“I think it’s communication and trust and sharing everything”: Qualitative evidence for a model of healthy intimate relationships in Black women living with HIV and men in KwaZulu‐Natal, South Africa

Abstract

In South Africa, couple‐based interventions (CBIs) have been used to increase HIV testing, reduce HIV transmission, and shift relationship dynamics. To understand local definitions of healthy relationships, this study sought to collect qualitative data on a model of healthy relationships in a semi‐rural area of KwaZulu‐Natal, South Africa. We conducted semi‐structured qualitative interviews with HIV‐positive women (n = 15) and men of mixed HIV status (n = 15) who were in heterosexual, monogamous relationships (not with each other). Thematic analyses guided coding. Three primary healthy relationship behaviour themes emerged, labelled open communication, couple‐level problem‐solving, and active relationship building, which were related to various relationship facets (trust, support, respect, commitment, and connection). We purposively explored contextual themes, namely the role of HIV, positive community involvement, and power dynamics, to better situate the healthy relationship behaviour themes. HIV was not central to relationship conceptualisations and three different power structures (shared power/flexible gender norms, shared power/traditional gender norms, male‐dominated power/traditional gender norms) were described as being healthy. This model of healthy relationships is similar to observed definitions in other African countries and in high‐income settings. Findings can inform HIV programming content for couples in KwaZulu‐Natal, particularly the active relationship building component.

INTRODUCTION

South Africa currently has the greatest number of people living with HIV globally—an estimated 7.8 million people (UNAIDS, ²⁰²⁰). The epidemic is concentrated in Black South Africans, with 16.6% of this racial group living with HIV, compared to 5.3% of those who identify as mixed race, 1.1% in Whites, and 0.8% in Indian/Asians (Simbayi et al., 2019). Findings from a meta‐analysis show that working with couples, as opposed to individuals, is more effective for improving a number of HIV‐related outcomes, including condom use and women's medication uptake to prevent mother to child transmission of HIV (Crepaz et al., 2015). In general, couple‐based interventions (CBIs) for health issues aim to improve the behaviour of interest, for example, getting the couple to use condoms to prevent HIV transmission, as well as enhance the relationship itself, such as by improving the couple's communication (Baucom et al., 2012). It is therefore necessary to understand what constitutes healthy relationships for particular cultural groups and communities at high risk for poor HIV outcomes, in order to effectively help couples move towards improved HIV‐related behaviours as well as healthier relationship functioning.

There is a small but growing body of literature linking various relationships constructs to relevant HIV outcomes in high risk communities. Early work in sub‐Saharan Africa, including in South Africa, showed that power inequity and intimate partner violence increased risk for women's HIV acquisition (Dunkle et al., 2004; Jewkes et al., ²⁰¹⁰). More recent work in South Africa and Eswatini examining predictors of sexual concurrency (the presence of outside sex partners) in monogamous relationships showed that several different relationship constructs including satisfaction, commitment, intimacy, and trust, were associated with a reduced likelihood of sexual concurrency, particularly for women (Belus et al., 2020; Ruark et al., ²⁰¹⁸). With regard to HIV testing, Gari et al. (2013) showed that higher levels of relationship cohesion, which included comfort in seeking partner support and feeling supported by one's partner, getting along, and trust, were related to increased likelihood of receiving an HIV test in Zambia. Furthermore, qualitative studies from South Africa have shown that partners can provide informational, tangible, and emotional support to promote antiretroviral therapy (ART) adherence, though relationship conflict and partner misinformation can interfere with adherence (Conroy et al., 2017; Conroy et al., ²⁰¹⁷). For people living with HIV, daily adherence to ART is critical to suppress the virus and lead a healthy life (Cohen et al., 2011).

Given the role that intimate partners and relationships play throughout the HIV care cascade, investigators have tried to understand how healthy relationships are defined in local communities so that relevant relationship constructs can be assessed and intervened upon in CBIs. In Eswatini and Rwanda for example, love was described as foundational to a successful relationship, which included providing tangible support to one's partner, giving gifts, spending time together, engaging in sexual fidelity, and verbal or written displays of affection (Ruark et al., 2017, 2019). An important next step of this line of research is to develop a model of healthy relationships that shows how these various relationship behaviours fit together. This was done using qualitative findings from a study conducted in a peri‐urban area of Cape Town (Belus et al., ^2019b). The proposed model of healthy relationships involved four primary relationship constructs: active relationship building, emotional support/display, communication, and problem‐solving. Each construct was hypothesised to be associated with various short‐ and long‐term relationship outcomes. However, participants in the study had previously participated in a CBI focussed on alcohol use reduction and HIV prevention (Wechsberg et al., 2016), which may have shaped their personal views of healthy relationships.

