In this episode of the Me, Myself, and AI podcast, Philips’s chief medical officer Carla Goulart Peron shares how artificial intelligence is reshaping health care — not by replacing clinicians but by expanding access, improving diagnostics, and freeing doctors to focus more time on patients. Drawing on her experience practicing medicine in Brazil’s strained public health system, she explains how technologies like AI-assisted imaging and remote collaboration can bridge critical gaps in care. Carla also explores the challenges of trust, bias, interoperability, and women’s health data in the next era of AI-enabled medicine. She offers a grounded, global perspective on how technology can make health care more human.
Subscribe to Me, Myself, and AI on Apple Podcasts or Spotify.
Transcript
Allison Ryder: How is one clinician thinking about applying AI to health care in additive ways that improve access to care, clinician confidence, and patient experience? Find out on today’s episode.
Carla Goulart Peron: I’m Dr. Carla Goulart Peron from Philips, and you are listening to Me, Myself, and AI.
Sam Ransbotham: Welcome to Me, Myself, and AI, a podcast from MIT Sloan Management Review exploring the future of artificial intelligence. I’m Sam Ransbotham, professor of analytics at Boston College. I’ve been researching data, analytics, and AI at MIT SMR since 2014, with research articles, annual industry reports, case studies, and now 13 seasons of podcast episodes. In each episode, corporate leaders, cutting-edge researchers, and AI policy makers join us to break down what separates AI hype from AI success.
Our guest today is Dr. Carla Goulart Peron, chief medical officer at Philips. Philips is a health care technology company behind imaging systems, patient monitoring, and a growing suite of AI-based clinical tools. What I find fascinating about Carla’s perspective is that she started her career as a physician in Sao Paulo’s public health system, where demand far outstrips resources. She’s carried that lens into the C-suite. She’s now leading medical strategy for a company that’s betting heavily on AI to close gaps in care. Carla, welcome to the show.
Carla Goulart Peron: Thank you very much, Sam.
Sam Ransbotham: Many listeners might still associate Philips more with consumer electronics, but can you tell us about the company in terms of health care and the kinds of things you’re doing?
Carla Goulart Peron: It’s a company that has [been around for] 130 years and has been in many areas, but the last few decades, Philips shifted into health care fully. We started with imaging — so the diagnostic area, X-ray, CT, MRI, ultrasound — and then we [went] into the interventional therapy, into the cath labs. We also are very much present in the ICU and any area of the hospital where you are monitoring signals coming from those patients. [We’re] also heavily invested in monitoring things outside of the hospital. And then the last piece, but not less important, [is] definitely AI that is supporting all those areas of care and then keeps growing [as] a very hot topic right now.
Sam Ransbotham: As you were listing these technologies, I was thinking, “Those are classic applications where AI has made huge advances with imaging and these sorts of things.” I mentioned earlier, you trained as a physician and you worked in both the public and the private sector, sometimes in the same day, I think, switching back and forth. What did that experience teach you about where health care breaks down in a way that maybe technology can help?
Carla Goulart Peron: You are right. In the morning I was working in the public sector in Brazil, where scarcity of health care workers, technology, information sometimes is very much present, so [you] need to work within the best of your capacity to offer care for those patients. But it’s universal health care, which means everybody has access. So there is a benefit there too.
Sometimes in the afternoon or in the night, I was working in the best of the best hospitals in the private sector, with everything available. I think that teaches you a lot of resilience, personally, as an individual, as a physician, but also gives you the chance to try different things and learn from those experiences. But [it] also makes so much clearer to an individual like me how much technology can actually build a bridge and help support more patients [who] are expecting to get access to health care overall … because it’s expediting the way we’re seeing those patients, because it’s connecting the data points of information, or even allowing collaboration across the specialties that may not be present in the public sector.
Sam Ransbotham: Was there a specific moment that you said, “This is something that technology could really help with or could help fix?” Was there anything that made you think technology might be the answer?
Carla Goulart Peron: Many times. I love sharing an example of ultrasounds [from] when I [was] coming out of residency, actually, and starting to see patients on my own. [The] ultrasound is one of the biggest diagnostic tools that we use in the OB-GYN practice. But we [did] not always have access to those machines in the hospital setup. Sometimes we had access to those machines, but we are not qualified to use them.
