Wearables, Hypertension Prediction, and the Patient–Physician Dyad with Sumbal Desai

In this episode of The Future of Medicine, we sit down with Dr. Sumbal Desai to explore how consumer health technology is reshaping medicine at scale — and why the human connection between patient and physician remains central to care. From breakthrough blood pressure sensing on wearables to a world-spanning hypertension predictor, Desai shares the design philosophy, scientific basis, and real-world impact behind Apple’s health initiatives. The conversation also delves into practicalities of bringing health data to the bedside — how physician-facing reports are crafted to be digestible in under a minute, how faxes and PDFs are being replaced by smarter EMR integration, and how data snapshots from outside the clinic enrich clinical decision-making.

Key themes and takeaways:

• The dyad of patient and physician: Why Apple emphasizes the patient–doctor relationship and how both sides should benefit from digital health data.
• Designing for the clinician: Reports that can be understood in 30 seconds to a minute, prioritizing clarity and usefulness without adding nuisance.
• From device to decision: Wearables moving beyond tracking to inform diagnoses, treatment decisions, and early warnings with novel sensing methods, including blood pressure measurements.
• Hypertension predictor: A first-of-its-kind deployment with global reach and the potential to identify hundreds of millions of cases. What success looks like and what it means for public health.
• Scale with responsibility: Delivering credible health insights at a population level while protecting privacy, safety, and scientific transparency.
• The Apple Heart Study and Stanford collaboration: A landmark public health study with hundreds of thousands of participants, findings on atrial fibrillation risk, and the importance of open publication to advance the field.
• Patient voices: Letters from customers about how devices have influenced their lives and opportunities to translate data into better care.
• The path forward: Opportunities and barriers as digital health technologies intersect with clinical care, research, and daily patient life.

What you’ll hear in this episode:

• A behind-the-scenes look at how Apple designs patient-facing health information that doctors can readily use in conversations with patients.
• The “snapshots of life” philosophy outside the clinic and how those insights support holistic care.
• Practical steps in moving from faxes to modern, integrated health data workflows.
• Real-world examples of how a novel blood pressure sensing approach can alter care trajectories.
• Reflections on large-scale health studies conducted with academic partners, and why openness matters for scientific progress.

Why this episode matters: This episode offers a candid, practically grounded look at how a tech-driven health platform is designed to support doctors, engage patients, and scale insights to millions — while preserving trust, nuance, and empathy at the heart of medicine.

Call to action: If you enjoy The Future of Medicine, subscribe for more conversations with leading scientists shaping the next era of healthcare. Please rate and review the podcast to help others discover these important discussions. Share with friends and colleagues who are curious about how science becomes medicine.

Chapters

• 0:00 Opening and the Future of Medicine
• 0:46 Apple’s Hypertension Predictor and Health Announcements
• 2:05 The BP Sensor and Its Scientific Promise
• 3:50 Designing for Physicians: Reports, PDFs, and EMR
• 5:15 The Doctor–Patient Dyad: Snapshots and the Watch
• 6:50 Roots and Education: From CS to Medicine
• 8:10 A Journey Through Media and Technology
• 9:50 Personal Catalyst: Mother’s Stroke and the Turn Toward Medicine
• 11:20 Returning to Medicine: Medical School and Residency
• 13:20 The Design Process: From Feasibility to Shipping
• 21:40 Closing Inspiration and Gratitude

Transcript

[00:00.1]
You should be able to look at it for 30 seconds to a minute as a physician and understand what it's trying to tell you. Dr. Sumbul Desai is Apple's Vice President of Health and a leader at the vanguard of digital health. From ECG on the wrist to headphones that double as hearing aids, and now high blood pressure detection, Desai's team is revolutionizing consumer health technology.

[00:17.9]
She joins us to discuss the science and design behind these amazing advancements. We think about the dyad between the patient and the physician. We're not looking to replace physicians. Now we may be biased. Welcome to Stanford Department of Medicine's inside look at the future of medicine. It is a really interesting time to have you here.

[00:34.7]
Many of us were watching last week as Apple made some pretty big announcements in the health space. And I was just thinking, as you announced a new predictor, first in the world for hypertension, I think I saw it was going to be rolled out in 150 countries and could potentially pick up a million cases.

