VR has been used for several mental health interventions. The underlying idea is that someone can stand and move in a computer-generated world, and feel immersed (Rothbaum et al., 2006; Botella et al., 2007; Fox et al., 2009; Baños et al., 2011; Bordnick et al., 2011; Repetto et al., 2013). That is, feel that they are in the real world while being somewhere else. VR treatment has been applied to tackle different psychiatric disorders, such as anxiety disorders, specific phobias, Post Traumatic Stress Disorder (PTSD), schizophrenia, Obsessive-Compulsive Disorder (OCD), eating disorders and autism, and has shown promising results (Turner and Casey, 2014; Valmaggia et al., 2016; Maples-Keller et al., 2017). VR as a treatment in mental healthcare settings has the ability to include dynamic yet controlled multisensory interactive situations, as well as register and analyze users’ behavioral responses. It therefore offers realistic situations, makes the treatment very personal through the adaptability of the VR-environment, motivates patients because it provides them with something new and different, and is a good tool for patients who have trouble with imagining or visualization (Riva, 2009; Maples-Keller et al., 2017). VR has the added benefit of facilitating learning, changing attitudes, and modifying behavior indirectly, as opposed to explicit interventions, where participants are directly given information and are aware of what is that they are learning. Thus, in this review we view VR interventions as implicit interventions.

VR has also been used in forensic settings. Forensic settings are defined, for the purpose of this paper, as correctional settings for people who have been accused of an offense. The characteristics of forensic mental healthcare patients (i.e., limited insight in their own health and diseases, not compliant to their treatment, heterogenous psychopathology and offending behavior, low levels of literacy, and low motivation for treatment) (Kip et al., 2020), make VR a potentially useful tool. Interactive, non-immersive, computerized case simulations have previously been explored in these settings and research has cautiously indicated that such programs may be beneficial (Arborelius et al., 2013; Kip et al., 2018; Sygel and Wallinius, 2021). However, immersive VR simulations can be an even bigger step forward for correctional rehabilitation (Ticknor, 2019). VR may facilitate treatment and reintegration of offenders into society through treatment of psychiatric disorders (e.g., using VR cognitive behavioral therapy), teaching necessary skills, providing feedback, and making virtual treatment groups possible (Sygel and Wallinius, 2021). Our aim with this paper is to see whether VR can also be an effective tool both in the prevention and treatment of IPV, within the perpetrators and the survivors. We want to find out whether the special benefits of VR produce a stronger effect than other types of interventions.

In the following pages we will review what has been done in terms of prevention and intervention of IPV in general. Regarding prevention, we will first review general interventions, followed by bystander interventions, and then delve into recent technological innovations in prevention. Shifting our focus to rehabilitation, we will explore interventions designed for both perpetrators and survivors separately, and end that section discussing technological advances in the treatment of IPV.

In order to evaluate which effective VR interventions exist for the prevention and rehabilitation of IPV, we established the search terms with the highest number of possible synonyms. The general search expression was (“Intimate partner violence (MeSH)” OR “Dating violence (MeSH)” OR “Spouse abuse (MeSH)” OR “Intimate partner abuse (MeSH)” OR “Domestic violence” OR “gender violence” OR “IPV”) AND (“Virtual Reality (MeSH)” OR “Virtual Reality, Educational (MeSH)” OR “Educational Virtual Realities (MeSH)” OR “Educational Virtual Reality (MeSH)” OR “Virtual Reality Immersion Therapy (MeSH)” OR “Virtual Reality Therapy (MeSH)” OR “Reality Therapies, Virtual (MeSH)” OR “Reality Therapy, Virtual (MeSH)” OR “Therapies, Virtual Reality (MeSH)” OR “Therapy, Virtual Reality (MeSH)” OR “Virtual Reality Therapies (MeSH)” OR “Virtual reality intervention” OR “VR intervention” OR “360 video”).

The search used several databases: PsycINFO by EBSCOHost, and Medline by PubMed. The specific syntaxes for the databases are in Supplementary Appendix S1.

Lastly, we reviewed the references in the selected studies in order to identify potential studies of interest that were missed during the database search, but no new ones were found.

Inclusion criteria for the current review were the following a) Participants must be over 14 years old; b) The study must be published in English, Spanish or Catalan; c) The study must have an experimental or quasi-experimental design d) The study must evaluate the effectiveness of VR interventions for the prevention or rehabilitation of IPV; e) The study must have been published after 2010, and f) the study must have used VR. Exclusion criteria included studies with participants under 14 years old, reviews, observations, single case studies, and studies with non-immersive videos. Location or sample size were not part of the exclusion criteria.

The selection process was carried out by the author of the review, doing a first selection based on the title and abstract of the studies and the final selection based on the full text of some studies and of those that did not give enough information in their title or abstract.

A data extraction protocol was designed for the selected studies. The registered information was the following: authors, publication year, research design, demographic characteristics of the target population (forensic or community, gender, age, and sexual orientation when available), number of participants, type of outcome, instruments that were used to analyze the main study variables, type of intervention (prevention, early intervention, or rehabilitation), conclusions, length of treatment with VR, length of control group intervention, and type of control group (waitlist, active, or delayed). Data extraction was carried out by the author.