Thus, the goal of this study was to build upon the work of Belus et al. (2019b) by assessing the relevance of their healthy relationships model to men and women who were naïve to CBIs and who were living in a semi‐rural area of KwaZulu‐Natal, South Africa. Prior work shows that working with couples in high‐risk communities is protective for a number of HIV‐related outcomes, yet what constitutes healthy relationships for these communities is poorly defined. Given that CBIs aim to shift relationship dynamics, it was necessary to determine how community members defined healthy relationships in KwaZulu‐Natal, the province with the highest HIV prevalence rate in South Africa, estimated to be 18.1% (Simbayi et al., 2019).

KwaZulu‐Natal is predominantly Zulu‐speaking and many areas are under a traditional authority. While monogamy is the predominant relationship structure, polygamous relationships account for ~12%–14% of marriages in this province, though this percentage is higher (~25%) for those who are 60 years or older (Hosegood et al., 2009). Level of unemployment is high (~28%) and over 20% of inhabitants in this area work in the informal sector (i.e., are not entitled to pension or other employment benefits; Statistics South Africa, ^2020a). It is necessary to collect data on how healthy relationships are defined in this community to inform locally relevant CBIs for HIV. Thus, we sought to determine whether the prior model of healthy relationships would apply to the current sample and whether there would be a need for any theoretical modifications, with the ultimate goal to inform the development of a CBI targeted to this location and population.

METHODS

Participants and procedures

This qualitative study was the formative work for the development of a CBI to improve HIV‐positive women's ART adherence. Participants were recruited from Vulindlela, a semi‐rural catchment area outside of Pietermaritzburg, in KwaZulu‐Natal, South Africa, in April and May 2019. The study recruited individuals (not couples) to complete individual semi‐structured interviews on the topic of healthy relationships and barriers and facilitators to the use of CBIs for women's HIV‐related health. Eligibility criteria for study participation were adults aged 18 and over, in a committed heterosexual and monogamous romantic relationship (i.e., not in a polygamous relationship), resided in one of the neighbouring communities to where the study took place, and conversant in either English or isiZulu, the dominant local language. Additionally, women were required to be living with HIV and have difficulty with ART adherence, given the study's larger focus to develop a CBI focussed on women's ART adherence. Men were included in the study because we wanted to ensure our understanding of healthy heterosexual relationships in this community was informed by both genders. Men were not required to have a particular HIV status, though we collected this information to understand both HIV‐negative and HIV‐positive partners’ perspectives on the proposed intervention. We focussed on heterosexual and non‐polygamous relationships as the prior theory of healthy relationships focussed on participants with a similar demographic profile. Rates of openly gay or lesbian relationships, or relationships with individuals identifying as transgender in this community, are also very low.

Women were recruited through local clinics. Study contact cards were given to clinic staff who provided study details to women attending the clinic who were HIV‐positive and had previously missed an HIV‐related clinic appointment or were nonadherent to their ART. The study, however, did not collect medical record data from participants and relied on self‐report for HIV‐related study eligibility. Men were randomly approached at various community locations within the Vulindlela area (e.g., taxi rank) and screened for eligibility. The study team approached 15 women and 15 men, all of whom were eligible. All agreed to participate and provided written consent. Participants were paid ZAR 120 (~8.33 USD) for participating in the individual interviews. The study was approved by a local ethics committee (#REC 3/19/09/18).

The individual semi‐structured interviews were conducted by trained research assistants who were bilingual in isiZulu and English and were from the local community. Research assistants were gender‐matched with participants. Interviews took place at the research site as well as in participants’ homes. All interviews were conducted in isiZulu and audio‐recorded. Interviews were subsequently translated into English and transcribed by research assistants.

It is important to mention the research team involved in the study so that our perspectives and potential influences on the data are made clear. Our team was comprised of Masters and PhD‐level academics from South Africa and North America. The team was diverse in terms of age and academic experience (ranging from junior scholars to senior academics), gender (both women and men), and racial and ethnic identity (Black African, White, Asian/Indian, and Mixed race).

Measures

Qualitative interviews were conducted in order to understand participants’ perspectives on romantic relationship behaviours and norms in the local community. The study used an emic framework (culturally specific) rather than an etic one (universal principles), although both can be valuable (Varjas et al., 2005). When creating the semi‐structured interview guide, we endeavoured to set aside any internal researcher and external contextual suppositions by relying only on the work that was previously conducted in Cape Town, which shares some important cultural similarities to the population of this study. The interview began with a description of the potential CBI, including a description of elements such as couples’ communication behaviours and working together to address issues pertaining to HIV, in order to determine the acceptability of this intervention in this population. Two bilingual (isiZulu and English), masters‐level researchers in KwaZulu‐Natal were trained on the semi‐structured interview guide. The content of the guide that centred on healthy relationship asked participants to define and describe healthy and committed romantic relationships, identify good qualities in a partner and in a relationship, as well as describe what unhealthy romantic relationships look like, all based on observations in their community and/or their own personal experiences. These open‐ended questions were not guided by a theoretical framework in order to draw on participants’ own personal and community conceptualisations of healthy relationships. The interview concluded with a vignette that was written to evoke the four constructs of the healthy relationships model put forth by Belus et al. (2019b). In response to the vignette, participants were asked to describe their impression of the relationship, indicate any healthy or unhealthy behaviours observed, examine the partners’ communication and decision‐making abilities, and indicate if trust, respect, and commitment were observed in the relationship. See Figure S1 for the vignette.

