Sexual Relationship Power and Periconception HIV‑Risk Behavior Among HIV‑Infected Men in Serodifferent Relationships

Abstract

Gender norms affect HIV risk within serodifferent partnerships. We assessed how the sexual relationship power described by men living with HIV (MLWH) associates with periconception HIV-transmission risk behavior. Quantitative surveys were conducted with 82 MLWH reporting a recent pregnancy with an HIV-negative or unknown-serostatus partner in KwaZulu-Natal, South Africa. Surveys assessed decision-making dominance (DMD) using the Pulerwitz et al. sexual relationship power scale; partnership characteristics; and HIV-risk behaviors. Multivariable logistic regression models evaluated associations between DMD score and HIV-risk behaviors. Higher male decision-making dominance was associated with non-disclosure of HIV-serostatus to pregnancy partner (aRR 2.00, 95% CI 1.52, 2.64), not knowing partner’s HIV-serostatus (aRR 1.64, 95% CI 1.27, 2.13), condomless sex since pregnancy (aRR 1.92, 95% CI 1.08, 3.43), and concurrent relationships (aRR 1.50, 95% CI 1.20, 1.88). Efforts to minimize periconception HIV-risk behavior must address gender norms and power inequities.

Introduction

Social determinants are the primary drivers of the HIV epidemic. Among the social determinants linked to the HIV epidemic are societal gender norms, the socially constructed roles and behaviors that a society considers appropriate for men and women [1–4]. Across many societies that ascribe to traditional gender roles, men control decisions within the partnership, particularly with regard to issues related to sex [5, 6]. This gendered nature of power dynamics within sexual relationships has profound implications for HIV prevention efforts [7], particularly in stable heterosexual, serodifferent relationships.

South Africa has the largest population of people living with HIV (PLWH) in the world [8]. Approximately 30% of PLWH in South Africa are in stable serodifferent relationships, or in a sexual partnership with a partner at high risk for acquiring HIV [9, 10]. This risk is even higher for HIV-uninfected women in serodifferent partnerships trying to conceive [11, 12]. Increased efforts are focused on reducing HIV-risk behaviors within these partnerships in an attempt to reduce both sexual and perinatal transmission of the virus [13–17]. However, risk behavior modification must be addressed within the context of structural factors that drive periconception HIV-risk behavior [18].

In South Africa, the masculine ideology, particularly that of men who are racially marginalized and socioeconomically disenfranchised, is often expressed through the exertion of power over women [19]. Norms of masculinity emphasize sexual behaviors that are thought to convey toughness, daring, and assertiveness such as having a greater number of partners and condomless sex [19–22]. Furthermore, since sex is viewed as a domain that men lead, women may be expected to cede control of sexual decisions and submit to their male partners’ wishes [22].

Numerous studies have examined the association between gender inequity and HIV-risk behavior and outcomes [23–29]. Many show that power imbalances within sexual relationships are associated with poor reproductive and sexual health outcomes for women; women who report having low control in their relationships are more likely to report higher-risk sexual behaviors like inconsistent condom use [30]. Negotiating condom use remains an important strategy for HIV harm reduction even amongst serodifferent couples who are trying to conceive, with safer conception guidelines recommending couples wait to initiate condomless sex until the HIV-infected partner achieves viral load suppression and, if possible, limiting condomless sex to peak fertility around the female partner’s predicted ovulatory days [14, 31, 32].

Most studies on gender norms and HIV risk behavior have examined power from the women’s perspective [23–28] and few examined the dynamics of sexual behavior decision-making in serodifferent couples [24, 33, 34]. Significantly fewer studies, however, have examined how the sexual relationship power of men is associated with HIV-risk behavior [5, 35]. This is despite the fact that men have been increasingly more recognized as underrepresented and inadequately engaged in the HIV services and have been labeled one of the “blind spots” in the HIV response [36–38].

