A South African Couple-Based HIV Prevention Program: Preliminary Evidence of the Long-Term Effects

Abstract

The goal of our study was to provide qualitative data on the long-term effects of a couple-based HIV prevention program, the Couples Health Co-Op (CHC), in South Africa. Qualitative focus group discussions were conducted with nonrandomly selected Black South African men (n = 27) and women (n = 23) who had participated in the Couples Health Co-Op 4 to 6 years prior to our study. The study evaluated: (a) salient content and skills learned, (b) long-term changes and challenges, and (c) recommendations for intervention improvement. Findings revealed (a) communication/problem-solving, safe sexual behaviors, and negative effects of alcohol were most salient; (b) long-term changes occurred in communication and healthier sexual behavior; alcohol use remained challenging; and (c) participants recommended continuing the couple format and suggested targeting teenage couples. We offer preliminary evidence of the strengths and weaknesses of the Couples Health Co-Op and provide a basis for future studies to build on these results.

Couples have become an important focus for HIV prevention efforts. The research to date on couple-based HIV prevention interventions, with both seroconcordant and serodiscordant couples, shows that such programs have a positive effect on HIV-risk behaviors, increasing condom usage and decreasing sexual concurrency (situations in which an individual has overlapping sexual relationships with more than one person; LaCroix, Pellowski, Lennon, & Johnson, 2013). Most studies, however, have only evaluated the efficacy of these interventions in the short term, with the longest follow-up times typically being 1-year postintervention (Jiwatram-Negrón & El-Bassel, 2014). Yet, HIV risk still exists for couples beyond 12 months after completing the intervention. Thus, it is imperative to know what the longer-term impact of HIV prevention interventions is on couples’ HIV risk behaviors.

The Couples Health Co-Op (CHC) is a couple-based intervention that was designed specifically for heterosexual South African couples who engaged in risky sexual behavior (i.e., inconsistent condom use) and where the male partner drank alcohol regularly (see Wechsberg et al., 2015, for a full description of the intervention). Briefly, the CHC was designed to target the intersection of alcohol, safe sexual practices, gender roles, and relationship violence and also included a focus on a number of relational aspects, such as communication and intimacy between partners. The CHC was adapted from several other evidence-based interventions: a South African woman’s intervention for HIV and alcohol use reduction (based in feminist and empowerment theories; Wechsberg, Luseno, Kline, Browne, & Zule, 2010); a South African men’s program targeting intimate partner violence, sexual health, and HIV prevention (based in an ecological framework, Peacock & Levack, 2004); and a U.S. couple-based program to reduce the risk of HIV and sexually transmitted infections (based in risk reduction and ecological frameworks; El-Bassel et al., 2003). The CHC was delivered by peer leaders from local communities in two 3-hr group workshops that took place 1 week apart in community settings. Participants attended the workshops with their intimate partners as well as other couples from their communities. In between the two intervention workshops, couples were given a homework assignment on an identified relationship goal specified in their action plans.

In the primary outcome study of CHC efficacy, study participants were followed for a period of 6-months postbaseline interview. Results from that study showed that, in comparison with the control interventions (a woman-only intervention or a same-gender intervention delivered to both partners in a couple), the CHC reduced men’s problem drinking, improved men’s and women’s condom use, and reduced the likelihood that women who were HIV-uninfected would seroconvert (Wechsberg et al., 2016). However, the short follow-up time period of the intervention meant it was difficult to draw conclusions about the sustainability of skills over time. Moreover, not all couples showed improvements in expected outcomes, suggesting an opportunity to refine the intervention by understanding individual experiences in participating in the CHC.

The aim of our study was to examine the long-term effects of the CHC. Specifically, the goals were to (a) understand sustainability of skills and content learned several years after couples had participated in the intervention, (b) determine whether the intervention had long-term positive effects, and (c) discover opportunities to improve the CHC. We also evaluated the importance of individual satisfaction with the intervention, particularly its focus on the couple as the point of intervention, as opposed to an individually focused intervention. No specific hypotheses were put forth for the study because the goal was to understand participant experiences over the long-term using skills learned in the CHC.

