Looking Ahead Toward Community-Level Strategies to Prevent Sexual Violence

Sexual violence (SV) is a serious public health problem affecting millions of women and girls around the world each year. Estimates from the National Violence Against Women Survey, collected in 1998, suggested that at least one in six women in the United States had experienced an attempted or completed rape in her lifetime, with more than half of all rapes occurring before the age of 18. ¹ Further, women and girls are frequently exposed to multiple victimizations; in the same study, women who had been raped in the prior 12 months had experienced 2.9 sexual assaults on average. ¹ A more recent national study estimated that 2.7 million American women had experienced unwanted sexual activity in the preceding 12 months. ² Using data from population-based studies of 48 countries, the World Health Report on Violence and Health also highlighted the global nature of SV. ³ Not only is SV pervasive in our society and around the world, but it is also associated with numerous short-term and long-term health consequences for victims, including posttraumatic stress disorder (PTSD), depression, chronic pain, and reproductive and sexual health problems. ^{4 –6} SV also imposes high costs on society, with estimates from the National Institutes of Justice suggesting that victim costs alone may total $126 billion annually in the United States. ⁷

The mission of the Division of Violence Prevention (DVP) within the Center for Disease Control and Prevention's (CDC) National Center for Injury Prevention and Control (NCIPC) is to prevent violence and reduce its consequences. ⁸ Applying a public health framework to the problem of SV, DVP's work includes surveillance efforts to assess and monitor the scope of the problem, etiologic research to identify risk and protective factors, development and rigorous evaluation of prevention strategies, and dissemination and implementation of evidence-based prevention approaches in communities. ^9,10 This work is typically framed within the social-ecologic model, ¹¹ which conceptualizes violence as a product of multiple, interacting levels of influence at the individual, relationship, community, and societal levels of the social ecology. ⁹ DVP is dedicated to preventing violent behavior before it occurs and has increasingly shifted the focus of research and prevention efforts from victims to perpetrators in order to maximize the likelihood of achieving population-level reductions in the prevalence of SV. ¹⁰

In an effort to advance research and programmatic activities toward the goal of identifying effective, evidence-based strategies for the primary prevention of SV perpetration, we recently undertook a process to systematically review the current evidence supporting programs and strategies to prevent SV. We conducted a comprehensive review of the literature spanning the last three decades, across populations and types of prevention strategies. All available published and unpublished outcome evaluations of primary prevention programs or strategies targeting perpetration of SV were included and coded by a team of DVP scientists for program content, study design, and evaluation outcomes. A full report of this systematic review is forthcoming and will include detailed conclusions regarding the current state of the evaluation literature in the SV prevention field, the effectiveness of currently evaluated strategies, and gaps in the prevention and evaluation literature.

In advance of this report, we would like to highlight one critical gap that we believe should be a focus of increasing attention in the next several years: the evaluation of community-level strategies to prevent SV. Community-level strategies function by modifying the characteristics of settings (e.g., schools, workplaces, neighborhoods) that increase the risk for violence victimization and perpetration. These include approaches that operate to change community-level norms, risk factors, or policies within communities, such as programs to improve school climate or the institution or enactment and enforcement of sexual harassment policies. Not surprisingly, the vast majority of programs evaluated for the primary prevention of SV to date have focused primarily on the individual, ¹² with a few strategies in the recent evaluation literature aimed at changing norms or behaviors at the peer group level of the social ecology. ^{13 –15} In contrast, we identified only a few prevention approaches in our review that included community-level components, such as schoolwide poster campaigns ¹⁶ or providing training for school staff or community service providers, ^17,18 and none evaluating programs, strategies, or policies directed solely at the community level. Strategies that target the societal level of the social ecology are also infrequently discussed in the SV prevention literature, with almost none of these strategies systematically evaluated; our review identified only one evaluation that examined the effects of funding provided by the Violence Against Women Act on violent crime, including rape. ¹⁹ We focus here on the need for community-level strategies, as identification of promising strategies at this level represents an important next step in the literature and may help inform the future development of farther reaching societal-level approaches. However, attention to implementing complementary strategies across levels is also needed if effective, multilevel approaches to SV prevention with potential for population-level impact are to be developed.