We are not aware of work to examine the sexual relationship power of HIV-infected men who report recent (partner) pregnancy with a serodifferent partner or of men in KwaZulu-Natal, arguably the epicenter of the HIV epidemic in South Africa. Understanding partnership dynamics within serodifferent relationships where the man is living with HIV has profound implications not only for sexual but also perinatal transmission rates [39]. In this manuscript, we assess the level of sexual relationship power described by men reporting a recent (partner) pregnancy with an HIV-uninfected or unknown-serostatus partner to evaluate how sexual relationship power affects their periconception HIV transmission risk behavior. Understanding sexual relationship power within this population may inform interventions to promote uptake of HIV-harm reductive sexual health strategies and behaviors.

Methods

Study setting and participant recruitment

This analysis was part of a larger study conducted between 2011 and 2012 in a large suburban area of eThekwini District in KwaZulu-Natal, South Africa to evaluate periconception HIV risk behavior among men and women reporting a recent pregnancy with a serodifferent partner. Though our larger study recruited women and men attending antenatal care (ANC) and antiretroviral (ARV) clinics within a large public hospital; in this manuscript, we only analyzed data from our male sample given our study aim to assess the sexual relationship power and HIV risk behavior described by men living with HIV who are in serodifferent relationships. Comparative analysis of both male and female periconception risk behavior can be found in separate publications [40, 41].

Screening was conducted in English and/or IsiZulu by trained, fluent research assistants. Eligible men were 18 years or older, living with HIV, reporting a partner pregnancy in the past 3 years with a pregnancy partner who was HIV-negative or unknown-serostatus at the time of conception. All HIV-serostatus data (self and partner) were based on self-report by participants enrolled in the study; to meet eligibility, all participants had to know their HIV serostatus prior to the referent pregnancy.

Measures and Procedures

Those who were eligible and provided informed consent were asked to complete a face-to-face structured interview administered by a trained research assistant. Administration of the quantitative questionnaire took approximately one hour; upon completion of interviews, participants were reimbursed 70 ZAR (valued at 8–9 USD at time of survey) for their time.

The study questionnaire assessed basic sociodemographic variables, reproductive history, HIV history, and partnership characteristics, in addition to items reported elsewhere [40, 42]. The primary outcome was power dynamics within their relationship with their most recent pregnancy partner, measured using the 8-item decision-making dominance (DMD) subscale of the sexual relationship power scale (SRPS) [43]. The DMD subscale assessed decision-making dynamics within the partnership, asking participants to refer to their referent pregnancy partner and report “who usually has more say” about couple activities such as going out with friends, when to have sex, types of sexual acts performed, and condom use. Participants were asked to designate whether their partner (“Your Partner” = DMD score 1), both of them equally (“Both of you equally” = DMD score 2), or they (“You” = DMD score 3) make those decisions in their relationship (Table 1). We used the South African adaptation of the SRPS [25], which has been used in prior studies examining sexual relationship power in South African men [35, 43–46]. Sociodemographic variables and power dynamics were measured at the time of survey.

Table 1.

Decision-making dominance subscale question items

	Your partner (1)	Both of you equally (2)	You (3)
Who usually has more say about whose friends you go out with?	1	2	3
Who usually has more say about whether you have sex?	1	2	3
Who usually has more say about what you do together?	1	2	3
Who usually has more say about how often you see one another?	1	2	3
Who usually has more say about when you talk about serious things?	1	2	3
In general who do you think has more power in your relationship?	1	2	3
Who usually has more say about whether you use condoms?	1	2	3
Who usually has more say about what types of sexual acts you do?	1	2	3

A DMD score was calculated by generating a mean of the responses to the eight questions related to DMD and rescaled to a range of 1 to 4 according to Pulerwitz et al. [43]. Rescaled DMD scores were categorized into low (score 1–2.430), medium (score 2.431–2.820), and high (2.821–4) levels of sexual relationship power. Higher DMD scores indicate that the participant perceives himself to have more dominance in making decisions within their partnership.

