‘Ghya Bharaari Ekatra’: study protocol for a cluster randomized trial to evaluate a couples-based intervention for the primary prevention of intimate partner violence in India

Abstract

Background

Intimate partner violence (IPV) is experienced by one-fourth of women in India and associated with poor mental, physical, sexual and reproductive health. Ghya Bharaari Ekatra (GBE) is an evidence-based, theory-driven, culturally tailored, dyadic intervention, developed and piloted for the primary prevention of IPV among newly-married, low-income couples residing in urban slums in Maharashtra, India. GBE is delivered over six weekly sessions by a team of lay peer educators to groups of five couples and makes use of engaging culturally-tailored delivery methods to challenge norms and build knowledge and skills to address key IPV determinants. This manuscript outlines the protocol for the quantitative evaluation of GBE’s total impact on IPV, mental health, and reproductive autonomy, with consideration of mediating and moderating factors.

Methods

We will conduct a cluster randomized trial among 850 newly married dyads (500 assigned to GBE and 350 controls) in two districts in Maharashtra, India to quantify the impact of the GBE intervention compared to the ethical standard of care (information on IPV and mental and reproductive/sexual health resources) in preventing IPV experience (primary outcome) at 12 months. Secondary outcomes include reproductive autonomy among the newly married women and mental health among both members of the dyad at 12 months. We will also conduct mediation and moderation analyses for the efficacy of the GBE intervention.

Discussion

Study findings will provide insight into GBE’s potential to prevent IPV, its mechanism of effect, and contexts that may modify its effects, thereby facilitating future scale-up across India and similar resource-limited, high-burden settings globally upon study completion.

Trial registration

Registered at ClinicalTrials.gov on 6th February 2025: NCT06811922. Registered at Clinical Trials Registry-India (CTRI) on 3rd December 2024 : REF/2024/12/095719.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-026-26803-9.

Background

Intimate partner violence (IPV) is experienced by over one in four women globally [1] and associated with poor mental, physical, sexual, and reproductive health [2–10]. A systematic review of longitudinal studies conducted globally found that women reporting IPV experience had two-fold higher odds of incident depression and significantly higher odds of attempted suicide [11]. IPV has also been associated with greater anxiety and post-traumatic stress disorder (PTSD) [12, 13]. Negative sexual and reproductive health outcomes of IPV include greater sexual risk-taking, sexually transmitted infections, restricted reproductive autonomy, unplanned pregnancy, and induced abortion [9, 10, 14]. Thus, the multifaceted deleterious health effects of IPV underscore the need for urgent, effective IPV prevention as both a human rights and public health strategy.

Historically, IPV interventions have focused on IPV support [15] and trauma treatment [16], leaving a critical gap in primary prevention [17]. Primary prevention, or intervention designed to prevent the first occurrence of IPV, is particularly important in resource-limited settings given the higher burden of IPV [1], associated healthcare and lost productivity costs [4, 18], and saturation of secondary and tertiary prevention resources (i.e., mental health, shelter, legal) [19, 20]. To date, there exist few evidence-based primary IPV prevention interventions globally, with existing interventions being largely developed in high-income settings or Sub-Saharan Africa [21]. Primary prevention in high-income countries typically begins in schools and focuses on dating relationships [21, 22]. However, such timing may be unsuitable for settings like South Asia where social taboos associated with discussing intimacy and sexual relations in schools and colleges [23–25] and elevated school dropout rates among girls of low socioeconomic status (SES) persist and hinder early prevention [26–28].

We previously developed and piloted a community-based, theoretically-rooted [29], dyadic intervention Ghya Bharaari Ekatra (Marathi for “Take a Flight Together”; GBE) for the primary prevention of IPV among newly married, low-income couples in India [28, 30]. Importantly, GBE times intervention during early marriage, when it is culturally appropriate [28, 31–33]. In a pilot study of 40 couples married < one year [30], GBE demonstrated high acceptance (with 85% of participants retained across all six intervention sessions), feasibility, safety, and preliminary efficacy in preventing IPV and improving mental health in the female partner at 3 months. We now seek to examine the longitudinal impact of GBE in preventing IPV (primary outcome) at 12 months, assess whether GBE also enhances downstream (secondary) outcomes, mental health and reproductive autonomy, and explore its mechanism of change.

Rationale for the proposed study

While the pilot study suggests GBE was effective in preventing IPV and enhancing mental health in women [30], larger scale testing is needed to determine whether GBE has sustained effects in preventing IPV and enhancing mental health in women, and whether it also has efficacy in enhancing reproductive autonomy in women and mental health in men. We expect GBE to impact these secondary outcomes as much of GBE’s content intentionally addresses determinants of mental health (i.e., self-esteem, resilience, recognition of social support) and reproductive autonomy (i.e., reproductive/sexual health knowledge, sexual communication, sexual relationship equity, and economic empowerment) as potential pathways to IPV [28]. The proposed mediation analyses will shed critical light on the mechanisms of GBE’s action by testing potential mediators suggested by Couples Interdependence Theory, (CIT, the theory that informed GBE’s development), as well as those suggested by the formative work that identified IPV correlates in the Pune community and informed GBE’s components [26, 28, 34]. The moderation analysis will enable the identification of subgroups in whom GBE works well and those in whom modification is necessary.

