Ep. 35: AI Revolution: From Calipers to Chatbots — with Dr. Sanjeev Bhavnani, Cardiologist

Artificial intelligence has a new solution for medicine seemingly every single day — so how do we use it without losing our minds (or our pockets)? In this episode, Dr. Regina Druz reconnects with her longtime friend and collaborator Dr. Sanjeev Bhavnani, a preventive cardiologist, prolific cardiovascular innovator, and a leader in national digital-health and AI policy. Together they trace how heart care moved ‘from calipers to chatbots,’ why we’re at a genuine inflection point, and the three guardrails — responsible use, real-world performance, and equitable access — that should steer innovation. They dig into the rise of ‘citizen doctors,’ the promise-versus-proof problem with consumer devices like continuous glucose monitors, why only a tiny fraction of FDA-cleared AI tools are actually used, and Dr. Bhavnani’s grounded, hopeful predictions for AI in heart care in 2026.

Watch on YouTube: A video version of this episode is available on the Own Your Heart Health YouTube channel. Subscribe to be notified of new episodes.

Episode Chapters

[00:00] Introduction: Meet Dr. Sanjeev Bhavnani
[03:27] From Toronto to the FDA: An Innovator’s Path
[06:39] The Cost of Curiosity: Becoming a Pioneer
[08:06] “AI Years”: How Fast the Field Is Moving
[13:50] Citizen Doctors & the Democratization of Data
[21:05] Why the U.S. Can’t Scale Innovation (and the EU Edge)
[25:41] 1,000 FDA-Cleared Algorithms — and How Few Are Used
[28:15] Promise vs. Proof: CGMs & Redefining “Healthy”
[37:58] Data, Decisions & Top-3 Predictions for 2026

Transcript

[00:00] Introduction: Meet Dr. Sanjeev Bhavnani

Dr. Regina Druz (00:02): Welcome to Own Your Heart Health. I’m Dr. Regina Druz, your holistic cardiologist. This week we’ll dive into common heart health concerns, uncovering root causes and unpacking scientific discoveries and controversies. The information provided does not constitute medical advice. Please contact your healthcare practitioner before making any changes that may impact your health.

Dr. Regina Druz (00:40): Hi everyone. A little secret: I’m always extra excited when my guests are fellow physicians, and fellow cardiologists, who’ve traveled an unexpected lane. Today’s guest is Dr. Sanjeev Bhavnani, whom I’ve known what feels like forever. He’s a self-described preventive cardiologist ‘adopted into integrative cardiology,’ and a very well-known cardiovascular innovator — he was the leading author and key opinion leader behind the roadmap-to-innovation document we collaborated on for the American College of Cardiology, back when ACC innovation was just beginning and then blossomed. He’s now on faculty at Scripps Clinic on the West Coast. I brought Sanjeev on to discuss what our listeners keep asking about and what we’re both doing with patients: digital health tools and artificial intelligence — how to leverage them without losing our minds or our pockets, to build something actionable, sustainable, and genuinely good. Welcome, Sanjeev.

Dr. Sanjeev Bhavnani (02:49): Thank you, Regina. It’s a pleasure to be with a friend — to talk about what we’ve done together, what we’re doing, and where we’re going. I’m excited, and I hope the audience finds it useful too.

[03:27] From Toronto to the FDA: An Innovator’s Path

Dr. Regina Druz (03:27): I ask all my guests: how did you grow up to be the cardiologist you are today? Who was little Sanjeev?

Dr. Sanjeev Bhavnani (03:27): My internal-medicine program director once joked that I was born with ECG calipers in my hands. I grew up in Toronto; my parents immigrated from India in the 1970s, and the ethos was public service. I wanted to build bridges as a civil engineer — the only bridge I built was to medical school. I had supportive family and great mentors who opened me up early to inquiry: how do you ask questions, and then go solve them? That brought me to the United States 25 years ago — New York, where you are, then Cleveland, Miami — and I learned research at places like the Cleveland Clinic and the University of Connecticut. I always say your patients ask the questions for you, and it’s on us to answer them; that’s how I became a clinical trialist. Then one day I got a call to move to San Diego, to the Qualcomm Institute, for an early-faculty opportunity with a friend of yours, Dr. Eric Topol. Back then it was called ‘mobile health,’ and I got to apply what we’d learned in medicine to new models of digital and AI-enabled care. That eventually brought me to the FDA to work on digital-health and AI policy. You never know where it’ll go — as long as you stay open to learning what you don’t know.

