Violence Assignment

Tel +41-22-791-4136 Fax +41-22-791-4171 Tel +41-22-791-2064 Fax +41-22-791-4489 www\fwho\fi\bt/viole\bce_i\bjury_preve\btio\b Preventing intimate partner and sexual violence against \fomen Taking action and generating evidence © World Health Organization 20\f0 All rights reserved.

\fublications of the \borld Health Organization can be obtained from \bHO \fress, \borld Health Organization, 20 Avenue Appia, 1211 Geneva 27, Switzerland (tel.: +41 22 791 3264; fax: +41 22 791 4857; e-mail: [email protected]). Requests for permission to reproduce or translate \bHO publications – whether for sale or for noncommercial distribution – should be addressed to \bHO \fress, at the above address (fax: +41 22 791 4806; e-mail: [email protected]).

The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of the \borld Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. Dotted lines on maps represent approximate border lines for which there may not yet be full agreement.

The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or recom- mended by the \borld Health Organization in preference to others of a similar nature that are not mentioned. Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters.

All reasonable precautions have been taken by the \borld Health Organization to verify the information contained in this publication. However, the published material is being distributed without warranty of any kind, either expressed or implied.

The responsibility for the interpretation and use of the material lies with the reader. In no event shall the \borld Health Organization be liable for damages arising from its use.

The named authors alone are responsible for the views expressed in this publication.

Designed by minimum graphics.

\frinted in Belgium \bHO Library Cataloguing-in-\fublication Data:

\freventing intimate partner and sexual violence against women: taking action and generating evidence / \borld Health Organization and London School of Hygiene and Tropical Medicine.

1.Spouse abuse – prevention and control. 2.Violence – prevention and control. 3.Sexual partners. 4.Risk assessment. 5.Battered women. 6.Sex offenses – prevention and control. I.\borld Health Organization.

ISBN 978 92 4 156400 7 (NLM classification: HV 6625) Suggested citation: \borld Health Organization/London School of Hygiene and Tropical Medicine. \freventing intimate partner and sexual violence against women: taking action and generating evidence. Geneva, \borld Health Organization, 2010. iii Contents Acknowledgements v E\becutive summary \f Foreword 3 Introduction 5 The hidden costs of intimate partner and sexual violence 5 A public health approach to prevention 6 A life-course perspective 8 \bhat is the purpose of this document? 8 \bho should use this document? 9 Key messages 10 Chapter \f The nature, magnitude and consequences of intimate partner and se\bual violence \f\f 1.1 \bhat are intimate partner and sexual violence? 11 1.2 How prevalent are intimate partner and sexual violence? 12 1.3 \bhat are the consequences of intimate partner and sexual violence? 15 1.4 Key messages 17 Chapter 2 Risk and protective factors for intimate partner and se\bual violence \f8 2.1 The ecological model of violence 18 2.2 Identifying risk factors 19 2.3 Risk factors associated with both intimate partner violence and sexual violence 20 2.4 Risk factors associated with intimate partner violence 26 2.5 Risk factors associated with sexual violence 29 2.6 \frotective factors for intimate partner violence and sexual violence 31 2.7 Gender norms and inequality 31 2.8 Key messages 32 iv Preventing intimate Partner and sexual violence against women Chapter 3 Primary prevention strategies – the evidence base 34 3.1 Introduction 34 3.2 Assessing the evidence for different prevention approaches 36 3.3 Summary tables of primary prevention strategies and programmes 38 3.4 During infancy, childhood and early adolescence 41 3.5 During adolescence and early adulthood 44 3.6 During adulthood 47 3.7 All life stages 51 3.8 Key messages 57 Chapter 4 Improving programme planning and evaluation 58 4.1 Step 1: Getting started 58 4.2 Step 2: Define and describe the nature of the problem 62 4.3 Step 3: Identify potentially effective programmes 66 4.4 Step 4: Develop policies and strategies 67 4.5 Step 5: Create an action plan to ensure delivery 68 4.6 Step 6: Evaluate and share learning 70 4.7 Key messages 75 Future research priorities and conclusions 76 Future research priorities 76 Conclusions 77 References 79 Anne\bes Annex A: Operational definitions 91 Annex B: \frevention activities that can be used for process evaluation and potential sources of information on such activities 94 v Acknowledgements This document was written by A\fexander Butchart, \b\faudia Garcia-Moreno, and \bhris- topher Mikton of the Wor\fd Hea\fth Organization, and bui\fds on an ear\fy draft produced by Joanna Nurse of the London Schoo\f of Hygiene and Tropica\f Medicine and the United Kingdom Department of Hea\fth, and Damian Basher of the United Kingdom Depart ment of Hea\ft h. Nata\fia Diaz- Granados of McMaster University drafted \bhap- ter 2: Risk and protective factors for intimate partner and sexua\f vio\fence , whi\fe Joanne K\fevens and Linda Anne Va\f\fe of the United States \benters for Disease \bontro\f and Prevention contributed to \bhapter 4: Improving programme p\fanning and eva\fuation . The document was edited by Tony Wadde\f\f.

The document has a\fso benefited great\fy from comments and suggestions provided by the fo\f\fowing peer reviewers: Linda Dah\fberg, Kathryn Graham, Rache\f Jewkes, Ho\f\fy Johnson, Ju\fia Kim, Karen Lang, \bhar\fotte Watts and A\fys Wi\f\fman.

The conceptua\f foundations for this document were out\fined in a May 2007 expert consu\ftation on the primary prevention of intimate partner and sexua\f vio\fence and the background paper for the meeting prepared by A\fison Harvey, \b\faudia Garcia-Moreno and A\fexander Butchart. \bonsu\ftation participants inc\fuded: Mark Be\f\fis, Pimpawun Boonmongkon, David Brown, \barme \b\fave\f-Arcas, Jane \bottingham, Pame\fa \box, Isabe\f\fe De Zoysa, Jane Ferguson, Vangie Foshee, Kathryn Graham, Henrica Jansen, Ho\f\fy Johnson, Faith Kasiva, Jackson Katz, Karen Lang, Lori Michau, Adepeju Ader- emi O\fukoya, Judith Ann Po\fsky, Jame\fa Sa\feh A\f-Raiby, Susan B\fair Timber\fake, Ber- nice Van Bronkhorst and \bhar\fotte Watts.

The deve\fopment and pub\fication of this document has been made possib\fe by the gen- erous financia\f support of the Government of Be\fgium and the United States \benters for Disease \bontro\f and Prevention. 1 E\becutive summary Intimate partner and sexua\f vio\fence affect a \farge proportion of the popu\fation – with the majority of those direct\fy experiencing such vio\fence being women and the majority perpetrating it being men. The harm they cause can \fast a \fifetime and span genera- tions, with serious adverse affects on hea\fth, education and emp\foyment. The primary prevention of these types of vio\fence wi\f\f therefore save \fives and money – investments made now to stop intimate partner and sexua\f vio\fence before they occur wi\f\f protect the physica\f, menta\f and economic we\f\f-being and deve\fopment of individua\fs, fami\fies, communities and who\fe societies.

This document aims to provide sufficient information for po\ficy-makers and p\fanners to deve\fop data-driven and evidence-based programmes for preventing intimate part- ner and sexua\f vio\fence against women and is divided into the fo\f\fowing chapters:

l Chapter 1 out\fines the nature, magnitude and consequences of intimate partner and sexua\f vio\fence within the broader typo\fogy of vio\fence.

l Chapter 2 identifies the risk and protective factors for such vio\fence and the impor - tance of addressing both risk and protective factors in prevention efforts.

l Chapter 3 summarizes the scientific evidence base for primary prevention strate- gies, and describes programmes of known effectiveness, those supported by emerging evidence and those that cou\fd potentia\f\fy be effective but have yet to be sufficient\fy eva\fuated for their impact.

l Chapter 4 presents a six-step framework for taking action, generating evidence and sharing resu\fts.

In the c\fosing section, severa\f future research priorities are out\fined and a number of key conc\fusions drawn.

The evidence-based prevention of intimate partner and sexua\f vio\fence is sti\f\f in its ear\fy days and much remains to be accomp\fished. At present, on\fy one strategy has evidence supporting its effectiveness – and this on\fy re\fates to intimate partner vio\fence. The strategy in question is the use of schoo\f-based programmes to prevent vio\fence within dating re\fationships. Evidence is, however, emerging of the effectiveness of a number of other strategies for preventing intimate partner and sexua\f vio\fence, inc\fuding micro- finance programmes for women combined with gender-equa\fity education; efforts to reduce access to and harmfu\f use of a\fcoho\f; and changing socia\f and cu\ftura\f gender norms. Many more strategies appear to have potentia\f, either on theoretica\f grounds or because they target known risk factors, but most of these have never been systematica\f\fy imp\femented – \fet a\fone eva\fuated. 2 Preventing intimate Partner and sexual violence against women T he pub\f ic he a \ft h approach to prevent ion t a ken i n t h i s doc u ment i s i ntended to comp\fe - ment crimina\f justice-based approaches. The approach re\fies upon the use of popu\fa- tion-based data to describe the prob\fem, its impact and associated risk and protective factors, whi\fe drawing upon the scientific evidence for effective, promising and theo- retica\f\fy indicated prevention strategies. Part of the approach is a\fso to ensure that a\f\f po\ficies and programmes inc\fude in-bui\ft monitoring and eva\fuation mechanisms. At the same time, taking a \fife-course perspective wi\f\f he\fp to identify ear\fy risk factors and the best times to disrupt the deve\fopmenta\f trajectories towards vio\fent behaviour using a primary prevention approach. For successfu\f primary prevention, ear\fy intervention is required that focuses on younger age groups.

A\fthough pressing, the need for evidence and further research in a\f\f these areas in no way prec\fudes taking action now to prevent both intimate partner vio\fence and sexua\f vio\fence. Those programmes that have evidence supporting their effectiveness shou\fd be imp\femented and, where necessary, adapted. Those that have shown promise or appear to have potentia\f can a\fso p\fay an immediate ro\fe – provided strenuous efforts are made to incorporate at the outset rigorous outcome eva\fuations. It is on\fy by tak- ing action and generating evidence that intimate partner and sexua\f vio\fence wi\f\f be prevented and the fie\fd of evidence-based primary prevention of such vio\fence wi\f\f suc- cessfu\f\fy mature. 3 Foreword Intimate partner vio\fence and sexua\f vio\fence are serious and widespread prob\fems wor\fdwide. Apart from being vio\fations of human rights, they profound\fy damage the physica\f, sexua\f, reproductive, emotiona\f, menta\f and socia\f we\f\f-being of individua\fs and fami\fies. The immediate and \fong-term hea\fth outcomes that have been \finked to these types of vio\fence inc\fude physica\f injury, unwanted pregnancy, abortion, gynae- co\fogica\f comp\fications, sexua\f\fy transmitted infections (inc\fuding HIV/AIDS), post- traumatic stress disorder and depression, among others. There are a\fso a number of pregnancy-re\fated comp\fications such as miscarriage, premature \fabour and \fow birth weight associated with vio\fence during pregnancy. In addition, high-risk behaviours such as smoking, harmfu\f use of a\fcoho\f and drugs and unsafe sex are significant\fy more frequent among victims of intimate partner and sexua\f vio\fence.

The overwhe\fming burden of intimate partner vio\fence (physica\f, sexua\f and emotion- a\f) and of sexua\f vio\fence is borne by women at the hands of men. In acknow\fedging the \farge sca\fe and extensive consequences of the prob\fem, scores of countries have passed \faws to crimina\fize intimate partner and sexua\f vio\fence, and many countries are increasing\fy providing \fega\f, hea\fth and socia\f services to abused women. At the same time, there have been remarkab\fy few efforts to prevent intimate partner or sexua\f vio\fence from occurring in the first p\face. We\f\f-designed and proper\fy imp\femented services for victims wi\f\f continue to be vita\f\fy important. However the sheer magnitude of the prob\fem means that preventing intimate partner and sexua\f vio\fence before it occurs wi\f\f be crucia\f not on\fy in reducing the burden of suffering but a\fso in reducing the \fong-term human, economic and pub\fic hea\fth costs of such vio\fence. Both primary prevention efforts and responding services need to be we\f\f documented and eva\fuated for their effectiveness.

The current paucity of evidence-based prevention approaches is part\fy due to the sepa- rate deve\fopment and imp\fementation of research and advocacy activities in this area.

It is a\fso due in part to the comp\fex array of factors that increase the \fike\fihood of such vio\fence occurring in the first p\face. These factors inc\fude gender inequa\fity and socia\f norms around mascu\finity, and other socia\f determinants such as economic inequa\fity; other prob\fem behaviours (such as harmfu\f use of a\fcoho\f); and other types of vio- \fence (such as chi\fd ma\ftreatment). There is \fitt\fe coordination between programmes or research agendas on chi\fd ma\ftreatment, a\fcoho\f and substance misuse, socia\f determi- nants, and intimate partner and sexua\f vio\fence. And yet a\f\f of these prob\fems regu\far\fy affect fami\fies, predictab\fy with greater frequency in economica\f\fy disadvantaged com- munities where there is often a\fso greater inequa\fity between women and men. 4 Preventing intimate Partner and sexual violence against women Preventing intimate partner and sexua\f vio\fence against \bomen: taking action and generating evidence aims to strengthen the fie\fd of intimate partner and sexua\f vio\fence prevention so that prevention efforts can comp\fement hea\fth, \fega\f and other service-de\fivery pro- grammes that respond to the prob\fem. Primary prevention po\ficies and programmes shou\fd be proactive, informed by the best avai\fab\fe research evidence and focus on the root causes of the prob\fem, inc\fuding gender inequa\fity, so that fewer women are affected by such vio\fence in the first p\face.

This document is a\fso bui\ft around a \fife-course perspective that recognizes how infant and ear\fy chi\fdhood experiences influence the \fike\fihood of \fater becoming a perpetra- tor or victim of intimate partner and sexua\f vio\fence. A review is provided of the \fatest avai\fab\fe scientific findings on effective, promising and theoretica\f\fy feasib\fe prevention strategies, and a pub\fic hea\fth mode\f used to high\fight the steps needed to address the under\fying determinants at the individua\f, fami\fy, community and societa\f \feve\f. Impor- tant\fy, it is a\fso strong\fy urged that future efforts to address intimate partner and sexua\f vio\fence are designed and conducted with the exp\ficit aim of expanding the scientific evidence base on the magnitude, consequences and preventabi\fity of these prob\fems.

This can be achieved for examp\fe by integrating outcome eva\fuations into prevention projects from the very start.

Reflecting the g\foba\f importance of preventing intimate partner and sexua\f vio\fence, this document is a joint pub\fication of the Wor\fd Hea\fth Organization and the Lon- don Schoo\f of Hygiene and Tropica\f Medicine. P\fease join us in acting to increase the number, qua\fity and effectiveness of intimate partner and sexua\f vio\fence prevention programmes in a\f\f of the wor\fd’s societies. This wi\f\f contribute to a significant improve- ment in the persona\f safety, we\f\f-being and happiness of gir\fs, boys, women and men everywhere.

Etienne Krug Director, WHO Department of Vio\fence and Injury Prevention and Disabi\fity \bichae\f \bbizvo Director ( ad interim), WHO Department of Reproductive Hea\fth and Research 5 Introduction The hidden costs of intimate partner and se\bual violence Because of their nature, the occurrence and impacts of intimate partner and sexua\f vio- \fence are frequent\fy “hidden” resu\fting in a significant underestimation of the rea\f \feve\f of harm caused. However, popu\fation-based surveys revea\f that these forms of vio\fence are common. For examp\fe, the WHO Mu\fti-country study on \bomen’s hea\fth and domestic vio\fence against \bomen (Garcia-Moreno et a\f., 2005) indicated that 15 –71% of women experience physica\f and/or sexua\f vio\fence by an intimate partner at some point in their \fives. Research a\fso shows that intimate partner and sexua\f vio\fence \fead to a wide range of short- and \fong-term physica\f, menta\f and sexua\f hea\fth prob\fems (Heise & Garcia- Moreno, 2002; Jewkes, Sen & Garcia-Moreno, 2002). In fact, the magnitude of poor hea\fth outcomes due to intimate partner and sexua\f vio\fence are comparab\fe to (and in some cases exceed) those associated with many other better-known hea\fth risk factors.

For examp\fe, a study using burden of disease methodo\fogy in Victoria, Austra\fia esti- mated that among women 18 – 44 years of age, intimate partner vio\fence was associated with 7% of the overa\f\f burden of disease. Such vio\fence was a \farger risk factor than others traditiona\f\fy inc\fuded, such as raised b\food pressure, tobacco use and increased body weight (Vos et a\f., 2006).

Intimate partner vio\fence a\fso very often has severe negative impacts on the emotiona\f and socia\f we\f\f-being of the who\fe fami\fy, with adverse effects on parenting ski\f\fs and on educationa\f and emp\foyment outcomes. Some chi\fdren from househo\fds where there is intimate partner vio\fence may exhibit increased rates of behavioura\f and emotiona\f prob\fems that can resu\ft in increased difficu\fties with education and emp\foyment, often \feading to ear\fy schoo\f drop out, youth offending and ear\fy pregnancy (Anda et a\f., 2001; Dube et a\f., 2002).

Intimate partner vio\fence a\fso has a significant adverse economic impact. For examp\fe, in the United Kingdom, one ana\fysis estimated that its annua\f cost to the economy in Eng\fand and Wa\fes was approximate\fy £22.9 bi\f\fion (Wa\fby, 2004). Another study in the United States of America (Arias & \borso, 2005) \fooked at sex differences in the use of services to treat injuries caused by physica\f vio\fence from a partner. \bompared to men:

l women were more \fike\fy to report physica\f intimate partner vio\fence and resu\fting injuries; l a \farger proportion of women reported seeking menta\f hea\fth services in response to experiencing physica\f intimate partner vio\fence, and reported more visits on aver- age; 6 Preventing intimate Partner and sexual violence against women l women were more \fike\fy to report the use of emergency department, inpatient hospi- ta\f and physician services; and l women were more \fike\fy to take time off from work and from chi\fdcare or househo\fd duties because of their injuries.

In addition, the average cost per person for women experiencing at \feast one occurrence of physica\f intimate partner vio\fence was more than twice that for men. When the costs of individua\fs not reaching their fu\f\f productive potentia\f are factored in, the overa\f\f costs to society wi\f\f be even greater.

A public health approach to prevention Different – but over\fapping and \farge\fy comp\fementary – approaches and perspectives have been adopted to understand intimate partner and sexua\f vio\fence. For instance, the gender perspective emphasizes patriarchy, power re\fations and hierarchica\f con- structions of mascu\finity and femininity as a predominant and pervasive driver of the prob\fem. These are predicated on contro\f of women and resu\ft in structura\f gender inequa\fity.

The human rights approach is based on the ob\figations of states to respect, protect and fu\ffi\f human rights and therefore to prevent, eradicate and punish vio\fence against women and gir\fs. It recognizes vio\fence against women as a vio\fation of many human rights: the rights to \fife, \fiberty, autonomy and security of the person; the rights to equa\fity and non-discrimination; the rights to be free from torture and crue\f, inhuman and degrading treatment or punishment; the right to privacy; and the right to the high- est attainab\fe standard of hea\fth. These human rights are enshrined in internationa\f and regiona\f treaties and nationa\f constitutions and \faws, which stipu\fate the ob\figations of states, and inc\fude mechanisms to ho\fd states accountab\fe. The \bonvention on the E\fimination of A\f\f Forms of Discrimination against Women (\bEDAW), for examp\fe, requires that countries party to the \bonvention take a\f\f appropriate steps to end vio\fence against women, as does the Inter-American \bonvention on the Prevention, Punishment and Eradication of Vio\fence against Women (\bonvention of Be\fem do Para, 1994), the on\fy regiona\f convention to specifica\f\fy address such vio\fence.

The crimina\f justice approach sees its main task as responding to intimate partner and sexua\f vio\fence after it has occurred by enforcing the \faw and ensuring that “justice is done”. This invo\fves proper\fy identifying perpetrators of intimate partner and sexua\f vio\fence, ascertaining their gui\ft and ensuring that they are appropriate\fy sentenced.

To prevent and respond to intimate partner and sexua\f vio\fence the crimina\f justice approach re\fies primari\fy on deterrence, incarceration and the punishment and reha- bi\fitation of perpetrators.

T h e pu b \f ic h e a \f t h appr o ac h 1 a dopt e d by t h i s do c u m e nt d r aw s on t h e s e ot h e r appr o ac h e s and perspectives, particu\far\fy the gender perspective. As described in the Wor\fd report on vio\fence and hea\fth (Krug et a\f., 2002), the pub\fic hea\fth approach is a science-driv- en, popu\fation-based, interdiscip\finary, intersectora\f approach based on the eco\fogica\f mode\f which emphasizes primary prevention. 1 Pub\fic hea\fth is “the science and art of preventing disease, pro\fonging \fife and promoting hea\fth through the organized efforts and informed choices of society, organizations, pub\fic and private, communities and individua\fs” ( Wins\fow, 1920). 7 Rather than focusing on individua\fs, the pub\fic hea\fth approach aims to provide the maximum benefit for the \fargest number of peop\fe, and to extend better care and safety to entire popu\fations. The pub\fic hea\fth approach is interdiscip\finary, drawing upon know\fedge from many discip\fines inc\fuding medicine, epidemio\fogy, socio\fogy, psy- cho\fogy, crimino\fogy, education and economics. Because intimate partner and sexua\f vio\fence are mu\fti-faceted prob\fems, the pub\fic hea\fth approach to these forms of vio- \fence emphasizes a mu\fti-sectora\f response. It has been proved time and again that cooperative efforts from such diverse sectors as hea\fth, education, socia\f we\ffare, and crimina\f justice are often necessary to so\fve what are usua\f\fy assumed to be pure\fy “crimina\f” or “medica\f” prob\fems. The pub\fic hea\fth approach considers that vio\fence, rather than being the resu\ft of any sing\fe factor, is the outcome of mu\ftip\fe risk factors and causes, interacting at four \feve\fs of a nested hierarchy (individua\f, c\fose re\fation- ship/fami\fy, community and wider society).

From a pub\fic hea\fth perspective, prevention strategies can be c\fassified into three types (Dah\fberg & Krug, 2002):

l Primary prevention – approaches that aim to prevent vio\fence before it occurs.

l Secondary prevention – approaches that focus on the more immediate responses to vio\fence, such as pre-hospita\f care, emergency services or treatment for sexua\f\fy transmitted infections fo\f\fowing a rape.

l Tertiary prevention – approaches that focus on \fong-term care in the wake of vio - \fence, such as rehabi\fitation and reintegration, and attempt to \fessen trauma or reduce \fong-term disabi\fity associated with vio\fence.

A pub\fic hea\fth approach emphasizes the primary prevention of intimate partner and sexua\f vio\fence, i.e. stopping them from occurring in the first p\face. Unti\f recent\fy, this approach has been re\fative\fy neg\fected in the fie\fd, with the majority of resources directed towards secondary or tertiary prevention. Perhaps the most critica\f e\fement of a pub\fic hea\fth approach to prevention is the abi\fity to identify under\fying causes rather than focusing upon more visib\fe “symptoms”. This a\f\fows for the deve\fopment and testing of effective approaches to address the under\fying causes and so improve hea\fth.

The pub\fic hea\fth approach is an evidence-based and systematic process invo\fving the fo\f\fowing four steps:

1. Defining the prob\fem conceptua\f\fy and numerica\f\fy, using statistics that accu - rate\fy describe the nature and sca\fe of vio\fence, the characteristics of those most affected, the geographica\f distribution of incidents, and the consequences of expo- sure to such vio\fence ( Chapter \f).

2. Investigating why the prob\fem occurs by determining its causes and corre\fates, the factors that increase or decrease the risk of its occurrence (risk and protective factors) and the factors that might be modifiab\fe through intervention ( Chapter 2).

3. Exp\foring ways to prevent the prob\fem by using the above information and designing, monitoring and rigorous\fy assessing the effectiveness of programmes through outcome eva\fuations ( Chapters 3, 4).

4. Disseminating information on the effectiveness of programmes and increasing the sca\fe of proven effective programmes. Approaches to prevent intimate partner and sexua\f vio\fence, whether targeted at individua\fs or entire communities, must be i ntroduction 8 Preventing intimate Partner and sexual violence against women proper\fy eva\fuated for their effectiveness and the resu\fts shared ( section 4.6). This step a\fso inc\fudes adapting programmes to \foca\f contexts and subjecting them to rigorous re-eva\fuation to ensure their effectiveness in the new setting. A life-course perspective This document views the causes, consequences and prevention of intimate partner and sexua\f vio\fence within a \fife-course perspective based upon understanding how influences ear\fy in \fife can act as risk factors for hea\fth-re\fated behaviours or hea\fth prob\fems at \fater stages (Davey Smith, 2000). For the purpose of understanding and preventing intimate partner and sexua\f vio\fence, the \fife course is divided into: infan- cy (0–4 years) , chi\fdhood and ear\fy ado\fescence (5 –14 years); ado\fescence and young adu\fthood (15 –25 years); adu\fthood (26 years and over); and a\f\f ages.

Each of these \fife stages represents a specia\f phase in the deve\fopment of risk and a unique opportunity to target deve\fopmenta\f\fy specific risk factors. Adopting such a \fife-course perspective therefore he\fps to both identify ear\fy risk factors and the best times to intervene using a primary prevention approach.

What is the purpose of this document?

This document provides a p\fanning framework for deve\foping po\ficies and programmes for the prevention of intimate partner and sexua\f vio\fence. Drawing upon sound pub\fic hea\fth princip\fes, and supported by avai\fab\fe evidence on risk factors for such vio\fence and on the effectiveness of prevention programmes, it out\fines how prevention strate- gies tai\fored to the needs, capacities and resources of particu\far settings can be deve\f- oped. The document emphasizes the importance of integrating scientific eva\fuation procedures into a\f\f prevention initiatives to continuous\fy monitor and improve their effectiveness, and to inform and expand the g\foba\f evidence base in this area.

At present, the bu\fk of the \fimited human and financia\f resources expended in the area of intimate partner and sexua\f vio\fence are devoted to the treatment, management and support of victims (who often do not seek out services unti\f they are o\fder) and the arrest and incarceration of perpetrators. Drawing \fessons from a pub\fic hea\fth perspec- tive means however that a simi\far\fy strong emphasis shou\fd now be p\faced upon the primary prevention of intimate partner and sexua\f vio\fence. The principa\f focus of this document is therefore on the use of a pub\fic hea\fth approach to preventing intimate partner and sexua\f vio\fence before it even occurs as this has the potentia\f to:

l enhance the hea\fth, we\f\f-being and productivity of individua\fs, communities and societies; l he\fp reduce gender-based inequities, in so far as intimate partner and sexua\f vio\fence – or the threat of them – contribute to maintaining such inequities; l reduce the antisocia\f behaviour, offending and other forms of interpersona\f vio\fence that can resu\ft from witnessing intimate partner vio\fence as a chi\fd; and l reduce the direct and indirect economic and socia\f costs and consequences associated with these forms of vio\fence. 9 Given – as indicated above – that the overwhe\fming burden of intimate partner and sexua\f vio\fence is borne by women at the hands of men, this document focuses on inti- mate partner and sexua\f vio\fence against women. Despite findings (main\fy from the United States), based on se\ff-reports, that men and women perpetrate intimate partner vio\fence at approximate\fy the same rate (Archer, 2000, 2006; \burrie, 1998; Strauss, 1998, 2009), women are over-represented in terms of deaths and severe injuries due to intimate partner and sexua\f vio\fence, and in emergency room and c\finica\f popu\fa- tions (Anderson, 2005; Archer 2000, 2006; Straus 1999, 2009). The findings of sex symmetry primari\fy concern the \fess-severe forms of physica\f intimate partner vio\fence and appear to app\fy main\fy to high-income western countries (Archer, 2006). Further- more, most of the current\fy avai\fab\fe evidence on intimate partner and sexua\f vio\fence comes from studies of how they affect women. Fina\f\fy, the scope of this document does not inc\fude the significant differences that exist in the nature of, and risk factors and specific prevention strategies for, intimate partner and sexua\f vio\fence perpetrated by women and ma\fe-to-ma\fe sexua\f vio\fence.

Evidence on the nature of intimate partner and sexua\f vio\fence and the effectiveness of prevention efforts has wherever possib\fe been se\fected on\fy from higher-qua\fity studies – for examp\fe, from systematic reviews and \farge studies with good methodo\fogies such as the use of randomized-contro\f\fed tria\fs. However, most of the high-qua\fity research studies come from the United States and the evidence of effectiveness may not a\fways be direct\fy transferab\fe to a \fow-income setting. The evidence reviewed a\fso inc\fudes examp\fes of promising practice from \fow- and midd\fe-income countries (LMI\b).

Who should use this document?

This document is primari\fy aimed at po\ficy-makers, programme deve\fopers and p\fan- ners and funding bodies in pub\fic hea\fth and re\fated sectors that aim to advance the pre- vention of intimate partner and sexua\f vio\fence against women. A \fthough pub\fic hea\fth practitioners are a key target audience, the document assumes no previous know\fedge of pub\fic hea\fth princip\fes. In addition to the principa\f audience, other interested parties wi\f\f inc\fude:

l those working in other government sectors such as education, chi\fd we\ffare, socia\f care, crimina\f justice and departments of women or gender equa\fity; l advocates for the prevention of intimate partner and sexua\f vio\fence, for examp\fe, from nongovernmenta\f organizations; l \foca\f authorities; l environmenta\f and urban p\fanners; and l researchers.

It is intended that practitioners and other professiona\fs working in these sectors wi\f\f find this document a usefu\f source of information on state-of-the-science efforts to enhance and strengthen mu\ftisectora\f co\f\faboration in the design, de\fivery and eva\fuation of pro- grammes for the primary prevention of intimate partner and sexua\f vio\fence. i ntroduction 10 Preventing intimate Partner and sexual violence against women Key messages n Intimate partner and sexual violence may affect everyone – though there is a strong gender pattern with the majority of those experiencing it being women and the majority perpetrating it being men; everyone can help to prevent it and can work together to stop the continuation of violence across generations.

n The harm caused by such violence can last a lifetime and span generations, with serious adverse affects on health, education, employment, crime and on the economic well-being of individuals, families, communities and societies.

n The primary prevention of intimate partner and sexual violence is likely to save lives and money – investment to stop intimate partner and sexual violence before they occur is crucial as this will protect and promote the well-being and development of individuals, families, communities and societies.

n This document takes a public health approach that draws on gender, human rights and criminal justice perspectives on prevention. The approach is based upon the use of population-based data to describe the problem, its impact, and its risk and protective factors, and draws upon the scientific evidence for proven, promising and theoretically indicated prevention strategies, while urging that all policies and programmes include in-built monitoring and evaluation mechanisms.

n Taking a life-course perspective helps to identify early risk factors and the best times to intervene using a primary prevention approach. For successful primary prevention, early intervention is required that focuses on younger age groups.

n Using evidence of what has already worked will increase the likelihood that prevention efforts are successful.

n Incorporating outcome evaluations into primary prevention interventions is crucial in generating further evidence. 11 CHA\fTER 1 The nature, magnitude and consequences of intimate partner and se\bual violence \f .\f What are intimate partner and se\bual violence?

This document adopts the definitions of intimate partner vio\fence and sexua\f vio\fence from the Wor\fd report on vio\fence and hea\fth (Heise & Garcia-Moreno, 2002; Jewkes, Sen & Garcia-Moreno, 2002). These definitions ( Bo\b \f) are now wide\fy used and pro- vide a usefu\f conceptua\f framework. However, they need to be operationa\fized for the purposes of monitoring progress and demonstrating how the magnitude of the prob\fem changes in response to specific prevention efforts. For more information on the use of operationa\f definitions see Anne\b A .

BOX 1 Definition of terms n Intimate partner violence – behaviour within an intimate relationship that causes physical\f sexual or psychological harm\f including acts o\b physical aggression\f sexual coercion\f psychological abuse and controlling behaviours .

This definition covers violence by both current and former spouses and partners.

n Se\bual violence – any sexual act\f attempt to obtain a sexual act\f unwanted sexual comments or advances\f or acts to traffic\f or otherwise directed against a person’s sexuality using coercion\f by any person regardless o\b their relationship to the victim\f in any setting including but not limited to home and work .

This definition includes rape, defined as the physically forced or otherwise coerced penetration of the vulva or anus with a penis, other body part or object – however the legal definition of rape may vary in different countries. Source: Heise & Garcia-Moreno (2002); Jewkes, Sen & Garcia-Moreno (2002) Intimate partner vio\fence occurs main\fy from ado\fescence and ear\fy adu\fthood onwards, most often in the context of marriage or cohabitation, and usua\f\fy inc\fudes physica\f, sexua\f and emotiona\f abuse as we\f\f as contro\f\fing behaviours. Sexua\f vio\fence can occur at any age – inc\fuding during chi\fdhood – and can be perpetrated by parents, caregivers, acquaintances and strangers, as we\f\f as intimate partners. Both forms of vio\fence are in the majority perpetrated by men against gir\fs and women; however the sexua\f abuse of ma\fe chi\fdren is a\fso common. Intimate partner vio\fence may a\fso be perpetrated by women against men and can occur in the context of same-sex re\fationships. 12 Preventing intimate Partner and sexual violence against women In preventing intimate partner and sexua\f vio\fence from occurring in the first p\face, it is important to keep in mind how such vio\fence is re\fated to other forms of interpersona\f, se\ff-directed and co\f\fective vio\fence. For examp\fe, experiencing chi\fd ma\ftreatment 1 – in particu\far physica\f, sexua\f and emotiona\f abuse by parents and caregivers – increases the \fike\fihood during ado\fescence and adu\fthood of both intimate partner and sexua\f vio\fence. Preventing chi\fd ma\ftreatment can therefore he\fp prevent both forms of vio- \fence – an association high\fighted in Chapter 3 as part of the review of evidence-based prevention efforts. However, for more detai\f on the preva\fence, consequences and causes of chi\fd ma\ftreatment (0 –14 years) and on how to prevent it see: Preventing chi\fd ma\f- treatment: a guide to taking action and generating evidence (WHO-ISP\bAN, 2006). 2 In addition, different forms of vio\fence have common under\fying risk factors, which inc\fude certain socia\f and cu\ftura\f norms, socia\f iso\fation, the harmfu\f use of a\fcoho\f and income inequa\fity. Prevention efforts that address these common factors thus have the potentia\f to decrease the occurrence of mu\ftip\fe forms of vio\fence.

\f.2 How prevalent are intimate partner and se\bual violence?

The preva\fence of intimate partner vio\fence and sexua\f vio\fence can be measured with reference to both victimization and perpetration. Owing to the sma\f\f proportion of cases recorded in routine\fy co\f\fected statistics from victim care faci\fities and the po\fice, both victimization and perpetration are most accurate\fy measured through popu\fation- based surveys that count se\ff-reports ( Chapter 4 of this document discusses data co\f- \fection in more detai\f). To date, few popu\fation-based surveys outside North America have measured the perpetration of intimate partner vio\fence and sexua\f vio\fence (but see Jewkes et a\f. 2009; Phi\fpart et a\f., 2009), and the comparabi\fity of their findings is unc\fear. Furthermore, \fow disc\fosure rates may compromise the findings of popu\fation- based perpetration surveys. This chapter thus focuses on victimization findings, which are more internationa\f\fy avai\fab\fe, more readi\fy comparab\fe and probab\fy \fess subject to \fow disc\fosure rates.

The WHO Mu\fti-country study on \bomen’s hea\fth and domestic vio\fence against \bomen (Garcia-Moreno et a\f., 2005) provides a comprehensive picture of the patterns of inti- mate partner vio\fence and sexua\f vio\fence victimization in \fow- and midd\fe-income settings. Over 24 000 women between the ages of 15 and 49 were interviewed in rura\f and urban areas in 10 countries. Its key findings inc\fude:

l between 1 and 21% of those interviewed reported experiencing chi\fd sexua\f abuse under the age of 15 years; l physica\f abuse by a partner at some point in \fife up to 49 years of age was reported by 13 – 61% of interviewees across a\f\f study sites; l sexua\f vio\fence by a partner at some point in \fife up to 49 years of age was reported by 6 –59% of interviewees; and 1 Chi\fd ma\ftreatment – refers to the physica\f and emotiona\f mistreatment, sexua\f abuse, neg\fect and neg\figent treatment of chi\fdren, as \be\f\f as to their commercia\f or other exp\foitation. It occurs in many different settings.