Data analytic plan

The study employed thematic analysis (Braun & Clarke, 2006) to identify relevant themes of healthy relationship behaviours. Thematic analysis is a qualitative method that identifies patterns of responses and can be used from a realist, constructionist, or contextualist theoretical framework. This study operates from a contextualist approach, trying to understand participants’ lived reality in terms of intimate relationships, but also recognises how societal factors impact on their reality. Thematic analysis is also appropriate to explore a wide variety of psychological, social, and health topics, which is fitting for this study (Braun & Clarke, 2006).

To carry out thematic analysis, transcripts were first reviewed by multiple study team members in order to familiarise themselves with participant responses. A codebook was developed, which identified codes, definitions, and example quotes. Researchers continuously sought to isolate and bracket out any existing preconceptions in this work throughout the analysis process, relying on existing research conducted with populations in urban and peri‐urban South Africa to guide the deductive analysis. Some examples of preconceptions that were discussed as part of bracketing included how the coding team conceptualised power dynamics and a tendency to focus on communication and problem‐solving skills as central to healthy relationships. Inductive codes were developed to capture participants’ responses to the open‐ended questions about how healthy relationships were defined in their community. Deductive codes were developed to capture the relationship constructs and outcomes described in the Belus et al. (2019b) model and only applied to the vignette. Each transcript was coded by two trained coders and discrepancies were resolved via consensus.

Once all transcripts were coded, data analysis was supported using NVivo (QSR International Pty Ltd, ²⁰¹⁸) by identifying code frequencies and the co‐occurrence of multiple codes. We first examined the coded responses to the open‐ended questions about how participants described healthy romantic relationships in their community. We then focussed on the codes that were mentioned in at least half of the sample by examining co‐occurrences between these codes and all other inductive codes. This was followed by summarising the responses for each code. For the vignettes, we followed a similar procedure and used both inductive and deductive codes. We examined the four major deductive codes, which mapped onto Belus and colleagues’ 2019 primary relationship constructs (active relationship building, emotional support/display, communication, and problem‐solving) and within each code, examined the co‐occurrence with all inductive codes. We then summarised each deductive code. We subsequently clustered codes into themes based on overlapping or related code summaries from both the open‐ended questions and the vignette.

Finally, after examining the dyadic relationship themes, and after receiving feedback on our manuscript from external experts on the topic that highlighted missing nuances in our conceptualisation, we purposively explored contextual factors that we viewed as culturally relevant to this setting but that did not emerge as relationship themes per se (but were still themes that emerged from the overall data). All contextual themes emerged inductively. During the process of identifying contextual themes, we further reviewed our original relational themes to ensure the findings were still accurate. We specify in text for the reader whether participant quotes came from inductive or deductive coding to increase transparency of the coding process.

FINDINGS

The demographics and clinical characteristics of sample are depicted in Table 1. Men and women were similar in a number of domains including age (both groups in their early to mid‐thirties on average), very few were married or cohabitating (~17%), and the majority had disclosed their HIV status to their partner (87% for women, 80% for men). Pronounced gender differences were observed in educational attainment (20% of women completed high school compared to 73% of men) and HIV positive status (100% of women vs. 27% of men).

TABLE 1.

Demographics and clinical characteristics of the sample

Variable	Total (N = 30)	Women (n = 15)	Men (n = 15)
Age, M (SD)	35.5 (8.9)	38.6 (7.8)	32.4 (9.3)
Relationship length in years, M (SD)	6.4 (6.5)	8.8 (7.9)	4.0 (3.3)
Married or cohabitating with partner, n (%)	5 (16.7)	3 (20.0)	2 (13.3)
Completed high school or above, n (%)	14 (46.7)	3 (20.0)	11 (73.3)
Cell phone access, n (%)	28 (93.3)	13 (86.7)	15 (100)
HIV‐positive, n (%)	19 (63.4)	15 (100)	4 (26.7)
Disclosed HIV status to partner, n (%)	25 (83.3)	13 (86.7)	12 (80.0)
Number of children, M (SD)	1.7 (1.3)	1.9 (1.3)	1.5 (1.3)

Abbreviations: M, mean; SD, standard deviation.