The risk of bias of the included articles in the review was evaluated by the author of the review applying the checklists recommended by The Joanna Briggs Institute (JBI) for Randomized Controlled Trials (RCT) and Quasi-Experiments (Table 1 and Table 2).

A narrative summary of the results was carried out of the studies regarding the effectiveness of VR or other interactive interventions for the prevention and rehabilitation of IPV. The main effectiveness results and conclusions were shown in a table, and future research opportunities were discussed.

First, virtually experiencing a domestic violence situation from the perspective of the victim can result in an enhancement of fear recognition in female faces, both in forensic and in community population (Seinfeld et al., 2018; Seinfeld et al., 2020). Moreover, this type of experience is able to physically change brain areas that are related to socio-cognitive processing and to the perception of emotional expressions (Seinfeld et al., 2018; Seinfeld et al., 2020; De Borst et al., 2020). Thus, VR could be a useful tool to increase empathy in those who might have it underdeveloped, and therefore reduce intimate partner violence. Moreover, the effect of perspective taking and identification with the victim is more easily achieved if the participant directly embodies the victim from first-person perspective than if they are merely observing the scene as a third-person (Gonzalez-Liencres et al., 2020). However, third-person perspective can still have an effect as well, especially if participants are indirectly affected by the violence even if it is not directly targeted at them. In fact, when participants embody a child that is put in a situation of abuse and violence from the father avatar to the mother avatar, where they are standing next to the mother, there is an increase in emotion recognition of angry faces and physiological reactions (Seinfeld et al., 2023). Emotion recognition outcomes are important as they can increase the ability to empathize with the victim and therefore decrease perpetration. The effect that VR has on the increasing empathy and perspective taking is higher than simply reading about a situation and imagining it is happening to them (Ventura et al., 2021). VR has also been used to attempt to change implicit gender bias, with participants embodying a gender different than their own and experiencing sexual harassment, and has shown promising results (Wu and Chen, 2022).

While the positive effect of VR in perspective taking, emotion recognition and identification with the victim is clear, evidence is mixed regarding a decrease on actual aggressive behavior. A study by Klein Tuente et al. showed that self-reported or observed aggressive behavior did not significantly decrease with VR compared to a control group. Moreover, while VR does seem to decrease hostility, anger control, and non-planning impulsiveness, these effects are not maintained in time (Klein Tuente et al., 2020; Neyret et al., 2020) suggested that by allowing participants to take part in being a woman victim of sexual harassment by a group of men, and then being asked by a group of men to harm another woman, they could psychologically identify with the woman, become aware of the illegitimacy of the behavior of the perpetrators of sexual harassment, and thus avoid doing such behavior.

VR is a useful tool to train individuals to act more assertively in situations of sexual violence, reducing the risk of sexual victimization, psychological victimization, and psychological distress (Rowe et al., 2015). However, no significant reduction of risk was observed for physical violence or threats. Future studies could expand this finding and apply VR to reduce other kinds of violence.

Regarding the effect of VR in bystander behavior, participants in the VR group were more likely than the video group to report intentions to engage in indirect, non-confrontational interventions, such as “intervening” (i.e., causing a distraction or removing the target from the situation), or “approaching” (i.e., offering social support to the target after the incident) (Rawski et al., 2022). In fact, a recent review by Park and Kim (2023) concluded that traditional offline methods, VR, and web-based methods significantly improved bystander outcomes (i.e., awareness, definition, responsibility, bystander efficacy, gender norms/attitudes, intention to intervene, and action to intervene), while videos did not.

There are several implications to this review.

First, studies that focused on rehabilitation interventions in the forensic context have shown that VR could be a potential tool to reduce recidivism by increasing empathy within the forensic population (Seinfeld et al., 2018; Gonzalez-Liencres et al., 2020; Seinfeld et al., 2018; Seinfeld et al., 2023; Ventura et al., 2021; De Borst et al., 2020). However, we would need future longitudinal studies following the forensic population to see whether these interventions directly have an effect on recidivism. Furthermore, it is important to mention the difficulty of running interventions within rehabilitation programs within the forensic setting, which is one explanation for the short number of articles found. Participants in most rehabilitation interventions included in this review are forensic population, that are usually in prison and difficult to be motivated to do the intervention voluntarily. Therefore, those participants typically want to be part of the rehabilitation program in general, and more specifically do the VR intervention, to show an improvement in their behavior and rehabilitation and thus, to obtain prison benefits (e.g., permission to go out). As such, this population tends to rate high on social desirability, reducing the reliability of the studies. In treating this kind of population, indirect interventions such as the VR we have seen in this review, become even more advantageous (Sanz Fernández et al., 2018; Seinfeld et al., 2018; Johnston, 2021). Future research should include more VR interventions in the context of rehabilitation programs in the community setting to be able to compare its impact in the community vs. in a forensic setting.

