Introduction

Over the past 40 years, feminist influence has brought a profound transformation to the way that violence against women is perceived. If domestic violence was once considered a private family matter, and discussing sexual violence was virtually taboo, violence against women is now recognised as both a serious human rights abuse and a social problem that demands urgent attention from both governments and communities (World Health Organization [WHO], 2005).

During that time, violence against women, such as domestic and family violence and sexual assault, has been conceptualised within various professional service systems and academic disciplines, including criminal law, family law and child protection, psychology, gender studies and public health. Differences in defining the issues by each discipline or approach have impacted the ways in which this social problem is conceptualised, how causes of violence against women are identified, and therefore the nature of policy and program solutions proposed (Jordan, 2011).

Most recently, the public health approach has emerged internationally and in Australia as a major influence on policy-making in the area of violence against women. An important aspect of the public health model is primary prevention, which aims to stop or prevent a health-related problem from occurring in the first place, by trying to address the underlying factors or causes of a problem.

This Issues Paper will reflect on the application of the public health model to the issue of violence against women in Australia and the way that fundamental public health principles have been implemented. A summary of what has developed so far is included, as are some examples of high quality prevention efforts that are gradually increasing our limited knowledge in this area. In looking to the future in the prevention of violence against women, we will examine where major challenges still exist, for example the methodological difficulties associated with evaluating immediate and longer-term impacts of primary prevention activities.

The public health approach

A conceptual approach

A public health approach focuses on preventing health problems in a way that extends better care and safety to entire populations rather than individuals (WHO, 2002).

The public health model is a conceptual model that has widespread acceptance across many disciplines, including health, education and welfare (Hunter, 2011). The aim of conceptualising a particular social problem or health issue within a public health framework is fundamentally one of prevention. Public health models aim to prevent problems from occurring in the first place by targeting key risk factors or social determinants and addressing these at a whole of population level (not just at-risk groups). This has been termed an "upstream" approach to prevention.

The four steps of a public health approach

Public health aims to provide the maximum benefit for the largest number of people. It adopts a particular process for achieving this using the following four steps:

1. Defining the problem

Define the problem through the systematic collection of information about the magnitude, scope, characteristics and consequences of violence. This step aims to understand the "who", "what", "when", "where" and "how". In relation to violence against women, this means understanding who the victims and perpetrators are, where the violence is occurring, and when and how the perpetrators are able to continue it.

2. Identifying risk and protective factors

Understand why violence occurs in terms of the causes and correlates of violence, the factors that increase or decrease the risk of violence, and the factors that might be modified through interventions. This step is about reviewing research to identify what factors put people at risk of violence against women (either perpetration or victimisation), and what protective factors may stop or reduce perpetration and victimisation.

3. Developing and testing prevention strategies and programs

With the information gathered in the previous step, programs and intervention strategies can be designed to target risk and protective factors. These interventions can then be evaluated for process, learning and effectiveness.

4. Ensuring widespread adoption by disseminating the information

Information about "what works" in addressing the factors influencing violence against women requires the evaluation of strategies and interventions with targeted analysis about what aspects of the interventions worked or didn't work. This builds the body of evidence around prevention and enables the information to be disseminated broadly, so that successful interventions can be replicated or spread widely. The aim is to "scale up" by implementing effective and promising interventions in a wide range of settings. The effects of these interventions on risk factors and targeted outcomes should be monitored, and their impact and cost-effectiveness should be evaluated.

Whilst the public health approach was originally designed for disease prevention, it has been modified for preventing violence against women. In the area of violence against women, key underlying determinants and contributing factors in the perpetration of violence against women include:

• lack of gender equality - e.g., rigid gender roles and identities, weak support for gender equality, masculine organisational cultures and masculine sense of entitlement;
• cultural norms around violence - e.g., social norms and practices that are violence-supportive, weak sanctions against violence/violence against women, and previous exposure to violence; and
• lack of access to resources and support systems - e.g., support for the privacy and autonomy of the family, unequal distribution of material resources, limited access to systems of support (VicHealth, 2007).

Primary prevention of violence against women therefore includes any program, campaign, policy or other measure that sets out to address these underlying factors and reduce the likelihood that violence against women will take place. Such actions may include strategies that:

• promote equal and respectful relations between men and women;
• promote non-violent social norms and reduce the effects of prior exposure to violence;
• promote access to resources and systems of support (VicHealth, 2007).

It is important to bear in mind that primary prevention is a relatively new approach to addressing the issue of violence against women. This means there are many things we don't yet know about the effectiveness of various types of interventions, so it is timely and important to critically reflect on the current state of practice and knowledge in this area.

The social-ecological model

An understanding of gender-based violence as having multiple causes is best conceptualised in a social-ecological model, demonstrating the interactive nature of a range of key factors (see Figure 1).

Figure 1: The social-ecological model

Source: Quadara & Wall, 2012, p. 4

Ecological models are a feature of public health based on the idea that influences and determinants are interrelated. Developed initially by Bronfenbrenner (1977), the social-ecological model describes the interrelatedness of different spheres of social life and the interactions between individuals and their environments.¹ Ecological conceptualisations of health or social problems aim to change behaviour by targeting the environmental factors that are most likely to influence people's decisions and actions (Di Clemente, Salazar, & Crosby, 2011)

The social-ecological model reinforces the importance of a comprehensive approach, in which actions at each level of the social ecology work to support the other levels. Because a public health approach is aimed at entire populations, its success also depends on action from more than one organisation or group simultaneously, thereby acknowledging the multiple causes of certain health or social problems (Centers for Disease Control and Prevention, 2004).