Three major healthy relationship behaviour themes emerged based on both the inductive and deductive components of the interview (see Table 2). The three themes were labelled open communication, couple‐level problem‐solving, and active relationship building. Each of the themes identified were viewed as dyadic behaviours, meaning that both partners needed to contribute to the behaviour in order for it to occur. The healthy relationship behaviour themes were related to various positive relationship facets including trust, respect, and commitment. Furthermore, three contextual themes emerged: the role of HIV, the positive role of community members, and power dynamics within the relationship. Figure 1 presents a visual depiction of the major relationship behaviour themes and the relationship facets that emerged, as well as the contextual factors that were examined to provide context for the relationship themes.

TABLE 2.

Healthy relationship behaviour themes based on inductive and deductive approaches

	Inductive approach a	Deductive approach b
Open communication	Able to discuss issues Open communication results in partners trusting each other Open communication results in feeling supported by one's partner Listening and hearing partner out is a sign of respect Couples who are in long‐term relationships are open and honest with each other	Both parties have the opportunity to explain their perspective c Listening and understanding each partner's perspective, particularly the man listening to the woman
Couple‐level problem‐solving	No yelling, calling names, or physical abuse Both parties listening to each other is critical to solving problems No interference from individuals outside of the relationship No arguments between partners is a sign of respect Solving problems together is key to lasting relationships	Both parties have the opportunity to explain their perspective c Both parties work towards solving the problem
Active relationship building	Spending time together, doing small everyday behaviours for one's partner Building a connection and bond	Spending time together Open about the relationship, including verbal and physical affection Prioritizing the relationship

^a Inductive approach was based on the open‐ended questions assessing personal and community definitions of healthy relationships.

^b Deductive approach was based on the four‐factor relationship model put forth by Belus et al. (2019b).

^c This facet emerged as part of open communication and couple‐level problem‐solving themes.

FIGURE 1.

Healthy relationship behavior themes, associated relationship facets, and contextual factors in defining healthy relationships. Note. The healthy relationship behavior themes were most closely aligned with relationships that had shared power/flexible gender norms but all three power structures were described as healthy, which is why the relationship facets span the continuum of the power context

Healthy relationship behaviour themes

Open communication

This construct centred on participants feeling comfortable discussing difficult topics with their partners (such as revealing their HIV status) or topics that required vulnerability (such as discussing the couple's future together). A key part of feeling comfortable discussing issues was knowing that the partner would listen and respond positively.

For me if a person is open and free to their partner and they are not always serious and my partner is not afraid of me, she can come at any time and talk to me about whatever is concerning her. … I think it’s communication and trust and sharing everything. You must not hide anything from each other, because if you do that, there will be problems in your relationship. Male participant #22, age 26, HIV‐negative, inductive coding

Being open allowed partners to receive support from each other on important issues and served as a foundation for participants to address any relationship issues directly with their partner, as opposed to seeking counsel from others.

So, when you are in a good relationship and you tell your partner that you’ve been to the clinic and found out that you have HIV, he understands. When either of you have problems, you handle those problems together. Not this thing of being afraid to even tell your partner and you end up telling your friends who might end up talking about it with everyone when it would have been better to tell your man, but you are afraid to tell him. Female participant #08, age 46, inductive coding

Finally, being able to communicate openly was also linked to trust. Exchanging cell phones or sharing a cell phone was a common example used to showcase that a couple trusted each other, particularly demonstrating that neither person had an outside relationship.

Firstly, the old[er] couple I’ve seen trusted each other. I’m saying this because sometimes when you called the other one, the call will be picked up by his wife and [she] tells me the phone is with her today. They are open to each other and they do not have secrets. Male participant #10, age 35, HIV‐negative, inductive coding

While most participants discussed open communication as a mutual endeavour between partners, with both partners acting as disclosers and listeners for each other at different times, instances of monodirectional communication dynamics were also highlighted as healthy behaviours. This monodirectional communication was raised mostly among female participants, who only highlighted that it was important for the female partner to have space to share and disclose, and the male partner would serve in the role of listener. For example, when asked what comprised a good relationship, one female participant outlined her ability to express herself and receive understanding in return. In another instance, a female participant discussed the support her partner provided to her by serving as a listener during times of difficulty. For these women, the importance of male partners having the opportunity to openly share and be understood by their female partners was not discussed.

Couple‐level problem‐solving

This dyadic theme comprised of three primary facets: solving problems between the two partners without outside interference, partners listening and understanding each other's perspective (which overlapped with open communication), and no physical or verbal abuse. Similar to open communication, a relationship where partners were able to solve issues between the two of them was viewed as healthy. Participants discussed that community members sometimes intentionally interfered in a couple's relationship, for example by gossiping about them. Participants saw that solving issues within the relationship would serve to protect the couple from this unwanted outside influence.