T h e p e r p et r a t o r s of c h i\f d m a \ft r e a t m e nt m ay b e p a r e nt s a n d oth e r fa m i\fy m e m b e r s ; c a r e g i v e r s ; f r i e n d s ; a c q u a i nt - ances; strangers; others in authority – such as teachers, so\fdiers, po\fice officers and c\fergy; emp\foyers; hea\fth care \borkers; and other chi\fdren ( W HO-ISP\bA N, 2006).

2 To obtain this guide, p\fease visit: www.who.int/vio\fence_injury_prevention /vio\fence /activities /chi\fd_ ma\ftreatment/en/index.htm\f 13 c\faPter 1. t\fe nature, magnitude and conse\buences of intimate Partner and sexual violence l sexua\f vio\fence by a non-partner any time after 15 and up to 49 years of age was reported by 0.3 –11.5% of interviewees.

Ta b l e \f summarizes the reported preva\fence rates of physica\f and sexua\f vio\fence against women in the study countries. These findings indicate that physica\f and sexua\f vio\fence frequent\fy co-occur within the context of intimate partner re\fationships, and high\fight the differences in preva\fence both between and within different countries. The rates of physica\f and/or sexua\f vio\fence by an intimate partner ranged from 15% in Japan to approximate\fy 70 % in Ethiopia and Peru, with most sites reporting rates of between 29 and 62% .

TABLE 1 Physical and se\bual violence against women by an intimate partner Site Physical \fiolenceSexual \fiolencePhysical or sexual \fiolence or both \b\fer (%) Current (%) \b\fer (%) Current (%) \b\fer (%) Current (%) Bangladesh City 3 7. 719.03 7. 420.2 53.430.2 Bangladesh \frovince 41.715.849.7 24. 2 61.731.9 Brazil City 2 7. 28.310.1 2.828.9 9. 3 Brazil \frovince 33.812.914.3 5.636 .9 14.8 Ethiopia \frovince 48.729.058.644.4 70.953.7 Japan City 12.93.16.2 1.315.4 3.8 Namibia City 30.615.916.5 9.135.9 19. 5 \feru City 48.616 .922.5 7. 151.2 19. 2 \feru \frovince 61.024.8 46.7 22.969.0 34.2 Samoa 40.51 7.919. 5 11.546.1 22.4 Serbia & Montenegro City 22.83.26.3 1.123.7 3.7 Thailand City 22.97.929.9 1 7. 141.121.3 Thailand \frovince 33.813.428.9 15.64 7. 4 22.9 Tanzania City 32.914.823.0 12.841. 3 21.5 Tanzania \frovince 46.7 18.7 30.7 18.355.9 29.1 Source: Garcia-Moreno et al. (2005) The study a\fso found that reported \feve\fs of sexua\f vio\fence by non-partners from the age of 15 years varied from under 1% in Ethiopia and Bang\fadesh (where a majority of women are married by age 15 years) to 10 –12% in Peru, Samoa and urban Tanza- nia. In most cases on\fy one perpetrator was mentioned, usua\f\fy an acquaintance or a stranger. In South Africa, a survey of over 1300 women in three provinces found that women’s first episode of rape was perpetrated in 42.5% of cases by a stranger, 20.8% by an acquaintance, 9.4% by someone from schoo\f, 8.5% by a re\fative, 7.5% by a partner and 11.3% by others ( Jewkes et a\f., 1999).

Information from the United States Nationa\f Vio\fence Against Women Survey (Tjaden & Thoennes, 2006) shows that most rape victims in the United States know their rapist.

Among a\f\f fema\fe victims identified by the survey, 16.7% were raped by a stranger, and 43% of a\f\f fema\fe victims were raped by a current or former intimate partner. 14 Preventing intimate Partner and sexual violence against women Prevalence in adolescence In the WHO Mu\fti-country study on \bomen’s hea\fth and domestic vio\fence against \bomen , 3 –24% of women reported that their first sexua\f experience was forced, and that for a majority of respondents this occurred during ado\fescence (Garcia-Moreno et a\f., 2005).

In 10 of the 15 settings investigated, over 5% of women reported that their first sexua\f experience was forced, with more than 14% reporting forced first sex in Bang\fadesh, Ethiopia, Peru (province) and Tanzania. A Nigerian study found that 45% of fema\fes and 32% of ma\fes aged 12–21 years reported having had forced sexua\f intercourse (S\fap, 2003). In South Africa, a survey of over 280 000 schoo\f pupi\fs showed that up to the age of 15 years around 9 % of both gir\fs and boys reported forced sex in the past year, rising to 13% for ma\fes and 16% for fema\fes by age 19 years (\bIETafrica, 2002).

Reviews of North American studies conc\fude that the preva\fence of \fifetime, past year, and current physica\f and sexua\f dating vio\fence 1 varies significant\fy across studies – ranging from 9 – 49% (G\fass et a\f., 2003). This makes conc\fusions about the preva\fence of teen dating vio\fence premature (Hickman, Jaycox & Aronoff, 2004). Internationa\f\fy, popu\fation-based studies of dating vio\fence are few but do suggest that this affects a substantia\f proportion of the youth popu\fation. For instance, one South African study (Swart, 2002) of 928 ma\fes and fema\fes aged 13 –23 years found that 42% of fema\fes and 38% of ma\fes reported being a victim of physica\f dating vio\fence at some point in their \fifetime. In Ethiopia, near\fy 16% of 1378 ma\fe co\f\fege students reported physica\f\fy abusing an intimate partner or non-partner, and 16.9 % reported perpetrating acts of sexua\f vio\fence (Phi\fpart et a\f., 2009). Prevalence in adulthood Popu\fation-based studies from various countries indicate that 10 – 69 % of women aged 15 – 49 years experience physica\f abuse by a ma\fe intimate partner at \feast once in their \fifetime (Heise, E\f\fsberg & Gottemoe\f\fer, 1999; Heise & Garcia-Moreno, 2002) whi\fe 6 – 47% of women report attempted or actua\f forced sex by an intimate partner in their \fifetime ( Jewkes, Sen & Garcia-Moreno, 2002). The WHO Mu\fti-country study on \bom- en’s hea\fth and domestic vio\fence against \bomen fo u n d t h at b e t we e n 6 % a n d 59 % of wo m e n reported experiencing sexua\f vio\fence by an intimate partner in their \fifetime, with the figures for most sites fa\f\fing between 10 % and 50 % (Garcia-Moreno et a\f., 2005).

Popu\fation surveys in high-income countries (HI\b) show that a significant propor- tion of men a\fso report experiencing physica\f abuse from an intimate partner (see for examp\fe, Mirr\fees-B\fack, 1999) a\fthough not necessari\fy with the same physica\f and emotiona\f consequences as in women. Studies that asked for more detai\f about fema\fe- to-ma\fe intimate partner vio\fence genera\f\fy found that men experienced vio\fence \fess frequent\fy, were \fess \fike\fy to be harmed by physica\f vio\fence perpetrated by women, and did not report \fiving in fear of their partner. Additiona\f\fy, a proportion of the vio- \fence experienced by ma\fe partners may be a direct resu\ft of women attempting to defend themse\fves (Wi\f\fiams et a\f., 2008).

1 Dating vio\fence refers to physica\f or sexua\f vio\fence occurring in the context of a “dating re\fationship”. A dating re\fationship is a term that frequent\fy refers in Europe and the United States to neither a mar- riage nor a \fong-term cohabiting re\fationship that occurs during ado\fescence and young adu\fthood.

These range from casua\f first encounters to \fonger-term sexua\f partnerships. In Asia and other parts of the wor\fd where marriage often takes p\face at a much younger age, the phenomenon of dating vio\fence is rare and intimate partner vio\fence begins at an ear\fier age. 15 Whi\fe most research has focused on intimate partner vio\fence in heterosexua\f re\fation- ships, studies have shown that it a\fso occurs in same-sex re\fationships. Some findings suggest that the rates of intimate partner vio\fence in same-sex re\fationships are simi\far to those in heterosexua\f re\fationships (Shipway, 2004), whi\fe others indicate higher rates of vio\fence in same-sex re\fationships. This may be due to increased \feve\fs of stress resu\fting from societa\f prejudice, and to the higher rates of a\fcoho\f and substance mis- use reported in same-sex re\fationships. Additiona\f\fy, due to the fear of discrimination, many peop\fe in vio\fent same-sex re\fationships may not seek he\fp (Shipway, 2004).

\f.3 What are the consequences of intimate partner and se\bual violence? \f Intimate partner and sexua\f vio\fence have a significant impact upon individua\fs, fami- \fies, communities and wider society. At an individua\f \feve\f, vio\fence occurring during the younger years – particu\far\fy chi\fd sexua\f abuse – may affect that individua\f and their fami\fy for the rest of their \fives. This can \fead to negative consequences in many spheres of \fife, inc\fuding educationa\f and economic under-performance, unsafe sexua\f practices, reduced abi\fity to bond as part of parenthood, increased uptake of hea\fth-risk behaviours (such as the harmfu\f use of a\fcoho\f and i\f\ficit drugs) and the perpetration of intimate partner and sexua\f vio\fence.

\fealth-risk behaviours and outcomes associated with sexual violence in childhood and adolescence A history of sexua\f abuse in chi\fdhood and ado\fescence has consistent\fy been found to be significant\fy associated with increased hea\fth risks and hea\fth-risk behaviours in both ma\fes and fema\fes (Mangio\foi, 2009). Many of these have been identified as risk factors for both sexua\f vio\fence victimization and perpetration in adu\fthood (see Chap- ter 2 for a fu\f\f discussion of the risk and protective factors for intimate partner vio\fence and sexua\f vio\fence). For examp\fe, a meta-ana\fysis of the preva\fence of chi\fd sexua\f abuse and its \fifetime hea\fth consequences showed that chi\fd sexua\f abuse contributes significant\fy to depression, a\fcoho\f and drug use and dependence, panic disorder, post- traumatic stress disorder and suicide attempts (Andrews et a\f., 2004). More recent\fy, a study of over 20 000 schoo\fchi\fdren aged 13 –15 years in Namibia, Swazi\fand, Uganda, Zambia and Zimbabwe (Brown et a\f., 2009) found that 23% reported having experi- enced sexua\f vio\fence (physica\f\fy forced to have sexua\f intercourse) at some point in their \fives. Such experiences were moderate\fy-to-strong\fy associated with poor menta\f hea\fth, suicida\f ideation, cigarette use, a\fcoho\f or drug misuse, mu\ftip\fe sexua\f partners and a history of a sexua\f\fy transmitted infection.

In a nationa\f survey in Swazi\fand that examined the preva\fence and circumstances of sexua\f vio\fence against gir\fs, some 33% of respondents reported experiencing an incident of sexua\f vio\fence before they reached 18 years of age. Sexua\f vio\fence was associated with significant\fy increased probabi\fity of reporting ever fee\fing depressed, thoughts of suicide, attempted suicide, unwanted pregnancy, pregnancy comp\fications or miscarriages, sexua\f\fy transmitted diseases, difficu\fty s\feeping and a\fcoho\f consump- 1 Most studies that inform the fo\f\fowing sections are cross-sectiona\f, and it is difficu\ft to be sure of the nature and direction of the interp\fay between intimate partner and sexua\f vio\fence and their associated hea\fth-risk behaviours and outcomes. For examp\fe, increased drinking increases the risk of experienc- ing or perpetrating sexua\f vio\fence – but at the same time, having previous\fy experienced sexua\f vio- \fence potentia\f\fy increases the risk of drinking. c \faPter 1. t\fe nature, magnitude and conse\buences of intimate Partner and sexual violence 16 Preventing intimate Partner and sexual violence against women tion. \bhi\fdhood sexua\f vio\fence was, however, not associated with increased probabi\fity of women reporting sexua\f vio\fence at 18 –24 years of age (Reza et a\f., 2009).

Additiona\f\fy, forced sexua\f initiation, intimate partner vio\fence and/or sexua\f vio\fence appear to increase the risk of pregnancy in ear\fy ado\fescence. In South Africa it was found that pregnant ado\fescents were over twice as \fike\fy to have a history of forced sexua\f initiation as non-pregnant ado\fescents ( Jewkes et a\f., 2001). Simi\far findings in the United States have a\fso been reported (Si\fverman, Raj & \b\fements, 2004).

\fealth-risk behaviours and outcomes associated with intimate partner and/or sexual violence in adulthood Intimate partner vio\fence and sexua\f vio\fence against women in adu\fthood can \fead direct\fy to serious injury, disabi\fity or death. They can a\fso \fead indirect\fy to a variety of hea\fth prob\fems, such as stress-induced physio\fogica\f changes, substance use and \fack of ferti\fity contro\f and persona\f autonomy as is often seen in abusive re\fationships.

\bompared to their non-abused peers, abused women have higher rates of unintended pregnancies and abor tions; sexua\f\fy transmitted infections, inc\fuding HIV; and menta\f disorders such as depression, anxiety, s\feep and eating disorders. When this vio\fence occurs during pregnancy, it is associated with adverse pregnancy events – such as mis- carriage, pre-term births and sti\f\fbirths (Ahmed, Koenig & Stephenson, 2006; As\fing- Monemi, Tabassum & Persson, 2008; Boy & Sa\fihu, 2004; \bampbe\f\f, 2002; \bampbe\f\f et a\f., 2008; Dunk\fe et a\f., 2004; P\fichta & Fa\fik, 2001; and Vos et a\f., 2006).

Intimate partner vio\fence not on\fy affects the women invo\fved, but may a\fso damage the hea\fth and we\f\f-being of chi\fdren in the fami\fy. This is in part due to increased rates of depression and traumatic stress in the abused mothers, and the destructive effects of intimate partner vio\fence on the qua\fity of their attachment and parenting capacities.

One review of studies examining the presence of both chi\fd ma\ftreatment and intimate partner vio\fence found that they occurred during the same period in 45 –70 % of stud- ies (Ho\ft, Buck\fey & Whe\fan, 2008). In the majority of such cases, a chi\fd witnessing intimate partner vio\fence appeared to precede the subsequent ma\ftreatment of chi\fdren by fami\fy members (McGuigan & Pratt, 2001). Studies have shown that the chi\fdren of abused mothers have \fower rates of immunization and higher rates of diarrhoea\f disease, and are more \fike\fy to die before the age of five years (As\fing-Monemi, Tabassum & Persson, 2008; Si\fverman et a\f., 2009).

Findings from a number of reviews show that witnessing intimate partner vio\fence can a\fso negative\fy affect the norma\f deve\fopment of chi\fdren in the fami\fy. For examp\fe, one meta-ana\fysis conc\fuded that chi\fdren’s exposure to vio\fence between parents is signifi - cant\fy corre\fated with chi\fd prob\fems in the areas of socia\f, emotiona\f, behavioura\f, cogni - t ive a nd genera \f hea \ft h f u nc t ion i ng ( K it zma n n et a \f., 20 03). However, not a \f \f ch i \fd ren a re simi\far\fy affected and a significant percentage of chi\fdren in some review studies showed no negative deve\fopmenta\f prob\fems despite witnessing repeated vio\fence (Ede\fson, 1999; Wo\ffe et a\f., 2003). This suggests that each chi\fd wi\f\f experience adu\ft intimate partner vio\fence in a unique way depending upon a variety of factors. Such factors are \fike\fy to inc\fude whether there was a\fso physica\f abuse of the chi\fd, the chi\fd’s sex and age, the time since exposure to vio\fence and the chi\fd’s re\fationship with adu\fts in the home.

Studies from severa\f countries have found that HIV-positive women report higher rates of intimate partner vio\fence (Dunk\fe et a\f., 2004; Maman et a\f., 2000) and there is increasing evidence that HIV risk is \finked to \fifetime exposure to vio\fence in comp\fex 17 ways (\bampbe\f\f et a\f., 2008). Rape is a potentia\f cause of direct infection with HIV for some women, yet even in high-preva\fence settings the \fow HIV transmission risk dur- ing a sing\fe sexua\f act makes it un\fike\fy that rape resu\fts in a substantia\f proportion of popu\fation-\feve\f HIV cases. Vio\fence and gender inequa\fity are more \fike\fy to increase HIV risk through indirect pathways, inc\fuding chronica\f\fy abusive re\fationships where women are repeated\fy exposed to the same individua\f, and are unab\fe to negotiate con- dom use for safer sex (WHO/ U NAIDS, 2010).

Qua\fitative research suggests that the intersections of HIV/AIDS, gender inequa\fity and gender-based vio\fence may invo\fve notions of mascu\finity that are predicated on the contro\f of women, and which va\fue ma\fe strength and toughness ( Jewkes & Morre\f\f, 2010). These idea\fs readi\fy trans\fate into sexua\f\fy risky behaviours, sexua\f predation and other acts of vio\fence against women as we\f\f as behaviours which increase the risk of acquiring HIV. Exposure to vio\fence, inc\fuding the contro\f\fing behaviour of a partner, is associated with high-risk sexua\f behaviour (such as mu\ftip\fe and concurrent sexua\f partnerships); higher number of overa\f\f partners; \fower \feve\fs of condom use; increased substance use and sex whi\fe intoxicated; and increased participation in transactiona\f sex as we\f\f as commercia\f sex work. Emerging evidence from India and South Africa indi- cates that men who perpetrate vio\fence are more \fike\fy to be infected with HIV (Decker et a\f., 2009; Jewkes et a\f., 2009).

\f.4 Key messages Pre\falence n The prevalence of intimate partner violence reported by girls and women varies markedly between and within countries, with higher rates tending to occur in lower-income countries.

n In some parts of the world, particularly where young people can have relationships outside of marriage, “dating violence” occurs frequently.

n Sexual violence affects children of both sexes and adults, particularly women.

Consequences n Sexual violence – particularly during childhood – is associated with increased health-risk behaviours – including smoking, drug and alcohol misuse and risky sexual behaviours.

n Sexual abuse and adverse experiences in childhood can cause lifelong health inequalities through increased risk behaviours.

n Intimate partner violence harms the physical and mental health of women and their children – women who have experienced intimate partner violence are more likely to report poor or very poor health, emotional distress and attempted suicide; and their children tend to have poorer health and educational outcomes.

n Intimate partner violence in pregnancy increases the likelihood of abortion, miscarriage, stillbirth, pre-term delivery and low birth weight.

n Intimate partner violence and sexual violence are both associated with increased vulnerability to HIV.

n Children who grow up in families where there is intimate partner violence suffer a range of behavioural and emotional disturbances that can be associated with the perpetration or experiencing of violence later in life. c \faPter 1. t\fe nature, magnitude and conse\buences of intimate Partner and sexual violence CHA\fTER 2 Risk and protective factors for intimate partner and se\bual violence 2 .\f The ecological model of violence As noted by the United States \benters for Disease \bontro\f and Prevention (\bD\b, 2004), in order to prevent sexua\f vio\fence it is crucia\f to understand the circumstances, and the risk and protective factors, that influence its occurrence. Many different theo- retica\f mode\fs attempt to describe the risk and protective factors for intimate partner and sexua\f vio\fence, inc\fuding those based upon bio\fogica\f, psycho\fogica\f, cu\ftura\f and gender equa\fity concepts. Each of these mode\fs contributes to a better understanding of intimate partner and sexua\f vio\fence and he\fps to bui\fd programmes that aim to reduce modifiab\fe risk factors and strengthen protective factors. Risk factors increase the \fike\fi- hood of someone becoming a victim and/or perpetrator of intimate partner and sexua\f vio\fence and their reduction shou\fd therefore be a key target of prevention efforts, as we\f\f as an integra\f concept in programme monitoring and eva\fuation efforts. Simi\far\fy, protective factors, which buffer against the risk of becoming a victim and/or perpetrator of intimate partner and sexua\f vio\fence, may need to be fostered – inc\fuding through structura\f and other interventions for achieving gender equa\fity and the empowerment of women.

In this document, WHO has chosen to use the “eco\fogica\f mode\f” ( Figure \f) as pre- sented in the Wor\fd report on vio\fence and hea\fth (Dah\fberg & Krug, 2002) because it a\f\fows for the inc\fusion of risk and protective factors from mu\ftip\fe domains of influ- ence. Thus, if there is evidence from psycho\fogica\f mode\fs on individua\f risk factors and from gender mode\fs on societa\f risk factors, this can be incorporated in the same eco\fogica\f mode\f.

FIGURE 1 The ecological model Societal Community Relationship Indi\fidual 18 19 c\faPter 2. risk and Protective factors for intimate Partner and sexual violence The eco\fogica\f mode\f organizes risk factors according to the fo\f\fowing four \feve\fs of influence:

l Individua\f: inc\fudes bio\fogica\f and persona\f history factors that may increase the \fike\fihood that an individua\f wi\f\f become a victim or perpetrator of vio\fence.

l Re\fationship: inc\fudes factors that increase risk as a resu\ft of re\fationships with peers, intimate partners and fami\fy members. These are a person’s c\fosest socia\f cir- c\fe and can shape their behaviour and range of experiences.

l Community: refers to the community contexts in which socia\f re\fationships are embedded – such as schoo\fs, workp\faces and neighbourhoods – and seeks to identify the characteristics of these settings that are associated with peop\fe becoming victims or perpetrators of intimate partner and sexua\f vio\fence.

l Societa\f: inc\fudes the \farger, macro-\feve\f factors that influence sexua\f and intimate partner vio\fence such as gender inequa\fity, re\figious or cu\ftura\f be\fief systems, soci- eta\f norms and economic or socia\f po\ficies that create or sustain gaps and tensions between groups of peop\fe.

Bui\fding such a mode\f offers a framework for understanding the comp\fex interp\fay of a\f\f the factors that influence intimate partner and sexua\f vio\fence, and can therefore provide key points for prevention and intervention (Dah\fberg & Krug, 2002).

The eco\fogica\f mode\f a\fso supports a comprehensive pub\fic hea\fth approach that not on\fy addresses an individua\f’s risk of becoming a victim or perpetrator of vio\fence, but a\fso the norms, be\fiefs and socia\f and economic systems that create the conditions for intimate partner and sexua\f vio\fence to occur. At the core of the approach is a strong emphasis on the mu\ftip\fe and dynamic interactions among risk factors within and between its different \feve\fs. For examp\fe, structura\f inequa\fities between women and men, socia\f const r uct ions of mascu\finit y and gender norms are risk factors for int i - mate partner and sexua\f vio\fence that wou\fd primari\fy be situated at the societa\f \feve\f of the mode\f. \b\fear\fy, however, they a\fso manifest themse\fves within other \feve\fs – for examp\fe, in communities and re\fationships – and are \fike\fy to be \finked with other risk factors such as the witnessing of vio\fence between parents and a\fcoho\f abuse by ma\fe perpetrators.

Using the eco\fogica\f mode\f a\fso he\fps to promote the deve\fopment of cross-sectora\f prevention po\ficies and programmes by high\fighting the \finks and interactions between different \feve\fs and factors. As a resu\ft, when designing comprehensive approaches to prevent intimate partner and sexua\f vio\fence, the embedding of effective strategies into mainstream programmes addressing such issues wi\f\f increase both their re\fevance and sustainabi\fity.

Viewing the ways in which these risk factors come together and influence patterns of behaviour throughout the \fife-course provides insights into the key points at which interventions to break the cyc\fe shou\fd be imp\femented ( Chapter 3).

2.2 Identifying risk factors The identification of risk factors is critica\f\fy important for informing strategies and programmes to ame\fiorate or buffer against risk – and u\ftimate\fy to guide prevention po\ficy. The causes of intimate partner vio\fence and sexua\f vio\fence are best investigated through the use of \fongitudina\f studies. These studies track peop\fe over time to docu- 20 Preventing intimate Partner and sexual violence against women ment their experiences of such vio\fence and how these experiences re\fate to other fac- tors at various stages of their \fife. Unfortunate\fy, few such studies exist – so much of the information in this section is derived from cross-sectiona\f popu\fation surveys. These are good at providing a snapshot of how frequent\fy something occurs and its associated factors, but they usua\f\fy cannot provide information on whether an observed associa- tion actua\f\fy “caused” a particu\far outcome.

This chapter draws on two main sources of information: first, the Wor\fd report on vio- \fence and hea\fth chapters on intimate partner vio\fence (Heise & Garcia-Moreno, 2002) and sexua\f vio\fence ( Jewkes, Sen & Garcia-Moreno, 2002) which reviewed the \fiterature unti\f 2002; and, second, a systematic review of more recent peer-reviewed \fiterature on risk and protective factors associated with either the perpetration or experiencing of intimate partner vio\fence, sexua\f vio\fence or both.

Most of the \fiterature is from high-income countries (HI\b), and it is unc\fear whether factors identified in HI\b a\fso app\fy to \fow- and midd\fe-income countries (LMI\b) due to differences in economies, eco\fogies, histories, po\fitics and cu\ftures. Any primary studies from LMI\b that identified factors associated with intimate partner vio\fence and/or sexua\f vio\fence were a\fso inc\fuded in the review of more recent \fiterature.

As a resu\ft of this process, over 50 risk factors were identified for intimate partner vio- \fence and/or sexua\f vio\fence – most at the individua\f and fami\fy/re\fationship \feve\fs. The paucity of risk factors that have current\fy been identified at the community and societa\f \feve\fs is probab\fy due to a \fack of research on risk factors at these \feve\fs rather than reflecting a true absence of risk factors.

2.3 Risk factors associated with both intimate partner violence and se\bual violence Ta b l e 2 \fists the risk factors identified for the perpetration by men and the experiencing by women of both intimate partner and sexua\f vio\fence. The risk factors are organized according to the individua\f, re\fationship, community and societa\f \feve\fs of the eco\fogica\f mode\f shown in Figure \f. Some of these risk factors can be modified (for examp\fe, the harmfu\f use of a\fcoho\f) whi\fe others cannot (for examp\fe, sex and age). Those factors that have the strongest reported effect or are consistent\fy reported across studies are high\fighted in bo\fd in the tab\fe.

Individual-level factors Young age Young age appears to be a risk factor for being either a perpetrator or victim of intimate partner vio\fence, and a victim of sexua\f vio\fence. Young age has consistent\fy been found to be a risk factor for a man committing physica\f vio\fence against a partner (B\fack et a\f., 2001) and for a woman experiencing intimate partner vio\fence (Harwe\f\f & Spence, 2000 ; Romans et a\f., 2007; Vest et a\f., 2002). Young women have been found to be more at risk of rape than o\fder women ( Jewkes, Sen & Garcia-Moreno, 2002). Accord- ing to data from justice systems and rape-crisis centres in \bhi\fe, Ma\faysia, Mexico, Papua New Guinea, Peru and the United States, between one third and two thirds of a\f\f victims of sexua\f assau\ft are aged 15 years or under, a\fthough this a\fso may reflect reporting bias. \bertain forms of sexua\f vio\fence, for instance, are very c\fose\fy associ- ated wit h young age, in par t icu\far vio\fence tak ing p\face in schoo\fs and co\f\feges, and t he 21 trafficking of women for sexua\f exp\foitation. Whether young age is a\fso a risk factor for the perpetration of sexua\f vio\fence by ma\fes is \fess c\fear. In South Africa, a generationa\f effect was suggested by findings that men aged 20 – 40 years were more \fike\fy to have raped than younger and o\fder men ( Jewkes et a\f., 2009).

education The re\fationship between individua\f educationa\f attainment and intimate partner vio- \fence and sexua\f vio\fence is comp\fex. Low \feve\f of education is however the most con- sistent factor associated with both the perpetration and experiencing of intimate partner vio\fence and sexua\f vio\fence across studies (Ackerson et a\f., 2008; Boy & Ku\fczycki, 2008; Boy\fe et a\f., 2009; Brown et a\f., 2006; \bhan, 2009; Da\fa\f, Rahman & Jansson, 2009; Gage, 2006; Jeyasee\fan et a\f., 2004; Johnson & Das, 2009; Koenig et a\f., 2006; Martin, Taft & Resick, 2007; Tang & Lai, 2008). For examp\fe, women who report \fower \feve\fs of education (primary or none) have a 2 to 5-fo\fd increased risk of intimate partner vio\fence compared to higher-educated women (Ackerson et a\f., 2008; Boy & Ku\fczycki, 2008; Da\fa\f, Rahman & Jansson, 2009; Koenig et a\f., 2004; Martin, Taft & Resick, 2007; Tang & Lai, 2008). Simi\far\fy, studies have found that \fower-educated ma\fes were 1.2 to 4.1 times more \fike\fy to perpetrate intimate partner vio\fence than higher-educated men (Ackerson et a\f., 2008; Da\fa\f, Rahman & Jansson, 2009).

A higher \feve\f of education may act as a protective factor, since women with a higher \feve\f of education, or married coup\fes with re\fative\fy equiva\fent education \feve\fs, report \fower \feve\fs of intimate partner vio\fence. Lower educationa\f attainment reduces a wom- c \faPter 2. risk and Protective factors for intimate Partner and sexual violence TABLE 2 Risk factors for both intimate partner violence and se\bual violence Perpetration by men Victimization of women Ind IVIdu Al l EVEl D EMOGRA\fHICS • Low income • Low education D EMOGRA\fHICS • Young age • Low education • Separated/divorced marital status E X\fOSURE TO CHILD MALTREATMENT • Sexual abuse • Intra-parental violence E X\fOSURE TO CHILD MALTREATMENT • Intra-parental \fiolence M E N TA L DISORDER • Antisocial personality M E N TA L DISORDER • Depression S U B S TA N C E USE • Harmful use of alcohol • Illicit drug use S U B S TA N C E USE • Harmful use of alcohol • Illicit drug use • Acceptance of violence • Acceptance of violence REl AT I O nSHIP lEVE l • Multiple partners/infidelity • Low resistance to peer pressure COMM unITy l EVE l • \beak community sanctions • \foverty • \beak community sanctions • \foverty S OCIETAl l EVEl • Traditional gender norms and social norms supportive of violence • Traditional gender norms and social norms supportive of violence 22 Preventing intimate Partner and sexual violence against women an’s exposure and access to resources, increases the acceptance of vio\fence and main- tains unequa\f gender norms.

It wou\fd however be premature to assume that the re\fationship between educationa\f attainment and intimate partner vio\fence is the same regard\fess of the type of intimate partner vio\fence invo\fved. In contrast to the above studies, which inc\fuded physica\f vio- \fence in their definition of intimate partner vio\fence, F\fake (2005) found that women with a higher \feve\f of education were at increased risk of sexua\f intimate partner vio- \fence. More research is needed on how educationa\f attainment is associated with the different types of intimate partner vio\fence.

A\fthough there is a paucity of evidence, a re\fationship between \fower fema\fe educationa\f \feve\f and their experiencing sexua\f vio\fence has a\fso been indicated. One study (Brown et a\f., 2006) showed that women with primary or no education were twice as \fike\fy to experience sexua\f vio\fence by a non-intimate partner compared to women with a sec- ondary \feve\f of education or higher.

exposure to child maltreatment One factor consistent\fy cited across countries as a risk factor for both the experiencing and perpetration of intimate partner vio\fence and sexua\f vio\fence is chi\fd ma\ftreatment.

A we\f\f-conducted systematic review summarized evidence from 10 studies to show that exposure to vio\fence during chi\fdhood increased the \fike\fihood of intimate part- ner vio\fence perpetration among men by 3 or 4-fo\fd, compared to men without chi\fd- hood exposure to vio\fence (Gi\f-Gonza\fez et a\f., 2007). Another meta-ana\fysis found that exposure to any chi\fdhood sexua\f abuse (but not physica\f abuse a\fone) increased ma\fe perpetration of sexua\f vio\fence towards women more than three-fo\fd ( Jespersen, La\fumiere & Seto, 2009). A number of primary studies in LMI\b found that chi\fdhood exposure to vio\fence (particu\far\fy intra-parenta\f vio\fence and sexua\f abuse) was posi- tive\fy associated with the experiencing of intimate partner vio\fence and sexua\f vio\fence by fema\fes (Martin, Taft & Resick, 2007; Söchting, Fairbrother & Koch, 2004; Vung & Krantz, 2009). Such exposure to vio\fence during chi\fdhood may increase the \fike\fi- hood of vio\fence acceptance either as a victim or perpetrator in future partnerships and high-risk situations.

antisocial personality Severa\f studies examined persona\fity characteristics and their association with the per- petration of intimate partner vio\fence and sexua\f vio\fence. Three narrative reviews (two from a high-income and another from a midd\fe-income country) reported a consistent association between antisocia\f persona\fity disorders and re\fated characteristics (such as impu\fsivity and \fack of empathy) and the perpetration of intimate partner vio\fence or sexua\f vio\fence. These studies suggest that perpetrators with these characteristics are more \fike\fy to disregard socia\f norms and have a tendency to become aggressive and impu\fsive, resu\fting in the perpetration of intimate partner vio\fence or sexua\f vio\fence (Abbey et a\f., 2004; \bhan, 2009; Marsha\f\f, Panuzio & Taft, 2005). More studies are needed to empirica\f\fy support a causa\f re\fationship between persona\fity disorders and the perpetration of intimate partner vio\fence and/or sexua\f vio\fence. 23 \farmful use of alcohol Harmfu\f use of a\fcoho\f and i\f\ficit drug use are other common\fy cited risk factors asso- ciated with the experiencing and perpetration of intimate partner vio\fence and sexua\f vio\fence (for examp\fe, see Graham et a\f., 2008). A\fthough severa\f studies have found an association between intimate partner vio\fence or sexua\f vio\fence and i\f\ficit drug use, most of the \fiterature has focused on the re\fationship between harmfu\f use of a\fcoho\f and intimate partner vio\fence or sexua\f vio\fence. Harmfu\f use of a\fcoho\f was strong\fy associated with the perpetration of intimate partner vio\fence in severa\f of the reviews, which inc\fuded studies from LMI\b (Abrahams et a\f., 2004; Da\fa\f, Rahman & Jans- son, 2009; F\fake, 2005; Gage, 2006; Gi\f-Gonza\fez et a\f., 2006; Johnson & Das, 2009; Koenig et a\f., 2004; Marsha\f\f, Panuzio & Taft, 2005; Martin, Taft & Resick, 2007; Ramiro, Hassan & Peedicayi\f, 2004; Tang & Lai, 2008; Testa, 2004). One systematic review poo\fed the resu\fts of 11 studies and found that harmfu\f use of a\fcoho\f was associ- ated with a 4.6 times increased risk of exposure to intimate partner vio\fence compared to mi\fd or no a\fcoho\f use (Gi\f-Gonza\fez et a\f., 2006). Two narrative reviews conc\fuded that an association exists between harmfu\f use of a\fcoho\f and the perpetration of sex- ua\f vio\fence (Abbey et a\f., 2004; Testa, 2004). \bross-sectiona\f studies from different LMI\b report that men who misuse a\fcoho\f are 1.6 to 4.8 times more \fike\fy to perpetrate intimate partner vio\fence (Abrahams et a\f., 2004; Da\fa\f, Rahman & Jansson, 2009; F\fake, 2005; Gage, 2006; Johnson & Das, 2009; Koenig et a\f., 2004; Ramiro, Hassan & Peedicayi\f, 2004).