[00:54.0]
I mean, I was sitting there thinking, what does that feel like to develop something that is rolled out to potentially, I know you don't release the numbers, but potentially hundreds of millions of people pretty much overnight. You know, it's funny, you don't actually, when you just said it, I thought about it. Day to day we don't necessarily think about it, but we do, and you'll see kind of a video come up, we often will talk about the letters we get from customers.

[01:20.9]
We literally get letters on health every day. Like, Tim gets letters every day from customers about how the devices have had some impact or potentially saved their life in a number of ways. And we don't use the word saving life, but they say that, and I think for us, the science that went into that product in particular, we're incredibly proud of.

[01:41.5]
It's a novel way to measure blood pressure. And we're very excited about the potential scientific opportunities that come out of it. And I can talk a little bit more about that in a bit, but, it's been a labor of love. That was not an easy journey. No, a lot of things went wrong at various points, but a lot of things went right.

[01:58.8]
We have so many teams at Apple that work on these products and just, it's a big celebration for all of our teams. It's funny, I never post on social media and this is the first time I ever posted on LinkedIn because I was really proud of our teams. We did an amazing job on this one. And that's really why I was interested to dig into and we can come to some of the details around that actual predictor in a little bit.

[02:17.1]
But just to be able to have a team, to have an effort that spans all that length of time, but then to have it happen at scale. So much of what we do in medicine is of course, us and a patient in front of us, which is just special. Just the ability to change the course of one patient's day and maybe their life is very special.

[02:35.5]
But to be able to do that at a scale where you're potentially affecting hundreds of millions of people in a pretty short period of time, as it goes from not being on the wrist to being on the wrist, that's pretty amazing. It's been amazing in ways that I would not have imagined.

[02:50.8]
I think the best is the same way that you all, as physicians will have a patient come up to you when you see them outside of the hospital and they thank you. We have a similar thing. We'll be out there. I mean, when I walk around normally you wouldn't recognize who I am. So, it doesn't happen to me that often, but it happens to me at events.

[03:06.2]
And they'll come up to you and say, hey, I got that AFIB notification and I went to the doctor and this is what happened. And those stay with you. Those are pretty remarkable. It's kind of the fuel. We talk about it as the fuel that drives our work. Yeah, amazing. Well, you talk about journeys and the journey it took to get to that hypertension predictor.

[03:22.4]
I would love some of the audience know you quite well, but some of them do not. So I'd love to talk about your journey, which is pretty remarkable. One of the things I learned actually that I didn't know about you when I was prepping for this yesterday was that your major in undergrad was computer science. So take us back though a little bit.

[03:38.6]
Right. Take us to pre undergrad if you like. How did you envision your life was going to unfold? I assume it wasn't in this direction. It wasn't this. No. Yeah, tell us about. It's funny. I'll give you a somewhat longer version, but I won't bore you guys. I grew up in an Indian family and there was only two choices when I was growing up.

[03:57.7]
I think, as an Indian child, was like engineer or doctor. That was it. And so, I actually got into a six year BS/MD program at Rensselaer. And and I literally cried the day I got in, because I knew I'd have to go because my dad was like, you are going to be a doctor.

[04:15.8]
And I was like, no, I'm not. And so in my first semester, and I have a little bit of this. Unfortunately, I have a little bit of this rebellious side to me, which is not great sometimes. But I purposely failed my whole first semester. You purposely failed? I purposely. I literally have every letter on the Alphabet.

[04:31.0]
There actually are Z's. If you don't take your final in a semester, and I have a Z on my transcript. And, and I went home, my dad was like, just do whatever you want. So he's like, just do it. So I changed to computer science because I wanted to still stay at an engineering school.

[04:47.5]
I actually like how you build things. I love the how you build things. Part of the reason I loved physiology so much is it's the why things work. I did computer science, but I minored in communications. And that's actually when I got exposed to media and had interned at the CBS affiliate back in Albany.

[05:04.3]
I went to RPI in Troy, New York, which, if anyone knows where that is. I ended up, starting when I was looking at jobs, got a job at ABC News, which I didn't, kind of just fell into it. But yeah, I mean, how did that happen? It happened because, So I've always, There's one thread that's very consistent in my work as I look at back now, is the ability to actually impact and empower people through information.

