BIPs with substance-abuse treatments, which, while also yielding similarly small results, can nevertheless have transformative effects on the lives of individuals undertaking such treatment. However, the impact of this transformation may not be captured in overall recidivism rates. Rather than enforcing a rigid, standardized curriculum, the authors urge policy makers to invest greater efforts in improving existing batterer intervention programs and maintaining a wide range of treatment options that can be tailored to specific individuals or groups. The potential harmful effects of batterer intervention programs on victims should also be addressed, along with consideration of the economic strain caused by charging families for such court-mandated interventions, since there have been reports of negative effects of mandated batterer intervention programs, especially when the intervention cost is not subsidized (Feder et al., 2008). Screening: A large number of screening evaluations take place in the context of health services, and involve pregnant women who are screened for violence during pre-natal care. Health care providers are uniquely positioned to identify and assist individuals in situations of violence by caring for their physical needs and referring them to shelters, counseling or legal services. Evaluations of screening programs have found statistically-significant positive results for identifying survivors of IPV, and recurrent screening throughout the pregnancy has further increased identification rates (Nelson, Bougatsos, & Blazina, 2012; Taft et al., 2013; O'Reilly, Beale, & Gillies, 2010). However, there is no evidence that screening alone increased referrals to support agencies (Nelson et al., 2012). Importantly, 33 14 evaluations included in one of the reviews concluded that screening itself was not harmful toward women (Nelson et al., 2012). The few screening evaluations that actually reported decreases in violence usually combined screening with psychosocial support or another type of survivor service. One of these, an RCT in Hong Kong, provided pregnant women (n=110) who screened positive for IPV with either a 30 minute “empowerment intervention,” 11 or put them in a control group that received the routine standard of care for abused women. Women in the intervention group reported significantly less psychological abuse and minor physical violence at six weeks post-partum (p=0.05), and significantly less post-partum depression than the control group (Tiwari et al., 2005). Significant disagreement remains regarding the ideal scope of screening interventions and regarding the impact and potential for mandatory reporting of VAWG cases to the police when VAWG is detected. The former, involving the use of universal versus targeted screening, involves issues of time and resources required for screening all patients irrespective of age and sex, while the latter raises ethical concerns regarding privacy and patient safety. In the United States, the U.S. Preventive Services Task Force (USPSTF) released a recommendation in 2013 for all clinicians to “screen women of childbearing age for intimate partner violence.” Overall, furt her research is needed to identify the most effective approaches to screening, to identify appropriate linkages to interventions, and to determine ensuing health outcomes. Survivor services: Other frequently studied interventions are “women-centered” programs that target known survivors or women newly identified through screening. These interventions use a combination of strategies, including survivor advocacy and psychosocial support to provide women with resources to reduce their future risk of violence, as well as to improve their health status. Basic psychosocial counseling may include providing danger assessments, safety planning, and referrals to specialized services. As mentioned earlier, providing screening alone has not been found to decrease IPV, although several of the screening evaluations report positive outcomes for both women and their children, such as decreased depression, lower stress, and greater knowledge and use of services. Advocacy interventions, on the other hand, include many of the same components as the psychosocial and home visitation programs. These programs provide additional support to women from a layperson trained in identifying and accessing services, often on behalf of the survivor. Sullivan and Bybee (1999) conducted an RCT to evaluate an intensive community-based advocacy intervention for 278 women leaving a battered women’s shelter in Michigan. The 10-week post-shelter intervention trained lay advocates to work one-on-one with women, helping them to access the community resources they needed to reduce their risk of future IPV. Women who worked with advocates over the course of two years experienced significantly less violence over time (p=0.03), reported higher quality of life and social support, and had less difficulty obtaining community resources. More than twice as many women receiving advocacy services experienced no violence across the two-year post-intervention period compared with women who did not receive such services. 11 The intervention consisted of advice in the area of safety, decision making and problem solving. It also included a “empathic understanding” component, derived from client -centered therapy. 34 Non-Partner Sexual Abuse - Rape, Sexual Assault and Harassment Overview of the Evidence Within the topic of non-partner sexual abuse, reviews of sexual assault prevention programs, particularly among North American university-level students, were the most numerous. The interventions focused on the primary prevention of acquaintance-perpetrated sexual assault, often termed “date rape”. Rigorous evidence on interventions for addressing workplace sexual harassment is particularly absent. While numerous impact evaluations found significant average effect sizes for improving rape-related knowledge and attitudes, only two impact evaluations reported positive findings of decreased non-partner sexual assault.12 The majority of the reviews, including twelve comprehensive reviews and three systematic reviews, were focused on interventions implemented in high-income settings. These reviews yielded 17 experimental or quasi-experimental evaluations measuring the results for the reduction in VAWG. It is worth noting that the systematic reviews averaged a score of six on the AMSTAR scale, meaning that the reviews for this topic are generally scored as moderate quality. Interventions Group training for women: Of 17 included intervention evaluations, only two reported significantly positive results in reducing non-partner sexual assault. They were both conducted in the United States among female students and consisted of university-based sexual assault prevention programs. Both interventions used training and video elements and utilized either discussion groups or role-playing activities. It is not clear to what extent these programs could be meaningfully applied to other settings or populations.13 Lessons highlighted in reviews of these interventions are that longer interventions are more likely to yield positive results than brief interventions, as are risk reducing versus empathy- focused programs (Anderson & Whiston, 2005). The same review also highlighted the scarcity of evaluations focusing on culturally and racially focused sexual assault education programs (Anderson & Whiston, 2005). Questions remain about whether programs targeting single-sex audiences result in more positive outcomes than mixed-sex audiences or a combination of both approaches, and about how results from programs using peer presenters compare to using professional presenters. The evidence seems to suggest that mixed-sex audiences are more productive, but that certain topics, such as rape myths/facts, are more likely to produce statistically significantly results when discussed in single-sex groups (Anderson & Whiston, 2005). There also seems to be more evidence pointing to professional presenters being preferred among college students. One study, (Anderson & Whiston, 2005), found that peer presenters are less successful in producing positive outcomes, although this may be a result of 12 Although not published in time to be included in this review’s search and analysis, a new systematic review on school-based interventions to reduce dating and sexual violence was released as this paper was submitted for editing. The review came to an overall similar conclusion: there is evidence of positive results for interventions’ effects on knowledge and attitudes, but little evidence of impacts on behavior (De La Rue, L., T. Pigott, et al. (2014). School-Based Interventions to Reduce Dating and Sexual Violence: A Systematic Review). 