These findings shou\fd however be interpreted with caution since evidence for a causa\f association between harmfu\f use of a\fcoho\f and vio\fence is weak (Gi\f-Gonza\fez et a\f., 2006). The ro\fe of a\fcoho\f use in the experiencing of intimate partner vio\fence and sexua\f vio\fence is \fess c\fear. Severa\f narrative reviews found a weak re\fationship or no re\fationship at a\f\f between a\fcoho\f use and the experiencing of intimate partner vio\fence or sexua\f vio\fence (Abbey et a\f., 2004; Gutierres & Van Puymbroeck, 2006; Söchting, Fairbrother & Koch, 2004; Tang & Lai, 2008; Testa, 2004).

acceptance of violence Men and women’s attitudes towards vio\fence are strong\fy corre\fated with exposure to intimate partner vio\fence and sexua\f vio\fence – both in terms of victimization and perpe- tration. Reviews and studies, inc\fuding severa\f from LMI\b, found a strong association between attitudes towards vio\fence and exposure to intimate partner vio\fence or sexua\f vio\fence (Abrahams et a\f., 2004; Boy\fe et a\f., 2009; Gage, 2006; Jewkes et a\f., 2006; Johnson & Das, 2009; Tang & Lai, 2008). Men who be\fieve that it is acceptab\fe to beat their wives have a two-fo\fd risk of intimate partner vio\fence perpetration (Abramson, 2004). This risk increased as acceptance of vio\fence increased. Men who be\fieve that it is a\fways acceptab\fe to beat their wives have a four-fo\fd increased risk of intimate partner vio\fence perpetration compared to a two-fo\fd increased risk among those who be\fieve it is sometimes acceptab\fe to beat their wives ( Johnson & Das, 2009). Women’s accept - ance of vio\fence is a\fso positive\fy associated with the experiencing of intimate partner vio\fence (Boy\fe et a\f., 2009; Uthman, Lawoko & Moradi, 2009). Women and men’s acceptance of intimate partner vio\fence; men’s attitudes towards women as inferior; restrictive gender ro\fes; and dominant patriarcha\f va\fues may a\f\f perpetuate the occur- rence of vio\fence. These attitudes may be transferred across generations through \fearn- ing processes, the media, schoo\fs, and witnessing and experiencing vio\fence throughout \fife – and can therefore be changed. c \faPter 2. risk and Protective factors for intimate Partner and sexual violence 24 Preventing intimate Partner and sexual violence against women Relationshi\f-level factors multiple partners Men who report having mu\ftip\fe sexua\f partners are a\fso more \fike\fy to perpetrate inti- mate partner vio\fence or sexua\f vio\fence. Mu\ftip\fe partnership and infide\fity (as per- ceived by fema\fe partners) were a\fso strong\fy associated with both the perpetration and experiencing of intimate partner vio\fence. (Abrahams et a\f., 2004; \bhan, 2009; Da\fa\f, Rahman & Jansson, 2009; Jewkes et a\f., 2006; Johnson & Das, 2009; Koenig et a\f., 2004; Tang & Lai, 2008; Vung & Krantz, 2009). A\f\f the studies reported a strong association between women’s perceived infide\fity or mu\ftip\fe sexua\f partnerships by their partners and intimate partner vio\fence or sexua\f vio\fence. Estimates ranged from a 1.5-fo\fd (India) to 17.1-fo\fd (South Africa) greater risk of the perpetration of intimate partner vio\fence and sexua\f vio\fence, and a 1.5-fo\fd (Uganda) to 2.4-fo\fd (Viet Nam) greater risk of experiencing intimate partner vio\fence ( Jewkes et a\f., 2006; Koenig et a\f., 2004; Vung & Krantz, 2009). It is thought that these men may seek out mu\ftip\fe sexua\f partners as a source of peer status and se\ff-esteem, re\fating to their fema\fe part- ners impersona\f\fy and without the appropriate emotiona\f bonding ( Jewkes et a\f., 2006).

A\fso, such men are more \fike\fy to engage in risky behaviours with mu\ftip\fe sexua\f part - ners by refusing to use condoms – thus exposing themse\fves and their partners to an increased risk of HIV infection. In most of the sites covered by the WHO Mu\fti-country study on \bomen’s hea\fth and domestic vio\fence against \bomen (Garcia-Moreno et a\f., 2005) women whose current or most recent partner was vio\fent were more \fike\fy than women in non-vio\fent re\fationships to report at \feast one refusa\f to use a condom.

Community-level factors weak community sanctions against intimate partner and sexual violence The Wor\fd report on vio\fence and hea\fth notes that how a community responds to partner vio\fence may affect the overa\f\f \feve\fs of abuse in that community (Heise & Garcia- Moreno, 2002). In a comparative study of 16 societies with either high or \fow rates of intimate partner vio\fence, \bounts, Brown & \bampbe\f\f (1992) found that societies with the \fowest \feve\fs of intimate partner vio\fence were those that had community sanctions against it, and where abused women had access to sanctuary – either in the form of she\fters or fami\fy support. \bommunity sanctions, or prohibitions, cou\fd take the form either of forma\f \fega\f sanctions or the mora\f pressure for neighbours to intervene if a woman was beaten. The “sanctions and sanctuary” framework suggests the hypothesis that intimate partner vio\fence wi\f\f be highest in societies where the status of women is in a state of transition. Where women have a very \fow status, vio\fence is not needed to enforce ma\fe authority. On the other hand, where women have a high status, they wi\f\f probab\fy have achieved sufficient power co\f\fective\fy to change traditiona\f gender ro\fes.

Intimate partner vio\fence is thus usua\f\fy highest at the transition point, as is the situa - tion in most LMI\b.

The \fike\fihood of sexua\f vio\fence is re\fated to the extent to which be\fiefs in ma\fe superi- ority and ma\fe entit\fement to sex are entrenched in a community, the genera\f to\ferance in the community of sexua\f assau\ft and the strength of sanctions, if any, against perpe- trators ( Jewkes, Sen & Garcia-Moreno, 2002). 25 Poverty Studies from a wide range of settings show that, whi\fe intimate partner vio\fence and sexua\f vio\fence cut across a\f\f socioeconomic groups, women \fiving in poverty are dispro- portionate\fy affected (Heise & Garcia-Moreno, 2002; Jewkes, Sen & Garcia-Moreno, 2002). It is not yet c\fear why poverty increases the risk of these forms of vio\fence – whether it is because of \fow income in itse\ff or because of other factors that accompany poverty, such as overcrowding or hope\fessness. For some men, \fiving in poverty is \fike\fy to generate stress, frustration and a sense of inadequacy for having fai\fed to \five up to their cu\ftura\f\fy expected ro\fe of providers. Poverty may a\fso provide ready materia\f for marita\f disagreements or make it more difficu\ft for women to \feave vio\fent or otherwise unsatisfactory re\fationships. Whatever the precise mechanisms, it is probab\fe that pov- erty acts as a “marker” for a variety of socia\f conditions that combine to increase the risks faced by women. Poor women and gir\fs may be more at risk of rape in the course of their dai\fy tasks than those who are better off, for examp\fe when they wa\fk home on their own from work \fate at night, or work in the fie\fds or co\f\fect firewood a\fone. \bhi\f- dren of poor women may have \fess parenta\f supervision when not in schoo\f, since their mothers may be at work and unab\fe to afford chi\fdcare. The chi\fdren themse\fves may be working and thus vu\fnerab\fe to sexua\f exp\foitation. Poverty forces many women and gir\fs into occupations that carry a re\fative\fy high risk of sexua\f vio\fence, particu\far\fy sex work. It a\fso creates enormous pressures for them to find or maintain jobs, to pursue trading activities and, if studying, to obtain good grades – a\f\f of which render them vu\fnerab\fe to sexua\f coercion from those who can promise these things ( Jewkes, Sen & Garcia-Moreno, 2002).

\bocietal-level factors traditional gender norms and social norms supportive of violence Research studies across cu\ftures have revea\fed a number of societa\f and cu\ftura\f fac- tors that might give rise to higher \feve\fs of vio\fence. For examp\fe, Levinson (1989) used statistica\f ana\fysis of coded ethnographic data from 90 societies to examine the cu\ftura\f patterns of wife beating – exp\foring the factors that consistent\fy distinguish societies where wife beating is common from those where the practice is rare or absent. This ana\fysis indicated that wife beating occurs more often in societies in which men have economic and decision-making power in the househo\fd, where women do not have easy access to divorce and where adu\fts routine\fy resort to vio\fence to reso\fve their conflicts.

Another strong predictor in this study of the frequency of wife beating was the absence of a\f\f-women workgroups. It was suggested that the presence of fema\fe workgroups offers protection from wife beating because they provide women with a stab\fe source of socia\f support and economic independence from their husbands and fami\fies.

One of the most common theories to exp\fain the perpetration and experiencing of inti- mate partner vio\fence and sexua\f vio\fence is the maintenance of patriarchy or ma\fe dominance within a society (Taft, 2009). Patriarcha\f and ma\fe dominance norms reflect gender inequa\fity and inequities at a societa\f \feve\f, and \fegitimize intimate part- ner vio\fence and sexua\f vio\fence perpetrated by men (Russo & Pir\fott, 2006). Whi\fe they are \focated at the societa\f \feve\f, these gender norms p\fay out at the \feve\f of com- munity, re\fationship and individua\f behaviours. Societa\f norms re\fated to gender are be\fieved to contribute to vio\fence against women and gender inequa\fity and other ineq- uities by creating power hierarchies where men are viewed by society as economica\f\fy c\faPter 2. risk and Protective factors for intimate Partner and sexual violence 26 Preventing intimate Partner and sexual violence against women and re\figious\fy superior, and of higher socia\f status compared to women – who are sometimes viewed as a \fiabi\fity (A\fi & Bustamante-Gavino, 2008). As such, men are socia\fized to be\fieve that they are superior to women, shou\fd dominate their partners and endorse traditiona\f gender ro\fes (Taft, 2009). Women’s subordination and submis- sion is then considered to be norma\f, expected, accepted and, in some cases, attrac- tive to men (Russo & Pir\fott, 2006). Women who are more competent or educated are often stigmatized or dis\fiked by society. This gender inequa\fity and ma\fe dominance reduces the opportunities for women to be invo\fved in decision-making at every \feve\f; decreases the resources avai\fab\fe to women; and increases acceptance of the use of vio- \fence against women. Furthermore, it contributes to gender-based inequities in hea\fth and access to hea\fth care; in opportunities for emp\foyment and promotion; in \feve\fs of income; in po\fitica\f participation and representation; and in education. Thus, macro- \feve\f interventions that increase structura\f supports and resources that decrease gender inequa\fity – as we\f\f as interventions to reduce gender inequa\fity at the community and individua\f \feve\fs – may serve to decrease intimate partner vio\fence and sexua\f vio\fence (Smith Fawzi et a\f., 2005).

In summary, there is a need for more research to identify modifiab\fe factors that can influence the perpetration or experiencing of intimate partner vio\fence and sexua\f vio- \fence at both community and societa\f \feve\fs. Potentia\f community-\feve\f factors inc\fude education, the avai\fabi\fity and accessibi\fity of resources, and the readiness of individu- a\fs to use avai\fab\fe community resources. Important societa\f-\feve\f factors inc\fude gen- der norms and other structura\f factors supportive of gender inequa\fity and vio\fence.

2.4 Risk factors associated with intimate partner violence Most factors associated with the perpetration and experiencing of intimate partner vio- \fence identified to date are at the individua\f \feve\f, fo\f\fowed by the re\fationship and com- munity \feve\f; fewer have been identified at the societa\f \feve\f ( Ta b l e 3). Those factors that are most strong\fy and consistent\fy associated with intimate partner vio\fence are high\fighted in bo\fd in the tab\fe, and any not a\fready discussed in Section 2.3 (Ta b l e 2 ) are examined be\fow.

Individual-level factors Past history of victimization or perpetration of violence Women who have previous\fy been abused by intimate or non-intimate partners during adu\fthood are more \fike\fy to experience future intimate partner vio\fence compared to those without prior exposure to vio\fence. For examp\fe, one study from India showed that women who reported previous non-intimate partner vio\fence were 3.8 times more \fike\fy to report intimate partner vio\fence compared to those without previous exposure (Boy\fe et a\f., 2009). A narrative review a\fso found that previous exposure to abuse may contribute to future victimization by changing a woman’s attitude towards vio\fence; decreasing her abi\fity to recognize risk; \fowering her se\ff-esteem; increasing her gui\ft, shame and embarrassment; and reducing her sexua\f assertiveness (Söchting, Fairbroth- er & Koch, 2004). Simi\far\fy, a prior history of perpetration among men was a strong risk factor for intimate partner vio\fence. Men with a history of abusive or vio\fent behav- iour were more \fike\fy to exhibit this behaviour in their future intimate partnerships, especia\f\fy during pregnancy and the postnata\f period (\bhan, 2009; Jewkes et a\f., 2006; Martin, Taft & Resick, 2007; Söchting, Fairbrother & Koch, 2004; Tang & Lai, 2008). 27 c\faPter 2. risk and Protective factors for intimate Partner and sexual violence TABLE 3 Risk factors for intimate partner violence a Perpetration by men Victimization of women Ind IVIdu Al l EVEl D EMOGRA\fHICS • Young age • Low socio-economic status/ income • Low education • Unemployment D EMOGRA\fHICS • Young age • Low socio-economic status/income • Low education • Separated/divorced marital status • \fregnancy E X\fOSURE TO CHILD MALTREATMENT • Intra-parental violence • Sexual abuse • \fhysical abuse E X\fOSURE TO CHILD MALTREATMENT • Intra-parental \fiolence • Sexual abuse M E N TA L DISORDER • Antisocial personality M E N TA L DISORDER • Depression S U B S TA N C E USE • Harmful use of alcohol • Illicit drug use S U B S TA N C E USE • Harmful use of alcohol • Illicit drug use • Acceptance of violence • Acceptance of violence • Past history of being abusive • E\bposure to prior abuse/victimization REl AT I O nSHIP lEVE l • Educational disparity • Educational disparity • Multiple partners/infidelity • Number of children R ELATIONSHI\f qUALIT Y • Marital dissatisfaction/discord • Gender role disputes• Marital duration R ELATIONSHI\f qUALIT Y • Marital dissatisfaction/discord C OMM unITy l EVE l • Acceptance of traditional gender roles • Acceptance of traditional gender roles N EIGHBOURHOOD CHARACTERISTICS • High proportion of poverty • High proportion of unemployment • High proportion of male literacy • Acceptance of violence • High proportion of households that use corporal punishment N EIGHBOURHOOD CHARACTERISTICS • High proportion of poverty • High proportion of unemployment • High proportion of female literacy • Acceptance of violence • Low proportion of women with high level of autonomy • Low proportion of women with higher education • \beak community sanctions • \beak community sanctions S OCIETAl l EVEl • divorce regulations by government • lack of legislation on intimate partner violence within marriage • Protective marriage law • Traditional gender norms and social norms supportive of violence • Traditional gender norms and social norms supportive of violence a Some of these factors are also risk factors for sexual violence (see Table 2 for risk factors for both intimate partner and sexual violence). 28 Preventing intimate Partner and sexual violence against women One high-qua\fity study reported that South African men with a previous abusive his- tory were a\fmost three times more \fike\fy to have perpetrated intimate partner vio\fence ( Jewkes et a\f., 2006).

Relationshi\f-level factors educational disparity Disparities in educationa\f attainment between ma\fe and fema\fe partners may a\fso resu\ft in higher rates of intimate partner vio\fence (Ackerson et a\f., 2008; \bhan, 2009). Men may use vio\fence to gain power within a re\fationship in which the woman’s \feve\f of edu- cation is higher. Ackerson et a\f. (2008) found that Indian women with a higher \feve\f of education re\fative to their partner were more \fike\fy to experience physica\f intimate part- ner vio\fence. Xu et a\f. (2005) found a simi\far re\fationship between re\fative educationa\f attainment between partners and any type of intimate partner vio\fence in \bhina.

r elationship quality/marital satisfaction Partnerships with \fow marita\f satisfaction, continuous disagreements and high mari- ta\f discord are more \fike\fy to be associated with intimate partner vio\fence compared to those without. Lack of marita\f satisfaction and marita\f discord are strong\fy corre- \fated with the occurrence of both the perpetration and experiencing of intimate part- ner vio\fence (Morrison, E\f\fsberg & Bott, 2007; Stith et a\f., 2004; Tang & Lai, 2008).

Disagreements often occur over traditiona\f gender ro\fes, contro\f in partnerships with status disparities (for examp\fe, in income, education or age) and sexua\f acts or refusa\fs.

Vio\fence is perpetrated against a partner as a way of dea\fing with conflict or reso\fving the disagreement. Additiona\f\fy, a woman’s risk of victimization may be heightened in situations where women start an argument or when they fight back.

Community-level and societal-level factors Fewer studies have examined community-\feve\f and societa\f-\feve\f risk factors for inti- mate partner vio\fence, especia\f\fy in LMI\b. However, a number of recent studies (Ack - erson et a\f., 2008; Boy\fe et a\f., 2009; Gage, 2005; Koenig et a\f., 2004; Koenig et a\f., 2006) show that severa\f neighbourhood-\feve\f factors are associated with higher rates of intimate partner vio\fence, inc\fuding:

l \fower proportion of women with a higher \feve\f of education; l higher neighbourhood poverty; l higher neighbourhood unemp\foyment rate; l higher proportion of ma\fe and fema\fe i\f\fiteracy; l higher proportion of individua\fs with a positive view of vio\fence; l \fower proportion of women with high \feve\f of autonomy; and l higher proportion of househo\fds that use corpora\f punishment.

One study in 17 sub-Saharan African countries showed that intimate partner vio\fence against women was wide\fy accepted under certain circumstances by men and women in a\f\f the countries studied (Uthman, Lawoko & Moradi, 2009). Women were more \fike\fy to justify it than men. “Neg\fecting the chi\fdren” was the most common reason 29 agreed to by both women and men for justifying intimate partner vio\fence fo\f\fowed by “going out without informing husband” and “arguing back with the husband”. The WHO Mu\fti-country study on \bomen’s hea\fth and domestic vio\fence against \bomen (Gar- cia-Moreno et a\f., 2005) found that the percentage of women who agreed with one or more justifications for “wife beating” varied from 6% to over 65% . Suspecting a wife of being unfaithfu\f was the most common\fy agreed justification. In a\f\f countries except Thai\fand the overa\f\f acceptance that wife beating cou\fd be justified for some reason was significant\fy greater among women who had experienced physica\f or sexua\f inti- mate partner vio\fence (or both) than among women who had never experienced such vio\fence. Increasing wea\fth, educationa\f attainment, urbanization, access to media and joint decision-making were a\f\f associated with decreased \feve\fs of justification of inti- mate partner vio\fence against women in most countries. A\fthough the strength of the associations between these factors and intimate partner vio\fence needs to be further va\fidated, they remain potentia\f targets for the deve\fopment of programmes at a com- munity \feve\f.

2.5 Risk factors associated with se\bual violence Of the factors associated with the perpetration and experiencing of sexua\f vio\fence, most were identified at the individua\f \feve\f ( Ta b l e 4). Those factors that have the strongest reported effect or are most consistent\fy reported are high\fighted in the tab\fe in bo\fd font. Many of the risk factors \fisted as strong risk factors of sexua\f vio\fence have been discussed above in previous sections, suggesting that few such factors are specific to sexua\f vio\fence.

family honour and sexual purity Fami\fy responses to sexua\f vio\fence that b\fame women without punishing men, and concentrate instead on restoring “\fost” fami\fy honour, create an environment in which rape can occur with impunity. Whi\fe such fami\fies wi\f\f often try to protect their women from rape and may a\fso put their daughters on contraception to prevent visib\fe signs shou\fd rape occur, there is rare\fy much socia\f pressure to contro\f young men or persuade them that coercive sex is wrong. Instead, in some countries, there is frequent\fy support for fami\fy members to do whatever is necessary – inc\fuding murder of the victim – to a\f\feviate the “shame” associated with a rape or other sexua\f transgression ( Jewkes, Sen & Garcia-Moreno, 2002).

ideologies of male sexual entitlement Sexua\f vio\fence committed by men is to a \farge extent rooted in ideo\fogies of ma\fe sex - ua\f entit\fement. These be\fief systems grant women extreme\fy few \fegitimate options to refuse sexua\f advances. Many men thus simp\fy exc\fude the possibi\fity that their sexua\f advances towards a woman might be rejected or that a woman has the right to make an autonomous decision about participating in sex. In many cu\ftures, women as we\f\f as men regard marriage as entai\fing an ob\figation on women to be sexua\f\fy avai\fab\fe virtua\f\fy without \fimit, though sex may be cu\ftura\f\fy proscribed at certain times, such as after chi\fdbirth or during menstruation. Societa\f norms around the use of vio\fence as a means to achieve objectives have been strong\fy associated with the preva\fence of rape.

In societies where the ideo\fogy of ma\fe superiority is strong – emphasizing dominance, c \faPter 2. risk and Protective factors for intimate Partner and sexual violence 30 Preventing intimate Partner and sexual violence against women TABLE 4 Risk factors for se\bual violence a Perpetration by men Victimization of women Ind IVIdu Al l EVEl D EMOGRA\fHICS • Low socio-economic status/income D EMOGRA\fHICS • Young age • Lower education • Separated/divorced and single women • Gang membership • Early exposure to sexual activity E X\fOSURE TO CHILD MALTREATMENT • Sexual abuse • Physical abuse • Intraparental violence E X\fOSURE TO CHILD MALTREATMENT • Intra-parental \fiolence • Sexual abuse M E N TA L DISORDER • Antisocial personality M E N TA L DISORDER • Depression S U B S TA N C E USE • Harmful use of alcohol • Illicit drug use S U B S TA N C E USE • Harmful use of alcohol • Illicit drug use • Prior victimization REl AT I O nSHIP lEVE l • Multiple partners/infidelity • Multiple partners • Low resistance to peer pressure • Family honour and sexual purity COMM unITy l EVE l • \beak community sanctions • \beak community sanctions • \foverty • \foverty S OCIETAl l EVEl • Traditional gender norms and social norms supportive of violence • Traditional gender norms and social norms supportive of violence • Ideologies of male sexual entitlement • Ideologies of male sexual entitlement • \beak legal sanctions • \beak legal sanctions a Some of these factors are also risk factors for intimate partner violence (see Table 2 for risk factors for both intimate partner and sexual violence).

physica\f strength and ma\fe honour – rape is more common. \bountries with a cu\fture of vio\fence, or where vio\fent conflict is taking p\face, experience an increase in other forms of vio\fence, inc\fuding sexua\f vio\fence ( Jewkes, Sen & Garcia-Moreno, 2002).

weak legal sanctions Factors operating at a societa\f \feve\f that influence sexua\f vio\fence inc\fude \faws and nationa\f po\ficies re\fating to gender equa\fity in genera\f and to sexua\f and intimate part- ner vio\fence more specifica\f\fy, as we\f\f as norms re\fating to the use of vio\fence. Whi\fe the various factors operate \farge\fy at \foca\f \feve\f (within fami\fies, schoo\fs, workp\faces and communities) there are a\fso influences from the \faws and norms working at nationa\f and even internationa\f \feve\f.

There are considerab\fe variations between countries in their approach to sexua\f vio\fence \fegis\fation. Some countries have far-reaching \fegis\fation and \fega\f procedures, with a broad definition of rape that inc\fudes marita\f rape, heavy pena\fties for those convicted, 31 and a strong response in supporting victims. A commitment to preventing or reducing sexua\f vio\fence is a\fso reflected in an emphasis on po\fice training and an appropriate a\f\focation of po\fice resources to the prob\fem; in the priority given to investigating cases of sexua\f assau\ft; and in the resources made avai\fab\fe to support victims and provide medico-\fega\f services. However, even in countries with the best \faws, the conviction rate for sexua\f vio\fence is minima\f.

At the other end of the sca\fe, there are countries with much weaker approaches to the issue – where conviction of an a\f\feged perpetrator on the evidence of the woman a\fone is not a\f\fowed, where certain forms or settings of sexua\f vio\fence are specifica\f\fy exc\fuded from the \fega\f definition, and where rape victims are strong\fy deterred from bringing the matter to court through the fear of being punished for fi\fing an “unproven” rape case ( Jewkes, Sen & Garcia-Moreno, 2002).

2.6 Protective factors for intimate partner violence and se\bual violence Most of the research on the perpetration and experiencing of intimate partner vio\fence and sexua\f vio\fence has focused on factors associated with an increased \fike\fihood of intimate partner vio\fence and/or sexua\f vio\fence (risk factors) rather than factors that decrease or buffer against risk (protective factors). However, severa\f studies have shown that women who were more high\fy educated (secondary schoo\fing or higher) were 20 –55% \fess \fike\fy to be victims of intimate partner vio\fence or sexua\f vio\fence com- pared to \fess-educated women (Brown et a\f., 2006; Fehringer & Hindin, 2009; F\fake, 2005). Simi\far\fy, men who were more high\fy educated were approximate\fy 40 % \fess \fike\fy to perpetrate intimate partner vio\fence compared to \fess-educated men ( Johnson & Das, 2009). Marita\f duration of more than 15 years was a\fso identified as a potentia\f protective factor against ma\fe perpetration of intimate partner vio\fence in Bang\fadesh ( Johnson & Das, 2009). Other factors that may decrease or buffer against risk inc\fude:

l having benefited from hea\fthy parenting as a chi\fd (protective against intimate part - ner vio\fence and sexua\f vio\fence); l having own supportive fami\fy (intimate partner vio\fence); l \fiving within extended fami\fy/fami\fy structure (intimate partner vio\fence); l be\fonging to an association; and l women’s abi\fity to recognize risk (sexua\f vio\fence) (E\f\fsberg et a\f. 1999; Gidicyz et a\f., 2006; Schwartz et a\f., 2006).

However, much more scientific research into these protective factors is required.

2.7 Gender norms and inequality Gender inequa\fity and inequity are important societa\f-\feve\f factors that remain signifi - cant\fy under-researched in terms of their associations with intimate partner vio\fence and sexua\f vio\fence. However, two specific risk factors appear to be strong\fy associated with intimate partner and sexua\f vio\fence – the unequa\f position of women in a particu- \far re\fationship and in society (which is underwritten by ideo\fogies of ma\fe superiority); and the normative use of vio\fence to reso\fve conflicts (and during po\fitica\f strugg\fes).

These factors are manifested by distinct and hierarchica\f gender ro\fes, notions of ma\fe sexua\f entit\fement, the \fow socia\f va\fue and power of women, and ideas of manhood c \faPter 2. risk and Protective factors for intimate Partner and sexual violence 32 Preventing intimate Partner and sexual violence against women \finked to the contro\f or “discip\fining” of women. These in turn are \finked to factors such as \fow \feve\fs of education among women; few pub\fic ro\fes for women; the \fack of fami\fy, socia\f and \fega\f support for women; and the \fack of economic power for women ( Jewkes, 2002).

For the effective prevention of intimate partner and sexua\f vio\fence, it is vita\f to shed a strong \fight on how gender norms and gender inequa\fity and inequity are re\fated to such vio\fence. A thorough understanding based on sound empirica\f evidence of how gender norms and gender inequa\fity and inequity function as risk and protective factors for, and as causes of, intimate partner and sexua\f vio\fence in different sociocu\ftura\f contexts is required. Acquiring such an understanding shou\fd be a top priority of research into the risk and protective factors for intimate partner and sexua\f vio\fence.

2.8 Key messages Risk and protecti\fe factors Intimate partner and sexual violence n \bomen and men with lower levels of education are at increased risk of experiencing and perpetrating, respectively, intimate partner violence.

n Exposure to child maltreatment is strongly associated with: — the perpetration by men of intimate partner and sexual violence; and — the experiencing by women of intimate partner and sexual violence.

n An antisocial personality disorder is a strong risk factor for the perpetration of both intimate partner and sexual violence.

n Harmful use of alcohol is frequently found to be associated with the perpetration of both intimate partner and sexual violence.

n Males who have multiple partners or are suspected by their partners of infidelity are more likely to perpetrate both intimate partner and sexual violence.

n Attitudes that are accepting of violence are strongly associated with both the perpetration and experiencing of intimate partner and sexual violence.

Ris\f factors specific to intimate partner violence n \fast history of violence as a perpetrator or victim is a strong risk factor for future intimate partner violence.

n Marital discord and dissatisfaction are strongly associated with both the perpetrating and experiencing of intimate partner violence.

Ris\f factors specific to sexual violence n Beliefs in family honour and sexual purity are associated with a lack of social pressure to persuade young men that coercive sex is wrong.

n Sexual violence committed by men is to a large extent rooted in ideologies of male sexual entitlement. These belief systems grant women extremely few legitimate options to refuse sexual advances.

n \beak legal sanctions for sexual violence send the message that such violence is condoned, and may even exclude certain forms of sexual violence from the legal definition. 33 Pre\fention implications n Different risk and protective factors may operate in different countries and settings. Hence, it is important to identify and then address those risk factors most strongly associated with intimate partner violence and sexual violence in each setting.

n \frimary prevention efforts should focus on younger age groups.

n \freventing all forms of violence and abuse, especially child maltreatment, will help to reduce the levels of intimate partner and sexual violence.

n Reducing overall alcohol consumption in a population may help to reduce the harmful use of alcohol and with it the perpetration and experiencing of intimate partner and sexual violence.

n There are several modifiable factors associated with intimate partner violence that can be targeted by primary prevention measures such as reducing acceptance of violence, increasing women’s access to education, changing laws that discriminate against women and implementing more gender equitable policies. Although targeting these factors will in all likelihood also reduce sexual violence, the required evidence is currently lacking. c\faPter 2. risk and Protective factors for intimate Partner and sexual violence 34 CHA\fTER 3 Primary prevention strategies – the evidence base 3 .\f Introduction Intimate partner and sexua\f vio\fence are not inevitab\fe – their \feve\fs vary over time and between p\faces due to a variety of socia\f, cu\ftura\f, economic and other factors. As out- \fined in Chapter \f , this can resu\ft in substantia\f differences between and within coun - tries in the preva\fence of intimate partner and sexua\f vio\fence. Most important\fy, this variation shows that such vio\fence can be reduced through we\f\f-designed and effective programmes and po\ficies. As discussed in Chapter 2, there are important factors re\fated to both perpetration and victimization – such as exposure to chi\fd ma\ftreatment, wit- nessing parenta\f vio\fence, attitudes that are accepting of vio\fence and the harmfu\f use of a\fcoho\f – that can be addressed.

At present, evidence on the effectiveness of primary prevention strategies for intimate partner and sexua\f vio\fence is \fimited, with the overwhe\fming majority of data derived from HI\b – primari\fy the United States. \bonsequent\fy, current high priorities in this fie\fd inc\fude adapting effective programmes from high-income to \fower-income settings; further eva\fuating and refining those for which evidence is emerging; and deve\foping and testing strategies that appear to have potentia\f, especia\f\fy for use in \fow-resource settings, with rigorous eva\fuation of their effectiveness. At the same time, the dearth of evidence in a\f\f countries means that the generating of evidence and the incorporation of we\f\f-designed outcome eva\fuation procedures into primary prevention programmes are top priorities everywhere. This wi\f\f he\fp to ensure that the efforts made in this area are founded upon a so\fid evidence base. Furthermore, programme deve\fopers shou\fd be encouraged to exp\ficit\fy base programmes on existing theoretica\f frameworks and mode\fs of behaviour change to a\f\fow under\fying mechanisms to be identified and to make rep\fication easier.

Most of the eva\fuated strategies aimed at preventing intimate partner and sexua\f vio- \fence have targeted proxima\f risk factors – primari\fy at the individua\f and re\fationship \feve\fs of the eco\fogica\f mode\f.

the need for upstream action In the pub\fic hea\fth framework, primary prevention means reducing the number of new instances of intimate partner and sexua\f vio\fence by addressing the factors that make the first-time perpetration of such vio\fence more \fike\fy to occur. Primary preven- tion therefore re\fies on identifying the “upstream” determinants and then taking action to address these. The impact of widespread, comprehensive programmes can then be measured at the popu\fation \feve\f by comparing the rates at which such vio\fence is either 35 experienced or perpetrated. Given the \fifetime preva\fence of intimate partner and sexu- a\f vio\fence shown in Ta b l e \f (Chapter \f ), the hundreds of mi\f\fions of women wor\fdwide in need of services wou\fd outstrip the capacity of even the best-resourced countries. A prob\fem on this sca\fe requires a major focus on primary prevention.

Upstream actions can target risk factors across a\f\f four \feve\fs of the eco\fogica\f mode\f presented in Chapter 2. To decrease intimate partner and sexua\f vio\fence at the popu\fa- tion \feve\f, it is particu\far\fy important to address the societa\f or outer \feve\f of the mode\f.

Such measures inc\fude nationa\f \fegis\fation and supportive po\ficies aimed at socia\f and economic factors – such as income \feve\fs, poverty and economic deprivation, patterns of ma\fe and fema\fe emp\foyment, and women’s access to hea\fth care, property, educa- tion, and po\fitica\f participation and representation. It is sometimes even argued that programmes that aim to reduce intimate partner and sexua\f vio\fence against women without increasing ma\fe-fema\fe equity wi\f\f u\ftimate\fy not succeed in reducing vio\fence against women. However, whi\fe many strategies invo\fving \fega\f and educationa\f reform and emp\foyment opportunities are being imp\femented to increase gender equa\fity, few have been assessed for their impact on intimate partner and sexua\f vio\fence, making the eva\fuation of such strategies a priority. Any comprehensive intimate partner and sexua\f vio\fence prevention strategy must address these sociocu\ftura\f and economic fac- tors through \fegis\fative and po\ficy changes and by imp\fementing re\fated programmes.

creating a climate of non-tolerance Addressing risk factors at the societa\f \feve\f may increase the \fike\fihood of successfu\f and sustainab\fe reductions of intimate partner and sexua\f vio\fence. For examp\fe, when the \faw a\f\fows husbands to physica\f\fy discip\fine wives, imp\fementing a programme to prevent intimate partner vio\fence may have \fitt\fe impact. Nationa\f \fegis\fation and sup- portive po\ficies shou\fd therefore be put in p\face to ensure that women have equa\f rights to po\fitica\f participation, education, work, socia\f security and an adequate standard of \fiving. They shou\fd a\fso be ab\fe to enter free\fy into a marriage or to \feave it, to obtain financia\f credit, and to own and administer property. Laws and po\ficies that discrimi - nate against women shou\fd be changed, and any new \fegis\fation and po\ficies shou\fd be examined for their impact upon women and men. Legis\fation and po\ficies that address wider socioeconomic inequa\fities are \fike\fy to reduce other forms of interpersona\f vio- \fence which wi\f\f in turn he\fp to reduce intimate partner and sexua\f vio\fence.

Legis\fation and po\ficies that address wider socioeconomic inequa\fities can make a vita\f contribution to empowering women and improving their status in society; to creating cu\ftura\f shifts by changing the norms, attitudes and be\fiefs that support intimate part- ner and sexua\f vio\fence; and to creating a c\fimate of non-to\ferance for such vio\fence.

The human rights of gir\fs and women need to be respected, protected and fu\ffi\f\fed as part of ensuring the we\f\f-being and rights of everyone in society. As a first step towards this, governments shou\fd honour their commitments in imp\fementing the fo\f\fowing internationa\f \fegis\fation and human-rights instruments:

l \bonvention on the E\fimination of A\f\f Forms of Discrimination Against Women (19 79 ) ; l The \bonvention on the Rights of the \bhi\fd (1991); l The Dec\faration on the E\fimination of Vio\fence Against Women (1993); l The Beijing Dec\faration and P\fatform for Action (1995); c \faPter 3. PrimarY Prevention strategies – t\fe evidence base 36 Preventing intimate Partner and sexual violence against women l The Mi\f\fennium Dec\faration (2000); and l The Inter-American \bonvention on the Prevention, Punishment and Eradication of Vio\fence against Women (\bonvention of Be\fem do Para, 1994).

Legis\fation and crimina\f justice systems must a\fso be in p\face to dea\f with cases of intimate partner and sexua\f vio\fence after the event. These systems shou\fd aim to he\fp prevent further vio\fence, faci\fitate recovery and ensure access to justice – for examp\fe through the provision of specia\fized po\fice units, restraining orders and mu\fti-agency sexua\f assau\ft response teams. Potentia\f\fy, \fega\f protection against intimate partner and sexua\f vio\fence he\fps to reinforce non-vio\fent socia\f norms by sending the message that such acts wi\f\f not be to\ferated. Measures to crimina\fize abuse by intimate partners and to broaden the definition of rape have been instrumenta\f in bringing these issues out into the open and dispe\f\fing the notion that such vio\fence is a private fami\fy matter. In this regard, they have been very important in shifting socia\f norms (Heise & Garcia- Moreno, 2002; Jewkes, Sen & Garcia-Moreno, 2002). However, the evidence surround- ing the deterrent va\fue of arrest in cases of intimate partner vio\fence shows that it may be no more effective in reducing vio\fence than other po\fice responses, such as issuing warnings or citations, providing counse\f\fing or separating coup\fes (Fagan & Browne 1994; Garner, Fagan & Maxwe\f\f, 1995). Some studies have a\fso shown increased abuse fo\f\fowing arrest, particu\far\fy for unemp\foyed men and those \fiving in impoverished areas (Fagan & Browne 1994; Garner, Fagan & Maxwe\f\f, 1995). Protective orders can be usefu\f, but enforcement is uneven and there is evidence that they have \fitt\fe effect on men with serious crimina\f records (Heise & Garcia-Moreno, 2002). In cases of rape, reforms re\fated to the admissibi\fity of evidence and removing the requirement for vic- tims’ accounts to be corroborated have a\fso been usefu\f, but are ignored in many courts throughout the wor\fd (Du Mont & Parnis, 2000 ; Jewkes, Sen & Garcia-Moreno, 2002).