Methods

Participants and Procedure

We employed focus group discussions (FGDs) to investigate the aims of the study. Participants were eligible for the study if they had participated in the previous CHC study (Wechsberg et al., 2016) and provided their consent to be recontacted for future research studies. All participants completed the intervention between 2010 and 2012. Men and women participated in FGDs separately to increase participant comfort with disclosure of information. Only limited demographic information was available on the sample due to restrictions placed by the U.S. ethics committee on collecting new demographics. Therefore, demographic information available was from the original data, which were collected 4–6 years prior to the current investigation. All participants were from Khayelitsha, a large disadvantaged community consisting mainly of informal housing outside of Cape Town, and were Black African. Women were 22.74 years on average (SD = 3.68 years), 13% (n = 3) had received a high school education or greater, and 43.5% (n = 10) were living with HIV. For men, the average age was 24.73 (SD = 4.28), 26% (n = 7) had received a high school education or beyond, and 7.4% (n = 2) were living with HIV at the time of the original study. Low rates of completed high school education for both genders (Statistics South Africa, 2011) and high rates of HIV among women (Shisana et al., 2014) were consistent with demographic characteristics in the region. Couple status was not systematically collected during the current FGDs, but participants reported informally during the focus groups that some were still with the same partners as during the CHC study, some participants were in new relationships, and others were single.

To select a sample for our study, a randomly chosen subset of 28 men and 28 women (out of a total 198 eligible participants) were identified as an initial point of contact for recruitment because this was anticipated to be an appropriate sample size based on other literature in this substantive area (Myers, Carney, Browne, & Wechsberg, 2018; Wecshberg et al., 2013). However, we were not able to reach all participants who were initially on our list for recontact. To aid in the recruitment of other participants, we used snowball sampling with participants who agreed to be in the study and were eligible for participation. A total of 29 men and 29 women were screened and eligible for the study although the final sample size consisted of 27 men and 23 women who participated in the FGDs (i.e., some men and women did not show up on the day of their scheduled FGD). To participate in the study, interested participants were required to consent to the FGD being audio-recorded and also had to be willing to share their experiences of being in the CHC study in a group. Eight FGDs were conducted (four for each gender) with four to eight participants in each group, as this appeared to be the point of data saturation. Recruitment took place in July 2016 and FGDs were carried out in July–August 2016.

FGDs were approximately 2 hr in length and took place at the field site in the community of Delft. FGDs were co-led by two female moderators (either two White North American women or one White North American woman and one White South African woman) and included one or two local Black South African staff for language assistance. The FGDs for men consisted of male and female staff, whereas the FGDs for women consisted of a female staff member only. The South African staff translated questions and responses to the local language spoken by participants (isiXhosa) for anyone who felt more comfortable in their native language rather than English (although all participants spoke English to some degree). The FGDs were audio-recorded and later transcribed for data analysis. All participants received a grocery voucher for ZAR100 (approximately $6.65 USD) for study participation. No adverse events were reported during the study. Participants signed a consent form to participate in the study. Ethics approval was obtained from RTI International Institutional Review Board in the United States and the South African Medical Research Council Ethics Review Board in South Africa.

Materials

An approved semistructured interview guide was used for the FGDs. The questions probed the following areas: (a) most salient skills and content domains sustained from the intervention including communication and problem-solving skills, sexual risk behaviors (e.g., condom use and sexual concurrency) and alcohol use; (b) maintained behavior changes and behaviors that were less amenable to change; and (c) desire to continue the intervention and recommendations on future changes.