13 The effective interventions are summarized in: Marx, B. P., Calhoun, K. S., Wilson, A. E., &Meyerson, L. A. (2001). Sexual revictimization prevention: An outcome evaluation. Journal of Consulting and Clinical Psychology, 69(1), 25-32. and Hanson, K. A., &Gidycz, C. A. (1993). Evaluation of a sexual assault prevention program. Journal of Consulting and Clinical Psychology, 61(6), 1046-1052 35 interventions where there had been insufficient investment of adequate time, training, support, and supervision concerning the peer educators, as well as of a “lack of appreciation of the complexity of the peer education process” (Anderson & Whiston, 2005). Furthermore, interventions with statistically significant positive results focused on each topic in-depth instead of briefly covering a range of topics. They also included content on risk-reduction, gender role socialization, or information and discussions on myths and facts about sexual assault. 36 Discussion To the authors’ knowledge, this study represents the first systematic review of reviews to synthesize a fragmented evidence base from specific reviews of interventions to prevent and reduce VAWG. The current evidence base from systematic and comprehensive reviews of impact evaluations suggests that knowledge of intervention impacts on VAWG prevention is growing, but still highly limited. The vast majority of empirical research to date has been devoted to describing and understanding the problem (for example victimization, perpetration, and risk and protective factors of VAWG), rather than testing potential solutions. Much of what has been evaluated has very limited generalizability to the poorest and most vulnerable populations in the world because a large proportion of the evidence comes from evaluations in a few high-income, industrialized countries, using narrowly-defined sample populations (for example university students), or comes from interventions involving implementation capacity that is rarely scalable in settings in low- and middle-income countries (for example trained and licensed clinical social workers or psychologists), because they use models that are often resource-intense and that require a more highly skilled labor force. Lessons Learned That said, there are lessons from high-income countries that could inform piloting and testing in low- resource settings. For instance, psychosocial support has, in some cases, decreased violence in high- income settings. Various modalities of psychosocial support are being increasingly implemented and tested in low- and middle-income settings (for example Murray et al., 2013; Robjant & Fazel, 2010), and which could be usefully applied towards individuals at risk of new or repeated violence exposure or perpetration. Lessons from the more limited evidence base in low- and middle-income country settings may also be instructive. For example, the focus on primary prevention in low- and middle-income settings is worth noting, and, despite fewer evaluations, several innovative programs with promising results were identified that resulted in a reduction of VAWG. Harnessing any applicable lessons learned from such programs and increasing the focus on primary prevention in high-income settings could be valuable. It is worth noting that, despite estimates suggesting that almost one third of women have experienced physical or sexual violence or both by their intimate partner, a large proportion of the research in low-income countries focused on reducing harmful traditional practices, rather than IPV. Whenever possible, when researchers are designing future studies or programs, they should keep in mind the epidemiologic data on the type and prevalence of VAWG. There are lessons to be learned from the different reviews that are in all likelihood applicable to most VAWG interventions. In the case of batterer intervention programs and sexual assault education programs, the reviews for each emphasize the poor quality of both program implementation and the absence of methodological rigor in the research undertaken. It may be unrealistic, for example, to expect only an hour-long video on sexual assault prevention to significantly change youth attitudes, much less reduce date rapes on a university campus. Similarly, failing to adapt or tailor a batterer intervention program to meet the needs and realities of perpetrators, even when the majority of participants drop out, indicates the need for a different approach. As suggested earlier by the Berg & Denison, 2012 review of harmful traditional practices, inconsistent and non-significant findings among impact evaluations may merely reflect inadequate responses to population needs. Unsurprisingly, the evaluations of programs such as those described above, often result in inconclusive reviews, especially when combined with factors such as an absence of control groups, short follow-up periods, selection 37 bias, or a failure to recognize the limitations of certain measures (such as self-reporting). For example, the Smedslund et al., 2007 review only identified six small RCTs and concluded that there were too few evaluations of BIPs that use an RCT design to draw any conclusions regarding this type of intervention. Future evaluations of any VAWG intervention, particularly BIPs and sexual assault education programs, should use an experimental or quasi-experimental study design, in view of the lack so far of evidence-based justification. Additionally, some evaluations have observed adverse effects. These include interventions meant to curb child sexual abuse by strangers and interventions that employ police officers as home visitors paired with social workers. Lessons from triggers of negative effects could inform better design of interventions to prevent and respond to VAWG and to avoid unintended harm. The results underscore the importance of having evaluations that carefully measure and report both positive and negative intervention effects. Several types of interventions suggested as promising by advocacy groups, as well as by the literature, have the potential to prevent VAWG. Yet according to the reviews conducted, many have not been rigorously evaluated. To overcome this limitation, there has been an increasing emphasis on implementing infrastructure-related interventions in order to try to minimize the circumstances that may put women at a greater risk of violence. These could include interventions to redesign special environments to make them less conducive settings for VAWG to occur—for instance, through improving the gender-sensitivity of public transport or increasing policing or community actions in specific “hot spots” (C. Garcia-Moreno & Chawla, 2011; McIlwaine, 2013; Moser & McIlwaine, 2006). Some have argued that energy and water-related projects may reduce the time women spend fetching firewood or water, and thus could minimize their exposure to assault and harassment (Ondeko & Purdin, 2004; Solhjell, Karlsrud, & Sande Lie, 2010). The team found no impact evaluations within reviews testing these sorts of interventions. Additionally, there is an emerging evidence-base suggesting both positive and negative outcomes associated with increased economic empowerment (Heath, 2014; Vyas & Watts, 2009). There are no reviews, however, examining the effects of economic empowerment interventions on VAWG. The World Bank Group review of Bank-supported impact evaluations found very limited evidence from three interventions (see Annex E), despite a great many more impact evaluations having evaluated the effects of economic empowerment interventions without including VAWG-related measures. (Kiplesund & Morton, 2014). Much more work is needed to look at the effects of both straightforward economic empowerment and economic transfer interventions on their own, as well as in programs such as IMAGES in South Africa, which combine economic empowerment intervention components with more specific intervention components designed to address gender- related attitudes and behaviors underlying violence. Strengths and Limitations Because this is a systematic review of reviews, the scope is too wide for meta-analysis. There would be too much heterogeneity in interventions and methods to conduct credible meta-analyses—for example, to pool individual trial data and produce a combined effect size and confidence intervals. This study does, however, synthesize basic information on results, covering all of the 290 relevant impact evaluations identified in the included reviews. The main limitation in the findings is that the team relied on evidence presented in either systematic or comprehensive reviews that met the inclusion criteria for this review. Therefore, some interventions 38 with statistically significant positive results may have been left out because they were not included in a systematic or comprehensive review. This is particularly likely in the case of “grey