FDA’s Troy Tazbaz is Changing the Narrative of Healthcare Technology

Melanie Marcus: Our guest today is Troy Tazbaz, who leads the Digital Health Center of Excellence at the Food and Drug Administration.

When the FDA called to see if he'd like the job, Troy was on a cross-country bike ride, literally across the country, pedaling from West to East Coast.

His story is compelling, and it spans from California to Turkey. From tech startups to large technology companies to the FDA. From operations to transformation. And from personal tragedy to a commitment to making things better for patients.

I don't want to give any more away, because hearing the story directly from Troy is a lesson for all of us in how to manage through deep personal grief. And the work he is doing with the FDA today is incredibly exciting. As we look to new technologies, including artificial intelligence, to help transform healthcare.

It's a pleasure to have you as our guest in our final episode of this season, Troy. And I'm excited for our listeners to hear about your work to modernize U.S. healthcare. Thank you so much for joining us today, Troy.

Troy Tazbaz: Thank you for having me. It's great to be here.

Marcus: I always love to start at the very beginning to just hear where you came from. So can you share a little bit about what your childhood was like? Start us at the beginning.

Tazbaz: I was born in Burlingame, California, [and lived in] San Mateo County, California, until about 9 or 10 years old. And then I went to school in Turkey and came back when I was 17.

Marcus: What brought you to Turkey?

Tazbaz: My mother wanted us to have a little more cultural kind of experience. And so I think that probably helped shape the way I've been thinking about things ever since, including the business world.

Marcus: And then you went to study at San Diego State University. What did you choose to study there?

Tazbaz: I ended up majoring in computer science and then business. I followed a traditional technology path working for multiple startups. After that, [I had] a variety of more technical roles, and then ended up coming to Oracle where I did come in also in the technical field. But this was an interesting time at Oracle where it was effectively trying to go into the cloud space, and it was a traditional software and hardware provider. And call it luck, call it timing, call it some of my background. I was able to start combining the technical aspects of delivering the service as well as the business problems that you had to solve to become a cloud service provider. So I learned a lot of my business skills that I still apply to this day at Oracle, and that's where I would say that I matured in my current trajectory of my career.

Marcus: Right. So you were at Oracle before Oracle acquired Cerner and kind of became Oracle Health as well, right? So that wasn't the trajectory to healthcare?

Tazbaz: No. So the trajectory to healthcare actually started with my wife being diagnosed with cancer back in 2019. And obviously, she was diagnosed with late-stage, Stage 4 colon cancer. And when you're being diagnosed with such a late stage, you have to immerse yourself in the treatment plan. And of course, because it was so severe, we spent a lot of time in the hospital. And so I ended up spending incredible amounts of time just sleeping next to her and also observing the care that she was getting as a fairly young person, a 37-year-old woman. And you can see the look of many of the healthcare professionals that are not used to seeing someone that young and looking so vibrant, dealing with something that is as horrific as cancer is. And so, throughout that experience I started observing what was happening within healthcare, where I believe technology could play in. Again, you know, my mind automatically goes to solving a problem. And in this instance, I was trying to solve a problem of caring and helping my wife get better.

And so when she passed away, this was about three months or so before the pandemic. And I had choices to make. And generally speaking, what I mean by “choices to make” is that I was dealing with a lot of grief. And you can approach grief in multiple ways if you are conscientious about it. And I chose to—I did struggle, but I always say that grief, you can apply it. You can be destructive and do something with yourself or to your community, or you can be constructive. And I chose to try to be constructive. And that was my kind of foray into healthcare.

And I was also one of those individuals at Oracle that when something was either going wrong or needed to be incubated, often they would call and say, “Hey, can you help with this?” And so when Oracle decided to build a healthcare business unit, I was asked, “Can you incubate this—the idea—what we should work on?” And so one of the products that we said we would want to implement, and at that point design and develop and deploy, was our electronic healthcare record. And it turns out it's a very complex product to implement. Not to mention the length of the implementation cycle. And that led to Oracle deciding to acquire Cerner. So, you know, my introduction to healthcare was effectively helping incubate that business unit.

Marcus: Well, I'm very sorry to hear about your journey and your wife, but thank you for sharing with us. Unfortunately, cancer is something that many of us can relate to in one way or another. And it certainly impacts our work in healthcare.

So, you were at Oracle thinking about healthcare, but it did not go into the electronic health record. [Oracle] bought Cerner. Where did you go from there?