Historical development of an ecological framework for violence against women prevention

Academic, Lori Heise of the Center for Health and Gender Equity was one of the early advocates for adopting an ecological, public health approach for understanding the origins of gender-based violence (Heise, 1998). Her theories supported an analysis of gendered violence as having multiple, interactive factors rather than individual explanations such as pathology or socio-political reasoning. Heise pointed out that an ecological approach had already been used to organise research findings on child abuse and could accommodate feminist and social science insights about violence against women. Heise's work was influential in guiding the move towards conceptualising violence against women as a public health issue and using the ecological framework to reflect the full complexity of its underlying causes. Heise's ecological explanation is sufficiently able to explain that although the male dominance of patriarchy provides a partial explanation of violence against women, without the incorporation of other social and individual factors, it cannot explain why only some men perpetrate violence even though they are all exposed to cultural messages of superiority and power. As Heise elaborated, any appropriate theory must account for why individual men, but not all men, are violent, but also why women as a group are targeted.

The US Centers for Disease Control and Prevention has raised relationship and sexual violence as a public health issue and worked to coordinate efforts by standardising definitions and identifying data needs, as well as directing a focus on the prevention of perpetration (DeGue et al., 2012). The Centers for Disease Control and Prevention outline a very clear conceptual breakdown of the steps to implementing the public health approach to prevention of violence. This is discussed in more detail below.

Levels of prevention

A key characteristic of the public health model is a focus on prevention, with the goal of developing prevention understandings, and translating this knowledge into practical application.

The public health framework locates strategies at three temporal points, according to when they occur in the timeline of violence against women occurring:

1. Primary prevention. This refers to strategies aimed at preventing violence before it occurs, including whole-of-population strategies.
2. Secondary prevention (early intervention). This refers to programs that involve early detection of risk or early manifestations of the problem. In terms of violence against women policy and programming, it refers to interventions that target individuals or population sub-groups showing early signs of engaging in violent behaviour, or becoming a victim of violence, or who may be particularly at risk of developing violent behaviours.
3. Tertiary prevention (response or intervention). These are the responses set in motion after the violence has occurred. They aim to reduce the consequences and impacts of violence and prevent recurrence.

Public health terminology is not always clear-cut. For example, it is sometimes argued that tertiary prevention should not be called prevention, as the violence has already happened (Chamberlain, 2008). However, rapid well-coordinated response systems directed at victims and perpetrators may prevent recurrence of violence and/or re-victimisation.

Some frameworks for violence against women, such as the Indigenous Family Violence Framework (Department of Human Services [DHS], 2012), use different terminology such as "early intervention" and "crisis intervention" rather than secondary or tertiary prevention (see Figure 2).

Figure 2: Indigenous Family Violence Framework

Source: DHS, 2012, p. 12

These levels of classification are not rigid nor mutually exclusive. For example, developing a strong criminal justice response or well resourced/integrated response system (tertiary prevention/intervention) can send a strong message to the community that violence is unacceptable and will not be tolerated, potentially influencing social norms that might otherwise condone or support violence (primary prevention).

Developing the public health model of violence prevention

Implementation of a public health model for prevention of violence followed on from successes in prevention of infectious diseases in the United States. As disease prevention expertise expanded to incorporate behavioural changes in lifestyle, successes in this area encouraged exploration of how to target underlying factors associated with interpersonal violence and suicide. During the 1970s, the health sector recognised the huge impact violent behaviour was having on the health of children and young people. As death rates from infectious diseases were dropping, it was noted that suicide and homicide were becoming more prominent as leading causes of death (Dahlberg & Mercy, 2009). This would see the emergence of violence prevention as a legitimate health issue in the United States by the 1980s (Dahlberg & Mercy, 2009). Early success in prevention of youth violence throughout the 1980s and 1990s saw the development of a public health approach to other problems of violence such as child maltreatment, intimate partner violence and sexual violence, with efforts to document the problem and understand the risk and protective factors for each type (Dahlberg & Mercy, 2009).

The World Health Organization followed this approach and adopted a public health strategy to combat the impacts of violence. The 2002 World Report on Violence and Health highlighted interpersonal violence as a leading cause of injury, disability and death worldwide (WHO, 2002)² and became a hallmark for violence prevention. This public health model of prevention has been increasingly influential in policy-making around the world in the prevention of violence against women.³

The work of VicHealth has also been highlighted in Australia and internationally in the development of the public health approach to preventing violence against women. The 2007 framework, Preventing Violence Before it Occurs: A Framework and Background Paper to Guide the Primary Prevention of Violence Against Women in Victoria (VicHealth, 2007) was the result of research looking at hundreds of international studies to identify factors that had been found to correlate with a higher likelihood of violence against women, and how these factors can be represented and targeted at the ecological levels. This framework has been acknowledged internationally as providing a strong basis for a shared understanding of the causes of violence against women (Fergus, 2013).

Benefits of applying the public health model

Public health models are characterised by an approach that emphasises prevention. Rather than "simply accepting or reacting to violence, its starting point is the strong conviction that violent behaviour and its causes can be prevented" (WHO, 2002, p. 45). This "upstream approach" has enabled violence against women to be framed as a threat to public health, acknowledging its impacts on physical and mental health as well as social wellbeing, and its impacts in other areas of social need such as housing and substance abuse. This acknowledges that the effects of violence against women extend beyond just the private and criminal justice spheres alone. This has provided new opportunities for the sector to be involved in guiding prevention work and offering their expertise to inform the necessary strategic investment that should, in the long term, reduce the need for criminal justice and other service responses.

Also advantageous is the public health emphasis on integrating evaluation into prevention programs to obtain vital evidence to compare the effectiveness of different strategies (Chamberlain, 2008). This builds a body of evidence that will be invaluable in preventing violence against women.

Considering violence against women as a public health problem has seen a significant shift, identified by Carmody et al. (2009, p. 15), in terms of the "confidence now expressed" that violence can be stopped; a new optimism that violence is not just random, but rather something that can be predicted and prevented.