literature,” such as government reports or community-based interventions that have not been published in peer-reviewed journals. Moreover, recently published articles are less likely to have been included in a systematic review. Although the team attempted to obtain missing information from the reviews by looking directly at the articles cited, in some cases this was not possible, as in the case of unpublished dissertations. Evaluations with incomplete information were not included in the analysis. While the authors verified the findings from the original articles of every evaluation classified as having “significant positive results,” the team relied on the systematic reviews or abstracts for information on the null or negative findings. Therefore, although the authors are confident that the findings do not include any “false positives,” there may be some evaluations that were incorrectly classified as not having statistically significant positive findings. Thus, although this review is not exhaustive, the team believes it represents an accurate view of the current state of evidence on what works to prevent violence against women and girls. Finally, the reviews were extracted using only terms in English. The group plans to update findings with searches in Spanish and French. Implications for Research As already noted this is the first systematic review of reviews that addresses all forms of VAWG. This may be the case in part because conducting systematic reviews of reviews is a very new field of research. The team found many gaps and weaknesses in the evidence base. In light of increasing evidence of the high victimization or perpetration, and severe health consequences, of VAWG, it is troubling that experimental and quasi-experimental trial data on what works to prevent violence is still so scarce. The evidence is highly skewed towards high-income countries, and focuses largely on secondary rather than primary prevention. The most frequently studied interventions, such as batterer’s treatment and screening programs, have largely not shown significant reductions in recidivism or revictimization. The primary prevention programs for non-partner sexual assault and IPV in high- income settings have been primarily conducted among college students, and therefore their value for informing rape prevention programs in low-resource settings is likely to be limited. Among the evaluations included in the review, the team also found many methodological weaknesses. A large number of the evaluations had very small sample sizes, and some of the copious amounts of null findings may be attributable to underpowered study designs. There was a very wide range of outcome measures and time frames, which made comparisons difficult. In addition, many of the quasi- experimental evaluations did not control for confounding factors, and this may lead to some bias in the results, leading to over- or under-estimation of effects. The vast majority of the evaluations identified did not include a long follow-up period, making it difficult to determine if changes are sustained over time. The outcomes used to measure reduction in violence and in perpetration often included self- reports of perpetration triangulated with reports of victimization. It is worth noting that studies that showed significantly positive effects in reported perpetration did not always find these reports verified by victimization reports. This discrepancy puts in question the actual attainment of behavior change and the reliability of outcomes based on self-reporting. 39 There are a number of areas where the evidence base is small or non-existent. For example, the team found only one review for trafficking, and it did not include any evaluations that met the author’s inclusion criteria. Also, there was only one review discussing interventions implemented in humanitarian situations, and there were no reviews on interventions tested in indigenous or ethnically diverse populations. With a few exceptions, the evaluations in this review did not measure cost effectiveness of interventions, which is a pivotal decision point for those who wish to implement and adapt an intervention. The adaptation of interventions to different settings is also undocumented, and information on the time and effort that this crucial step takes is also missing. There was no discussion in the reviews of the potential conflict of interest between the evaluators and the program implementers. In cases where there is lack of resources, self-evaluated interventions may exist and this could bias results. Donors can help correct this situation by providing more adequate funding and incentives for organizations to implement rigorous evaluations. Despite the shortcomings of the current evidence base, some promising trends have emerged. Several evaluations have shown that it is possible to prevent VAWG, with large effect sizes over a relatively short time-frame. The interventions with the most positive findings used multiple approaches and engaged with multiple stakeholders over time. They also addressed underlying risk factors for violence, including social norms regarding gender dynamics and the acceptability of violence. These examples point to the imperative of greatly increasing investment both in innovative programming in primary prevention, as well as experimental and quasi-experimental evaluations to guide international efforts to end VAWG. 40 Annexes Annex A - Full Search Strategy Search Terms: The following terms were searched with Boolean operator and wildcard variants depending on the databases’ demands: violence against girls OR violence against women OR VAW* OR domestic violence OR GBV OR gender violence OR gender-based violence OR femicide OR feminicide OR human trafficking OR trafficking of persons OR partner violence OR abuse of women OR wife abuse OR abuse of wives OR wife battering OR battering of wives OR battering of women OR spouse abuse OR family violence OR murdering of women OR homicides of women OR honor killing OR acid attack* OR acid throwing OR sex selective abortion OR missing women OR missing girls OR widow burning OR witch-craft OR witchcraft OR stoning of women OR rape OR sexual violence OR sexual abuse OR sexual assault OR sexual harassment OR coerced sex OR unwanted sex OR unwanted fondling OR unwanted touching OR harmful traditional practices OR FGM* OR FGC OR female genital mutilation OR female genital cutting OR child marriage OR force marriage OR early marriage OR sexual trafficking OR sexual exploitation OR forced prostitution OR sexual slavery AND review OR meta-analysis OR overview OR summary OR synthesis AND prevent* OR intervention* OR program* OR approaches OR trial* OR evaluation* OR response* OR evidence OR impact* OR effect* OR efficacy OR what works Annex B - Websites Consulted for Systematic or Comprehensive Reviews ● Virtual Knowledge Center to End Violence Against Women and Girls ● Intercambios ● Population Reference Bureau ● Eldis - Gender-based violence (GBV) ● Complete Evidence Base ● What Works to Prevent Partner Violence? An Evidence Overview | STRIVE LSHTM ● GBV Prevention Network ● Publications - Center for the Study and Prevention of Violence - Institute of Behavioral Science ● SVRI Website 41 ● DFID Data base ● 3IE Database ● Campbell Systematic Reviews ● UNICEF Online Library ● WHO Publications ● National Online Resource Center on VAW ● Domestic Violence Evidence Project | A project of the National Resource Center on Domestic Violence ● Violence Against Women - The Gender and Development Network ● Who We Are | Violence is not our Culture Annex C - Data Extraction (selection and coding) Two reviewers (one at the World Bank Group and one at the Global Women’s Institute) independ ently screened all abstracts using an inclusion screening form, and recommended exclusion or pass for further review. At this initial stage, reviewers were blinded of the publisher, journal, and authors; only the titles, years of publication, and abstracts were screened. Any discrepancies were discussed in a meeting with all authors from both institutions for a final decision. Full papers were reviewed for all abstracts passed for further review. The full papers were all reviewed independently by the same two reviewers, and any discrepancies were discussed in a meeting with all authors from both institutions for a final decision. Data for all categories was then extracted by the two reviewers for the eligible systematic reviews. Data was extracted according to a standardized coding and extraction form. Any concerns with data extraction decisions were discussed in a meeting with all authors from both institutions prior to final decisions being made. Data extraction differed for systematic reviews and comprehensive reviews, as they use fundamentally different approaches to collecting information. If further information was needed from primary evaluations because it was unavailable in the review, this data was extracted by one reviewer. For comprehensive reviews, reviews were split evenly for data extraction between two reviewers, and only one reviewer extracted the data for each; any uncertainties were discussed in a meeting with all review authors. Data extracted from systematic reviews include the following: ● Methodological quality of appraisal results of systematic reviews based on the AMSTAR instrument (Shea et al., 2007). ● Objective(s) of the review, type(s) of violence addressed, population(s), intervention(s), evaluation design(s), and outcome(s) eligible in the inclusion criteria, and the numbers of evaluations that met each review’s inclusion criteria. 