Data Analytic Plan

To examine the study aims, we undertook an inductive thematic analysis (Schreier, 2012), which allowed us to identify themes that emerged throughout the FGDs. Specifically, transcripts from the FGDs were read multiple times to become familiar with the data. Then, transcripts were coded twice—first using a line-by-line coding strategy, which helped to keep the initial codes tied as close to the data as possible, then coded a second time as iterative changes to the codes were made as more transcripts became available for coding. Two independent coders (both authors of the study) completed the coding of all transcripts. The coders independently read the transcripts, met to discuss emerging themes, and conducted coding based on initial impressions. Discrepancies in codes were discussed, and a consensus was reached as to the appropriate codes. Interrater reliability was not calculated because of the iterative nature of the developed codes. A codebook for data analyses was maintained that included the name of the code, a definition of the code, and example text that supported the code. The data analytic software ATLAS.ti (version 7) was used for data analyses. Specifically, the software allowed for the examination of code frequencies, co-occurrence of multiple codes, and the extent to which codes were densely connected to other codes in the network. All tools aided data organization and management to support identification of themes.

Results

Participants were asked to address three primary areas related to long-term follow-up of the CHC intervention: (a) content and skill areas they most remembered learning about, (b) longer-term changes and challenges to using the skills since participating in the intervention, and (c) recommendations for intervention improvement. A diagram of these domains and the related themes are provided in Figure 1.

Figure 1.

Information learned, long-term effects, and intervention recommendations for the Couples Health Co-Op.

Most Salient Content and Skills Learned

Men and women participants reported learning skills and content in three primary domains: communication and problem-solving, safe and healthy sexual behaviors, and negative consequences of alcohol use. Within these domains, there was convergence and divergence between genders about which facets were salient; the similarities and differences are highlighted.

Communication and problem-solving.

Men and women discussed communication skills, specifically selfdisclosure, listening, and understanding one’s partner, as some of the most valuable aspects of the intervention. Listening to one’s partner was viewed as a way to show respect by giving the partner an opportunity to express himself or herself. It was described as incompatible with dogmatically pushing one’s viewpoint onto one’s partner.

It’s always better to talk with your partner, and not only talking, but to listen at what she’s going to come up with, to give you that sort of respect of hearing, or listening to what she wants to say as well. Not only talking more about being pigheaded and telling yourself that you are a man in this relationship…

(Male participant, FGD #2)

The other major aspect of communication mentioned by participants was self-disclosure to their partners. Participants learned that it was important to talk about problems with their partners, rather than keeping feelings to themselves and making assumptions about how their partners would respond.

You will have something to say to your partner, but you don’t want to say it, because you know that he’s going to say—he’s not going to be on the same path with you. But what we’ve learned here is to sit down with your partner….I tell him what I feel. And then at some point, he had something else also to say to me, but he now felt comfortable just because we sat down and then we talked. That’s why I can say, that’s why I’m still in this relationship with my partner.

(Female participant, FGD #5)

As demonstrated by the quote, mind-reading (i.e., assuming one knows what another person is thinking) often occurred regarding how one’s partner would respond in a given situation, which inhibited self-disclosure and perpetuated feelings of discomfort between partners. However, participants said that, ultimately, learning to share their thoughts and feelings made them feel more comfortable with their partners and helped them get to know each other better.

It’s about communication. The more I communicate with her, the more I find that to be a help. To be more, like, to be all free when she opens up to me also, so that I can start to know what she really [is like] deep down, what kind of things that are hurting her. So that I can help her more freely open up…

(Male participant, FGD #2)

Moreover, salient communication behaviors for men also included more adaptive problem-solving strategies. Men reported learning that physically violent behaviors were an inappropriate means to solve problems. Instead, they learned that verbal discussions would help them solve problems more effectively. Encouragingly, no men discussed keeping in their negative feelings as a new alternative strategy to solving problems or disagreements in the relationship.

Safe and healthy sexual behaviors.

Within the domain of sex, men and women both discussed the importance of learning that alcohol affected the ability to engage in safe sexual behaviors, particularly with regard to sex partners outside of the relationship and condom use.

