Unlocking the Future of Health: How Tech and AI Are Transforming Patient Care

In this special episode, recorded onsite at Aspen Ideas: Health 2025, Mohit Manrao, Senior Vice President and Head of the US Oncology Business Unit at AstraZeneca, sits down with Karen DeSalvo, Chief Health Officer at Google to explore how cutting-edge technology—including AI and Google Cloud—is revolutionizing healthcare. Discover how these advancements are making care more personalized, accessible, and effective. Take a behind-the-scenes look at real-world innovations and learn what the future holds for patients in a world where health meets the power of tech.

Transcript

Unknown: 0:04

Brendan, welcome to That's understandable. I'm your host. Brendan McAvoy, US head of corporate reputation and digital communications at AstraZeneca, as we kick off for third season, we're introducing a new twist. Some episodes will be hosted by my talented colleagues bringing new perspectives to our mission of making healthcare and science more understandable. Today, I'm excited to hand things over to Mohit man Rayo, Senior Vice President and head of the US oncology business unit. Enjoy the episode. Hello everyone. Welcome to another episode of That's understandable. And I'm here on ground at the Aspen health ideas. And today I have a very special guest with me, Karen. And Karen, great to have you here with us. And thank you for joining me to begin. Could you please introduce yourself and talk a bit about your current role and how you got into this field? Well, thank you for having me today, mohith, I'm really excited for the conversation. It's great to be here at Aspen health Ideas Festival. It always brings new ideas, new perspectives, into my brain. So this will be another fun way to have a conversation. So I am currently the Chief Health Officer at Google. I'm actually the first Chief Health Officer we've had at the company. I've been there over five years, and my job has been evolving since the moment I walked in the door, but in essence, I'm leading a global health team that works to unlock our technology and platforms to help everyone, everywhere live a longer, healthier life. That means we work with governments and healthcare systems and pharma partners to use our AI technology, use our cloud platforms, but also many other ways that we that we help advance their goals. And the same thing for individual people every day who come to us on surfaces like search. I came to this job because it was a place where I could apply all in one role, the three worlds that I have had professional experiences in across my career, the world of medicine, the world of public health and the world of technology, and so every day I get to blend those worlds and think about how together they can help create for people all over the planet. Your experience and skillset is amazing, because I kind of think about the golden era of science we are in. But science can also be accelerated today with the technology that is unlocking the power, and if the public policy is right, we can have exponential impact. And I think you having had an experience in all those three and taking a role at Google, which was so instrumental, and I want to come back to it at the end on your leadership lessons there, it's been great. But today for our audience, I want to start with you mentioned about AI and emerging technologies. The last decade has seen a huge shift in how consumers digitally engage, and here in 2024 in the United States, half of the population at least had one telehealth visit. From your vantage point, what are some of the opportunities and challenges you've seen from that evolution of tech in healthcare. You know, I started medicine in a time in the early 90s when we still had digital pagers, you know, that spoke to us or bleeped out numbers to call back. And so I've seen quite a lot of technology introduced into medicine across my career, the internet, smartphones, cloud devices and AI increasingly in these last few years, I think that the view that I have from being in a company like Google that I don't think I saw when I was working in government or academia or nonprofit world, is how much People actually begin their health journey with digital tools. So most people start their health journey online on a, you know, a digital surface, a question and answer tool like search or some of these new AI interactive bots that people are using Ours is called Gemini, but but around the world, we get about 500 million health questions every day from people who are asking about something, you know, their own health, their family's health, what they're trying to get more clarity about what they were told when they went to the doctor's office. And I think it really helped me understand that by the time they someone does arrive in the doctor's office or in the hospital, they've been trying to understand and learn, and that, I think helping the health system recognize that, be respectful of that, feels like a responsibility we have at Google, but I think it's all. Also a way that we try to help, help the healthcare system and all of the health ecosystem understand that, that there is knowledge work that we're not meeting, that need, sometimes on the healthcare sector. And so people are using digital tools just to get information, much less what you mentioned, to connect with the care system, and that is, you know, both synchronous, like with telehealth, or asynchronously communicating with clinicians. I think what we're the world that we're in already now is one where health systems of all of all ilk, but especially those are providing care, are starting to use technology bots to interact with people between the care experience or to help them navigate the care experience. So it's not even asynchronous communication with a human or telehealth visit. It's actually a circumstance