Second, in this review we have seen how different usages of VR can be advantageous for specific population groups. Specifically, using VR as a prevention tool has proven to be effective with young adults (Rowe et al., 2015; Neyret et al., 2020; Seinfeld et al., 2020; Lee and Cha, 2021; Rawski et al., 2022). Institutions could benefit from using VR in educational settings in order to reduce psychological violence (increasing empathy), prevent victimization (assertiveness training), decrease sexual harassment, and increase bystander behavior.

Third, studies mentioned in this review have shown that VR can change implicit gender bias (Wu and Chen, 2022). This is a crucial result to design interventions that would tackle gender stereotypes and promote gender equality. More studies are needed that use VR to change implicit gender bias. It is important that those studies include both genders in order to be able to detect the impact of the intervention and possible gender differences. Similarly, only men were recruited as offenders when applying the intervention with forensic participants, and only women for studies with survivors. Future research should attempt to recruit both genders for studies with offenders and with survivors in order to assess any gender differences.

Fourth, and along the same lines, studies should aim at being more diverse, especially in terms of sexual orientation. Future research should take sexual orientation into account, include participants who are homosexual and bisexual, and study the impact of the intervention and possible sexual orientation differences.

Fifth, while quantitative studies inform us of the results in a general sense, future research should also include qualitative studies to be able to specifically see the impact of VR on an individual level, and thus have a more comprehensive idea of the effect of VR on IPV. Along the same lines, it is crucial to include behavioral changes measures to assess reduction in violent behavior (in terms of both aggression and victimization).

Lastly, both the context and the duration of the VR intervention are important factors to consider, as they might have an impact on the final results. In regards to the context, we would need to see whether the intervention mainly only uses VR, or if it is included in a longer/more comprehensive intervention in the context of a rehabilitation or prevention program. In terms of the duration of the VR intervention, we have to consider whether it is a one-off intervention lasting a few hours or whether it is a longer intervention, spanning multiple days. Future research should include longer interventions to obtain a higher effect, as well as tailor them to specific target groups, according to the results found in Klein Tuente et al. (2020).

A similar narrative review by Johnston et al. (2023) was recently published about this topic, focusing on the effectiveness of embodiment and perspective taking in general, and more specifically in IPV. Johnston et al. (2023) concluded that the integration of validated VR embodiment technology inspired by neuroscience could positively contribute to IPV rehabilitation processes. Our present study has built on the result of Johnston et al. (2023) by conducting a systematic review to analyze the studies that have used VR in the context of IPV. Both Johnston et al. (2023)’s study and the present study give a comprehensive understanding of the use of VR in contexts of IPV.

Overall, there are few interventions using VR to tackle the prevention or rehabilitation of IPV, and many still use more traditional approaches. The specific strategies and content used to reduce IPV is different in VR compared with more traditional IPV interventions. As mentioned previously, in most VR interventions the main focus was more on indirectly changing attitudes and enhancing emotional recognition by experiencing being a victim themselves, or part of an abusive scene, through embodiment (i.e., implicit interventions) (Seinfeld et al., 2018; Gonzalez-Liencres et al., 2020; Neyret et al., 2020; Seinfeld et al., 2023; Ventura et al., 2021; De Borst et al., 2020).

In contrast, as mentioned in the introduction, most of the traditional interventions or online programs, were focused on directly training participants on the identification of healthy relationships, and on the development of techniques of conflict resolution, empathy, anger management, making decisions, assertiveness, communication skills, psychoeducation about dating violence and emotion regulation strategies (i.e., explicit interventions) (Arango et al., 2014; Sygel et al., 2014; Levesque et al., 2016; Ellsberg et al., 2018; Kearney, 2018; Draucker et al., 2019; Peskin et al., 2019; Doss et al., 2020; Gibbs et al., 2020). Specifically, some of these studies that do not use VR are still using technology, showing several videos (mostly as part of a broader online or computer program) in order to model behavior, train perspective taking, or educate participants in what should not be done (Sygel et al., 2014; Levesque et al., 2016; Kearney, 2018; Draucker et al., 2019; Peskin et al., 2019; Doss et al., 2020). These technological interventions have overall proven more useful to tackle intimate partner violence than traditional non-technological educational interventions, although the effects have mostly been seen as a reduction of psychological violence, and not physical violence (Sygel et al., 2014; Levesque et al., 2016; Kearney, 2018; Draucker et al., 2019; Peskin et al., 2019; Doss et al., 2020).

However, although these programs have effect sizes similar to VR interventions (Sygel et al., 2014; Rowe et al., 2015; Levesque et al., 2016; Kearney, 2018; Seinfeld et al., 2018; Draucker et al., 2019; Peskin et al., 2019; Doss et al., 2020; Gonzalez-Liencres et al., 2020; Klein Tuente et al., 2020; De Borst et al., 2020; Seinfeld et al., 2020; Lee and Cha, 2021; Ventura et al., 2021; Rawski et al., 2022; Seinfeld et al., 2023), future research should evaluate whether the immersive aspect of VR, and more specifically the increased outcomes of perspective taking achieved with embodiment, has a significant added value in reducing aggressive behavior compared to those more traditional non-immersive video programs by comparing them in equal conditions.