Implementation of couples’ voluntary HIV counseling and testing services in Durban, South Africa

Almost two-thirds of incident HIV infections in sub-Saharan Africa occur among stable concordant HIV negative or HIV discordant (one partner testing HIV seropositive and one partner testing HIV seronegative) couples [7]. In an analysis of Demographic and Health Surveys data from 24 sub-Saharan African countries, HIV incidence among concordant HIV negative couples contributes to roughly 30 % of all new infections while incidence among discordant couples contributes another 30 % [7]. About half of all HIV-infected persons in the region are in discordant relationships [13], and the majority of transmissions between serodiscordant couples originate with the HIV-positive index, not concurrent, partners [6].

Couples’ voluntary HIV counseling and testing (CVCT) is associated with a two-thirds reduction in incidence in discordant couples, reduced incidence of new infections in concordant HIV negative couples, and reductions in sexually transmitted infection (STI), mother-to-child transmission (MTCT), and unplanned pregnancies for all couples [3–5, 10, 15, 17, 18, 20, 24].

In South Africa, most HIV is transmitted heterosexually. Nation-wide HIV prevalence among persons aged 15–49 is estimated at 23.3 % for women and 13.3 % for men and varies considerably by region. KwaZulu-Natal has the highest HIV prevalence among those aged 15–49, estimated at 27.6 % by the 2012 HSRC survey [33]. Additionally, the Partners in Prevention HSV-2/HIV-1 Transmission study group reported a very high prevalence of serodiscordance in South African couples of 27.4 %, much higher than other countries in the region [28]. A recent study in Soweto (in the Johannesburg metropolitan area) comprised of 11,163 sexually-active HIV-infected men and women in an urban primary care program indicated that 40 % of participants had a known positive partner, while 40 % of participants had a partner of unknown HIV status and 20 % had an HIV-negative partner. Participants in this study had been in their relationships for a substantial period of time (median >3 years), and knowledge of partner HIV status did not vary by the duration of relationships [38].

These high rates of HIV prevalence and HIV serodiscordance reinforce WHO recommendations for couples-based HIV testing, which is not yet widely available in South Africa. The Rwanda Zambia HIV Research Group (RZHRG) has been providing CVCT services in Rwanda and Zambia for over 20 years and has provided support and training for CVCT under a CDC Centers of Excellence grant to over 20 countries, including South Africa [1, 32]. RZHRG has a successful history of recruiting couples for CVCT and building capacity within local communities to provide the service [2, 27, 39].

The implementation of CVCT was recently piloted in five health facilities in Durban, South Africa. During this pilot, RZHRG provided on-site support, training, and technical assistance to health facility staff; trained local clinic promotions staff to mobilize their communities to attend CVCT; assisted with implementation of the service; and helped monitor and evaluate patient-level data including testing outcomes, antiretroviral treatment (ART) use, and pregnancy status. Here, we present the implementation methods as well as CVCT client-level outcomes among couples receiving CVCT in Durban during this pilot. The “pilot” refers to all activities related to CVCT promotions and service provision.

This is the first publication reporting the results of implementing the RZHRG-developed CVCT intervention in an African country outside of Rwanda or Zambia, where the intervention originated and is now a standard of care. Using standardized implementation developed by RZHRG and adapted with the CDC, we were able to quickly and successfully increase capacity for CVCT in Durban. Both counselors and promoters achieved a high level of success in the trainings and in the implementation of CVCT promotion and provision. Each counselor provided services to 7–8 couples/day which was considered to be a feasible workload. The project was well received by the community and health authorities, who saw the importance of encouraging and offering HIV testing to couples and not just individuals.

Though there has been debate about the impact of couples’ HIV testing specifically in South Africa—a country where multiple concurrent partnerships were hypothesized to be a driving factor in the epidemic—a recent population-based data from KwaZulu-Natal suggests that prevalence of multiple or concurrent partnerships among men is currently decreasing [28] and that concurrent partnerships are not an important driver of high HIV incidence in this area [36]. A recent publication of a new randomized controlled trial protocol which aims to increase uptake of CVCT via a group and couple counseling intervention highlights the growing interest in couples-focused HIV prevention. This is the first couples-based intervention with CVCT uptake as its outcome, the results of which may bolster CVCT in South Africa [9]. Couples-focused HIV testing and counseling interventions have also been found acceptable among men who have sex with men in South Africa [33–35]. Additionally, the interest in couples-focused interventions in South Africa is illustrated by the recent results of a pilot study of Project Connect, a HIV/STI prevention intervention developed for US minority couples [11, 12]. Project Connect found that intervention is acceptable among couples in Johannesburg, though not entirely feasible due to stringent eligibility criteria based on previous history of intimate partner violence (IPV) and the number of counseling sessions involved [29]. The investigators of that study recommend that IPV not be an exclusion criterion and that the number of sessions be reduced in order to provide feasible couples-based HIV services. We support both of these recommendations. IPV is a discussion topic in the CVCT trainings as is emphasis on the voluntary nature of the CVCT service.