A healthy relationship is about sitting down with your partner and talking about your problems privately. Not going around and telling people about your problems. When you have a problem, you must not run to your friend, you must start by talking to your partner. Sit down with them and try and find a solution together. That is a good relationship. Female participant #2, age 46, inductive coding

Participants described that a key part of solving problems together was listening to each other's perspective and not using any verbal or physical abuse to get one's way. The phrase “treat each other well” came up often as participants described how partners should treat each other when facing a disagreement.

Like if I do something wrong, now they are shouting and saying all kinds of things [to me]…I think when you enter a relationship, your way of thinking is grown. A person must know that to solve a problem, we must sit down and talk and not shout at each other, throwing words at each other. Female participant #15, age 36, inductive coding

Participants described being committed to the relationship and feeling respected as a result of both couple‐level problem‐solving and open communication. Commitment came from being united in discussing difficult topics and making decisions together, which involved actions from both partners. Respect came from being listened to and understood. As the quote below describes, respect was an integral part of a healthy relationship, and a facet of the relationship that could emerge in many different ways.

Respect is crucial, you see that respect between the couple about everything [they do]. You must trust each other as well and enjoy each other’s company. You should not have secrets with each other, you must also treat each other very well. You must not swear at each other and treat each other badly. You should not fight to a point where neighbors can hear you or the kids can hear that their parents are fighting. Male participant #01, age 54, HIV‐positive, inductive coding

As the quote describes, open communication (not having secrets) and couple‐level problem‐solving (no abuse and treating each other well) were key to feeling respected in the relationship.

While most participants described couple‐level problem‐solving as a collaborative endeavor, some male participants discussed a more unilateral problem‐solving and decision‐making dynamic, in which the male partner holds the responsibility to solve problems or make decisions. While this stance occurred infrequently in our findings, it is notable that an alternative version of healthy problem‐solving in romantic relationships was raised, and this spoke to the complex dynamics of power within the relationship, which we elaborate on further below (see section on Power dynamics within the relationship).

Active relationship building

In the last healthy relationship behaviour theme that emerged, participants described relationship behaviours where both partners were intentional about building their relationship. The primary behaviours were spending lots of time together, prioritizing the relationship over other competing demands, and being open about one's partner and the relationship to other people in the community. Participants described relationships where couples engaged in task‐oriented behaviours, such as grocery shopping together, as well as pleasure‐oriented behaviours like going to the movies, as showing that the couple operated as a unit. Participants also described how the couple spending time together allowed them to build their relationship and lead to feeling connected. Furthermore, relationships where there was an element of being “friends first” or a sense of camaraderie between the partners further demonstrated the bond between partners.

There are couples when you look at them, you can see that they are friends, they do everything together, they are always together. Even when they are amongst other people you can never tell that they are in a relationship, they play and laugh together like friends. Female participant #07, age 32, inductive coding

In addition, prioritizing the partner was part of active relationship building because it showed that the couple would make time for their relationship. The fictitious couple in the vignette, Themba and his wife Londiwe, encountered a situation where friends of Themba came over unannounced, just as the couple was about to have an evening together after not having spent much time alone for a while. The vast majority of participants saw it as important for Themba to choose spending time with his wife over his friends. They described how protecting the couple's personal time together would enhance the couple's bond, because the couple would be able to spend time together and both partners would also see that the relationship was valued. Although most participants emphasised the importance of Themba prioritizing Londiwe when his friends made an unexpected visit, two male participants discussed the importance of Londiwe respecting Themba's time spent with friends, accommodating their visit, and Themba making it up to Londiwe at another time, if Londiwe was displeased.

Finally, being open about the relationship (i.e., not hiding that one is in a relationship with their partner) was also viewed as key to helping the couple build their connection. In the vignette, participants heard that the couple, Themba and Londiwe, would show verbal and physical affection to each other in public, like saying “I love you” and holding hands. Participants described how both types of affection demonstrated that the partners were not embarrassed that other people knew of their relationship and that it was an expression of genuine love. Participants described how it was also good for the community to see such positive demonstrations of love.

It is a good thing because he [Themba] is showing others that he loves her [Londiwe], and he is not afraid to let people know this, and he is not ashamed of her. Which means those who see them will get a good lesson from him, that here is a man who loves his woman. Female participant #09, age 31, deductive coding

A few participants were uncomfortable with the idea of the couple sharing a small kiss in front of their children, describing it as non‐traditional, but they saw value in the other types of affection. A small subset of participants strayed from the more widely held stance that affection was important part of active relationship building, but all agreed on the importance of intentional relationship building behaviours.

Contextual themes

HIV and relationships

It was important to understand the role of HIV on the health of romantic relationships, given the prevalence of HIV in KwaZulu‐Natal communities and the study aim of developing a CBI targeting ART adherence. While participants were not directly prompted to discuss HIV in relation to their observations or experience of healthy relationships, we nevertheless observed how infrequently HIV emerged as a topic. When HIV was discussed in connection to relationship health, the importance of comfort in disclosing one's HIV status and receiving acceptance by one's partner were most commonly described.