like a Hippocratic AI, which makes a nurse bot that reaches out and does an outbound call to a patient to see how they're doing after a hospital discharge. And they're not the only company doing that. We're seeing lots of that grow around the world, and I think, all to say, the opportunities that we've that we've been seeing, especially in the last few years, are that we're getting better at meeting people where they are, because they were already trying to come meet the health system. They're trying to learn about new medications they were on. They were trying to figure out where the science was, where they can enroll in a trial. It was really hard. It's gotten easier with the admin and the Internet, with the digitization of their records, of other records, they have a patient portal, but also other ways that the system is now starting to be able to use those tools to reach out and help people, even in their own home, 500 million is a number, and that is absolutely kind of a signal of the unmet need, of how we people are looking for information. And knowledge is king, as it relates to you know when you are going through any treatment and you want to understand more about it, and empowered patients lead to better outcomes, for sure, having said that, you also talked about med Gemini, and I know alpha forward came from Google's shop. So from that perspective, what are the challenges of integrating these healthcare tech solutions or AI innovations in the real world, whether it is at the hospital side, where you have experience in managing a health system, but also at a patient level, like, how do you how do you integrate that? What are the challenges there? So this is a one of the very interesting things about a company like ours being involved in in building solutions for consumers, but also for enterprises. So on the consumer side, patient, pre patient, pick the language, and then for individuals, we are because of the scale of the several 100 million, because of the more than 200 billion views of health videos we have a year on YouTube, because of all the interactions we have on surfaces like maps, we're we're getting a lot we get a good feeling of the anonymous experience of people, and what are the priority questions that they're asking about? How hard is it for them to find an answer about a new condition that emerges on the scene, like COVID? And so we think about ways that we can we can make that easier, but we're able to do that almost in real time. So our user research for consumers is constantly happening and iterating. And it's a, it's a, it's quite exciting and great. And it was, I think, during the pandemic, which was the early part of when I was at Google. I just couldn't believe how helpful it was. Because when you're in public health, I had been, you know, on the front lines in public health, you like, what are the questions people are asking? How can I help them get more information? How do I help them navigate to the right sources? We as a as a company, are able to amplify what the CDC or the local health officer or any organization is trying to get out to folks and help people navigate because they're starting on that, that digital platform. But we can also see like we work with, say, a global entity that's trying to provide information like Gavi, which did vaccines? We were able to have our marketing teams work with them and to create ads on the search homepage that were interesting and invited people to click and navigate to GAVI to learn more about vaccines. It was one of our philanthropic programs that we did during during the pandemic, and so we can take what we know about how users interact with the platform and apply that to a public health good. It's harder on the enterprise side, because we don't run hospitals. We don't, you know, do directly research into building new new pharmaceuticals. We're not going to we'll use alpha fold in the wet lab, if you will, but we're not using it in the real world. And we learned pretty early on that we need to not only build AI that works in the laboratory, if you will, that works in in the research environment, but work with partners to put it in the real world. And one of our early examples of that was some work we did on. Mammography with one of the trusts in the UK, where we built a model that could read a mammogram, do that on par with a with a radiologist. But what was needed in the UK still is, is a second reader. So if you have a mammogram, you need to have two, two readings of it before the woman can get the report, but they have a shortage of radiologists to read mammograms, so they said, Could you help us, you know, figure out if we could apply this in the field? And we learned so much by just saying, Okay, we built the tool. Now let's put it in the, you know, the environment, and see how the clinicians respond to that as a second read. Figure out when to push that information into the system. What's the right timing? We learned other things with related work that we did, like with Northwestern that reading mammograms to help a radiologist, you can do you can have the the AI read it, and push the abnormal scans to the top of the cue, so that's the first thing, first ones that the fresh eyes of the radiologist read and or they get read before the the patient leaves the the radiology suite. So if they have to do additional views, or do ultrasound, whatever, they're already there. So those are examples of how we we test and learn it in the real world. I want to say maybe, maybe, if there's time, I can give you a more detailed example about Gen AI, but we have a work, some work in the field right now, with a medically tuned model called on me, which does essentially Q and A with people about, oh, you have a fever, have you been traveling, etc, and we're doing that with one of the Harvard systems in the real world, because