Artificial intelligence (AI) has been a hot topic in health care, capturing headlines for its ability to solve diagnostic mysteries and even outperform trained physicians in specific tasks. While we are still far from going to your local AI “doctor” for diagnosis and treatment, generative AI is already capable of intervening in meaningful ways—particularly in cardiovascular education for patients. This potential remains underutilized despite the significant public health need for cardiovascular disease prevention.
Generative AI offers personalized, accessible, and scalable tools that can address key challenges in our current cardiovascular prevention approach. Prevention relies on three pillars: identifying at-risk individuals, intervening—a combination of education, lifestyle changes, and medications—and monitoring metrics like blood pressure, cholesterol, and blood sugar. While electronic health records (EHRs) are equipped to trigger reminders for screenings based on robust guidelines, the largest obstacles often involve proper patient education and implementing sustainable lifestyle interventions. Here, generative AI stands out as a game-changer.
Studies show that half of patients with a diagnosed cardiovascular disease are still not following recommended prevention strategies, which include taking prescribed medications and lifestyle interventions. Patient education is critical for driving favorable outcomes but is often fragmented and inadequate. For example, many patients I see are not consistently taking their medications and cannot explain why they are prescribed certain therapies. I have found that it is powerful to explain to a patient if a medication is for symptom relief, preventing major events (like heart attacks or strokes), or helping them live longer. Many of the patients I see deeply appreciate when I take the time to go through their medication list but need additional education beyond a quick conversation at the end of an appointment.
This issue becomes particularly pronounced in cardiovascular prevention, which hinges on individual decisions regarding diet, exercise, and other lifestyle factors. Unfortunately, the reality of patient “education” during visits often consists of vague advice or standardized after-visit summary instructions: “lose weight, exercise 150 minutes weekly, avoid fried foods, and eat more fruits and vegetables.” While well-meaning, such guidance lacks the specificity and support needed to drive sustained change. Time constraints—15-minute visits—further limit the depth of these conversations.
Generative AI offers a way to fill these gaps. By synthesizing information and delivering it in a personalized, engaging format, AI can enhance education in ways previously unimaginable. Generative AI excels at tailoring content to individual needs on the patient’s schedule. It can provide explanations in a patient’s preferred language and at their literacy level, breaking down complex medical information into digestible insights. This is particularly beneficial for underserved populations, where language barriers or limited health literacy often contribute to disparities in care.
For instance, I entered a prompt into ChatGPT: “I have high blood pressure and high cholesterol. I want to change my diet and start exercising to improve these. Based on the ACC/AHA guidelines, can you give me a one-week schedule for meals and exercise that works around a 9-to-6 workday?”
In seconds, I received this detailed plan for Monday:
- Lunch: Grilled chicken salad with mixed greens, cherry tomatoes, cucumbers, olive oil, and balsamic vinegar. Side of whole-grain bread.
- Dinner: Baked salmon with roasted Brussels sprouts and quinoa.
- Exercise: 30-minute brisk walk after work (6:30–7:00 p.m.).
When I asked for Indian cuisine suggestions, the tool adapted accordingly, providing a menu of lentil soup (dal), vegetable khichdi, and tandoori chicken. It even offered grocery lists and estimated costs for these meals. This level of personalization and accessibility is unprecedented in traditional health care settings.
Despite its promise, generative AI is not without challenges. Ensuring the accuracy of AI-generated content is paramount, as misinformation could erode patient trust or lead to harm. Rigorous validation of training datasets and models is essential to guarantee that recommendations are evidence-based and aligned with literature and guidelines. Another concern is bias. AI models trained on incomplete or unrepresentative data could perpetuate existing health care disparities. For example, dietary recommendations that fail to account for cultural preferences or socioeconomic constraints may alienate diverse patient populations. Developers must prioritize inclusivity in training data and design to mitigate these risks. Privacy is yet another critical issue. Patient-facing generative AI tools must adhere to stringent data security standards to protect sensitive health information.
Imagine a world where patients receive a personalized cardiovascular prevention plan complete with meal suggestions, exercise routines, and reminders tailored to their schedules and preferences. Picture an AI assistant that collaborates seamlessly with health care providers, offering supplementary education and motivation between clinic visits. It could monitor medication adherence and provide on-demand education about specific medications. Generative AI has the potential to make this vision a reality. By addressing gaps in education and lifestyle intervention,
Anand Shah is a cardiology and preventive medicine fellow.