\burrent\fy, on the who\fe, sufficient evidence of the deterrent effect of crimina\f justice system responses on intimate partner and sexua\f vio\fence is sti\f\f \facking (Dah\fberg & Butchart, 2005).

Dismant\fing hierarchica\f constructions of mascu\finity and femininity predicated on the contro\f of women, and e\fiminating the structura\f factors that support inequa\fities are \fike\fy to make a significant contribution to preventing intimate partner and sexua\f vio\fence. However, these are \fong-term goa\fs. Strategies aimed at achieving these \fong- term objectives shou\fd be comp\femented by measures with more immediate effects that are informed by the evidence base presented in this chapter.

3.2 Assessing the evidence for different prevention approaches From the perspective of pub\fic hea\fth, a fundamenta\f question is “do intimate partner and sexua\f vio\fence prevention programmes work”? That is to say, are there certain programmes or strategies that are effective in preventing or reducing intimate partner and sexua\f vio\fence? Effectiveness can on\fy be demonstrated using rigorous research designs, such as randomized-contro\f\fed tria\fs or quasi-experimenta\f designs. These typica\f\fy compare the outcomes of an experimenta\f group (which receives the pro- gramme) with a contro\f or comparison group (which is as equiva\fent as possib\fe to the experimenta\f group but which does not receive the programme). One major concern is to be ab\fe to ru\fe out a\fternative exp\fanations for any observed changes in outcome in order to be confident that the changes rea\f\fy were due to the programme and not some other factor. This issue is discussed further in section 4.6 of this document. 37 A\fthough “testimonia\fs” are not a sound basis for eva\fuating the effectiveness of a pro- gramme, they can provide insights into its running and on whether participants find it worthwhi\fe. However, approaches that are based upon testimonia\fs might expend signif- icant resources and capacity on programmes that may be ineffective or may even make things worse (Dah\fberg & Butchart, 2005). Various criteria have now been proposed to more systematica\f\fy eva\fuate the effectiveness of different programmes. The most strin- gent criteria invo\fve programme eva\fuation using experimenta\f or quasi-experimenta\f designs; evidence of significant preventive effects; evidence of sustained effects; and the independent rep\fication of outcomes.

In spite of the emphasis on and visibi\fity of efforts to promote gender equa\fity and prevent intimate partner and sexua\f vio\fence, very few of the programmes reviewed in this chapter meet a\f\f of these criteria, whi\fe others have not been subjected to any kind of scientific eva\fuation. Rigorous scientific eva\fuation of programmes for preventing intimate partner and sexua\f vio\fence are even rarer in LMI\b. The fie\fd of intimate partner and sexua\f vio\fence prevention must therefore be considered to be at its ear\fiest stages in terms of having an estab\fished evidence base for primary prevention strategies, programmes and po\ficies. The \fimited evidence base for intimate partner and sexua\f vio\fence prevention has three important imp\fications for this chapter.

First, the chapter extrapo\fates, when re\fevant, from the stronger evidence base for chi\fd ma\ftreatment and youth vio\fence prevention, but c\fear\fy signa\fs that these extrapo\fa- tions remain specu\fative. Much, however, can be \fearned from the \fiterature on youth vio\fence and chi\fd ma\ftreatment prevention.

Second, the chapter describes those primary prevention programmes which have the potentia\f to be effective either on the grounds of theory or know\fedge of risk factors – even if there is current\fy \fitt\fe or no evidence to support them or where, in certain cases, they have not yet been wide\fy imp\femented. In the process, an attempt is made to draw attention to the under\fying theories, princip\fes and mechanisms on which the programmes are based. However, it is noted that a firm theoretica\f base and consistency with identified risk factors does not guarantee the success of a programme.

Third, the chapter inc\fudes programmes deve\foped in LMI\b settings on condition that they have some supporting evidence (even if it is weak) or are current\fy in the process of being eva\fuated, that they appear to have potentia\f on theoretica\f grounds, or that they address known risk factors. The inc\fusion criteria are designed on the one hand to avoid setting the bar of methodo\fogica\f standards too high – which wou\fd \fead to the exc\fu- sion of many of the programmes deve\foped in \fow-resource settings on the grounds that they have no or \fow-qua\fity evidence supporting them. On the other hand, setting the bar too \fow wou\fd run the risk of appearing to endorse programmes unsupported by evidence. However, the \fimitations of the evidence presented are c\fear\fy spe\ft out and the need for rigorous outcome eva\fuation studies emphasized.

A\fthough sti\f\f in its ear\fy stages, there are sound reasons to be\fieve that this fie\fd is poised to expand rapid\fy in coming years. Some programmes have been demonstrated to be effective fo\f\fowing rigorous outcome eva\fuations, evidence is beginning to emerge to support the effectiveness of many more, and suggestions for potentia\f strategies have pro\fiferated. Furthermore, tried and tested methods for deve\foping effective evidence- based primary prevention programmes and po\ficies for other forms of interpersona\f vio\fence have been reported (for examp\fe, F\fay et a\f., 2005; Mzarek & Haggerty, 1994; O\fds, Sad\fer & Kitzman, 2007). The fie\fd of evidence-based intimate partner and sex- c \faPter 3. PrimarY Prevention strategies – t\fe evidence base 38 Preventing intimate Partner and sexual violence against women ua\f vio\fence prevention now requires an open mind to promising approaches, and to innovative new ideas at a\f\f stages of the \fife cyc\fe.

3.3 Summary tables of primary prevention strategies and programmes \f Ta b l e 5 summarizes the strength of evidence for the effectiveness of those strategies to prevent intimate partner vio\fence and sexua\f vio\fence for which some evidence is avai\f- ab\fe. Strategies are grouped according to \fife stage. An important distinction must be drawn between a strategy and a specific programme. A\fthough specific programmes may have been demonstrated to be effective, this in no way imp\fies that a\f\f other pro- grammes categorized under the same strategy are a\fso effective. For examp\fe, the Nurse Fami\fy Partnership, deve\foped in the USA, is a home-visitation programme that has been demonstrated to be effective in preventing chi\fd ma\ftreatment. Neverthe\fess, it is the on\fy programme within the broader strategy of home visitation (which inc\fudes a mu\ftitude of different programmes) that is supported by so\fid evidence of its effective- ness (MacMi\f\fan et a\f., 2009).

Strategies are ranked for their effectiveness in preventing intimate partner vio\fence and sexua\f vio\fence as fo\f\fows:

 Effective: strategies which inc\fude one or more programmes demonstrated to be effective; effective refers to being supported by mu\ftip\fe we\f\f-designed studies showing prevention of perpetration and/or experience of intimate partner and/or sexua\f vio\fence; m Emerging evidence: strategies which inc\fude one or more programmes for which evidence of effectiveness is emerging; emerging evidence refers to being supported by one we\f\f-designed study showing prevention of perpetration and/or experience of intimate partner and/or sexua\f vio\fence or studies showing positive changes in know\fedge, attitudes and be\fiefs re\fated to intimate partner vio\fence and/or sexua\f vio\fence; ? Effectiveness unc\fear: strategies which inc\fude one or more programmes of unc\fear effectiveness due to insufficient or mixed evidence; X Emerging evidence of ineffectiveness: strategies which inc\fude one or more programmes for which evidence of ineffectiveness is emerging; emerging evidence refers to being supported by one we\f\f-designed study showing \fack of prevention of perpetration and/or experience of intimate partner and/or sexua\f vio\fence or stud- ies showing an absence of changes in know\fedge, attitudes and be\fiefs re\fated to intimate partner vio\fence and/or sexua\f vio\fence; X Ineffective: s t r a t e g i e s w h i c h i n c \f u d e o n e o r m o r e p r o g r a m m e s s h ow n t o b e i n e f fe c - tive; ineffective refers to being supported by mu\ftip\fe we\f\f-designed studies show- ing \fack of prevention of perpetration and/or experience of intimate partner and/or sexua\f vio\fence; XX Probab\fy harmfu\f: strategies which inc\fude at \feast one we\f\f-designed study showing an increase in perpetration and/or experiencing of intimate partner and/ 1 Since the deve\fopment of prevention strategies has not mirrored the identification of risk and protective factors, the organization of this chapter does not para\f\fe\f that of Chapter 2. However, the previous\fy identified risk factors are high\fighted in this chapter. 39 or sexua\f vio\fence or negative changes in know\fedge, attitudes and be\fiefs re\fated to intimate partner and/or sexua\f vio\fence; As shown in Ta b l e 5, there is current\fy on\fy one strategy for the prevention of intimate partner vio\fence that can be c\fassified “effective” at preventing actua\f vio\fence. This is the use of schoo\f-based programmes to prevent vio\fence within dating re\fationships.

However, on\fy three such programmes – described be\fow – have been demonstrated to be effective and these findings cannot be extrapo\fated to other schoo\f-based pro- grammes using a different approach, content or intensity. At present, there are no cor- responding\fy eva\fuated effective programmes against sexua\f vio\fence.

Ta b l e 6 \fists those strategies for which there is current\fy no evidence or very weak evi- dence but which appear to have potentia\f on the grounds of theory, known risk factors, or outcome eva\fuations that are methodo\fogica\f\fy of \fower qua\fity; it a\fso inc\fudes some promising strategies that are current\fy undergoing eva\fuation.

A\f\f the strategies reviewed have been organized according to the main \fife stages. When strategies are re\fevant to more than one stage of \fife, they have been categorized under the stage at which they are most often de\fivered. Strategies re\fevant to a\f\f \fife stages are described \fast. In contrast to the forma\f definitions in section \f.\f and because of the way programmes are organized, intimate partner vio\fence is considered here to inc\fude instances of sexua\f vio\fence that occur within an intimate partnership, whi\fe sexua\f vio- \fence is used here to refer to sexua\f vio\fence occurring outside intimate partnerships (i.e.

perpetrated by friends, acquaintances or strangers). Dating vio\fence can be considered to incorporate both possibi\fities since dating partners can range from being \fitt\fe more than acquaintances to more intimate partners. However, in Ta b l e 5 and Ta b l e 6 , dating vio\fence is c\fassified for the sake of convenience under intimate partner vio\fence. BOX 2 Outcome measures of effecti\feness The effectiveness of a programme can be evaluated in terms of three different types of outcome – each of which can be measured at different intervals after the programme:

1. Changes in knowledge, attitudes and beliefs regarding intimate partner and sexual violence. This is the weakest of the three outcomes because changes in knowledge, attitudes and beliefs do not necessarily lead to changes in violence behaviour. In this respect, even successful programmes in this area cannot be assumed to be effective at preventing actual intimate partner or sexual violence without further research demonstrating corresponding reductions in violent behaviour.

2. Reductions in the perpetration of intimate partner or sexual violence.

3. Reductions in the experiencing of intimate partner or sexual violence.

Intimate partner violence is not a unitary construct and can take different forms, including physical, sexual and psychological violence. Despite this, outcome evaluations generally do not examine effectiveness in relation to these different types of violence – nor are programmes generally designed to address specific types of intimate partner violence in particular. It is possible that programmes considered to be effective or promising may only be so for certain forms of intimate partner violence (\bhitaker, Baker & Arias, 2007). c \faPter 3. PrimarY Prevention strategies – t\fe evidence base 40 Preventing intimate Partner and sexual violence against women TABLE 5 Primary prevention strategies for intimate partner violence and se\bual violence for which some evidence is available STRATEGy In T I M AT E PART nER VIO lEn CE SE xu Al VIO lEn CE duRInG In FAnCy, CHI ldHOO d And EAR ly AdOl ESCEnCE Interventions for children and adolescents subjected to child maltreatment and/or exposed to intimate partner violence m ?

School-based training to help children recognize and avoid potentially sexually abusive situations ?

m du RInG AdOl ESCEnCE A nd EAR ly Adul THOO d School-based programmes to prevent dating violence  nA Sexual violence prevention programmes for school and college populations nA?

Rape-awareness and knowledge programmes for school and college populations nA X Education (as opposed to skills training) on self-defence strategies for school and college populations nA X Confrontational rape prevention programmes nAXX du RInG A dulTHOO d Empowerment and participatory approaches for addressing gender inequality: Microfinance and gender-equality training m ?

Empowerment and participatory approaches for addressing gender inequality: Communication and relationship skills training (e.g. Stepping Stones) m ?

Home-visitation programmes with an intimate partner violence component ??

A ll lIFE STAGES Reduce access to and harmful use of alcohol m ?

Change social and cultural gender norms through the use of social norms theory ?

m Change social and cultural gender norms through media awareness campaigns m ?

Change social and cultural gender norms through working with men and boys m ?

 Effective: strategies which include one or more programmes demonstrated to be effective; effective refers to being supported by multiple well-designed studies showing prevention of perpetration and/or experiencing of intimate partner and/or sexual violence; m Emerging evidence of effectiveness: strategies which include one or more programmes for which evidence of effectiveness is emerging; emerging evidence refers to being supported by one well-designed study showing prevention of perpetration and/or experiencing of intimate partner and/or sexual violence or studies showing positive changes in knowledge, attitudes and beliefs related to intimate partner violence and/or sexual violence; ? Effectiveness unclear: strategies which include one or more programmes of unclear effectiveness due to insufficient or mixed evidence; X Emerging evidence of ineffectiveness: strategies which include one or more programmes for which evidence of ineffectiveness is emerging; emerging evidence refers to being supported by one well-designed study showing lack of prevention of perpetration and/or experiencing of intimate partner and/or sexual violence or studies showing an absence of changes in knowledge, attitudes and beliefs related to intimate partner violence and/or sexual violence; X Ineffective: strategies which include one or more programmes shown to be ineffective; ineffective refers to being supported by multiple well-designed studies showing lack of prevention of perpetration and/or experiencing of intimate partner and/or sexual violence; XX Probably harmful: strategies which include at least one well-designed study showing an increase in perpetration and/or experiencing of intimate partner and/or sexual violence or negative changes in knowledge, attitudes and beliefs related to intimate partner and/or sexual violence; nA Not applicable. 41 TABLE 6 Primary prevention strategies for intimate partner violence and se\bual violence with potential STRATEGy du RInG In FAnCy, CHI ldHOO d And EAR ly AdOl ESCEnCE Home-visitation programmes to prevent child maltreatment \farent education to prevent child maltreatment Improve maternal mental health Identify and treat conduct and emotional disorders School-based social and emotional skills development Bullying prevention programmes du RInG AdOl ESCEnCE A nd EAR ly Adul THOO d School-based multi-component violence prevention programmes du RInG A dulTHOO d United States Air Force multi-component programme to prevent suicide Empowerment and participatory approaches for addressing gender inequality – SASA!

3.4 during infancy, childhood and early adolescence \fome-visitation and parent-education programmes to prevent child maltreatment As noted in ear\fier sections of this document, a history of chi\fd ma\ftreatment substan- tia\f\fy increases the risk of an individua\f becoming either a perpetrator or victim of inti- mate partner vio\fence and of sexua\f vio\fence. It is therefore reasonab\fe to assume that preventing chi\fd ma\ftreatment has the potentia\f to reduce subsequent intimate partner and sexua\f vio\fence (Foshee, Reyes & Wyckoff., 2009). However, direct evidence of the effect of such programmes on the \feve\fs of intimate partner vio\fence is current\fy sti\f\f \fack ing.

In genera\f however, reducing the risk of the different forms of chi\fd ma\ftreatment reviewed in Preventing chi\fd ma\ftreatment: a guide to taking action and generating evidence (WHO-ISP\bAN, 2006) 1 can contribute to reducing the intergenerationa\f transmission of vio\fence and abuse. The most promising strategies for preventing chi\fd ma\ftreat- ment in this area inc\fude home visitation and parent education programmes (Mikton & Butchart, 2009). However, neither type of programme has been eva\fuated for its \fong-term effects on the prevention of intimate partner and sexua\f vio\fence among the grown-up chi\fdren of parents who were invo\fved in such programmes. improve maternal mental health Materna\f depression (which affects at \feast 1 in 10 new mothers) can interfere with good bonding and attachment processes. This in turn increases the risk of persistent conduct disorders in chi\fdren (a key risk factor for the \fater perpetration of vio\fence) by as much as five-fo\fd (Me\ftzer et a\f., 2003). Effective approaches for addressing materna\f depression inc\fude ear\fy recognition (antenata\f\fy and postnata\f\fy) fo\f\fowed by peer and c \faPter 3. PrimarY Prevention strategies – t\fe evidence base 1 To obtain this guide, p\fease visit: www.who.int/vio\fence_injury_prevention /vio\fence /activities /chi\fd_ ma\ftreatment/en/index.htm\f 42 Preventing intimate Partner and sexual violence against women socia\f support, psycho\fogica\f therapies and antidepressant medication (Nationa\f \bo\f- \faborating \bentre for Menta\f Hea\fth, 2007). The \fong-term effects on the chi\fdren of mothers treated for materna\f depression in terms of their \fater invo\fvement in intimate partner and sexua\f vio\fence have not been assessed, but the approach appears to have potentia\f.

identify and treat conduct and emotional disorders in children \bonduct disorders in chi\fdhood and ado\fescence – a precursor of antisocia\f persona\f- ity disorder – are associated with an increased risk of experiencing and/or perpetrating intimate partner and sexua\f vio\fence. Additiona\f\fy, emotiona\f disorders are associated with \fater depression and anxiety in adu\ft years and can increase the risk of postnata\f depression and persistent materna\f depression. As out\fined above, these in turn con- tribute to as much as a five-fo\fd increased risk of emotiona\f or conduct disorders in the chi\fdren of mothers with poor menta\f hea\fth (Me\ftzer et a\f., 2003). The ear\fy identifica- tion and effective treatment of conduct and emotiona\f disorders in chi\fdhood and ado- \fescence cou\fd therefore be expected to reduce the occurrence of subsequent intimate partner and sexua\f vio\fence.

Good evidence exists of the \finks between ear\fy conduct disorder and \fater invo\fvement in vio\fence as both victim and perpetrator, and of the effectiveness of interventions to reduce conduct disorder and youth offending. However, despite their potentia\f, there is at present no evidence showing that the strategy of identifying and treating conduct and emotiona\f disorders in chi\fdhood or ear\fy ado\fescence \feads to reductions in intimate partner and sexua\f vio\fence during \fater ado\fescence and adu\fthood.

interventions for children and adolescents subjected to child maltreatment and/or exposed to intimate partner violence As discussed in section 2.3, because chi\fdren or ado\fescents who have been subjected to chi\fd ma\ftreatment or exposed to parenta\f vio\fence are at increased risk of becoming the perpetrators and victims of intimate partner and sexua\f vio\fence, interventions in this area are particu\far\fy important.

One meta-ana\fysis (Skowron & Reinemann, 2005) examined 21 programmes invo\fving psycho\fogica\f interventions targeted at chi\fdren and ado\fescents who had experienced chi\fd ma\ftreatment. Resu\fts suggested that psycho\fogica\f treatments for chi\fd ma\ftreat- ment yie\fded improvements among participants: some 71% of treated chi\fdren appeared to be functioning better than their non-treated counterparts. A\f\f of the interventions were designed to improve cognitive, emotiona\f and behavioura\f outcomes, with 11 of the studies considered to be experimenta\f. A randomized tria\f of one of these pro- grammes (the Youth Re\fationship Project – section 3.5 ) used ado\fescent dating vio\fence as an outcome and found a reduction in the experiencing and perpetration of physica\f and emotiona\f abuse (Wo\ffe et a\f., 2003).

Psycho\fogica\f interventions for chi\fdren and ado\fescents subjected to chi\fd ma\ftreat- ment and/or exposed to intimate partner vio\fence therefore appear to represent a strat- egy for the prevention of intimate partner vio\fence supported by emerging evidence.

Their effect on sexua\f vio\fence remains unc\fear at present. 43 school-based social and emotional skills development Factors such as impu\fsiveness, \fack of empathy and poor socia\f competence – which may be indicative of conduct disorder, a precursor of antisocia\f persona\fity disor- der – are important individua\f risk factors for perpetrating various forms of vio\fence, inc\fuding intimate partner and sexua\f vio\fence. \bognitive-behavioura\f ski\f\fs training programmes and socia\f deve\fopment programmes that address these factors in chi\fdren a nd you n g ado\fescent s a re t herefore prom isi n g st r ateg ies for prevent i n g subsequent v io- \fence. These programmes seek to promote pro-socia\f behaviour and to provide socia\f and emotiona\f ski\f\fs such as prob\fem-so\fving, anger management, increased capacity for empathy, perspective-taking and non-vio\fent conflict reso\fution. They can either be popu\fation-based or targeted at those at high risk, and are typica\f\fy de\fivered in schoo\fs.

Whi\fe there is strong evidence that such programmes can be effective in reducing youth vio\fence and improving socia\f ski\f\fs (e.g. Löse\f & Bee\fmann, 2003), there is current\fy no evidence that they can reduce sexua\f and dating vio\fence among ado\fescents and young adu\fts, or intimate partner and sexua\f vio\fence \fater in \fife. Nonethe\fess, they appear to have potentia\f in preventing subsequent intimate partner vio\fence and sexua\f vio\fence.

school-based training to help children to recognize and avoid potentially sexually abusive situations Schoo\f-based programmes to prevent chi\fd sexua\f abuse by teaching chi\fdren to recog - nize and avoid potentia\f\fy sexua\f\fy abusive situations are run in many parts of the wor\fd, but eva\fuated examp\fes come main\fy from the United States.

A recent systematic review of reviews (Mikton & Butchart, 2009) found that a\fthough schoo\f-based programmes to prevent chi\fd sexua\f abuse are effective at strengthening know\fedge and protective behaviours against this type of abuse, evidence showing wheth - er such programmes reduce its actua\f occurrence is \facking. Two studies that measured future experience of sexua\f abuse as an outcome reported mixed resu\fts (Finke\fhor, Asdi - gian & Dziuba-Leatherman, 1995; Gibson & Leitemberg, 2000). Nonethe\fess, emerg - ing evidence of their effectiveness in preventing subsequent sexua\f abuse victimization appears to support the use of such programmes. Further research on the \fong-term impact on actua\f sexua\f abuse victimization is however required (Finke\fhor, 2009).

bullying prevention programmes Bu\f\fying has both immediate and \fong-term consequences on perpetrators and victims, inc\fuding socia\f iso\fation and the exacerbation of antisocia\f behaviour that can \fead to juveni\fe and adu\ft crime (for perpetrators) and depression, suicida\f ideation, socia\f iso- \fation and \fow se\ff-esteem (for victims). Some of these consequences may increase the risk of \fater invo\fvement in intimate partner and/or sexua\f vio\fence either as perpetrator or victim. A number of reviews have conc\fuded that bu\f\fying prevention programmes are effective in reducing bu\f\fying (Ba\fdry & Farrington, 2007; Smith et a\f., 2004). A systematic review and meta-ana\fysis of schoo\f-based programmes to reduce bu\f\fying and victimization (Farrington & Ttofi, 2009) showed that, overa\f\f, schoo\f-based bu\f\fy- ing prevention programmes are effective in reducing both bu\f\fying and being bu\f\fied.

On average, bu\f\fying perpetration decreased by 20 –23% and the experiencing of being bu\f\fied decreased by 17–20 % . c \faPter 3. PrimarY Prevention strategies – t\fe evidence base 44 Preventing intimate Partner and sexual violence against women A\fthough such programmes are \fike\fy to have broader potentia\f benefits, evidence of their effect on the experiencing or perpetrating of intimate partner and/or sexua\f vio\fence \fater in \fife is \fimited. A number of studies, however, have demonstrated an association between bu\f\fying and sexua\f harassment. Some sexua\f vio\fence prevention programmes in the United States inc\fude bu\f\fying prevention components for e\femen- tary and midd\fe schoo\f age chi\fdren (Basi\fe, Hertz & Back, 2009).

3.5 during adolescence and early adulthood \bchool-based \frogrammes to \frevent dating violence Dating vio\fence is an ear\fy form of partner vio\fence, occurring primari\fy in ado\fescence and ear\fy adu\fthood, and experienced within a “dating re\fationship”. Dating vio\fence prevention programmes have been the most eva\fuated of a\f\f intimate partner vio\fence prevention programmes with 12 eva\fuations of ado\fescent dating vio\fence prevention programmes, inc\fuding five randomized tria\fs (Foshee et a\f., 2008). Targeted at ear\fy sexua\f re\fationships, in contexts where marriage is usua\f\fy entered into from about 20 years of age, these programmes have been shown to prevent dating vio\fence and sexua\f vio\fence. Furthermore, dating vio\fence appears to be a risk factor for intimate partner vio\fence \fater in \fife (Smith, White & Ho\f\fand, 2003) and is a\fso associated with injuries and hea\fth-compromising behaviours, such as unsafe sex, substance abuse and suicide attempts (Wo\ffe et a\f., 2009). According\fy, the prevention of dating vio\fence can be assumed to be preventive of intimate partner and sexua\f vio\fence in \fater \fife (Foshee, Reyes & Wyckoff., 2009).

One dating vio\fence prevention programme that has been we\f\f eva\fuated using a rand- omized-contro\f\fed design is Safe Dates. Positive effects were noted in a\f\f four pub\fished eva\fuations (Foshee et a\f., 1998; Foshee et a\f., 2000 ; Foshee et a\f., 2004; Foshee et a\f., 2005). Foshee et a\f. (2005) examined the effects of Safe Dates in preventing or reduc - ing perpetration and victimization over time using four waves of fo\f\fow-up data. The programme significant\fy reduced psycho\fogica\f, moderate physica\f and sexua\f dating vio\fence perpetration at a\f\f four fo\f\fow-up periods. The programme a\fso significant\fy reduced severe physica\f dating abuse perpetration over time, but on\fy for ado\fescents who reported no or average prior invo\fvement in severe physica\f perpetration at base- \fine. Programme effects on the experiencing of sexua\f dating vio\fence over time were margina\f. Safe Dates did not prevent or reduce the experiencing of psycho\fogica\f dating abuse. Programme effects were primari\fy due to changes in dating vio\fence norms, gen- der ro\fe norms and awareness of community services. The programme did not affect conflict-management ski\f\fs. The programme was found to have had a greater impact upon primary prevention as opposed to preventing re-abuse among those with a history of previous abuse (Foshee et a\f., 1996; 1998; 2000 ; 2004; 2008).

Two schoo\f-based programmes for preventing dating vio\fence in Ontario, \banada have a\fso been eva\fuated (Wo\ffe et a\f., 2003; Wo\ffe et a\f., 2009). An outcome eva\fuation of The Fourth R: Ski\f\fs for Youth Re\fationships used a c\fuster-randomized design and found that, based on se\ff-reported perpetration at 2.5-year fo\f\fow-up, rates of physica\f dating vio\fence were 7.4% in the programme group and 9.8% in the contro\f group – a differ- ence of 2.4% . However, for reasons not fu\f\fy understood, this decrease of se\ff-reported perpetration was found in boys (7.1% in contro\fs versus 2.7% in intervention students) 45 but not in gir\fs (12.1% versus 11.9 %). 1 The programme – eva\fuated by samp\fing over 1700 hundred students aged 14 –15-years from 20 pub\fic schoo\fs – was integrated into the existing hea\fth and physica\f education curricu\fum, and taught in sex-segregated c\fasses. An under\fying theme of hea\fthy, non-vio\fent re\fationship ski\f\fs was woven throughout the 21 \fessons, which inc\fuded extensive ski\f\fs deve\fopment using graduated practice with peers to deve\fop positive strategies for dea\fing with pressures and the reso- \fution of conflict without abuse or vio\fence. The cost of training and materia\fs averaged 16 \banadian do\f\fars per student (Wo\ffe et a\f., 2009).

The other \banadian schoo\f-based programme which has been eva\fuated (Wo\ffe et a\f., 2003) is the Youth Re\fationship Project . This community-based programme aimed to he\fp 14 –16 year-o\fds who had been ma\ftreated as chi\fdren to deve\fop hea\fthy non-abu- sive re\fationships with dating partners. The programme educated participants on both hea\fthy and abusive re\fationships and he\fped them to acquire conflict reso\fution and communication ski\f\fs. A randomized-contro\f\fed tria\f showed that the programme had been effective in reducing incidents of physica\f and emotiona\f abuse, and the symptoms of emotiona\f distress over a 16 -month period after the programme (Wo\ffe et a\f., 2003).

These three schoo\f-based programmes therefore appear to be effective for the preven- tion of physica\f, sexua\f and emotiona\f vio\fence in dating re\fationships in ado\fescents, and may a\fso he\fp to prevent intimate partner and sexua\f vio\fence among adu\fts. How- ever, there are a number of necessary caveats concerning dating vio\fence prevention programmes. A\fthough high-qua\fity eva\fuations of the three programmes described above found reduced vio\fence at moderate\fy \fong fo\f\fow-up periods, the eva\fuations of most other programmes have been of poor qua\fity, used short fo\f\fow-up periods, and on\fy inc\fuded know\fedge and attitude changes as outcomes (for which some positive effects were found). Whether changes in know\fedge and attitudes \fead to corresponding changes in behaviour is uncertain (Whitaker et a\f., 2006). Moreover, further research is needed to eva\fuate the effectiveness of dating vio\fence prevention programmes in the \fonger term, when integrated with programmes for the prevention of other forms of vio\fence, and when de\fivered outside North America and in resource-poor settings.

A particu\far concern that has been raised about programmes such as Safe Dates is the extent to which they are cu\fture-bound to North America and hence may be of \fimited va\fue in LMI\b. school-based multi-component violence prevention programmes Universa\f mu\fti-component programmes are the most effective schoo\f-based vio\fence prevention programmes (Adi et a\f., 2007; Dusenbury et a\f., 1997; Hahn et a\f., 2007).

Such programmes are de\fivered to a\f\f pupi\fs and go beyond the norma\f components of curricu\fum-based teaching to inc\fude teacher training in the management of behaviour, parenting education and peer mediation. There can a\fso be after schoo\f activities and/or community invo\fvement. One systematic review (Hahn et a\f., 2007) estimated that on average, universa\f mu\fti-component programmes reduced vio\fence by 15% in schoo\fs that de\fivered the programmes compared to those that did not. c \faPter 3. PrimarY Prevention strategies – t\fe evidence base 1 These rates of perpetration by boys and gir\fs are broad\fy in keeping with findings of sex-symmetry in young men and women’s se\ff-reported partner assau\fts in North America mentioned in the Introduc- tion . 46 Preventing intimate Partner and sexual violence against women Schoo\f-based mu\fti-component vio\fence prevention programmes have most\fy focused upon bu\f\fying and youth vio\fence as outcomes. Given that the risk factors for youth vio\fence and intimate partner and sexua\f vio\fence are to some extent shared, such pro- grammes wou\fd appear to have some potentia\f for preventing these \fatter forms of vio- \fence. However, there is current\fy no evidence of their effectiveness in these areas.

sexual violence prevention programmes for school and college populations In the United States, the majority of programmes for the primary prevention of sexua\f vio\fence by strangers, acquaintances and non-intimate dating partners have focused on co\f\fege students – though they have a\fso increasing\fy been de\fivered to high schoo\f and midd\fe schoo\f pupi\fs. In settings where few go into higher education this approach has obvious \fimitations. Deve\fopmenta\f\fy, it makes sense to educate young peop\fe in appro- priate and inappropriate sexua\f behaviour at a time when their sexua\f identities are forming and their attitudes to romantic partners are beginning to take shape. However, once again there is a severe paucity of evidence to confirm the effectiveness or otherwise of such programmes (Schewe, 2007).

Two recent systematic reviews in the United States have eva\fuated the effectiveness of specific primary prevention programmes in this area. The first of these (Morrison et a\f., 2004) inc\fuded both co\f\fege, high-schoo\f and midd\fe-schoo\f popu\fations, and found that programmes usua\f\fy inc\fuded severa\f components (most often the cha\f\fenging of rape myths; information on acquaintance and date rape; statistics on rape; and risk- reduction and protective prevention ski\f\fs). Of the 50 studies reviewed, seven (14%) showed exc\fusive\fy positive effects on know\fedge and attitudes, but none used the actua\f experiencing or perpetration of vio\fence as outcomes; 40 (80 %) reported mixed effects; and three (6%) indicated no effect. The studies a\fso had a number of serious meth- odo\fogica\f \fimitations that \fed the reviewers to conc\fude that the effectiveness of such programmes remains unc\fear. These \fimitations inc\fuded the use of know\fedge and attitude as the on\fy outcome measures; studies of higher-qua\fity design showing poorer resu\fts; and the positive effects of the programmes being found to diminish over time.

The second systematic review (Anderson & Whiston, 2005) examined 69 education programmes for co\f\fege students on sexua\f assau\ft, and found \fitt\fe evidence of the effectiveness of such programmes in preventing such assau\fts, or in increasing \feve\fs of rape empathy (the cognitive-emotiona\f recognition of a rape victim’s trauma) or aware- ness. However, the programmes eva\fuated were found to increase factua\f know\fedge about rape and to beneficia\f\fy change attitudes towards it. The acute shortage of studies that use behaviour as outcomes \fed the authors to conc\fude that more research using such outcomes was needed before definitive conc\fusions cou\fd be reached. The effec- tiveness of such programmes, on the basis of these two reviews, is current\fy unc\fear.

It has been fou nd t hat t he prov ision of “fact ua \f” i n for mat ion 1 as part of addressing rape myths appears to have no effect on attitudes to rape or on the \feve\fs of empathy for its victims (Schewe, 2007). Eva\fuation studies indicate that rape awareness and know\fedge programmes based on imparting such information rare\fy work. Simi\far\fy, educating women on effective se\ff-defence strategies without teaching them actua\f se\ff-defence 1 Such information typica\f\fy inc\fudes: the \fega\f definition of rape and associated \fega\f terms; statistics on the preva\fence of rape; descriptions of typica\f perpetrators and of the rape trauma syndrome; and information on avai\fab\fe resources for rape victims. 47 ski\f\fs has been found to be of questionab\fe va\fue, and may even be potentia\f\fy harmfu\f in some contexts (Schewe, 2007). Two eva\fuations of programmes that focused on a discussion of se\ff-defence strategies without teaching the corresponding ski\f\fs found no reduction in sexua\f assau\ft risk at fo\f\fow-up (Breitenbecher & Gidycz, 1998; Breiten- becher & Scarce, 2001). Rape prevention programmes that use a sty\fe of persona\f con- frontation with participants actua\f\fy appear to be harmfu\f. One study eva\fuating such a programme found that it resu\fted in greater to\ferance among men of the justifiabi\fity of rape (Fisher, 1986).

A number of other approaches have been tried for which there is present\fy very \fimited evidence of effectiveness. Encouraging victim empathy has been associated with both improvements and worsening of attitudes towards sexua\f vio\fence and the acceptance of rape myths (Schewe, 2007). Educating women on how to avoid high-risk situations (such as hitchhiking, abusing a\fcoho\f or becoming invo\fved with o\fder men) has a\fso \fed to mixed resu\fts and it too has been associated with greater acceptance of rape myths. To avoid the encouragement of victim-b\faming, it is crucia\f that such education is de\fivered to fema\fe-on\fy audiences. There have a\fso been mixed indications of the effectiveness of programmes that emphasize the negative consequences of sexua\f vio\fence to men, and that try to persuade them to see such sex as \fess rewarding than consensua\f sex.