patients don't always read the textbooks. That's what we learned in medical school, right? And so it things show up in different ways in the real world environment. So we'd never say it's done when we've just built it in the lab. But you know, I think the one of the things that I have been really intent on since I came to Google is to build health expertise for our health team. Because we support all of we build health use cases into all of our products at the company, all the ones I've mentioned, from search all the way through to cloud and in the research teams that underlie all of that work. And what I need for our team members are people who have experience in coordinary academic environments or in small clinics. I need people who have worked all over the world. I need nurses. I need clinical psychologists. I need physicians. I need people who are regulatory experts. So I really have looked for people with with different walks of life, you know, who have these different lived experience, because they also will really understand the environment. You were sharing a little bit about your background earlier, and it reminded me of one of the physicians I have who grew up in India, ran a rural hospital there. Then he worked in the Middle East for a while, then he worked in the UK for a while, and then he came to work in the US for a while. And so, you know, it's just Joel. Is his name. He has this really interesting global perspective. Then I have others who have gone very deeply in the Japanese market for years. And so they really understand research and work there. So I'm counting on our team, but we know it's a beginning, not the end. So we have to work with partners, and then we have to see how the how these tools work out in the field. No, absolutely. And I think diverse people in the team who have had different lived experiences, they help us innovate and push the boundaries to learn more and adopt more. And technology can be actually a leveler, whether it is, as you said, in terms of driving efficiency or filling in infrastructure gaps that exist, it could help leapfrog systems and and you talked about partnerships. I want to also come back to we just came from ASCO, and ASCO is is kind of the Super Bowl for oncology, where new clinical data gets presented. And every year, plenaries get presented, there, where big data presentations happen, that will transform care. And I kind of say, I wish that transforms care is for everyone tomorrow. It doesn't at zip code level, gaps could exist, and whether it is underserved population, whether it is vulnerable populations. So at Google, how do you think about partnerships with or initiatives that help address specific vulnerable populations, underserved populations, with the needs that they may have with regard to driving equitable outcomes for those or education for those. It's a priority for us, and we build with inclusion by design. On the other hand, we also know that sometimes you have to build customized solutions, because one size won't always fit all. In fact, ASCO is one of our partners. That of cloud was there and talked about some work we're doing to help bring make it easier, to bring evidence based guidelines to the fingertips or brains of the clinicians serving individuals. And our CFO was there, and she talked about her lived experience as a cancer survivor, and some work we'd done with the American Cancer Society on the consumer piece of that equation to help them find the information they need. So you'll often find us on both sides of that of that equation of helping the consumer and the clinician. So I mean to say that. That inclusion by design and that we're thinking about how to help everyone everywhere find health on their own terms, is how we say it. It's it's yes about vulnerabilities based upon things like geography, rural populations or things like income or literacy levels. But it's also true that sometimes we want to build solutions for highly vulnerable populations or help protect them from from finding misinformation on some of our surfaces. I'll give you an example of that. Mental health is has been a priority for us, you know, in all of our areas. So we think about, okay, someone's going to come on YouTube or search or platform and look for information about how to commit more effective suicide or how to if they have disordered eating, how to find ways to eat a 300 calorie a day diet and and what we don't want to do is satisfy that user intent that. But typically, that's how our company is built. Like, Oh, you're asked a question about Greece, we're going to get fined to the answer, but we want to interrupt in those cases, so we use AI to try to, especially if the question's not crystal clear of what they're asking to see. Oh, this is a person who seems to be trying to get more information about suicide tasks or something like disorder eating. And instead, in say, suicide, what we would do is not help. We would not direct them to find the information they're looking for about the more effective ways to harm themselves. But we interrupt that user intent with a with what we call one box. And in the US, for example, that would be 988, but based on the Geo, we would that's big box with the big numbers shows up in front if you need help. This is where you can call or you can do a chat and sort of give them, give them that information. So when we think about vulnerable populations, we are thinking about the whole as much as we can, the whole spectrum. But we do prioritize life and death in those kinds of circumstances, and we've done that, for example, with Gen AI and our new our new services like Bard or Gemini, is what we call it, because we want to make sure that as we're putting out new tools and technology in the world that we don't, we don't always want to just satisfy what people need. There's a flip side to that, where