Technology [is coming] into reality today — I’m very jealous about the people [who] are learning today [in] their own clinical practice — in a way that you have clear collaboration. So you can really open the technology, open the ultrasound machine, get access to an expert [who] can be anywhere [on] the planet, let’s say in the same city, just to make it easier from the clinical practice perspective, to guide you, to see what you are seeing [in] that same imaging, help you to capture the right imaging, and expedite the technology. In some other places, like the ones that were practicing in the public sector, I would need to transfer that patient, sometimes to another facility, which means call an ambulance, be stuck in traffic, just to get the image captured and then [take the] patient back for you to be able to take a final diagnostic and initiate therapy. I mean that was like wine and water, unfortunately, between those two worlds that I was living in.
Sam Ransbotham: You mentioned the traffic. Not too long ago I was in Sao Paulo, and that was a big thing I remember there — just how long it took to get from one place to another.
But actually, ultrasounds [are] expensive machines that can’t move, but then there [are] also the other parts that information can move. I think you were sort of making a distinction between those. You’ve got some aspects of what you’re doing that seem to rely very heavily on sophisticated equipment. On the other hand, you also have information [that] can flow without that bottleneck of Sao Paulo traffic.
Carla Goulart Peron: I mentioned the ultrasound, which is highly dependent on the imaging that you are actually seeing on time, right? So you use that imaging on time to make the diagnostic. But when we think about a CT or MRI, they produce hundreds of images that the radiologist can see from anywhere, and that is also facilitating drastically the way that we are actually reporting imaging. Also, you are highly dependent on the users [who] are actually placing the patient into the machine, making sure that the patient is well positioned, [who] is holding [their] breath if needed, or [who] is kind of moving accordingly. And now with AI coming on board, you also can get that done very quickly, precisely, with not much support or training from the technical perspective. So that has also been a game changer.
I would say the way we are leveraging technology, AI, automation, to position those patients, to make sure that the exam can be as fast as possible, but also how we are processing the imaging that is coming out of those big machines, has been very different.
Sam Ransbotham: I think I was too simple. I was talking about the machine and then the results of the machine, but actually, you’re bringing up an important point, which is there’s also a knowledge transfer and an information transfer about how to get the best image in the first place. I think I glossed over that.
You started [with] bedside [care], but now you’re running medical strategy for Philips. How does that inform what you think about the health care strategy and how that works, your deep background in actually doing medicine?
Carla Goulart Peron: It’s interesting because I think when you go into med school, most of the people [who] decide to take this pathway don’t think about anything differently than just seeing patients on a one-on-one basis. At least when I started med school, that was the reality — I think today is very different.
But in that journey of understanding the health care system, how companies that are developing drugs, medical devices, or other types of equipment work, I [learned] that there is a role to play [on the] industry side, [on] the corporate side that can be as rewarding and as interesting as seeing patients on a one-on-one basis, probably [on] a much bigger scale. So I think I needed to convince myself that by moving from the bedside [to] corporate, I was not changing my background as a physician or I was not leaving my professional [training] behind. I was actually just applying that knowledge in a different way.
I’m fascinated by innovation. I’m fascinated [with] ensuring that whatever innovation we are investing in, in the corporate environment, can actually reach the patients [who] are going to benefit most. And that’s really the biggest part of my job: ensuring that the ideas that our engineers are developing in partnership with hospitals and physicians are going to meet the requirements of the regulation, because we need to prove that it’s safe and it’s effective, but also [ensure] that we will have a good plan in how this technology can be incorporated by the health care system in the way that can actually reach the patients [who] are going to benefit most.
It’s very different, but it’s fascinating, because I keep learning every day because it’s new technologies, new areas of care, new types of health care systems. If you think about a company like Philips that has a global presence, it’s very different to think about commercializing something in Brazil than it is in the U.S. or in Europe or in Africa or in Asia, so you need to have that globalized thought in mind when you are thinking about developing technology.
Sam Ransbotham: Let’s talk about some of these specifics. I think Philips just got [Food and Drug Administration] clearance for SmartHeart, which is an automated cardiac MR [magnetic resonance] planning tool. First, explain to people like me what that actually means, and then how does that actually change a radiologist’s day? What’s different?
Carla Goulart Peron: SmartHeart is a great example. As you think about an MR machine, it is a technology that can capture imaging from your entire body. In order for you to do that, you need to have a technician [who] understands exactly why you are actually being requested as a patient to do that exam.
In this case, [the] physician — a cardiologist most likely — wants to actually see how your heart is functioning. So imagine that a technician needs to know exactly how he or she should be positioning you on the MR table, at which angle, if you are tall or short, if you are someone [who] is big or small, if it’s a kid or if it’s a female or a male — there are so many different data points that a technician needs to understand in order to capture the right level of imaging with the right quality [so] that a radiologist can actually do a diagnostic out of it.
