We performed a facility-based cross-sectional study in Kicukiro District, Kigali, Rwanda, from February to April 2025. Eleven health centers providing HIV care were included.

The target population was HIV-discordant married or cohabiting couples (one partner HIV-positive, the other HIV-negative) enrolled in care at these centers. Based on district surveillance data, we estimated ~2,344 discordant couples in Kicukiro. Using Fisher's formula (assuming 50% violence prevalence, 95% confidence, 5% margin of error), the required sample was 384 couples.

Stratified systematic sampling was employed across eleven health centers providing HIV care in Kicukiro District. The sample was proportionally allocated based on the number of discordant couples registered at each facility. Eligible participants were HIV discordant couples (one HIV-positive, one HIV-negative partner) who were married or cohabiting and provided informed consent.

Trained interviewers administered a pre-tested structured questionnaire during face-to-face interviews with the HIV-positive partner (ensuring privacy). The questionnaire collected information on socio-demographic factors, relationship characteristics, and experiences of domestic violence in the past year (physical, sexual, and emotional). Emotional abuse was assessed by questions on insults, humiliation, and controlling behaviors. We also asked about behavioral factors (alcohol/substance use, jealousy, conflict resolution style) and contextual factors (reaction to HIV disclosure, perceived stigma, social support, access to care). Data collection took place in a private room at each facility.

The primary outcome was any domestic violence, categorized by type (physical, sexual, emotional). Key exposures included gender, age, education, employment, financial status, substance use, and contextual variables as above.

Data analysis was performed using Stata version 19.5 (StataCorp LLC, College Station, Texas, USA). The following analytical approaches were adopted:

Participant characteristics, including socio-demographic (age, gender, education level, employment status), economic factors (household income, financial difficulty), behavioral factors (alcohol and substance use), and contextual factors (reaction to HIV disclosure, stigma, social support), were summarized using frequencies, percentages, means, and standard deviations where appropriate. The prevalence of domestic violence was presented separately for physical, sexual, and emotional abuse, using percentages with 95% confidence intervals (CIs).

Chi-square tests were utilized to identify statistically significant associations between domestic violence types (physical, sexual, emotional) and each independent variable. Variables considered potentially significant were those with a p-value less than 0.05. These variables were retained for further analysis in the multivariate regression model to control for potential confounding factors.

Ethical approval was obtained from Mount Kigali University Ethics Committee and Kigali City authorities. Written informed consent was secured from all participants, with emphasis on confidentiality and voluntary participation. Referral mechanisms were established for participants requiring support services.

A total of 384 HIV discordant couples participated in the study (Table 1). The sample included 197 females (51.3%) and 187 males (48.7%). The largest age group was 35-44 years (34.4%), followed by 25-34 years (31.8%). Most participants had completed secondary education (41.9%) or primary education (32.3%), while 9.9% had no formal education. Nearly half (47.4%) were in informal employment, while 27.6% were unemployed and 25.0% in formal employment. Regarding financial status, 43.2% described their situation as somewhat difficult, and 23.2% as very difficult. Most participants (71.9%) were formally married, while 28.1% were cohabiting.

The study revealed a high prevalence of domestic violence among HIV discordant couples (Table 2). Overall, 42.7% of participants (164 out of 384) experienced domestic violence in the past year. When examining specific types of violence, physical violence was reported by 41.1% of participants (158 individuals: 13.5% experienced it often, 27.6% sometimes), sexual coercion by 34.1% (131 participants: 8.9% experienced it often, 25.3% sometimes), and emotional abuse by 52.3% (201 participants: 20.3% experienced it often, 32.0% sometimes). Additionally, 36.2% reported that their partner controlled their social interactions, and 34.9% reported partner control over mobile phone or internet use.

Regarding frequency patterns, 11.5% experienced violence frequently, 27.6% occasionally, 23.2% rarely, and 37.8% never experienced violence in the past year. The study also revealed that 42.4% had experienced partner violence in previous relationships, and 35.7% witnessed violence during childhood. Most participants (58.1%) reported having access to social support networks.