we also will work with local organizations, local governments to try to understand, hey, how can our tools be more useful on the front lines? We have some ways. We've done this with Android, something called Open health stack. We work with frontline community organizations that are serving mental health or other needs. We're not going to build the tool, but we'll help them build it so it makes more sense for them on the front lines. Maybe the easiest way to say it is that this is major priority for us across all our products, in all of our areas of work as a company, but in health, major priority, because it is about life and death, and because if you don't attend to it when you're creating the model or creating a product, you could exacerbate disparities, and tech actually has the potential to eliminate disparities, And that's the direction that we would want to go as a company. That's great, and that resonates with me. Karen as as part of my second hat I wear. Of course, I run the oncology business for AstraZeneca in the US, but I also chair and run our foundation, and we work with a lot of NGOs on the grassroot level to understand the barriers, and use local ambassadors to educate on those barriers or provide local solutions to drive that. Because to your point, we can, we can. We cannot solve everything alone, and we need to depend on partners, and those partnerships are key, but also the right point of care intervention, like you gave this example of people wanting to find information about harming themselves. That's where you need to intervene. And intervene with trust is where it makes the biggest impact. So absolutely, I think the technology could be an enabler with a trust that comes alongside from if they hear from those whom they trust. So that's something I think we are working on. You also, can I just this trust thing? I was just looking at a survey that Edelman did, and that still surprises me, which essentially said that about a third of people, this is 116,000 people, are surveyed in 16 countries around the world, and about a third of people said that there they are more likely to trust something they read online and or something they hear from family and friends than their doctor. And in some age groups like this, 18 to 34 it's about almost half. It's like 45% so this notion of trust, by the way, going back to medicine, like I always thought, coming up in medicine, right, that patients are going to trust doctors and but we certainly saw this increasing in the pandemic, in this surveys from just a few months ago. And I think it's this word Trust has different meanings for different people and and so I think your your point about partnership also matters, because sometimes people are going to trust you. Their local pastor. Sometimes they're going to trust their doctor. Sometimes they're going to trust a sports figure. I think all the ways that we can get the right information in people's hands to help people navigate is certainly what we'd want to do. But I think what that I've had to expand the way I think about who people trust a lot in the last few years, I have to sometimes tell my mother what she reads on a messaging service is not maybe the only thing, and she needs to double check, but yes, that's true. At the same time, you've had the opportunity to work in government, both at state and federal level, and talking again about partnership and building trust. How can health systems, industry, government, communities come together to kind of drive not only trust, but impact and bring especially this technological wave of innovation to life. You know, it's, I think, coming out of the pandemic, globally, there's been more distrust of traditional institutions. And it is, I think it's something that we're all going to have to lean, lean our heads together, and think about how to restore some of that trust. You know, when I was a local public health officer, if we had in New Orleans, we often had challenges with our water plant, and so we didn't need a water we'd have to have a boil water notice, because there would have been a power outage or something. And I used to always think that the most important thing was, was that when I went on the television, because that was, you know, whatever, back in the day, or the radio, and said to people, boil your water, they needed to trust me and do that. And so I needed to be really careful about everything I ever said to them, because I didn't want it to ever get gray, because that was become sort of life and death, especially for some some populations in that environment. And I think the part of the reason the pandemic, I think, was hard for folks, is because information did change as we learned about the virus, as we learned about transmission, as we learned about new therapeutics, things that evolve. But I do believe it can be restored, but I also believe that we have to recognize sort of like this Edelman survey results that I mentioned, which is that that you cannot take for granted that people are going to trust the same resources and sources. Some people will trust government. Some people won't. I one of the things that we did learn a lot during about during the pandemic was that that it's not just the message, but sometimes the messenger, which is essentially what I'm saying, about who to trust, and we learned things like, if people are already watching a sports figure that they trust, and we can and that sports figure has an interest in an area could be something like infectious disease, like the pandemic, could be mental health. We've used it in those spaces, and we can provide them with evidence based information that comes from not we Google, but, you know, link, link them up with the CDC or the NHS, or, you know, whatever government around the world wants to work with them, then they can be the messenger, and people are going to