SmartHeart is an AI-driven, one-click automation that plans all setups that drive how the cardiac imaging needs to be captured. This happens in 30 seconds. So that sounds simple, but for the operator [who] is actually doing multiple exams in different parts of the body with different indications, that can be from 15 minutes to 30 seconds. It makes the machine a lot more accessible. It makes the technician capable of doing a lot more exams. It also reduces the dependency of having someone [who’s] highly trained. The burden on the technician is also something that today is a big issue.
Sam Ransbotham: I like the idea that we have a very expensive machine that we can increase the throughput for. You said 15 minutes — that’s four exams an hour, even without any setup and putting on the gown. But you talked about 30 seconds, and suddenly I feel like, “Hey, we have an expensive machine that we can use a lot more.”
Carla Goulart Peron: It is definitely [about the] right speed. There is a burden on health care providers and nurses, physicians overall. But [it] also [involves] “first time right.” So sometimes if you don’t have something like this, you’ll go through the exam, send the images to the radiologist, and they’ll say, “You need to call this patient back because we are missing one or two views.” With something like that, this doesn’t happen.
Sam Ransbotham: Nobody likes that. Nobody likes to go back and forth.
Carla Goulart Peron: Especially on a machine — you don’t want to be there, it’s small, and it may not be that comfortable for the patients overall.
Sam Ransbotham: Radiology is one of the areas that I think a decade ago people were saying, “Oh gosh, we’re never going to have radiologists again. The machines are going to do everything.” You know, that narrative has really not played out at all. But I think it’s a great example of, in general, how artificial intelligence might affect the future of work and what we do. But is there a risk that if this process works too well, hospitals are going to start thinking they need fewer radiologists? How’s that going to play out, do you see?
Carla Goulart Peron: I personally believe that AI is here to add, not to take over. Maybe this conversation will be very different five years from now because I think we’re learning that environment.
Sam Ransbotham: Predicting is so tough in this world.
Carla Goulart Peron: But I think [in] the radiology space for now, radiologists are wasting their time on things that are not valuable at all, reviewing images that were not captured precisely, or doing reports, or reassessing a lot of normal images. Interesting enough, a few weeks ago, I was with one of the radiologists’ medical societies, and they were talking about what if we could have AI defining all normal images, and then radiologists will be looking at only abnormal [images]? What’s fascinating about AI and [its] potential, someone in the audience raised their hand and said, “Well, how are we going to get the radiologists trained in what is abnormal if they are not going to be seeing normal?”
So I think the answer for your question is still TBD [for] what the future will look like, to your point, but I don’t see AI taking over. I think it’s actually helping us to see more patients because there is a big gap out there [and] also to be dedicating our time to things that are unique that we can do as clinicians.
Sam Ransbotham: One way I think about this is we would have a very different story if every possible patient in the world was completely satisfied with a perfect supply of radiology. If everybody that needed this treatment was currently getting it, then I think we maybe could be talking about this replacement type of thing. But you know far better than I do that’s just not the case. There’s a massive undersupply of these types of technologies, and a lot of it is driven by the certain return on investment and the cost structure now, which I believe this can change. How do you see that aspect changing, in terms of serving more?
Carla Goulart Peron: I personally believe that there is such a big gap out there in access to care that as we incorporate technology, we’ll be able to do more with the same, not with less. Because I was born and raised in Brazil and practiced in Brazil, I feel very comfortable in going [along] that pathway [in] low-income countries.
In those areas, the gap is really huge. You can be waiting in line for months in order to have, as an example, an MRI or CT scan done, right? Sometimes too late in the game, that actually can change the patient outcome. Well, we cannot fool ourselves. That reality also exists in the U.S. and in Europe and in Asia. We have deserts, right? We have areas where people don’t have any access to care. So I don’t personally see a timeline where I’ll say, “Well, this will quickly fix that gap in a way that is going to be reduced.”
Sam Ransbotham: I love the idea that what this leads to is better matching of need with supply.
I’ve got a beef to raise with you. You had this Future Health Index that you put together. Normally I like to skim these things before a call, but it was actually quite interesting. So it cost me a lot more time looking at it than I wanted. But I think part of your finding in the Future Health Index was that 79% of the health care professionals are optimistic about AI, but half of the patients are worried that it’s going to reduce their face-to-face time. How are we going to reconcile these two different perspectives from a market [where] both parties are important?
Carla Goulart Peron: I think the perspective of the patients is very important in this one, right? If the patients start to feel too uncomfortable and reject AI, it may become a challenge.