The public health approach to preventing violence against women in Australia

In Australia, two significant reports have had a major impact on policy development relating to violence against women, by revealing the extent of impacts violence against women has on the health of women and the national economy.

Firstly, in 2004, a Victorian Health Promotion Foundation (VicHealth) report identified the significant "burden of disease" caused by intimate partner violence:

Intimate partner violence is the leading preventable contributor to death, disability and illness in Victorian women aged 15-44, responsible for more of the disease burden than many well-known risk factors such as high blood pressure, smoking and obesity. (VicHealth, 2004, p. 10)

This analysis reflected a growing acknowledgement internationally of the heavy health-related burden in every country brought about by gendered violence (Dyson, 2012; WHO, 2002, 2013). VicHealth's analysis focused attention on violence against women as a health burden that was "prevalent, serious, preventable" (VicHealth, 2004). Importantly, this led the organisation to develop its framework for preventing violence against women (VicHealth, 2007). Whilst prevention activities have occurred in other Australian states and territories, only Victoria has thus far implemented such "concentrated government policy and leadership" for the prevention of violence against women (Murray & Powell, 2011, p. 144).

Secondly, in 2009, a comprehensive analysis of the economic burden of violence against women and children was outlined in a report commissioned by the National Council to Reduce Violence against Women and their Children. This report estimated the economic cost burden to the Australian economy was to be $13.6 billion in 2009, rising to $15.6 billion by the year 2021-22 unless appropriate action was taken (KPMG, 2009).

The combination of new evidence about the extent of health and economic burdens of this violence provided the impetus for renewed vigour in a national effort to tackle violence against women in Australia. The National Plan to Reduce Violence Against Women and their Children 2010-2022 (the National Plan) (Dept. of Families, Housing, Community Services and Indigenous Affairs [FaHCSIA], 2013) would be the most high profile plan Australia has yet seen to tackle gendered violence. It promotes prevention as a cornerstone of strategic effort in this area.⁴

The Foundation to Prevent Violence against Women and their Children (the Foundation) is a new national organisation launched in July 2013 by the Australian and Victorian Governments (The Foundation, 2013). The role of the Foundation is to drive cultural and attitudinal change to prevent violence against women and their children from the ground up through community engagement and advocacy.

The Foundation will, at a national level, bring together the vital work being done to prevent violence against women and their children. It aims to strengthen engagement on these issues across the broader Australian community, provide opportunities for community leadership as well as support future work in primary prevention.

The establishment of the Foundation will complement the work of Australia's National Research Organisation for Women's Safety (ANROWS), another key initiative of the National Plan. ANROWS is focused on building a rigorous evidence-base regarding the issue of violence against women and their children.⁵

Reflecting on the current status of primary prevention in Australia

Australian frameworks, policy development and standards for primary prevention

An extensive body of evidence about what works in the primary prevention of violence against women is not yet available either internationally or in Australia. At this stage, the evidence is disparate, owing to the diverse nature of programs or the specificity of the particular cultural setting. However, efforts are being made to try to establish frameworks and standards to guide developers of programs for primary prevention, based on the best knowledge we have to date.

Population specific frameworks and policies

The Victorian Health Promotion Foundation has provided strong leadership in Australia by developing a comprehensive framework for primary prevention of violence against women (VicHealth, 2007).

Other frameworks and policy documents have been developed to guide activities in particular intervention settings or populations. The Indigenous Family Violence Framework (DHS, 2012) sets out guiding principles and key strategies for initiatives to prevent violence in Indigenous communities, appropriate to the particular needs and challenges they face. Another important piece of policy work developed by the Multicultural Centre of Women's Health is designed to guide development of violence prevention efforts appropriate to the particular needs of immigrant and refugee communities (Poljski, 2011). Australia's National Disability Strategy 2010-2020 (COAG, 2011) contains policy goals around enabling people with a disability to be safe from violence, exploitation and neglect. The Commonwealth Government has also funded the Stop the Violence research project through Women with Disabilities Australia, aimed at improving service responses to women with a disability. Although not primarily a prevention project, research will include mapping of good policy and practice models to help prevent violence against women.

Practical step-by-step guides and toolkits have also been developed to assist communities or organisations wishing to design or develop high-quality prevention programs suitable for particular settings, such as the workplace or rural and remote settings (National Rural Women's Network, 2012; Women's Health Victoria, 2012).

Guidelines and good practice principles for schools prevention education

In the area of school education, much work has been done to establish best practice principles for violence prevention programs (Carmody et al., 2009; Flood, Fergus, & Heenan, 2009). Good practice principles such as these are considered essential for both designers of primary prevention programs, and to provide criteria against which to evaluate program quality and effectiveness.

Arguably, Australia is at step three in the public health process, as described earlier, where the steps must be applied iteratively. This means that prevention and intervention strategies are still being developed for testing and evaluated for process learning and effectiveness. Despite a lack of evaluated evidence on what does work, there are resources to guide the efforts. Some work has been done on what constitutes effective primary prevention and what elements are required to be present for primary prevention to work.⁶ VicHealth has worked towards changing the paradigm through which violence against women is viewed and supporting this by producing locally relevant resources.⁷

Promising practice examples in primary prevention

Building the evidence base is unavoidably a gradual process. As knowledge about preventing violence develops with each aspect of programming, we can utilise what evolves to shape prevention efforts.

A range of settings and delivery formats are used for primary prevention activities. Examples include face-to-face classroom programs, television commercials for the general public, social media, community action, as well as legislative and policy reform. Each intervention may target different levels of the ecological model and therefore be designed for a specific setting or population. Programs and interventions may be designed to target sexual violence against women, domestic and family violence, or violence against women in general.

This section highlights some promising programs across various formats, targeting a range of different settings and influences within the ecological framework.