receive that, and and and trust it. Well, it, it sounds simple when I say it like that, but a lot of medicine still feels like they should be the messenger of the message. And I think the reality is, is that we have to let that go, because the most important thing is people hear the evidence, hear the hear the right data driven, you know, information and decisions and who delivers that? We have so many more options, basically, in this day and age, if it's done, if the pairing is done properly, with messengers on the different surfaces and places where people receive care in public health there are, there's not a new idea. This goes back, you know, I worked the Ebola outbreak, and it's very similar thing. You know, just couldn't be that it was what the World Health Organization wanted to tell people locally in the community, they had to find local faith leaders, etc, to help carry that message. We see this with community health workers all around the world, who are often the most trusted individual. So it's a flexibility notion that I think we're all going to have to get comfortable with. But what I do hope is we'll, as a health ecosystem broadly, start to do a better job of having a shared set of facts, even if we were comfortable having separate messengers. Yeah, and I think that message on the right messengers for the audience, where they connect well with and trust is key, because I think the intent is everywhere good. But if the messenger is not the right one, it may drive people in the opposite direction. Actually, we, as part of AstraZeneca bold ambition to one day eliminate cancer as a cause of death. We believe we cannot do it alone. We have partnerships. But we also believe not only pharmaceutical interventions will drive it. We can drive it by also driving early detection of cancer, and we've got a partnership with NHL, where actually the NHL and NHL, you know, the player Association, and everybody involved in that, are the voices sharing their stories of how early detection helped them to communicate to the hockey community and then encourage them to talk to their plus one to go and get screened. So it's. Yes, it's society has a role to play, but finding the right messenger so it absolutely brings home the point that there may be more trust there, versus XYZ companies standing up and saying that, go get your screening done. So I do relate to that. And I think Google has is in a position where people do do searches and trust a lot of this and the work you just talked about with the information is really key. Yeah, and even there, we've learned we recently started out a thing like people like you, because people, when they search on a topic, it's easy on YouTube, because they're looking for people like them, but on search, they also want to find people who have the condition they have and have had the experience. So we've done a thing with Reddit where we can connect folks and people really that experience is really positive, because they can all sometimes they just identify with another individual who has the condition. But I love this NHL example. I want to hear a little more about it later. Yeah, okay, just don't talk to me about hockey terminology. I'm still picking or me either. I mean, I would say yes one, one of the things we said is, go get your body checked against cancer, which has been very successful because the hockey community relates to that line, yes, and we are trying to do more with that. And again, everybody has a role to play. Talking about role you started off with introducing and talking about, when you came to Google, you were the first Chief Health Officer. You started that in 2019 and very recently, you announced that you're going to retire as of August 1. And I'll come to that what you're going to do after it. But as you think about all you have accomplished, what advice would you pass on to your next person, but also people listening in from your leadership lessons, not only at Google, but broadly across clinical government and all the work you've done today. Well, you know, one of the things that I will do in retirement is I'm gonna do a lot of processing. It's interesting, you know, I've I about what I've learned. I'm one of those people that went from thing to thing to thing without break. And I started working when I was in high school, actually, my first job, I was 13, I taught tap so I could pay for my own dance lessons, and I paid my way through college. I, you know, sort of, I've been on a go, go, go. And I think I'm gonna, if you ask me, in a year, I'll probably have processed a lot more. But what I would say is, I think that, I think people need to enjoy the journey. Sometimes people get really fixated on, I want this job or this role at some point. And I sometimes I get asked, Did you ever think you'd work at Google and be the Chief Health Officer? And my response is, there was no Google when I went to med school. Therefore I would have never imagined. You know, there's so many things that have evolved that I would have never imagined are ways to help people, which is the second category, and that is for me, the choices I've made about where I've gone, where I've said yes to taking on a role as a place where I felt like I could help people. And medicine very much was that way I practiced for 20 years. I loved caring for patients, but I always was thinking about the systems in which they were trying to navigate, and how, frankly, they were more built for doctors than for them. And you know, how could you change that so that it was really built around the people we were here to serve? And that story is repeated for me over and over and over again in my roles of what are the ways to get information in people's hands, like when I was national coordinator, and changing the way that we got gave people access to their electronic health records, or ways that I