42 ● Whether or not the review included meta-analyses with this review’s primary or secondary outcomes, and the results. ● Whether the review was published as a peer-reviewed paper in an academic journal or published as grey literature. ● The numbers and references of impact evaluations reviewed (experimental or quasi- experimental trials that measure this review’s primary outcomes), their evaluation designs, interventions evaluated, frequency and duration of interventions, sample sizes, sample population characteristics, country/-ies of evaluation, measures used for this review’s primary outcomes, and results for this review’s primary and secondary outcomes. The review of reviews also coded whether impact evaluations were complemented by any process or implementation evaluations according to the primary review. ● Key conclusions—including implications for practice and/or research—from each review. Data extracted from comprehensive reviews include the following: ● Objective(s) of the review and type(s) of violence addressed. ● Methodology or procedures used, if known, to identify and synthesize impact evaluations. ● Whether the review was published as a peer-reviewed paper in an academic journal or published as grey literature. ● Whether or not the review included meta-analyses with this review’s primary or secondary outcomes, and the results. ● The numbers and references of impact evaluations reviewed (experimental or quasi- experimental trials that measure this review’s primary outcomes), their evaluation designs, stage of the evaluation, interventions evaluated, frequency and duration of interventions, sample sizes, sample population characteristics, country/-ies of evaluation, measures used for this review’s primary outcomes, and results for this review’s primary and secondary outcomes. Any data not available in the review will be listed as “unknown” in the review of reviews. 43 Annex D – All Included Reviews Evaluation Name Complete Citation Systematic Reviews Anderson & Whiston, 2005 Anderson, L. A., & Whiston, S. C. (2005). Sexual Assault Education Programs: A Meta-Analytic Examination of Their Effectiveness. Psychology of Women Quarterly 29 (4): 374-388. Ashman & Duggan, 2004 Ashman, L., & Duggan, L. (2004). Interventions for Learning Disabled Sex Offenders: A Systematic Review. Campbell Systematic Reviews (3). Berg & Denison, 2012 Berg, R.C. & Denison, E. (2012). Interventions to reduce the prevalence of female genital mutilation/cutting in African countries. Campbell Systematic Review (9). Bilukha et al., 2005 Bilukha, O., Hahn, R. A., Crosby, A., Fullilove, M. T., Liberman, A., Moscicki, E.,…Briss, P.A. (2005). The Effectiveness of Early Childhood Home Visitation in Preventing Violence: A Systematic Review. American Journal of Preventive Medicine 28 (2): 11-39. Coulthard et al., 2010 Coulthard, P., Yong, S., Adamson, L., Warburton, A., Worthington, H. V., & Esposito, M. (2004). Domestic violence screening and intervention programmes for adults with dental or facial injury. Campbell Systematic Reviews (12). Davis &Weisburd,2008 Davis, R. C., Weisburd, D., & Taylor, B. (2008). Effects of Second Responder Programs on Repeat Incidents of Family Abuse. Washington, DC: U.S. Department of Justice. Feder et al., 2008 Feder, L., Austin, S., & Wilson, D. (2008). Court Mandated Interventions for Individuals Convicted of Domestic Violence. Campbell Systematic Review (12). Jahanfar et al., 2013 Jahanfar, S., Janssen, P. A., Howard, L., & Dowswell, T. (2013). Interventions for preventing or reducing domestic violence against pregnant women. Cochrane Database of Systematic Reviews (2). 44 Kataoka et al., 2004 Kataoka, Y., Yaju, Y., Eto, H., Matsumoto, N., & Horiuchi, S. (2004). Screening of domestic violence against women in the perinatal setting: a systematic review. Japan Journal of Nursing Science 1 (2): 77-86. Morrison et al., 2004 Morrison, S., Hardison, J., Mathew, A., & O’Neil, J. (2004). An evidence-based review of sexual assault preventive intervention programs. Washington, DC: Department of Justice. Nelson et al., 2012 Nelson, H. D., Bougatsos, C., &Blazina, I. Screening women for intimate partner violence: a systematic review to update the U.S. Preventive Services Task Force recommendation. Annals of Internal Medicine 156 (11): 796-808. O'Reilly et al., 2010 O'Reilly, R., Beale, B., &Gillies, D. (2010). Screening and Intervention for Domestic Violence During Pregnancy Care: A Systematic Review. Trauma, Violence, and Abuse 11 (4): 190-201. Ramsay et al., 2002 Ramsay, J., Richardson, J., Carter, Y. H., Davidson, L. L., &Feder, G. (2002). Should health professionals screen women for domestic violence? Systematic Review. British Medical Journal 325 (7359): 314-318. Ramsay et al., 2005 Ramsay, J., Rivas, C., &Feder, G. (2005). Interventions to reduce violence and promote the physical and psychosocial well-being of women who experience partner abuse: A systematic review. London, UK: Queen Mary’s School of Medicine and Dentistry. Ramsay et al., 2009 Ramsay, J., Rivas, C., Davidson, L., Dunne, D., Eldridge, S., Feder, G., Hegarty, K… Warburton, A. (2009). Advocacy interventions to reduce or eliminate violence and promote the physical and psychosocial well-being of women who experience intimate partner abuse. Cochrane Database of Systematic Reviews (3). Ricardo et al., 2011 Ricardo, C., Eads, M., & Barker, G. (2011). Engaging boys and young men in the prevention of sexual violence: A systematic and global review of evaluated interventions. Pretoria, South Africa: Oak Foundation. 45 Smedslund et al., 2007 Smedslund, G., Dalsbø, T. K., Sterio, A. K., Winsvold, A., & Clench-Aas, J. (2007). Cognitive behavioral therapy for men who physically abuse their female partner. Cochrane Database of Systematic Reviews (2). Spangaro et al., 2013 Spangaro, J., Zwi, A., Adogu, C., Ranmuthugala, G., Davies, G., & Steinacker, L. (2013). What is the evidence of the impact of initiatives to reduce risk and incidence of sexual violence in conflict and post-conflict zones and other humanitarian crises in lower and middle-income countries? A systematic review. London, UK: EPPI- Centre, Social Science Research Unit, Institute of Education, University of London. Taft et al., 2013 Taft. A., O’Doherty, L., Hegarty, K., Ramsay, J., Davidson, L., &Feder, G. (2013). Screening Women for Intimate Partner Violence in Healthcare Settings. Cochrane Database of Systematic Reviews (4). Van Der Laan et al., 2011 Van der Laan, P. H., Smit, M., Busschers, & Aarten, P. (2011). Cross-border trafficking in human beings: Prevention and intervention strategies for reducing sexual exploitation: A Systematic Review. Campbell Systematic Reviews (9). Wathen & Macmillan, 2003 Wathen, C. N., & MacMillan, H. L. (2003). Interventions for violence against women: scientific review. JAMA 289(5): 589-600. Whitaker et al., 2006 Whitaker, D. J., S. Morrison, Lindquist, C., Hawkins, S. R., O'Neil J. A., Nesius, A. M., … Reese L R. (2006). A critical review of interventions for the primary prevention of perpetration of partner violence. Aggression and Violent Behavior 11 (2): 151-166. Zwi et al., 2007 Zwi, K., Woolfenden, S., Wheeler, D. M., O’Brien, T., Tait, P., & Williams, K. J. (2007). School based education programmes for the prevention of child sexual abuse. Cochrane Database of Systematic Reviews (3). Comprehensive Reviews Babcock et al., 2004 Babcock, J. C., Green, C. E., & Robie, C. (2004). “Does batterers’ treatment work? A meta-analytic review of domestic violence treatment.” Child Psychology Review 23 (8): 1023-1053. Bachar & Koss, 2001 Bachar, K., & Koss, M. (2001). “From Prevalence to Prevention: Closing the Gap Between What We Know About Rape and What We Do.” In Sourcebook on Violence Against Women. Eds: C. 46 Renzetti et al. Sage Publications, Thousand Oaks, 2001. Barker et al., 2007 Barker, G., Ricardo, C., & Nascimiento, M. (2007). Engaging men and boys in changing gender-based inequity in health: Evidence from programme interventions. Geneva, Switzerland: World Health Organization. Blackwell et al., 2004 