School-based education programs

The greatest level of primary prevention activity in Australia has been implemented in the school setting. Evaluation has also provided the strongest evidence of effectiveness within this setting, both internationally and in Australia (Carmody et al., 2009; Cornelius & Resseguie, 2006; Flood et al., 2009; Foshee & Reyes, 2009). The provision of dedicated funding for Respectful Relationships projects under the National Plan has also stimulated activity in this area. The Respectful Relationships funding stream supports projects in each state and territory designed to prevent sexual assault and domestic and family violence through education of students and the whole school community. Such projects work to target potential perpetrators, victims and bystanders, and raise their awareness of ethical behaviour; to develop protective behaviours; and to develop their skills in conducting respectful relationships.

Advantages of school programs include accessibility, affordability and broad reach, as well as the fact that adolescence is considered an ideal time to influence attitudes and behaviours (Flood et al., 2009). Programs that address an entire school grade or take a whole-school approach may avoid stigmatising high risk individuals or groups; however, more intensive effort beyond the school environment may also be needed, given that teenagers at high risk may already be disengaged from school (Whitelaw et al., 2001).

Community-based education programs

Although school-based education is the most common location of primary prevention activity in Australia today, there is sound rationale for more focus on community-based programs that reach beyond adolescence to try to engage, support and equip people with the knowledge, skills and resources they may need as they transition into different types of relationships, such as cohabitation, becoming a parent, or taking on caring responsibilities for others. We know from research that a high-risk period for domestic and family violence is pregnancy, for example, and family violence is increasingly recognised as taking place within care relationships for the disabled and elderly (Campbell, Garcia-Moreno, & Sharps, 2004; Dowse, Soldatic, Didi, Frohmader, & van Toorn, 2013).

Community action campaigns

Working to change a community environment rather than focus on individuals, community action taps into community resources, social settings and local knowledge to design appropriate programs (Fanslow, 2005). Community buy-in and leadership are considered essential components of successful design and implementation. Promising examples are evident in the setting of sporting clubs, where interventions aim to change macho, violence-supportive attitudes and behaviours, and provide leadership role-models for the broader community.

Social marketing/public education campaigns

Social marketing applies the techniques of commercial marketing to campaigns aiming for positive social outcomes. It generally involves qualitative research to understand the target audience, and to test the design of key messages and delivery formats in advance of the campaign roll out.

Social marketing, or public education campaigns, may involve use of television commercials and other media to convey messages universally to the general public. Such campaigns may be ideally placed to raise community awareness of legislative or policy changes, increase community understandings of what constitutes domestic and family violence or sexual violence, and consequently encourage people to seek help and report violence (Horsfall, Bromfield, & McDonald, 2010; Saunders & Goddard, 2002).

However, advertisements are an expensive public investment and may be unlikely to contribute to the significant behaviour changes needed to reduce levels of violence. In fact, it is important to note that while public education campaigns may be effective in increasing knowledge or community awareness, the deeper attitudinal and behaviour changes required to stop violence against women are likely to require more intensive, direct forms of intervention (Fanslow, 2005; Davies, Hammerton, Hassall, Fortune, & Moeller, 2003).

Targeted social marketing approaches may be more successful than universal public education campaigns, but depend on the existence of media networks willing and able to air them.

Social media

The advantages of online social media include its accessibility, reach and access, younger appeal, potential for interactivity, and the fact that it offers a degree of confidentiality for discussion of sensitive topics. Social media campaigns may work best in collaboration with other campaigns or face-to-face program elements that push or invite participants to go to the website for further discussion or information on an issue.

"Edutainment"

This form of primary prevention may involve integrating information and messages within episodes of existing popular television series or creating new entertainment media products to appeal to particular audiences.

Legislative and policy reform

Reforms in policy and law provide both tertiary prevention (acting to prevent and respond to violence that has occurred) and primary prevention, in that they communicate powerful messages to the community about the unacceptability of violence against women and the penalties or sanctions that will be applied. Changes to law and policy may be usefully paired with public education campaigns to promote new penalties and/or improved support for victims.

Saturation approach: A place-based trial

An innovative primary prevention approach that is undergoing trial in Victoria at present involves a suite of well-developed, trialled and tested programs being simultaneously implemented in one geographic location, to evaluate the impact of a place-based saturation approach. Theoretically, this comprehensive approach to implementing primary prevention, which saturates a particular community with messages, programs and interventions across different settings, should have a greater impact than programs implemented in single settings. Five key settings identified for intervention are: the workplace, maternal/child health services, faith organisations, the youth sector, and local government.

Significant effort and resources have been applied by VicHealth to develop, scale-up and implement the setting-based approach and to establish a place-based saturation trial, including carefully designed evaluation methods and measures. It is therefore predicted that this trial will provide important new evidence for prevention of violence against women within the next couple of years. The multi-setting nature of this approach and the fact it targets different life course stages, beyond school-age adolescence, may hold particular promise in terms of primary prevention of domestic and family violence.

Current challenges in primary prevention

The challenge of effective program design

US research has identified components that must be present for primary prevention to be effective in the sexual assault sector (Casey & Lindhorst, 2009), and this is backed up by Australian findings that concluded similar necessary elements (Flood et al., 2009). It is important that the sectors involved in a public health model of prevention have a similar understanding of what is required for effective primary prevention to ensure that resources are most strategically directed⁸

At its worst, poor program design can be dangerous rather than just ineffective. An example is oversimplifying the task of being a bystander and the potential increased safety risk for vulnerable women and children:

Some evaluations have found that members of the public have been encouraged to directly confront perpetrators of violence against women in a way that increases the danger to victims. (Davies et al., 2003, p. 62)

Another example is education programs and training delivered by people not adequately trained or resourced to know how to respond to possible disclosures from recipients (for example from school students). For effective delivery to be present, contextualised programming that is capable of responding to local perspectives and be relevant to all stakeholders is an essential factor. This includes being able to respond appropriately to disclosures where the program may be a catalyst to recipients disclosing.