[05:28.2]
And I initially thought it was communication. And, you know, with the media, you can potentially use it to really drive education and drive change. And so that was my interest in the media business back then. And I went to work for CBS thinking that I could do that. And then I moved to ABC News thinking that I could, like, have that impact.

[05:44.8]
What you learn very quickly is, like, the media business, obviously, like many businesses, is driven by P&L. And so you make certain decisions as a result. So I. And I also just wanted more of an analytical job. So I moved over to the business side of Walt Disney company that owned ABC. And what was neat about that is you actually understand, like, what goes into the business that drives all of our work and why that's important.

[06:06.9]
And in 2001, my mother had a massive hemorrhagic stroke. And, it was in August of 2001, and so it was in New York. So it was the month before September 11, so she was in the unit. And I was not a doctor at this point. She had a cavernous malformation that bled, since all of you would know what that is compared to when I usually tell the story.

[06:27.0]
And she got moved on the day of 9/11 to a rehab facility. So she went from an ICU to rehab. And so basically I had to be her advocate and almost like her nurse because the care was just so different. That's actually when I first fell in love with medicine for, like, the first time. And I think when multidisciplinary care comes together, it's pretty remarkable.

[06:45.0]
And I got to see that happen with my mom's case. She went from a coma to being able to be rehabbed. She obviously changed quite a bit, but, you know, came out of that. And that's when I decided to go back to medical school for exactly what you talked about, that impact that you have on the one on one. But one of the best pieces of advice that stayed with me during that time was a neurosurgeon.

[07:05.3]
And they can have inspiring comments occasionally. Said to me, you have to advocate for your mom because no one else will. That thread has stayed with me now through my career because I think when I was practicing, and obviously I owe so much to Stanford, this is still a very special place in my heart.

[07:24.4]
Did my residency here, I learned everything. You know, that empathy in the bedside manner that I want to bring to our products, I learned here. And I want to treat. I want to make sure patients feel empowered and families feel empowered to kind of speak up for themselves. And so did my residency.

[07:41.8]
And while I was here, I got very involved with the technology side as well and wanted to see that intersection of how do you drive efficiency through healthcare? There was a moment there. So you decided to go to medical school? I decided to go back to medical school. I was 30 when I went to med school. And I decided to go back to medical school because.

[07:58.8]
So I went back to Disney after my mom had the coma. And, I took care of her for a year, and there was a meeting that I had to go to. I used to work for Strap Planning in Disney. I decided to go back and do my postbac and then applied to medical school, went to school in Virginia at MCV, and then did my residency program very gratefully here, for this team up here.

[08:20.2]
And, during residency, if you all recall, for those that are here, I recall you being a resident. Yeah, EPIC was one of the first rollouts. So we kind of, like, were just getting into the EMR world back then. And, I think the experience here and where we are, you're at the center of technology.

[08:40.0]
It's so easy to see those intersections. And that's kind of what drove my interest in that space. And then fast forward and here I am. I know, amazing. One of the other things I learned, yesterday, or I'd forgotten, I must have known all these things at some point, was that you were the vice chair for strategy and innovation for Bob.

[08:55.3]
Yeah. In the department. So we also now have another vice chair for strategy and innovation, Vicky Parikh, within the department. But, I remember that you were working on strategy and innovation for the department. So talk a little about what you did in the department here before your next chapter.

[09:13.4]
So my role here was for the department as well as for the hospital. So I had worked, as an ACMO for strategy and innovation. Not sure really what that means anymore. But anyways, we did a lot of, if you all recall, when video visits were first rolled out, we actually worked with the CIO and the CSO back in the day to one make sure that the codes and the reimbursement would happen for those visits, but actually rolled out that technology through Stanford.

[09:44.6]
When I was here, in the department side, that's actually when we launched the Center of Digital Health. So thinking about how do we actually do more research in that space and connect startups and technology companies to the institution so we can actually look and explore and see what technology actually works and what doesn't. I think it was during those years here and I reference it a lot to the work that I do now.

[10:08.7]
I really learned, being on the hospital side, you all get exposed to startups all the time that tell you, I can solve this for you, but it's a technology looking for a use case. And that is the absolute, it's one of my biggest pet peeves. And I tell our teams at Apple we will not do use cases for technology.

