May 29, 2026
Heart disease affects millions of people and often hits hardest in communities with fewer resources. Clinicians know heart health is shaped by more than what shows up in a lab test — the daily realities patients face at home, at work and in their neighborhoods can influence risk long before a crisis occurs, yet many widely used cardiovascular risk tools fail to account for these social and environmental factors.
That’s why OCHIN is partnering with UMass Chan Medical School to develop a new artificial intelligence tool that more accurately predicts heart disease risk by incorporating non-clinical drivers of health outcomes alongside clinical data. This project seeks to improve risk estimates by combining clinical information with patients’ lived experiences, and to ensure the research translates into better care.
Why current cardiovascular risk tools fall short
Traditional calculators of heart disease risk focus on age, cholesterol, blood pressure and smoking status. But they often ignore contextual factors influencing health outcomes, such as housing instability, food insecurity, chronic stress and financial strain — factors that can make it harder to keep appointments, take medications or follow care plans. Over time, these stressors can contribute to higher blood pressure, worsening diabetes and other conditions that increase heart risk.
When that broader context is not included, traditional calculators can underestimate cardiovascular risk. That can mean missed chances to provide earlier support, connect patients to resources or tailor a plan. For care teams already stretched thin, having a stronger indicator of who may need extra attention can make a real difference.
How OCHIN is addressing gaps in traditional heart risk prediction
OCHIN is using shared electronic health record data from community health organizations across the United States, including OCHIN member clinics, to build a practical AI tool that supports frontline heart care. The project draws on a real-world dataset that reflects the communities our members serve — enabling research that no single clinic could do alone and solutions that work in busy, resource-limited settings.
The AI tool will capture a broad range of contextual factors known to impact health outcomes and raise cardiovascular risk. Examples include housing instability, food insecurity, financial strain and stress, all of which can affect sleep, diet, medication adherence and access to care. It may also flag related challenges such as transportation barriers, social isolation, relationship safety concerns and difficulty paying for utilities.
A key part of this project is using information that already exists in clinical notes. Notes often include details that matter for health but don’t fit neatly into checkboxes, such as “lost housing,” “ran out of food,” “can’t afford meds” or “high stress at home.” These details can be essential to understanding contextual factors influencing health outcomes, but they are hard to use consistently across thousands of patients.
To help, researchers are using natural language processing, a type of AI that can identify meaning in written text, to find these clues and map them to medical and nonmedical health-related needs. Because the model draws from routine documentation, it should not require new forms, additional screening steps or major workflow changes.
The goal is to use what care teams already document without adding to provider burden. That matters because many clinics are balancing complex patient needs with limited staffing, and this project is designed to deliver value while respecting day-to-day operations.
Why this matters for clinical care
A more accurate risk estimate can help care teams prioritize prevention. If a patient’s clinical numbers look normal but the notes show serious social strain, the tool may suggest closer follow-up. That can help providers intervene earlier, before a preventable, costly emergency visit or hospitalization.
This tool’s emphasis on nonclinical drivers of health outcomes may also surface additional modifiable factors, meaning issues that can change with the right support. If financial strain is affecting medication adherence, a plan might include lower-cost options or enrollment in assistance programs. If food insecurity is contributing to poor diabetes control, a faster connection to local food resources could become part of the prevention strategy.
The goal is not to replace clinical judgment. Instead, it supports decision-making by making it easier to see patterns across a patient panel. If clinics identify higher-risk patients earlier, they may prevent some serious events or slow disease progression. This helps care teams focus limited time where it can have the most impact and can reduce avoidable emergency care and hospitalizations.
For rural and community health organizations, even small gains in prevention can relieve pressure on care teams. Fewer crises can mean more capacity to manage chronic conditions proactively. It may also lower costs and reduce the burden of care for patients and clinics.
A commitment to responsible innovation
This project reflects OCHIN’s broader commitment to using data responsibly to help our members transform care delivery and improve patient outcomes. As a trusted nonprofit with more than 20 years of experience supporting practice-based research in partnership with community health organizations, OCHIN has a long track record of stewarding health data in ways that prioritize patient privacy, ethical research and real-world impact.
Central to this work is OCHIN’s leadership of the ADVANCE Clinical Research Network, a nationally recognized network that uses electronic health record data from community health organizations to support research aimed at improving care for patients receiving care in community health organizations. All research conducted through ADVANCE follows rigorous governance standards and HIPAA‑compliant protocols, with multiple layers of oversight to ensure data is used anonymously and securely.
With these privacy protections in place, network EHR data helps OCHIN and trusted research partners test ideas that can strengthen clinical practice and transform health care delivery at the scale needed to address today’s challenges. The focus is on practical automation that supports community health and helps clinics deliver the highest standard of care, even with limited resources.
When nonclinical drivers of health outcomes contribute to cardiovascular disease, ignoring them can worsen access to care. Including them in prediction is one step toward more responsive accessible prevention. It is also a way to recognize what care teams have long known: context matters.
OCHIN will continue to share updates as this work moves forward, along with what the team is learning about building better ways to predict cardiovascular disease risk and how AI tools may help care teams deliver more proactive, patient-centered care.
Help shape the future of heart risk prediction
OCHIN members play a critical role in keeping this work relevant. If you are an OCHIN member and have feedback about which contextual factors most affect cardiovascular health in your community, we invite you to share your insights with our research team.