Fina\f\fy, severa\f programmes for preventing sexua\f vio\fence have been proposed that have as yet been neither wide\fy imp\femented nor eva\fuated. These inc\fude providing universa\f rape prevention education, and parent education in sexua\f vio\fence preven- tion, throughout schoo\fs and workp\faces; educating teachers and coaches about sexua\f vio\fence and its prevention; and changing organizationa\f practices to inc\fude activities such as mandatory training in the prevention of vio\fence against women.

3.6 during adulthood empowerment and participatory approaches to reduce gender inequality Empowerment is an approach that he\fps individua\fs and communities to identify their own prob\fems and to deve\fop, through participatory methods, the resources, ski\f\fs and confidence needed to address them. This approach emphasizes the ro\fe of individua\fs and communities as agents of change and prioritizes community ownership and \fead- ership of the entire process. \bomprehensive programmes dea\f with the community as a who\fe or with mu\ftip\fe subgroups of the popu\fation; have severa\f components; and are designed to effect socia\f change by creating a supportive environment for chang- ing individua\f and community attitudes and behaviour. Such approaches often uti\fize a combination of participatory rapid needs assessment, education or training, pub\fic awareness campaigns and community action (Lankester, 1992; Mor\fey et a\f, 1983).

Two examp\fes of empowerment approaches for preventing intimate partner vio\fence are the use of microfinance with gender-equalit\f training and t he Stepping Stones training package.

A number of initiatives invo\fving microfinance have now been estab\fished to increase the economic and socia\f power of women. These initiatives provide sma\f\f \foans to mobi- \fize income-generating projects that can a\f\feviate poverty. Stand a\fone credit and rura\f deve\fopment programmes such as Grameen Bank and the Bang\fadesh Rura\f Advance - ment \bommittee target women and appear to show some promise in reducing inti- mate partner vio\fence. However, the eva\fuation of such programmes needs to take into c \faPter 3. PrimarY Prevention strategies – t\fe evidence base 48 Preventing intimate Partner and sexual violence against women account reports of \fenders exp\foiting disadvantaged borrowers with very high rates of interest which can trap peop\fe in debt and contribute to further poverty (Rhyne, 2001) as we\f\f as reports of increases in intimate partner vio\fence (Kabeer, 2001). Disagree- ments over the contro\f of new\fy acquired assets and earnings, combined with wom- en’s changing attitudes towards traditiona\f gender ro\fes, improved socia\f support and greater confidence in defending themse\fves against ma\fe authority, has sometimes \fed to marita\f conflicts and vio\fence against women perpetrated by their partners (Schu\fer et a\f., 1996). Increases in vio\fence fo\f\fowing participation in credit programmes have a\fso been reported e\fsewhere, at \feast in the initia\f stages of membership (Ahmed, 2005; Rahman, 1999). Pre-existing gender ro\fes appear to affect the vio\fence-re\fated outcomes of credit programmes – in communities with rigid gender ro\fes, women’s invo\fvement can resu\ft in increased \feve\fs of intimate partner vio\fence not seen in communities with more-flexib\fe gender ro\fes (Koenig et a\f., 2003). The outcome eva\fuations conducted to date of such stand a\fone microfinance programmes have not been as rigorous as that of the Intervention \bith Microfinance for AIDS and Gender Equity ( I M AGE ) prog ra m me described in Bo\b 3 .

BOX 3 Intervention with Microfinance for AID\b and Gender Equity (IMAG\b) One of the most rigorously evaluated and successful microfinance and women’s empowerment programmes to date has been the Intervention with Microfinance \bor AIDS and Gender Equity (IMAGE) in South Africa. This programme targets women living in the poorest households in rural areas, and combines a microfinance programme with training and skills-building sessions on preventing HIV infection, and on gender norms, cultural beliefs, communication and intimate partner violence.

The programme also encourages wider community participation to engage men and boys. It aims to improve women’s employment opportunities, increase their influence in household decisions and their ability to resolve marital conflicts, strengthen their social networks and reduce HIV transmission.

A randomized-controlled trial found that two years after completing the programme, participants reported experiencing 55% fewer acts of violence by their intimate partners in the previous 12 months than did members of a control group. In addition, participants were more likely to disagree with statements that condone physical and sexual violence towards an intimate partner (52% of participants versus 36% of the control group). Whi\fe microfinance programmes can operate as discrete entities, IM AGE is an exam- p\fe of such a programme which a\fso incorporates education sessions and ski\f\fs-bui\fding workshops to he\fp change gender norms, improve communication in re\fationships and empower women in other ways and has been shown to be effective at reducing intimate partner vio\fence (Kim et a\f., 2009). Through education and ski\f\fs-bui\fding for women and engagement with boys and men, and the broader community, IM AGE was effec- tive in reducing intimate partner vio\fence and supporting women. This was achieved without producing the type of negative effects seen in other settings where cu\ftura\f shifts and other changes have taken p\face in the absence of efforts to engage men. 49 The Stepping Stones training package 1 is another participatory approach that promotes communication and re\fationship ski\f\fs within communities. Training sessions are run in para\f\fe\f for sing\fe-sex groups of women and men. Origina\f\fy designed for the preven- tion of HIV infection, severa\f communities have now incorporated e\fements of vio\fence prevention. The approach has been used in 40 LMI\b in Africa, Asia, Europe and Latin America. Versions of the programme have now been eva\fuated in a number of countries (We\fbourn, 2009). 2 T he most t horough eva \fuat ion to date has been a random i zed- con - tro\f\fed tria\f in the Eastern \bape province of South Africa, with participants aged 15 –26 years. This study indicated that a \fower proportion of men who had participated in the programme committed physica\f or sexua\f intimate partner vio\fence in the two years fo\f\fowing the programme, compared with men in a contro\f group ( Jewkes et a\f., 2008).

Furthermore, an eva\fuation in Gambia compared two vi\f\fages where the programme was carried out with two contro\f vi\f\fages, and fo\f\fowed participating coup\fes over one year. It found that compared to coup\fes not receiving the programme, communication was improved and quarre\f\fing reduced in participating coup\fes. In addition, participat - ing men were found to be more accepting of a wife’s refusa\f to have sex and \fess \fike\fy to beat her (Paine et a\f., 2002).

SASA! is an “act ivist k it” for mobi\fizing com munit ies to prevent vio\fence against wom - en, focusing in particu\far on the connection between HIV/AIDS and vio\fence against women. 3 “Sasa” is a Kiswahi\fi word meaning “now” and the kit inc\fudes practica\f resources; activities monitoring and assessment too\fs to support \foca\f activism, media and advocacy activities; and communication and training materia\fs. It targets commu - nity norms and traditiona\f gender ro\fes and aims to change know\fedge, attitudes, ski\f\fs and behaviour to redress the power imba\fance between men and women. It was created by Raising Voices, a Uganda-based nongovernmenta\f organization that works in the Horn of Africa, and Southern Africa. The London Schoo\f of Hygiene and Tropica\f Medicine, Raising Voices, the Kampa\fa-based \benter for Domestic Vio\fence Preven- tion and Makerere University are current\fy conducting a joint randomized-contro\f\fed tria\f to eva\fuate the effectiveness of the approach.

Thus evidence is emerging of the effectiveness in LMI\b of empowerment and partici- patory approaches in preventing intimate partner vio\fence through microfinance com- bined with gender-equa\fity training, and through the Stepping Stones training package.

The resu\fts of the SASA! eva\fuation are expected to provide further evidence on the effectiveness of this type of programme which seems to have potentia\f for reducing inti- mate partner vio\fence. There is a need to rep\ficate and sca\fe up this type of approach.

Severa\f other participatory and community-empowerment strategies to prevent intimate partner vio\fence may be of va\fue, a\fthough these have se\fdom been imp\femented as pri- mary prevention strategies or rigorous\fy eva\fuated. \boup\fes counse\f\fing focuses on vio- \fence and/or substance abuse, and may be effective for coup\fes who have not resorted to intimate partner vio\fence but who may be at risk. Fami\fy programmes to promote posi- tive communication and hea\fthy re\fationships, and prevent fami\fy vio\fence, might a\fso be effective in preventing both intimate partner and sexua\f vio\fence given the impor- tance of fami\fy factors in their deve\fopment. In Ecuador, one intimate partner vio\fence prevention programme that was imp\femented (but not eva\fuated) consisted of c\fose c \faPter 3. PrimarY Prevention strategies – t\fe evidence base 1 See: www.steppingstonesfeedback.org2 See a\fso: www.creativexchange.org/chd /case33 See: www.raisingvoices.org/sasa/index.php 50 Preventing intimate Partner and sexual violence against women friends or re\fatives being assigned to “monitor” new\fyweds and to intervene shou\fd seri- ous conflict arise. There is a\fso some initia\f evidence that socia\f cohesion among resi- dents increases a community’s capacity to manage crime and vio\fence (by increasing “co\f\fective efficacy”) \feading to decreases in both \fetha\f and non-\fetha\f intimate partner vio\fence. Such community-\feve\f interventions can beneficia\f\fy change community-\feve\f characteristics and warrant further eva\fuation.

\fome visitation programmes to prevent intimate partner violence A systematic review of home visitation programmes (Bi\fukha et a\f., 2005) identified on\fy one eva\fuation study (Eckenrode, 2000 in the United States) which examined the effect of home visitation on \feve\fs of intimate partner vio\fence. No significant differ- ence in the incidence of such vio\fence among the programme and contro\f groups was fou nd.

A five-year project (2007–2012) funded by United States \benters for Disease \bontro\f and Prevention is current\fy under way, which wi\f\f deve\fop, test and eva\fuate a pro- gramme to reduce intimate partner vio\fence among \fow-income women enro\f\fed in the Nurse Fami\fy Partnership during pregnancy and in the first two years postpartum.

The Nurse Fami\fy Partnership is a nurse home visitation programme of demonstrated effectiveness in reducing chi\fd ma\ftreatment. The primary aims are to deve\fop a mode\f for an in-home intimate partner vio\fence prevention programme for enro\f\fed mothers at risk of such vio\fence; to test the feasibi\fity and acceptabi\fity of the programme; and in a randomized-contro\f\fed study to compare the effectiveness of the approach to that of the Nurse Fami\fy Partnership a\fone.

An eva\fuation of the Ha\baii Hea\fthy Start Program – an ear\fy chi\fdhood home visita- tion programme – found that when compared with a contro\f group, the participation of mothers was associated with reduced perpetration and experiencing of intimate partner vio\fence. The effect persisted for the first three years of a chi\fd’s \fife, with sma\f\f decreas- es in both the perpetration and experiencing of materna\f intimate partner vio\fence at fo\f\fow-up when the chi\fd was seven and nine years-o\fd (Bair-Merrit et a\f., 2010). Evi- dence for the effectiveness of such programmes can current\fy thus be considered to be unc\fear.

u nited states air force ( usaf) multi-component programme to prevent suicide This programme was primari\fy aimed at reducing the rate of suicide among USAF per- sonne\f, but was a\fso shown to reduce “fami\fy vio\fence” which inc\fuded both intimate partner vio\fence and chi\fd ma\ftreatment. The programme was based upon:

l the fu\f\f invo\fvement of the USAF \feadership to ensure the programme had the sup - port of the entire service; l incorporation of suicide prevention into professiona\f mi\fitary education; l community education and training of mi\fitary personne\f to identify risk factors, pro- vide appropriate intervention, and refer individua\fs who were potentia\f\fy at risk of suicide; and l the creation of a mu\ftidiscip\finary team consisting of menta\f hea\fth providers, medi - ca\f providers and chap\fains who cou\fd respond to traumatic events at the community \feve\f, inc\fuding suicides. 51 The programme reduced the rate of suicide by 33% and the rates of severe and mod- erate fami\fy vio\fence by 54% and 30 % respective\fy. Due to the combining of intimate partner vio\fence and chi\fd ma\ftreatment in the same outcome measure, it is not pos- sib\fe to determine the effect of the programme on intimate partner vio\fence specifica\f\fy (Knox et a\f., 2003), hence this programme is considered to have potentia\f, rather than being supported by emerging evidence.

3.7 All life stages reduce access to and harmful use of alcohol As described in section 2.3, harmfu\f use of a\fcoho\f is associated with the perpetration of intimate partner and sexua\f vio\fence. It can therefore be hypothesized that reducing both access to a\fcoho\f and its harmfu\f use wi\f\f \fead to reductions in intimate partner and sexua\f vio\fence. However, the re\fationship between harmfu\f use of a\fcoho\f and vio\fence is comp\fex – not everyone who drinks is at equa\f\fy increased risk of committing vio- \fence, and intimate partner and sexua\f vio\fence can occur at high rates in cu\ftures where a\fcoho\f use is taboo. Furthermore, there is disagreement among experts on whether or not a\fcoho\f can be considered to be a “cause” of intimate partner and sexua\f vio\fence or whether it is better viewed as a moderating or contributory factor. It seems c\fear, however, that individua\f and societa\f be\fiefs that a\fcoho\f causes aggression can \fead to vio\fent behaviour being expected when individua\fs are under the influence of a\fcoho\f, and to a\fcoho\f being used to prepare for and excuse such vio\fence. To date, research focusing on the prevention of a\fcoho\f-re\fated intimate partner and sexua\f vio\fence is scarce. There is, however, some emerging evidence suggesting that the fo\f\fowing strat- egies aimed at reducing a\fcoho\f consumption may be effective in preventing intimate partner vio\fence.

l Reducing a\fcoho\f avai\fabi\fity: In Austra\fia, a community intervention that inc\fuded restricting the hours of sa\fe of a\fcoho\f in one town reduced the number of domestic vio\fence victims presenting to hospita\f (Doug\fas, 1998). In Green\fand, a coupon-based a\fcoho\f rationing system imp\femented in the 1980s that entit\fed adu\fts to a\fcoho\f equiva\fent to 72 beers per month saw a subsequent 58% reduction in the number of po\fice ca\f\f outs for domestic quarre\fs (Room et a\f, 2003).

l Regu\fating a\fcoho\f prices: Increasing the price of a\fcoho\f is an effective means of reducing a\fcoho\f-re\fated vio\fence in genera\f (\bha\foupka, Grossman & Saffer, 2002).

A\fthough research eva\fuating the effectiveness of this approach in reducing intimate partner vio\fence specifica\f\fy is scarce, one study using economic mode\f\fing estimated that in the United States a 1% increase in the price of a\fcoho\f may decrease the proba- bi\fity of intimate partner vio\fence towards women by about 5% (Markowitz, 2000).

l Treatment for a\fcoho\f-use disorders: In the United States, treatment for a\fcoho\f dependence among ma\fes significant\fy decreased husband-to-wife and wife-to-hus- band intimate partner vio\fence 6 and 12 months \fater (Stuart et a\f., 2003), suggesting that such treatment may a\fso be an effective primary prevention measure.

Intimate partner and sexua\f vio\fence may a\fso be reduced through primary prevention programmes to reduce the more genera\f harms caused by a\fcoho\f (Anderson, \bhisho\fm & Fuhr, 2009). Approaches for which effectiveness is we\f\f supported by evidence inc\fude: c \faPter 3. PrimarY Prevention strategies – t\fe evidence base 52 Preventing intimate Partner and sexual violence against women l \baking a\fcoho\f \fess avai\fab\fe: This can be achieved by introducing minimum purchase-age po\ficies, and reducing the density of a\fcoho\f retai\f out\fets and the hours or days a\fcoho\f can be so\fd. Such an approach has been shown to \fead to fewer a\fco- ho\f-re\fated prob\fems, inc\fuding homicide and assau\fts (Duai\fibi et a\f., 2007).

l Banning of a\fcoho\f advertising: A \fcoho\f i s m a rketed t h rou g h i nc re a si n g \fy soph i s - ticated advertising in mainstream media, through the \finking of a\fcoho\f brands to sports and cu\ftura\f activities; through sponsorships and product p\facements; and through direct marketing via the Internet, podcasting and mobi\fe te\fephones. The strongest evidence for the \fink between a\fcoho\f advertising and consumption comes from \fongitudina\f studies on the effects of various forms of a\fcoho\f marketing – inc\fuding exposure to a\fcoho\f advertising in traditiona\f media and promotion in the form of movie content and a\fcoho\f-branded merchandise – on the initiation of youth drinking, and on riskier patterns of youth drinking (Anderson et a\f., 2009). However, evidence showing that such measures reduce intimate partner and sexua\f vio\fence is current\fy \facking.

l Individua\f\fy directed interventions to drinkers a\fready at risk: These inc\fude screening and brief interventions. A\fcoho\f screening (such as AUDIT; Babor et a\f., 2001) and brief interventions in primary hea\fth care settings have proven effective in reducing \feve\fs and intensity of consumption in LMI\b and HI\b (Room et a\f., 2003).

However, their direct effect on a\fcoho\f-re\fated intimate partner vio\fence has not been measured. Evidence indicates that drinkers may reduce their consumption by as much as 20 % fo\f\fowing a brief intervention, and that heavy drinkers who receive such an intervention are twice as \fike\fy to reduce their a\fcoho\f consumption as heavy drinkers who receive no intervention. Brief interventions inc\fude the opportune provision of advice and information in hea\fth or crimina\f justice settings (typica\f\fy during a 5 –10 minute period) but can a\fso extend to severa\f sessions of motivationa\f interviewing or counse\f\fing (FPH, 2008; Sheehan, 2008).

Schoo\f-based education on a\fcoho\f does not appear to reduce harm, but pub\fic-informa- t ion a nd educat ion prog ra m mes (wh i \fe aga i n appa rent \fy i nef fec t ive at reduc i ng a \fcoho\f- re\fated harm) can increase the attention given to a\fcoho\f on pub\fic and po\fitica\f agendas (Anderson et a\f., 2009).

As with most primary prevention programmes to prevent intimate partner and sexua\f vio\fence, programmes to reduce access to and harmfu\f use of a\fcoho\f have main\fy been conducted and eva\fuated in HI\b and \fitt\fe is known of their suitabi\fity or effectiveness outside such countries. For many LMI\b, programmes such as efforts to strengthen and expand the \ficensing of out\fets cou\fd be of great va\fue in reducing a\fcoho\f-re\fated inti- mate partner and sexua\f vio\fence. In many deve\foping societies, a \farge proportion of a\fcoho\f production and sa\fes current\fy takes p\face in unregu\fated informa\f markets. One study in São Pao\fo, Brazi\f (Laranjeira & Hink\fy, 2002) found that just 35% of a\fcoho\f out\fets surveyed had a \ficence of some form, and that a\fcoho\f vendors (whether \ficensed or not) faced few apparent restrictions on trading. Furthermore, in many LMI\b there are far fewer specia\fist hea\fth faci\fities, reducing the opportunities for a\fcoho\f treatment or screening. In such settings, it may instead be beneficia\f to deve\fop the ro\fe of primary hea\fth care workers or genera\f practitioners in identifying and a\f\feviating the harmfu\f use of a\fcoho\f.

A\fthough evidence for the effectiveness of measures to reduce access to and harm- fu\f use of a\fcoho\f is on\fy beginning to emerge, and high-qua\fity studies showing their 53 impact on intimate partner and sexua\f vio\fence are sti\f\f \farge\fy \facking, a\fcoho\f-re\fated programmes for the prevention of intimate partner vio\fence and sexua\f vio\fence appear promising. The strong association between a\fcoho\f and intimate partner and sexua\f vio\fence suggests that primary prevention interventions to reduce the harm caused by a\fcoho\f cou\fd potentia\f\fy be effective. Approaches to preventing a\fcoho\f-re\fated intimate partner and sexua\f vio\fence shou\fd a\fso address the socia\f acceptabi\fity of excessive drinking as a mitigating factor in vio\fence, whi\fe a\ftering normative be\fiefs about mas- cu\finity and heavy drinking. There remains a pressing need for additiona\f research to eva\fuate the effectiveness of such approaches in reducing intimate partner and sexua\f vio\fence, especia\f\fy in LMI\b.

change social and cultural norms related to gender that support intimate partner and sexual violence \bu\ftura\f and socia\f gender norms are the ru\fes or “expectations of behaviour” which regu\fate the ro\fes and re\fationships of men and women within a specific cu\ftura\f or socia\f group. Often unspoken, these norms define what is considered appropriate behav- iour, govern what is and is not acceptab\fe, and shape the interactions between men and women. Individua\fs are discouraged from vio\fating these norms through the threat of socia\f disapprova\f or punishment, or because of fee\fings of gui\ft and shame in con- travening interna\fized norms of conduct. Often traditiona\f socia\f and cu\ftura\f gender norms make women vu\fnerab\fe to vio\fence from intimate partners, p\face women and gir\fs at increased risk of sexua\f vio\fence, and condone or support the acceptabi\fity of vio\fence (Bo\b 4).

Efforts to change socia\f norms that support intimate partner and sexua\f vio\fence are therefore a key e\fement in the primary prevention of these forms of vio\fence. Approach- es have been adopted, a\fthough rare\fy eva\fuated, throughout the wor\fd to break the c \faPter 3. PrimarY Prevention strategies – t\fe evidence base BOX 4 \bxamples of social and cultural norms that support \fiolence against women n A man has a right to assert power over a woman and is considered socially superior – e.g. India (Mitra & Singh, 2007); Nigeria (Ilaka, 2005); and Ghana (Amoakohene, 2004).

n A man has a right to physically discipline a woman for “incorrect” behaviour – e.g. India (Go et al., 2003); Nigeria (Adegoke & Oladeji, 2008); and China (Liu & Chan, 1999).

n \fhysical violence is an acceptable way to resolve conflict in a relationship – e.g. the United States (Champion & Durant, 2001).

n Intimate partner violence is a “taboo” subject – e.g. South Africa (Fox et al., 2007).

n Divorce is shameful – e.g. \fakistan (Hussain & Khan, 2008).

n Sex is a man’s right in marriage – e.g. \fakistan (Hussain & Khan, 2008).

n Sexual activity (including rape) is a marker of masculinity – e.g. South Africa (\fetersen, Bhana & McKay, 2005).

n Girls are responsible for controlling a man’s sexual urges – e.g. South Africa (Ilika, 2005; \fetersen, Bhana & McKay, 2005). 54 Preventing intimate Partner and sexual violence against women si\fence that often surrounds intimate partner and sexua\f vio\fence; to try to inform and influence socia\f attitudes and socia\f norms on the acceptabi\fity of vio\fence; and to bui\fd po\fitica\f wi\f\f to address the prob\fem. The use of research findings for advocacy has been shown to be promising in bringing attention to, and raising awareness of, the prob\fem and in contributing to the shaping of reforms and po\ficies (for examp\fe, see E\f\fsberg, Li\fjestrand & Winkwist, 1997). \burrent\fy the three main approaches for changing socia\f and cu\ftura\f norms t hat support intimate partner and sexua\f vio\fence are: socia\f norms theory (i.e. correcting misperceptions that the use of such vio\fence is a high\fy preva\fent normative behaviour among peers); media awareness campaigns ; and working with men and boys . Often severa\f approaches are used in one programme.

S oc i a \f nor m s t he or y assumes that peop\fe have mistaken perceptions of other peop\fe’s attitudes and behaviours. The preva\fence of risk behaviours (such as heavy a\fcoho\f use or to\ferance of vio\fent behaviour) is usua\f\fy overestimated, whi\fe protective behaviours are norma\f\fy underestimated. This affects individua\f behaviour in two ways: (i) by increasing and justifying risk behaviours; and (ii) by increasing the \fike\fihood of an individua\f remaining si\fent about any discomfort caused by risky behaviours (thereby reinforcing socia\f to\ferance). The socia\f norms approach seeks to rectify these misper- ceptions by generating a more rea\fistic understanding of actua\f behavioura\f norms, thereby reducing risky behaviour.

In the United States, the socia\f norms approach has been app\fied to the prob\fem of sexua\f vio\fence among co\f\fege students. Among such students, men appeared to underestimate both the importance most men and women p\face on sexua\f consent and the wi\f\fingness of most men to intervene against sexua\f assau\ft (Fabiano et a\f., 2003). A\fthough the evidence is \fimited, some positive resu\fts have been reported. In one university in the United States, the A Man Respects a Woman project aimed to reduce the sexua\f assau\ft of women, increase accurate perceptions of non-coercive sexua\f behaviour norms and reduce se\ff-reported coercive behaviours by men. The project used a socia\f norms mar- keting campaign targeting men, a theatre presentation addressing socia\fization issues and ma\fe peer-to-peer education. Eva\fuation of the campaign two years after its imp\fe- mentation found that men had more accurate perceptions of other men’s behaviour, and improved attitudes and be\fiefs regarding sexua\f abuse. For examp\fe, a decreased percentage of men be\fieved that the average ma\fe student has sex when his partner is intoxicated; wi\f\f not stop sexua\f activity when asked to if he is a\fready sexua\f\fy aroused; and, when wanting to touch someone sexua\f\fy, tries and sees how they react. However, the percentage of men indicating that they have sex when their partner is intoxicated increased (Bruce, 2002).

\bedia awareness campaigns are a common approach to the primary prevention of intimate partner and sexua\f vio\fence. \bampaign goa\fs might inc\fude raising pub\fic awareness (for examp\fe, about the extent of the prob\fem, about intimate partner vio- \fence and sexua\f vio\fence as vio\fations of women’s human rights and about men’s ro\fe in ending vio\fence against women); providing accurate information; dispe\f\fing myths and stereotypes about intimate partner vio\fence and sexua\f vio\fence; and changing pub\fic opinion. Such campaigns have the potentia\f to reach \farge numbers of peop\fe.

An examp\fe of a media-awareness campaign is Sou\f City in South Africa. 1 This mu\fti- media hea\fth promotion and change project examines a variety of hea\fth and deve\f- opment issues, imparts information and aims to change socia\f norms, attitudes and 1 www.sou\fcity.org.za 55 practice. It is directed at individua\fs, communities and the socio-po\fitica\f environment.

One of its components aims to change the attitudes and norms that support intimate partner and sexua\f vio\fence. This mu\fti-\feve\f intervention was \faunched over six months and consisted of a series of te\fevision and radio broadcasts, print materia\fs and a he\fp- \fine. In partnership with a nationa\f coa\fition on preventing intimate partner vio\fence, an advocacy campaign was a\fso directed at the nationa\f government with the aim of achieving imp\fementation of the Domestic Vio\fence Act of 1998. The strategy aimed for impact at mu\ftip\fe \feve\fs from individua\f know\fedge, attitudes, se\ff-efficacy and behaviour; to community dia\fogue, shifting socia\f norms and the creating of an ena- b\fing \fega\f and socia\f environment for change. An independent eva\fuation of the pro- gramme inc\fuded nationa\f surveys before and after the intervention, focus groups and in-depth interviews with target audience members and stakeho\fders at various \feve\fs. It found that the programme had faci\fitated imp\fementation of the Domestic Vio\fence Act of 1998, had positive\fy impacted on prob\fematic socia\f norms and be\fiefs (such as that intimate partner vio\fence is a private matter) and had improved \feve\fs of know\fedge of where to seek he\fp. Attempts were a\fso made to measure its impact on vio\fent behaviour but there was insufficient data to determine this accurate\fy (Usdin et a\f., 2005).

As the Sou\f City project indicates, evidence is emerging that media campaigns com - bined with other educationa\f opportunities can change know\fedge, attitudes and be\fiefs re\fated to intimate partner and sexua\f vio\fence. Whi\fe good campaigns can increase know\fedge and awareness, influence perceptions and attitudes and foster po\fitica\f wi\f\f for action, evidence of their effectiveness in changing behaviour remains insufficient (Whitaker, Baker, & Arias, 2007).

Working w ith men and boys – t here has been an increasing tendency to focus effor ts to change socia\f and cu\ftura\f norms on ado\fescent ma\fes or younger boys using univer - sa\f or targeted programmes that are de\fivered through a variety of mechanisms, inc\fud- ing schoo\f-based initiatives, community mobi\fization and pub\fic awareness campaigns.

Objectives typica\f\fy inc\fude increasing an individua\f’s know\fedge, changing attitudes towards gender norms and vio\fence and changing socia\f norms around mascu\finity, power, gender and vio\fence. Some programmes a\fso aim to deve\fop the capacity and confidence of boys and young men to speak up and intervene against vio\fence, with the goa\f of changing the socia\f c\fimate in which it occurs (Katz, 2006). Fai\fure to engage men and boys in prevention may resu\ft in the type of negative effects seen in some set- tings where cu\ftura\f shifts and other changes have taken p\face in the absence of efforts to engage them ( Bo\b 5).

A review of programmes that work with men and boys to prevent vio\fence against women (Barker, Ricardo & Nascimento, 2007) inc\fuded 13 primary prevention programmes, five of which were imp\femented in LMI\b. Four of these programmes were judged by the reviewers to be “effective”; six “promising”; and three “unc\fear”. For examp\fe, one community outreach and mobi\fization campaign in Nicaragua judged to be effective was ca\f\fed Vio\fence against \bomen: a disaster \be can prevent as men (So\fórzano, Abaunza & Mo\fina, 2000). This was aimed at men aged between 20 –39 years who were affect- ed by Hurricane Mitch. The campaign’s main messages addressed men’s abi\fity and responsibi\fity to he\fp prevent or reduce vio\fence against their partners. Constructing mascu\finity \bithout intimate partner vio\fence was a group-education programme aimed at men in periurban districts of Managua, Nicaragua (We\fsh, 1997). The effect of the programme was, however, unc\fear due to the weakness of the outcome eva\fuation.

Indeed, the methodo\fogica\f qua\fity of most of the outcome eva\fuations was very \fow c \faPter 3. PrimarY Prevention strategies – t\fe evidence base 56 Preventing intimate Partner and sexual violence against women and outcome measures consisted main\fy of attitude changes and se\ff-reported rates of gender-based vio\fence, often using on\fy sma\f\f samp\fe sizes.

One campaign in New South Wa\fes in Austra\fia – Vio\fence against \bomen: it’s against a\f\f the ru\fes 1 – targeted 21–29 year-o\fd men and aimed to influence their attitudes. Sports ce\febrities de\fivered the message that vio\fence towards women is unacceptab\fe and that a masc u \f i ne ma n is not a v io\fent ma n. It a \fso soug ht to en ha nce t he com mu n it y ’s capac - ity to cha\f\fenge and address vio\fence against women. A post-campaign survey indicated that the campaign achieved some positive resu\fts: 83% of the respondents reported that the message of the campaign was that vio\fence against women is “not on” and 59 % of respondents cou\fd reca\f\f the campaign s\fogan. However, 91% of the target group reported that the issue was not one they wou\fd ta\fk about with their peers, irrespective of the campaign.

Simi\far\fy, in the United States, Men can stop rape 2 runs a pub\fic education campaign for men and boys with the message: “my strength is not for hurting”. This campaign runs in conjunction with Men of Strength (MOST) c\fubs – a primary prevention pro- BOX 5 Nicaraguan backlash shows the need to engage men as well Since 2000, Nicaragua has pioneered a number of initiatives to protect women against domestic violence. These have included:

n a network of police stations for women ( Comisaria de la Mujer) where women who have been abused can receive psychological, social and legal support; n a ministry for family affairs ( Mi Familia) which among other responsibilities ensures that shelter is available to women and children who suffer domestic violence; and n reform of the national reproductive health programme to address gender and sexual abuse.

During the same period, civil society groups have campaigned to promote the rights of women and to empower them to oppose domestic abuse. Because of these efforts, the reported frequency of intimate partner violence and sexual violence against women has increased dramatically. The more advocacy and awareness, the more likely women will report violence against them. For example, the number of reported cases of sexual violence received by the Comisaria de la Mujer rose from 4174 (January to June 2003) to 8376 (January to June 2004).

Researchers at the Universidad Centro Americana and the Institute for Gender Studies say a number of factors explain this increase – growing awareness among women that the cultural traditions that foster violence are no longer acceptable under international law and the Nicaraguan Domestic Violence Law; and better reporting of cases as women are encouraged to speak out. However, as Nicaraguan women have more actively opposed male hegemony, domestic conflicts have also increased and more men have resorted to intimate partner violence.

These findings suggest that responses to intimate partner violence must not focus exclusively on women, but must also target men to prevent this type of backlash (Schopper, Lormand & \baxweiler, 2006). 1 http:// \faw\fink.nsw.gov.au / \faw\fink/vaw/ \f\f _vaw.nsf/vwPrint1/vaw_vaw_ iaatrcampaign2 www.mencanstoprape.org 57 gramme which provides high-schoo\f-age young men with a structured and supportive space to \fearn about hea\fthy mascu\finity, and the redefining of ma\fe strength.

Whi\fe programmes to a\fter cu\ftura\f and socia\f norms are among the most visib\fe and ubiquitous of a\f\f strategies for preventing intimate partner and sexua\f vio\fence, they remain one of the \feast eva\fuated. Even where eva\fuations have been undertaken, these have typica\f\fy measured changes in attitudes and be\fiefs rather than in the occurrence of the vio\fent behaviours themse\fves, making it difficu\ft to draw firm conc\fusions on their effectiveness in actua\f\fy preventing intimate partner and sexua\f vio\fence. None- the\fess, some evidence is emerging to support the use of the three types of programmes reviewed above in changing the socia\f and cu\ftura\f gender norms that support intimate partner and sexua\f vio\fence. However, these must now be taken to sca\fe and more rigor- ous\fy eva\fuated.

3.8 Key messages n To achieve change at the population level it is important to target societal-level factors in the primary prevention of intimate partner and sexual violence. Approaches include the enactment of legislation and the development of supporting policies that protect women; addressing discrimination against women; and helping to move the culture away from violence – thereby acting as a foundation for further prevention work.

n Currently, there are no strategies of demonstrated effectiveness for preventing sexual violence outside intimate partner or dating relationships. Only one strategy has been demonstrated to be effective in preventing intimate partner violence, namely school-based programmes for adolescents to prevent violence within dating relationships – and this still needs to be assessed for use in resource-poor settings.

n Although it is too early to consider them proven, evidence is emerging of the effectiveness of several other strategies for the prevention of intimate partner and sexual violence, particularly the use of microfinance with gender equality training and of programmes that promote communication and relationship skills within communities.

n Developing the evidence base for programmes for the primary prevention of intimate partner and sexual violence is still very much in the early stages. But there is every reason to believe that rigorous outcome evaluations of existing programmes and the development of new programmes based on sound theory and known risk factors will lead to a rapid expansion in coming years. c \faPter 3. PrimarY Prevention strategies – t\fe evidence base 58 CHA\fTER 4 Improving programme planning and evaluation This chapter is based upon the six-step framework shown in Figure 2 for imp\fementing intimate partner and sexua\f vio\fence prevention po\ficies and programmes.

4 .\f Step \f: Getting started identify key partners and develop partnership working arrangements Partnership invo\fves different sectors and agencies working co\f\faborative\fy to de\fiver intimate partner and sexua\f vio\fence prevention programmes in a coordinated way. Pro- grammes are \fike\fy to cut across the activity areas of many different sectors and organi- zations, and can be \farge-sca\fe and/or mu\fti-component, often running for extended periods. As a resu\ft, effective prevention wi\f\f require the p\fanning of actions together, and the sharing of funding or other resources according to the strengths and abi\fities of each partner.

It is important to understand the different activities and agendas of the various sectors working to prevent intimate partner and sexua\f vio\fence, and to define c\fear ro\fes and contributions according\fy. Working in partnership to prevent intimate partner and sexua\f vio\fence has many benefits, and wi\f\f enhance the understanding of how such vio\fence affects different sectors, and how each sector can contribute to its preven - tion.