[06:39] The Cost of Curiosity: Becoming a Pioneer

Dr. Regina Druz (06:39): A common theme among my guests — and our patients — is that drive to ask questions and keep looking for better solutions. But there’s a cost: once you open those doors, you can’t close them, and you keep deviating from the established path for a physician, a cardiologist, or a patient. Over time, though, that defines a new path — it transitions the whole construct of medical care into a better reality everyone can benefit from. It’s just tough to be a pioneer; there are a lot of arrows in the pioneer’s back. I’ve always been impressed with how you navigated it with grace and real accomplishment. So take us through where we were a decade ago versus where we are now with digital health and AI.

[08:06] “AI Years”: How Fast the Field Is Moving

Dr. Regina Druz (08:06): I’ve started calling these ‘AI years’ — like dog years. GPT-4 already feels like the 1980s, because we’re now in the agentic and ‘vibe coding’ era. So before we lose our minds, give us the story of where we started and where we’re heading.

Dr. Sanjeev Bhavnani (09:16): We have to look back to know how much we’ve learned. Even five years ago, before COVID, the conversations about change in healthcare were completely different — and now they include so many more people, inside and outside healthcare, and the people we serve are pulling us forward. Consider this: when generative AI arrived, it took less than two months to reach a billion users, and now there are a couple of billion queries a day. People are using innovation in daily life and to inform their health. So this isn’t just a challenge to us as practitioners — it’s a responsibility, to figure out how to best use it for the next generation of care. We’re at an inflection point: people are wearing devices, doing telehealth, and we even worked on earbuds that can detect heart disease. If we shape it well now, the trajectory forward is very different from if we don’t adopt and take steps back. And it’s the people who are demanding it of us.

[13:50] Citizen Doctors & the Democratization of Data

Dr. Regina Druz (13:50): I recently wrote a short Substack post on ‘citizen doctors.’ AI and digital health came of age in the generative-AI era because we finally have tools that democratize access. Whether you’re wearing an Oura Ring or any other device, you can plug your data into a generative-AI model and get a reasonable first-level interpretation — maybe not 100% correct, but some guidance — without going through endless barriers. That’s the democratization, and that’s the evolution of citizen doctors: patients taking their own data and making sense of it. With that comes our responsibility to help patients, and ourselves, use these tools transparently, accountably, and safely, because outputs in this fast-moving field may or may not be valuable, and could sometimes put a person at risk. From your vantage point shaping innovation — including at the FDA — what are the top concerns and the top opportunities?

Dr. Sanjeev Bhavnani (16:40): You took the words out of my mouth: how do we ensure responsibility? We wouldn’t adopt a new tool at home unless we knew it would work well. So in healthcare we need to define responsible use, ensure it functions well in the real world, and — third — make sure everyone has access. There are a finite number of hospitals, clinics, doctors, x-rays, echoes, and ECGs. If we democratize new ways of delivering care with AI, are we expanding only where resources already exist — urban areas like New York or San Diego — or reaching people who lack access? Those three levers — responsible use, real-world performance, and equitable access — are priorities the country is working on, and I had the good fortune of working on them at the FDA. But I’d rephrase it as a society, because people live in their environments. Physician-innovators can be the bridge between policy and people — the responsible entity, the ones ensuring performance and equity — and that turns innovation into operations.

[21:05] Why the U.S. Can’t Scale Innovation (and the EU Edge)

Dr. Regina Druz (21:05): I always struggle with whether innovation should be top-down or bottom-up; the answer is usually a happy medium. Many European countries, with unified or federated national health systems and one electronic record per resident, have built registries and huge, linkable datasets — so in some ways they’re applying AI faster than we are, because we’re still siloed: every health system has its own electronic record, and even if you and I both use Epic on opposite coasts, our systems may not talk. We have lots of opportunity but the operational layer is missing. There’s no easy way for a practice that isn’t part of Scripps or Mayo to access their AI or digital-health tools — a cardiologist with 200 patients can’t just opt in from the outside.