HIV-risk behaviors were defined as behaviors that influenced risk (increased or decreased) of HIV transmission to the uninfected partner. HIV-risk behaviors of interest included: condom use before and after partner pregnancy, number of concurrent sexual partners in the past year, anal sex practices, and self-reported ARV adherence (i.e., able to take ARV “all”, “most”, “some”, or “none of time” over the past 30 days). Disclosure of HIV-serostatus and knowledge of partner’s HIV-serostatus were also included as HIV-risk behaviors given that disclosure is often an important first step to initiating many HIV preventative strategies and has been associated with greater linkage to care and reduced risk of HIV transmission [47, 48]. We also assessed sexual and reproductive health communication behavior within partnerships in asking how often (“often”, “sometimes”, “rarely”, or “never”) participants had discussions with their partners about sex, HIV, having children, or contraception.

HIV-risk behavior responses were dichotomized into “risky” or “not risky” categories. Risky behavior was defined as: not having disclosed HIV-serostatus to partner; having no knowledge of partner’s HIV-serostatus; reporting sex without condoms “all of the time,” “most of the time,” or “some of the time” since pregnancy; endorsing at least one concurrent sexual partner; and/or reporting “never” or “rarely” having discussions with partner about sex, protection from HIV, having children, and/or contraception. These data were collected prior to HPTN052 and public health messaging about the prevention benefit of HIV-RNA suppression [49]. We also did not collect HIV-RNA data to further categorize the actual risk associated with self-reported condomless sex. Therefore, we categorized condomless sex as “risky” because this behavior would have been deemed risky according to the counseling these men received at the time, regardless of their HIV-RNA suppression status. Anal sex practices and safer conception behaviors were not included in analyses given infrequent reports of behaviors in our sample. In addition, ART adherence was not considered as men on ART uniformly endorsed high adherence.

In addition to these general HIV prevention behaviors, participants were asked about their use of other safer conception strategies that lower the risk of HIV transmission while allowing for pregnancy (i.e., timed sex to peak ovulatory window, sperm washing and artificial insemination); findings on our larger sample’s reported safer conception behavior are described elsewhere [41].

Analysis

Descriptive statistics summarized demographic information, partner characteristics, HIV-risk behaviors, and DMD scores. Unadjusted linear regression models were used to evaluate the association between DMD score and each of the following covariates of interest, including: education, employment status, income, age difference between participant and partner, type of partnership.

Poisson regression with robust standard errors [50] assessed the relationship between the DMD score and HIV-risk behaviors. The DMD scaled score was the primary predictor and was modeled as a continuous variable and the dichotomized HIV-risk behaviors were the primary outcome. Separate unadjusted Poisson regression models were used to examine associations between DMD score and HIV-risk behaviors of interest as estimated by relative risk ratios (RRs) and their corresponding 95% confidence intervals (CI). Multivariable Poisson regression models assessed association between DMD and HIV-risk behavior adjusted for partner age and education of the participant. Questionnaire data was recorded using REDCap (Research Electronic Data Capture); analyses were conducted using SAS version 9.4 (SAS Institute Cary, N.C., USA).

Ethics

We obtained ethical approval from the Institutional Review Board at Partners Healthcare (Boston, USA) and the Human Research Ethics Committee at the University of the Witwatersrand (Johannesburg, South Africa). We also obtained permission to conduct the research from local study site authorities.

Results

A total of 83 MLWH who reported a recent pregnancy with an HIV-uninfected or unknown-serostatus partner completed our study questionnaire. One study participant who reported contradictory partner-serostatus was removed from the analysis; thus, our current analysis is based on findings from 82 men (Table 2). Participants were a median age of 34 years (range 22, 44) and the majority identified as black South Africans (n = 81, 99%). The greatest proportion (n = 44, 54%) listed secondary school as the highest education achieved, and 74% (n = 61) were employed. Most described their referent pregnancy partner as a casual partner (n = 53, 65%), and 94% (n = 75) reported that their relationship with their pregnancy partner was ongoing. At the time of the survey, 54% of reported that their partner was currently pregnant. Participants whose pregnancy partners had already delivered their baby reported that the age of their youngest child was a median of 1-year-old at the time of the survey. The majority did not know their recent pregnancy partner’s serostatus (n = 57, 70%).