Many of the recognized potentia\f so\futions re\fy upon good working partnerships as intimate partner and sexua\f vio\fence affect peop\fe at every stage of \fife. Estab\fishing joint working practices and ensuring that processes are in p\face from the beginning for continued co\f\faboration wi\f\f he\fp to increase ownership and engagement in the issue, and wi\f\f he\fp to ensure that the approaches taken are both comprehensive and sustain- ab\fe. According\fy a\f\f the steps out\fined in this chapter shou\fd be conducted as a partnership of a\f\f the key agencies which are \fike\fy to inc\fude:

l the hea\fth sector ( Bo\b 6); l those working with chi\fdren and young peop\fe, inc\fuding the educationa\f sector; l the po\fice and crimina\f justice sector; l \foca\f government; l communities; and l nongovernmenta\f organizations. 59 c\faPter 4. imP roving Programme Planning and evaluation FIGURE 2 Steps for implementing intimate partner and se\bual violence prevention policies and programmes STEP \f: GETTInG STARTEd n Identify key partners and develop partnership working arrangements n Develop a shared vision across all sectors of how intimate partner and sexual violence can be prevented n Develop skills and capacity in leadership and advocacy STEP 2: d EFInE A nd dESCRIBE THE n ATuRE OF THE PROBl EM n Define intimate partner and sexual violence, and describe the size and nature of each problem n Develop capacity to assess health needs and health impacts STEP 3: IdEn TIFy POTE nTIA l ly EFFECTIVE PROGRAMMES n Identify potentially effective programmes with reference to the nature of the problem and to the evidence base for prevention STEP 4: d EVElOP POlICIES A nd STRATEGIES n Agree upon a framework for joint policy and strategy development, and prioritize effective programmes STEP 5: CREATE An ACTIO n PlAn TO EnSuRE dElIVERy n Agree upon the process and timetable for implementation n Agree upon and define the roles and responsibilities of partners n Develop professional skills, undertake further training and establish effective networks STEP 6: EVA luATE A nd SHARE l EARnInG n \flan and implement appropriate evaluation n Learn – and then share evidence and promising practice Factors that are \fike\fy to increase the \fike\fihood of success in this area are the deve\fop- ment of a shared vision and objectives, c\fear \feadership and estab\fished ro\fes as part of a joint strategy and action p\fan that brings together a\f\f these and other partners.

develop a shared vision across all sectors of how intimate partner and sexual violence can be prevented \breating a c\fimate for prevention wi\f\f require a c\fear and shared vision of how primary prevention programmes can stop intimate partner and sexua\f vio\fence occurring in the first p\face. Primary prevention wi\f\f necessari\fy invo\fve different sectors and new ways of working together. As part of this, a good “vision statement” can influence peop\fe in different sectors and encourage their participation and shou\fd be:

l created by a representative group of senior peop\fe from key sectors, organizations and affected groups and/or popu\fations; 60 Preventing intimate Partner and sexual violence against women l a statement of a future that everyone wants; l a statement that shows the benefits for a\f\f who participate; l easy to communicate and understand; l easi\fy remembered by the widest range of peop\fe; l verifiab\fe in terms of knowing when it has been achieved; and l reviewed to check that it is strategica\f\fy a\figned with the re\fevant po\ficies.

develop skills and capacity in leadership and advocacy Leadership ski\f\fs are essentia\f in supporting those who wi\f\f turn the shared vision for the prevention of intimate partner and sexua\f vio\fence into a rea\fity. Good \feadership can a\fso he\fp sustain continued ownership of a programme by a\f\f sectors and organiza - tions. This wi\f\f be vita\f if programmes are to be sustained \fong enough to achieve the behaviour changes required to reduce the rates of intimate partner and sexua\f vio\fence.

Leadership ski\f\fs inc\fude:

l strong communication and interpersona\f ski\f\fs; l an abi\fity to accommodate mu\ftip\fe viewpoints and work co\f\faborative\fy across differ- ent sectors with different and sometimes competing needs and interests; l an abi\fity to adapt to emerging and changing needs; l maintaining a focus on the rea\fistic de\fivery of outcomes; and l he\fping peop\fe to \fook for and share creative so\futions.

Intimate partner and sexua\f vio\fence often disproportionate\fy affect disadvantaged and vu\fnerab\fe peop\fe who, due to the nature of the vio\fence and their re\fative \fack of power in society, may find it difficu\ft to speak out. \bdvocac\f invo\fves working with and on beha\ff of those who experience intimate partner and sexua\f vio\fence to cha\f\fenge the socia\f and cu\ftura\f acceptabi\fity of such forms of vio\fence and to promote their preven - tion.

Advocacy can be defined as raising the awareness of an issue for the purpose of affect- ing the po\ficies, programmes and resources devoted to it. It is a fundamenta\f compo- BOX 6 The role of the health sector in primary pre\fention The health sector has a key role to play in creating a climate for primary prevention by:

n sharing knowledge and raising awareness of the acute and long-term damage to health caused by intimate partner and sexual violence; n raising awareness of the wider health impacts of intimate partner and sexual violence on the development of children and adults, and on the functioning of families and communities; and n applying and sharing public health skills including knowledge of the patterns of, and risk and protective factors for, intimate partner and sexual violence; the critical appraisal of evidence; a whole-population focus; evaluation skills; and working in partnership. 61 nent of efforts to prevent intimate partner and sexua\f vio\fence, and is often needed when there is insufficient \feadership and/or \fitt\fe or no preventive action being taken. Such advocacy can be carried out by a range of individua\fs and sectors inc\fuding vo\funtary community groups, and nongovernmenta\f and governmenta\f organizations.

Government-sponsored advocacy campaigns can inform peop\fe about the magnitude, determinants and consequences of intimate partner and sexua\f vio\fence in their country and about how these can be prevented. They can a\fso correct pub\fic misconceptions surrounding the causes and preventabi\fity of such vio\fence. Informationa\f campaigns can then accompany prevention efforts, high\fighting for examp\fe the unacceptabi\fity of vio\fence against women, or the importance of \faws to reduce the harmfu\f use of a\fco- ho\f, or the need to ensure women’s rights to property and inheritance. The \faunch of new po\ficies, programmes, research findings or pub\fications on intimate partner and sexua\f vio\fence often provides a good opportunity for governments to conduct advo- cacy efforts. Research findings can be trans\fated into simp\fe figures and pie charts and used to deve\fop advocacy messages for different target audiences. Partnerships between researchers and advocates can be usefu\f in this regard.

Advocacy campaigns can a\fso be bui\ft around high-profi\fe events on the g\foba\f ca\fendar, such as the Internationa\f day for the e\fimination of vio\fence against \bomen (25 November) and the annua\f 16 Days of activism against gender vio\fence campaign 1 (25 November –10 December). Invo\fving prominent pub\fic figures and the \foca\f and nationa\f media in campaigns bui\ft around these events can boost their impact. Advocacy can be made more powerfu\f by using \foca\f data, combined with internationa\f agreements such as the United Nations reso\futions on the e\fimination of vio\fence against women, human-rights conventions, and Wor\fd Hea\fth Assemb\fy and WHO regiona\f committee reso\futions on vio\fence prevention.

Another approach to raising the profi\fe of intimate partner and sexua\f vio\fence preven- tion efforts is to identify and cu\ftivate a \feading figure in business, po\fitica\f or com- munity \fife who has demonstrated a strong and authentic interest in championing the prevention of such vio\fence. To be effective this champion must: l c\fear\fy exp\fain the benefits of preventing intimate partner and sexua\f vio\fence, and encourage others to co\f\faborate in achieving this; l have credibi\fity or authority at nationa\f, regiona\f and \foca\f \feve\fs; l maintain good re\fationships with peop\fe across different organizations, sectors and areas; and l work co\f\faborative\fy with a\f\f sectors that share the goa\f of preventing intimate partner and sexua\f vio\fence.

Nongovernmenta\f organizations a\fso p\fay an important advocacy ro\fe in he\fping to prevent intimate partner and sexua\f vio\fence. Their focus may vary and inc\fude, for examp\fe, \fobbying the government to adopt new \fegis\fation and po\ficies. Many non- governmenta\f organizations a\fso raise awareness at community \feve\f and may run edu- cation programmes to support this work. c \faPter 4. imP roving Programme Planning and evaluation 1 www.cwg\f.rutgers.edu/16days/home.htm\f 62 Preventing intimate Partner and sexual violence against women 4.2 Step 2: define and describe the nature of the problem define intimate partner and sexual violence, and describe the size and nature of each problem I n many pa r ts of t he wor\fd, i n for mat ion on i nt i mate pa r t ner and sexua \f v io\fence is \fack - ing. As a resu\ft, po\ficies and p\fans to address such vio\fence are often made on the basis of very \fitt\fe information. For examp\fe, po\ficies are sometimes created in response to sing\fe cases that create significant pub\fic and po\fitica\f concern, or are based upon issues that receive a \fot of media attention. However, deve\foping po\ficy on this basis can resu\ft in misdirected p\fans that focus on the symptoms rather than the under\fying causes, and that consequent\fy have no impact on the prob\fem. Myths and misconceptions sur- rounding the risk factors, the characteristics of victims and perpetrators, and the \fike\fy effects of such vio\fence may a\fso exist. It is on\fy through obtaining accurate informa- tion and rep\facing conjecture with facts that such misconceptions can be dispe\f\fed. The study of how often and for what reasons a hea\fth prob\fem occurs in specific groups of peop\fe is known as “epidemio\fogy” and epidemio\fogica\f information is crucia\f in p\fan- ning and eva\fuating strategies for the prevention of intimate partner and sexua\f vio\fence ( Bo\b 7 ).

Idea\f\fy, intimate partner and sexua\f vio\fence wi\f\f be defined by way of authoritative epi- demio\fogica\f and other reports based upon regu\far\fy updated statistics on the nature, extent and consequences of the prob\fems, and their under\fying causes and risk fac- tors. Such reports wi\f\f provide the basis for designing prevention efforts and a base\fine against which to measure prevention effectiveness. Data co\f\fection shou\fd be conducted with input from experts inc\fuding epidemio\fogists, statisticians and socia\f scientists.

BOX 7 What are the benefits of accurate epidemiological information? Accurate epidemiological information on intimate partner violence and sexual violence, and on their causes and consequences, can contribute directly to preventing such forms of violence by:

n making visible and providing a quantitative description of a problem that for various cultural and social reasons may otherwise remain hidden; n providing ongoing and systematic data on the incidence, causes and consequences of intimate partner and sexual violence at local, regional and national levels; n suggesting priority prevention strategies based on those risk factors for perpetrating and experiencing intimate partner and sexual violence identified in epidemiological studies; n enabling the early identification of emerging trends and problem areas in intimate partner and sexual violence so that appropriate prevention programmes can be established; n giving an overview of the geographical distribution of such violence that can help in targeting programmes aimed at reducing the number of new cases, and in planning the provision of future victim-support services; n providing a baseline against which the impact of prevention efforts can be measured; and n monitoring changes over time in the prevalence and characteristics of intimate partner and sexual violence, and in their associated risk factors. 63 Estab\fishing the precise nature and extent of intimate partner and sexua\f vio\fence in a specific setting wi\f\f require comp\fementing the conceptua\f definitions provided in Chapter \f of this document with operationa\f definitions. An operationa\f definition is a c\fear, concise and detai\fed specification of the concept of interest that enab\fes its meas- urement. This he\fps to ensure that different peop\fe in different settings and at different points in time wi\f\f be measuring the same precise\fy defined phenomenon. Further infor- mation on the deve\fopment and use of operationa\f definitions is provided in Anne\b A.

Different types of data are needed to produce information for different purposes, inc\fuding: describing the magnitude and impact of intimate partner and sexua\f vio- \fence; understanding the risk factors invo\fved; and determining the efficacy of preven- tion programmes. Ta b l e 7 \fists a number of data categories a\fong with the potentia\f sources of such data and the type of information that can typica\f\fy be co\f\fected for each category.

As shown in Ta b l e 7, potentia\f data sources range from individua\f, agency or institu- tiona\f records, to \foca\f programme, community and government records and to pop- u\fation-based and other surveys and specia\f studies. A\fmost a\f\f these sources inc\fude basic demographic information (such as age and sex). Some sources (such as medica\f records, po\fice records, death certificates and mortuary reports) inc\fude information specific to the vio\fent event or injury. Data from emergency departments, for examp\fe, may provide information on the nature of an injury, how it was sustained, and when and where the incident occurred (though often not on the perpetrator invo\fved). Data co\f- \fected by the po\fice may inc\fude information on the re\fationship between the victim and the perpetrator, whether a weapon was invo\fved, and other circumstances re\fated to the event. Surveys and specia\f studies can provide detai\fed information about the victim or perpetrator, and their background, attitudes, behaviours and possib\fe previous invo\fve- ment in vio\fence. Such sources can a\fso provide a truer estimate of the preva\fence of vio\fence compared to records of what is reported to the po\fice or other agencies.

The majority of victims of intimate partner and sexua\f vio\fence do not seek he\fp. Inju- ries and other physica\f and menta\f hea\fth prob\fems resu\fting from these types of vio- \fence are either se\ff-treated or treated by primary care or other hea\fth providers such as pharmacists – without necessari\fy an awareness of the under\fying cause. Large-sca\fe popu\fation surveys are therefore another particu\far\fy important source of data as they provide a more comp\fete picture of the occurrence and consequences of intimate part- ner and sexua\f vio\fence. In some settings, an approach has been to forge internationa\f partnerships with academic institutions and nongovernmenta\f organizations to obtain funding for an in-country survey or study, a\fthough efforts to engage governments in nationa\f or sma\f\fer surveys are a\fso under way. In many countries, however, accurate\fy describing the nature of the prob\fem remains a major difficu\fty, especia\f\fy if there is \fitt\fe routine data or previous studies to draw upon. Time pressures in some settings can a\fso mean that “shortcuts” must be taken. In these situations, information from internationa\f studies in simi\far settings can be used. For examp\fe, resu\fts obtained from comparab\fe countries in the WHO Mu\fti-country study on \bomen’s hea\fth and domestic vio\fence against \bomen (G a rc ia- Moreno et a \f., 2 0 05 ), t he Demog r aph ic a nd He a \ft h Su r - veys or in other studies presented in Chapter \f of this document can be used. c \faPter 4. imP roving Programme Planning and evaluation 64 Preventing intimate Partner and sexual violence against women develop capacity to assess health needs and health impacts Most countries \fack routine\fy avai\fab\fe information on intimate partner vio\fence and sexua\f vio\fence. Where this is the case, there are a number of options:

l Dedicated surveys on intimate partner and sexua\f vio\fence provide the most re\fi- ab\fe and in-depth data as these can co\f\fect information on preva\fence and frequency, impact on hea\fth, contextua\f factors, and on risk and protective factors. However, they are cost\fy and not a\fways easy to repeat. Integrating a modu\fe on intimate partner and sexua\f vio\fence into existing information co\f\fection systems and surveys to a\f\fow the sustainab\fe and repeated co\f\fection of data is a usefu\f a\fternative, but consideration needs to be given to ensuring ethica\f and safety standards (E\f\fsberg & Heise, 2005).

\baution is required, however, as integrating such modu\fes into very broad hea\fth surveys may resu\ft in underestimates of the prob\fem. Adding on\fy a sma\f\f number of questions on vio\fence is not recommended as this has been shown to \fead to such underestimates and other mis\feading resu\fts. If new research is being deve\foped, it is important to obtain specia\fized advice and to fu\f\fy consider safety, confidentia\f- ity, ethica\f and other issues. WHO has deve\foped guidance on the ethica\f and safety considerations in researching intimate partner vio\fence (WHO, 2001). A manua\f for researchers and activists on researching vio\fence against women is a\fso avai\fab\fe and provides practica\f advice on imp\fementing surveys and other vio\fence against women studies (WHO/ PATH, 2005).

l Data from registers that re\fy upon professiona\f or pub\fic reporting (for examp\fe, hospita\f or po\fice data) are a potentia\f\fy usefu\f ongoing source of information, and have been used to monitor trends in chi\fd abuse in the United States. However, cau- tion shou\fd be taken in interpreting this data, as they are subject to changes in report- TABLE 7 Potential sources of data and information by data category d ata category Potential data sources E\bamples of collected information MortalityDeath certificates; vital statistics registries; and reports from mortuaries, medical examiners or coroners Individual characteristics; cause of death; and time, place and location of death Morbidity and health related Hospital, clinical and medical records Information on disease, injuries, and physical or mental health; and on the circumstances and severity of injury Self-reported attitudes, beliefs and practices Surveys; focus groups; and the media Information on attitudes, beliefs and practices; the experiencing and perpetration of violence; exposure to violence in the community and home; and risk behaviours Community based Demographic records; and local government records \fopulation counts; income levels; educational levels; and unemployment rates Criminal data \folice records; judiciary records; prison records; and crime laboratories Offence type; characteristics of offenders and victims; and circumstances of event Socioeconomic Institutional or agency records; and special studies Health expenditures; use of services; access to health care; costs of treatments; personal and household income; and distribution of income Policy or legislative Government and legislative records Information on laws and decrees; and on institutional policies and practices Adapted from: Dahlberg & Krug (2002) 65 ing po\ficies and tend to under-represent an issue when compared to the findings of representative popu\fation surveys. l Rapid appraisa\f techniques 1 are \fow-cost methods for describing the nature of a prob\fem and can be quick\fy imp\femented in a \foca\f setting. They usua\f\fy invo\fve a combination of methods – for examp\fe, interviews with stakeho\fders; focus groups; mapping exercises; and co\f\fection and ana\fysis of avai\fab\fe data from hea\fth faci\fi - ties, nongovernmenta\f organizations and \foca\f authorities (Pa\fmer, 1999). The who\fe approach c a n be comp\feted i n 2 – 3 mont h s, a nd t he resu \ft s prov ided to t he com mu n it y to stimu\fate \foca\f action and inform p\fanning.

I n order to advocate for i nvest ment i n prevent ion, i n for m t he desig n of i nt i mate pa r t ner and sexua\f vio\fence prevention po\ficies and programmes, and monitor the effects of such programmes, it is high\fy usefu\f to produce information in the fo\f\fowing areas: l Preva\fence and incidence of intimate partner and sexua\f vio\fence – how com- mon\fy they affect the popu\fation and their trends over time.

l Distribution of such vio\fence – for examp\fe, by age, sex, socioeconomic status, set- ting and \focation.

l Hea\fth consequences – mortalit\f : the number of deaths re\fated to intimate part- ner or sexua\f vio\fence. Data can be used to monitor changes over time in fata\f vio- \fence, identify individua\fs and groups at high risk of death from such vio\fence and to make comparisons within and between countries. Morbidit\f/health outcomes :

data on diseases, injuries and other hea\fth conditions arising from being a victim of intimate partner and/or sexua\f vio\fence; and on the impact of these on the abi\fity to work and carry out basic \fiving activities.

l Risk factors – at the individual \feve\f, for examp\fe, sex, young age, history of ma\f- treatment, a\fcoho\f misuse and conduct and antisocia\f behaviours; at the relation- ship \feve\f, for examp\fe, mu\ftip\fe partners /infide\fity and marita\f satisfaction/discord; at the communit\f \feve\f, for examp\fe, re\fated to popu\fation characteristics and \fev- e\fs of income, education and unemp\foyment; and at the societal \feve\f, for examp\fe, socia\f norms re\fated to gender re\fationships and to the acceptabi\fity of vio\fence and \faws and po\ficies re\fated to vio\fence.

l Protective factors – for examp\fe, schoo\f attendance, vo\funteering and community action and invo\fvement in group activities. This can a\fso inc\fude the mapping of services and po\ficies that are effective in prevention.

l Crime data – on the characteristics and circumstances of vio\fent events and vio\fent offenders that come to the attention of the po\fice; on the progress of cases through the courts; and on the proportion of events that end in conviction.

l Economic data – re\fated to the costs of treatment and socia\f services; the economic burden on hea\fth care systems; and the possib\fe savings arising from prevention pro- grammes.

l Po\f ic ie s a nd \feg i s \f at ion – inc\fuding the introduction of po\ficies; and the enactment and enforcement of \faws that address a popu\fation’s experience of risk factors (such as a\fcoho\f misuse). c \faPter 4. imP roving Programme Planning and evaluation 1 For further information, inc\fuding free resources and examp\fes of rapid-appraisa\f techniques, see: The Managers E\fectronic Resource \bentre – part of Management Services for Hea\fth at: http://erc.msh.

org/mainpage.cfm?fi\fe=2.1.1b.htm&modu\fe=chs&\fanguage=Eng\fish 66 Preventing intimate Partner and sexual violence against women Avai\fab\fe data wi\f\f be essentia\f in informing the next steps. Given \fimited resources, the data can he\fp to decide upon priority groups (for examp\fe, by showing who has the highest number of risk factors for perpetration) and on potentia\f entry points for action.

This wi\f\f then inform the next step of identifying interventions.

4.3 Step 3: Identify potentially effective programmes identify potentially effective programmes with reference to the nature of the problem and to the evidence base for prevention In Chapter 3 , a summary of the evidence for the effectiveness or otherwise of a broad range of programmes was presented. As noted, most of this evidence comes from HI\b and it is therefore important that po\ficy-makers and p\fanners decide what is \fike\fy to be most re\fevant and effective in their own \foca\f context, informed by avai\fab\fe data.

Evidence of effectiveness in a high-income setting may for examp\fe assume access to resources and infrastructure that wou\fd not be avai\fab\fe in a \fow-income setting, so a\f\f programmes wi\f\f need to be adapted to the \foca\f context. In addition, some programmes which are often imp\femented (such as rape education programmes) have been shown to be ineffective.

Given the scarce resources avai\fab\fe for addressing intimate partner and sexua\f vio- \fence, it is crucia\f that programmes are carefu\f\fy p\fanned. Imp\fementation shou\fd a\fso be stepwise, carefu\f\fy monitored and on a sufficient sca\fe and over a \fong enough period to a\f\fow for the documentation and eva\fuation of changes. \bapacity to p\fan, monitor and eva\fuate programmes wi\f\f need to be deve\foped in order to ensure this happens.

Prior to the se\fection and imp\fementation of specific programmes there wi\f\f be consid- erab\fe advantages in undertaking a process of “formative eva\fuation”. This process wi\f\f he\fp to better understand the needs and resources avai\fab\fe in a particu\far setting, and to inform the best approach to take. Together with an understanding of risk factors, this process wi\f\f he\fp to guide the se\fection and deve\fopment of prevention activities (such as whether to emp\foy a media campaign, po\ficy initiative or other prevention strategy).

Once a programme is identified, the broader issues to address wi\f\f typica\f\fy inc\fude:

l Wi\f\f the perceived needs of the community be met?

l Are the \fanguage and cu\ftura\f approaches se\fected respectfu\f of the anticipated audi- ence?

l Are resources being used in the most efficient way to support activities that achieve prevention goa\fs?

l \ban the programme be improved?

l Are there ways to make the programme more effective?

l How are peop\fe \fike\fy to respond to prevention efforts?

l How are peop\fe \fike\fy to change and benefit from these efforts?

l What rea\f effect is the programme \fike\fy to have on preventing intimate partner and sexua\f vio\fence?

Sources of information wi\f\f inc\fude \foca\f or nationa\f statistics; surveys or questionnaires; focus groups, interviews, or debriefings; and pre-testing or tria\f runs. The answers to the questions above can then inform the theory under\fying the programme, its goa\fs 67 and activities and the imp\fementation of the programme itse\ff. Engaging in formative eva\fuation can be a high\fy practica\f and cost-effective way of ensuring the qua\fity, fea- sibi\fity and re\fevance of a prevention approach or message before committing resources on a broad sca\fe.

Different sectors can then work in partnership to identify which programmes are most appropriate for their area, and those approaches that cou\fd be deve\foped further to maximize their prevention impact. Taking the sectors \fisted above in section 4 .\f, a t y pi- ca\f distribution of activity areas might then be:

l The hea\fth sector – materna\f menta\f hea\fth, parenting, a\fcoho\f misuse treatment; l Those working with chi\fdren and young peop\fe, inc\fuding the educationa\f sector – vio\fence and abuse prevention ski\f\fs for chi\fdren and young peop\fe, socia\f and emotiona\f ski\f\fs deve\fopment, ear\fy intervention for conduct and emotiona\f dis - orders; l T he p o\f ice a nd c r i m i n a \f ju st ice se c tor – deve\fopment and enforcement of \fegis\fa- tion; l Loca\f government – reduce access to a\fcoho\f and poverty-reduction programmes; and l Communities and nongovernmenta\f organizations – change cu\ftura\f norms, microfinance projects, education programmes.

4.4 Step 4: develop policies and strategies agree upon a framework for joint policy and strategy development and prioritize effective programmes Working in partnership, efforts wi\f\f need to be made to incorporate the outcomes of the three previous steps of the framework into po\ficies and strategies for the prevention of intimate partner and sexua\f vio\fence. Deve\foping such po\ficies and strategies wi\f\f be easier where there is c\fear \feadership driving the process, and where working arrange- ments are deve\foped in meaningfu\f partnership with a\f\f key stakeho\fders as described in Step \f . Deve\foping a joint po\ficy and strategy wou\fd then invo\fve the fo\f\fowing sequen- tia\f steps (see a\fso: Schopper, Lormand & Waxwei\fer, 2006):

l Agree upon a shared vision and overa\f\f framework for preventing intimate partner and sexua\f vio\fence – for examp\fe, the promotion of equa\f and hea\fthy re\fationships.

l \b\fear\fy set out the nature and importance of the prob\fem – for examp\fe, how many peop\fe are affected, their age, sex, risk factors, and the impact on individua\fs (inc\fud- ing on their hea\fth and education), and on fami\fies, communities and the economy.

l Pr ior it ize effect ive prog ram mes – work w it h a\f \f par t ners to ident ify a sma\f \f set of pr i - ority programmes for further action, se\fected for their appropriateness in a particu\far setting.

l Ensure that po\ficy \feads to action – for examp\fe, by gaining endorsement from a\f\f the different partners, setting c\fear priorities and reaching high-\feve\f agreement on the steps to take.

This sequentia\f process of deve\foping a specific po\ficy and strategy for the prevention of intimate partner and sexua\f vio\fence wi\f\f a\fso he\fp to engage different sectors and c \faPter 4. imP roving Programme Planning and evaluation 68 Preventing intimate Partner and sexual violence against women raise awareness of the issues among decision-makers and the genera\f pub\fic. Addition- a\f\fy, efforts are \fike\fy to be far more powerfu\f if the re\fevant activity areas are integrated into wider po\ficy and strategies (for examp\fe, on materna\f or menta\f hea\fth, a\fcoho\f use, education, crime, economic deve\fopment and urban p\fanning). The degree to which a programme cou\fd feasib\fy be incorporated into such mainstream activities is one of the guiding princip\fes that shou\fd be considered and agreed upon by the key partners as part of the prioritizing process ( Bo\b 8).

4.5 Step 5: Create an action plan to ensure delivery agree upon the process and timetable for implementation Fo\f\fowing the deve\fopment of po\ficies and an agreed strategy it wi\f\f be necessary to for- mu\fate an action p\fan – once again idea\f\fy in partnership with other sectors. An action p\fan provides the detai\f of how to de\fiver the programmes identified in the strategy.

Some strategies contain an action p\fan at the end of the written report. The fo\f\fowing suggested out\fine areas might usefu\f\fy be adapted for the deve\fopment of an action p\fan according to \foca\f context and setting:

l Timetab\fe: specify a timetab\fe for managing a\f\f the various steps required to achieve the programme aims (for examp\fe, training staff before they provide parenting ski\f\fs to mothers).

BOX 8 Factors to consider when prioritizing programmes n \bhat is the size of the population that would benefit from the programme?

n \bhat is the estimated level of health gains among this population after the programme – in the short, medium and long term?

n \bhat would be the wider gains of the intervention – such as educational, social or economic improvements?

n How feasible is mainstreaming the programme within existing established services (for example, for identifying and managing maternal depression as part of antenatal care services)?

n \bhat are the approximate costs of implementing the programme based upon current capacity and resources?

n Is there any potential for harm caused by implementing the programme?

n Does the implementing agency have the necessary capacity? If not, can it be built?

n Is the community ready for the programme or are certain steps needed to enhance community readiness?

n Are there any financial or cultural barriers, or other obstacles, to implementing the programme, and how easy would it be to overcome these?

n Are there any opportunities available that would make a specific programme easier to implement? For example, new policy, funding or resources available in a particular activity area. 69 l Define ro\fes and responsibi\fities, and identify the resources and ski\f\fs need- ed: nominate a \fead from each sector to take responsibi\fity for ensuring the de\fivery of each recommended programme and ensure they have the capacity and ski\f\fs to do it. Different partners wi\f\f have different resources and capacities that may need to be avai\fab\fe at various points in the project. It is important to agree who wi\f\f provide what input at which time. For examp\fe, hea\fth-sector staff may train education-sector staff to teach vio\fence prevention ski\f\fs whi\fe \foca\f community deve\fopment workers may be responsib\fe for ensuring that vu\fnerab\fe chi\fdren attend events.

l P\fan communications: it is important to agree on a p\fan for communicating the progress and expected benefits of the programme. This may be achieved through the use of media, community meetings or discussions with key individua\fs (for examp\fe, community \feaders).

l Engage communities: community engagement emphasizes the ro\fe of individu- a\fs and communities as agents of change, rather than passive programme recipients.

It a\fso p\faces a priority on ensuring community ownership and influence over the change process. A\f\f programmes stand to benefit significant\fy from the increased support and receptiveness to change that resu\fts from successfu\f community engage- ment.

agree upon and define the roles and responsibilities of partners \b\fear\fy identifying who wi\f\f take the \feadership ro\fe, and coordinating the specific ro\fes and responsibi\fities of each of the sectors and organizations invo\fved wi\f\f great\fy faci\fitate the imp\fementation of the agreed strategies and programmes. Each sector shou\fd p\fay its part in coordinating efforts so as to ensure synergy and avoid inadvertent\fy under- mining the efforts of other partners. This wi\f\f a\fso require identifying the resources and ski\f\fs needed to ensure effective imp\fementation.

develop professional skills, undertake further training and establish effective networks Deve\foping the necessary ski\f\fs and training is one of the e\fements that may need to be put in p\face. There is a need genera\f\fy to bui\fd capacity for the prevention of intimate partner and sexua\f vio\fence through, for examp\fe, inc\fuding these issues in the cur- ricu\fa for nurses, doctors and pub\fic hea\fth practitioners, as we\f\f as in-service training for hea\fth personne\f. In addition, there may be a need, for examp\fe, for technica\f ski\f\fs to carry out research and set up survei\f\fance systems, or for advocacy and \feadership work, inc\fuding ski\f\fs in fundraising, media and communications.

Specific training is a\fso \fike\fy to be needed in order to imp\fement whichever of the strategies discussed in Chapter 3 are se\fected for the prevention of intimate partner and sexua\f vio\fence. Where feasib\fe, organizations shou\fd support training with other partners, for examp\fe, through exchange visits, mentoring and co\f\faboration with other institutions. Such technica\f exchanges can speed up the uptake of best practices, he\fp formu\fate po\ficy agendas and create effective networks. These exchanges can a\fso \fink practitioners with potentia\f technica\f resources and other forms of assistance. c \faPter 4. imP roving Programme Planning and evaluation 70 Preventing intimate Partner and sexual violence against women 4.6 Step 6: Evaluate and share learning Plan and implement appropriate evaluation Monitoring and eva\fuation – and dissemination of the resu\fts – are core components of the pub\fic hea\fth approach to preventing intimate partner and sexua\f vio\fence out\fined in the Introduction to this document. Both are necessary to promote evidence-based programme p\fanning and imp\fementation, and programme eva\fuation is key to pro- gramme improvement. Eva\fuations of a programme can be designed to answer different types of questions – which are \fike\fy to vary depending upon the approach taken and the stage reached. A\fthough incorporating eva\fuation as an integra\f part of a programme a\f\fows for constant monitoring and adjustment as needed, different types of eva\fua- tion may be more appropriate at different stages as different questions arise. Eva\fuation components can then be added or modified as the programme evo\fves. These compo- nents can be divided into the fo\f\fowing three principa\f categories of eva\fuation:

l Process eva\fuation to document whether a programme can be (or is being) imp\fe- mented as p\fanned; l Outcome eva\fuation to deter m i ne whet her a prog r a m me h as t he i ntended ef fec t on intimate partner and sexua\f vio\fence (or on its risk and/or protective factors); and l Economic eva\fuation to identify the resources needed to conduct, rep\ficate or expand programmes, and to assess the costs and benefits to he\fp determine if the benefits of a programme outweigh its costs.

Process eva\fuation assesses specific e\fements of a programme (such as the curricu\fum, training activities or practica\f support) to find out if it is being de\fivered as p\fanned, and to identify gaps between its intended and actua\f de\fivery. It can, for examp\fe, identify whether the programme is reaching the intended audience, describe what is provided (and how we\f\f) and assess the perceptions of audiences and participants and their \feve\f of satisfaction with the programme. Process eva\fuation can a\fso document programme accomp\fishments.

Information is obtained both on what is working we\f\f and what may require refinement.

\bha\f\fenges and barriers can then be identified as they emerge to permit improvements in future responses, or to document those that cou\fd not be overcome. If a programme is eventua\f\fy viewed as successfu\f in preventing intimate partner and sexua\f vio\fence, then documenting exact\fy what occurs a\f\fows the approach to be continued in the same way. It a\fso increases the \fike\fihood that other groups wi\f\f be ab\fe to achieve simi\far suc- cess if the programme is used e\fsewhere. If there has been on\fy \fimited or no success, then process eva\fuation wi\f\f he\fp to determine if this can be attributed to prob\fems with the way in which the programme was imp\femented. This wi\f\f great\fy reduce the risk of discarding a potentia\f\fy effective programme simp\fy because of imp\fementation issues.

With newer programmes, the questions in a process eva\fuation might usefu\f\fy focus on aspects of estab\fishing and de\fivering programme components. With more estab\fished programmes, eva\fuation cou\fd focus on assessing whether or not the current processes are working optima\f\fy. With ongoing programmes, process eva\fuation may be incor- porated as a routine part of dai\fy activities so that the programme can be continua\f\fy monitored and reviewed. The frequency of monitoring can a\fso be adjusted as needs and resources change. 71 Regard\fess of how fu\f\fy estab\fished a programme is, the first step in eva\fuating whether it is being imp\femented as intended is to identify the key components (for examp\fe, critica\f content areas or activities required to imp\fement a prevention curricu\fum, make changes in po\ficy, or conduct a media campaign). The second step is to then deve\fop criteria against which the degree of imp\fementation of these components can be judged.

In many cases, these criteria wi\f\f a\fready have been discussed as part of the priority- setting process – as out\fined above in section 4.5. Fina\f\fy, information wi\f\f need to be co\f\fected to monitor and document each component to determine how we\f\f the actua\f imp\fementation matches the proposed p\fan.

Anne\b B provides examp\fes of the type of e\fements that can be considered during the process eva\fuation of different types of prevention activities, and \fists some of the poten- tia\f sources of this information. U\ftimate\fy, the process eva\fuation methods se\fected wi\f\f need to be informed by the type of programme and the questions to be answered – and the degree of confidence needed in the resu\fts, as we\f\f as avai\fab\fe resources and the \feve\f of expertise required by different eva\fuation approaches.

Outcome eva\fuation – the u\ftimate outcomes that intimate partner or sexua\f vio\fence prevention approaches are trying to achieve are reductions in the perpetration of these types of vio\fence, or reductions in the injuries, non-injury hea\fth impacts and deaths they cause. However, assessing these \fong-term changes can be difficu\ft and cost\fy. It often takes time to observe such changes, requiring that data be co\f\fected over \fong periods. This is particu\far\fy true when the behaviour or outcome of interest occurs re\fative\fy infrequent\fy. A more rea\fistic objective for many prevention programmes may therefore be to change the risk factors that contribute to vio\fence ( Chapter 2) with the u\ftimate goa\f of preventing vio\fent behaviour. Whatever outcomes are se\fected shou\fd be based upon the theory under\fying the prevention approach; the programme objec- tives; the extent to which the programme and its activities are fu\f\fy estab\fished; and the resources avai\fab\fe for eva\fuation.

Outcome eva\fuation specifica\f\fy seeks to determine if a programme was successfu\f in bringing about the intended changes (for examp\fe, in risk factors, behaviours, or \feve\fs of injuries, i\f\fness and deaths). The main concern when designing an outcome eva\fua- tion is to ensure that any a\fternative exp\fanations for the changes observed can be ru\fed out to provide confidence that such changes were due to the programme and not to some other factor. This wi\f\f he\fp to ensure that resources are not wasted on ineffective or \fess-effective programmes.

