The Complexities of a Definitive Answer: The Case of the Tathmini GBV Evaluation

Speeches Shim

A group of young African women dance together. Photo credit: Tash McCarroll/USAID
According to the U.S. Centers for Disease Control and Prevention Violence Against Children Surveys, nearly a third of girls in Kenya, Swaziland, Tanzania and Zimbabwe reported sexual violence before the age of 18.
Photo credit: Tash McCarroll/USAID
headshot of Amelia Peltz

Amelia Peltz, MA, senior gender advisor, Office of HIV/AIDS

Earlier this month, the global community recognized International Women’s Day. It is one of those occasions that brings out mixed emotions. It is a moment to celebrate the gains made in the struggle for gender equality, to reflect on new developments and directions that inspire us, and to renew calls for action on issues that continue to dominate our lives.

And there is no issue more detrimental than gender-based violence.

The numbers speak for themselves. Global estimates published by the World Health Organization indicate that about 1 in 3 (35 percent) of women worldwide have experienced either physical and/or sexual intimate partner violence or non-partner sexual violence in their lifetime.

And for too many girls, sexual violence features far too prominently in their young lives. According to the U.S. Centers for Disease Control and Prevention Violence Against Children Surveys, nearly a third of girls in Kenya, Swaziland, Tanzania and Zimbabwe reported sexual violence before the age of 18. And perhaps most strikingly, research noted in the Greentree II Report Violence against Women and Girls, and HIV, shows that women who experience intimate partner violence have a 50 percent higher risk of contracting HIV.   

Clearly, the need for HIV service delivery platforms to prevent and respond to gender-based violence – in particular intimate partner violence and violence against children – is urgent.

The U.S. President’s Emergency Plan for AIDS Relief (PEPFAR) Gender-Based Violence Initiative (GBVI) aimed to do just that: to demonstrate that gender-based violence prevention and response activities could be integrated into existing HIV platforms.  GBVI also aimed to understand the best approaches for achieving a successful integrated response through existing HIV prevention, care and treatment services  using a complementary set of multi-level interventions across many different sectors. Over the course of three years, GBVI interventions and services reached more than 1.3 million individuals across the Democratic Republic of the Congo, Mozambique and Tanzania; clinic- and community-level services included gender-based violence screening, post-gender-based violence care, provision of HIV post-exposure prophylaxis, and HIV/gender-based violence prevention programs that aimed to promote gender equity in communities. Efforts also included the development of national guidelines, training programs and resources (such as gender-based violence screening tools) and the standardization of procedures, protocols, and frameworks for integrating gender-based violence into HIV services.

Complementing the program in Tanzania was an evaluation called Tathmini GBV [PDF, 803KB]. Designed under USAID’s Project SEARCH and completed under Project SOAR, it set out to definitively answer the “did it work?” question using a rigorous evaluation design. The full engagement of policy, program and community stakeholders was essential to answering the evaluation’s two central questions: Did the comprehensive gender-based violence program lead to 1) increased care for gender-based violence survivors; and 2) changes in community-level attitudes and behavior toward violence.

So…did the GBVI work?




The results were disappointing, interesting and hopeful all at once.

Perhaps most disappointing was that the program had only small effects on women’s reported experience of intimate partner violence and other forms of gender-based violence. Furthermore, exposure to and participation in the gender-based violence program was quite low, possibly explaining the lack of program impact on intimate partner violence and gender-based violence prevalence. As disappointing as these findings were, our disappointment was tempered by the fact that, although not statistically significant, the program effects were in the intended direction. Perhaps the most interesting finding was that over time, significant declines in women’s reported intimate partner violence experiences were found among both the control and intervention clusters. Attempts to pin down exactly why this was the case were unsuccessful, but it was most likely due to unexplained factors outside the program.

Despite these lackluster results, there were other findings that inspired hope and demonstrated that change is possible. The evaluation was able to show that the comprehensive program created a decline in women’s acceptance of intimate partner violence for refusing to have sex with her partner; a shift to more gender equitable norms among women; more widespread knowledge of gender-based violence and informed beliefs about violence against children; increased likelihood that women would take action to intervene in gender-based violence cases they encounter; and greater community-wide responses to gender-based violence including on the part of local community leaders.

So, did it work? As with many evaluated public health interventions trying to address complex problems, such as gender-based violence and HIV, sometimes the most interesting answers are the ones that illuminate how the “real world” intersects with our carefully crafted interventions. In the case of Tathmini GBV, while at first the mixed results left us disappointed, it also led to rich discussions about what other factors contributed to a reduction in reported experience of violence, as well as other behavior and attitude changes among women. While meaningful changes were measured across both intervention and control communities, particularly with the noted changes in attitudes and behavior, the evaluation prodded us to seek greater understanding about the process of change at a community level and what is needed to sustain long-term social change. Such insights have the potential to help us deliver more effective interventions.

Last updated: September 16, 2019

Share This Page