The challenge of acknowledging gender in violence against women prevention

Primary prevention approaches must recognise gender inequality and gender stereotyping as key underlying factors associated with men's violence against women. To ignore the gendered nature of such violence excludes essential elements that must be tackled for effective primary prevention:

… the erasure of gender from the theoretical frameworks that guide public health efforts may have serious consequences, namely the development of misguided and ineffective prevention and intervention programs. (Reed et al., 2010, cited in Western & Mason, 2013, p. 349)

As Casey and Lindhorst (2009) noted, what is required is an explicit theory of change that links prevention activities to theory about the causes of the problems and theory of behavioural and community change. Without a feminist-informed theory of change and program logic to frame the problem, primary prevention interventions will lack the fundamental link between gendered power/inequality and violence against women. This is described in more detail in the report on violence prevention programs in schools, Respectful Relationships Education: Violence Prevention and Respectful Relationships Education in Victorian Secondary Schools (Flood et al., 2009). An appropriate change theory will address culturally structured beliefs and norms about heterosexuality, masculinity and femininity, along with a feminist-informed conceptual theory to understand these links.

The challenge of evaluation

A key aspect of a public health approach to social problems is to expand and draw on a strong evidence base to direct the most effective use of resources. As discussed earlier, the evidence at this stage of the process is quite disparate and because still new, much of the evidence has not yet been measured over a longer term to ensure the impacts of change are lasting. This doesn't mean that primary prevention won't work, it means there is much to be done in evaluating and documenting change. One of the major barriers to doing this is the lack of resources provided or allocated for evaluation. This results in only short-term changes such as attitudinal change or increased awareness of issues, for example, the ability to recognise and challenge sexist behaviour, being measured rather than the actual levels of violence (Fergus, 2013).

For evaluators of primary prevention of violence against women there are three key challenges:

• the social silence around issues of violence against women and complications in terms of evaluation relying on self-reporting;
• that awareness is not the same as behaviour change, so programs need to focus on evaluating changes to behaviours, not just the acquisition of new knowledge or even changed attitudes; and
• the long period of time required to measure the impacts of violence against women primary prevention, more so than for other public health issues.

Particular challenges for evaluators of primary prevention arise because of the social silence around the problems of sexual violence, domestic and family violence. This is compounded by the privacy and ethical issues evaluators in this area must take into account when seeking to gather reliable data, as well as problems inherent in relying on self-reports about the perpetration or experience of violence. These problems are not insurmountable but need to be acknowledged in evaluation design.

Evidence may point to or detect awareness-raising/increased knowledge rather than actual changes in behaviour, as it is well known that people don't always act according to what they know, even when they believe in and understand it (e.g., smokers don't necessarily quit just because they know smoking is carcinogenic; rather, a range of coordinated levers, supports and skill developments are called for to bring about smoking behaviour change). It is relatively easy to detect changes in beliefs and attitudes that come about as a result of violence against women primary prevention; however, it is much more difficult to evaluate behaviour changes that may be effective in reducing violence against women.

Furthermore, the nature of primary prevention means there will be a relatively long time period between the intervention and the desired outcome of behaviour change. The need to address underlying factors or social determinants to have an effect on behaviour means that no change will be instant, and often it won't be one factor alone that will influence change. An expectation that primary prevention interventions will be immediately recognisable as "successful" or "unsuccessful" is limiting for the process of developing an evidence base. It does not allow for the time required to build an evidence base grounded in the experiences of particular interventions that work on particular aspects of the problem. An incremental approach to primary prevention is inherent in the concept of a public health approach that relies on evidence.

Concerns are sometimes raised about a lack of research evidence in primary prevention, that implementing prevention programs is therefore potentially misdirected and, in any case, that the task of trying to measure required social changes is too big a hurdle. But deeper analysis of these problems shows it is possible and important to consider the specifics of what to measure, to show that an intervention has had a particular impact, and to assess whether that impact is a sign of success.

By bringing focus back to the purpose of the evaluation, and recognition that the values emphasised by stakeholders in an evaluation will affect the type of knowledge produced, then a clear understanding of what may constitute "success" can be formed, and evaluation designed to capture this information (Wall, 2013).

By its nature, primary prevention is innovative and aimed at addressing underlying determinants of large social problems. The purpose of evaluation of primary prevention is to tell us whether there was an impact with a particular intervention and what that impact was. This includes whether the changes planned within the "logic theory" of a particular innovation have been achieved. Evaluation should give an understanding of what occurred, noting any additional "ripple" effects that were unintended but may be positive or negative, and what the process of the intervention can tell us about change (Trickett, 2009).

"Utilisation-focused evaluation" is strongly advocated by prominent evaluation academic, Michael Patton (2008). It means that in planning to evaluate, the key purpose of designing the evaluation should be to extract the information that will be most useful to the various stakeholders involved and to make it useable.

By focusing on the purpose of the evaluation, one aspect that becomes important to acknowledge is the need for the goals and objectives of an intervention to be clear. All stakeholders should be informed about what the intervention is trying to achieve and what their role in supporting the evaluation is. This type of participatory approach to evaluation is important for evaluating complex social change, such as violence against women prevention, because it can acknowledge the different levels and sectors impacting on the problem at hand and can increase the knowledge and skills of those involved in the evaluation to identify and contribute relevant information (Wall, 2013).

In planning an evaluation of multi-level primary prevention programs, it is important to acknowledge interventions at each level. Without this, it would be impossible to assess the interaction influence each has had on the problem at hand (Nastasi & Hitchcock, 2009).