get to do that at Google, where we give more knowledge and insights, and I'd say power to people about their own health. We're building personal large language models that that our hardware, like Fitbit users, can have to help them know what are the right questions to ask when I go into the clinical environment, or what are the ways that I can get better sleep? That is that side of the equation, the patient, the consumer, is the part that I think it's left behind a lot when we're designing systems or thinking about the best way to pay or do technology or implement Gen AI, and we often think about, what's the what's the way that we're going to make the doctor's life better? Now I'm a doctor. I'm married to one. I love doctors. Don't get me wrong. I taught a lot of them, however, we're just a piece of the puzzle, quite honestly, and I think I do hope that of the things that I've learned, and I'm seeing more of is that people are realizing that it's team based care, and part of that team is the patient or the really, most people don't want to be thought of as a patient. They want to be just a person, right? And they want to be healthy most of the time. And I think that this era that we're in with Nai is allowing more of that. So it's very exciting. So journey focusing on what you really care about and what you choose east of your job. And for me, it's been like figuring out how to equalize some of that knowledge and power so people have more access to where they want to navigating their own health journey on their own terms. And I think, I think the other piece that's been thematic for me is it's really it is possible to both do. Good and do well, that sometimes people think it's two choices. Oh, I can go work in private sector. I'm gonna go work in the nonprofit, and I've done all that. But I also know that that's very true, that you can help people, but you can also do that in a way that that has makes the business a going concern. I'm gonna give you a specific example about that that was taught to me by patients after Katrina, Hurricane Katrina in New Orleans, we were remaking the health system there. It was a one that was very much about hospital based. Most people were uninsured, so about a quarter of the Louisiana populations, about a million people were uninsured, and most in New Orleans area, most individuals. So most of the time, what happened is, if somebody was sick or had a chronic condition, even they used the emergency room for care because there was no community health so we were taking the opportunity from the tragedy to build a community health network and infrastructure. And after a little bit of time, I had one patient say to me, you know, it was free for the beginning, because we had a lot of philanthropy and volunteerism, and we were trying to figure out how to give people access to community based, neighborhood based care. And a patient say to me, y'all should start charging. And this is a service worker in New Orleans, very low income, person without health insurance. They said, Just charge something, because if you don't, you're not going to be around to take care of us in the future. And this is our medical home in the neighborhood, and we want to keep it and so I think, predicated on that, some other conversations we had with the community organizations, with when we worked across the area, we did start charging a little bit on a sliding scale. And people really felt a sense of dignity. They felt a sense of ownership, and when they could come in and it was $1 or whatever it was, and, and it was a good lesson for me about also, you know, you think you're doing good for the community, but, but they want to be a part of this. They want to feel like they've got some some knowledge, that you're going to be there for the long haul, right? Because sometimes philanthropy comes and goes and they also they're smart. They know that that a business has to pay its employees. Many of their friends and family were working there. So just one example of many I've had across my career that it is possible to do good and do well, and I would encourage people to think that way and not think of it as just a choice in one direction or the other. Love it. Karen, thank you. I'm going to take that with me. You can you can do well and you can do good. I think that's the job you have right now. Quite honestly, I think this is something which will drive our audience as well. One last thing I'll say is what I took away from what you said, amongst many things, was journey is critical and destination incidental. And as we think about our patient's journey, team based care for them, thinking of them as individuals is key. Destination is incidental. I know you are a bit hiker. What's your next destination? As we wrap this up, well, I'm really looking forward to handing the baton to the guy who's currently my deputies, tremendous human being and physician and leader, and I'm very excited to feels great to be able to hand it off to someone I really trust. I'm also very excited to put on hike and choose and we have a number of trips planned in the Southwest, and then planning up a bigger trip to do the Tahoe Rim Trail, which is to hike around Lake Tahoe. It's, it's about 175 miles, so just a couple of weeks of through hike. But I'm gonna do that one as an intro before I decide if I'm taking on anything bigger. But mostly I just want to have some time to reflect and relax and and build up more gratitude for this career that I've had. Yeah, thank you. And thank you. You had an amazing career, 35 plus years of contributing in so many places. We wish you the very best. And thank you for being here. And I'm sure you'll country contribute further on, even in during your retirement, into many things and many companies on this tech journey in healthcare. So thank you so much for being with us. Thank you very much. You