Study objectives and hypotheses

We aim to quantify the impact of the GBE intervention compared to the ethical standard of care (information on IPV and mental and reproductive/sexual health resources) in (1) preventing IPV experience among newly married women at 12 months (primary outcome), and (2) reproductive autonomy among newly married women and mental health among both members of the dyad at 12 months (secondary outcomes). We hypothesize that participation in GBE will lead to greater (1) prevention of IPV in women, (2) reproductive autonomy in women, and (3) mental health in men and women at 12 months, relative to controls (Fig. 1). Further, we aim to investigate mechanisms underlying GBE’s effectiveness, as well as potential heterogeneity of effects by participant characteristics.

Fig. 1.

Conceptual Framework.M=male; F=female; *using Table 1 potential moderators; CIT=Couples Interdependence Theory

Methods

Study Design

We will employ a multi-site parallel design cluster randomized trial (Fig. 2) to test the impact of the GBE intervention compared to the ethical standard of care (information on IPV and mental and reproductive/sexual health resources) in preventing IPV experience (primary outcome) and enhancing reproductive autonomy among newly married women (secondary outcome) and mental health among both members of the dyad (secondary outcome) at 12 months. Eligible couple dyads will be randomly assigned to the intervention (n = 500 dyads) or control arm (n = 350 dyads); (total n = 1,700 participants). Dyads assigned to the GBE arm will be sequentially assigned into a group composed of 5 newly married (married ≤ 1 year) couples. Intervention groups will receive the 6-session GBE intervention (over 6 weeks) plus ethical standard of care (information about IPV resources) while the control groups will receive the ethical standard of care (information about IPV resources alone). GBE will be delivered in 5 implementation waves (each comprised of 100 intervention dyads (20 intervention groups) and 70 control dyads) over the course of 1.25 years (as opposed to simultaneous implementation across all clusters) to facilitate inter-site coordination and assure standardization and quality of peer educator training and GBE delivery.

Fig. 2.

Overall Trial Design. a Eligible dyads will be randomly assigned within each geographic strata (slum community) of Mumbai and Pune to the intervention or control arm. Dyads assigned to GBE will be sequentially grouped into intervention groups of five dyads. b The implementation wave will be repeated over time for a total of five implementation waves (170 dyads x 5 waves = 850 dyads or 1,700 participants)

Primary and secondary outcomes will be assessed at the individual and dyad level, controlling for group cluster. Baseline, 6-, and 12-month follow-up surveys will be conducted during each wave with dyadic members (independently) in the intervention and control groups. Survey tools are being developed using the Fig. 1 framework and Table 1 variables, translated into Marathi and Hindi, and piloted and revised based on feedback. They will be made publicly available on Open Science Framework.

Table 1.

Study Measures. GBE=Ghya Bharaari Ekatra; Mos=months; F=female; M=male; psych=psychological; NFHS=National Family Health Survey; IFVCS=Indian Family Violence and Control Scale; CIT=Couples Interdependence Theory; std=standard;

Variable	Level of Variable	GBE Participant	Time Assessed (mos.)	Assessment Tool
Primary Predictor Variable
Receipt of GBE versus control	Intervention, control	F, M	0
Primary and Secondary Outcomes
Physical, sexual, and/or psych IPV experience (1o)	Yes/no	F	0, 6, 12	NFHS Modified Conflict Tactics Scale-2 [35] IFVCS Psych Abuse Subscale [32]
Mental Health (2o)	Total score	F, M	0, 6, 12	General Health Questionnaire-12 [36]
Reproductive autonomy (2o)	Total score	F	0, 6, 12	Reproductive Autonomy Scale [14, 37]
Potential Mediators
Perception re: mutual joint health benefit of violence-free relationship (CIT transformation of motivation)	Important for health of self, spouse, both, or neither	F, M	0, 6, 12	Single item derived from CIT “transformation of motivation” [29]
Joint communication and decision-making regarding behavior change needed for violence-free relationship (CIT communal coping)	Total number of IPV protective behaviors (GBE components) the couple has discussed and developed a shared action plan to adopt	F, M	0, 6, 12	Single item derived from CIT “communal coping” [29]
Execution of cooperative action plan for violence-free relationship (CIT behavior change)	Total number of IPV protective behaviors the couple has actively worked to modify using their shared action plan	F, M	0,6,12	Single item derived from CIT “behavior change.” [29] Protective behaviors assessed are those GBE was designed to change (dedicated relationship time, promotion of self and partner, conflict resolution, communication, sexual communication, goal setting/implementation, redefining IPV) [26, 34]
Self-esteem	Not at all, little, somewhat, great extent	M	0, 6, 12	Perceived attainment of gender ideal [34]
Relationship prioritization	Total time (hours)	M	0, 6, 12	Time dyad spends together alone/week
Perceived social support	Total score	M	0, 6, 12	Multidimensional Scale of Perceived Social Support [38, 39]
Resilience	Total score	M	0, 6, 12	Connor-Davidson Resilience Scale-10 [40]
IPV conceptualization	Total Score	F, M	0, 6, 12	IFVCS [26, 32, 34]
IPV acceptance	Total Score	F, M	0, 6, 12	NFHS Attitudes Toward Wife Beating Scale [33]
Conflict negotiation and communication skills	Total Score	F, M	0, 6, 12	Conflict Tactics Scale-2 Negotiation Subscale [35]
Sexual/reproductive health knowledge and sexual communication skills	Total Score	F, M	0, 6, 12	Knowledge assessment of Session 5 sexual and reproductive health information; Sexual Communication Self Efficacy Scale [41]
Potential Moderators
Income	Income quintiles	F, M	0
Education	≤ 7th, 8-10th, > 10th std	F, M	0
Employment	Yes/No; hours	F, M	0
Husband alcohol use	Total score	F, M	0	Alcohol Use Disorders Identification Test [42]
Baseline mental health	Total score	F, M	0	General Health Questionnaire-12 [36]
Baseline psych IPV & control	Total score	F	0	IFVCS Psych Abuse and Control Subscales [32]
Duration of premarital acquaintance	Total months	F, M	0
Family type	Joint, nuclear	F, M	0
Perceived family support	Very little, somewhat, great extent	F, M	0	Perceived family support item [34]
Acceptance of IPV	Total score	F, M	0	NFHS Attitudes Toward Wife Beating Scale [33]