[10:25.0]
We need to solve a problem and see how the technology actually solves a problem. And I learned that from here and being here and working in the department which was focused on looking at those use cases as well as working on the hospital side. Yeah, you're the founding director of the Center for Digital Health and of course it's in very good health today.

[10:42.1]
It's in very good health. It's in very capable hands here with Dr Linos. In medicine, you actually have to show your work. And so you have to kind of be open to sharing research and talking about it. So getting Apple to actually do what we did with the Apple Heart Study with Stanford was actually groundbreaking for us as a company.

[11:00.8]
You know, we did the largest public study with Stanford. It was 400,000 participants. But then we published on it. Right. Which is unheard of for us, in the past. So just maybe remind everyone quickly what that was. Yeah, so we had, developed an irregular heart rhythm notification.

[11:16.0]
And what it does is it uses the PPG sensor on the back of the watch and it will notify you if it sees patterns consistent with atrial fibrillation. They were like, okay, we're going to put this out there. But we had never had the conversations around sensitivity and specificity. How do we avoid false positives?

[11:32.9]
There was a big concern that we were going to drive the medical community to frustration with a bunch of notifications. So what happened is we worked with Stanford and we talked about, hey, how do we actually design a study where we can show how this technology works in the wild, and release the feature kind of in a study environment?

[11:53.0]
Which, by the way, again, was something we've never done before. And the intention was so we could understand the false positive rates. The intention was so that we could understand what the impact on healthcare would be. So we got together with Stanford and the Center and the Department of Medicine at that time and said, hey, let's roll out this, let's figure out how to do this.

[12:09.9]
For a number of reasons, and this is one of the reasons I love Stanford, we were able to get it done quickly. We published on it quickly. We launched the feature as a research feature and then got FDA clearance pretty shortly after and then launched it the subsequent year. And we got Apple to get comfortable with doing studies like this, which now we have three large public studies in cardiology.

[12:31.3]
The Heart and Move Study. We have the Hearing Study, which led to our hearing aid feature, and then, the Women's Health Study. And now it's like something that we're, like, very comfortable with. But that was a journey. Right. And white papers, which were not a thing before, we actually have white papers on the website with some of the data behind some of the features.

[12:48.0]
And we can directly thank Euan for that because there was a number of emails of like, the heart rate's not right on the treadmill. I remember those emails. I remember those. And, we do white papers on, so, for example, for the hypertension feature. Well, when we release features, we release the white paper for the medical community to look at and then we will ultimately publish on it as well.

[13:08.3]
But the way the hypertension feature works. And the reason it's so important is just to give you a little bit of inside baseball of some of the things we think about when you deliver at this scale is what this does is it looks at the PPG signal, which is the green light on the back of your watch, looks at your vessels, basically every beat of the heart.

[13:27.8]
The blood flow, we look at the blood flow and its impact on the vessels. We gave them traditional blood pressure cuffs and we gave them a watch and we kind of correlated the signal. And so over years we've done machine learning and algorithmic work to be able to get to a level of being able to say, hey, you have possible hypertension.

[13:44.0]
Some of the trade offs we had to make was when you are notifying people at a large scale, you really need to think about again, avoiding false positives. So the trade off between sensitivity and specificity was really key for us. When you get a notification, our specificity is about 92%.

[13:59.5]
You will more than likely be in stage one or stage two. But the trade off we had to make is we have lower sensitivity. So we're at, in the 40s, on our sensitivity. And so it's interesting because the reason I share some of that is like we have to have these conversations at the highest levels to explain why we're making these trade offs, why we're not at a higher level on certain numbers.

[14:18.2]
And then when we put it through regulatory clearance, we obviously get very public about it. But I think what's neat about this product is you will get a notification and then we lead you to the blood pressure log. So for two weeks you can actually measure your blood pressure, put it into the log and the hope is then that's when you go to the doctor.

[14:36.2]
We didn't want people rushing to get to the doctor right after the notification. So you all will have to tell me if that worked out as intended. With the hope that you can potentially eliminate one of the visits of traditionally you get white coat hypertension, you go home and check your log, you come back in.