In relation to violence against women, the lack of an extensive evidence base from which to draw knowledge for intervention design provides the sector with a challenge. It is also an opportunity to design evaluations to inform an evidence base that can provide stakeholders with the most useful information possible. Ideally, program evaluation includes a baseline evaluation of attitudes and behaviours for a target group, as well as follow-up evaluations with participants at later stages to monitor lasting effects on awareness and behaviours in the years that follow. For effective evaluation to occur, appropriate measurement tools and methods must also be developed and applied.⁹

The challenge of resources

Many primary prevention interventions have the effect of increasing awareness of violence against women, which can increase reporting. Where this can be problematic is where there is no corresponding increase in adequately resourced services to respond to crises and those seeking help. This may be one of the so-called "ripple" effects of an intervention. Increased primary prevention activity is likely to lead to increased demand for secondary and tertiary service provision, as it appears that an increase in help-seeking and reporting of violence against women may occur as societies "break the silence" on issues of interpersonal violence:

(As) public awareness campaigns can increase disclosure of violence and abuse … services must be sufficiently resourced to be able to meet any increased demand generated … Awareness campaigns risk endangering women and children's safety, if requests for help are not responded to promptly and appropriately. (Davies et al., 2003, p. 62)

Coinciding with the state of Victoria's focus on primary prevention interventions over the past decade, numbers of reported incidents of domestic and family violence have more than doubled, with a 21.6% increase in just the last 12 months (Victoria Police, 2013). These figures challenge us to consider whether they represent increases in actual levels of violence taking place within the community, or rather, whether they are an indication of rising community awareness enabling people to seek assistance with increasing confidence about what help is available (Bucci, 2013; Lay & Murdoch 2013).

Better qualitative and quantitative research will be necessary to truly understand what is going on. However, increased reporting indicates that as governments embark on primary prevention, there is a corresponding need to ensure that tertiary level services can respond adequately to people seeking help or justice. Lives could be endangered if people are unable to be made safe or be adequately assisted in a time of crisis.

In fact, there is evidence that well-resourced services and structural reform to enhance support and strengthen safety (tertiary prevention) ultimately contribute to creating community and social environments that are less tolerant of violence (primary prevention) (Chamberlain, 2008). Thus, providing access to resources and supports adds both tertiary level assistance for those victimised and makes an impact on primary-level prevention.

Where to from here?

As Australia enters an exciting phase in primary prevention, with a clearly articulated goal to reduce violence against women and a national emphasis on prevention, it is timely to consider the next steps.

National Standards, definitions, guidelines

The challenges outlined above provide a good starting point for moving forward. The first step forward must be to agree on what primary prevention is and what elements will need to be present to make it most effective. This agreement could come from national guidelines, definitions and standards that work as a resource and checklist for program design, program implementation and program evaluation. By providing a blueprint of what is and what isn't primary prevention, and what elements are necessary to make it most effective, a strong basis is laid for the next phase of understanding violence against women prevention.

Considering other fields of public health to identify risk factors protective elements

Violence against women, in its various forms, shares risk factors that can be present in other fields of public health. For example, developmental experiences of violence or sex, attitudes and beliefs around peer behaviour, social norms and behavioural change can all be factors relevant to other social problems such as alcohol or substance misuse,¹⁰ HIV and unsafe sexual practices,¹¹ and bullying. Existing frameworks such as the National Safe Schools Framework¹² could be drawn on to provide transferable information. These fields may provide us with a rich source of information that could be applied to violence against women prevention.

Incorporate strategies to address structural change

The societal and community levels of the ecological model are the areas of program design that seem to be the most difficult to target for change. This is an area where social norms, cultural beliefs and violence-supportive attitudes can be addressed, but it is also the area where fewest programs are available and where changes are most difficult to measure. The broader changes required are usually intertwined with influences from different spheres and from other levels and require more complex programming and deeper consideration of how to evaluate them. The changes made to date, in beginning the awareness raising and addressing individual and interpersonal factors that support violence against women, can be a building block to extension of prevention messages and structural change to the wider sphere needed to ensure reform. This may require, for example, legal reform and policy measures to eliminate discrimination and other barriers to gender equality, as well as actively encouraging increased gender equity in the workplace and other spheres.

Translating effective primary prevention evidence into a variety of settings

Reaching agreement on what is needed for primary prevention to be effective is only a start. Being able to incorporate this effectiveness into a variety of settings will then be paramount in reproducing successful program design. This is about contextualising programs to enable similar goals, such as "promoting respectful relationships", to be transferable to contextually different settings, such as an Indigenous remote community versus an urban school setting, a rural organisation or city-based corporate workplace. Contextual relevance is already recognised as one of the requirements for effective primary prevention (Quadara & Wall, 2012) but extracting the essential elements of a program and creating sufficient flexibility to ensure it can be reproduced in each different setting will be vital.

Understanding what works

Evaluation and measurement of interventions are the only way we can understand whether primary prevention interventions are working. However, this may require further research and debate to establish what is understood by "success". If the goal is "reduction", how much reduction should there be, and what will be the indicators to show that it has occurred? More specific analysis of the outcome measures, such as having respectful relationships, is also in need of clear agreement. Without a clear understanding of what we are looking for to articulate success or failure, it is difficult to interpret success or failure as sufficient, complete or in need of refinement. An important starting point to understanding change has been the development of national indicators, which can be used for tracking change and evaluating whether primary prevention efforts are effective:

• Australian Gender Indicators - first published by the Australian Bureau of Statistics (ABS) on 26 August 2011, now a biannual product.
• National Survey of Community Attitudes to Violence Against Women - results from the 2013 survey will be published soon. The survey gauges contemporary attitudes held about violence against women and tracks changes in attitudes since 1995.
• Personal Safety Survey - on 11 December 2013, the ABS released the Personal Safety Survey 2012. This is the second time the Personal Safety Survey has been conducted. The survey was last run by the ABS in 2005, with a previous comparable survey on Women's Safety undertaken in 1996. The Personal Safety Survey collected information about the nature and extent of violence experienced by men and women since the age of 15, including their experience of violence in the 12 months prior to the survey. It also collected detailed information about men and women's experiences of current and previous partner violence, lifetime experiences of stalking, physical and sexual abuse before the age of 15, and general feelings of safety.