Table 2.

Demographic and partnership characteristics of men living with HIV and who have HIV serodifferent partners with recent pregnancy (n = 82)

Characteristic	No. (%) of participantsa
Median age	34 (range 22, 44)
Black South African	81 (99%)
Education
Primary and some secondary	27 (33%)
Completed secondary	44 (54%)
Post-secondary	11 (13%)
Employment
Not employed	13 (16%)
Full-time employed	44 (54%)
Part-time employed	13 (16%)
Self employed	4 (5%)
Monthly incomeb
1000–4999 ZAR	31 (38%)
> 5000 ZAR	32 (39%)
Head of household
Respondent	29 (36%)
Adult relative (i.e., parent, aunt)	47 (58%)
Other	5 (6%)
Median number of living children	2 (range 1, 10)
Pregnancy partner typec
Wife	5 (6%)
Main partner	22 (27%)
Casual partner	53 (65%)
One-time sexual encounter	2 (2%)
Age of pregnancy partner	28 (range 17, 46)
Living with partner	15 (18%)
Partner HIV-serostatus
Negative serostatus	25 (30%)
Serostatus unknown	57 (70%)
Timing of referent pregnancy
Current	44 (54%)
If not current, age of youngest child in years	1 (IQR 0.6, 2)
Relationship ongoing (N = 80 men responded to this item)	75 (94%)

^aMissing data was not detailed in the table

^bAt the time of survey, 1 USD = 7–8 ZAR

^cParticipants selected ‘All that apply’ and could choose multiple responses (e.g., married and have a girlfriend). Long-term relationship was defined as main partner for at least one year or living with partner. Girlfriend was defined as main partner for less than one year, not living together. Casual partner was defined as “Kwapheni,” a local term that corresponds to a casual partner who is concurrent to a main partner

HIV Transmission Risk Behavior

Only 27% (n = 22) of men reported disclosing their HIV-serostatus to their partner prior to the referent pregnancy. At the time of the survey, 33% of men (n = 27) reported that their partner knew their status. Most men (n = 75, 91%) were on ART at the time of the survey and self-reported excellent adherence (99% reported being able to take their medication “all of the time”). The majority of men indicated that they had concurrent sexual partners at time of survey (n = 62, 76%). Nearly 30% of participants reported having condomless sex since pregnancy “some” or “all of the time.” Anal sex practice was rarely reported (n = 2). Discussions with partner on how to protect from HIV rarely or never occurred for 65% (n = 53) of our sample (Fig. 1).

Fig. 1.

HIV risk behavior of men living with HIV and reporting HIV serodifferent partners with recent pregnancy (n = 82), missing items not included

*Men reporting HIV-positive serostatus pregnancy partners were excluded from our study

Decision-Making Dominance Score

The median DMD score of our sample was 3.06 (IQR 2.88, 4.00), corresponding to high sexual relationship power. Seventy-seven percent (n = 63) of participants were in the “high” level of power category while 20% (n = 16) were in the “medium” level of power and 4% (n = 3) scored as having “low” level of power in their pregnancy partnership. With regard to particular responses within the DMD, participants generally indicated that decisions in their relationship were primarily determined by either themselves or both themselves and their partners equally (Fig. 2). The majority (n = 61, 74%) of men indicated they had the most power in the relationship and that the decision to have sex was either determined by themselves (n = 40, 49%) or agreed upon by both partners equally (n = 38, 46%); however, they primarily made the decisions around condom use (n = 46, 56%).

Fig. 2.

Distribution of responses to DMD scale

Sexual Relationship Power and HIV Transmission Risk Behavior

Factors associated with DMD score were assessed in a linear regression model; however, only type of partnership was found to be significantly associated. Men in casual partnerships had higher DMD scores than those in stable partnerships (median DMD score 3.41 vs. 2.91 respectively, p ≤ 0.01). There was not a statistically significant association between DMD score and participant education status, employment status, and income (data not shown).