Blackwell, L. M., Lynn, S. J., Vanderhoff, H., & Gidycz, C. (2004). Sexual assault revictimization: Toward effective risk-reduction programs. In L. J. Koenig, L. S. Doll, A. O’Leary, & W. Pequeqnat (Eds.) From child sexual abuse to adult sexual risk: Trauma, revictimization, and intervention (269-295). Washington, DC: American Psychological Association. Blanc et al., 2013 Blanc, A. K., Melnikas, A. Chau, M., & Stoner, M. (2013). A Review of the Evidence on Multisectoral Interventions to Reduce Violence against Adolescent Girls. London: Girl Hub. Bowen, 2011 Bowen, E. (2011). The Rehabilitation of Partner-Violent Men. United Kingdom: Wiley-Blackwell. Brecklin & Forde, 2001 Brecklin, L. R., & Forde, D. R. (2001). A Meta-Analysis of Rape Education Programs. Violence and Victims 16(3): 303-321. Breitenbecher, 2000 Breitenbecher, K. H. (2000). Sexual assault on college campuses: Is an ounce of prevention enough? Applied & Prevention Psychology 9(1): 23-52. Campbell & Manganello, 2006 Campbell, J. C., & Manganello, J. (2006). Changing the Way People Think about Intimate Partner Violence: Changing Public Attitudes as a Prevention Strategy to Reduce Intimate Partner Violence. Journal of Aggression, Maltreatment & Trauma 13(3-4): 13-39. Cornelius &Resseguie,2007 Cornelius, T. L., & Resseguie, N. (2007). Primary and secondary prevention programs for dating violence: A review of the literature. Aggression and Violence Behavior 12(3): 364-375. Evanson, 2006 Evanson, T. A. (2006). Addressing domestic violence through maternal-child health home visiting: what we do and do not know. Journal of Community Health and Nursing 23 (2): 95-111. 47 Foshee et al., 2009 Foshee, V., McNaughton Reyes, H. L., Wyckoff, &S. C.. (2009). Approaches to preventing psychological, physical, and sexual partner abuse. In K. D. O’Leary & E. M. Woodin (Eds.), Psychological and physical aggression in couples: Causes and interventions (165-189). Washington, DC: American Psychological Association. Garrity, 2011 Garrity, S. E. (2011). Sexual assault prevention programs for college-aged men: a critical evaluation. Journal of Forensic Nursing 7(1): 40-48. Guedes, 2004 Guedes, A. (2004). Addressing gender-based violence from the reproductive health/HIV sector: A literature review and analysis. Washington, DC: USAID Interagency Gender Working Group. Heise, 2011 Heise, L. L. (2011). What Works to Prevent Partner Violence? An Evidence Overview. Working Paper, STRIVE Research Consortium, London School of Hygiene and Tropical Medicine, London, UK. Hickman et al., 2004 Hickman, L. J., Jaycox, L. H., & Aronoff, J. (2004). Dating Violence among Adolescents: Prevalence, Gender Distribution, and Prevention Program Effectiveness. Trauma, Violence, & Abuse 5(2): 123-142. Leen et al., 2013 Leen, E., Sorbring, E., Mawer, M., Holdsworth, E., Helsing, B., & Bowen, E. (2013). Prevalence, dynamic risk factors and the efficacy of primary interventions for adolescent dating violence: An international review. Aggression and Violent Behavior 18(1): 159- 174. Lee-Rife et al.,2012 Lee-Rife, S., Malhotra, A., Warner, A., & Glinksi, A. M. (2012). What Works to Prevent Child Marriage: A Review of the Evidence. Studies in Family Planning 43 (4): 287-303. Lund, 2011 Lund, E. M. (2011). Community-based services and interventions for adults with disabilities who have experienced interpersonal violence: a review of the literature. Trauma, Violence and Abuse 12(4): 171- 182. McCollum &Stith,2008 McCollum, E. E., & Stith, S. M. (2008). Couples Treatment for Interpersonal Violence: A Review of Outcome Research Literature and Current Clinical Practices. Violence and Victims 23(2): 187-201. 48 Murphy & Ting, 2010a Murphy, C. M., & Ting, L. (2010). The effects of treatment for substance use problems on intimate partner violence: A review of empirical data. Aggression and Violent Behavior 15(5): 325-333. Murphy & Ting, 2010b Murphy, C. M, & Ting, L. (2010). Interventions for Perpetrators of Intimate Partner Violence: A Review of Efficacy Research and Recent Trends. Partner Abuse 1(1): 26-44. Paluck & Ball, 2010 Paluck, E. L., & Ball, L. (2010). Social norms marketing aimed at gender based violence: A literature review and critical assessment. New York: International Rescue Committee. Sartin et al., 2006 Sartin, R. M., Hansen, D. J., & Huss, M. T. (2006). Domestic violence treatment response and recidivism: A review and implications for the study of family violence. Aggression and Violent Behavior 11(5): 425-440. Sharps et al. 2008 Sharps, P.W., Campbell, J., Baty, M. L., Walker, K. S., & Bair- Merritt, M. H. (2008). Current evidence on perinatal home visiting and intimate partner violence. Journal of Obstetric, Gynecologic, and Neonatal Nursing 37 (4): 480-491. Stith et al., 2012 Stith, S. M., McCollum, E. E., Amanor-Boadu, Y., & Smith, D. (2012). Systematic Perspectives on Intimate Partner Violence Treatment. Journal of Marital and Family Therapy 38(1): 220-240. Stover et al., 2009 Stover, C. S., Meadows, A. L., & Kaufman, J. (2009). Interventions for intimate partner violence: Review and implications for evidence- based practice. Professional Psychology: Research and Practice 40(3): 223-233. Sullivan, 2012 Sullivan, C. M. (2012). Advocacy Services for Women with Abusive Partners: A Review of the Empirical Evidence. Harrisburg, Pennsylvania: National Resource Center on Domestic Violence. Sullivan, 2012 Sullivan, C. M. (2012). Domestic Violence Shelter Services: A Review of the Empirical Evidence. Harrisburg, Pennsylvania: National Resource Center on Domestic Violence. Thakker & Gannon, 2010 Thakker, J., & Gannon, T. A. (2010). Rape treatment: An overview of current knowledge. Behaviour Change 27(4): 227-250. 49 Townsend & Campbell, 2005 Townsend, S. M. & R. Campbell (2005). School-Based Sexual Violence Prevention Programs: Current Evaluation Findings and Policy Implications. H.E. Fitzgerald, R. Zucker, & K Freeark. The Crisis in Youth Mental Health: Critical Issues and Effective Programs (177-196).Westport, Connecticut: Praeger Publishers/Greenwood Publishing Group. Ullman et al., 2007 Ullman, S. E. (2007). A 10-Year Update of “Review and Critique of Empirical Studies of Rape Avoidance”. Criminal Justice and Behavior 34(3): 411-429. Wacker et al., 2009 Wacker, J., Macy, R., Barger, E., & Parish, S. (2009). Sexual assault prevention for women with intellectual disabilities: A critical review of the evidence. Intellect Dev Disabil 47(4): 249-62. WHO & LJMU 2010 World Health Organization, Liverpool JMU – Centre for Public Health. (2010). Violence prevention: the evidence. Geneva: World Health Organization. WHO & LSHTM 2010 World Health Organization, London School of Hygiene and Tropical Medicine. (2010). Preventing intimate partner violence and sexual violence against women: taking action and generating evidence. Geneva, Switzerland: World Health Organization. 50 Annex E – Lessons from World Bank Group Impact Evaluations The World Bank Group (WBG) has spearheaded an initiative, enGENDER IMPACT, to compile and synthesize WBG gender-related evaluations, and launched a Web-based gateway through which the public can access information about each impact evaluation and obtain a summary of the general findings. Evaluations are organized around five outcome areas: (i) Economic Opportunities and Access to Assets; (ii) Education and Skills; (iii) Health; (iv) Voice and Agency; and (v) Gender-Based Violence. Over 160 impact evaluations have thus far been identified and included in the resource-point. The three impact evaluations supported or led by the WBG that measured victimization or perpetration of VAWG all took place in Africa. The three programs primarily targeted women or adolescent girls, and attempted to empower their target populations, both economically and socially. One included addressing VAWG as a primary programmatic focus. The intervention was aimed at female Village Savings and Loans Associations (VSLA) participants and their male partners. VSLA is a methodology aimed at increasing saving opportunities and capital acquisition through the creation of groups. In addition to the standard VSLA model, half of the groups also participated in a Gender Dialogue Group, designed to help participants (both male and female) discuss norms and attitudes regarding financial decisions, the value of women in the household, gender equality and the use of violence. Adding the Gender Dialogue Groups to the VSLA program showed statistically significant increases in control over household economic resources. Physical, sexual, and emotional IPV also decreased, although findings were not statistically significant. Among women and men who attended the Gender Dialogue Groups regularly, physical IPV was significantly reduced. An issue brief examining these three impact evaluations was recently published. While it concludes that there