Contributing to the knowledge base

Commonwealth activities under the First Action Plan of the National Plan are still underway. Some of these activities offer an early opportunity to add understanding to the primary prevention space. These include, for example, an evaluation of the Australian Government's Respectful Relationships Program being undertaken by the Institute for Social Science Research at the University of Queensland, due for completion in April 2014 (FaHCSIA, 2013).

Conclusion

Australia has strongly committed to a public health approach to the problem of violence against women. This follows global developments in treating gendered violence as a public health issue. Australia has the benefit of a knowledge base provided by organisations and countries that are further along in the primary prevention process, such as the Centers for Disease Control and Prevention in the US, the United Nations and the World Health Organization. The emphasis is now clearly on prevention of violence against women before it occurs.

However, despite having articulated a focus for strategic effort on prevention and increasing the evidence base in the National Plan to Reduce Violence against Women and their Children, this paper's reflection on the Australian journey so far shows that there are still challenges to overcome. These include developing a shared understanding of definitions, strategies and priorities for change. Another difficulty is the long timeframe required to produce evidence of what works in primary prevention, which is an impediment to developing effective interventions in the short term. Evaluation is imperative, as is the need to incorporate agreed indicators of change and outcome measurement so that "success" has a shared meaning for all stakeholders.

As breaking the silence on the topic of violence against women, via a range of primary prevention measures, appears to have the effect of increasing reporting, there is an evident need for well-resourced services to be ready to respond to those seeking help or reporting violence. In addition, it will be essential to ensure that primary prevention of violence against women retains its focus as a gendered issue.

The goal of preventing the perpetration of violence is an exciting prospect that has generated an unprecedented energy in the public health sector and beyond. It is important that the next steps taken maintain this positivity and that Australia can benefit from innovation in intervention and evaluation design inherent in the public health approach.