Tazbaz: After the Cerner acquisition, I was asked to help integrate the company itself, and I had realized that I never had a chance to take time off. Like, you know, after you're dealing with grief and then the pandemic forced you into this box. And I never really had a chance to kind of just check in on myself. And what Oracle was really good [about], is they understood what I was going through as an individual, but they also knew that in order to make sure that I don't go in a direction that was probably unhealthy, they kept me very intellectually stimulated. They understood that that is one of the things that I truly value in anything that I do is that intellectual stimulation. And so I, of course, poured my entire existence into work. And I had also this incredible creative streak at that point in the business world. And it's kind of interesting how trauma tends to drive a lot of creativity. And, because I enjoy writing so much, I thought about even writing a book around why does trauma actually result into a creative streak? And if you look historically speaking, there were a lot of very creative periods that followed a traumatic event.

And so I had this creativity. But when the integration discussion started happening, I realized that at major corporations, that integration is going to take multiple years. And I had recently found this Excel sheet that I had created almost ten years earlier around biking across the country. And I couldn't stop thinking about this idea of biking across the country. And when the time came to commit to this integration versus “I really need to commit to my mental health,” I at that point chose my mental health fully knowing that I really never had a chance to reset myself. And I'm not sure if biking across the country helped with that, but certainly it changed the trajectory. And when I was asked to take a sabbatical, I said, “I don't know where I will be when I come back, in my mental state.” And so I didn't want to string along the company who had done so much for me. But I also didn't want to be tethered to it. And so when I came back from a bike ride, I was like, this is the first time I haven't worked since I was 17 years old. And, might as well. I mean, that was nearly 30 years of nonstop working. And so I was like, “Let me take some time off. Let me explore. Let me look at what's going on.” And I had an idea of taking a year off.

And about a month and a half into me returning back, FDA called. They had received my name from a CEO in Silicon Valley as they were looking for someone with my background that had executive skills [and a] technology background. And I was so curious about this. I'm saying, “You know, I have no medical background. I'm not a researcher. I'm not a regulatory affairs person. I'm not a lawyer. I am nothing like what you traditionally hire for as a regulatory agency.” And they said, “That's exactly what we're looking for. We have lawyers, we have medical professionals, we have researchers, we have regulatory affairs folks and policy folks. And we need someone who's going to think about how the industry thinks and helps us determine how do we engage with industry.” And I thought that was a very progressive leaning kind of approach.

And given what was going on with artificial intelligence and all these emerging technologies, I felt that it was a really great time to join and help make sense of all of this and come with a perspective that combines both my technical background and my business acumen. This is what's happening in the industry and this is what we have to do to either enable that industry with safeguards to ensure that really safe and effective medical products are getting into the hands of the medical professionals who have to use them and the patients who have to effectively benefit from them. And so I thought it was the right time. I was the right person for it by combining all of that and then the right problem statements. And I think that's sometimes how life works.

Marcus: Wow. What an amazing story to prioritize mental health. And that is really important. And then it is true also, you say trauma drives creativity. But also, somehow opening yourself up to possibilities often brings possibilities that you might not have thought of. Right?

Tazbaz: Yeah. Manifestation is not this miracle thing. Manifestation is awareness. It's just that you're aware of something all of a sudden and you make it happen because it's now in front of you. Like when I decided that I was going to move to D.C. for work, all of a sudden I started seeing all these D.C. license plates or Virginia license plates in San Francisco, right? They probably were there before, but all of a sudden there was an awareness to it. So I was almost manifesting.

Marcus: Let's turn to your operations and ultimately your role at the FDA, because I really want to dig in there and what you're doing in general. How is digital transformation unfolding in healthcare?

Tazbaz: I think the basic example that I can give is customer relations or patient relations, right? In the old days you had to call an office, or you had to call a hospital, you had to call a customer service representative. You had to speak to someone. And what digital transformation [has done] is that it automated that process. And so, now you can resolve your inquiry without talking to someone, right? It's either automated or through a chatbot. And so it's effectively digitizing a business process.

So how I always define digital health is, it's integrating technology into the various points of a value chain of delivering healthcare. If you think about delivering healthcare, it could be simply a wellness app that someone engages with to monitor their wellness. Or it could be your smartwatch that is monitoring your heart rate. That's digital health technology, right? Or if you start getting to more acute settings, it could be an artificial intelligence application that looks at someone's biomarkers and compares it against, let's say, a population health and determines progression. But now all of a sudden you're not using simply a medical professional’s knowledge base to say whether someone at any snapshot in time, what the progression of their illness is. But you're also using technology to determine whether it's within a range against the population health or not. So now all of a sudden you're augmenting what a medical professional does. And so it's effectively using technology at every point in delivering healthcare. Whether it's someone who's already healthy or someone who is going through medical care.

Marcus: Got it. So the digital transformation process is actually the same thing as digital health. But digital health is the process of automating certain elements of the health diagnosis or throughout the process of delivering care.

Tazbaz: I would look at it as digital health is the vehicle that allows a digital transformation to happen.