Second, to assess mechanisms of effect of the GBE intervention on IPV reductions in women using a mediation analysis. Specifically, we will test whether intervention effects are mediated by enhanced self-esteem, relationship prioritization, perceived social support, and resilience in the male partner; reconceptualization of IPV, challenging of norms of IPV acceptance, enhanced conflict negation and communication, sexual/reproductive health knowledge, and/or sexual communication in both dyadic members at 12 months. The mediation analysis will contribute mechanistic understanding of how GBE prevents IPV and identify “core components” as well as components that can be modified to meet an adopting agency’s contextual needs. We hypothesize that GBE will prevent IPV through the following theory-informed pathways: (1) enhancing dyadic perception of the mutual benefits of a violence-free healthy relationship (CIT “transformation of motivation”; focus of GBE Sessions 1 and 6); (2) enabling vision of a shared action plan for, and (3) dyadic adoption of behaviors that we previously identified to be protective of a healthy, supportive, violence free relationship (CIT “communal coping” and “behavior change”; focus of GBE Sessions 1–6). We additionally hypothesize that GBE will prevent IPV through the empirically-derived mediators listed in Fig. 1/Table 1. Lastly, GBE’s effect on IPV may be mediated through peer learning, collective challenging of social norms around IPV, and social support within the intervention group.

Third, we will investigate whether intervention effects are moderated by individual, relationship, familial and social norm factors (i.e., socio-economic status, duration of premarital acquaintance, joint/nuclear family type, IPV acceptance). The moderation analysis is guided by the hypothesis that the effect of GBE on IPV will be moderated by factors that interact with intervention mechanisms (Table 1). Specifically, we expect attenuated GBE effects among participants who have less time or control over their ability to internalize and/or practice GBE content outside of the sessions. We expect this attenuation to occur in dyads with (1) lower vs. higher SES (due to competing importance of income-generating activities), (2) joint vs. nuclear families (due to less control over time and more competing household responsibilities), and (3) lower vs. higher family support (due to family being less supportive of their spending time alone). Similarly, GBE effects may be attenuated among participants with psychosocial conditions that diminish motivation for change +/- impair executive function. We hypothesize lower intervention effects in dyads with (1) poor vs. strong baseline mental health, (2) those where the husband reports higher vs. lower alcohol use, (3) longer vs. shorter premarital acquaintance (due to less relationship plasticity), (4) higher vs. lower IPV acceptance, and (5) higher vs. lower baseline relationship psychological IPV and control. Moderation analyses will thus identify subgroups and contexts where GBE may need more tailoring for optimal effectiveness.

Study setting, target population, and eligibility criteria

The study will be conducted in slum communities in Mumbai and Pune Cities, the two largest metropolises in the western state of Maharashtra. In the most recent census 42% (5.2 million) and 22% (690,545) of each respective city’s populations resided in slums [43, 44]. National estimates demonstrate IPV prevalence to be 2.5x higher in the lowest wealth quintile relative to the highest [33]. We will use the eligibility criteria developed and employed during GBE pilot studies: couples wherein both members are age ≥ 18 years, married ≤ 1 year, in their first marriage, cohabiting in a slum, and fluent in Marathi and/or Hindi. As GBE was designed for primary IPV prevention, couples in which the female member screens positive for physical and/or sexual IPV (using the modified Conflict Tactics Scale (CTS-2)) at baseline will be excluded (and referred to support services) as will those in the third trimester of pregnancy (as women traditionally return to their natal home during this perinatal period).

