Tulane researchers seek to advance cardiovascular disease prevention through artificial intelligence and community-based care

Tulane University researchers are developing an initiative that uses artificial intelligence to support cardiovascular disease prevention by combining digital tools with community-based care.

Known as Heartline, the initiative brings together AI, community health workers and faith-based organizations to help individuals make sustainable lifestyle changes.  

The project is led by Farah Allouch, assistant professor at the Celia Scott Weatherhead School of Public Health and Tropical Medicine, and Nicholas Mattei, associate professor at the School of Science and Engineering. It builds on their years of experience working in community-centered health interventions and AI development.  

The project recently received support through a joint seed grant from the Connolly Alexander Institute for Data Science and the Celia Scott Weatherhead School of Public Health and Tropical Medicine. The seed grant is focused on advancing AI and machine learning in public health research.

The concept for Heartline emerged from Allouch’s work during the past four years on a church-based cardiovascular disease prevention trial in New Orleans.

“I’ve been managing community health workers directly and seeing the impact their health education and care management approach has had on our participants in the study,” Allouch said.

While participating in the cardiovascular disease prevention trial, Allouch developed an interest in AI and began considering how these approaches could intersect. This ultimately led to the creation of Heartline.  

Heartline focuses on a persistent challenge in cardiovascular care: helping individuals adopt and maintain lifestyle changes. In the church-based trial, approximately 80% of participants were hesitant to begin medication, even when it was clinically recommended. This resistance often reflects a broader concern about how treatment is approached.  

“A lot of the time, people feel like medications are just putting a Band-Aid on the problem,” Allouch said.

In response, Heartline is designed to make lifestyle-based interventions more practical, personalized and accessible. Heartline features a digital platform intended to provide real-time, practical support for users as they make everyday health decisions. Rather than offering only general recommendations, the system can generate personalized guidance aligned with established dietary patterns that are proven to reduce the risk of cardiovascular disease.  

“You can say, [as an AI prompt] ‘Give me a weeklong grocery list for a family of two that aligns with the DASH diet or the Mediterranean diet,’ which are the two diets with strong evidence supporting their role in reducing cardiovascular disease,” Allouch said.

By translating clinical guidance into concrete, day-to-day actions, the tool aims to reduce barriers that often prevent individuals from following through with lifestyle changes.

Heartline also uses retrieval-augmented generation (RAG) — an AI framework that boosts accuracy — to ensure that responses are grounded in curated, trusted materials provided by the research team. RAG improves reliability and limits “AI hallucinations.”

Rather than replacing clinicians or community health workers, the Heartline platform is designed to enhance the impact of these professionals by supporting patients between visits and offering them guidance in real time.

Community engagement is a core component of the project, particularly through partnerships with the churches involved in the original trial. The research team, in collaboration with the Center for Community-Engaged Artificial Intelligence, is continuing to work directly with community members throughout the development process to ensure the tool reflects their needs, preferences and lived experiences.

“After each session, we will go back and incorporate the feedback that we got into the design of the prototype,” Allouch said.

The seed grant is also supporting this iterative, community-informed design process, allowing the team to refine Heartline based on real-world input before scaling the platform.

Heartline aims to support not only immediate behavior change but also long-term engagement with cardiovascular health by empowering individuals to take an active role in their health.

“I can’t think of anything that is more motivational than realizing that you can do it yourself,” Allouch said.