And alcohol, when you are drinking alcohol, you get drunk you do something that you… maybe you find another partner on top of that partner you already have, and then you go and sleep with the other partner and then that partner maybe doesn’t want to use condoms.

(Female participant, FGD #8)

This was a fairly common narrative, mostly from men but also reported by some women, which involved having sex with an outside partner when under the influence of alcohol. In some instances, participants were unable to remember whether a condom was used and, therefore, whether they had the potential to infect their main partner with HIV or another sexually transmitted infection. The intervention appeared to help men and women coalesce these experiences into a uniform narrative (i.e., alcohol impacts judgment and decision-making and can result in unprotected sex with an outside sex partner). The intervention not only highlighted the importance of using condoms with outside partners, but also with main partners. “I have learned how not to abuse alcohol and to use a condom all the way, even if I have one partner because you don’t know what he is doing behind your back” (Female participant, FGD #7). An important tenet of the intervention was to encourage couples to use condoms consistently and not equate condom use with mistrust.

Additionally, some unique facets emerged as important for men versus women in the domain of sex. Men primarily remembered the importance of being faithful to one partner and the role that outside partners played in increasing HIV risk. “I saw it as an advantage that it’s not good to have more than one partner…because it’s easy to get some disease, because it’s not all of us who uses condoms” (Male participant, FGD #1). On the other hand, women learned how to communicate with their partners about sex or sexually related issues, including HIV. These topics were often difficult to discuss, but the women appeared to embrace the idea that it was important to have open conversations with their partners. “We were not comfortable to speak about sex before the workshop, but after the workshop, we were free to talk to each other. Negotiated the condom use, on how to make sex pleasurable…” (Female participant, FGD #5).

Negative consequences of alcohol.

In addition to learning how alcohol interfered with sexual decision-making, men and women recalled learning how ubiquitous alcohol was in their communities. Men also discussed the availability of alcohol and the extent to which peers and community members used alcohol, which made it difficult to distance themselves from it. “In the communities that we live in, people are, we are so close to drugs, we are so close to alcohol. And people tend to use alcohol and they tend to mistreat their partners” (Male participant, FGD #1). Women primarily remembered the impact that alcohol had during pregnancy on an unborn baby. For example, “When you drink when you are pregnant, it can damage a baby, according to… come out before time. And it can affect the baby in the brain; it can damage the joint system of the body” (Female participant, FGD #7). Women reported not having this knowledge prior to the intervention, but since the intervention, some reported taking it on themselves to teach this fact to other women in their communities.

Long-Term Changes and Challenges

With regard to long-term changes, participants most commonly reported significant shifts in their intimate relationships since participating in the intervention, albeit in different ways. For men, most sustained changes occurred in their abilities to use more effective communication skills, specifically speaker and listener skills as well as adaptive problem-solving skills, as discussed earlier. But improvements noted were not limited to communication skills; they were often part of broader changes in the importance men placed on intimate relationships.

One of the reasons why I still maintain my relationship with my partner is that I’ve realize[d] that before, if I want to hang around with friends I would hang around with friends and drink and not care about her, and not spend time with her. But now if she phones me, if she wants to spend some time with me I prioritize her. And, that’s the reason why I stick to her.

(Male participant, FGD #3)

As this quote shows, men reported that they began to prioritize their female partners more and understood that relationships required effort. Men’s newfound commitment postintervention not only influenced how they conceptualized their roles as partners, but also their roles as fathers.

So, before the workshop I used to drink a lot and spend all the money, and not really care about my kid. But now, I know for sure that I need to take care of my kid. I need to set money aside to take care of her needs.

(Male participant, FGD #4)

Overall, the most notable shifts for men after participating in the intervention were in how they viewed their intimate relationships and the ways in which they communicated with their partners.

For women, the most notable long-term changes occurred in their sense of empowerment to be treated well by their intimate partners. Women reported speaking up more when they observed behaviors from their partners that were harmful to their own health or behaviors they were uncomfortable with, such as partner abuse (physical or verbal) or secretive behavior.