Sampling, recruitment, enrollment, consent, and retention

We will first map and identify slum communities in Mumbai and Pune and re-establish partnerships with community-based organizations (CBOs) with whom ICMR-NITVAR and National Institute for Research in Reproductive and Child Health (NIRRCH) have previously collaborated in recruitment and implementation of clinical trials [45]. Next, in the first quarter of each of the five implementation waves, we will use the multi-staged community sensitization and recruitment process used successfully in the GBE pilot study that led to high rates of enrollment and 85% retention: (1) study staff and CBOs will contact key community leaders to inform them about GBE, (2) CBOs and key community leaders will lead community sensitization meetings within each slum community to bring awareness of GBE’s aim of “fostering healthy relationships,” build community buy-in, and prevent potential community non-cooperation that could result from erroneous community speculation of GBE’s intent (i.e., to preach family planning, publicly disclose familial violence), (3) CBOs and key leaders will identify potential peer educators to be intervention facilitators and potentially-eligible couple participants, (4) study staff will meet with families of the couples at their homes to seek initial permission (as per cultural norms), and then (5) with the couples individually at a private venue of convenience to them to obtain written informed consent, confirm eligibility, and conduct baseline surveys with each member individually (Fig. 3). We will recruit 50 intervention and 35 control couples per site per implementation wave, with the aforementioned steps being followed across all five waves. To facilitate retention of participants across the intervention and follow-up period, all participants will (1) provide detailed locator information including other personal contacts that we may contact if we are unable to reach the participant between visits, and (2) receive up to three reminders of upcoming study visits via phone calls and/or text messages followed by one home visit if necessary between each study visit.

Fig. 3.

Participant Screening and Enrollment Process. Potentially eligible participants will be identified in partnership with key community leaders and agencies and undergo initial demographic screening. Dyads considered eligible through the demographic screening will be asked to provide informed consent and undergo the baseline survey. The second part of participant eligibility screening (assessing for physical and sexual intimate partner violence in the female partner) is nested in the baseline survey. Eligible couples who screen eligible through the two-step screening process are then randomized to the intervention or control condition

Intervention pre-implementation and implementation teams

A Translational Advisory Board, comprised of leaders from governmental and community-based organizations who bring real-world program implementation, will be established at study onset. They will be regularly updated on study intent, progress, and findings to enable adoption of GBE upon study completion if the intervention is deemed effective. GBE delivery will employ a pre-implementation (community leaders, CBOs) and implementation team (peer educators, government medical officers). Community leaders (local government/political figures) and CBOs will facilitate community sensitization and entry and help recruit potential participants and peer educators. Peer educators (one male, one female per group), who will facilitate GBE sessions, are lay community people who themselves are married. They will be recruited from local agencies (i.e., anganwadis, CBOs, education departments) and selected based on strength of oratory, group facilitation, and critical thinking skills and community involvement. Session 5 “Sexual Communication and the Sexual Relationship” is the only session delivered in gender-concordant groups and jointly facilitated with a medical officer. While it may seem counterintuitive to separate couples for the session, this strategy was noted as imperative for GBE acceptance during the formative work [46]. Medical officers will be recruited from government Reproductive Health and Family Planning Centers.

The GBE Intervention

GBE will be delivered to groups of 5 newly married couples in weekly two-hour sessions over a six-week period. GBE is delivered at a group level to foster peer-learning, richness of discussions challenging IPV social norms, community resilience (through group building and social support), and cost-effectiveness [28]. GBE will be facilitated by a male/female pair of trained peer educators and held at a community-based venue (i.e., school, community hall, CBO, anganwadi (government child care/preventive health center)), in the slum communities in which the couples reside. The intervention is highly participatory, making use of reflections, discussions, role plays, games, films, and competitions. Session 1, “You, Me, and Us: Spending Meaningful Time Together,” aims to increase quality time spent together and employs a series of self, couple, and group reflections on (1) the benefits of a marital relationship, (2) the current amount of the time couple spends together, barriers and facilitators to spending quality time with one another, and planning of strategies to increase quality time. Session 2, “I’m a Champion: I Can’t Be Broken and I Don’t Accept Defeat,” aims to increase self-esteem and resilience and employs: (1) facilitated group exercises in which participants brainstorm available community and human resources and methods to address common adversities (i.e., job loss, health, relationship difficulties), and (2) a guided interview between the two members of a couple wherein each gets to know the other on a deeper level (i.e., background, hobbies, sources of pride and worry, strengths, weaknesses) and culminates with recognition of the need to see each other as a unit, supporting growth of one another. Session 3, “Building Communication and Conflict Management Skills,” aims to improve communication and conflict management skills and employs two short films that describe the initial hopes, dreams, and expectations of a newly married bride and groom and subsequent challenges they face adjusting to married life in the context of a larger joint, low-income family. This is followed by a peer-led discussion that culminates with the group brainstorming strategies for handling each of the situations. Session 4, “Empowerment of the Couple: Planning Ahead,” aims to empower both members of the couple through fostering goal setting and goal implementation skills. It employs (1) couples-based reflections on their dreams, followed by peer-assisted planning to achieve goals, and (2) role-plays demonstrating skills to prepare for and participate in a job interview. Session 5, “Sexual Communication and the Sexual Relationship,” aims to improve sexual communication and the sexual relationship, and is the only session co-led by peer educators and government medical officers and delivered in gender-concordant groups (to ensure acceptance and participation as suggested during GBE formative work). It employs (1) medical officer-delivered lectures on sexual and reproductive health, (2) a quiz with facilitated discussion that addresses commonly held reproductive and sexual health myths, (3) a “Chutes and Ladders” game adapted to methods for fostering romance, and (4) a lecture regarding methods for sexual communication with a follow-up question and answer period. Session 6, “A Lens into Domestic Violence,” aims to expand participant definitions of behaviors constituting IPV and to challenge subjective norms of IPV occurrence. It employs (1) a facilitated discussion of the different forms and effects of IPV on all members of the family, (2) an exercise wherein each participant ranks examples of violence by severity to initiate a discussion about definitions of violence, highlight differences in conceptualizations of violence, and respectfully challenge one another to expand definitions of IPV and commit to a life of non-violence. The session ends with a video of scrolling participant photos, presentation of the couple with a framed photo of themselves, announcement of the winners of the GBE competitions, and final reflections.