I also think that this is slowly changing. We’re going to be launching a new version of the assessment a little later this year. But for me, the ability to reconcile those two things from the health care professional perspective is ensuring that we are validating, we are getting good access to data that is actually being used to train the machines. So physicians are on board but really a little bit cautious about how much evidential bias do [they] get in the AI. Can I fully trust AI? How much do I need to review what’s happening? I think [from] the physician’s perspective or the health care worker’s perspective, it’s more towards data-driven.
While [from] the patient’s perspective, I think it’s more about the experience. And I think AI became a big buzzword, and so people don’t know exactly what to expect. There is this misperception that physicians are going to be substituted by machines. The reality is, the physician’s time is actually being freed up to be actually dedicated more to the empathy piece, to the touch, to that one-on-one, eye-to-eye, which I think is going to make a big difference.
But to the second point, where I think it’s very interesting, there are some studies already out there that show if you are talking to a real physician or to an AI version of that physician, sometimes the AI can learn how to be more empathetic than the physician. So I think this is going to be a journey of us as individuals actually learning how to incorporate AI into our lives and trusting a little bit that help that I think we’re going to start getting [in the] future.
Sam Ransbotham: In our research, we found that individuals who trust AI are twice as likely to use it regularly. That trust is an important part, but when we were studying it, we were sort of thinking in general [about the] use of AI. It strikes me that’s different about trusting Netflix’s recommendation versus health care, but perhaps there is something transferable between the consumer levels of artificial intelligence and the greater understanding there, and health care. Is there something unique about building trust in this clinical setting that is different than my Netflix example?
Carla Goulart Peron: I think people are more concerned about their health than eventually getting advice of which movie I should be watching or which series. But one of the areas that I think AI is actually already doing for the patients is enabling. I think even the name we gave to that individual [who] is in the center of everything we do as clinicians — patient — it’s almost like, “You stay there, be patient, and wait until somebody tells you what to do.”
While now with, first of all, AI enabling interoperability data points, you are giving more visibility to the overall health. Also, what are the options that those patients may have in front of them? And I think this is, in my view, going to change drastically the way that we, physicians and patients, will be embracing health care for the future, because I think [when] I went to med school many years ago, patients didn’t have a say.
It was really, that’s the protocol, and that’s what we’re going to do, and you just follow it. I think more and more we’re starting to talk about precise medicine, where patients will be able to be offered one, two, three potential treatment pathways with pros and cons and the ability to choose. And I think AI may enable those patients to make more informed decisions at least.
Sam Ransbotham: Actually, I really like the framing of the word patient because I feel like I often am not. We had Josh Weiner from CVS Health on a couple of episodes ago, and one of the things we got into was, “Hey … forget all this AI stuff; I’d just be happy if I didn’t have to put my name in over and over again and wait for a long time.” You can talk about all the cool AI stuff you want to, but let’s get some of those simple things done.
I think you’re very focused on women’s cardiac health. Maybe orient us. What are some of the specific gaps in care for women, and how do you see AI perhaps connecting and either exacerbating or helping with that?
Carla Goulart Peron: I’m very passionate about the topic because I always say we need to remind people that we’re not mini men; we are different as women, and we do have a heart, and it needs to function like any other heart in the man’s body.
There is a big gap. Cardiac cases [have] the highest mortality rate present in females. Despite that, females have a much longer waiting time until [they] get the diagnostic, because we experience symptoms differently sometimes, because most of the protocols that have been designed have been designed based on studies that included only males. So there is a big gap out there that needs to be covered.
What technology definitely can help in covering that gap is ensuring that the nuances and the differences in the physiology in the type of response that females used to present is actually incorporated into the way we are designing the diagnostic tools.
We talked about MR. The position of the heart is slightly different in [a woman’s] chest compared with the men’s chest. It’s a very detailed, minor thing, but it can impact the way you are capturing your imaging. If you are capturing cardiac rhythm, for example, the algorithms need to understand that the female heart has a pattern that’s slightly different from the male heart, and so I think AI will quickly get that information into those algorithms because of the speed, and be able to equalize that.
Sam Ransbotham: That sort of presumes the fact that we’ve got these perhaps underrepresented or undertreated populations within our data sets. We had Ziad Obermeyer of University of California, Berkeley, on the show a couple of seasons ago, and he was talking about how these algorithms can actually build up this equity. But it was depending on having that raw data to start with. What can we do to try to get better sampling on those underrepresented populations?
Carla Goulart Peron: The first thing is what we are doing right now — talk about it, right? So there is an opportunity for us to do better now. You are spot-on. Most of the drugs are developed based on a very limited number of females or even other diverse types of population. You name it.