He supported me in every way, even with my HIV. He knew from the onset that I have HIV, I told him. He did not undermine me in anyway and that made me comfortable to be with him. Female participant #13, age 40, inductive coding

Participants also described joint HIV‐related health behaviours as positive relationship practices, such as reminding each other to take their ART, getting tested for HIV, and picking up their ART at the clinic together, as well as coordinated decision‐making around sexual health practices, including condom use.

I am going to make an example about me, the way I look at it a good relationship would be, committing ourselves to each other, take medication, use condoms, so that we can protect ourselves. Female participant #04, age 47, inductive coding

Overall, HIV did not feature prominently when participants autonomously described facets of relationship health. When HIV was discussed, the focus was on dyadic behaviours of communication, disclosure, and acceptance, as well as collaboratively engaging in healthcare behaviours together.

Positive role of community members

We intentionally explored how participants described the positive role of community members with regard to healthy relationships, since we observed the negative role of community members emerge as part of couple‐level problem‐solving. Participants described the importance of being seen in public with one's partner and showing the community that the couple was happy together. This was a form of active relationship building, but centred around public displays of togetherness and openness about the relationship, as opposed to openness with one's immediate circle, like family or friends. Behaviours that showed to the broader community the couple's togetherness, which included the couple spending quality time (like buying groceries or spending time with friends) and visibly having fun together, were indicators of a healthy relationship.

They took each other out, spent time together most of the time, you could see from their actions that these people love each other a lot. Male participant #20, age 31, HIV‐negative, inductive coding

At the same time, participants were conscious of showing respect in the community, particularly in respecting elders by avoiding overt displays of affection. Many participants described appropriate public relationship behaviours as an absence or limiting of physical affection, citing that such displays were both disrespectful to others and a poor model of behaviour for impressionable children in the community. Couples therefore sought to balance their general couple togetherness in public as a way to receive tacit community support for the relationship, while being respectful with their actions.

I would say it’s a good thing, to be taught to respect; you don’t go around kissing a girl in front of elders. You only show elders that you are in love only when you are planning to get married to this person or after you are married. Male participant #19, HIV‐negative, age 23, inductive coding

Power dynamics within the relationship

We observed that in the three healthy relationship behaviour themes that emerged, there was, for the most part, an inherent assumption that the couple would share power. When we purposively explored power dynamics in the data, we found a continuum of power dynamics emerge that were described as healthy: shared power and flexible gender norms (where the three healthy relationship behaviour themes were embedded), mix of shared power and traditional gender norms, and male‐dominated power and traditional gender norms. These are displayed in our conceptual model in Figure 1.

Many participants described a mix of shared power and traditional gender norms that were both seen as part of a healthy relationship. In the vignette, many participants stated that Londiwe and Themba should make the decision together about whether or not Londiwe should take the job promotion she was offered. Many participants agreed that Londiwe should take the job, even if it required the family to adopt more non‐traditional gender roles, such as Themba helping with child care. At the same time, these same participants described other healthy relationship behaviours that involved traditional gender norms such as women benefiting materially from their romantic relationship and men being the one to make certain decisions independently in the relationship.

Let me talk about what I have been through, there is a situation where you try and speak to a woman and she doesn’t listen. When you try to advise her, she doesn’t listen… I have experienced things like that and I would sometimes tell her what I want her to do and not allow her to have a say. Male participant #05, age 45, HIV‐positive, inductive coding

In the male‐dominated power and traditional gender norms, only men described this as being healthy. Men's independent decision‐making was again highlighted as a factor for healthy relationships. In addition, several participants described that women having non‐traditional roles like working outside the home would harm the relationship because the woman would end up neglecting the relationship. In the case of the vignette, one participant described Londiwe's new job opportunity as a threat to the stability and health of the relationship.

I still think Themba should not allow Londiwe to take the job. Yes, they need the money but that should affect their relationship at the end of the day. […] It will ruin their marriage. Yes, money is needed but it should affect the kids and the husband. She should not take it. Male participant #01, age 55, HIV‐positive, deductive coding

Keeping the family structure stable was described as being healthy, and having pre‐defined roles in the relationship was a strategy to achieve that outcome.

DISCUSSION

This study sought to assess the relevance of a model of healthy relationships, originally developed in a peri‐urban area of Cape Town, and applied to women living with HIV and men of mixed HIV status in a semi‐rural area in KwaZulu‐Natal. Through the use of inductive and deductive approaches, men and women identified three primary behaviours that were viewed as essential to healthy relationships using evidence from their own personal experiences, observations in the community, and relationship ideals: open communication, couple‐level problem‐solving, and behaviours that involved actively building a connection (termed active relationship building). These three themes overlap with the model put forth by Belus et al. (2019b), though we did not find direct support that the emotional support/display construct is its own domain, as it was in the original model. Further, we purposively explored three contextual themes, namely the role of HIV, positive role of community members, and power dynamics, to better situate the healthy relationship behaviour themes that emerged. The study provides qualitative evidence of relevant relationship domains in an HIV‐endemic setting as well as a basis for guiding CBIs that seek to address HIV‐related problems in this community.