are too few WBG impact evaluations on this topic to establish firm policy recommendations, the results are consistent with broader evidence. The conclusions point to a few recommendations for research and suggesting items for inclusion in program designs: - Consider adding to economic empowerment interventions features that address gender-based violence and/or underlying norms. These features can and should address underlying social and cultural norms enabling VAWG, such as decision-making dynamics, household power relations, and social sanctions or stigmas or both related to specific behaviors. - Engage men and boys to increase the likelihood of success. Rather than engaging men only as potential perpetrators of violence, programs should leverage men’s influence as critical decision-makers and potential agents of change, while at the same time recognizing men’s susceptibility to violence themselves. - Consider the duration and intensity of components aimed at reducing VAWG. It takes time to change deep-rooted norms and behaviors, and few short-term or one-off interventions have proven effective in reducing violence. - Explore active ingredients. The design of the three impact evaluations made it difficult to determine the effect of different components, since it is unclear which programmatic elements, and in what 51 sequence and combination, are essential for reducing VAWG. Complementary process evaluations should be considered, since they can help assess process and implementation factors. - Include valid and reliable VAWG-related measures following ethical guidelines. Two of the impact evaluations showed that it is possible and useful to include VAWG measures and ethically collect data, in interventions that do not focus solely on VAWG. It is important to also include researchers with experience working on VAWG, since they can provide technical expertise and further help ensure the safety of participants and research staff. - Conduct more impact evaluations that use experimental and quasi-experimental designs. Currently, the vast majority of evidence from experimental and quasi-experimental evaluations comes from a few high-income countries and has limited generalizability. Therefore, there is a need to test in different contexts programs designed for preventing VAWG as well as to conduct other interventions that the literature suggests may have positive or negative spillover effects on VAWG. For example, a better understanding of the impacts of cash transfers, microfinance, education, leadership training, and public works on VAWG is needed. Source: Kiplesund, S. & Morton, M., 2014, Gender-based violence prevention: lessons from World Bank impact evaluations, enGENDER IMPACT Issue Brief Series, Washington DC. World Bank Group The enGENDER IMPACT resource point is available at: www.worldbank.org/engenderimpact 52 Bibliography Anderson, L.A.,& Whiston, S. C. (2005). Sexual Assault Education Programs: A Meta-Analytic Examination of Their Effectiveness. Psychology of Women Quarterly 29(4): 374-388. Ashman, L., & Duggan, L. (2004).Interventions for Learning Disabled Sex Offenders: A Systematic Review. Campbell Systematic Reviews (3). Babcock, J.C., Green, C. E., & Robie, C. (2004).“Does batterers’ treatment work? A meta-analytic review of domestic violence treatment.” Child Psychology Review23(8): 1023-1053. Bachar, K., & Koss, M. (2001). “From Prevalence to Prevention: Closing the Gap Between What We Know About Rape and What We Do.” In Sourcebook on Violence Against Women. Eds: C. Renzetti et al. Sage Publications, Thousand Oaks, 2001. Bair-Merritt, M., Jennings, J., Chen, R., Burrell, L., McFarlane, E., Fuddy, L., & Duggan, A. (2010). Reducing maternal intimate partner violence after the birth of a child A randomized controlled trial of the Hawaii healthy start home visitation program. Archives of Pediatrics & Adolescent Medicine, 164(1), 16- 23. Bandiera, O., Buehren, N., Burgess, R., Goldstein, M., Gulesci, S., Rasul, I., &Sulaiman, M. (2012).Empowering Adolescent Girls: Evidence from a Randomized Control Trial in Uganda. Working Paper. Bass, J. K., Annan, J., McIvor Murray, S., Kaysen, D., Griffiths, S., Cetinoglu, T.,…Bolton, P. A. (2013). Controlled Trial of Psychotherapy for Congolese Survivors of Sexual Violence. New England Journal of Medicine 368(23): 2182-2191. Barker, G., Nascimento, M., Pulerwitz, J., & Segundo, M. (2006).Promoting gender-equity among young Brazilian men as an HIV prevention strategy. Washington, DC: Population Council. Barker, G., Ricardo, C., & Nascimiento, M. (2007). Engaging men and boys in changing gender-based inequity in health: Evidence from programme interventions. Geneva, Switzerland: World Health Organization. Berg, R.C. & Denison, E. (2012).Interventions to reduce the prevalence of female genital mutilation/cutting in African countries. Campbell Systematic Review (9). Bilukha, O., Hahn, R. A., Crosby, A., Fullilove, M. T., Liberman, A., Moscicki, E., …Briss, P.A. (2005). The Effectiveness of Early Childhood Home Visitation in Preventing Violence: A Systematic Review. American Journal of Preventive Medicine 28 (2): 11-39. Blackwell, L. M., Lynn, S. J., Vanderhoff, H., & Gidycz, C. (2004).Sexual assault revictimization: Toward effective risk-reduction programs. In L. J. Koenig, L. S. Doll, A. O’Leary, &W. Pequeqnat (Eds.) From child sexual abuse to adult sexual risk: Trauma, revictimization, and intervention (269-295). Washington, DC: American Psychological Association. Blanc, A. K., Melnikas, A. Chau, M., & Stoner, M. (2013).A Review of the Evidence on Multi-sectoral Interventions to Reduce Violence against Adolescent Girls. London: Girl Hub. 53 Blattman, C., Green, E., Annan, J., &Jamison, J. (2013). Building Women’s Economic and Social Empowerment Through Enterprise An Experimental Assessment of the Women’s Income Generating Support (WINGS) Program in Uganda. Innovations for Poverty Action Working Paper. Bott, S., Morrison, A., & Ellsberg, M. (2005). Preventing and responding to gender-based violence in middle and low-income countries: A global review and analysis. Working Paper. Bowen, E. (2011). The Rehabilitation of Partner-Violent Men. United Kingdom: Wiley-Blackwell. Brecklin, L. R., & Forde, D. R. (2001).A Meta-Analysis of Rape Education Programs. Violence and Victims 16(3): 303-321. Breitenbecher, K. H. (2000). Sexual assault on college campuses: Is an ounce of prevention enough? Applied & Prevention Psychology 9(1): 23-52. Butchart, A., & Mikton, C. (2009). Child maltreatment prevention: A systematic review of reviews. Bulletin of the World Health Organization, 87(5): 353–361. Calhoun, K. S., Gidycz, C. A., Loh, C., Wilson, A., Lueken, M., &Outman, R. C. (2001). Sexual assault prevention in high-risk women. Paper presented at the Association for Advancement of Behavior Therapy's 35th Annual Convention, Philadelphia, PA. Campbell, J. C., & Manganello, J. (2006). Changing the Way People Think about Intimate Partner Violence: Changing Public Attitudes as a Prevention Strategy to Reduce Intimate Partner Violence. Journal of Aggression, Maltreatment & Trauma 13(3-4): 13-39. Cornelius, T. L., & Resseguie, N. (2007). Primary and secondary prevention programs for dating violence: A review of the literature. Aggression and Violence Behavior 12(3): 364-375. Coulthard, P., Yong, S., Adamson, L., Warburton, A., Worthington, H. V.,& Esposito, M. (2004). Domestic violence screening and intervention programmes for adults with dental or facial injury. Campbell Systematic Reviews (12). Davis, R. C., Weisburd, D., & Taylor, B. (2008).Effects of Second Responder Programs on Repeat Incidents of Family Abuse. Washington, DC: U.S. Department of Justice. Davis, R. C., Weisburd, D., & Taylor, B. (2008).Effects of Second Responder Programs on Repeat Incidents of Family Abuse. Duflo, E., Dupas, P., Kremer, M., & Sinei, S. (2006). Education and HIV/AIDS prevention: Evidence from a randomized evaluation in western Kenya. (No. Working Paper Number 4024). Washington, DC: World Bank. Duggan, A., McFarlane, E. C., Windham, A. M., Rohde, C. A., Salkever, D. S., &Fuddy, L. (1999). Evaluation of Hawaii’s healthy start program. The Future of Children, 3(1): 66-90. Dupas, P. (2011). Do Teenagers Respond to HIV Risk Information? Evidence from a Field Experiment. American Economic Journal: Applied Economics 3(1): 1-34. Ellsberg M., & Heise L. (2005). Researching Violence Against Women: A Practical Guide for Researchers and Activists. Washington, DC: World Health Organization, PATH. 54 Ellsberg, M., Bradley, C., Egan, A., & Haddad, A. (2008).Violence against women in Melanesia and East Timor; building on global best practices. Canberra, Australia: Office of Development Effectiveness, AusAID. Ellsberg, M., Arango, D. J., Morton, M., Gennari, F., Kiplesund, S., Contreras, M., & Watts, C. (2014). Prevention of violence against women and girls: what does the evidence say? The Lancet . Advance online publication: http://dx.doi.org/10.1016/S0140-6736(14)61703-7 Erulkar, A., & Muthengi, E. (2007). Evaluation of BerhaneHewan: A pilot program to promote education and delay marriage in rural Ethiopia. New York: Population Council. Erulkar, A., & Muthengi, E. (2009). Evaluation of BerhaneHewan: A program to delay child marriage in rural Ethiopia. International Perspectives on Sexual and Reproductive Health, 35(1): 6-14. Evanson, T.A. (2006). Addressing domestic violence through maternal-child health home visiting: what we do and do not know. Journal of Community Health and Nursing 23(2): 95-111. Feder, L., Austin, S., & Wilson, D. (2008). Court Mandated Interventions for Individuals Convicted of Domestic Violence. Campbell Systematic Review (12). Fergus, L. (2012). Background paper for the UN Women Expert Group Meeting: Prevention of Violence Against Women and Girls in Bangkok, Thailand, 17-20 September 2012. Bangkok, Thailand: UN Women. Foshee, V., McNaughton Reyes, H. L., Wyckoff, & S. C..