Statistical power and sample size

We powered the study for the primary outcome of IPV measured by the incidence of physical/sexual IPV in women. Psychological IPV is much more commonly reported [26, 34], so we used the lesser reported physical/sexual IPV as the basis for a conservative power calculation. We assumed 13.5% prevalence of IPV in the control group (from national data of individuals married < 2 years) [33] and estimated a relative risk for the treatment effect at 0.5 (based on other similar interventions: DIL MIL, SASA!, EMAP, IMAGE) [47–50]. We calculated a minimal sample size of 313 dyads per arm, using the POWER command in STATA v15 (College Station, TX), and 80% power with α = 0.05. GBE is delivered in a group setting of 5 dyads; thus, we applied an intra-class correlation coefficient of 0.1 (based on similar intervention: SASA! ) to accommodate clustering of treatment effects within groups (e.g., based on group dynamics) [47]. This resulted in a design effect of 1.4, leading to an estimate of 440 dyads (88 groups). To allow for possible attrition (~ 10%), we will include 500 dyads (100 groups) in the intervention group and 350 dyads in the control group (1,700 total participants).

Randomization

Eligible dyads will be sequentially enrolled and randomly assigned to the GBE intervention (n = 500 dyads) or control arm (n = 350 dyads); total n = 1,700 participants within each geographic strata (slum community) of Mumbai and Pune (equal numbers in both cities) using a computer-generated randomization scheme. Unbalanced allocation is due to clustering in the intervention arm. Dyads assigned to GBE will be sequentially grouped into intervention groups of five dyads. Randomization at the level of the dyad (as opposed to the group/cluster) was selected because of the efficiency of design and the risk of contamination of the control arm being low (risk of interaction and sharing of materials between participants in a large slum community is low; further, GBE requires active participation for impact). The 10:7 random assignment was selected to optimize sample size while accommodating statistical clustering related to peer-group intervention delivery in the intervention arm. There is no peer-group clustering in controls as the ethical standard of control is delivered individually.

Blinding

The participants, facilitators, and study team members involved in data collection, entry, and data management will not be blinded as it is not practically feasible. The data manager will prepare and provide a clean analytic dataset masking intervention assignment to two external analysts. The external analysts will independently conduct analyses for the primary and secondary outcomes and remain masked to intervention assignment until concurrence on model specification and results. Additional analyses examining intervention effects will be conducted after analyses that examine the impact of the intervention on our primary and secondary outcomes have been completed. Purely observational analyses that do not require information on intervention groups may be completed before unmasking occurs.

Intervention facilitation training: safeguarding fidelity

Peer educators will undergo a weeklong training at respective implementation centers in Mumbai and Pune by the study staff and the implementation leadership team. The manualized peer educator training, developed at the time of GBE piloting and updated with lessons learned from the pilot [28], includes training on IPV and gender equality, GBE intent, manualized content, intervention facilitation methods, methods for establishing rapport, privacy, and confidentiality, and safety protocols. The second half of each training day is dedicated for peer educators to practice intervention delivery and receive real-time feedback from the study team. Additionally, during the six-week intervention delivery period, peer educators will have one-hour weekly paired retraining meetings to practice delivery of the specific session and have session-specific questions answered. These meetings will also enable space for peer educators to notify study staff of concerns about the safety of particular participants and/or discuss emotional trauma they themselves are experiencing as facilitators. Before delivering Session 5, medical officers will receive a two-hour training by the research team on session content.