The same thing for medical devices. And most of the protocols and guidelines that are created are also created based on trials that were developed a long time ago, which didn’t necessarily include the right level of variety. That’s the first point.
The second point is using technology and AI to capture that information pretty quickly and reiterate rather than starting from scratch. I learned a big lesson two years ago when I attended the WEF [World Economic Forum] for the first time, and we were talking about women’s health. I’m an OB-GYN, I have more than 20 years of clinical practice. Someone was telling me the story that when a woman gives birth, there is a standard that defines that 500 milliliters of blood loss is normal. So if you have that, you don’t need to take any actions.
Then I heard a question from the audience during the WEF discussion. “How was that established?” I paused because I had never asked that question. I never had the curiosity. That was in the book. I just kind of assumed that a very good methodology was put in place. And that standard was developed based on nine females in Germany, and exploited and used [on] the entire planet. Can you imagine how that translates to India or China, where there are much smaller bodies? So they quickly were able to iterate that with data points and create the correlation, and established that in India, that number should be 300 milliliters. So that makes a huge difference in how you’re going to be treating your patients.
I think technology, that interconnectivity, that not only the fact that we’re going into automation, but that now AI can analyze such a big data set so quickly, can really improve the way we are practicing medicine.
Sam Ransbotham: Actually, that’s an interesting example because it strikes me as a slightly different approach, which is, in your example, you had to go to the World Economic Forum to have that question get raised. But somehow I can also imagine a very simple job for agents would be, “Hey, go through all of our clinical practices in every area, and find the root study for that, and assess how that plays out.” I would feel like your nine people in Germany sample should rise pretty quickly to the top of that list. That seems pretty exciting.
Carla Goulart Peron: Pretty exciting, exactly. It’s fascinating what can be done.
Sam Ransbotham: Let’s look forward for a minute, though. You started in public health in Brazil. There’s a lot of resource constraints there about technology. If you could pick one AI capability to deploy globally that would make the biggest difference — you get to choose right now, wish list — what would you push out to the world? What do you think [is] the greatest application of the use of artificial intelligence in health care that we could push globally?
Carla Goulart Peron: Interoperability. That is going to change completely the way we practice medicine. Because today we’re very much closed or restricted to the health care system that you are operating. So the ability to see the patients longitudinally without those barriers, I personally believe, is going to change outcomes significantly. If I need to pick one, that will be my choice.
Sam Ransbotham: That’s not at all where I thought you’d go with that. That’s pretty fascinating. It’s so cross-cutting, and it affects everything. All right, so I’ll bite. What’s the biggest barrier to that actually happening?
Carla Goulart Peron: That’s a big question. That’s why it’s a dream. You said you can pick anything, don’t be restricted. I think there are many. The first one is making sure that we have access to good quality data, but also that we start thinking about that from the get-go. If you don’t have some level of standardization, it’s very difficult to think about interoperability. I think that’s the first piece, which is science. It’s how we drive this for [the] future.
The second is how we think about incorporating this new era. How do we think about incorporating reimbursement and access to technology into the discussion about AI? We are still restricted by the reimbursement systems. What kind of code coverage do we have? What is the incentive? We may have something that can actually pretty quickly take the patients out of the hospital, reduce lengths of stay. But if that’s not the incentive from the health care system perspective, this is not going to happen. So for me, that’s another big thing that we need to think about.
The other piece, which I think is how we need to partner, is regulation. Regulation will need to evolve with this new environment that we are getting into. The type of regulation that brought us here is not going to take us to the future because the future is very different than the one we are playing today.
Sam Ransbotham: That seems really hard because, yes, we built this regulatory system and we built these reimbursement processes, and so many processes we built off of the way things used to work.
Thank you for joining us. I think what really comes through so clearly is that this isn’t abstract for you. This is something that you know deeply, and you’ve seen what happens when care isn’t accessible. And I liked some of the ideas that you mentioned about how do we make the use of technology, not just AI, but technology in general, part of the solution and not just sort of a headline. Thanks for joining us.
Carla Goulart Peron: My pleasure, Sam. [I] really appreciate the conversation.
Sam Ransbotham: Thanks for tuning in today. On the last episode of Season 13, I’ll be joined by Bernard Hampton, a corporate learning leader at Bank of America. Speak to you then.
Allison Ryder: Thanks for listening to Me, Myself, and AI. Our show is able to continue, in large part, due to listener support. Your streams and downloads make a big difference. If you have a moment, please consider leaving us an Apple Podcasts review or a rating on Spotify. And share our show with others you think might find it interesting and helpful.