I would say, yes, it did change a lot, because before [the workshop] I was like whatever the boyfriend says is okay, but now I’m like, “Okay, no it is not okay”…[The workshop] did change [me] because before I was scared to speak up. Even if he’s older or drunk. Even if he is wrong, I don’t get acknowledged that he is wrong. But now if what he says is wrong, I speak up. It’s 50-50.

(Female participant, FGD #8)

As this quote demonstrates, following the intervention this woman no longer passively agreed with what her partner said. Instead, she asserted herself in order to be in an equal relationship, even in situations where her partner may have been more dominant. Additionally, women spoke of becoming more empowered to leave relationships that were not meeting their needs.

Most of them [women from the community who were in the workshop] broke up with their boyfriends. Because when you trust someone, you ask that person if he is not answering you and he is abusing you, it goes altogether. You tend to leave that relationship because there is no trust, there’s no honesty, there’s no love. So you move on with your life. That’s how, like, because you [are] educated.

(Female participant, FGD #6)

As noted by this participant, some of the women in her community who participated in the intervention were no longer with their partners because they had ended maladaptive and violent relationships. Consistent with changes in the women’s sense of empowerment, they noted long-term improvements in the ability to talk about sex and use condoms with their partners.

Although there were notable long-term changes for both men and women after participating in the intervention, there were also areas that were less amenable to change. Most notably, both genders reported that alcohol use was an area that required further intervention. Some men reported improvements in this area, but many men and women said it was still a problem in their own lives and in their communities. Alcohol affected men’s individual functioning, often by fueling anger, which at times led to violence with other men.

So, there are times where I act out of character. For instance, when I’m angry. I go outside because I still use drugs. I abuse alcohol… In order to get money, I get into the thug life and hustle and get money. I don’t want to say that it’s right. It’s not right. I want someone who can help me. It only happens when I’m drunk.

(Male participant, FGD #3)

Alcohol use also caused relationship problems for men because many of their female partners did not approve of their drinking habits, which created conflict for the couple.

Yes, I’m drinking each and every weekend, knowing that my behavior is still good. But, the problem is at home. I clearly understand that that person there at home needs me. She needs that quality time to spend [with me]. Now, if I go get drunk, she’s already disappointed, my relationship goes down.

(Male participant, FGD #3)

For women, a similar picture emerged where alcohol use impacted their individual functioning, such as nonadherence to antiretroviral medication if living with HIV and poor decision-making (e.g., not using condoms with a new partner). It also resulted in relationship problems, such as arguments and relationship dissolution. “I also have the same problem with alcohol. I ended up breaking up with my partner because of alcohol and I lose control when I am drunk” (Female participant, FGD #8).

Intervention Recommendations

Participants in the FGDs reported a strong desire to have the intervention continue in the community. The topics desired were primarily those already covered in the intervention, such as the role of alcohol and safe sexual behavior. When asked about whether the intervention would be better in a couples format with both male and female partners present (as was the case with the original intervention) or completed separately by gender, participants overwhelmingly reported that the intervention with couples would be most effective. For men, participating in the intervention with their partners would ensure that both members of the couple heard the same information; partners would be held accountable to each other, and each person would better understand their partner’s perspective. A few men wanted the intervention to continue as separate groups for each gender because they were shy and/or uncomfortable sharing information with their partners. However, other men understood the merits of both formats (separate by gender and as a couple) and saw that without participating as a couple, the goal of couples interacting more adaptively would not be achieved.

Look, for me it’s about both ways. The workshops with you, the men apart from the women, and the workshop, you know, with partners in… I get the point that the other men are shy to open up in front of their women, or in front of women, but that is what we are trying to solve in a way. So for me, I think, we can do men’s only here… But again, should you stop there, that problem of men not being able [to talk], because they end up resulting to violence… you see, because they can’t deal with their problems, because they can’t talk… Men must learn to sit with their women and being able to talk with them, and being able to listen to them.

