The Bottom Line
Hello Heart's peer-reviewed research shows that combining clinical data with real-world home monitoring may help identify short-term cardiovascular risk more clearly than traditional calculators alone. Built on more than a decade of evidence, the Hello Heart platform pairs an FDA-cleared blood pressure monitor with AI-powered coaching grounded in peer-reviewed research.
- Research Finding: A peer-reviewed study in Frontiers in Digital Health evaluated a Hello Heart AI model for short-term ASCVD risk prediction across 51,127 adults with hypertension.
- Data Approach: The model combined medical record data with home monitoring measurements and may help identify short-term cardiovascular risk more clearly than traditional risk calculators alone.
- Research Stage: This ASCVD model is a research finding and is not a currently deployed feature of the Hello Heart platform.
- Design Guardrail: Nia, Hello Heart's AI heart health assistant, does not diagnose conditions, change medications, or replace a physician. She supports people between clinical visits.
Why heart health is a natural fit for AI
Conversations around artificial intelligence (AI) applications in healthcare have reached a fever pitch, but I’ve noticed that many of them aren’t centered on people or their health.
Much of the attention is going to tools that make our healthcare system more efficient: clinical note summaries, scheduling automation, patient routing, coding support, and other workflow improvements that can give time back to clinicians and care teams.
That work is important. The healthcare sector has never been known for its efficiency.
But the opportunity that excites me the most is how we can use AI to help people understand their health risk earlier, make sense of their data, and take action before a problem becomes more serious.
That’s what we’re focused on at Hello Heart.
Heart health is a natural place to apply AI because the signals are measurable and the stakes are high. Blood pressure, cholesterol, medication routines, heart rate, weight, activity, symptoms, and engagement patterns can all help create a clearer picture of cardiovascular health over time.
Traditional care often sees those signals in pieces.
A person may take blood pressure readings at home for weeks, talk about medications during a doctor visit months later, skip doses during a stressful period, or receive a long-term risk estimate that feels too abstract to change daily behavior.
Hello Heart is using AI to connect those moments in a more useful way.
Why heart health risk can't be measured in a single snapshot
Traditional cardiovascular risk tools have played an important role in preventive care. Tools such as the Pooled Cohort Equations and PREVENT™ scores help estimate longer-term cardiovascular risk, over 10 years or more.
Cardiovascular risk tools help clinicians think about prevention, treatment, and risk at the population and patient level.
But many people managing high blood pressure want to understand something more immediate: what is changing in my health right now?
Blood pressure isn't a static number. It shifts hour to hour—sometimes minute to minute—in response to activity, food, stress, sleep, and daily routine. Medication adherence adds another layer of variability. Many people miss doses during asymptomatic periods when they don't feel the immediate effect.
A single office reading captures none of that. It offers a snapshot, not the longitudinal context primary care providers need to assess risk and guide treatment.
That is where shorter-term risk prediction can add an important layer, for both patients and clinicians.
A peer-reviewed study published in Frontiers in Digital Health evaluated a Hello Heart AI model designed to predict short-term atherosclerotic cardiovascular disease, or ASCVD, risk over 90 and 365 days. ASCVD includes devastating and costly cardiac events such as heart attacks and strokes.
The retrospective study included 51,127 adults with hypertension who enrolled in the Hello Heart cardiovascular risk self-management program between January 2015 and January 2024. Our team developed an AI-driven machine learning model using electronic health record data and mobile health measurements, including blood pressure and heart rate from at-home blood pressure monitoring.
To predict who may have a heart attack or stroke within the next 90 or 365 days, the model drew on a broader clinical picture than traditional risk calculators—combining medical record data with real-world home monitoring measurements like blood pressure and heart rate. In the study, that approach identified short-term cardiovascular risk more accurately than Pooled Cohort Equations and PREVENT scores alone.
