The Power and Process of Shifting Gender Norms: Insights from a Randomized Controlled Trial in South Africa

"The study demonstrated that shifts in views toward gender can be fostered in a meaningful way at the community level, and that such shifts may also provide an important boost at the individual level. Further, decreases in IPV [intimate partner violence] could be a path to testing for women and men."

In the global HIV field, policy and programme discussions increasingly recognise the importance of addressing harmful gender norms across the HIV care continuum. This Project SOAR (Supporting Operational AIDS Research) report summarises findings from a sub-study of a randomised controlled trial (RCT) of "Tsima ra rihanyu" ("working together for health"), a three-year (2015 to 2018) community mobilisation programme for treatment as prevention (CM for TasP) that aimed to address the social barriers to HIV testing and treatment in Mpumalanga Province, South Africa. Results from the main trial are under review (Lippman et al.), and in this report, the authors explore the types of gender norms that are salient for women and men, how these relate to HIV testing and treatment, and the pathways through which gender norms operate to affect women's and men's HIV service use and experience and perpetration of intimate partner violence (IPV).

The study used a cluster randomised controlled design with eight intervention and seven control communities. Findings presented in this report come from three sources: pre-post cross-sectional surveys conducted before the three-year intervention began and just after it ended, in all intervention and control communities; qualitative data gathered at three time points over the course of the study (in the beginning, toward the end, and after preliminary analysis of endline results); and a costing analysis.

The intervention was implemented by Sonke Gender Justice, a South African non-governmental organisation. Described in detail elsewhere (see Related Summaries, below)), Tsima is based on the theory that the barriers to broad use of treatment as prevention are primarily social and structural, and that to create social change requires CM. Over the course of the three-year intervention, the team conducted over 9,000 activities and 1,000 community action team (CAT) meetings in the eight intervention villages. Over 8,000 of the activities were outreach/small group activities such as door-to-door education, mural painting, open houses, and health talks. Other activities included: 35 young women's groups that met approximately nine times each and reached 421 young women; 591 leadership/stakeholder engagement meetings; 46 events such as soccer tournaments, men's events, and fun days; and 373 two-day workshops.

Survey participants were men and women ages 18-49 randomly selected from intervention and control villages (final evaluation samples were 1,149 baseline [2014] and 1,189 endline [2019]). Endorsement of equitable/inequitable gender norms was measured using an adapted version of the GEM Scale (which measures attitudes towards gender norms), including four subdimensions: (i) norms condoning men's violence and control over women; (ii) norms around men as decision-maker in the couple; (iii) norms around men's toughness and avoidance of help seeking; and (iv) norms around women's primary responsibility as family caretaker. The qualitative component involved in-depth interviews (IDIs) and focus group discussions (FGDs) conducted with community members, community leaders, and Tsima community mobilisers and CAT members.

Key findings:

• The four subdimensions of inequitable gender norms differed in terms of how strongly and widely they were endorsed. For the baseline survey, responses to the subdimension regarding women's primary responsibility as family caretaker reflected the most inequitable views (mean of 1.80 on a range of 1.0 to 3.0, for both men and women - with lower scores representing less equitable views). The subdimensions relating to men's violence/control and men as decision-makers in a couple also reflected endorsement of inequitable norms. In contrast, responses to the subdimension around men's toughness and avoidance of health-seeking were the most equitable (mean score of 2.51 for women and 2.47 for men).
• Endorsement of equitable gender norms was associated with HIV testing and treatment in various ways. Analysis of baseline data showed that for women living with HIV, more equitable views on gender norms was significantly associated with increased odds of reported current antiretroviral treatment (ART) use for the composite GEM Scale, as were three out of four gender norm subdimensions: men's violence and control, men as decision-makers in the couple, and men's toughness and avoidance of health seeking. Among men living with HIV, equitable scores regarding the subdimension of men as decision-makers in the couple were significantly associated with increased odds of treatment use. Among women, more equitable views on gender norms decreased the odds of reported HIV testing - this finding was significant for the composite score, as well as norms condoning men's violence and women's primary responsibility as family caretaker. Men with more equitable scores regarding condoning men's violence and control over women were less likely to report an HIV test in the past year.
• There were broad, substantial shifts toward more equitable gender norms across control and intervention communities over the study period. Women's and men's composite GEM Scale score increased (i.e., became more equitable) from baseline to endline in multivariate analysis, and women's and men's scores on each of the four subdimensions also significantly increased. However, there were no significant differences between study arms, and the change in participants' gender views cannot be attributed to the intervention. Qualitative findings confirmed a shift in gender norms. The round 3 interviews and FGDs, which were conducted after the endline, suggested that the broad shifts may be due in part to rapidly increasing access to TV (via satellite dish) and smartphones across the study area (as elsewhere in South Africa), exposing more people to serial dramas and media that negatively portray IPV and model couples communicating equitably and resolving conflicts.
• There are diverse pathways through which changes in gender norms affected IPV and HIV testing and treatment. The intervention led to a significant decrease in women's experience of IPV. Men's reports of IPV perpetration also decreased, but did so in both intervention and control communities. Qualitative data indicate that Tsima encouraged and built skills around more equitable, constructive communication between intimate partners. Such shifts helped men find alternatives to violence and facilitated treatment uptake and retention. New knowledge about the benefits of TasP was more actionable, for example, because of reduced endorsement of inequitable norms regarding men's health seeking and men's toughness.
• Attitudes and behaviours appeared to be more difficult to influence among older men (i.e., ages 30-49). Older men were less likely than younger men to participate in Tsima, experienced a smaller magnitude increase in their GEM Scale score in intervention vs. control communities (although at endline, scores were higher, i.e., more equitable, in both), and did not differ over time or between intervention and control communities in terms of reductions in IPV perpetration.
• The estimated total programmatic cost of Tsima was US$429,000 for the three-year period, or US$143,000 per year, and the average cost was US$4.68 per contact.

Based on the findings, recommendations for those implementing gender-transformative work to address HIV and IPV include:

1. Conduct research to determine the extent to which different gender norms influence what - and whose - behaviour, and tailor interventions accordingly.
2. Combine activities to shift gender norms with skill building to ensure translation into practice, such as more equitable couple communication and conflict negotiation as a way to prevent IPV and encourage HIV testing and treatment.
3. Simultaneously address gender-related barriers to HIV service use, such as gender norms and IPV that prevent disclosure and treatment seeking, while providing concrete information about the benefits of early testing and treatment and TasP.
4. Explore and address possible impediments to service utilisation that may not always be gendered, such as clinic hours, wait time, and concerns about confidentiality.
5. Conduct additional studies to explore implementation questions, such as optimal exposure and length of interventions to change gender norms and efficient combinations of intervention components.
6. Follow up with participants over the longer term to observe possible additional/ripple effects of interventions like Tsima on, for example, new HIV infections and retention in care.
7. Pay greater attention in research, strategic planning, and intervention development to critical, broader, social changes underway in a programme's setting - including rapidly expanding media and internet access across sub-Saharan Africa - and how these changes can be leveraged to improve health behaviours.