How University Hospitals Harrington Heart & Vascular Institute is Harnessing Artificial Intelligence to Improve Cardiovascular Health

Innovations in Cardiovascular Medicine & Surgery | June 2026

Experts at University Hospitals Harrington Heart & Vascular Institute in Cleveland, Ohio, are collaborating with peers at Case Western Reserve University to harness artificial intelligence (AI) in pursuit of timely, precise and equitable solutions aimed at revolutionizing health care delivery.

Sanjay Rajagopalan, MD, MBA

The HEAL initiative — Health Enhancement through Artificial Intelligence — is a multifaceted collaboration among thought leaders at UH Harrington Heart & Vascular Institute and Case Western Reserve University’s Departments of Biomedical Engineering and Computer and Data Sciences.

The vision for the HEAL initiative is to develop trustworthy AI innovations that bridge data and compassion.

“Too often, AI models are trained on data sources that may not generalize across a broader range of settings,” says Sanjay Rajagopalan, MD, MBA, Chief of Cardiovascular Medicine and Chief Academic and Scientific Officer at the UH Harrington Heart & Vascular Institute, Director, Case Cardiovascular Research Institute and the Herman K. Hellerstein, MD, Professor of Cardiovascular Medicine at Case Western Reserve University School of Medicine.  “Compassion, which is a UH value, is necessary to ensure that AI is driven by human-centered design and sculpted by the stakeholders where it will be implemented.”

Team members are utilizing large language models (LLMs), geospatial imaging and other AI capabilities to better understand a range of cardiovascular health topics:

• Unleashing intelligent search and phenotyping

A promising approach to clinical AI is the use of retrieval augmented generation (RAG) combined with graph-based models layered onto systems like Epic Systems, where structured data such as labs, diagnoses and medications are organized into knowledge graphs that capture relationships between patients, conditions and outcomes. In this Graph RAG framework, models query these interconnected data rather than isolated fields, enabling more accurate and context-aware reasoning. This can be extended to unstructured data using natural language processing to extract insights from clinical notes and reports, progressively enriching the graph. Together, this approach shifts care away from static dashboards toward dynamic, conversational intelligence that allows clinicians to directly query complex data and receive integrated, patient-specific insights.

• Mining EKG and imaging data to improve cardiovascular risk assessment

Electrocardiograms (EKGs) capture distinct features of the heart’s electrical activity. Together with imaging features, LLMs are helping identify the top features that may best predict major adverse cardiovascular events or mortality. “Reasoning AI models can help us analyze data, such as electrical desynchrony or delayed ventricular depolarization and other imaging and laboratory attributes, to identify causal cardiovascular relationships and predict adverse events,” Dr. Rajagopalan says.

• Reduce unnecessary rehospitalization

“Rehospitalization is expensive and reportable from a quality perspective,” Dr. Rajagopalan says. “Our multi-modal approach is helping uncover reasons for readmissions.” The team has employed LLMs to analyze thousands of admission and discharge notes, laboratory data and other parameters, including images and EKGs, to help understand the likelihood of 30-day readmissions.

• Observing contextual health outcomes

Using geospatial AI, including satellite imagery and street-level views, the researchers are examining neighborhood features that may reveal external health risks. “While patient demographics can provide a fairly accurate portrayal of a person’s health, they do not account for factors such as proximity to green spaces, access to grocery stores or social stressors,” Dr. Rajagopalan says. “We believe we need to understand patients’ environments to unlock a deeper understanding of human health trajectories over time.”

Navigating a Data-rich Future

Dr. Rajagopalan and his colleagues published a clinical outlook on emerging AI in Nature Reviews Cardiology that highlights how AI technologies are revolutionizing cardiovascular health. “We now have the capacity to investigate environmental factors at an unprecedented scale to better identify complex risk patterns and guide precision interventions,” he says.

The team is partnering with academic institutions worldwide to leverage large data sets. “Globally, we are collecting an unbelievable amount of information,” Dr. Rajagopalan says. “The challenge is that the data we collect often suffers from a lack of connected and interpretable understanding.”

“Utilizing AI, big data and computational resources requires bringing together intelligent minds and creating environments that foster the development of unique solutions,” Dr. Rajagopalan says. “The collaborative academic medicine framework underpinning University Hospitals and Case Western Reserve University enables us to lead the advancement of AI tools that will empower caregivers and patients.”

For more information, contact Dr. Rajagopalan at Sanjay.Rajagopalan@UHhospitals.org.

Contributing Expert: Sanjay Rajagopalan, MD, MBA