(2009). Approaches to preventing psychological, physical, and sexual partner abuse. In K. D. O’Leary & E. M. Woodin (Eds.), Psychological and physical aggression in couples: Causes and interventions (165-189). Washington, DC: American Psychological Association. Fulu, E., Jewkes, R., Roselli, T., & Garcia-Moreno, C. (2013). Prevalence of and factors associated with male perpetration of intimate partner violence: Findings from the UN multi-country cross-sectional study on men and violence in Asia and the pacific. Retrieved from http://linkinghub.elsevier.com/retrieve/pii/S2214109X13700743\par Garcia-Moreno, C. & Chawla, M. (2011).“Making Cities Safe for Women and Girls: Integrating a Gender Perspective into Urban Health and Planning.” In Women’s Health and the World's Cities, ed. A. I. Meleis, E. L. Birch, and S. M. Wachter. Philadelphia, PA: University of Pennsylvania Press. Garcia-Moreno, C., Jansen, H. A., Ellsberg, M., Heise, L., &Watts, C. H. (2006). Prevalence of intimate partner violence: findings from the WHO multi-country study on women's health and domestic violence. The Lancet 368(9543): 1260. Garrity, S. E. (2011). Sexual assault prevention programs for college-aged men: a critical evaluation. Journal of Forensic Nursing 7(1): 40-48. Green, S., Prof, & Higgins, J. P. T. (Eds.).(2009). Cochrane handbook for systematic reviews of interventions. Hoboken, NJ: Wiley-Blackwell. Guedes, A. (2004). Addressing gender-based violence from the reproductive health/HIV sector: A literature review and analysis. Washington, DC: USAID Interagency Gender Working Group. Hanson, K. A., & Gidycz, C. A. (1993).Evaluation of a sexual assault prevention program. Journal of Consulting and Clinical Psychology 61(6): 1046-1052 55 Heath, R. (2014). Women’s Access to Labor Market Opportunities, Control of Household Resources, and Domestic Violence: Evidence from Bangladesh. World Development 57(0): 32-46. Heise, L.L. (2011). What Works to Prevent Partner Violence? An Evidence Overview. Working Paper, STRIVE Research Consortium, London School of Hygiene and Tropical Medicine, London, UK. Hickman, L. J., Jaycox, L. H., & Aronoff, J. (2004).Dating Violence among Adolescents: Prevalence, Gender Distribution, and Prevention Program Effectiveness. Trauma, Violence, & Abuse 5(2): 123-142. Jahanfar, S., Janssen, P. A., Howard, L., &Dowswell, T. (2013).Interventions for preventing or reducing domestic violence against pregnant women. Cochrane Database of Systematic Reviews 2. Jewkes, R., Nduna, M., Levin J., Jama, N., Dunkle, K., Puren, A., & Duvvury, N.(2008).Impact of stepping stones on incidence of HIV and HSV-2 and sexual behavior in rural South Africa: Cluster randomized controlled trial. British Medical Journal337: a506. Jewkes, R., Fulu, E., Roselli, T., & Garcia-Moreno, C. (2013). Prevalence of and factors associated with non-partner rape perpetration: findings from the UN Multi-country Cross-sectional Study on Men and Violence in Asia and the Pacific. The Lancet Global Health 1(4): e208-e218. Kandpal, E., Baylis, K., & Arends-Kuenning, M. (2013).Measuring the effect of a community-level program on women's empowerment outcomes: evidence from India. World Bank Policy Research Working Paper 6399. Kataoka, Y., Yaju, Y., Eto, H., Matsumoto, N., &Horiuchi, S. (2004). Screening of domestic violence against women in the perinatal setting: a systematic review. Japan Journal of Nursing Science1(2): 77-86. Kim, J. (2007). Developing an integrated model for post-rape care and HIV post-exposure prophylaxis in rural south Africa. Washington, DC: Population Council. Kiplesund, S., & Morton, M. (2014), Gender-based violence prevention: lessons from World Bank impact evaluations, enGENDER IMPACT Issue Brief Series, Washington DC. World Bank Group Klugman, J., Hanmer, L., Twigg, S., Hasan, T., McCleary-Sills, J., & Santamaria, J. (2014). Voice and Agency: Empowering Women and Girls for Shared Prosperity. Washington, DC: World Bank. Knerr, W., Gardner, F., &Cluver, L. (2011). Parenting and the prevention of child maltreatment in low and middle income countries: A systematic review of interventions. Sexual Violence Research Initiative (SVRI), Medical Research Council, South Africa. Leen, E., Sorbring, E., Mawer, M., Holdsworth, E., Helsing, B., & Bowen, E. (2013). Prevalence, dynamic risk factors and the efficacy of primary interventions for adolescent dating violence: An international review. Aggression and Violent Behavior 18(1): 159-174. Lee-Rife, S., Malhotra, A., Warner, A., & Glinksi, A. M. (2012). What Works to Prevent Child Marriage: A Review of the Evidence. Studies in Family Planning 43(4): 287-303. Lund, E. M. (2011). Community-based services and interventions for adults and disabilities who have experienced interpersonal violence: a review of the literature. Trauma, Violence and Abuse 12(4): 171- 182. Marx, B. P., Calhoun, K. S., Wilson, A. E., & Meyerson, L. A. (2001). Sexual revictimization prevention: An outcome evaluation. Journal of Consulting and Clinical Psychology 69(1): 25-32. 56 McCollum, E. E., & Stith, S. M. (2008). Couples Treatment for Interpersonal Violence: A Review of Outcome Research Literature and Current Clinical Practices. Violence and Victims 23(2): 187-201. McIlwaine, C. (2013). Urbanization and gender-based violence: exploring the paradoxes in the global South. Environment and Urbanization 25(1): 65-79. Mikton, C., and Butchart, A. (2009). Child maltreatment prevention: a systematic review of reviews. Bulletin of the World Health Organization (87): 353-361. Miller, E., Decker, M. R., McCauley, H. L., Tancredi, D. J., Levenson, R. R., Waldman, J., . . . Silverman, J. G. (2011). A family planning clinic partner violence intervention to reduce risk associated with reproductive coercion. Contraception, 83(3), 274-280. Morrison, S., Hardison, J., Mathew, A., & O’Neil, J. (2004).An evidence -based review of sexual assault preventive intervention programs. Washington, DC: Department of Justice. Moser, C. O. N. and McIlwaine, C. (2006). Latin American Urban Violence as a Development Concern: Towards a Framework for Violence Reduction. World Development 34(1): 89-112. Murphy, C. M., & Ting, L. (2010). The effects of treatment for substance use problems on intimate partner violence: A review of empirical data. Aggression and Violent Behavior 15(5): 325-333. Murphy, C. M, & Ting, L. (2010). Interventions for Perpetrators of Intimate Partner Violence: A Review of Efficacy Research and Recent Trends. Partner Abuse 1(1): 26-44. Murray, L. K., Familiar, I., Skavenski, S., Jere, E., Cohen, J., Imasiuk, M., …Bolton, P. (2013). An evaluation of trauma focused cognitive behavioral therapy for children in Zambia. Child Abuse & Neglect 37(12): 1175-1185. Nelson, H. D., Bougatsos, C., & Blazina, I. Screening women for intimate partner violence: a systematic review to update the U.S. Preventive Services Task Force recommendation. Annals of Internal Medicine 156(11): 796-808. Ondeko, R., & Purdin, S. (2004). Understanding the causes of gender-based violence. Forced Migration Review, (19): 30. O'Reilly, R., Beale, B., & Gillies, D. (2010). Screening and Intervention for Domestic Violence During Pregnancy Care: A Systematic Review. Trauma, Violence, and Abuse 11(4): 190-201. Paluck, E. L., & Ball, L. (2010). Social norms marketing aimed at gender based violence: A literature review and critical assessment. New York: International Rescue Committee. Pande, R., Kurz, K., Walia, S., MacQuarrie, K., & Jain, S. (2006). Improving the reproductive health of married and unmarried youth in India: Delaying age at marriage in rural Maharashtra, India. Washington, DC: ICRW. Ramsay, J., Richardson, J., Carter, Y.H., Davidson, L.L., & Feder, G. (2002). Should health professionals screen women for domestic violence? Systematic Review. British Medical Journal 325(7359): 314-318. Ramsay, J., Rivas, C., & Feder, G. (2005). Interventions to reduce violence and promote the physical and psychosocial well-being of women who experience partner abuse: A systematic review. London, UK: Queen Mary’s School of Medicine and Dentistry. 