In the Poisson model, higher DMD scores were independently associated with non-disclosure of HIV-serostatus, not knowing partner’s HIV status, engaging in condomless sex since pregnancy, and having concurrent sexual relationships (Table 2). A one-point increase in DMD score was associated with twice the risk of reporting that the pregnancy partner did not know the participant’s HIV status (aRR 2.00, 95% CI 1.52, 2.64), 64% increased risk of the participant not knowing the pregnancy partner’s HIV status (aRR 1.64, 95% CI 1.27, 2.13), 92% increased risk of engaging in condomless sex since pregnancy (aRR 1.92, 95% CI 1.08, 3.43), and 50% increased risk of having concurrent relationships (aRR 1.50, 95% CI 1.20, 1.88). Higher DMD scores were also associated with poor reproductive health communication. Men with a one point higher DMD score were over three times more likely to rarely or never talk about sex with their partner, greater than two times more likely to rarely or never discuss having children or contraception with their partner, and approximately twice as likely to rarely or never talk to their partner about HIV prevention (Table 3).

Table 3.

Decision-making dominance and HIV-risk behavior

Associations between higher DMD score and HIV-risk behaviors	Adjusted* RR (95% CI) per point increase on DMD scale	p value
Non-disclosure of HIV status to partner	2.00 (1.52, 2.64)	< 0.01
Not knowing pregnancy partner’s HIV status	1.64 (1.27, 2.13)	< 0.01
Condomless sex since pregnancy	1.92 (1.08, 3.43)	0.03
Concurrent sexual partners	1.50 (1.20, 1.88)	< 0.01
Rarely/never discuss sex with partner	3.64 (2.28, 5.81)	< 0.01
Rarely/never discuss HIV prevention with partner	1.98 (1.47, 2.67)	< 0.01
Rarely/never discuss having children with partner	2.54 (1.69, 3.81)	< 0.01
Rarely/never discuss contraception with partner	2.72 (1.81, 4.10)	< 0.01

^*Adjusted for type of partner

Discussion

Our study found that men living with HIV and reporting a recent pregnancy with an HIV-serodifferent partner had high sexual relationship power, with 74% of men reporting they had the most power in their relationship. High decision-making dominance of men within their partnerships was significantly associated with HIV-risk behaviors such as non-disclosure of HIV-serostatus, not knowing partner’s HIV status, engaging in condomless sex, having concurrent sexual relationships, and poor reproductive health communication (Table 2). Only 33% of men in our sample had disclosed their HIV-serostatus to their partner at the time of survey, and over one third endorsed having condomless sex after their partner became pregnant.

Our findings align with those of studies that assess sexual decision-making dominance within South African couples [33, 34] as well as those studies that have examined power from women’s perspective [23–28]. In a study of sexual decision-making amongst serodifferent couples in Uganda, couples wherin a single partner dominated sexual decision-making had significantly lower odds of condom use compared to couples with shared decision-making [24]. In a 2004 study that sampled South African women presenting for antenatal care, Dunkle et al. found that women who reported male dominance on the SRPS were more likely to be infected with HIV. Our findings are also consistent with studies that measure sexual decision-making dominance and HIV-risk behavior from the male perspective [35, 51, 52]. Shai et al. examined relationship dominance among rural South African men of unknown HIV status using a different gender attitudes and partner control scale and found that men who reported having more equitable relationships were more consistent condom users [29].

Men’s dominance in sexual decision-making within heterosexual partnerships is theorized to relate to the practice of hegemonic masculinity, or the culturally-idealized social roles and behaviors prescribed to men in a society [53, 54]. Masculinity is constructed through the exertion of power or dominance over other men and women in society through various means such as through wealth, social status, or physical strength (i.e., sexual behavior, violence) [55]. Black South Africans have faced a long history of economic marginalization. Until a few decades ago, black men were legally restricted to low-paying employment, could not vote, could only live in rural land reserves or overcrowded urban townships, and could not own land [55, 56]. To date, structural factors such as economic inequality, poverty, and intermittent work still limit marginalized men’s access to economic and social power structures [19, 20]. These structural factors disempower men and drive them to turn to sexuality and the exertion of power over women as a path to achieving social standing and having a sense of power in their lives [57–60].