(Male participant, FGD #2)

The overall sentiment was that men and women should participate together so that they could better understand each other and be comfortable listening and sharing.

Many women also preferred the couples format instead of the intervention divided by gender. Women felt that participating as a couple would increase the likelihood that both partners would be on the same page and better understand each other. Women also noted that if their male partners did not participate in the intervention, they were less apt to believe the information if it only came from the female partner.

[If] I have this partner who hasn’t attended the program, then when like, I prefer that partner to have the information from the horse’s mouth, from the professionals, because when we [women] like dishing out all the information, he will never trust you. Instead, he will make like more chaos.

(Female participant, FGD #7)

Many women felt that their partners disregarded their perspectives, especially if the women were unemployed. However, they felt that men would be willing to listen to new information if it came from professionals or from other men. Furthermore, the women who preferred to have the intervention separated by gender believed that their male partners would be too shy to talk in a group with other women (consistent with what some men reported) or because their male partners held different viewpoints from their own, making it too challenging to have the intervention in a couples format.

Finally, participants felt that the intervention could be used to address some of the community’s most challenging problems. Throughout the discussion, men and women noted many structural issues in the community, such as unemployment or lack of community resources, but a major theme that emerged was teenage pregnancy. Teenage pregnancy was identified in each FGD as a major problem plaguing the community. Participants noted that alcohol and drug use was a major factor contributing to teenage pregnancy.

You see, like, where I come from in my community, like, eish, I’m very worried about the youngsters. Because, they are exposing many things to them, there’s drug-alcohol abuse, and then, they get in a relationship in the early stages. So I think, interventions like this then can work, especially starting from schools. Because there’s too much teenage pregnancy.

(Male participant, FGD #2)

Participants thought a couple-based program could be used to prevent teenage pregnancy in their communities by working with teenage couples, teaching them about alcohol, safe sex, and HIV.

Overall, the FGDs revealed that the CHC intervention had the greatest impact on relationship knowledge and skills for both men and women and that these changes were some of the most robust over time. Communication skills were reported by all participants to have been central to improving their relationship functioning. In addition, men also noted shifts in how they viewed and prioritized their partners, whereas women became more empowered to ask for healthier relationship behaviors as well as to end relationships that were not meeting their needs.

Discussion

Long-term feedback on intervention sustainability is a scarce, but essential, element for intervention developers and policy makers. The goal of our qualitative study was to provide preliminary data on the long-term effects of the CHC, a couple-based HIV prevention intervention in South Africa, where heterosexual HIV transmission is high. Participants reported that the three major areas of the intervention that were most salient to them were communication and problem-solving skills, learning how to engage in safe and healthy sexual behavior, and the negative effects of alcohol use. These domains map onto the foci of the CHC (Wechsberg et al., 2015), suggesting that the intended main goals were communicated effectively. Although these three areas were remembered exceptionally well a number of years later, participants reported a differential impact of the intervention in these areas. The most notable long-term changes, according to both men and women, occurred in their relationships, albeit in different ways.

Men reported behavioral shifts in the way they communicated with their partners, primarily listening more and trying to solve problems using a more collaborative approach. Also striking were the cognitive shifts that men described in the ways they viewed their intimate relationships. After the intervention, many men noticed that they prioritized their relationships more; instead of going out with friends, men reported willingly choosing to spend time with their partners. This shift also extended to prioritizing their roles as fathers, which is consistent with limited research on South African men as fathers that described a good father as someone who devoted quality time and was able to provide emotional support (Ratele, Shefer, & Clowes, 2012).

For women, cognitive and behavioral long-term changes were also reported in their intimate relationships, but these differed from male participants. The cognitive changes occurred in women’s standards about what was an acceptable way to be treated. Behaviorally, this translated into women voicing their opinions more often and a willingness to discuss sex with their partners. This is consistent with research showing the positive effects of women’s empowerment interventions, including increases in the empowerment of women living with HIV to use condoms in low- and high-income countries (Robinson et al., 2017), improvements in women’s relationship quality with their partners in Burkina Faso (Ismayilova et al., 2017), and increased confidence for women to leave abusive relationships in South Africa (Kim et al., 2007).