Dr. Sanjeev Bhavnani (23:44): You’re absolutely right — we don’t have the infrastructure to change quickly. My start in digital health was designing smartphone clinics in rural Uganda and India, where we could look for heart disease precisely because there was no existing infrastructure — we leapfrogged. What we did there ten years ago we’re finally doing here. The pace of change will only be as fast as our ability to implement it. That gap — lots of excitement and development but very little real use — is actually an opportunity to slow down and be intentional. COVID was a pivotal lesson: telehealth was adopted virtually overnight through a coordinated national effort. Those are lessons I rely on.

[25:41] 1,000 FDA-Cleared Algorithms — and How Few Are Used

Dr. Sanjeev Bhavnani (25:41): Let me ask you a question. There are over 1,000 AI and machine-learning algorithms authorized for use in the United States by the FDA — available right now. How many do you think are actually used by doctors and their patients in practice — used, documented, and acted upon?

Dr. Regina Druz (26:43): I’d say somewhere between one and three percent — and that’s my optimistic guess.

Dr. Sanjeev Bhavnani (26:49): You’re right on. Two years ago it was less than 1%; now it’s a bit more — one to three percent. When you share that with people excited about AI, it brings everyone back to reality: there’s a lot we hear about, but very little being used. That’s changing at this inflection point, but US healthcare is so varied — different geographies, records, clinics, workflows, and access — that this is a moment to develop, decide what you actually want to do, and then push it forward. The pace is dictated by how fast, and how intentionally, we implement.

[28:15] Promise vs. Proof: CGMs & Redefining “Healthy”

Dr. Regina Druz (28:15): You hit a key point: new technology offers a lot of promises, but promises are not proof. Take continuous glucose monitors — Dexcom released a consumer CGM you can buy without a prescription. They clearly help people with diabetes, and Medicare reimburses CGMs there. But using them preventively — for impaired glucose tolerance, insulin-resistance concerns, or body-composition optimization — we just don’t have strong data yet. There’s a lot of promise and a paucity of proof. In medicine we want to know things work, that benefit outweighs harm. With AI and digital biomarkers, the technology moves so much faster than the proof. How do you reconcile that?

Dr. Sanjeev Bhavnani (31:08): Two approaches. First, innovation is here and people are already acquiring and using devices to make their own decisions about what’s healthy — that itself is a transformation we didn’t have ten years ago. Second, and this needs to change: how we use these tools will challenge how we defined conditions before. We defined insulin resistance and diabetes with unified measures from decades of research — but now healthy people wearing CGMs will show patterns no one has seen, so we’ll have to ask, what defines ‘healthy’? I’d anchor it in how the person is defining and driving their own health — then layer on what we’ve done for 50 years: diet, exercise, sleep, and stress.

Dr. Regina Druz (33:33): I call it stress resiliency, not stress reduction.

Dr. Sanjeev Bhavnani (33:38): I like ‘contentment.’ If we look at health that way and let people use their CGMs and watches to identify their own patterns, we’ll find data that challenges how we define insulin resistance, diabetes, and hypertension. So I support both — we shouldn’t be a barrier to use; we should be facilitators who bring that data in a coordinated way.

Dr. Regina Druz (34:53): I’m laughing because you and I aren’t writing the guidelines — and cardiology guidelines, for all the useful work the authors do, aren’t very flexible. They’re a rigid, population-based, top-down framework — a good framework, but hard to personalize. The opposite is patient-crowdsourced data and looking for new patterns — what we call N-of-1 or small data, which I’m a big fan of. It’s cognitively and resource-intensive — in our work at Holistic Heart Centers, personalizing care so we don’t just default to guideline minimums is intense — but it’s how we redefine specific patient segments and achieve better outcomes. Part of innovation is reframing how we as doctors think, and accepting that our patients are citizen doctors now, going to ChatGPT and other models for personalized, actionable guidance — not just for disease control, but for health optimization.

