Partner Abuse Uncommon Among HIV-Serodiscordant Couples in Africa

A. Kott

First published online:

Only a minority of HIV-serodiscordant heterosexual couples who have ongoing access to couples' HIV counseling experience intimate partner abuse, according to a multinational study from Africa.

Only a minority of HIV-serodiscordant heterosexual couples who have ongoing access to couples' HIV counseling experience intimate partner abuse, according to a multinational study from Africa.1 Although most couples did not experience verbal or physical abuse, the odds of being abused were slightly elevated among HIV-positive partners (odds ratios, 1.3 for women and 2.2 for men), and women—regardless of their infection status—were more likely than men to suffer abuse. Overall, 18% of HIV-positive and 14% of HIV-negative women reported experiencing intimate partner abuse during the two-year study, compared with 7% and 5%, respectively, of infected and uninfected men.

The study was a secondary analysis of data from a seven-country clinical trial, conducted in East and Southern Africa in 2004–2008, that examined the potential benefits of using an antiviral drug to prevent HIV transmission within serodiscordant couples co-infected with herpes simplex virus type 2. Participants had to be at least 18 years old, sexually active and planning on staying in their current relationship for the duration of the study. Most couples had learned that they were serodiscordant in the month prior to enrollment. HIV-positive partners attended monthly visits at an HIV clinic, and uninfected partners had quarterly visits; however, partners were encouraged to attend each other's clinic visits so that they could receive ongoing counseling to address the struggles of living as a serodiscordant couple.

At enrollment and every three months thereafter, participants were asked—separately and privately—whether they had been abused verbally (yelled at, called names or threatened) or physically (hit, slapped or forced to have sex) by their partner during the previous three months. The researchers used Kaplan-Meier analysis to determine the cumulative proportion of individuals who had reported intimate partner abuse at any point in the study, and univariate and multivariate generalized estimating equations to assess potential correlates of intimate partner abuse, including a wide range of social, demographic and behavioral factors assessed at initial and follow-up visits.

In 67% of the 3,408 serodiscordant couples, the woman was the partner with HIV. Two-thirds of couples were from East Africa; more than half were married and cohabiting. At enrollment, couples in which the woman was infected had been together for an average of five years, while those in which the man was infected had been together for seven. Most respondents had at least one biological child, eight or more years of education, and little or no income. The median number of times couples had had sex in the month before enrollment was four; one-quarter of respondents had had unprotected sex during that time.

At enrollment, 3% of infected women and an identical proportion of uninfected women reported having been a victim of intimate partner abuse in the previous three months, compared with 2% of infected and 1% of uninfected men. About 36–37% of the incidents in which HIV-positive partners experienced abuse were probably or definitely related to the couples' learning that they were serodiscordant, in the view of study staff, compared with 13–15% of incidents involving abuse of uninfected partners.

Participants attended more than 39,000 follow-up visits. Intimate partner abuse was reported at 3% of infected women's visits, 2% of uninfected women's visits, and 1% each of infected and uninfected men's visits. After two years, the proportions of infected and uninfected women who had ever reported intimate partner abuse (18% and 14%, respectively) were more than twice those among men (7% and 5%). For both sexes, having reported abuse at enrollment was associated with reporting it at follow-up (odds ratios, 5.8 for women and 7.9 for men). However, the incidence of violence declined over the course of the study.

Overall, one or both partners reported intimate partner abuse in 16% of couples. Reports of such abuse came from the woman in 69% of these couples, from the man in 13% and from both in 18%. In 67% of the couples in which the man and the woman each reported intimate partner abuse at some point in the study, both partners reported abuse during the same visit at least once, indicating that the abuse was mutual.

The majority of reports involved verbal abuse, either alone or in tandem with physical violence. A larger proportion of women than men reported physical violence (11% vs. 3%), or a combination of verbal abuse and physical violence (48% vs. 20%). Abuse was not associated with seroconversion during the trial.

In multivariate analyses, a woman's odds of intimate partner abuse were elevated if she was HIV-positive (adjusted odds ratio, 1.3), unmarried (1.4) or living with her partner (2.6), or had had unprotected sex with her partner in the past month (1.9). Odds were elevated for a man who was HIV-positive (2.2) or had had sex with other partners in the past month (2.6); in addition, they were higher among men from East Africa than among those from Southern Africa (2.2), and greater among men aged 35–44 than among those aged 45 or older (1.9). For both sexes, the odds of having experienced abuse in the past three months were lower during the second year of study (0.3–0.7) than at enrollment.

The authors acknowledge several study limitations. The sample may not fully represent HIV-serodiscordant couples, especially those who are unaware of their serodiscordant status or are not in couples' counseling. Investigators did not collect data on alcohol use, which has been associated with intimate partner violence in other settings. Because of cultural norms about masculinity, men may have underreported being a victim of intimate partner abuse.

Although couples' awareness of serodiscordance may help them avoid transmitting HIV to the uninfected partner, the authors note that "becoming aware of being in a serodiscordant relationship itself could lead to intimate partner violence." However, the decrease in partner abuse observed during the trial could be "an indicator of the effectiveness of ongoing counseling in addressing couples' need for support." The authors suggest that future studies examine the extent to which HIV counseling can minimize the occurrence of intimate partner abuse.—A. Kott


1. Were E et al., A prospective study of frequency and correlates of intimate partner violence among African heterosexual HIV serodiscordant couples, AIDS, 2011, 25(16):2009–2018.