Table 3 presents the behavioral characteristics and contextual factors among study participants. Regarding alcohol consumption, 49.5% never consumed alcohol, 37.0% consumed occasionally, and 13.5% frequently. Substance use other than alcohol was reported by 16.7% of participants. For conflict resolution methods, 37.2% discussed issues calmly, 44.0% used shouting or arguing, and 18.8% resorted to physical confrontation. Decision-making in households was shared equally by 49.0% of couples, while 27.3% reported that their partner made most decisions and 23.7% made decisions themselves. Partner jealousy or accusations were experienced by 45.3% of participants, and 38.0% reported that their partner did not allow social activities. Religious beliefs were used in conflict resolution by 46.4% of participants.

Concerning HIV-related factors, 88.8% of participants had disclosed their HIV status to their partner. Among those who disclosed, 67.4% received supportive reactions, 18.5% faced hostile reactions, and 14.1% received indifferent responses. HIV-related stigma was experienced by 34.4% of participants, while 94.0% had access to HIV care services.

Bivariate analysis revealed significant associations between various socio-demographic factors and domestic violence experience (Table 4). Gender was strongly associated with violence, with women experiencing significantly higher rates (56%) compared to men (29%) (p<0.001). Age showed a progressive increase in violence prevalence, from 32% in the 18-24 age group to 61% among those ≥55 years (p=0.004). Education level was inversely associated with violence, with 53% of those with no formal education experiencing violence compared to 29% with higher education (p=0.034).

Employment status significantly influenced violence risk, with unemployed participants experiencing the highest rates (55%) compared to those in formal employment (34%) (p=0.014). Financial situation was a strong predictor, with those in very difficult financial situations reporting 63% violence prevalence versus 25% among those financially comfortable (p=0.002). Economic dependency patterns showed that participants whose partners were primary income providers experienced higher violence rates (56%) compared to those who earned equally (41%) or were self-reliant (38%) (p=0.016). Access to personal money was protective, with those having free access experiencing 30% violence compared to 56% among those with no access (p=0.001).

Table 5 demonstrates significant associations between behavioral characteristics and domestic violence. Alcohol consumption showed a dose-response relationship, with frequent drinkers experiencing 58% violence prevalence versus 38% among non-drinkers (p=0.021). Substance use other than alcohol was strongly associated with violence, with users reporting 62% prevalence versus 35% among non-users (p<0.001). Conflict resolution methods were critical predictors, with violence highest among those using physical confrontation (72%), followed by shouting/arguing (51%), versus calm discussion (25%) (p<0.001).

Decision-making patterns significantly influenced violence risk, with those in relationships where the partner made most decisions experiencing 52% violence compared to 37% when respondents made decisions themselves and 43% with shared decision-making (p=0.034). Partner jealousy or accusations were strongly predictive, with affected participants reporting 63% violence prevalence versus 29% without such behaviors (p<0.001). Control behaviors showed similar patterns, with social activity control associated with 61% versus 27% violence prevalence (p<0.001), and mobile/internet control linked to 59% versus 34% violence rates (p<0.001). Notably, religious coping appeared protective, with those using religious beliefs in conflict resolution experiencing lower violence rates (27%) compared to those who did not (54%) (p<0.001).

Table 6 reveals significant associations between HIV-related factors and domestic violence. Paradoxically, HIV status disclosure was associated with higher violence risk, with those who disclosed experiencing 56% violence prevalence compared to 42% among those who had not disclosed (p<0.001). Among those who disclosed, partner reactions were critical determinants of violence risk. Hostile reactions were associated with 68% violence prevalence, compared to 34% for supportive reactions and 45% for indifferent reactions (p<0.001).

HIV-related stigma was a strong predictor of violence, with affected participants reporting 60% violence prevalence versus 35% among those without stigma experiences (p<0.001). Access to HIV care services also influenced violence risk, with those having limited access experiencing higher rates (52%) compared to those with access (41%) (p<0.001). These findings highlight the complex interplay between HIV-related stressors and domestic violence in discordant relationships.

Ethical approval was obtained from the Mount Kigali University Ethics Committee and Kigali City authorities. Written informed consent was obtained from all participants.

Not applicable.

The datasets used during the current study are available from the corresponding author on reasonable request.

This research was conducted as part of a Master's thesis with no funding.

NU conceived and designed the study, collected and analyzed the data, and drafted the manuscript. MM and OR supervised the study design, methodology, and analysis, and critically reviewed the manuscript. JDH provided technical input and reviewed the final manuscript. All authors read and approved the final manuscript.