Despite long-term changes, there was consistency in participant reports that alcohol use was still a major concern. Some men noted longer-term changes in this domain, but many said it was still a problem. In the original CHC study, men demonstrated encouraging improvements in problem drinking 6 months following completion of the intervention (Wechsberg et al., 2016); our findings suggest that some men are unable to sustain these changes. Participants noted that continued alcohol use impacted individual function, primarily through difficulties in self-control and decision-making. This is consistent with the Morojele et al. (2006) model of alcohol use in South Africa that suggests one of the major ways in which alcohol use impacts function is through impaired decision-making. Research has found that the most effective approach to impacting alcohol use disorders is making changes to policy, for example regulating the demand and availability of alcohol through means such as taxation and advertising bans (Patel et al., 2016). However, psychological treatments are needed for refractory cases (Patel et al., 2016), although evidence has suggested that brief interventions targeting individuals with high alcohol use are only minimally effective (Anderson, Chisholm, & Fuhr, 2009). This suggests that further research is needed to develop culturally appropriate effective treatments for individuals with refractory and high levels of alcohol use in South Africa.

Finally, with regard to recommendations for the intervention going forward, both men and women wanted to see the intervention continue using a couples format, rather than each gender individually, because they felt it was necessary for both partners to have the same information in order for the couple to shift behaviors. This recommendation was not surprising, given the notable long-term shifts described by participants in how they understood and viewed relationships. Moreover, it is encouraging to have consistency among men’s and women’s desires for couple-based interventions. Men’s preferences for continued opportunities for couple-based interventions contrast with some of the ways in which South African men have been portrayed in terms of their masculinity. Research has documented difficulties in recruiting men for HIV prevention purposes in sub-Saharan Africa because the role of masculine identity might interfere with seeking services (Hensen, Taoka, Lewis, Weiss, & Hargreaves, 2014; Sharma, Barnabas, & Celum, 2017). However, Rosenberg et al. (2015) showed that when male partners of women living with HIV in Malawi were offered “family-focused services,” 75% subsequently attended a clinic appointment. This suggests that framing services as couple-based or family-based may be more appealing to men if they are able to identify a need in their female partners. In addition, our own recent quantitative dyadic analysis showed that couples who received the CHC intervention demonstrated more egalitarian gender norms than the control condition at 6-month follow-up (Speizer et al., 2018), further supporting the notion that masculinity is not an intractable barrier from engaging men in effective HIV treatment and prevention efforts.

Finally, the other major recommendation that emerged from the intervention, unexpectedly, was a call for help on a major community issue—teenage pregnancy. Participants discussed the intergenerational transmission of social problems in their communities and how many young children are not cared for by their families. Participants viewed alcohol use as a major cause of teenage pregnancy. Participants recommended a couple-based intervention to curb unsafe sexual behaviors and alcohol use in this population, given that many unexpected pregnancies occur in the context of adolescent committed relationships. Research has documented the risk behaviors of South African teens, including early sexual debut and multiple partners (Shisana et al., 2014). Working with young couples may be an appropriate and novel approach to reduce pregnancy and HIV risk in this population.

