In this review, we did not take a strict definition of primary prevention.  Specifically, studies were included as primary prevention study if the intervention targeted IPV, and did not select a sample of known victims or perpetrators.  Studies included may have delivered interventions universally to a population, and that population may have included some prior victims and perpetrators. Or, the studies included may have targeted high risk, or “selected” populations for intervention some of whom may have already been victims or perpetrators.

Electronic databases were searched for IPV prevention studies.  Two reviewers identified 19 articles published between 1993 and 2012 that were included in this review. Studies included (1) contained one or more interventions targeting physical or sexual partner violence perpetration or victimization, (2) used a experimental or quasi-experimental design study design that included a comparison or control group, and (3) measured at least one outcome relevant to IPV including behavior, knowledge, attitude, belief, or another related construct.

Of the 19 studies, 15 used experimental designs, the strongest design for inferring causation. All but two studies tested a single intervention against a control group.  One tested two interventions against a control group, and another tested a short versus long version of the same intervention.  All studies used some form of a curriculum-based intervention to effect IPV outcomes.  Curriculum approaches as IPV prevention change strategies included:  focusing on IPV knowledge, attitudes, and beliefs from a feminist and/or cognitive behavioral perspective; using social norms to change behavior; promoting help seeking and peer support; promoting the development of specific relationship skills; and, focusing on the legal and judicial aspects of IPV.  Several studies included important non-curriculum based activities (e.g., community activities, a microloan program), but no studies were designed to examine the different effects of curriculum vs. non-curriculum based activities.  About two thirds (n = 13) of the interventions were conducted in school settings, and the rest were conducted in community settings. There was large variation in sample size with samples ranging from 37 to 2310 participants.

Of the 19 studies, 9 were determined to be methodologically strong in most aspects: use of randomized designs, acceptable retention rates, sufficient follow-up assessments, and use of valid measures of IPV behavior. Four of the studies were conducted in school settings, and five were conducted in non-school settings. Of the five school-based studies, only one found unqualified positive results on IPV behavior. Over four years, the Safe Dates program was shown to reduce IPV perpetration (psychological abuse, mild physical abuse, and sexual abuse) and victimization (physical IPV).  The program was equally effective for boys and girls, for all race/ethnicities included, and for teens who had experienced IPV and those who had not.

Of the five non-school based studies, each showed some positive effect on IPV behaviors. The five studies included two community-based interventions with group curricula and non-curriculum based activities (one set in Limpopo Kenya), two interventions that worked with couples (one in groups, one one-on-one), and one family-based intervention in which parents and teens discussed dating violence.  Each of the five interventions found some reduction IPV following the intervention.

This review found several programs that were effective in preventing IPV.  Community-based programs were particularly effective in this review.  Although many research questions remain regarding prevention programming, it is not too early to consider implementing some programs broadly. Prevention activities have traditionally been underutilized compared to programs for identified perpetrators and victims. Because prevention is generally cost-effective, programming is badly needed to prevent IPV before it begins.

There are several areas of need for future work of IPV prevention studies. First, although several programs were found that affected IPV behavior, no studies were replicated. Second, several of the effective programs included multiple components (curriculum plus community activities) but no analyses were reported that determined which components accounted for the positive study findings.  Third, future research will need to examine whether IPV prevention can be delivered with prevention programs that targeted other risk behaviors that emerge in adolescence such as risky sexual behavior, substance use, and peer violence. Last, if prevention programs will be implemented broadly, implementation and dissemination research is needed to understand how best to implement those programs with fidelity to maintain program effectiveness.