[14:52.8]
Our hope is that you could just have this information and take it to the physician and have a conversation about treatment options. The white paper shows kind of like the performance and some of the specifics around performance so that physicians understand what we're putting out there. I think the two step screening is a really important part that was also part of the atrial fibrillation detection, that it's a PPG sensor, but the ECG is available on the wrist to do a better job.

[15:17.2]
And then the other part, and I get asked a lot about this when I talk about the Apple Heart Study and atrial fibrillation, is that how do you judge where to set the sensitivity and specificity? Because obviously that's a dial. It's a dial. And then you have a demographic that is, you know, again, not all that data is released, but I think we know that it skews a bit towards younger people, maybe even skews a little male versus female.

[15:37.6]
And yet the people who are at risk of hypertension and atrial fibrillation are the, you know, are certainly age wise, skewing probably in a different direction. So yeah, how do you end up making those decisions? The number of kind of feasibility studies that we do in.

[15:53.3]
And so I know the pain of recruitment very well. Because we will run studies in certain subpopulations to see what our performance is. And when we see the through line across those sub studies, then we kind of make a decision for our clinical validation studies of like, obviously you need to have a, a balanced demographic when you're going for regulatory clearance anyway.

[16:14.2]
It often honestly comes down to a lot of the trade off of performance. We look at performance across subgroups of ethnicity, skin color, age, weight, all of those things play a role and we try to find where do we feel like our performance is most consistent with being accurate if you get a notification.

[16:33.9]
So what do I mean by that is we tend to lean towards trading off on sensitivity and this is the first time we've ever had to do it at this level to be candid, because like for example, our sleep apnea feature, the sensitivity is in the 60s, it's about 68%. But we did that because as a consumer company, if you get a notification, we can't have you not believe that notification because we're at scale.

[16:55.9]
Right. And we're talking directly to the consumer. There's not a physician on the loop. And so a lot of times when we notify you, we lean towards dialing, towards the perspective of if you get a notification, it's accurate. Yeah, the highest, best, highest best. And that is because we need you to believe what you're getting and we need physicians to believe it too.

[17:14.2]
We actually don't want to unnecessarily notify people. Well, and if the baseline is that this has never existed before and this was the case with atrial fibrillation and for this kind of hypertension predicts the same, then any cases you pick up, you could argue. That's exactly the conversations we have. Like the way we talked about hypertension is when you look at it worldwide, a billion people, about 50% of them don't know they have hypertension.

[17:33.5]
And so the idea was those people are not getting anything. So this is at least better than zero. And so it's interesting because a lot of times reporters will ask us about, well, it's a sensitivity this, but you think about it, there's nothing out there like this. This is a pre screening mechanism anyway and you're just getting it by wearing your watch.

[17:50.0]
So it's at least notifying somebody and having impact on a certain population. Is this a philosophy that you're generally thinking as you develop for the next five or ten years – I mean we talk about it a lot here – moving from a disease care system to a health care system, being proactive, preventive.

[18:06.2]
Is that part of the philosophy you bring? 100%. Those are the two words that I actually talk about a lot. I think we, by virtue of being with you, we have the ability to sense things. And so the more that we can get you to pick up disease before it progresses, that's really our goal. And we're very focused on being proactive and preventative around your health.

[18:25.4]
And that's everything from the sensing mechanisms, like what we did with hypertension, but also even just how we engage with you. We're trying to do a lot more in getting you to think about your mental health, getting you to think about sleep, getting you to move or more. How do people start caring for their health as young as in the 20s and 30s so that they can have less disease later on?

[18:46.5]
I think one of the challenges is we're reactive in healthcare, not proactive. And the one thing that we feel like we can give you because we're in your pockets or on your wrists, is driving you to one, be more proactive and preventative, but also do it in a really delightful way. We work really hard from a design standpoint to not make it scary, not be anxiety provoking, make it kind of a fun moment to engage in your health care.

[19:09.8]
If you think about it, I think one of my things, as a physician that I've always found frustrating is that people really don't want to deal with their health until something's wrong. And I would love for people to find health more interesting to engage with on a day to day basis so that we can avoid something going wrong.

[19:25.0]
And that's kind of how we design. How do you balance the idea that you're providing data to the patient or future patient with the idea that we also are all here trying to doctor our best to the patients who come through our door. How do you think about communicating data both to the patient, your consumer, the person who's bought your device, but also to their doctor?

