Natasha Martin, DPhil, professor and vice chief of Global Public Health in the Division of Infectious Diseases and Global Public Health at UC San Diego School of Medicine, has been awarded a five-year, $5.6 million Avant-Garde Award from the National Institute on Drug Abuse (NIDA), one of the most prestigious and competitive NIH research awards. The funding will support the development of innovative artificial intelligence tools designed to improve public health responses to Human Immunodeficiency Virus (HIV), hepatitis C virus (HCV) and overdose among people who use drugs across the United States.
For too long, local intervention preferences of people who use drugs have been absent in the data that guide responses to HIV, hepatitis C and overdose. This project will combine community perspectives with cutting-edge AI tools to help health departments better understand local needs, allocate resources more effectively and ultimately save lives. The project is called 'AMPLIFY,' because it is designed to amplify the voices of people who use drugs and ensure their priorities shape the interventions delivered in their communities."
Natasha Martin, DPhil, professor and vice chief of Global Public Health, Division of Infectious Diseases and Global Public Health, UC San Diego School of Medicine
The NIDA Avant-Garde Award supports exceptionally creative scientists pursuing bold, high-impact research with the potential to transform HIV and substance use disorder research. The award is given to only one to three researchers each year and is intended to support innovative ideas that could open entirely new areas of scientific inquiry.
Bringing AI to public health decision-making
Despite major advances in HIV prevention and treatment, including long-acting therapies, the U.S. continues to face persistent challenges in preventing HIV, HCV and overdose among people who use drugs. One obstacle is that public health agencies often lack timely information about the preferences, needs and behaviors of the populations they serve, particularly in communities with limited surveillance and research infrastructure.
Martin's project aims to address that gap by creating realistic, community-informed "digital twins" - AI-powered simulations that reflect the intervention preferences and decision-making patterns of people who use drugs. Using data from multiple cohorts and a community-based participatory approach, the research team will train large language models to generate digital twins that can represent people in different regions and circumstances across the country.
The digital twins will then be integrated into sophisticated models that simulate HIV and HCV transmission, overdose and related health outcomes. Researchers will use these models to test different prevention and treatment strategies before they are implemented in the real world, helping public health leaders identify approaches that deliver the greatest benefit at the lowest cost for their communities.
"Applying AI to predict preferences of people who use drugs could be transformative," Martin said. "Beyond this project, this approach could help design new health services, identify barriers to care and tailor interventions for populations whose needs are poorly understood."
From research to real-world impact
The project builds on years of collaboration between UC San Diego researchers and public health agencies, including work conducted through the Resilient Shield initiative, a Centers for Disease Control-funded center, led by Martin and Eliah Aronoff-Spencer, MD, PhD, professor of medicine at UC San Diego School of Medicine, that provides outbreak analytics and modeling support to health departments. Through that work, researchers observed that strategies successful in one location did not always translate effectively to another because intervention preferences differed, highlighting the need for more personalized and locally informed approaches.
A central component of the new project is engaging people with lived and living experience of substance use to help guide the development of the AI models. Community members will play a key role in shaping how the digital twins are designed, tested and applied, helping establish safeguards, build trust and ensure the technology addresses the needs and priorities of the communities it is intended to serve.
The team will ultimately develop an interactive dashboard that public health departments can use to evaluate prevention strategies, assess resource allocation options and respond more effectively to emerging health challenges. By providing rapid, data-driven insights, the platform is designed to help decision-makers direct limited resources where they can achieve the greatest impact.
In addition to Martin, the project includes UC San Diego collaborators: Ravi Goyal, PhD; Eli Aronoff-Spencer, MD, PhD; Steffanie Strathdee, PhD; Annick Bórquez, PhD; Laramie Smith, PhD; Britt Skaathun, PhD; and John Ayers, PhD, all of the Division of Infectious Diseases and Global Public Health at UC San Diego School of Medicine, and Angela Bazzi, PhD, of the Herbert Wertheim School of Public Health and Human Longevity Science.
New study tracks HIV care after hospital discharge statewide