The Project
In Brazil, a striking majority of gender-based violence (GBV) remains uncounted and unseen. Brazilian public health systems data indicates that only 13% of GBV records come from primary care units. Mostly, the notification of suspect cases only happens when the victim is admitted to a hospital and the record fulfillment is mandatory for health professionals.
It means the health services are missing opportunities to address violence against women earlier!
Vital Strategies is aware of the fact that health systems can play a fundamental role in preventing and responding to gender-based violence. We partnered with FrameNet Brasil, a Natural Language Processing lab at the Federal University of Juiz de Fora, to pioneer a project that combines AI and expertise in public health surveillance.
The project uses linguistic analysis methods to read and map patterns in the open-text fields of electronic medical records filled out in health units. This information, which adds a more subjective factor to the analysis by including observations made by health professionals at the time of care, are cross-checked with other databases: violence reporting, hospitalization, and death systems.
By identifying correlations within the data and by crossing records of the same person through different systems (using methodologies that guarantee anonymization and privacy), it is possible to identify risk factors for violence and patterns that suggest women registered in the health systems are candidate victims of violence.
The goal is to create tools for early warning and intervention that better equip health teams, local authorities and policymakers to design and apply best practices and public policies for identifying and responding to potential cases of GBV. To ensure the tools will work according to the government’s structure to assist victims with the service network, the initiative counts on the partnership with public health authorities in Brazil to build its usage protocols.
When GBV is identified early, it is possible to ensure the root causes of the problem are addressed and the cycle of gender-based violence can be broken before it is too late.
Read more about the project:
- Blog: Leveraging AI to Help Prevent Gender-Based Violence in Brazil
- Clinton Global initiative Commitment to Action | Innovation in Data Analysis to Prevent Gender-based violence
- Paper: Building a Frame-Semantic Model of the Healthcare Domain: Towards the identification of gender-based violence in public health dataing tasks
- Paper: Evaluating the contribution of framenet to gender-based violence identification – How semantic annotation can be used as a resource for identifying patterns of violence
About us
Vital Strategies believes every person should be protected by an equitable and effective public health system. We partner with governments, communities and organizations around the world to reimagine public health so that health is supported in all the places we live, work and play. The result is millions of people living longer, healthier lives.
Learn more:
For more information, please visit: vitalstrategies.org and our capability statement on Gender-Based Violence.
Contact
Email: vsbrazil@vitalstrategies.org
Meet the team
Vital Strategies
Pedro de Paula
Country Director at Vital Strategies
Sofia Reinach
Senior Project Manager for Violence Prevention and Public Health Systems at Vital Strategies
Renato Teixeira
Technical Advisor – Data Scientist and Epidemiologist at Vital Strategies
Fatima Marinho
Principal Technical Advisor at Vital Strategies
Olívia Guaranha
Program Officer for Violence Prevention and Public Health Systems at Vital Strategies
Luiza Borges
Institutional Communications Senior Manager at Vital Strategies
Nayme Bizaio
Institutional Communications Senior Manager at Vital Strategies
Arthur Lorenzi
PhD Candidate in Linguistics at the Federal University of Juiz de Fora. Consultant at Vital Strategies Brasil
FrameNet Brasil
Tiago Torrent
Linguist, PI at the FrameNet Brasil Lab
Lívia Dutra
Masters student in Language Technology at Gothenburg University
Fred Belcavello
Linguist, Researcher at the FrameNet Brasil Lab
Ely Matos
Computational Linguist, Researcher at the FrameNet Brasil Lab
Marcelo Viridiano
Designer, PhD Candidate in Linguistics at the Federal University of Juiz de Fora
Lorena Tasca
Masters student in Linguistics at the Federal University of Juiz de Fora
Amanda Pestana
Undergraduate student researcher at the Federal University of Juiz de Fora
Kenneth Brown
Undergraduate student researcher at the Federal University of Juiz de Fora
Mariana Machado
Undergraduate student researcher at the Federal University of Juiz de Fora
Victor Herbst
Undergraduate student researcher at the Federal University of Juiz de Fora