Control/comparator

There is presently no established standard of care for IPV prevention in India. Across all clusters, the control group will receive from the research field staff an “ethical standard of care” designed in consultation with WHO IPV research guidelines [51]. It consists of a list of IPV and mental and sexual/reproductive health resources concealed in a booklet of other services. Booklets will be provided once to all female participants at time of consent regardless of whether they disclose IPV and field staff will provide a verbal overview of the booklet content to the participant. To isolate the effect of GBE specifically, the participants assigned to the intervention arm (across all clusters) will also receive the booklet provided to the control arm participants. Fidelity to the control condition (i.e., provision and review of the booklet at time of consent) will be documented in the checklist by the field staff at time of consent.

Study outcomes

The primary outcome, IPV experience, is defined as physical, sexual, and/or psychological IPV in women at 12 months. Specifically, it is ≥ 1 act of physical and/or sexual IPV and/or a score of ≥ 24 on the IFVCS Psychological Abuse Subscale. Physical and sexual IPV will be assessed using the modified Conflict Tactics Scale-2 (α = 0.81–0.93) [52] used in the Indian National Family Health Survey and other demographic health surveys worldwide. This measure was selected to enable comparisons of efficacy of GBE with other global IPV interventions. Psychological IPV will be measured using the culturally-tailored IFVCS Psychological Abuse Subscale (α = 0.81) [32], given the strong influence of culture and context on how psychological IPV manifests and is conceptualized [53–55]. Per WHO recommendations [51], we do not assess IPV perpetration by men out of safety concern for the linked female, but rather use IPV experience in the female partner as a surrogate. Secondary outcomes: change in (1) mental health among women and men at 12 months will be measured using the General Health Questionnaire-12 (GHQ-12, α = 0.90) [36], and (2) reproductive autonomy among women at 12 months, will be measured using the Reproductive Autonomy Scale (α = 0.86) [14, 37].

Process evaluation measures

We will examine key elements proposed by Saunders et al. and Moore et al. [56, 57]: fidelity, participation rate, recruitment and context. To assess fidelity, trained study staff will directly observe and document, using methods similar to the GBE pilot [28, 30], the extent to which GBE activities were implemented as intended during actual delivery. They will explore challenges faced with delivering GBE activities through exit interviews with peer educators after Session 6. To assess the participation rate, they will track session attendance and contact participants who missed sessions to assess barriers to participation. To assess recruitment, they will record numbers of subjects approached, consented, assessed for eligibility, barriers to participation among those not enrolled, those completing baseline assessment, and randomized. To assess context, they will assess, through structured observation, responsiveness of community leaders, community members during sensitization meetings, and family members at time of initial permission. Exit interviews with the implementation team after Session 6 will be examined for contextual barriers and facilitators encountered in delivering GBE.

Data management/data quality/data sharing

Each participant will be provided with a study identification number (SID). Only the consent and locator forms will have identifying information. All other forms will be labeled with the SID and be absent of personal identifying information. All documents will be locked in a filing cabinet at their respective research institute, with consent and locator forms kept separately. All computerized documents will be password-protected. The study team will check data for completeness before leaving the interview site and upon data entry and run a weekly quality check of the entered data. Datasets housed at NITVAR and NIRRCH will be merged upon completion of Aim 1 data collection and undergo a third quality check by NITVAR Data Manager. Data will be converted to a SAS dataset for analysis.

De-identified dataset(s) that can be shared, with some restrictions to be outlined in the Restricted Use Data Agreement to preserve participant confidentiality and safety, will be deposited in the Data Sharing for Demographic Research (DSDR).

Statistical analysis

The study protocol and statistical analysis plan will be available at Open Science Framework. Initial analyses will assess the robustness of our measures and sample. We will assess the psychometric properties of all measures [58]. To evaluate and mitigate the impact of attrition on internal validity, we will test for differences in baseline and follow-up variables between drop-outs and retainees. Bivariate tests appropriate to the distribution of the variables will be applied (e.g., Chi-square; Wilcoxon, t-tests). Procedures outlined by Hansen [59] will be used to assess threats to internal and external validity due to attrition. Statistical adjustments, as described by Lepkowski [60], will be used to correct for effects of attrition as needed. Standard multiple imputation techniques as outlined by Little and Rubin [61], will be used to account for missing data if needed.

We will qualitatively evaluate imbalances in participant characteristics between the control and intervention groups. Characteristics that display potentially meaningful differences at baseline will be considered in the analyses of the primary outcome. Trial effects will be assessed through intention-to-treat analysis of incident IPV in women (primary outcome), reproductive health in women, and mental health in men and women separately (secondary outcomes). We will calculate risk difference between intervention and control groups at endline using generalized estimating equations, accounting for clustering (at the group level) using robust standard errors. Secondary analysis will include an as-treated analysis, in which the fidelity of group delivery and participation rates will be treated as independent variables, to understand if delivery of GBE as intended yielded similar outcomes to those with incomplete fidelity and/or participation, as we have done elsewhere [62]. Process measures will be examined alongside the outcome measures to contextualize and explain findings.