[37:58] Data, Decisions & Top-3 Predictions for 2026

Dr. Sanjeev Bhavnani (37:58): We do need data — but let’s take old data and make it new through coordinated collection: mobile devices, electronic records, and device developers all hold data. A friend of yours, Dr. Geoff Ginsburg, used to point out that direct-to-consumer genomics did this a decade earlier — people acquiring and acting on their own genetic tests — and that data became part of their health and then part of drug development and policy. I hold onto this: data creates decisions, evidence creates policy. To drive a decision for one person, use the data available; to do it for everyone, you need evidence. The physician-patient partnership in the middle is where we figure out how to share data responsibly and let it inform guidelines. There’s now a hypertension-screening tool on watches — a chance to study what happens when millions use it at scale, and to generate data that informs guidelines and decisions. I’m confident in our ability to both generate and analyze that data — it’s on us to do it.

Dr. Regina Druz (41:57): I’d add a small caveat: sometimes it’s not data driving decisions first, but decisions driving data. Patients seeking health optimization — rethinking lifestyle, supplements, and which medications truly help — make a decision to own their health, and that human agency opens the door to incorporating data from screening tools, wearables, and genetic and blood markers. When I suggest tools, patients may not know them, but their mindset is already set to take ownership, and the tools become what’s in their toolbox. That’s why I see this explosion of AI tools and the rise of citizen doctors — with safety, accountability, and partnership, it makes the decision to transform a little easier.

Dr. Sanjeev Bhavnani (44:42): Beautifully put. And here’s a thought: right now we treat AI as a tool — can we diagnose or analyze better? But if AI becomes part of the process of care delivery — detection, decision, diagnosis, treatment — that’s a systems shift. I believe patients will drive it, and younger physicians born into AI will help, so that within five years AI becomes part of the process and we have a new model of delivery.

Dr. Regina Druz (46:36): So to wrap up — your top three predictions for healthcare AI in 2026?

Dr. Sanjeev Bhavnani (47:13): First, we’ll start seeing the operational successes and challenges of AI — enough institutions are using it that we’ll learn what’s working and what can work better in care delivery and patient-physician interactions. Second, in cardiology, we’ll see more innovative, AI-enabled digital-health and wearable devices that do things we didn’t think possible for understanding the heart. And third — hopefully — aside from steps toward payment and reimbursement, we’ll start better understanding the equitable implementation of digital and AI; any step in that direction gets us onto the inflection point toward where we need to go. So: optimism, hope, and a bit of reality — we can do this in 2026.

Dr. Regina Druz (49:36): I love that — optimism, hope, and a reality check. Sanjeev, thank you so much for being a guest on Own Your Heart Health. A delightful conversation.

Dr. Sanjeev Bhavnani (49:52): Thank you, Regina, and thank you everyone — it was an absolute pleasure, dear friend.

Dr. Regina Druz (49:59): Thank you for tuning in to Own Your Heart Health with Dr. Regina Druz. This podcast is powered by Holistic Heart Centers. If you enjoyed the show, please rate and review us on your favorite podcast platform. To learn more about our services, visit holisticheartcenters.com and subscribe to our YouTube channel — the link is in the show notes. See you next week.

Frequently Asked Questions

What is a “citizen doctor,” and is using ChatGPT for my health a good idea?

Dr. Druz uses ‘citizen doctor’ to describe patients who take their own health data — from wearables like an Oura Ring or smartwatch, or from lab results — and run it through generative-AI tools to get a first-level interpretation. Both physicians see this as a real democratization of access, but with caveats. The output may be reasonable and helpful, or it may be incomplete or wrong, and in a fast-moving field it could occasionally point someone down the wrong path or create risk. Their guidance is to use these tools transparently, accountably, and safely — as one input in a partnership with your clinician, not a replacement for professional evaluation. AI is best treated as a starting point for questions to bring to your doctor, not a final answer. This is educational commentary, not medical advice.

Should healthy people use a continuous glucose monitor (CGM)?