The promoters were at liberty to give CVCT invitation to individual/couple they encountered in the community regardless of which neighborhood or part of Umlazi the couple came from. The promoter training did not prescribe the areas for invitations to be given but rather put emphasis on strategies for successful invitation. As in our previous studies of CVCT promotions [2, 26, 38, 39], promoters would typically invite individuals and couples within their social network.

Our client-level findings highlight the need for couple-focused HIV prevention services. We found a very high HIV prevalence (41.8 %), higher than the adult HIV prevalence estimates published in the 2012 HSRC [32]. The high prevalence of couple serodiscordance (29.5 %) was very similar to that found by the Partners in Prevention HSV-2/HIV-1 Transmission study group in this region (27.4 %) [27]. It was much higher than other countries in sub-Saharan Africa. Additionally, we found that HIV prevalence and serodiscordance differed by CVCT test site. These differences are likely explained in part by age distribution, with higher HIV prevalence and discordance found in clinics attending more couples in the 25–34 age group. Though all testing sites were located in the same general area, these differences in HIV testing outcomes can occur and implementers should be aware of this possibility. The number of couples tested also differed by site; the location of Gateway clinic on the grounds of the PMMH may have contributed to seeing the largest number of couples. A client-level measure of length of relationship was not assessed in this pilot study. This would likely be a meaningful measure especially given that 67 % of couples were not cohabiting. For a study of this nature in future, length of relationship will be useful and should be collected.

For several years, many researchers, clinicians, and sub-Saharan Africa country Ministries of Health have expressed concern regarding HIV risk and age-disparate relationships [19, 40]. We found in this study that the age disparity between couples was greatest for concordant HIV positive couples. However, due to the cross-sectional nature of our study, the temporality of the relationship between age-disparity and HIV status cannot be established. Other cross-sectional studies in Kenya, Uganda, Zimbabwe, and South Africa found that women with older partners were more likely to be HIV positive or at risk for HIV [16, 22, 24, 30]. In a population-based open cohort of 2,444 young women in KwaZulu-Natal, age disparity with their most recent sexual partner did not predict HIV acquisition among women [19]. Partner serostatus was not included in that analysis.

Previous individual HIV testing was low (50 % for men, 63 % for women). Compared to HIV positive non-pregnant women, HIV positive pregnant women were not more likely to have previously been tested. Most HIV positive persons were not on ART, including HIV positive pregnant women who had previously tested for HIV. As CVCT has been shown to increase uptake of ART among HIV positive pregnant women (PMTCT Prong 3) [14], this intervention could serve as an effective entry point to care and treatment services aimed at reducing both vertical and perinatal HIV transmission.

This pilot study compliments the GHRI-funded Canadian Sub-Saharan Africa HIV/AIDS Network, which aims to build multi-disciplinary capacity for HIV prevention research and clinical trials at seven sites in five African nations. The combination of clinic-based skills development, sociological research, and community engagement significantly enhanced Network activities that are similarly designed to use collaborative approaches to support ‘practice-based’ knowledge and skills training. Notably, this project used African experts to train Africans, an efficient South-South strategy that promotes leadership skills and learning opportunities for young African investigators. Since the Network’s focus is to build capacity by linking clinical, biomedical, and sociological researchers in Africa, it has not initiated efforts to engage HIV counselors and community volunteers or to assess intervention strategies; however, it is well situated to disseminate ‘proven’ methods to other sites in South Africa and Lesotho, who struggle with similarly devastating HIV epidemics.

Implementation of standard CVCT services was found to be feasible and acceptable in Durban clinics, where burden of HIV and couple serodiscordance was found to be extremely high, using a South-South strategy. In a representative subset of the couples tested, 70 % did not know discordance was possible prior to pre-test counseling [25], and 29 % proved to be discordant couples. Long-term follow-up of couples would help to assess the effectiveness of this strategy over a period of time. The cohort of highly skilled South African counselors, researchers, and community workers generated by this project will provide an excellent local resource for continued CVCT efforts in the region.