The interpretation of our findings needs to be viewed in light of how interviewer demographics (i.e., gender, race) may have interfaced with participant characteristics and shaped the results of the study, especially for the male FGDs. As described by Broom, Hand, and Tovey (2009), gender dynamics between a female researcher and male interviewee can lead to men providing narratives consistent with idealized masculinity (as a way to assert power over female interviewers) and/or the opposite, with men providing more feminized accounts of their experiences, which includes more emotionally laden or vulnerable content, either intentionally to please female interviewers or unintentionally because they feel more comfortable disclosing such experiences. Men in the study did not use hypermasculine language to describe their experiences. When they described sexual relationships, they often did so cautiously and at times appearing embarrassed to discuss certain sexual dynamics. Male participant perspectives, for the most part, converged with female participant perspectives, suggesting that men were comfortable sharing their personal experiences, which may have been enhanced by having female interviewers. Moreover, because the female interviewers were all White, it was obvious to participants that the interviewers had not had the same experiences as the participants in their community (South African communities are often divided on racial lines). Throughout the FGDs, it seemed that men and women were intentional about explaining relationship or community dynamics in sufficient detail so the interviewers would have an adequate understanding of their lived experiences.

Our study contributes to the literature by providing qualitative data on the long-term effects of a couple-based HIV prevention intervention but several limitations exist that must be considered. First, follow-up data were not collected from the couples who received an intervention similar to the CHC but in same-gender groups (one of the control interventions in the primary study). Without these data, it is difficult to ascertain whether the changes described by participants were specific to the couple-based intervention or whether the results would be similar in any of the active intervention groups. Moreover, it is also possible that the changes described by participants were due to the passage of time (e.g., gaining maturity and therefore valuing relationships more) and not as a result of the intervention. This possibility cannot be ruled out and is a limitation of our data interpretation. However, to our knowledge, this is the first study that provides data on the long-term effects (more than 4 years after treatment has ended) of a couple-based alcohol and HIV intervention. These results can be used as a benchmark for future studies.

Another limitation of our study is that only limited demographic data were collected on the sample, including relationship status and HIV status. These demographics are especially relevant to the study because intervention efficacy may be differentially affected by these characteristics; the lack of this data limits the specificity of more precisely knowing who the intervention works for and under what circumstances (e.g., unclear if the skills transfer over into new relationships). That said, our study has a moderate amount of external validity because it assessed the long-term effects of a couple-based intervention provided in community settings to men and women from a high-risk South African community. The demographic characteristics of the study sample (e.g., employment, HIV status) are reflective of the larger community’s vulnerabilities. Finally, it is important to note that because data were collected in FGDs with the researchers present, this may have biased participants’ accounts of their past experiences.

In spite of these limitations, the study provides novel qualitative data on the long-term effects of an HIV prevention program in South Africa, the CHC. The findings provide preliminary evidence that couples who received the CHC identified value in relationship-based interventions and experienced sustained impact to their relationships over the longer term. The nuanced descriptive data acquired through our study would not be possible with other research designs (e.g., quantitative investigations). We do, however, recommend future investigators conduct follow-up studies with controlled designs (qualitative and quantitative) to strengthen the evidence base for the long-term effects of couple-based HIV prevention and treatment programs. However, one of the key findings points to the ongoing difficulties in the community related to alcohol use, which complicates issues related to HIV prevention, individual well-being, relationship functioning, and community wellness. This finding highlights the need to expand the availability of evidence-based interventions at the individual, couple, family, community, and policy levels. The use of couple-based interventions is one strategy that has the potential to support individual behavior change in areas that are often more difficult to change, such as alcohol use.

Conclusion

Overall, we have provided preliminary qualitative data on the long-term efficacy of a couple-based HIV-risk behavior intervention. We add to a growing body of evidence that shows couple-based interventions for HIV prevention and treatment are both desirable and effective. Researchers and policy makers need to work together to translate this knowledge into practice by making evidence-based couples interventions widely available for those at-risk or already living with HIV.

Key Considerations.

• Intimate partners should be incorporated into HIV prevention and treatment efforts where appropriate.

• Skills can be taught to couples/other dyads to enhance communication and help couples/other dyads work more effectively together in the context of HIV prevention and treatment.

• Providers working with individuals living with HIV or those at risk of HIV should assess for the presence of alcohol use and alcohol-related problems. Providers should consider environmental factors (e.g., availability of alcohol, prevalence of use) when helping individuals who are trying to reduce alcohol use.