The SASA study: a cluster randomised controlled trial to assess the impact of a violence and HIV prevention programme in Kampala, Uganda

Publication Details

Watts, C, Devries, K, Kiss, L, Abramsky, T, Kyegombe, N and Michau, L, 2014. The SASA! study: a cluster randomised trial to assess the impact of a violence and HIV prevention programme in Kampala, Uganda, 3ie Impact Evaluation Report 24. New Delhi: International Initiative for Impact Evaluation (3ie)

Link to Source
Charlotte Watts, Karen Devries, Ligia Kiss, Tanya Abramsky, Nambusi Kyegombe, Lori Michau
Institutional affiliations
None specified
Grant-holding institution
None specified
Sub-Saharan Africa (includes East and West Africa)
Health Nutrition and Population
HIV and AIDS, Sexual Behavior
Gender analysis
HIV and AIDS, Sexual Behavior
Gender analysis
Equity Focus
Evaluation design
Randomised Control Trials (RCT)
3ie Series Report
3ie Funding Window
Open Window Round 3


The study aims to assess the impact of an intervention on gender norms, and past-year levels of physical and sexual violence against women, HIV risk behaviours and community responses to violence.


Violence against women is increasingly recognised as an important social, development, health and human rights problem. In Uganda, 59 per cent of women aged 15-49 who have ever been married, have been physically or sexually abused by their partners. Recent analysis also suggests that women who have experienced violence face twice the odds of HIV infection compared to women who have not experienced violence. Recognition of this scale of violence has highlighted the urgent need for HIV programmes that address violence against women, and the underlying causes of gender inequality. 

This evaluation aims to measure the impact and effectiveness of SASA! - an Activist Kit for Preventing Violence against Women and HIV. The SASA! approach aims to prevent violence against women and HIV by focusing on the root cause of women's vulnerability to both: the imbalance of power in intimate relationships. The intervention aims to mobilise communities to reassess their acceptability of violence and gender inequality. It supports the development of community activists, and engages with key stakeholders, including local and religious leaders, the police and health workers. 

The impact evaluation of this intervention is a randomised controlled trial that uses both a quantitative and a qualitative approach. The evaluation explore the impact of the intervention on gender-related norms and behaviours and on the community’s responses to incidents of violence against women.

Website :


Intervention design

As outlined in the SASA! trial protocol, the trial has six primary outcomes: attitudes towards the acceptability of violence against women and the acceptability of a women refusing sex (among female and male community members); past year experience of physical intimate partner violence and past year experience of sexual partner violence (among women); appropriate community response to women experiencing violence (among women reporting past year physical or sexual partner violence); and past year concurrency of sexual partners (reported by men with a primary partner).

Theory of change

The SASA! approach aims to prevent violence against women and HIV by focusing on the root cause of women's vulnerability. The intervention aims to mobilise communities to reassess the acceptability of violence and gender inequality, to recognise the links between these issues and HIV and AIDS, to take effective action to change their own relationships, and to support women experiencing violence or HIV and AIDS.

Evaluation design

The SASA! study was conducted between 2008 and 2012 in two administrative divisions of Kampala (Makindye and Rubaga). The trial involved a pair matched cluster randomised trial with randomisation conducted within matched community pairs. Four communities were randomly selected to receive the full intervention from baseline and four communities wait-listed to receive the intervention at the end of the trial. Two cross-sectional surveys were conducted, one at baseline before implementation of the survey (2008) and one four years later (2012). A total of 1583 respondents were interviewed at baseline (717 women, 866 men) and 2532 respondents were interviewed at follow-up (1130 women and 1402 men). 

Main findings

SASA! reduced the reported social acceptance of physical violence in relationships among both women (AOR 0.54 (0.37 – 0.80) and men (AOR 0.13 (1.01 – 1.14)), and also increased the social acceptance of the belief that there are circumstances when a women can refuse sex from her partner (AOR 1.31 (1.01 – 1.70) for women, 1.28 (1.06 – 1.54) for men). The levels of physical partner violence occurring in the past year reported by women were 52 per cent lower in the SASA! intervention communities compared to the control (AOR 0.48 (0.16 – 1.39)). Women in intervention communities who had experienced violence were over twice as likely as women in control communities to report that they experienced a supportive community response, although this was not significant due to small numbers in each cluster. The findings also suggest that SASA! impacted significantly on the reported levels of sexual concurrency, with 27 per cent of partnered men in intervention communities reporting having had other sexual partners in the past year, compared to 45 per cent of men in the control communities (AOR 0.57 (0.36 – 0.89). 

These quantitative findings are strongly supported by the qualitative findings.

Additional publications

Media coverage



Scroll to Top