The three healthy relationship behaviour themes that emerged from this study were situated within the shared power/flexible gender norms dynamic, although two other power structures were also identified as being healthy: traditional gender norms coupled with either shared power or male‐dominated power. This suggests there is more than one avenue through which relationship power and gender norms can be experienced as healthy, though the male‐dominated decision‐making coupled with traditional gender norms was described as healthy only by men. This conceptualisation dovetails work by Conroy et al. (2020) that shows young Malawians use traditional gender narratives as well as shared power to describe ideal relationships. Similar narratives that highlight the complexity of power dynamics in intimate relationships are also observed in Central and Eastern Africa, where studies demonstrate inconsistencies in gender norm beliefs as well as inconsistencies between gender beliefs and behaviours (McLean et al., 2020; Vaillant et al., ²⁰²⁰). While prior CBIs for HIV have incorporated gender norms and power into the interventions and demonstrate that these constructs are modifiable (Nikolova & Small, 2018), interventions should account for the shared preferences of both partners with regard to the constitution of power dynamics in the relationship.

At the same time, there is an abundance of evidence demonstrating that hegemonic masculinity and gender norms that promote men's dominance are associated with poorer HIV outcomes for both men and women (Jacques‐Aviñó et al., 2019; Jewkes et al., ²⁰¹⁰). Thus, the same behaviour that is considered healthy in the relational domain may actually put both men and women at risk for poorer health outcomes in the physical health domain. CBIs should, at a minimum, provide this information to couples in order to facilitate discussion and help couples make informed decisions about desired power structures and gender norms within their own relationship. Interventionists delivering CBIs can also use a variety of strategies, either subtle or more direct, to address relational power dynamics (D’Arrigo‐Patrick et al., ²⁰²⁰). Future research should explore how such strategies can be used effectively by interventionists delivering CBIs in diverse cultural contexts, such as in South Africa or other low‐resource global settings.

It is important to note that men in this study had considerably higher educational attainment than women. One of the biggest risk factors for girls’ high school non‐completion is teenage pregnancy, which is high in KwaZulu‐Natal (Statistics South Africa, ^2020b), though we did not assess when women had their first child in this study. It is possible that if men in the study had lower levels of education (more similar to women's), they would have more strongly identified with healthy relationship behaviours that are consistent with traditional gender norms and/or male‐dominated decision‐making, as this association has been shown in other contexts (Barker et al., 2012; Fattah & Camellia, ²⁰²⁰).

Despite the high rate of HIV in this region, and discussion about a CBI aiming to address HIV behaviours within the relationship occurring early in the interview, participants rarely raised the topic of HIV when depicting relationship dynamics in their community. While all women and 27% of men in the study were HIV‐positive, the lack of discussion around HIV seems to suggest that participants in this sample conceptualised healthy relationships independently of HIV. This speaks to earlier work that advocated for examining relationships not solely through a lens of HIV and focusing more on positive aspects of relationships (Thomas & Cole, 2009). When we did purposively explore HIV, we observed that healthy relationship behaviours included being comfortable to disclose one's HIV status and jointly engaging in healthcare behaviours, such as the couple encouraging each other to take ART and attending clinic visits together. CBIs can incorporate this as a strategy to both improve HIV‐related outcomes and foster healthy relationships.

The healthy relationship behaviour themes that emerged all have a direct impact on the content of CBIs for HIV. First, the emphasis on open communication and couple‐level problem‐solving skills is consistent with existing CBIs for HIV treatment and prevention that have focussed on skill‐building in these domains in South Africa, Kenya, and Zambia (Chitalu et al., 2016; Hatcher et al., ²⁰²⁰; Wechsberg et al., ²⁰¹⁶). Several studies have documented positive changes in these domains after skill‐building interventions (Belus et al., ^2019a; Stern et al., ²⁰¹⁹). These prior CBI studies have either overtly or latently articulated that communication and problem‐solving are important for healthy relationships; this study adds to the existing literature by using community members’ voices to describe how dyadic communication and problem‐solving fit together and relate to other relationship facets. The use of communication and problem‐solving skills as part of a potential CBI was discussed early in the interview, potentially prompting participants to think of these areas in the context of their relationships. However, the responses presented in this article are the result of open‐ended questions about defining healthy relationships in participants’ community, which occurred later in the interview, and soliciting specific perspectives and salient examples not previously discussed. Future studies will need to evaluate whether improvements in communication and problem‐solving result in the hypothesised relationship changes in the domains of support, trust, respect, and commitment.