This is the heart of the ‘promise versus proof’ discussion. CGMs clearly help people with diabetes — the evidence is strong and Medicare reimburses them in that setting — and consumer versions (such as Dexcom’s over-the-counter device) are now available without a prescription. But for preventive use in people without diabetes — to explore impaired glucose tolerance, insulin-resistance concerns, or body composition — Dr. Druz notes the data are still thin. Dr. Bhavnani’s view is that healthy people wearing CGMs will reveal patterns we’ve never catalogued, which may eventually reshape how we define ‘healthy,’ insulin resistance, and related conditions — but that’s a research opportunity, not settled science. If you’re considering a CGM for non-diabetic use, discuss the rationale and interpretation with a clinician. This is general education, not a recommendation.

How many FDA-cleared AI tools are actually used in care?

Surprisingly few. Dr. Bhavnani notes that more than 1,000 AI and machine-learning algorithms are authorized by the FDA for use in the United States, yet only roughly one to three percent are actually used in day-to-day practice — meaning a clinician applies the tool, documents the result, and acts on it. Two years earlier, the figure was under one percent. The gap reflects how fragmented US healthcare is — different geographies, electronic records, workflows, and access — which makes implementation, not invention, the bottleneck. He frames this as an opportunity to slow down, be intentional, and choose what’s genuinely worth deploying. (These figures are as stated on the episode and worth verifying against current FDA data.)

What are the top predictions for AI in heart care in 2026?

Dr. Bhavnani offered three, with a tone of ‘optimism, hope, and a reality check.’ First, the field will start seeing real operational results — enough institutions are now using AI that we’ll learn what’s working and what needs improvement in care delivery and patient-physician interactions. Second, cardiology will see more innovative, AI-enabled wearables and digital-health devices that reveal things about the heart we couldn’t easily detect before. Third — and more aspirational — beyond steps toward payment and reimbursement, he hopes for progress on the equitable implementation of digital and AI tools, so the benefits reach people where they are. He also predicts a longer-term shift from AI as a ‘tool’ to AI as part of the care ‘process.’ These are forward-looking opinions, not guarantees.

Show Notes & Resources

Guest: Dr. Sanjeev Bhavnani, MD

Dr. Sanjeev Bhavnani is a preventive cardiologist and nationally recognized cardiovascular innovator on the faculty of Scripps Clinic in San Diego. He helped launch digital-health and ‘mobile health’ research at the Qualcomm Institute (alongside Dr. Eric Topol), was the lead author behind the American College of Cardiology’s roadmap for innovation, and has shaped national policy on digital health and artificial intelligence through work with the FDA. His focus is translating AI and connected devices into responsible, high-performing, and equitable cardiovascular care.

Dr. Sanjeev Bhavnani — Scripps Clinic (San Diego)

Resources Mentioned in This Episode

FDA-cleared AI/ML in medicine — over 1,000 algorithms are authorized, but only ~1–3% are actually used in day-to-day care today (as stated on the episode; verify against current FDA data)
Continuous glucose monitors (CGMs) — now available over the counter (e.g., Dexcom’s consumer device); strong evidence in diabetes, far less for healthy/preventive use
Wearables — the Oura Ring and smartwatches now offering ECG and a hypertension-screening feature; an opportunity to study real-world use at scale
Citizen doctors — patients bringing their own device data and generative-AI analyses to visits (see Dr. Druz’s Substack post on the topic)
The ACC roadmap for innovation — the American College of Cardiology document Dr. Bhavnani led
Small-data / N-of-1 research — a personalized approach to finding patterns that population-level guidelines miss
‘Chat with the podcast’ on NotebookLM — Google’s public notebook loaded with OYHH episodes (holisticheartcenters.info/notebook)

Holistic Heart Centers

holisticheartcenters.com
HeartWell.ai — AI-powered cardiovascular risk assessment
Address: 55 Bryant Avenue, Suite #6, Roslyn, NY 11576
Phone: 877-511-5166
YouTube: @reginadruzmd
Instagram: @dr.reginadruz
Podcast: Own Your Heart Health — available on Apple Podcasts, Spotify, and all major platforms

Listen & Subscribe

If you enjoyed this episode, please rate and review us on your favorite platform — it helps more people find the show.

Apple Podcasts Spotify YouTube