57 Ramsay, J., Rivas, C., Davidson, L., Dunne, D., Eldridge, S., Feder, G., Hegarty, K… Warburton, A. (2009). Advocacy interventions to reduce or eliminate violence and promote the physical and psychosocial well-being of women who experience intimate partner abuse. Cochrane Database of Systematic Reviews (3). Renner, L.M., Slack, K. S. (2006). Intimate partner violence and child maltreatment: Understanding intra- and intergenerational connections. Child Abuse and Neglect30 (6):599-617. Ricardo, C., Eads, M., & Barker, G. (2011).Engaging boys and young men in the prevention of sexual violence: A systematic and global review of evaluated interventions. Pretoria, South Africa: Oak Foundation. Robjant, K. & Fazel, M. (2010). The emerging evidence for Narrative Exposure Therapy: A review. Clinical Psychology Review 30(8): 1030-1039. Sartin, R. M., Hansen, D. J., & Huss, M. T. (2006). Domestic violence treatment response and recidivism: A review and implications for the study of family violence. Aggression and Violent Behavior 11(5): 425- 440. Schlosser, R. W. and J. Sigafoos (2009)."‘Empty’ reviews and evidence-based practice." Evidence-Based Communication Assessment and Intervention 3(1): 1-3. Sharps, P.W., Campbell, J., Baty, M. L., Walker, K. S., & Bair-Merritt, M. H. (2008). Current evidence on perinatal home visiting and intimate partner violence. JOGNN—Journal of Obstetric, Gynecologic, and Neonatal Nursing 37(4): 480-491. Shea, B., Andersson, N., Henry, D. (2009). Increasing the demand for childhood vaccination in developing countries: a systematic review. BMC International Health and Human Rights. 9 (Suppl 1):S5. Shea, B. J., Grimshaw, J. M., Wells, G. A., Boers, M., Andersson, N., Hamel, C., . . .Bouter, L. M. (2007). Development of AMSTAR: A measurement tool to assess the methodological quality of systematic reviews. BMC Med Res Methodol, 7, 10. Smedslund, G., Dalsbø, T. K., Sterio, A. K., Winsvold, A., & Clench-Aas, J. (2007).Cognitive behavioral therapy for men who physically abuse their female partner. Cochrane Database of Systematic Reviews (2). Solhjell, R., Karlsrud, J., & Sande Lie, J. H. (2010).Protecting civilians against sexual and gender-based violence in Eastern Chad. Oslo, Norway: The Norwegian Institute of International Affairs. Spangaro, J., Zwi, A., Adogu, C., Ranmuthugala, G., Davies, G., & Steinacker, L. (2013). What is the evidence of the impact of initiatives to reduce risk and incidence of sexual violence in conflict and post- conflict zones and other humanitarian crises in lower and middle-income countries? A systematic review. London, UK: EPPI-Centre, Social Science Research Unit, Institute of Education, University of London. Stith, S. M., McCollum, E. E., Amanor-Boadu, Y., & Smith, D. (2012).Systematic Perspectives on Intimate Partner Violence Treatment. Journal of Marital and Family Therapy 38(1): 220-240. Stöckl, H., Devries, K., Rotstein, A., Abrahama, N., Campbell, J., Watts, C., & Moreno, C. G. The global prevalence of intimate partner homicide: a systematic review. The Lancet 382(9895): 859-865. Stover, C. S., Meadows, A. L., & Kaufman, J. (2009). Interventions for intimate partner violence: Review and implications for evidence-based practice. Professional Psychology: Research and Practice 40(3): 223-233. 58 Sullivan, C. M. (2012). Advocacy Services for Women with Abusive Partners: A Review of the Empirical Evidence. Harrisburg, Pennsylvania: National Resource Center on Domestic Violence. Sullivan, C. M. (2012). Domestic Violence Shelter Services: A Review of the Empirical Evidence. Harrisburg, Pennsylvania: National Resource Center on Domestic Violence. Taft. A., O’Doherty, L., Hegarty, K., Ramsay, J., Davidson, L., & Feder, G. (2013).Scr eening Women for Intimate Partner Violence in Healthcare Settings. Cochrane Database of Systematic Reviews (4). Thakker, J., & Gannon, T. A. (2010). Rape treatment: An overview of current knowledge. Behaviour Change 27(4): 227-250. Tiwari, A., Leung, W. C., Leung, T. W., Humphreys, J., Parker, B., &Ho, P. C. (2005). A randomised controlled trial of empowerment training for Chinese abused pregnant women in Hong Kong. BJOG: An International Journal of Obstetrics & Gynaecology 112 (9): 1249. Townsend, S. M. & R. Campbell (2005). School-Based Sexual Violence Prevention Programs: Current Evaluation Findings and Policy Implications. H.E. Fitzgerald, R. Zucker, & K Freeark. The Crisis in Youth Mental Health: Critical Issues and Effective Programs (177-196).Westport, Connecticut: Praeger Publishers/Greenwood Publishing Group. Ullman, S. E. (2007). A 10-Year Update of “Review and Critique of Empirical Studies of Rape Avoidance”. Criminal Justice and Behavior 34(3): 411-429. UN General Assembly. (1993). Declaration on the Elimination of Violence Against Women. Accessed on May 15, 2014 at http://www.un.org/documents/ga/res/48/a48r104.htm. UN Women, ESCAP, UNDP, UNFPA, UNICEF, and WHO.(2012). “Report of t he Expert Group Meeting on Preventing Violence Against Women and Girls.” Bangkok, Thailand: UN Women. Van der Laan, P. H., Smit, M., Busschers, & Aarten, P. (2011). Cross-border trafficking in human beings: Prevention and intervention strategies for reducing sexual exploitation: A Systematic Review. Campbell Systematic Reviews (9). Verma, R., Pulerwitz, J., Mahendra, V., Khandekar, S., Singh, A., Das, S., . . . Barker, G. (2008). Promoting gender equity as a strategy to reduce HIV risk and gender¬‐based violence among young men in India. Washington, DC: Population Council. Vyas, S. and Watts, C. (2009). How does economic empowerment affect women's risk of intimate partner violence in low and middle income countries? A systematic review of published evidence. Journal of International Development 21(5): 577-602. Wacker, J., Macy, R., Barger, E., & Parish, S. (2009). Sexual assault prevention for women with intellectual disabilities: A critical review of the evidence. Intellect Dev Disabil47(4): 249-62. Walsh, M., Srinathan, S. K., McAuley, D. F., Mrkobrada, M., Levine, O., Ribic, C., …Devereaux, P. J. (2014). The statistical significance of randomized controlled trial results is frequently fragile: a case for a Fragility Index. Journal of clinical epidemiology 67(6): 622-628. Wathen, C. N., & MacMillan, H. L. (2003). Interventions for violence against women: scientific review. JAMA 289(5): 589-600. 59 Whitaker, D. J., S. Morrison, Lindquist, C., Hawkins, S. R., O'Neil J. A., Nesius, A. M., … Reese L R. (2006). A critical review of interventions for the primary prevention of perpetration of partner violence. Aggression and Violent Behavior 11 (2): 151-166. Wolfe, D., Crooks, C., Jaffe, P., Chiodo, D., Hughes, R., Ellis, W., & Donner, A. (2009). A school-based program to prevent adolescent dating violence: A cluster randomized trial. Archives of Pediatrics & Adolescent Medicine 163(8): 692-699. Wolfe, D. A., Wekerle, C., Scott, K., Straatman, A., Grasley, C., & Reitzel-Jaffe, D. (2003). Dating violence prevention with at-risk youth: A controlled outcome evaluation. Journal of Consulting and Clinical Psychology 71(2): 279-291. Women’s Refugee Commission. (2014). Livelihoods. New York City, New York: Women’s Refugee Commission. Access online: http://womensrefugeecommission.org/programs/livelihoods. World Bank.(2011) World development report 2012 - Gender Equality and Development. Washington, DC: World Bank. Access online: http://siteresources.worldbank.org/INTWDR2012/Resources/7778105- 1299699968583/7786210-1315936222006/Complete-Report.pdf. World Health Organization, London School of Hygiene and Tropical Medicine, and the Medical Research Council. (2013). Global and regional estimates of violence against women: Prevalence and health effects of intimate partner violence and non-partner sexual violence, 2. World Health Organization, Liverpool JMU – Centre for Public Health. (2010). Violence prevention: the evidence. Geneva: World Health Organization. World Health Organization, London School of Hygiene and Tropical Medicine. (2010). Preventing intimate partner violence and sexual violence against women: taking action and generating evidence. Geneva: World Health Organization. Yaffe, J., P. Montgomery, et al. (2012). "Empty Reviews: A Description and Consideration of Cochrane Systematic Reviews with No Included Studies." PLoS ONE 7(5): e36626. Zwi, K., Woolfenden, S., Wheeler, D. M., O’Brien, T., Tait, P., &Williams, K.J. (2007). School based education programmes for the prevention of child sexual abuse. Cochrane Database of Systematic Reviews (3). 60