One way to address this is to co\f\fect data on the outcome(s) of interest at mu\ftip\fe time interva\fs before, after (and sometimes during) imp\fementation to see at what point changes occur (“time-series design”). This approach is often used to eva\fuate broad- based community programmes such as po\ficy changes. The data co\f\fected are often derived from community-\feve\f sources such as community surveys or from communi- ty-\feve\f records such as the information routine\fy co\f\fected in survei\f\fance or monitor- ing systems (see section 4.2 for a fu\f\fer discussion of potentia\f data sources).

Another approach is to use “experimenta\f” and “contro\f” groups as part of a “contro\f- \fed tria\f” in which the experimenta\f group receives the programme and the contro\f group does not. A\fthough differences over time between the two groups cou\fd sti\f\f resu\ft from other events happening simu\ftaneous\fy (such as other prevention activities, po\ficy changes or community events) the use of the contro\f group(s) wi\f\f he\fp ru\fe out this possibi\fity. However, this wi\f\f re\fy upon both groups being equa\f\fy exposed to the c \faPter 4. imP roving Programme Planning and evaluation 72 Preventing intimate Partner and sexual violence against women outside events and other activities, and it is therefore necessary to monitor the degree to which each group was exposed to these externa\f influences.

It is a\fso possib\fe that the programme might “seep” into the contro\f group and distort the resu\fts. For examp\fe, students who take part in a programme might share informa - tion they \fearned with students in the contro\f group. Or a contro\f community may be exposed to a media campaign or \fegis\fation. These effects can be minimized by se\fect- ing a contro\f group that is geographica\f\fy distant from (or otherwise un\fike\fy to interact with) the experimenta\f group. For examp\fe, students may be \fess \fike\fy to interact with students from a distant schoo\f than with students in a different c\fass in the same schoo\f or in a different schoo\f in the same neighbourhood.

\bo\f\fecting data as soon as a programme ends or very short\fy afterwards wi\f\f provide information on its immediate effects. \bo\f\fecting “fo\f\fow-up” data at some interva\f (or interva\fs) after the immediate post-programme data have been co\f\fected can he\fp to determine whether changes are maintained or diminish once the programme ends.

Fo\f\fow-up data can a\fso provide information on \fater changes that were not apparent immediate\fy after the programme (sometimes ca\f\fed “s\feeper effects”). The decision to co\f\fect fo\f\fow-up data wi\f\f be guided by the questions to be answered. The capacity of the programme to co\f\fect further information from participants and the avai\fabi\fity of eva\fuation resources need to be addressed from the start.

Because of the severe\fy \fimited evidence base for effective or promising programmes for the primary prevention of intimate partner and sexua\f vio\fence, the inc\fusion of outcome eva\fuation in eva\fuation p\fanning can provide va\fuab\fe and much needed information.

Programme eva\fuations for programme improvement requires some technica\f capac- ity but can be done using existing sources of information discussed above. However, a rigorous outcome eva\fuation to bui\fd an evidence base for prevention requires scientific expertise from the very beginning of the p\fanning phase to ensure:

l the appropriate se\fection of prevention objectives; l the deve\fopment of a \fogic mode\f; 1 and l the correct choice of research design, study popu\fation(s), outcomes, data sources and methods of ana\fysis.

As a resu\ft, partnership with an academic or research institution with the re\fevant exper- tise is necessary. Bo\b 9 describes how a pub\fic hea\ft h agency reached out to a grassroots organization to improve the evidence base for one particu\far programme for preventing intimate partner and sexua\f vio\fence.

U\ftimate\fy, the se\fection of which outcome eva\fuation approach to take wi\f\f be guid- ed by the questions to be answered and the degree of certainty required in attribut- ing observed effects to the programme. The benefits of each approach wi\f\f need to be weighed against the practica\f and financia\f constraints of the programme. Differ- ent types of outcome eva\fuation may a\fso be he\fpfu\f at different programme stages.

However, whatever eva\fuation approach is used, the routine incorporation of ongoing outcome-data co\f\fection wi\f\f be needed to monitor the continued effectiveness of the programme. For examp\fe, at the beginning of a new prevention strategy, co\f\fecting pre- and post-programme imp\fementation data from participants can he\fp to determine if 1 See: \bD\b Eva\fuation Working Group: www.cdc.gov/eva\f/resources.htm #\fogic%20mode\f and \bentre for Hea\fth Services and Po\ficy Research (\bHSPR): A Resu\fts-Based Logic Mode\f for Primary Hea\fth \bare: www.chspr.ubc.ca/ fi\fes /pub\fications /2004/chspr04 -19.pdf 73 the approach \fooks promising. Positive indications wou\fd inc\fude that changes appeared to be in the right direction, of sufficient magnitude, or simi\far or better than other pro- grammes targeting simi\far outcomes. The findings can then be used to he\fp improve the programme. Later, a contro\f\fed tria\f (or randomized contro\f\fed tria\f) design can be used to provide more evidence that the programme was indeed contributing to the desired changes, thus bui\fding the evidence base on primary prevention.

Economic eva\fuation is a systematic method of assessing the impact of strategies, po\ficies and programmes on outcomes and costs (Haddix, Teustch & \borso, 2003).

Economic eva\fuation can he\fp to identify and understand the use of resources by a pro - gramme, determine what is feasib\fe given avai\fab\fe resources, a\f\fow for the more efficient use of resources, and p\fan or prioritize activities shou\fd resources be reduced or removed.

In addition, administrators, funding agencies and the pub\fic are increasing\fy concerned about how programme costs re\fate to programme benefits. Economic eva\fuation can he\fp demonstrate that a programme is worth the investment. For examp\fe, practitioners in the fie\fd of youth vio\fence have been ab\fe to te\f\f their po\ficy-makers that a programme based on nurse home visits costs US $ 6142 per fami\fy. However, by reducing the risk of youth invo\fvement in vio\fence and crime by 16% the programme generates savings per prog ra m me pa r t icipa nt of about US $ 460 0 for t he t a x payer a nd US $ 8100 for potentia\f victims (Aos, Mi\f\fer & Drake, 2006). Such cost-benefit ana\fyses require the effective - ness of a programme to have been determined beforehand. Economic eva\fuation is usu- a\f\fy carried out by researchers and requires methodo\fogica\f expertise. c \faPter 4. imP roving Programme Planning and evaluation BOX 9 Impro\fing the e\fidence base through impro\fed e\faluation Concerned with the lag between the discovery of effective programmes and their widespread adoption, the United States Centers for Disease Control and \frevention decided to reach out to sexual violence prevention programmes with demonstrated real- world feasibility in order to build a scientific evidence base for the interventions used. First, existing programmes were identified based on published and unpublished documents, web- site searches and interviews with experts, practitioners and programme agencies. \fublicly available information on each programme was supplemented with structured telephone interviews to compile summary descriptions of each programme identified.

Following an open competition for funds, four programmes, each delivered in a multi- session format, were selected based on their focus on first-time male perpetration, prior experience with evaluation, commitment to using evaluation for programme improvement, and their ability to commit staff to the evaluation process. An outside evaluator then served as a facilitator, trainer and coach to help programme stakeholders design and conduct their own evaluations. This process utilized empowerment principles to build evaluation capacity among the programmes and to generate information about programme activities.

\farticipating programmes subsequently reported substantial enhancements to their evaluation capacity, to the extent and sophistication of their evaluation practices and greater allocation of resources devoted to evaluation. Numerous examples were then provided of the successful application of evaluation findings to programme-improvement efforts.

For further information, see:

www.cdc.gov/ncipc/delta/ www.cdc.gov/ncipc/DELTA/DELTA _ AAG.pdf 74 Preventing intimate Partner and sexual violence against women learn – and then share evidence and promising practice Prevention partners interested in receiving the resu\fts of eva\fuations shou\fd be identi- fied in the preparatory stages. During the course of imp\fementing and eva\fuating the programme, other parties may a\fso have \fearned of the programme and demonstrated an interest. Presenting the resu\fts of the eva\fuation to various audiences may garner wider support from decision-makers and the community for a programme and for the \farger goa\f of preventing vio\fence. By sharing the evidence supporting the use of a par- t ic u \fa r prog ra m me w it h ot her prevent ion pa r t ner s, \foca \f capacit y ca n be bu i \ft a nd prac - tice improved. The scope, degree of detai\f and method of reporting eva\fuation resu\fts wi\f\f vary depending on the interest and needs of the various audiences.

Interna\f audiences such as staff, vo\funteers, the board of directors and other stakeho\fd - ers c\fose\fy invo\fved in the programme wi\f\f be fami\fiar with the objectives and expecta- tions of the eva\fuation and may instead want detai\fed information. In this case, a verba\f presentation (with accompanying visua\f materia\f to i\f\fustrate key points or summarize d at a) t h at a \f \fow s fo r d i s c u s s ion of t he fi nd i n g s , a nd a \ft e r n at ive i nt e r pr e t at ion s a nd app\f i - cations of the findings to the programme may be most usefu\f. Sharing and discussing the resu\fts with the interna\f audience promotes investment in the programme and in the process of eva\fuation. To avoid the misuse of eva\fuation resu\fts, it is he\fpfu\f to remind stakeho\fders of ear\fier agreements on the purpose of the eva\fuation and on how the resu\fts and recommendations wou\fd be shared and used to improve the programme.

Ve r b a \f b r i e fi n g s m ay a \f s o b e appr opr i at e fo r d i f fe r e nt ex t e r n a \f aud i e nc e s , s uc h a s \f e g i s \f a - tors and funding agencies. However, such audiences often require a written report with a description of the aim(s), rationa\fe, activities, eva\fuation design and findings of the programme, possib\fy supp\femented with anecdotes, case histories or graphics. Reports that are succinct and c\fear are more \fike\fy to be read. Depending on the audience, an executive summary or brief fact sheet may suffice, with interested individua\fs being referred to a \fonger and more detai\fed report.

\bommunicating findings to the media – and through it to the pub\fic at \farge – may require a s\fight\fy different approach. Working with the media is not comp\ficated, but it does take p\fanning. First, it is necessary to reach out to news media contacts in the com- munity to begin to bui\fd re\fationships, and to estab\fish the group as an expert source of information on intimate partner and sexua\f vio\fence. Usefu\f, time\fy and accurate new information wi\f\f capture a reporter’s attention. For examp\fe, offer \foca\f statistics if possib\fe, or show how new services might be he\fping to reduce intimate partner and sexua\f vio\fence for the benefit of the community. Deve\fop a media contact \fist (using the Internet or phone book) and then determine the best method to reach reporters with the avai\fab\fe information. Examp\fes of products for news media outreach inc\fude:

l Press re\fease – for time\fy and new information; key findings are presented at the beginning fo\f\fowed by supporting descriptive information; and it shou\fd be easi\fy formatted for sending e\fectronica\f\fy.

l \bedia advisory – used to advise reporters to attend a photo opportunity or other event.

l Letters to the Editor – for experts sharing a specific point of view; each newspaper and other media out\fet wi\f\f have guide\fines avai\fab\fe on their web site on the submis- sion process and required \fength. 75 l Opinion-Editoria\f – some newspapers accept guest opinion editoria\fs (“op-eds”) and the media web sites shou\fd again be checked for detai\fs.

l P ress k it – shou \fd be kept si mp\fe a nd i nc\fude one -page fac t sheet s, flyer s a nd cha r t s ; consideration shou\fd be given to making the press kit avai\fab\fe on\fine.

New media innovations such as b\fogs and Twitter can a\fso be usefu\f for disseminating intimate partner and sexua\f vio\fence prevention messages. The Internet has increasing- \fy become a source of up-to-date information. A brief notice can be posted to a re\fevant programme or agency web site, with a \fink provided to the fu\f\f report(s). \bo\f\faborating with co\f\feagues in the sharing of \finks to various programmes and reports may a\fso be a usefu\f option.

Sharing the programme and eva\fuation findings both within and between countries is an important e\fement of deve\foping and disseminating the know\fedge and ski\f\fs required for the prevention of intimate partner and sexua\f vio\fence. Networks require good coor - dination to connect peop\fe whose prevention activities can be distributed across a range of pub\fic and private sectors. Networks can a\fso be set up to meet \foca\f needs, and these too can \fink across sectors for the more efficient exchange of information on p\fanning and action. How eva\fuation findings are shared – and with whom – wi\f\f depend upon the focus of a programme and the community or country context. A key point in a\f\f cases is to be succinct and c\fear, and to tai\for the reporting appropriate\fy for each audience.

4.7 Key messages To implement prevention policies and programmes, the following steps are essential:

Step 1: Getting started Identify key partners and develop partnerships Develop a shared vision Develop skills and capacity in leadership and advocacy Step 2: Define and describe the nature of the problem Define intimate partner and sexual violence Describe the size and nature of each problem Develop capacity to assess health needs and health impacts Step 3: Identify potentially effecti\fe programmes with reference to the nature of the problem and the e\fidence base for pre\fention Step 4: De\felop policies and strategies Agree upon a framework for joint policy and strategy development \frioritize effective programmes Step 5: Create an action plan to ensure deli\fery Agree upon the process and timetable for implementation Agree upon and define the roles and responsibilities of partners Develop professional skills, undertake further training and establish effective networks Step 6: \b\faluate and share learning \flan and implement appropriate evaluation Learn – and then share evidence and promising practice c \faPter 4. imP roving Programme Planning and evaluation 76 Future research priorities and conclusions Future research priorities Research is urgent\fy needed to bui\fd the evidence base and address the current \fack of information on effective programmes for primary prevention. The major constraints on research in this area are the \fong time sca\fes and high costs of producing resu\fts.

However, without such research it wi\f\f be difficu\ft to deve\fop new po\ficy that is both effective and cost effective. Gathering data on the preva\fence and hea\fth outcomes of both intimate partner vio\fence and sexua\f vio\fence wi\f\f be important in settings where these data are not avai\fab\fe. It can serve to raise awareness of the existence and impact of these re\fative\fy hidden issues and to provide base\fine information that can eventu- a\f\fy be used for monitoring. To strengthen the evidence base and he\fp deve\fop urgent\fy required guide\fines and recommendations, the fo\f\fowing areas shou\fd be addressed as a priority:

l Eva\fuate promising programmes and further deve\fop and test potentia\f programmes – especia\f\fy in LMI\b where rates of intimate partner vio\fence are often high.

l Prioritize programmes and eva\fuations that focus on younger age groups since the evidence c\fear\fy suggests that this may be a more effective primary prevention approach.

l \bonsider new settings for primary prevention programmes. At present a dispropor-tionate number of interventions are designed to be de\fivered in schoo\fs, co\f\feges and universities. However, the individua\fs most at risk of perpetrating or experiencing intimate partner and sexua\f vio\fence may no \fonger be attending schoo\f or may have difficu\fty engaging in schoo\f-based activities. Far more consideration shou\fd be given to the use of other venues and settings, such as the home, hea\fth-care settings, and the faci\fities avai\fab\fe to community-based (inc\fuding faith-based) organizations and others.

l Once the effectiveness of a programme has been determined, obtain more informa-tion on the re\fative cost effectiveness and the wider socia\f and emotiona\f benefits of specific programmes in a variety of settings. The information such studies can provide is crucia\f in convincing po\ficy-makers of the importance of investing in pri- mary prevention programmes. Idea\f\fy, such cost-effectiveness eva\fuations shou\fd a\fso examine the re\fative cost effectiveness and benefits of programmes for the primary prevention of intimate partner and sexua\f vio\fence at each of the major \fife stages.

Having re\fative cost-effectiveness evidence wi\f\f then he\fp po\ficy-makers to decide which programmes to prioritize according to their beneficia\f effects at different ages and in different settings. 77 future researc\f Priorities and conclusions l Eva\fuations of different po\ficies – for examp\fe, those that combine a variety of pro-grammes – in terms of their feasibi\fity, cost and outcomes wi\f\f great\fy acce\ferate their deve\fopment and imp\fementation in other settings.

l A\fthough some evidence now exists in most regions on the preva\fence, consequences of, and risk factors for both intimate partner and sexua\f vio\fence, sti\f\f more is needed.

In particu\far, far more evidence is needed on community and societa\f \feve\f risk fac- tors, such as gender norms and gender inequa\fity; the impact of such vio\fence on chi\fd deve\fopment, hea\fth and associated outcomes using \fongitudina\f studies, espe- cia\f\fy in LMI\b; and on educationa\f and economic outcomes.

m ethodological issues l Programme deve\fopers shou\fd be encouraged to exp\ficit\fy base programmes on c\fear\fy articu\fated theoretica\f frameworks to a\f\fow the under\fying mechanisms to be identified and to make rep\fication easier.

l Determining whether or not changes in know\fedge, attitudes, be\fiefs and behavioura\f intention actua\f\fy resu\ft in reduced \feve\fs of intimate partner and sexua\f vio\fence is an urgent need. Too often, outcome eva\fuations have re\fied upon the former measures instead of measuring the effects of a programme on the actua\f \feve\fs of vio\fence.

l Fo\f\fow-up periods shou\fd be \fonger to ensure that any beneficia\f effects of pro-grammes do not fade.

l Efforts shou\fd be made to bui\fd research capacity and to work co\f\faborative\fy – inc\fud-ing with nationa\f and internationa\f agencies – to support \foca\f research and to identify the research priorities for different types of settings.

Conclusions The evidence-based prevention of intimate partner and sexua\f vio\fence is sti\f\f in its ear\fy days and much remains to be accomp\fished. At present, on\fy one approach has evidence supporting its effectiveness – and this on\fy re\fates to intimate partner vio\fence. The approach in question is the use of schoo\f-based programmes to prevent vio\fence within dating re\fationships. Evidence is, however, emerging of the effectiveness of a number of other approaches to preventing intimate partner and sexua\f vio\fence, inc\fuding micro - finance programmes for women combined with gender-equa\fity education, efforts to reduce access to (and harmfu\f use of ) a\fcoho\f and changing socia\f and cu\ftura\f gender norms. Many more approaches appear to have potentia\f, either on theoretica\f grounds or because they target known risk factors, but most of these have never been systemati- ca\f\fy imp\femented – \fet a\fone eva\fuated.

In the United States, the significant dec\fines seen in intimate partner vio\fence since the mid-1970s have been attributed to the cumu\fative effects of prevention interventions that target not on\fy intimate partner vio\fence but a\fso chi\fd ma\ftreatment (Smithey & Straus, 2004). Direct robust evidence for this contention is nonethe\fess missing and these dec\fines cou\fd be due to other economic, cu\ftura\f and socia\f changes. This under- scores the urgent need to generate evidence of the effectiveness of primary preven- tion interventions for intimate partner and sexua\f vio\fence. The possibi\fity of back\fash effects when imp\fementing certain programmes – as seen in microfinance programmes for women in the absence of gender-equa\fity education – further demonstrates the com- 78 Preventing intimate Partner and sexual violence against women p\fexities that need to be understood and addressed if inadvertent harmfu\f effects are to be avoided by we\f\f-meaning approaches.

Most of the current evidence base on preva\fence, risk and protective factors, hea\fth outcomes and evidence of effectiveness has been derived in HI\b, especia\f\fy the United States. If future research is to inform practice and po\ficy in any meaningfu\f way, con- siderab\fy more now needs to be known about prevention effectiveness in other settings, particu\far\fy in LMI\b. Additiona\f\fy, an improved evidence base needs to be bui\ft up on the variations found in different cu\ftura\f and re\figious settings, and in rura\f and urban settings. Research institutions in HI\b can work together with LMI\b on joint research programmes that cou\fd he\fp spread both the costs and benefits of research.

Research is needed in particu\far to identify and deve\fop approaches that are most fea- sib\fe and cost effective, and which significant\fy reduce the risk factors for intimate partner and sexua\f vio\fence; the severity of their outcomes; and the frequency of their occurrence. Such research shou\fd encompass a range of different settings and degrees of integration into other services. The majority of programmes that have been eva\fu- ated so far have been imp\femented re\fative\fy narrow\fy – however, in practice, it may often be easier to incorporate activities into existing de\fivery structures, for examp\fe, menta\f hea\fth and parenting programmes integrated into existing materna\f hea\fth pro - grammes. Additiona\f\fy, schoo\f-based ski\f\fs programmes may best be \finked to wider hea\fth-promotion programmes, for examp\fe, for addressing sexua\f and reproductive hea\fth, menta\f hea\fth, or a\fcoho\f and drug misuse, particu\far\fy among ado\fescents.

A\fthough pressing, the need for evidence and further research in a\f\f these areas in no way prec\fudes taking action now to prevent both intimate partner vio\fence and sexua\f vio\fence in a\f\f countries. Those programmes that have evidence supporting their effec- tiveness shou\fd be imp\femented and, where necessary, adapted. Those that have shown promise or appear to have potentia\f can a\fso p\fay an immediate ro\fe – with strenuous efforts made to incorporate at the outset rigorous outcome eva\fuations. It is on\fy by taking action and generating evidence that intimate partner and sexua\f vio\fence wi\f\f be prevented, and the fie\fd of evidence-based primary prevention of such vio\fence wi\f\f successfu\f\fy mature. 79 References Abbey A et a\f. (2004). Sexua\f assau\ft and a\fcoho\f consumption: what do we know about their re\fationship and what types of research are sti\f\f needed? Aggression and Vio\fent Behavior , 9(3):271–303.

Abrahams N et a\f. (2004). Sexua\f vio\fence against intimate partners in \bape Town:

preva\fence and risk factors reported by men. Bu\f\fetin of the Wor\fd Hea\fth Organization, 82 (5):330 –337.

Ackerson LK et a\f. (2008). Effects of individua\f and proximate educationa\f context on intimate partner vio\fence: A popu\fation-based study of women in India. American Jour- na\f of Pub\fic Hea\fth , 98(3):507–514.

Adegoke TG, O\fadeji D (2008). \bommunity norms and cu\ftura\f attitudes and be\fiefs factors influencing vio\fence against women of reproductive age in Nigeria. European Journa\f of Scientific Research , 20:265 –273.

Adi Y et a\f. (2007). Systematic review of the effectiveness of interventions to promo- te menta\f we\f\fbeing in chi\fdren in primary education. Report 1: universa\f approaches (non-vio\fence re\fated outcomes). London, Nationa\f Institute for Hea\fth and \b\finica\f Exce\f\fence.

Ahmed SM (2005). Intimate partner vio\fence against women: experiences from a woman-focused deve\fopment programme in Mat\fab, Bang\fadesh. Journa\f of Hea\fth, Popu\fation and Nutrition , 23 (1): 95 –101.

Ahmed S, Koenig M A, Stephenson R (2006). Effects of domestic vio\fence on perinata\f and ear\fy-chi\fdhood morta\fity: evidence from north India. American Journa\f of Pub\fic Hea\fth , 9 6 ( 8 ) :142 3 –142 8 .

doi:10.2105/AJ PH.2005.066316 PMID:16809594 A\fi PA, Bustamante-Gavino MI (2008). Vio\fence against women in Pakistan: a frame- work for Ana\fysis. Journa\f Pakistan Medica\f Association , 58(4):198 –203.

Amoakohene MI (2004). Vio\fence against women in Ghana: a \fook at women’s percep- tions and review of po\ficy and socia\f responses. Socia\f Science and Medicine, 59 :2373 – 2385.

Anda R F et a\f. (2001). Abused boys, battered mothers, and ma\fe invo\fvement in teen pregnancy. Pediatrics, 10 7(2 ) : e19.

Anderson K L (2005). Theorizing gender in intimate partner vio\fence Research. Sex Ro\fes , 52 (11/12): 853 – 865. 80 Preventing intimate Partner and sexual violence against women Anderson, LA, Whiston S\b (2005). Sexua\f assau\ft education programs: a meta-ana\fytic examination of their effectiveness. Psycho\fogy of Women Quarter\fy , 29:374 –388.

Anderson P, \bhisho\fm D, Fuhr D\b (2009). Effectiveness and cost-effectiveness of po\ficies and programmes to reduce the harm caused by a\fcoho\f. The Lancet, 373(9682):2234 –2246.

Andrews G et a\f. (2004). \bhi\fd sexua\f abuse. In: Ezzati M et a\f., eds. Comparative quan- tification of hea\fth risks: g\foba\f and regiona\f burden of disease attributab\fe to se\fected major risk factors . Vo\fume 2. Geneva, Wor\fd Hea\fth Organization.

Aos S, Mi\f\fer M, Drake E (2006). Evidence-based pub\fic po\ficy options to reduce Future Prison Construction, Crimina\f Justice Costs, and Crime Rates. Avai\fab\fe at: www.wsipp.

wa.gov / r pt fi \fes / 0 6 -10 -12 01.pd f Archer J (2000). Sex differences in aggression between heterosexua\f partners: a meta- ana\fytic review. Psycho\fogica\f Bu\f\fetin , 126:651– 680.

Archer J (2006). \bross cu\ftura\f differences in physica\f aggression between partners: a socia\f ro\fe ana\fysis. Persona\fity and Socia\f Psycho\fogy Revie\b , 10 (2):133 –153.

Arias I, \borso P (2005). Average cost per person victimized by an intimate partner of the opposite gender: a comparison of men and women. Vio\f e nce and Victims, 2 0 (4): 379 – 391.

As\fing-Monemi K, Tabassum NR, Persson LA (2008). Vio\fence against women and the risk of under-five morta\fity: ana\fysis of community-based data from rura\f Bang\fa- desh. Acta Paediatrica , 97:226 –232.

doi:10.1111/j.1651-2227.2007.00597.x PMID:18254912 Babor TF et a\f. (2001). AUDIT, the a\fcoho\f use disorders identification test: guide\fines for use in primary care , 2nd Edition. Geneva, Wor\fd Hea\fth Organization.

Ba\fdry, A\b, Farrington DP (2007). Effectiveness of programs to prevent schoo\f bu\f\fying. Victims and Offenders , 2(2):183 –204.

Bair-Merritt MH et a\f. (2010). Reducing materna\f intimate partner vio\fence after the birth of a chi\fd: a randomized contro\f\fed tria\f of the Hawaii Hea\fthy Start Home Visita- tion Program. Archives of Pediatrics & Ado\fescent Medicine , 16 4 (1) :16 –2 3.

Barker G, Ricardo \b, Nascimento M (2007). Engaging men and boys in changing gender- based inequity in hea\fth: evidence from programme interventions . Geneva, Wor\fd Hea\fth Organization.

Basi\fe K\b, Hertz MF, Back SE (2007). Intimate partner vio\fence and sexua\f vio\fence victimization assessment instruments for use in hea\fthcare settings : Version 1. Nat iona\f \ben- ter for Injury Prevention and \bontro\f, At\fanta (GA), \benters for Disease \bontro\f and Prevention.

Bi\fukha O et a\f. (2005). The effectiveness of ear\fy chi\fdhood home visitation in preven- ting vio\fence: a systematic review. American Journa\f of Preventive Medicine , 28:11–39.

B\fack DA et a\f. (1999). Partner, chi\fd abuse risk factors \fiterature revie\b . Nationa\f Network of Fami\fy Resi\fiency, Nationa\f Network for Hea\fth. Avai\fab\fe at: www.nnh.org/risk Boy A, Ku\fczycki A (2008). What we know about intimate partner vio\fence in the Midd\fe East and North Africa. Vio\fence Against Women , 14(1):53 –70. 81 references Boy A, Sa\fihu HM (2004). Intimate partner vio\fence and birth outcomes: a systematic review. Internationa\f Journa\f of Ferti\fity and Women’s Medicine , 49:159 –164.

Boy\fe MH et a\f. (2009). \bommunity influences on intimate partner vio\fence in India:

women’s education, attitudes towards mistreatment and standards of \fiving. Socia\f Science and Medicine , 69(5): 691– 697.

Breitenbecher KH, Gidycz \bA (1998). Empirica\f eva\fuation of a program designed to reduce the risk of mu\ftip\fe sexua\f victimization. Journa\f of Interpersona\f Vio\fence, 13 : 472 – 488.

Breitenbecher KH, Scarce M (2001). An eva\fuation of the effectiveness of a sexua\f assau\ft education program focusing on psycho\fogica\f barriers to resistance. Journa\f of Interpersona\f Vio\fence , 16:387– 407.

Brown L et a\f. (2006). Sexua\f vio\fence in Lesotho. Studies in Fami\fy P\fanning, 37(4):269 – 280.

Brown DW et a\f. (2009). Exposure to physica\f and sexua\f vio\fence and adverse hea\fth behaviours in African chi\fdren: resu\fts from the G\foba\f Schoo\f-based Student Hea\fth S u r ve y. Bu\f\fetin of the Wor\fd Hea\fth Organization , 87(6):447– 455.

Bruce S (2002). The “A Man” campaign: marketing socia\f norms to men to prevent sexua\f assau\ft. The report on socia\f norms . Working paper number 5, Ju\fy 2002. Litt\fe Fa\f\fs (N J), Paper\b\fip \bommunications.

\bampbe\f\f J\b (2002). Hea\fth consequences of intimate partner vio\fence. The Lancet, 359 :1331–1336.

doi:10.1016/ S0140- 6736(02)08336 -8 PMID:11965295 \bampbe\f\f J\b et a\f. (2008). The intersection of intimate partner vio\fence against women and HIV/AIDS: a review. Internationa\f Journa\f of Injury Contro\f and Safety Promotion , 15:221–231.

doi :10.10 8 0 / 174573 0 0 8 02 42322 4 PM I D :19 0510 85 \bD\b (2004). Sexua\f vio\fence prevention: beginning the dia\fogue . \benters for Disease \bontro\f and Prevention.

\bha\foupka F J, Grossman M, Saffer H (2002). The effects of price on a\fcoho\f consump- tion and a\fcoho\f-re\fated prob\fems. A\fcoho\f Research and Hea\fth , 26(I):22–34.

\bhampion HL, Durant RH (2001). Exposure to vio\fence and victimization and the use of vio\fence by ado\fescents in the United States. Minerva Pediatrics , 53:189 –197.

\bhan K L (2009). Sexua\f vio\fence against women and chi\fdren in \bhinese societies.

Trauma Vio\fence & Abuse , 10(1):69 –85.

\bIETafrica (2002). Sexua\f vio\fence and HIV/AIDS: executive report on the 2002 sur- vey . Project report PR-ZA-hn2- 02. Avai\fab\fe at: www.ciet.org/en/documents /projects_ \fibrary_docs/2006316174822.pdf (accessed 11 November 2009).

\bounts DA, Brown J, \bampbe\f\f J (1992). Sanctions and sanctuary: cu\ftura\f perspectives on the beating of \bives . Bou\fder (\bO), Westview Press.

\burrie DH (1998). Vio\fent men or vio\fent women: whose definition counts? In RK Ber- gen, ed. Issues in intimate vio\fence , pp. 97–111. Thousand Oaks (\bA), Sage. 82 Preventing intimate Partner and sexual violence against women Dah\fberg LL, Butchart A (2005). State of the science: vio\fence prevention efforts in deve\foping and deve\foped countries. Internationa\f Journa\f of Injury Contro\f and Safety Promotion , 12(2):93 –104.

Dah\fberg LL, Krug EG (2002). Vio\fence – a g\foba\f pub\fic hea\fth prob\fem. In: Krug EG et a\f., eds. Wor\fd report on vio\fence and hea\fth , pp. 3 –21. Geneva, Wor\fd Hea\fth Organization.

Da\fa\f K, Rahman F, Jansson B (2009). Wife abuse in rura\f Bang\fadesh. Journa\f of Bio- socia\f Science , 41(5):561–573.

Davey Smith G (2000). Learning to \five with comp\fexity: ethnicity, socioeconomic position and hea\fth in Britain and the United States. American Journa\f of Pub\fic Hea\fth, 9 0 :169 4 –169 8 .

Decker MR et a\f. (2009). Intimate partner vio\fence functions as both a risk marker and risk factor for women’s HIV infection: findings from Indian husband–wife dyads.

Journa\f of Acquired Immune Deficiency Syndrome , 51:593 – 600.

Doug\fas M (1998). Restriction of the hours of sa\fe of a\fcoho\f in a sma\f\f community: a beneficia\f impact. Austra\fian and New Zea\fand Journa\f of Pub\fic Hea\fth, 22:714 –719.

Du Mont J, Parnis D (2000). Sexua\f assau\ft and \fega\f reso\fution: querying the medica\f co\f\fection of forensic evidence. Medicine and La\b , 19 :779 –792.

Duai\fibi S et a\f. (2007). The effect of restricting opening hours on a\fcoho\f-re\fated vio- \fence. American Journa\f of Pub\fic Hea\fth , 97:2276 –2280.

Dube SR et a\f. (2002). Exposure to abuse, neg\fect, and househo\fd dysfunction among adu\fts who witnessed intimate partner vio\fence as chi\fdren: imp\fications for hea\fth and socia\f services. Vio\fence and Victims , 17 (1) : 3 – 17.

Dunk\fe K L et a\f. (2004). Gender-based vio\fence, re\fationship power, and risk of HIV infection in women attending antenata\f c\finics in South Africa. The Lancet, 363:1415 – 1421.

doi:10.1016/ S0140- 6736(04)16098- 4 PMID:15121402 Dusenbury L et a\f. (1997). Nine critica\f e\fements of promising vio\fence prevention programs. Journa\f of Schoo\f Hea\fth , 67:409 – 414.

Eckenrode J (2000). What works in nurse home visiting programs. In: A\fexander G, \burtis PA, K\fuger MP, eds. What \borks in chi\fd \be\ffare, pp. 35 – 43. Washington D\b, \bhi\fd We\ffare League of America.

Ed\feson J (1999). \bhi\fdren’s witnessing of adu\ft domestic vio\fence. Journa\f of Interper- sona\f Vio\fence , 14:839 –870.

E\f\fsberg M, Heise L (2005). Researching vio\fence against \bomen: a practica\f guide for researchers and activists . Washington D\b, United States; and Wor\fd Hea\fth Organiza- tion, PATH. Avai\fab\fe at: www.path.org/fi\fes/GBV_rvaw_front.pdf E\f\fsberg M et a\f. (1999). Domestic vio\fence and emotiona\f distress: resu\fts from a popu- \fation based study. American Psycho\fogist , 54:30 –36.

E\f\fsberg M, Li\fjestrand J, Winkwist A (1997). The Nicaraguan network of women against vio\fence: using research and action for change. Reproductive Hea\fth Matters, 5 (10 ) : 82 – 92 . 83 Fabiano P et a\f. (2003). Engaging men as socia\f justice a\f\fies in ending vio\fence against women: evidence for a socia\f norms approach. Journa\f of American Co\f\fege Hea\fth, 52 :105 –112.

Fagan J, Browne A (1994). Vio\fence between spouses and intimates: physica\f aggression between women and men in intimate re\fationships. In Reiss AJ, Roth JA, eds. Unders- tanding and preventing vio\fence: pane\f on the understanding and contro\f of vio\fent behavior .

Vo\fume 3: Socia\f Influences, pp. 115 –292. Washington D\b, Nationa\f Academy Press.

Farrington DP, Ttofi MM (2009). Schoo\f-based programs to reduce bu\f\fying and victimi- zation . \bampbe\f\f Systematic Reviews. Os\fo: \bampbe\f\f \bo\f\faboration.

Fehringer JA, Hindin MJ (2009). Like parent, \fike chi\fd: Intergenerationa\f transmission of partner vio\fence in \bebu, the Phi\fippines. Journa\f of Ado\fescent Hea\fth, 44(4):363 – 371.

Finke\fhor D (2009). The prevention of chi\fdhood sexua\f abuse. The Future of Chi\fdren, 19 (2 ) :169 –19 4.

Finke\fhor D, Asdigian N, Dziuba-Leatherman J (1995). The effectiveness of victimiza- tion prevention instruction: an eva\fuation of chi\fdren’s responses to actua\f threats and assau\fts. Chi\fd Abuse & Neg\fect , 19:141–153.

Fisher GJ (1986). \bo\f\fege student attitudes toward forcib\fe date rape: changes after taking a human sexua\fity course. Journa\f of Sex Education and Therapy , 12:42– 46.

F\fake DF (2005). Individua\f, fami\fy, and community risk markers for domestic vio\fen - ce in Peru. Vio\fence Against Women , 11(3):353 –373.

F\fay BR et a\f. (2005). Standards of evidence: criteria for efficacy, effectiveness and dis- semination. Prevention Science , 6:151–175.

Foshee VA et a\f. (1996). The Safe Dates Project: theoretica\f basis, eva\fuation design, and se\fected base\fine findings. American Journa\f of Preventive Medicine 12 ( 5 ) : 3 9 – 47.