Marcus: Can you describe your role at the FDA around digital health?

Tazbaz: Yeah. So I am the Director for Digital Health Center of Excellence, and this was an office that was created, I believe, in 2020. I'm the second director for it. And it's comprised of four divisions.

We have a policy division that, every time FDA is writing and publishing a guidance around specific laws and what the medical industry has to use to develop products with guardrails and the process of coming to FDA for regulatory approval. That's what that team works on. It's a lot of the guidance and policies that we have developed as an agency.

And then you have a technology team that is working on a variety of evaluation frameworks. So like, how do we actually evaluate these? They work on understanding the industry itself and the technological makeup and the designs of the products. And then they work with a variety of other centers as well as offices within FDA to ensure that we are developing evaluation frameworks proper to the technology that is coming down.

And then we have a strategic partnership team that works with a variety of both strategic initiatives internally, but also developing public-private partnerships. As a lot of the work that we do requires us working with some sort of a public-private partnership, and that has multiple flavors to it. So it doesn't have to be just a single entity. It could be collaborative communities. It could be standards organizations. So we contribute to that.

And then finally, we have a clinical team. All of this technology is great, but [not] if you don't know how to integrate that into what you do from the process of delivering care. That's workflow innovation you want. So we have a clinical team that is actually looking at “How do we actually integrate these things into what we do as physicians?” And so that's the four divisions of digital health.

Marcus: One of your priorities is fostering collaboration. And that is a key thing for us here around convening different parts of the network and different parts of the healthcare industry to work together. What do you mean by that and the FDA? And can you give an example of how that works?

Tazbaz: Yeah. So, there are different mechanisms that we use to establish partnerships. We have public-private partnerships working with, let's say, industry organizations and a few of them that are on our website is something like the Medical Device Consortium, MDIC. We work with certain artificial intelligence consortiums. And what we do is we align on a set of objectives that we need as a regulatory agency and that gives us a vehicle into working directly with the industry around these shared objectives. An example would be, recently we published an artificial intelligence lifecycle management framework. But the reality is we need the industry to own that. We're not a standards organization. And so what we wanted to do was have a starting point and from an incubation perspective, but we need the industry to take in and own that and collectively agree on a consensus-driven framework that any developer can use when they're developing AI-enabled medical products. So that's one avenue.

Then we have collaborative communities. Think of collaborative communities on very specific themes that people want to collaborate on. And FDA at that point becomes an observer to that collaboration because it can help inform us of how the industry is really thinking about a certain issue. And that collaborative community can be something like a patient advocacy and engagement consortium. It could be—so like, the patients talk about what they would like to see. Our patient advocates talk about how they would like to see technology being used and how they're engaging with that technology. And their perspectives are being taken into consideration, and both from a technology development but also from a regulatory perspective. It could be a diabetes community, right?

And then the last one is what we call CERSI, which is for regulatory science development. It stands for the Center for Excellence of Regulatory Science. And we work with academic medical centers around the country coming up with tools that you could use to validate products as they're going through a regulatory submission. And so we want to make these tools available for the developer community to say, “Here's what you can use to validate your product as you're going through the regulatory process.”

Marcus So let's talk about ROI and how this fits. I've heard you talk about this before, so I'd love to hear it for our audience, which is what is different about addressing AI-empowered devices—AI-empowered any digital health mechanism—than non AI?

Tazbaz: I think in order to answer that, first you have to be very specific when we're talking about AI. I think that, as a broad technological kind of category, we say artificial intelligence, but the issues that you have to address are both the value proposition and the risks associated. You have to be very specific between what we call predictive AI versus generative AI. When we're talking about AI and the risks associated to that, and what we're also seeing with media coverage, has been mostly focused on generative AI. And to give perspective, FDA has approved close to a thousand artificial intelligence-enabled devices. The first one was back in 1995, if I recall.

Marcus: Not generative.

Tazbaz: Not generative, no. And so, we've been playing in this space for quite some time. Now the difference is it’s actually very specific. With predictive AI or any medical product, there's always a human intervention. And once you develop and deploy it, you want to make changes to it unless it's learning a little bit from new data that is coming in. But nonetheless, if you want to add anything and you want to change its logic that it's using to reason with the data, you have to have human intervention. With adaptive technologies or generative AI, it's adapting with either new information or a new behavior that you're using to interact with the application. And that's been a risk. So all of a sudden we, as not just a regulatory agency, have to think very, very differently. And this concept of making sure that medical products are not just brought onto the market with safety and effectiveness, but they remain safe and effective as they're deployed into the real world.