Despite strong evidence that men dominate reproductive decision making, engagement of men in safer conception counselling remains a challenge [36]. Most periconception HIV-risk reduction interventions to date have primarily engaged women or mutually-disclosed serodifferent couples. In a prospective cohort study of a safer conception intervention offered to 400 HIV-infected individuals in Uganda who expressed fertility intentions with their partners, only 25% of the participants were men [61]. Similarly, in developing a safer conception service for serodifferent couples in Johannesburg, South Africa, Schwartz et al. were able to recruit significantly more women than men; moreover, all of the men in the study were accompanied by their partners whereas 45% of women participants attended the program alone, without their male partners [62]. These existing safer conception interventions face challenges in recruiting an equal proportion of men because reproductive health has traditionally been approached as the purview of women by providers [63, 64] and men have demonstrated reluctance to seek healthcare services [65, 66]. Men are less likely to be offered reproductive health counseling, the least likely to present for HIV testing, and are more likely to be lost to follow-up and die even if they initiate antiretroviral therapy [59–63].

Men’s poor engagement with HIV services is also related to hegemonic masculinity. Men face pressure to conform to representations that convey control, competitiveness, invulnerability, strength and agency [67]. These pressures result in heightened stigma and fear associated with HIV testing and treatment [68, 69] because having HIV is perceived as lacking physical strength and failure to be in control of one’s circumstances. Men’s perceptions of heightened stigma in turn make disclosure challenging. Only a minority of men in our study were able to disclose their status to their pregnancy partner, likely because living openly with HIV requires them to contradict their gender expectations and risk social ostracism by doing so [19, 67, 70, 71]. As a result, HIV incidence continues to rise among women in South Africa in the era of test and treat, likely in part because their male partners avoid testing and are not on treatment [72]. Beyond prevention of transmission, MLWH’s reluctance to engage in HIV-related services has resulted in substantial gaps in health outcomes between men and women; thus, though incidence of HIV in South Africa is higher in women, men are twice as likely to die of HIV-related causes [73, 74].

Yet, particular norms of masculinity, such as those that promote responsibility for building and supporting healthy families, can be harnessed to develop interventions that increase men’s engagement in reproductive and HIV care [67, 75–79]. Our group has developed an intervention to support MLWH who want to have children in the adoption of HIV risk-reduction behaviors such as HIV-serostatus disclosure and uptake of and adherence to antiretroviral therapy [7, 80]. Ideally, interventions for men should go beyond leveraging gender norms to encourage testing and treatment uptake [70]. Gender transformative interventions aim to foster gender equality while encouraging protective sexual behaviors and reducing incidence of STI/HIV [70]. In South Africa, the “One Man Can Campaign” developed by Sonke Gender Justice Network demonstrates the effectiveness of such interventions. This campaign created structured workshops encouraging men’s active involvement in fatherhood and was found to shift men’s attitudes toward greater gender equality and lead to greater engagement in safer sex habits and reductions in male violence against women [81].

Our study is subject to a number of limitations. We only include survey-reported HIV-risk behaviors and could not objectively verify participant-reported behaviors such as ART adherence. Social desirability bias may have led to underestimation of HIV-risk behavior and imprecise estimations of decision-making dominance. Our small sample size may not be representative of the larger population of South African men living with HIV. Additionally, we enrolled men who already knew their HIV-serostatus and were engaged in HIV treatment; thus, there are limitations in generalizing the results of our study to men who are not linked to HIV services or clinical care.

Given the high level of male dominance in sexual decision-making, programs must address existing gender norms in order to support the safe achievement of HIV-serodifferent couples’ reproductive goals. In addition, strategies to reach female partners of MLWH who may not access ART or safer conception care are needed to minimize periconception- and pregnancy- related HIV risks. Safer conception programming that addresses gendered partnership communication challenges regarding HIV disclosure, reproductive goals, acceptable HIV risk, and relationship commitment, alongside technical safer conception counselling is needed.