[19:45.8]
We really think about it as a dyad and I think that's one of the things that I'm most proud of at at Apple. And I think it's actually different than a lot of other companies, is that we think about the dyad between the patient and the physician.

[20:01.9]
We're not looking to replace physicians. Now we may be biased because it's a group of physicians on our end that are making some of those decisions. Like what you see here is like our physician reports. With all of our consumer facing products, we design a physician report and the design principle behind it is that you should be able to look at it for 30 seconds to a minute as a physician, and understand what it's trying to tell you.

[20:20.8]
Because we don't want to take a lot of time, but we need it to be digestible and we know patients are going to want to share this information. So how do we make that a delightful experience for the physician? Sadly right now it's PDFs because physicians still use PDFs and faxes. But we're going to figure that out with the EMR system.

[20:38.7]
We've got to get rid of those faxes. You got rid of those faxes. But we do really think about, we focus a lot on what we're telling you to the consumer, but we want it to enrich your conversation with the doctor. And so we want to make sure that the physician is just as informed on their end about what the product is.

[20:56.5]
And when they get information, it's easy to understand and it's not a nuisance. It's more of like, oh, this is additional information. Just to give you like one kind of analogy that I use is if you think about all of our lives now we take snapshots of our lives with the photo and the camera, right? So we think of watch and those features as more of those like snapshots of your life outside of the clinic so that you have a more holistic picture when you're sitting down with a patient of what's going on with them.

[21:21.6]
And so that's really how we think about it to help you with that. How do you know when something's ready to ship? That's a really good question by the way. So there's one. How do we decide what we're going to work on? Right? That then leads to how we're going to ship. So we kind of have three core principles when we build features. One, we do want to have impact overall.

[21:38.0]
So that impact at scale is really important. So we think a lot about that. And it also ties to, I will not, I refuse to have us do technology with use cases. Like, we want to solve a problem with the technology. So that also drives a lot of the thinking. And we try to use differentiated technology.

[21:54.1]
So like, we also really do think about what can we do that's differentiated with our technology. So those things are important parts of the puzzle. The other three core principles that we always think of when we think about products in specific is one, it has to be scientifically based and validated. So we need to be able to stand behind the science, behind what we do, so that we'll always come back to that throughout the product process.

[22:16.9]
We also want to never give users data for data's sake. We want it to be actionable information. So that goes into the decision of what we're going to ship. So like, is it something I can actually tell you to do? You know, a lot of times people will be like, well, can't you tell me? Like with your heart rate being elevated, can't you tell me I'm sick? And I'm like, no, I can't necessarily.

[22:34.0]
Because it could be a number of things. And so we want to make sure it's actionable. And then obviously we go to great lengths sometimes to the expense of our speed and privacy being core. So the one thing to point out to all of you is that all of this processing happens on your device.

[22:49.1]
Apple sees none of it. We do not collect data in a cloud. And that's the reason our studies take so long is we actually never see identifiable data at all, for customers. And so that's a key part of our process, in deciding what we're going to do.

[23:06.0]
Can we actually execute in doing that. And then to make the decision on what we ship? To be honest, we have a lot of the conversations around sensitivity and specificity. Do we feel like we've hit our specs at a certain level and then also do we feel like the design, ultimately we can provide a design that's clear, understandable, delightful.

[23:25.7]
And so we kind of go through the engineering process. It starts with feasibility, then it's actually testing it. Then we do a little bit of the mock up of the designs. Do we feel good about that. We then go into product development, and then if we've hit our specs in our validation studies, then we decide, yes, we're going to go ahead and Okay to ship is what we call it.

[23:41.2]
With all of the intellect sitting in this room, like, you guys can drive your own destiny. So don't feel like even though the world may be kind of, there's a lot of challenges today, I think you have to find those little, turn those challenges into opportunities and drive yourself to be uncomfortable, to make change. These are really inspiring words.

[23:56.5]
And, I just, We've come to the end of our time. Just to remind everyone you are still part of our department. Yes, you're trained here, but you're quite literally part of our department. So thank you so much for coming to your department. Thank you so much to chat to us about these really exciting things that you do.

[24:11.8]
Thank you. The preceding program is copyrighted by the Board of Trustees of the Leland Stanford Junior University. Please visit us at med.stanford.edu.