You might have heard of this concept called data drift, and drift is effectively, as new data comes in, the behavior might change and it might actually provide responses that may not be as accurate as it was. And so this is where the challenge with generative AI comes in. Where, if you asked the same question to a specific generative AI application, you might get different answers, but the answers might not all be wrong. It might just give you a different wording of the answer and slightly different sentence structure. The question now becomes, how do you know when it's going wrong, when it's starting to provide inaccurate information? It's a little bit easier to do that with predictive AI or any other kind of software as a medical device technology. Right. Because it's not changing. And so you can do a lot more analysis and monitoring that is a little more straightforward. With generative AI, it becomes much more challenging. Now you have to take something that is subjective in many ways, and you have to apply an objective monitoring question around, “Is it still accurate?”

Marcus: So what are you most excited about in health IT right now?

Tazbaz: That's a great question. I believe that the transformation opportunities in AI is not just the capability of artificial intelligence. It's going to be the forcing function of—in order to take advantage of its capability, we have to fix a lot of the things that had been plaguing our system. So I believe it's going to be that vehicle.

You mentioned interoperability. Now, we've been talking about interoperability for a very long time. There are some business considerations why we have this problem, why we are where we are, right? It takes a lot of capital to fix this problem. It takes companies to align on interoperability standards. It requires the companies working with the various device manufacturers and EHR developers to align on a standard. That is all resource-intensive problems that you have. Not to mention, the resource could also be financial. AI, no matter what we say, at the very basics of it, requires data. Which means that you really have to fix the data layer to truly take advantage of its capability. And so what I hope and I'm excited for is that I think it's going to be a forcing function to address some of the critical issues that we've been talking [about] as a healthcare system for a very long time.

Marcus: These are the last questions. Healthcare is hurting—what's your diagnosis for what’s wrong with it? And then how would you help healthcare heal itself? What's your prescription?

Tazbaz: Have you ever heard of the Quintuple Aim?

Marcus: Yes, of course.

Tazbaz: When I was working on the healthcare business unit at Oracle, what we did was we sat down and we said, OK, what are the healthcare Quintuple Aims? And these are effectively strategic priorities. What they were was better patient outcome, lower cost. Then they added medical professional satisfaction. Then they added health equity, right? And so on and so forth.

And so I always ask the question, “Are you addressing any of one of those goals?” And here's where I actually do believe we have to change the narrative. There's been this narrative. They're saying that AI is coming for our jobs. And anyone who has studied healthcare-specific jobs will know that we have a significant labor shortage. And a significant medical professional burnout.

Marcus: Absolutely.

Tazbaz: We have to change the narrative. To me, this is an existential thing. No matter what technologies you have, if you don't have medical professionals applying that technology, it's kind of useless. And so we have to address that. What do we do? Well, you have to fix the burnout issue by making healthcare fun again for them, by reducing some of that administrative work that they have to effectively deal with. The labor issue could maybe be addressed because more people will be attracted to going into this field.

Now, here are some statistics that I've read and you can quote me on this one. I think I read it in American Medical School Association. Something like 130,000 fewer radiologists are expected in the next ten years. I read something like 8,000 fewer pathologists than is what's needed. And so we need to stop talking about replacing people that we don't have to replace to begin with, but more around how do we augment? How do we ensure that technology can help alleviate the gap while still providing the same quality of care? That to me is where you have to start. Without that, all of the upstream effects just will not happen. I think that we have to always address why did someone go into the profession to begin with? It was for patient care. It was for them wanting to make someone's life better.

Marcus: Well, thank you so much for a wonderful discussion and for all you're doing at the FDA to help healthcare advance and use some amazing technologies in doing so. We really appreciate that work. Thanks so much.

Tazbaz: Thank you so much for having me.

Marcus: How can we help healthcare heal itself? That's the question we've been asking throughout the podcast this season. As for our guest, Troy Tazbaz, his prescription is to let medical professionals be medical professionals, to let nurses be nurses and doctors be doctors. To let them focus on their patients and ease burnout by reducing the administrative burden.

But getting there won't be an overnight revolution. As Troy predicts, it will be a measured, step-by-step evolution where one change leads to the next. We'll get there by the industry coming together to align on data, for example, and by using technology like artificial intelligence as a “forcing function to fix the data layer of healthcare.” Then, before we know it, these changes will add up. They'll add up to a digital transformation. One where technology is fully integrated into the healthcare value chain. This will alleviate care gaps caused by burnout and provider shortages, all with the same high-quality standard for patient care.

In short, Troy's prescription is to make healthcare less about drudgery and more about fun. Or at least the kind of fun that comes from making a real difference in a patient's life. After all, as he puts it, why does someone go into the healthcare profession in the first place? They go into it for patient care.

On that note, Troy, thank you for helping make a wonderful end to our podcast this season, and I wish you all the success in your work to modernize U.S. healthcare.