Existing approaches to SV prevention, which focus mainly on the individual level, have often demonstrated small or short-lived effects. ¹² Although these strategies likely represent an important piece of the prevention puzzle, enacting individual behavior change within an environmental context that continues to support, facilitate, or encourage those behaviors is challenging, and traditional strategies aimed at changing individual attitudes and behavioral intentions may be insufficient when implemented in isolation. Indeed, researchers have argued that individual-level approaches, even when brought to scale and implemented widely may be unlikely to achieve desired impacts on overall rates of violence. ²⁰ Thus, a move toward the implementation of strategies that operate across the individual, relationship, community, and societal levels is needed, ^10,21,22 with the development and evaluation of community-level strategies representing a critical next step toward this end.

The field faces several challenges to expanding the evidence base around community-level strategies for SV prevention. First, current knowledge of community-level and societal-level risk factors for SV perpetration is very limited. Additional etiologic research identifying these risk and protective factors is critical and will guide the development of strategies for SV prevention that impact these factors, as well our ability to measure these factors as potential mediators or outcomes of existing prevention strategies. Unfortunately, the SV literature provides minimal guidance about promising factors at these levels; thus, it may be helpful to examine community-level and societal-level risk factors that have been established in other research areas that share commonalities with SV, such as youth violence or sexual health. ²³

A second, and related, challenge involves the lack of theoretical or empirical guidance in the SV literature for identification of promising programs, strategies, or policies that may impact SV behavior at the community level. Again, looking at successful approaches in other areas of health behavior may provide some direction. For example, in a recent article, Casey and Lindhorst ²¹ identified several factors associated with effective multilevel prevention strategies in other public health domains, such as HIV prevention and alcohol use, as a starting place for the SV field to move toward an ecologic model of prevention. One possible strategy that has already gained some traction in the field but has not been sufficiently evaluated is a community-level social norms approach that focuses on changing not just individual attitudes but also perceptions within a larger community that violent behavior is socially unacceptable or that prosocial behavior, such as actively intervening to prevent SV or SV-supportive behaviors, is expected and encouraged. ^{13 –15} In contrast, some policy and environmental strategies that have proven effective in other fields (e.g., condom availability for HIV prevention, tax policies for tobacco use prevention) may not be easily translated to SV prevention. As a result, theoretical approaches that consider what effective community-level practices, environmental change strategies, or policies (e.g., organizational/institutional policies, legislation) might look like for SV are needed. One approach could involve exploring other literature fields for promising approaches that might be more easily adapted or evaluated for SV outcomes and to identify existing policies or other community-level strategies that have yet to be evaluated but that may hold potential for SV prevention.

Third, once such policies or strategies are identified, challenges exist in terms of conducting the outcome evaluations necessary to establish effectiveness. Rigorous evaluation to establish effectiveness is a key step in the identification of prevention approaches that can be disseminated to communities and ensures that limited resources can be devoted to strategies that work. However, estimating the impact of community-level strategies on SV may be somewhat more difficult than for behaviors with more accessible and reliable records data, such as health or hospital records or crime data. As SV is significantly underreported and prosecuted at lower rates than other forms of violence, ²⁴ traditional crime report data may not capture changes in SV prevalence and incidence that result from prevention efforts. This is particularly problematic for evaluations of policy, environmental, or structural interventions, as the most rigorous methods for conducting such evaluations typically rely on administrative outcome data (such as hospital or crime data). These outcomes may need to be supplemented or replaced with self-report data, which may be costly or logistically challenging in larger communities or jurisdictions (e.g., neighborhoods, counties, states). Population-based surveillance systems that monitor SV prevalence rates provide valuable data for monitoring change over time, but data collection must be ongoing and available at a level (e.g., local, state, national) that maps on to the prevention strategy being implemented.

Rather than serving as a barrier, acknowledgment of these gaps and challenges should provide direction for future research and an opportunity for dialogue aimed at moving the field toward prevention strategies at the outer levels of the social ecology. More work is needed to expand our knowledge of community-level and societal-level risk and protective factors for SV. Potential approaches for evaluation need to be identified from the SV theoretical literature, as well as other fields of prevention. Efforts to create or identify new and innovative data collection methods and data sources for the reliable assessment of SV perpetration and victimization at the community and societal levels are also needed. Progress toward these goals will move us significantly closer to being able to identify community-level and societal-level strategies with the potential to maximize prevention resources and toward a significant reduction in the burden of SV on society.