Although communication and problem‐solving have both been discussed in the literature, much less attention has been paid to the domain of active relationship building, which is focussed on how couples can build and sustain an emotional connection and bond. Particularly in the HIV context, where sex without condoms can be a display of love and trust (Parker et al., 2014), there is a need to provide couples with alternative behaviours to demonstrate their care for each other. Based on findings from this study, this could involve helping couples plan activities to engage in together, a common strategy in behavioural couple interventions (Baucom et al., 2008), as well as ways for partners to show that the relationship is a priority. This also builds upon theoretical research that suggests the emotional component of couples’ relationships must be addressed in African couples (Thomas & Cole, 2009) and includes suggestions for how to work with couples around HIV to address the central emotional aspects such as trust and intimacy (Belus, 2020).

When it came to participants’ description of healthy relationships, they provided as evidence a mix of actual relationship behaviour, either observed in the community or experienced, as well as idealised behaviours of how couples “should” behave. Although we cannot determine the extent to which the healthy behaviours described were actively experienced in participants’ own relationships or in their community, our findings shed light on what is important to participants in this domain, which was the goal of the study. However, future research should descriptively examine the frequency of these relationship behaviours, which will highlight what is relevant and needed in terms of CBI content.

This study's findings support the relevance of three (out of four) healthy relationship factors from the Belus et al. (2019b) model, in a sample of CBI‐naïve participants. The emotional display/support construct did not emerge as its own primary theme in the conceptualisation of healthy relationships. That said, the key elements of emotional support/display from the original model, which included verbal and physical displays of affection and provision of emotional support, did emerge in this study, albeit as a byproduct of other relationship domains. Specifically, affection emerged as a behaviour to showcase one's feelings and comfort about the relationship, which was instead categorised as active relationship building. Furthermore, receiving emotional support was described as something that emerged from having open communication. In spite of emotional support/display not emerging as its own construct, both studies similarly described public displays of affection as non‐traditional and historically disrespectful to elders, which has been previously documented (Miller, 2013). Participants described a tension between demonstrating togetherness (an indication of active relationship building), such as being seen together in public, but also doing so in a way that was respectful to elders. Yet, despite these cultural norms, both studies revealed that at the individual level, displays of affection are often accepted and desired, but there is concern about being perceived as disrespectful to others. Thus, it is important to understand how family and community structures can impact relationship dynamics. CBIs need to help couples adopt behaviours that fit within their own couple‐level beliefs and those of their broader community; several existing couple and family interventions are very attentive to the role of environmental factors in delivery of the intervention (McDowell et al., 2017).

This study adds to the literature on how healthy relationships are defined, using community members’ voices, in the region of KwaZulu‐Natal, South Africa, and provides a model that may be used to guide CBIs for HIV. However, there are several limitations in this study that need to be considered. First, the findings generated via qualitative methods, which provide rich information and evidence on how community members define healthy relationships, are not widely generalisable. To address this, future studies should employ quantitative designs with women of mixed HIV status to better understand the relevance of the proposed relationship model to a more diverse set of couples. Furthermore, the study's design utilised individuals, not couples, to assess the constructs of interest. Though there are merits to both approaches, one limitation of collecting data from individuals is that descriptions of couple‐level behaviour rely on only one partner's perception. However, the lack of data on intimate relationships in sub‐Saharan Africa still make the findings useful and can be built upon in future research with couples. Finally, the study only recruited individuals who were in opposite‐sex and monogamous relationships. Polygamous relationships account for 12%–25% of marriages in KwaZulu‐Natal, depending on the age group (Hosegood et al., 2009). In addition, rates of individuals who identify as having same‐sex attraction are low due to heteronormativity, stigma, and violence against individuals with diverse sexual orientations (Reygan & Lynette, 2014). Future research should investigate how couples in such relationships define healthy intimate relationships and subsequently ensure that appropriately tailored interventions are available for these populations.

Overall, this study provides evidence on how women and men from a semi‐rural area of KwaZulu‐Natal define healthy, heterosexual, and monogamous relationships, including relevant contextual factors, and provides support for a number of the domains observed in Belus et al. (2019b) prior research. Study findings are in line with prior research on healthy relationships in sub‐Saharan Africa, and also extend our understanding of this domain by showing how various relationship constructs fit together. Results from this study and future research are expected to have a direct impact on how CBIs for HIV intervene on relationship functioning, for the purpose of not only enhancing health outcomes for HIV but also improving the relationship well‐being of participating couples.

Supporting information

Fig S1

Belus, J. M. , Bradley V. D., van Heerden A., Msimango L. I., Barnabas R. V., & van Rooyen H. (2022). “I think it’s communication and trust and sharing everything”: Qualitative evidence for a model of healthy intimate relationships in Black women living with HIV and men in KwaZulu‐Natal, South Africa. Family Process, 61, 1507–1524. 10.1111/famp.12744