The ASCVD prediction model is a research finding that supports our philosophy at Hello Heart: everyday heart health data can help create a more personalized view of cardiovascular risk.
Catalyzing risk prediction into real-world action
A risk model is useful only when it helps someone do something with the information.
That could mean helping people adhere to their prescribed medications and treatment plans, preparing more informed questions for a doctor visit, recognizing when a symptom is unusual or needs attention, or building sustainable health habits that reduce risk over time.
This is where product design matters as much as data science.
At Hello Heart, we’ve built AI into our preventive heart health technology from the start. Our app and FDA-cleared blood pressure monitor have helped more than 350,000 people track and manage their heart health, and they receive personalized insights and coaching to help them better understand their numbers.
Nia, our AI heart health assistant, helps users get evidence-based answers about heart health questions, prescriptions, side effects, drug interactions, and medication routines. It supports people between clinical visits, when many everyday questions and decisions happen, and where breakdowns in treatment plan adherence tend to happen most often.
The experience is designed to help people feel informed and capable. Heart health can feel scary, especially when someone sees a high reading or does not understand what their numbers mean. Good technology should reduce confusion and empower people to take control of their health.
Predictive AI has the greatest value when it becomes part of an experience that is clear, personal, timely, and responsibly designed for the everyday user.
What responsible AI in heart health should look like
Healthcare needs a higher standard for AI.
People are already asking general-purpose AI tools personal health questions. Many of those tools were not built for cardiovascular care, may not understand a patient’s context, and may not operate within clear clinical boundaries.
An issue as serious as heart health requires more care than that.
Responsible AI in this space should be evidence-informed, clinically guided, and privacy-conscious. It should help people ask better questions, understand their health more clearly, and guide them to human care providers and resources when they need additional support or intervention.
At Hello Heart, that means building AI with defined guardrails. Nia does not diagnose conditions, change medications, or replace a physician or pharmacist. Instead, she is complementary to clinical care: there for people on the 360+ days a year they’re not at the doctor’s office. Human clinical expertise, including input from highly specialized cardiologists and pharmacists, remains central to how we design, evaluate, and improve the Hello Heart experience.
AI in healthcare must be tested, published, scrutinized, and improved over time. At Hello Heart, we've worked on this for more than a decade, and the evidence continues to support that approach.
A more useful future for healthcare AI
The next chapter of healthcare AI should be more ambitious than automation. It should personalize preventive healthcare and impact lives on a global scale.
Predictive AI will not solve cardiovascular disease by itself. No technology can without the action of humans, both patients and clinicians. But when AI is built responsibly and connected to a thoughtful user experience, it can help people understand their risk earlier and take steps that can radically improve their heart health over time and help them live healthier, longer lives with the ones they love.
For a condition as common, costly, and consequential as heart disease, earlier action is exactly where healthcare AI should be headed.
To learn more about Hello Heart’s work in AI, check out our solutions here.
Frequently asked questions
Can AI predict heart attacks and strokes?
AI cannot predict heart attacks or strokes with certainty, but predictive models may help identify people at higher short-term risk. In a peer-reviewed study, a Hello Heart AI model combined medical record data with home blood pressure and heart rate readings to estimate short-term cardiovascular risk. This model is a research finding, not a currently deployed clinical feature.
Why isn't a single blood pressure reading enough to assess heart health risk?
Blood pressure is not static. It shifts hour to hour in response to activity, food, stress, and sleep, and medication adherence varies over time. A single office reading captures one moment, not the longitudinal pattern clinicians need to assess risk and guide treatment. Home monitoring over time offers a fuller picture of what is changing in someone's health.
Does Hello Heart's AI replace a doctor?
No. Nia, Hello Heart's AI heart health chatbot, is built with defined guardrails. She does not diagnose conditions, prescribe medications, or replace a physician or pharmacist. Instead, she supports users between clinical visits, helping them understand their numbers and ask better questions, and guides them to human care providers when additional support is needed.