To identify mechanisms of the intervention’s effect, we will conduct mediation analyses [63]. Mediators will be lagged relative to the outcome (that is, mediator data will be taken from t-1 for an outcome measured at time t). We will first apply the traditional product-moment approach [64] to estimate the effect of the GBE directly on IPV experience (in women), as well as the indirect effect of GBE through each mediator (Table 1). This analysis will also allow us to quantify the effect of GBE on the mediator. We will next apply a counterfactual approach to mediation [65], which allows for simultaneous examination of multiple mediators, flexible outcomes, and mediator specification (binary, categorical) [66, 67]. The clustered design will be addressed using robust variance estimators.

To assess potential moderation of intervention effects according to individual and environmental factors, we will estimate GBE effects using the models specified in the primary analysis with the additional inclusion of an interaction term between intervention assignment and each hypothesized moderating factor separately. The potential moderators listed in Table 1 will be evaluated. A statistically significant interaction term (α = 0.05) will be treated as evidence of moderation. Subgroup effects will be reported alongside the statistical significance of the interaction term.

In exploratory analyses, we will evaluate potential moderated effects of statistically significant mechanisms linking GBE and IPV using established methods compatible with causal mediation analysis [68, 69]. We will consider both socioeconomic and psychosocial moderators.

Ethics and safety

The safety protocol for the present study mirrors the protocol used in the pilot study as no safety concerns/events were noted in the pilot study. First, all research staff will be required to undergo ethics training, CITI certification, and a week-long specialized IPV training, as recommended by the WHO Ethical and Safety Recommendations for Research on Domestic Violence against Women. To address potential emotional trauma incurred by the study staff, all will have the opportunity to participate in weekly debriefing meetings and offered facilitated referral to support services if necessary.

Privacy/confidentiality

We will take extra measures to ensure privacy during the conduct of the surveys and delivery of the intervention, given both will take place in often-crowded slum communities. First, study staff will identify secure community-based venues for the conduct of interviews and the intervention. Additionally, during the consent process, participants will be notified that if there is a concern for breach of privacy, the study staff will either pause the interview or switch to a “safer” topic (i.e. nutrition). During intervention delivery, research staff will actively monitor surroundings for bystanders and redirect them elsewhere, explaining the session is to foster healthy relationships and that only enrolled participants can partake. The investigator team and research staff will be the only study team members with access to participant identifying information. Surveys will be de-identified and labeled with a study identification number and kept in a locked filing cabinet, separate from the consent and locator forms. All computerized study documents will be password-protected and stored on a secure server.

Safety

To facilitate the safety of participants and research staff, the following measures will be taken: (1) All female study participants will be provided with the contact information for local IPV and mental health and sexual/reproductive health resources and information regarding the India Domestic Violence Act regardless of their disclosure of IPV at the time of provision of informed consent; (2) Emphasis is placed on pre-enrollment community sensitization efforts to dispel myths regarding the intervention and reduce associated harm. Parents-in-law or others in the participants’ household are notified of the study’s intent “to promote healthy relationships.” Local police and key leaders are formally notified of planned survey and intervention conduct prior to the research team’s community entry, so they can provide support if necessary; (3) During delivery of the GBE intervention, participants will never be directly questioned regarding IPV in their relationship; (4) To enhance detection of IPV, a “Women’s Day” will be hosted after the third and sixth GBE session wherein female participants have the option to attend confidentially to convey safety concerns to the study team, female participants will be questioned regarding IPV and threats to their safety resulting from the intervention in private at six- and twelve-month follow-up survey, and all female participants will be provided with study staff contact information to alert study staff if a violence or safety concern arises; (5) if it comes to the attention of the study staff that a participant is experiencing IPV or in danger, study staff will meet independently with the participant to ensure she is aware of available community support services and confidentially refer her to local mental health services, spiritual counseling, financial and legal aid, medical care, and/or emergency shelter, and offer her individual-level follow-up outside of the intervention. The remainder of the intervention sessions will follow as planned to reduce the chance of the husband learning of the participant’s disclosure of IPV. These safety procedures will be explained to the female participant in private during the initial informed consent process and the contact information of both spouses will be updated at each session. Male participants will also be told during the initial informed consent process that study staff may contact either him or his spouse in between visits and that is per protocol.

All detected IPV and other adverse safety events will be recorded (in accord with data privacy protocols) and reported to the ICMR-NITVAR and ICMR-NIRRCH Ethics Committee, the Emory IRB, and Data Safety Monitoring Board (DSMB). The DSMB will be comprised of individuals with expertise in gender-based violence intervention research and clinical trials implementation who are independent from the study funder. The DSMB will meet annually at minimum to review (1) adherence to the study protocol, (2) adequacy of compliance with goals for recruitment and retention, (3) data quality, completeness, and timeliness of data collected, (4) interim and cumulative data for evidence of study-related adverse events, and (5) factors that might affect study validity or compromise confidentiality of the data. Accordingly, the DSMB will make recommendations concerning the continuation, modification, or termination of the proposed trial. Changes to the protocol will be made per their recommendations, including seizure of the study if necessary. Additionally, on an annual basis, the study and study findings will be presented to the ICMR-NITVAR and ICMR-NIRRH Ethics Committees, the DSMB, and Emory IRB. Each will review and evaluate study safety and effectiveness data, and make recommendations regarding whether the study should be continued, modified, or terminated. Significant protocol modifications will be communicated to study staff, co-investigators, DSMB and Translational Advisory Board members.