References

• Bucci, N. (2013, 28 August). Domestic violence drives state crime rate higher, Victoria Police say. The Age. Retrieved from <www.theage.com.au/victoria/domestic-violence-drives-state-crime-rate-higher-victoria-police-say-20130828-2sp9n.html>.
• Campbell, J., Garcia-Moreno, C., & Sharps, P. (2004). Abuse during pregnancy. Violence Against Women, 10, 770-789.
• Carmody, M., Evans, S., Krogh, C., Flood, M., Heenan, M., & Ovenden, G. (2009). Framing best practice: National standards for the primary prevention of sexual assault through education. Hobart: National Association of Services Against Sexual Violence.
• Casey, E., & Lindhorst, T. (2009). Toward a multi-level, ecological approach to the primary prevention of sexual assault: Prevention in peer and community contexts. Trauma, Violence and Abuse, 10(2), 91-114.
• Centers for Disease Control and Prevention. (2004). Sexual violence prevention: Beginning the dialogue. Atlanta, GA: Centres for Disease Control and Prevention.
• Chamberlain, L. (2008). A prevention primer for domestic violence: Terminology, tools, and the public health approach. Pennsylvania: The National Online Resource Center on Violence Against Women. VAWnet.
• Cornelius, T. L., & Resseguie, N. (2006). Primary and secondary prevention programs for dating violence: A review of the literature. Aggression and Violent Behavior, 12, 364-375.
• Council of Australian Governments. (2011). National Disability Strategy 2010-2020. Canberra: COAG. Retrieved from <www.coag.gov.au/sites/default/files/national_disability_strategy_2010-2020.pdf>
• Dahlberg, L., & Mercy, J. (2009). The history of violence as a public health issue. Atlanta: Centers for Disease Control and Prevention. Retrieved from <www.cdc.gov/violenceprevention/pdf/history_violence-a.pdf>.
• Davies, E., Hammerton, H., Hassall, I., Fortune, C. A., & Moeller, I. (2003). How can the literature inform implementation of Action Area 13 of Te Rito: Public education and awareness. Wellington, NZ: Ministry of Health and Ministry of Social Development.
• DeGue, S., Simon, T., Basile, K., Lin Yee, S., Lang, K., & Spivak, H. (2012). Moving forward by looking back: Reflecting on a decade of CDC's work in sexual violence prevention, 2000-2010. Journal of Women's Health, 21(12), 1211-1218.
• Department of Families Housing Community Services and Indigenous Affairs. (2013). National Plan to Reduce Violence Against Women and Their Children 2010-2022: Progress report to the Council of Australian Governments. Canberra: FACHSIA.
• Department of Human Services. (2012). Indigenous Family Violence Primary Prevention Framework. Melbourne: Department of Human Services.
• Di Clemente, R., Salazar, L., & Crosby, R. (2011). Health behaviour theory for public health principles, foundations and applications. Burlington, MA: Jones and Barret Learning.
• Dowse, L., Soldatic, K., Didi, A., Frohmader, C., & van Toorn, G. (2013). Stop the violence: Addressing violence against women and girls with disabilities in Australia (Background paper) Hobart: Women with Disabilities Australia.
• Dyson, S. (2012). Preventing violence against women and girls: From community activism to government policy. Bangkok: Partners for Prevention.
• Fanslow, J. (2005) Beyond zero tolerance: key issues and future directions for family violence work in new Zealand. Wellington, NZ: Families Commission.
• Fergus, L. (2013). International evidence base on prevention of violence against women and their children. Unpublished paper. Foundation to Prevent Violence Against Women and Their Children, Melbourne.
• Flood, M., Fergus, L., & Heenan, M. (2009). Respectful relationships education: Violence prevention and respectful relationships education in Victorian secondary schools. Melbourne: Department of Education and Early Childhood Development.
• Flynn, D. (2011). Baby Makes 3: Project report. Melbourne: Whitehorse Community Health Centre.
• Foshee, V. A., & Reyes, H. L. M. (2009). Primary prevention of adolescent dating abuse perpetration: when to begin, whom to target, and how to do it. In D. J. Whitaker, & J. R. Lutzker (Eds.), Preventing partner violence: Research and evidence-based intervention strategies, (pp. 141-168). Washington, DC: American Psychological Association.
• Foundation to Prevent Violence Against Women and their Children. (2013). Welcome. Melbourne: Foundation to Prevent Violence Against Women and their Children. Retrieved from <www.preventviolence.org.au/index.htm>
• Heise, L. (1998). Violence against women, an integrated, ecological framework. Violence Against Women, 4(4), 262-290.
• Horsfall, B., Bromfield, L., & McDonald, M. (2010). Are social marketing campaigns effective in preventing child abuse and neglect? (NCPC Issues Paper No. 32). Melbourne: Australian Institute of Family Studies.
• Hunter, C. (2011). Defining the public health model for the child welfare services context (NCPC Resource Sheet). Melbourne: Australian Institute of Family Studies. Retrieved from <www.aifs.gov.au/nch/pubs/sheets/rs11/>.
• Jordan, C. E. (2011). Building academic research centers to advance research on violence against women: An empirical foundation. Violence Against Women, 17(9), 1123-1136.
• KPMG. (2009). The cost of violence against women and their children: Report commissioned by the National Council to Reduce Violence against Women and their Children. Canberra: FaHCSIA
• Lay, K. & Murdoch, M. (2013, May). Activating institutions in the primary prevention agenda. Paper presented at the Global to Local: White Ribbon Australia international conference, Sydney.
• McFerran, L. & Walden, I. (2012). Keeping women safe at work: Why domestic violence is a workplace issue, Human Rights Defender, 21(3), pp. 151-157.
• Morgan, J. & Politoff, V (2012). Victorian print media coverage of violence against women: Summary report. Melbourne: University of Melbourne & VicHealth.
• Murray, S., & Powell, A. (2011). Domestic violence: Australian public policy. Melbourne: Australian Scholarly Publishing.
• National Rural Women's Network (2012). Stopping violence before it happens: A practical toolkit for communities. Kyneton,Vic.: National Rural Women's Network & National Rural Women's Coalition. Retrieved from <www.nrwc.com.au/Portals/43/Documents/WAV%20Tool%20Kit/NRWN-Family-Violence-Toolkit-Web.pdf>
• Nastasi, B., & Hitchcock, J. (2009). Challenges of evaluating multilevel interventions. American Journal of Community Psychology, 43, 360-376.
• Patton, M. (2008). Utilization-focused evaluation (4th ed.). St Paul, MN: Sage Publications Inc.
• Poljski C. (2011). On her way: Primary prevention of violence against immigrant and refugee women in Australia. Melbourne: Multicultural Centre for Women's Health.
• Powell, A. (2010). Review of bystander approaches in support of preventing violence against women. Melbourne: VicHealth.
• Quadara, A., & Wall, L. (2012). What is effective primary prevention in sexual assault? Translating the evidence for action (ACCSA Wrap No. 11). Melbourne: Australian Insitute of Family Studies.
• Saunders, B., & Goddard, C. (2002). The role of mass media in facilitating community education and child abuse prevention strategies (Child Abuse Prevention Issues No. 16). Melbourne: Australian Institute of Family Studies.
• Trickett, E. J. (2009). Multilevel community based culturally situated interventions and community impact: An ecological perspective. American Journal of Community Psychology, 43, 257-266
• Upston B, Durey R, & Women's Health Victoria. (2012). Everyone's business: A guide to developing workplace programs for the primary prevention of violence against women. Melbourne: Women's Health Victoria.
• VicHealth. (2004). The health costs of violence: Measuring the burden of disease caused by intimate partner violence. A summary of findings. Melbourne: Department of Human Services.
• VicHealth. (2007). Preventing violence before it occurs: A framework and background paper to guide the primary prevention of violence against women in Victoria. Melbourne: VicHealth. Retrieved from <www.vichealth.vic.gov.au/Publications/Freedom-from-violence/Preventing-violence-before-it-occurs.aspx>.
• VicHealth. (2012). Guide to reporting on violence against women. Melbourne: VicHealth. Retrieved from <www.vichealth.vic.gov.au/reportingVAW>.
• Victoria Police. (2013). Crime statistics 2012/13. Melbourne: Victoria Police.
• Wall, L. (2013). Issues in evaluation of complex social change programs for sexual assault prevention (ACSSA Issues No. 14). Melbourne: Australian Institute of Family Studies.
• Western, D., & Mason, R. (2013). Gender-based violence in Australia: A state-based joined-up approach. In K. Nakray (Ed.), Gender-based violence and public health: International perspectives on budgets and policies (pp. 79-92). UK: Routledge.
• Whitelaw, S., Baxendale, A., Bryce, C., Machardy, L., Young, I., & Witney, E. (2001). "Settings" based health promotion: A review. Health Promotion International, 16(4), 339-353.
• World Health Organization. (2002). World report on violence and health. Geneva: WHO
• World Health Organization. (2005). Multi-country study on women's health and domestic violence against women, summary report: Initial results on prevalence, health outcomes and women's responses. Geneva: WHO.
• World Health Organization. (2013). Global and regional estimates of violence against women: Prevalence and health effects of intimate partner violence and non-partner sexual violence. Geneva: WHO.