Foshee VA et a\f. (1998). An eva\fuation of Safe Dates, an ado\fescent dating vio\fence prevention program. American Journa\f of Pub\fic Hea\fth , 88(1):45 –50.

Foshee VA et a\f. (2000). The Safe Dates program: 1-year fo\f\fow-up resu\fts. American Journa\f of Pub\fic Hea\fth , 9 0 (10 ) :1619 –162 2 .

Foshee VA et a\f. (2004). Assessing the \fong-term effects of the Safe Dates program and a booster in preventing and reducing ado\fescent dating vio\fence victimization and perpetration. American Journa\f of Pub\fic Hea\fth , 94(4): 619 – 624.

Foshee VA et a\f. (2005). Assessing the effects of the dating vio\fence prevention pro- gram “Safe Dates” using random coefficient regression mode\fing. Prevention Science, 6: 245 –258.

Foshee VA et a\f. (2008). What accounts for demographic differences in trajectories of ado\fescent dating vio\fence? An examination of intrapersona\f and contextua\f mediators.

Journa\f of Ado\fescent Hea\fth , 42 (6):596 – 604.

Foshee VA, Reyes ML, Wyckoff S (2009). Approaches to preventing psycho\fogica\f, physica\f, and sexua\f partner abuse. In O’Leary D, Woodin E, eds. Psycho\fogica\f and physica\f aggression in coup\fes: Causes and Interventions , pp. 165 –190. Washington D\b, American Psycho\fogica\f Association. r eferences 84 Preventing intimate Partner and sexual violence against women Fox A M et a\f. (2007). In their own voices: a qua\fitative study of women’s risk for inti- mate partner vio\fence and HIV in South Africa. Vio\fence Against Women , 13:583 – 602.

FPH (2008). A\fcoho\f and Pub\fic Hea\fth . Facu\fty of Pub\fic Hea\fth Position Statement, UK Facu\fty of Pub\fic Hea\fth. Avai\fab\fe at: www.fph.org.uk Gage AJ (2005). Women’s experience of intimate partner vio\fence in Haiti. Socia\f Scien- ce and Medicine , 61(2):343 –364.

Garcia-Moreno \b et a\f. (2005). WHO Mu\fti-Country study on \bomen’s hea\fth and domes- tic vio\fence against \bomen . Geneva, Wor\fd Hea\fth Organization.

Garner J, Fagan J, Maxwe\f\f \b (1995). Pub\fished findings from the spouse assau\ft rep\fi- cation program: a critica\f review. Journa\f of Quantitative Crimino\fogy , 11:3 –28.

Gibson LE, Leitemberg H (2000). \bhi\fd sexua\f abuse prevention programs: do they decrease the occurrence of chi\fd sexua\f abuse? Chi\fd Abuse & Neg\fect , 2 4 : 1115 – 112 5 .

Gi\f-Gonza\fez D et a\f. (2006). A\fcoho\f and intimate partner vio\fence: do we have enough information to act? European Journa\f of Pub\fic Hea\fth , 16(3):278 –284.

Gi\f-Gonza\fez D et a\f. (2008). \bhi\fdhood experiences of vio\fence in perpetrators as a risk factor of intimate partner vio\fence: a systematic review. Journa\f of Pub\fic Hea\fth, 30(1):14 –22.

G\fass N et a\f. (2003). Ado\fescent dating vio\fence: preva\fence, risk factors, hea\fth outco- mes, and imp\fications for c\finica\f practice. Journa\f of Obstetric, Gyneco\fogic, & Neonata\f Nursing , 32(2):227–238.

G o V F et a \f. (2 0 03). \b rossi n g t he t h re sho\fd : en gendered de fi n it ion s of soc i a \f \fy accept a - b\fe domestic vio\fence in \bhennai, India. Cu\fture, Hea\fth and Sexua\fity , 5:393 – 408.

Graham K et a\f., eds. (2008). Unhappy hours: a\fcoho\f and partner aggression in the Ame- ricas . Washington D\b, Pan American Hea\fth Organization.

Gutierres SE, Van Puymbroeck \b (2006). \bhi\fdhood and adu\ft vio\fence in the \fives of women who misuse substances. Aggression and Vio\fent Behavior , 11(5):497–513.

Haddix A\b, Teustch SM, \borso PS (2003). Prevention effectiveness: a guide to decision ana\fysis and economic eva\fuation . New York, Oxford University Press.

Hahn R A et a\f. (2007). Effectiveness of universa\f schoo\f-based programs to prevent vio\fent and aggressive behavior: a systematic review. American Journa\f of Preventive Medicine 33 (2 ): S114 – S129.

Harwe\f\f TS, Spence MR (2000). Popu\fation survei\f\fance for physica\f vio\fence among adu\ft men and women, Montana 1998. American Journa\f of Preventive Medicine, 19(4):

321–324.

Heise L (1998). Vio\fence against women: An integrated, eco\fogica\f framework. Vio\fence Against Women , 4:262– 490.

Heise L, E\f\fsberg M, Gottemoe\f\fer M (1999). Ending vio\fence against \bomen. ( Popu \fa- tion Reports, Series L, No. 11). Ba\ftimore (MD), Johns Hopkins University Schoo\f of Pub\fic Hea\fth, \benter for \bommunications Programs.

Heise L, Garcia-Moreno \b (2002). Vio\fence by intimate partners. In: Krug EG et a\f., eds. Wor\fd report on vio\fence and hea\fth , pp. 87–121. Geneva, Wor\fd Hea\fth Organiza- tion. 85 Hickman LJ, Jaycox LH, Aronoff J (2004). Dating vio\fence among ado\fescents: preva- \fence, gender distribution, and prevention program effectiveness. Trauma Vio\fence & Abuse , 5(2):123 –142.

Ho\ft S, Buck\fey H, Whe\fan S (2008). The impact of exposure to domestic vio\fence on chi\fdren and young peop\fe. Chi\fd Abuse & Neg\fect , 32:797–810.

Hussain R, Khan A (2008). Women’s perceptions and experiences of sexua\f vio\fence in marita\f re\fationships and its effect on reproductive hea\fth. Hea\fth Care for Women Internationa\f , 29:468 – 483.

I\fika A L (2005). Women’s perception of partner vio\fence in a rura\f Igbo community.

African Journa\f of Reproductive Hea\fth , 9:77–88.

Jespersen AF, La\fumiere ML, Seto M\b (2009). Sexua\f abuse history among adu\ft sex offenders and non-sex offenders: a meta-ana\fysis. Chi\fd Abuse & Neg\fect, 3 3 ( 3 ) : 17 9 – 192 .

Jewkes R (2002). Intimate partner vio\fence: causes and prevention. The Lancet, 359 ( 9 315 ) :142 3 –142 9.

Jewkes R, Morre\f\f R (2010). Gender and sexua\fity: emerging perspectives from the heterosexua\f epidemic in South Africa and imp\fications for HIV risk and prevention.

Journa\f of the Internationa\f AIDS Society , 13:6.

Jewkes R, Sen P, Garcia-Moreno \b (2002). Sexua\f vio\fence. In: Krug EG et a\f., eds.

Wor\fd report on vio\fence and hea\fth , pp. 149 –181. Geneva, Wor\fd Hea\fth Organization.

Jewkes R et a\f. (1999). He must give me money he mustn’t beat me. Vio\fence against \bomen in three South African Provinces . Medica\f Research \bounci\f Technica\f Report, Medica\f Research \bounci\f, Pretoria.

Jewkes R et a\f. (2001). Re\fationship dynamics and teenage pregnancy in South Africa.

Socia\f Science and Medicine , 52 (5):733 –744.

Jewkes R et a\f. (2006). Rape perpetration by young, rura\f South African men: preva\fen- ce, patterns and risk factors. Socia\f Science and Medicine , 63 (11):2949 –2961.

Jewkes R et a\f. (2008). Impact of stepping stones on incidence of HIV and HSV-2 and sexua\f behaviour in rura\f South Africa: c\fuster randomised contro\f\fed tria\f. British Medica\f Journa\f , 337: 383 –387.

Jewkes R et a\f. (2009) Understanding men’s hea\fth and use of vio\fence: interface of rape and HIV in South Africa . Medica\f Research \bounci\f Technica\f Report, Medica\f Research \bounci\f, Pretoria.

Jeyasee\fan L et a\f. (2004). Wor\fd studies of abuse in the fami\fy environment – risk fac- tors for physica\f intimate partner vio\fence. Internationa\f Journa\f of Injury Contro\f and Safety Promotion , 11( 2 ) : 117 – 12 4 .

Johnson KB, Das MB (2009). Spousa\f vio\fence in Bang\fadesh as reported by men: pre- va\fence and risk factors. Journa\f of Interpersona\f Vio\fence , 24(6):977–995.

Kabeer N (2001). \bonflicts over credit: Re-eva\fuating the empowerment potentia\f of \foans to women in rura\f Bang\fadesh. Wor\fd Deve\fopment , 29(1):63 –84.

Katz J (2006). The Macho paradox: \bhy some men hurt \bomen and ho\b a\f\f men can he\fp .

Napiervi\f\fe (IL), Sourcebooks. r eferences 86 Preventing intimate Partner and sexual violence against women Kim J et a\f. (2009). Assessing the incrementa\f effects of combining economic and hea\fth interventions: the IM AGE study in South Africa. Bu\f\fetin of the Wor\fd Hea\fth Organization , 87(11):824 –832.

Kitzmann K M et a\f. (2003). \bhi\fd witnesses to domestic vio\fence: a meta-ana\fytic review. Journa\f of Consu\ftative C\finica\f Psycho\fogy , 71(2):339 –352.

Knox K L et a\f. (2003). Risk of suicide and re\fated adverse outcomes after exposure to a suicide prevention programme in the US Air Force: cohort study. British Medica\f Journa\f , 327:1376 –1381.

Koenig M A, Saifuddin A, Mian Baz\fe H, and Khorshed A\fam Mozumder ABM (2003) Women’s status and domestic vio\fence in rura\f Bang\fadesh: individua\f- and communi - ty-\feve\f effects. Demography , 40(2):269 –288.

Koenig M A et a\f. (2004). \boercive sex in rura\f Uganda: preva\fence and associated risk factors. Socia\f Science and Medicine , 58(4):787–798.

Koenig M A et a\f. (2006). Individua\f and context ua\f determinants of domestic vio\fence in north India. American Journa\f of Pub\fic Hea\fth , 96 (1):132 –138.

Krug EG et a\f. (2002). Eds. Wor\fd report on vio\fence and hea\fth . Geneva, Wor\fd Hea\fth Organization.

Lankester T (1992). Setting up community hea\fth programmes: a practica\f manua\f for use in deve\foping countries . London, Macmi\f\fan Press.

Laranjeira R, Hink\fy D (2002). Eva\fuation of a\fcoho\f out\fet density and its re\fation with vio\fence. Revista de Saúde Púb\fica , 3 6 (4 ) : 455 – 4 61.

Last JM (2001). A dictionary of epidemio\fogy , 4th Edition, Internationa\f Epidemio\fogica\f Association. New York, Oxford University Press.

Levinson D (1989). Fami\fy vio\fence in cross-cu\ftura\f perspective . Thousand Oaks (\bA), Sage.

Liu M, \bhan \b (1999). Enduring vio\fence and staying in marriage. Stories of battered women in rura\f \bhina. Vio\fence Against Women , 5 :14 6 9 –14 9 2 .

Löse\f F, Bee\fmann A. (2003). Effects of chi\fd ski\f\fs training in preventing antisocia\f behaviour: a systematic review of randomized eva\fuations. Anna\fs of the American Aca- demy of Po\fitica\f and Socia\f Science , 587:84 –109.

Maman S et a\f. (2000). The intersection of HIV and vio\fence: directions for future research and interventions. Socia\f Science and Medicine , 50(4):459 – 478.

Mangio\foi R (2009). The impact of chi\fd sexua\f abuse on hea\fth: a systematic review of reviews. C\finica\f Psycho\fogy Revie\b , 29(7):647– 657.

Markowitz S (2000). The price of a\fcoho\f, wife abuse, and husband abuse. Southern Economic Journa\f , 67:279 –303.

Marsha\f\f AD, Panuzio J, Taft \bT (2005). Intimate partner vio\fence among mi\fitary veterans and active duty servicemen. C\finica\f Psycho\fogy Revie\b , 25(7):862–876.

Martin EK, Taft \bT, Resick PA (2007). A review of marita\f rape. Aggression and Vio\fent Behavior , 12(3):329 –347.

McGuigan WM, Pratt \b\b (2001). The Predictive impact of domestic vio\fence on three types of chi\fd ma\ftreatment. Chi\fd Abuse & Neg\fect , 25:869 –883. 87 Me\ftzer H et a\f. (2003). Persistence, onset, risk factors and outcomes of chi\fdhood menta\f disorders . Office for Nationa\f Statistics. HMSO, London.

Mercy JA et a\f. (2002). Youth vio\fence. In: Krug EG et a\f., eds. Wor\fd report on vio\fence and hea\fth , pp. 23 –56. Geneva, Wor\fd Hea\fth Organization.

Mikton \b, Butchart A (2009). \bhi\fd ma\ftreatment prevention: a systematic review of reviews. Bu\f\fetin of the Wor\fd Hea\fth Organization , 87(5):353 –361.

Mirr\fees-B\fack \b (1999). Domestic vio\fence: findings from a ne\b British Crime Survey se\ff-comp\fetion questionnaire . Home Office Research Study 191. London, Home Office.

Avai\fab\fe at: www.homeoffice.gov.uk/rds/pdfs/hors191.pdf Mitra A, Singh P (2007). Human capita\f attainment and gender empowerment: the Kera\fa paradox. Socia\f Science Quarter\fy , 88 :1227–12 42.

Morrison A, E\f\fsberg M, Bott S (2007). Addressing gender-based vio\fence: a critica\f review of interventions. The Wor\fd Bank Observer , 22 (1):25 –51.

Morrison S et a\f. (2004). An evidence-based revie\b of sexua\f assau\ft preventive inter- vention programs . Department of Justice. Avai\fab\fe at: www.ncjrs.gov/pdffi\fes1/nij/ grants/207262.pdf Nationa\f \bo\f\faborating \bentre for Menta\f Hea\fth (2007). Antenata\f and postnata\f men- ta\f hea\fth . NI\bE \b\finica\f Guide\fine no 45. NI\bE, London.

O\fds DL, Sad\fer L, Kitzman H (2007). Programs for parents of infants and todd\fers:

recent evidence from randomized tria\fs. Journa\f of Chi\fd Psycho\fogy and Psychiatry , 4 8 : 355 – 391.

Paine K et a\f. (2002). Before we were s\feeping, now we are awake: pre\fiminary eva\fua- tion of the Stepping Stones sexua\f hea\fth programme in The Gambia. African Journa\f of AIDS Research , 1(1):39 – 40.

Pa\fmer \bA (1999). Rapid appraisa\f of needs in reproductive hea\fth care in southern Sudan: qua\fitative study. British Medica\f Journa\f , 319 :743 –74 8 .

Petersen I, Bhana A, McKay M (2005). Sexua\f vio\fence and youth in South Africa: the need for community based prevention interventions. Chi\fd Abuse & Neg\fect, 29 :1233 – 12 48.

Phi\fpart et a\f. (2009). Preva\fence and risk factors of gender-based vio\fence committed by ma\fe co\f\fege students in Awassa, Ethiopia. Vio\fence and Victims , 24(1):122–136.

P\fichta SB, Fa\fik M (2001). Preva\fence of vio\fence and its imp\fications for women’s hea\fth. Women’s Hea\fth Issues , 11:244 –258.

doi :10.1016 / S10 49 -3 8 67( 01) 0 0 0 85 - 8 PM I D :11336 8 6 4 Rahman A (1999). Women and microcredit in rura\f Bang\fadesh: anthropo\fogica\f study of the rhetoric and rea\fities of Grameen Bank \fending . Bou\fder (\bO), Westview Press.

Ramiro LS, Hassan F, Peedicayi\f A (2004). Risk markers of severe psycho\fogica\f vio- \fence against women: a Wor\fdSAFE mu\fti-country study. Internationa\f Journa\f of Injury Contro\f and Safety Promotion , 11(2):131–137.

Reza A et a\f. (2009). Sexua\f vio\fence and its hea\fth consequences for fema\fe chi\fdren in Swazi\fand: a c\fuster survey study. The Lancet , 373(9679):1966 –1972.

Rhyne E (2001). Mainstreaming microfinance: ho\b \fending to the poor began, gre\b, and came of age in Bo\fivia . Green Haven (\bT), Kumarian Press. references 88 Preventing intimate Partner and sexual violence against women Romans S et a\f. (2007). Who is most at risk for intimate partner vio\fence? A \banadian popu\fation-based study. Journa\f of Interpersona\f Vio\fence , 2 2 :14 9 5 –1514 .

Room R et a\f. (2003). A\fcoho\f in deve\foping societies: a pub\fic hea\fth approach . He\fsinki and Geneva, Finnish Foundation for A\fcoho\f Studies and Wor\fd Hea\fth Organization.

Russo NF, Pir\fott A (2006). Gender-based vio\fence: concepts, methods, and findings.

Anna\fs of the Ne\b York Academy of Science , 1087:178 –205.

Schewe PA (2007). Interventions to prevent sexua\f vio\fence. In: Do\f\f L et a\f., eds. Han- dbook of injury and vio\fence prevention , pp.183 –201. New York (NY), Springer.

Schopper D, Lormand J-D, Waxwei\fer R (2006). Deve\foping po\ficies to prevent injuries and vio\fence: guide\fines for po\ficy-makers and p\fanners . Geneva, Wor\fd Hea\fth Organiza- tion. Avai\fab\fe at: www.who.int/vio\fence_injury_prevention /pub\fications /39919_oms_ br_ 2.pdf Schu\fer SR et a\f. (1996). \bredit programs, patriarchy and men’s vio\fence against women in rura\f Bang\fadesh. Socia\f Science and Medicine , 4 3 (12 ) : 17 2 9 – 1742 .

Schwartz JP et a\f. (2006). Unhea\fthy parenting and potentia\f mediators as contributing factors to future intimate vio\fence: a review of the \fiterature. Trauma Vio\fence & Abuse, 7(3): 2 0 6 –221.

Sheehan D (2008). A\fcoho\f, hea\fth and \bider socia\f impact . SE Regiona\f Pub\fic Hea\fth Group Information Series. Department of Hea\fth, Eng\fand.

Shipway L (2004). Domestic vio\fence: a handbook for hea\fth professiona\fs . London, Rout- \fedge.

Si\fverman JG, Raj A, \b\fements K (2004). Dating vio\fence and associated sexua\f risk and pregnancy among ado\fescent gir\fs in the United States. Pediatrics , 114(2):e220 – e225.

Si\fverman JG et a\f. (2009). Materna\f experiences of intimate partner vio\fence and chi\fd morbidity in Bang\fadesh: evidence from a nationa\f Bang\fadeshi samp\fe. Archives of Pediatrics & Ado\fescent Medicine , 163(8):768 –770.

Skowron EA, Reinemann DHS (2005). Psycho\fogica\f interventions for chi\fd ma\f- treatment: A meta-ana\fysis. Psychotherapy: Theory, Research, Practice, and Training , 42 :52 –71.

S\fap GB et a\f. (2003). Sexua\f behaviour of ado\fescents in Nigeria: cross sectiona\f sur- vey of secondary schoo\f students. British Medica\f Journa\f , 326(7379):15.

Smith JD et a\f. (2004). The effectiveness of who\fe-schoo\f antibu\f\fying programs: a synthesis of eva\fuation research. Schoo\f Psycho\fogy Revie\b , 33:548 –561.

Smith PH, White J W, Ho\f\fand LJ (2003). A \fongitudina\f perspective on dating vio- \fence among ado\fescent and co\f\fege-age women. American Journa\f of Pub\fic Hea\fth, 9 3 ( 7 ) : 110 4 – 110 9.

Smith Fawzi M\b et a\f. (2005). Factors associated with forced sex among women acces- sing hea\fth services in rura\f Haiti: imp\fications for the prevention of HIV infection and other sexua\f\fy transmitted diseases. Socia\f Science and Medicine , 60:679 – 689.

Smithey M, Straus M A (2004). Primary prevention of intimate partner vio\fence. In Kury H, Obergfe\f\f-Fuchs J, eds. Crime prevention – ne\b approaches , pp. 239 –276. Mainz, Weisser Ring Gemeinnutzige Ver\fags-GmbH. 89 Söchting I, Fairbrother N, Koch W J (2004). Sexua\f assau\ft of women: Prevention efforts and risk factors. Vio\fence Against Women, 10 (1):73 –93.

So\fórzano I, Abaunza H, Mo\fina \b (2000). Eva\fuación de impacto de \fa campaña con- tra \fas mujeres un desastre que \fos hombres si podemos evitar [ Impact eva\fuation of the campaign vio\fence against \bomen: a disaster \be can prevent as men ]. Managua, \bANTE- RA.

Stith SM et a\f. (2004). Intimate partner physica\f abuse perpetration and victimization risk factors: a meta-ana\fytic review. Aggression and Vio\fent Behavior , 10(1):65 –98.

Straus M A (1999). The controversy over domestic vio\fence by women: a methodo\fogi - ca\f, theoretica\f, and socio\fogy of science ana\fysis. In Arriaga X, Oskamp S, eds. Vio\fence in intimate re\fationships , pp. 17– 44. Thousand Oaks (\bA), Sage.

Straus M A (2009). Gender symmetry in partner vio\fence: the evidence and the imp\fi- cations for primary prevention and treatment. In Lutzker JR, Whitaker DJ, eds. Preven- tion of partner vio\fence . Washington D\b, American Psycho\fogica\f Association.

Stuart GL et a\f. (2003). Reductions in marita\f vio\fence fo\f\fowing treatment for a\fcoho\f dependence. Journa\f of Interpersona\f Vio\fence , 18 : 1113 – 1131.

Swart LA et a\f. (2002). Vio\fence in ado\fescents’ romantic re\fationships: findings from a survey amongst schoo\f-going youth in a South African community. Journa\f of Ado\fes- cence , 25(4):385 –395.

Taft \bT et a\f. (2009). Intimate partner vio\fence against African American women: an examination of the socio-cu\ftura\f context. Aggression and Vio\fent Behavior , 14:50 –58.

Tang \bS, Lai BP (2008). A review of empirica\f \fiterature on the preva\fence and risk markers of ma\fe-on-fema\fe intimate partner vio\fence in contemporary \bhina, 1987– 2006. Aggression and Vio\fent Behavior , 13(1):10 –28.

Testa M (2004). The ro\fe of substance use in ma\fe-to-fema\fe physica\f and sexua\f vio- \fence: a brief review and recommendations for future research. Journa\f of Interpersona\f Vio\fence , 19(12):1494 –1505.

Tjaden P, Thoennes N (2006). Extent, nature, and consequences of rape victimiza- tion: findings from the Nationa\f Vio\fence Against Women Survey. Washington D\b, US Department of Justice. Avai\fab\fe at: www.ojp.usdoj.gov/nij/pubs-sum/210346.htm Usdin S et a\f. (2005). Achieving socia\f change on gender-based vio\fence: a report on the impact eva\fuation of Sou\f \bity’s fourth series. Socia\f Science and Medicine, 61(11):2434 – 2 4 45.

Uthman OA, Lawoko S, Moradi T (2009). Factors associated with attitudes towards intimate partner vio\fence against women: a comparative ana\fysis of 17 sub-Saharan countries. BMC Internationa\f Hea\fth and Human Rights , 9:14.

Vest JR et a\f. (2002). Mu\ftistate ana\fysis of factors associated with intimate partner vio\fence. American Journa\f of Preventive Medicine , 22(3):156 –164.

Vos T et a\f. (2006). Measuring the impact of intimate partner vio\fence on the hea\fth of women in Victoria, Austra\fia. Bu\f\fetin of the Wor\fd Hea\fth Organization , 84(9):739 – 74 4 .

Vung ND, Krantz G (2009). \bhi\fdhood experiences of interparenta\f vio\fence as a risk factor for intimate partner vio\fence: a popu\fation-based study from northern Vietnam.

Journa\f of Epidemio\fogy and Community Hea\fth , 63:708 –714. references 90 Preventing intimate Partner and sexual violence against women Wa\fby S (2004). The cost of domestic vio\fence. Women and Equa\fity Unit, DTI. Lon- don. Avai\fab\fe at: www.\fancs.ac.uk/fass/socio\fogy/papers/wa\fby-costdomesticvio\fence.

pdf We\fbourn A (2009). Stepping Stones – \fist of surveys and reports to 2006 and some quotes from Stepping Stones users around t he wor\fd. Some brief notes prepared for t he U NAIDS pre-Think Tank Meeting on Eva\fuation Strategies for Prevention Interven- tions, Geneva. Avai\fab\fe at: www.steppingstonesfeedback.org/resources/22/ We\fbourn_ Quotes_U NA IDS_ Presentation _ 2009.pdf We\fsh P (1997). Hacia una mascu\finidad sin vio\fencia en \fas re\faciones de pareja [Towards mascu\finity without partner vio\fence]. Managua, \bANTER A.

Whitaker DJ et a\f. (2006). A critica\f review of interventions for the primary prevention of perpetration of partner vio\fence. Aggression and Vio\fent Behavior , 11:151–166.

Whitaker DJ, Baker \bK, Arias I (2007). Interventions to prevent intimate partner vio- \fence. In: Do\f\f L et a\f., eds. Handbook of injury and vio\fence prevention , pp.203 –223. New York (NY), Springer.

Wins\fow \bEA (1920). The unti\f\fed fie\fds of pub\fic hea\fth. Science , 51(1306):23 –33.

WHO-ISP\bAN (2006). Preventing chi\fd ma\ftreatment: a guide to taking action and gene- rating evidence . Avai\fab\fe at: www.who.int/vio\fence_injury_prevention /vio\fence /activi - ties/chi\fd_ma\ftreatment/en/index.htm\f WHO (2001). Putting women first: ethica\f and safety recommendations for research on domestic vio\fence against women. Geneva, Wor\fd Hea\fth Organization. Avai\fab\fe at:

www.who.int/gender/vio\fence /en/womenfirtseng.pdf WHO/ U NAIDS (2010). Addressing vio\fence against women and HIV/AIDS: What works? Geneva, Wor\fd Hea\fth Organization. Avai\fab\fe at www.who.int.reproductive- hea\fth /topics /vio\fence.

Wi\f\fiams JR et a\f. (2008). Fema\fe perpetration of vio\fence in heterosexua\f intimate re\fa- tionships: ado\fescence though adu\fthood. Trauma Vio\fence & Abuse , 9(4):227–249.

Wo\ffe DA et a\f. (2003). The effects of chi\fdren’s exposure to domestic vio\fence: a meta- ana\fysis and critique. C\finica\f Chi\fd and Fami\fy Psycho\fogy Revie\b , 6(3):171–187.

Wo\ffe DA et a\f. (2003). Dating vio\fence prevention with at-risk youth: a contro\f\fed outcome eva\fuation. Journa\f of Consu\fting and C\finica\f Psycho\fogy , 71(2):279 –291.

Wo\ffe DA et a\f. (2009). A schoo\f-based program to prevent ado\fescent dating vio\fence:

a c\fuster randomized tria\f. Archives of Paediatrics & Ado\fescent Medicine , 163 ( 8 ) : 69 2 – 69 9.

Xu X et a\f. (2005). Preva\fence of and risk factors for intimate partner vio\fence in \bhina.

American Journa\f of Pub\fic Hea\fth , 95:78 – 85. 91 ANNEX A Operational definitions For measuring the magnitude of intimate partner vio\fence and sexua\f vio\fence, it is important to have operationa\f definitions based upon specific behaviours. This he\fps to avoid subjective interpretations and a\f\fows for comparabi\fity across sites. Having opera- tiona\f definitions is necessary both for popu\fation-based epidemio\fogica\f surveys, and for monitoring cases that present to hea\fth faci\fities, the po\fice or other services.

Population-based surveys In popu\fation-based surveys of intimate partner and sexua\f vio\fence, respondents are usua\f\fy asked from a \fist of specific acts whether they have experienced any abuse.

Research has shown that behaviour-specific questions – such as: have you ever been forced to have sexua\f intercourse against your \bi\f\f? – produce higher rates of positive respons- es than questions asking peop\fe whether they have been “abused” or “raped”. They a\fso a\f\fow for an assessment of re\fative severity. Fo\f\fow-up questions asking how often these behaviours take p\face and for how \fong a\fso a\f\fow researchers to better assess the frequency of the abuse. Physica\f acts that are more severe than s\fapping, pushing or throwing an object at a person are genera\f\fy defined in studies as “severe vio\fence”, though some observers object to the defining of severity so\fe\fy according to the act. A series of practica\f too\fs has been deve\foped (Basi\fe, Hertz & Back, 2007) that he\fps to quantify both a victim’s experience of abuse and perpetrator behaviour with standard- ized and va\fidated questionnaires designed for cross-sectiona\f surveys. Bo\b \f0 shows the operationa\f definitions used in the WHO Mu\fti-country study on \bomen’s hea\fth and domestic vio\fence against \bomen (Garcia-Moreno et a\f., 2005). The WHO instrument has a set of core questions that have been used in many countries in a comparab\fe way, with some countries introducing variations according to \foca\f need. The set of core questions and variations re\fated to domestic vio\fence can be found at: www.who.

int/gender/vio\fence /who_mu\fticountry_study/en /. Simi\far behaviour-specific questions can be deve\foped for physica\f and sexua\f abuse by perpetrators other than intimate partners. This quantitative data shou\fd be comp\femented by qua\fitative data that can he\fp to contextua\fize and better understand the vio\fence.

service-based monitoring and icd -10 codes For the purposes of survei\f\fance and monitoring the use of services, standardized defi- nitions for case reporting are necessary. The internationa\f standard diagnostic c\fassifi- cation for a\f\f genera\f epidemio\fogica\f purposes and many hea\fth management purposes is the Internationa\f \b\fassification of Diseases (I\bD), tenth edition (11th revision cur- rent\fy in progress). 92 Preventing intimate Partner and sexual violence against women BOX 10 Operational definitions used to measure physical and sexual \fiolence, emotional abuse and controlling beha\fiours by an intimate partner P\bysical violence by an intimate partner Moderate violence:

n \bas slapped or had something thrown at her that could hurt her n \bas pushed or shoved or (in 2 sites) had hair pulled Severe violence:

n \bas hit with fist or something else that could hurt n \bas kicked, dragged or beaten up n \bas choked or burnt on purpose n \ferpetrator threatened to use or actually used a gun, knife or other weapon against her Sexual violence by an intimate partner n \bas physically forced to have sexual intercourse when she did not want to n Had sexual intercourse when she did not want to because she was afraid of what partner might do n \bas forced to do something sexual that she found degrading or humiliating Emotionally abusive acts by an intimate partner n \bas insulted or made to feel bad about herself n \bas belittled or humiliated in front of other people n \ferpetrator had done things to scare or intimidate her on purpose, e.g., by the way he looked at her, by yelling or smashing things n \ferpetrator had threatened to hurt someone she cared about Controlling be\baviours by an intimate partner n Tried to keep her from seeing friends n Tried to restrict contact with her family of birth n Insisted on knowing where she was at all times n Ignored her and treated her indifferently n Got angry if she spoke with another man n \bas often suspicious that she was unfaithful n Expected her to ask permission before seeking health care for herself Source: Garcia-Moreno et al. (2005) 93 annex a . oPerational definitions Most countries have issued officia\f guide\fines for coding and reporting, based on either the 9th or 10th revision of I\bD. However, such guide\fines for the c\fassification of inti- mate partner vio\fence and sexua\f vio\fence cases might be used on\fy in a \fimited number of faci\fities and then very se\fective\fy – the resu\fting coding may therefore not be re\fiab\fe.

In such cases, it wi\f\f probab\fy be he\fpfu\f to set up a working group to improve c\fassifica- tion and to deve\fop an agreed set of guide\fines on how to assign I\bD codes to known and suspected cases of intimate partner vio\fence and sexua\f vio\fence.

The assignment of I\bD codes to a case is performed on\fy after medica\f professiona\fs have provided care and have done their best to reach a definitive conc\fusion as to the externa\f causes of the presenting prob\fem. \b\fassification of cause of death is usua\f\fy based on coroner reports and inquest findings. For non-fata\f cases, the c\fassification of cause of injury is genera\f\fy based on information from the fi\fes that record hospita\f inpatient discharges and emergency department visits.

The two main components of an I\bD c\fassification concern the nature of the present - ing condition (for instance, traumatic subdura\f haemorrhage) and its externa\f cause.

Information on the re\fationship of the perpetrator to the victim is often not inc\fuded in reporting making the documentation of, for examp\fe, injuries caused by an intimate partner particu\far\fy cha\f\fenging.

Intimate partner vio\fence and sexua\f vio\fence can have a wide range of hea\fth conse- quences ( Chapter \f). Whi\fe it may be possib\fe to identify certain presenting conditions that are more suggestive of intimate partner vio\fence and sexua\f vio\fence than others, it is through the a\f\focation of an externa\f cause – or E-\bode – that an individua\f case is c\fassified as known or suspected intimate partner vio\fence or sexua\f vio\fence.

I\bD-10 codes for assau\ft are X85 –Y09 and inc\fude:

l YO5 : Sexua\f assau\ft by bodi\fy force, inc\fudes rape (attempted), sodomy (attempt- ed); l Y06 : Neg\fect and abandonment, inc\fudes by spouse or partner, by parent, by acquaint- ance or friend, by other specified persons, by unspecified person; and l Y07: Other ma\ftreatment syndromes, inc\fudes menta\f crue\fty, physica\f abuse, sexua\f abuse, torture by: spouse or partner, by parent, by acquaintance or friend, by officia\f authorities, by other specified persons, by unspecified person, exc\fudes neg\fect and abandonment and sexua\f assau\ft by bodi\fy force.

Those assigning I\bD codes and comp\feting death certificates must be encouraged to document the re\fationship between the perpetrator and the victim where known. 94 ANNEX B Prevention activities that can be used for process evaluation and potential sources of information on such activities Activity Potential sources of information Ad VOCAC y TO dEVE lOP OR IMPROVE InT I M AT E PART nER And SE xuAl VIO lEn CE lEGIS lAT I O n • Number of letters sent • Number of people contacted • \fublic information meetings held and number of attendees • \fress releases prepared/released • Files of letters • Records of phone and in-person contacts • Calendars • Records and files Pu BlIC In FORMATIO n CAMPAIG n • Messages developed and tested • Subscribers/listeners/viewers when each message was placed • Number of times message was run • \fress releases distributed/picked up by the press • Interviews/talks requested • \feople exposed to and recalling the information • Files and focus-group results • Institution reports • Institution reports • Log and media searches • Log • Community surveys C OA lITIO n BuI ldInG • Groups/agencies/communities/people contacted • Meetings held, and members participating • Number/type of actions generated/implemented • Interactions among members of coalition • Log • Calendars and meeting minutes • Meeting minutes • Surveys and interviews M ICROFI nAnCE PROGRAMME • Number of applications • Number, size, location and business type of loans approved • Number of loans repaid • \fre-designed data-collection forms Cu RRIC uluM • q ualifications and training of staff • Recruitment activities and coverage • Inquiries received and how enquirers had heard • Number/demographics of those enrolled, attending and dropping out • Differences between non-participants, participants, drop-outs and target population • Barriers to participation • Number of sessions/modules delivered • Coverage of, and time allotted to, key content areas in each session • q uality of delivery • \farticipant satisfaction/recommendations for improvement • \farticipant recall of messages • Staff satisfaction with programme and other staff, and recommendations for improvement • Files and hiring records • Observation • Log and public-awareness surveys • Telephone log • Enrolment-application and attendance sheets • Follow-up survey of drop-outs/non-participants • Staff log /reports • Observation • Survey • Observation or survey • Survey or interview • Survey or interview • Survey or interview Tel +41-22-791-4136 Fax +41-22-791-4171 Tel +41-22-791-2064 Fax +41-22-791-4489 www\fwho\fi\bt/viole\bce_i\bjury_preve\btio\b