Timeline

Since the onset of the study (September 2024), we have completed necessary US and India trial registrations, received study ethical approvals at all participating institutes, completed US and India government regulatory approvals, and developed and translated the survey tools. The study received foreign clearance on April 22, 2025, so our team began the process of establishing subcontracts between Emory and ICMR-NITVAR and NIRRCH so that the India field staff could be hired and trained and trial recruitment could begin in August 2025. On May 1st, 2025, NIH issued a new policy prohibiting foreign subawards from being nested under parent grants with a plan to initiate a new award structure directly to foreign organizations before September 2025 [70]. Assuming the new NIH award structure is established as proposed [70], field staff hiring/training would take place in Fall 2025 and the GBE trial would begin January 2026. Trial implementation and data collection for Aims 1 and 2 would be completed within three years of initiation, and analyses and dissemination of results to the scientific community and general public within a year thereafter.

Discussion

The proposed trial will be the first to evaluate the impact of a contextually-tailored couples-based primary IPV prevention intervention on incident IPV, mental health, and reproductive autonomy among newly-married dyads in India. The mediation and moderation analyses will shed light on its mechanism of effect and contexts that may modify its effects.

The proposed study is innovative because it will (1) test the impact of the first published couples-based intervention developed in an LMIC for the primary prevention of IPV. (2) While most primary IPV prevention targets teen dating in schools [21, 22], GBE innovatively prevents marital IPV through work with young married couples in community-based settings; (3) The intervention design makes use of lay community members as intervention facilitators, was developed by the community, and was delivered in the very slum communities in which participants reside. GBE employs evidence-based, engaging delivery strategies including couples-based reflections, culturally-inspired games, and films depicting commonplace conflicts encountered by newly married couples (with scripts inspired by formative research); and (4) The efficient study design enables simultaneous evaluation of intervention impact and process outcomes, elucidates GBE’s mechanism of change, and identifies subgroups likely to benefit most and those in whom GBE will need adaptation for greater effect, thereby enabling rapid adoption. If deemed effective, future steps would include integration into the India national health program or non-governmental sector to enable scaling and sustainability.

Potential limitations, operational issues and alternative approaches

If we do encounter difficulty meeting recruitment targets, we will expand enrollment to a third site in urban West India or extend the recruitment timeline. If conflicts arise between couples, peer educators will first try to facilitate discussion and resolution, and then, if needed, refer to outside counseling. We acknowledge that by our excluding from the study couples wherein both members of the dyad do not express willingness to participate (i.e., as is required for a couples-level intervention), we may select for couples in low-conflict relationships or relationships where there may be low family support for participation. Lastly, if the new NIH foreign sub-agreement policy proves prohibitive to conduct this study meaningfully in India, we will seek alternative funding sources.

Abbreviations

CIT

Couples Interdependence Theory

GBE

Ghya Bharaari Ekatra

IPV

intimate partner violence

HIV

human immunodeficiency virus

ICMR

Indian Council of Medical Research

LMICs

low- and middle-income countries

NITVAR

National Institutes of Translational Virology and AIDS Research

PSTD

post-traumatic stress disorder

SES

socioeconomic status

US

United States

WHO

World Health Organization

YLD

years lived with disability

Authors’ contributions

A.S. Kalokhe contributed to study conceptualization, methodology, funding acquisition, supervision, and original draft preparation. S. Begum contributed to methodology, supervision, and writing (review and editing). S. Bangar contributed to methodology and writing (review and editing). M. Bhise contributed to methodology and writing (review and editing). M.C. Freeman contributed to study conceptualization, methodology and writing (review and editing). T. Katendra contributed to methodology and writing (review and editing). G. Rahane contributed to methodology and writing (review and editing). U. Ravi contributed to writing (reviewing and editing). S.A. Patel contributed to study conceptualization, methodology, and writing (review and editing). R. Prusty contributed to methodology and writing (reviewing and editing). J.M. Sales contributed to study conceptualization, methodology, and writing (review and editing). S. Wirtz contributed to methodology, supervision, and writing (review and editing). S. Sahay contributed to study conceptualization, methodology, funding acquisition, supervision, and writing (review and editing).

Funding

The study is funded by the National Institutes of Health National Institute of Child Health and Human Development (NIH/NICHD) grant #1R01HD114862. Program Official: Ronna Popkin (email: ronna.popkin@nih.gov). NIH/NICHD has provided financial support for the project, but has/will not influence the study design, execution, or interpretation of results.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

This study received ethical approval from Emory University Institutional Review Board (study #00007022) on 19 December 2023, the Indian Council of Medical Research National Institute of Translational Virology and AIDS Research (study #EC/2024-04) on 10th July 2024, and the Indian Council of Medical Research National Institute for Research in Reproductive and Child Health (study #600/2024) on 1st January 2025. All study participants will provide informed consent prior to participating in study activities.

Competing interests

The authors declare no competing interests.

Consent for publication

Not applicable.