The SASA! Study: An Evaluation of a Community Intervention to Address Gender-based Violence and Reduce HIV/AIDS Risk in Uganda (SASA!)

Background : Physical and sexual violence against women is widespread in Uganda, and is increasingly recognized as a major global social, public health and human rights problem. Epidemiological evidence demonstrates that violence may also be intimately tied in with HIV risk, with violence being both a risk factor for and a consequence of HIV infection.Behaviours that often form the cornerstone of HIV prevention messages such as refusal of sex, inquiring about other partners, or suggesting condom use, have all been identified as triggers of intimate partner violence (IPV) in various settings.

There is now growing interest in integrated HIV/violence prevention strategies to tackle the underlying notions of masculinity that condone both male infidelity and male control over women. Promising research from South Africa and Brazil suggests that interventions which explicitly aim to challenge gender inequalities (either through economic empowerment of women or discussion with men and women about gender inequalities, norms and their implications), may influence levels of violence and HIV risk behaviours. However, such interventions remain few in number, as do scientifically rigorous evaluations of them.

Intervention Design :

SASA! uses a community mobilization approach to address gender inequity and try to change attitudes and behaviours that support both HIV risk behaviours and the perpetration of violence against women.

SASA! supports communities through a four-phase process of social change (based on a social-level adaptation of the Stages of Change Theory):

1. Start to think about violence against women and HIV/AIDS as interconnected issues and the need to personally address these issues
2. Awareness raising about how communities accept men's use of power over women, fueling violence against women and HIV/AIDS
3. Support women and men directly affected by/involved in these issues to change
4. Action to prevent violence against women and HIV/AIDS

The intervention team engages with four major groups of actors: community volunteers selected from the general public; community leaders (e.g. religious, cultural and local council leaders); resource persons (health care providers, police, etc); and institutional leaders. The community volunteers are key agents to raise awareness about power imbalances between men and women, discuss the impacts of HIV and violence against women, discuss the benefits of non-violence and gender equity, and most importantly to mobilize others in the community to take action thereby promoting sustainable change.

Study design:

The SASA! Study randomized four pairs of communities that were matched on population density and stability/transience. The intervention communities received the SASA! intervention from early 2008. Comparison communities will receive the intervention approximately three years later.

A cross-sectional survey will be conducted among community members prior to the implementation of SASA! in their communities. Another cross-sectional survey was initially planned for three years post-implementation, but later delayed to four years post-implementation due to political disturbances that caused interruptions to programme activity. The primary outcomes are: past year experience of physical IPV (among partnered women); past year experience of sexual IPV (among partnered women); acceptability of violence against women; acceptability of a woman refusing sex; community response to women experiencing violence (among women reporting past year physical and/or sexual IPV); and past year concurrent sexual partners (among partnered men). Data analysis will be appropriate to the cluster-randomized study design. The target sample size per arm at each round of data collection is 800.

Complementary qualitative and quantitative data will also be collected (from community volunteers and resource people as well as from community members) to document processes of change and potential pathways of impact.