Chris Boone of Pfizer on being a data hippie

EPISODE SUMMARY

This week Harry talks with Chris Boone, a leader of Pfizer’s effort to use new types of real-world data on patients—from insurance claims to lab tests to molecular profiles to data from wearable health sensor—to speed up drug discovery, development, and testing.

EPISODE NOTES

Transcript

Harry Glorikian: Welcome to this special series of Moneyball medicine focused on AI machine learning and analytics applied to drug discovery and development. This special series was recorded as part of the AI applications summit. Produced by Corey lane partners. I’m your host Harry in this series. I will interview different speakers from the event and we will hear their experiences.

We will dive into the challenges and opportunities they’re facing and their predictions for the years to come welcome to Moneyball medicine.

My next speaker has been described in many ways, a health futurist, a social entrepreneur. Executive professor patient advocate and self-proclaimed data hippie. Christopher Boone. PhD has a career long history as a dynamic innovative thought leader and a public voice on the power of health, informatics, and big data analytics and its ability to radically transform the global healthcare system into a learning health system.

Boone currently serves as the vice president and lead for global medical epidemiology and big data analysis at Pfizer. An adjunct assistant professor of health administration at the New York. University’s Robert F wagger graduate school of public service and active board member of several influential organizations and a co-founder of a few startup companies.

More recently, he served at the vice president and global head of real world data and analytics at Pfizer. Dr. Boone has been recognized as a 2019 top 100 innovator in data and analytics, a 2018 emerging pharma leader by pharmaceutical executive and a 2017, top 40 under 40 leader in minority health by the national minority quality forum.

Dr. Boon holds. Or has held appointments to some of the most influential national committees focused on health data. Patient-centricity including the board of director for the stores of change Institute, the executive board of director for the patient advocate foundation, the executive board of directors for the national patient advocate foundation, the board of directors for share for cures.

The Robert Wood Johnson foundation’s data access across sectors for health care initiative, the interoperability committee for the national quality forum, the national committee on vital and health statistics working group on HHS data access and use the health it policy committee, federal advisory committee, and the advisory group for the American society of clinical oncologists cancer link initiative.

Dr. Boon is earned his BS from the university of Tulsa and Ms. From the university of Texas at Arlington, a PhD from the university of Texas at Dallas and two executive certificates from the Harvard Kennedy school. He is a fellow of the American college of health executives and a fellow of the healthcare information management and system society.

Chris, that was. A hell of a mouthful. Ha ha ha. I don’t know. How do you keep up? I’m I’m surprised I like set all that stuff straight and didn’t stumble upon myself, um, that, you know, I I’m really happy to have you on the show. I mean, I really do want to delve into sort of your experiences, I guess, along the way.

And now you’re at Pfizer. Can you give us some examples of how you’re utilizing data to accomplish the goals you’ve set out for, you know, even yourself or your organization?

Chris Boone: Oh, wow. That’s a, that’s a very, very, very, very good question. But I will stop the part with a faster response with addressing your part of how I ended up at Pfizer, because I think it’s a.

It just adds to the story and I think gives people a better sense of my perspective on the world. Um, you know, and as you highlighted in the, um, uh, in my bio sketch, You know, if people look at my career trajectory, I mean, I started off very much in the health system in the hospital. And, um, and the reason that I did like my masters in healthcare administration, or even, uh, or even became a fellow of the American college of healthcare executives, because I, for all, I knew I was going to be a CEO of the health system and I was pretty adamant about that.

But I think there was a certain point in my career when I really embraced my true self. Um, I have spent, you know, I started in informatics, um, before we called it technically informatics, it was just it. Uh, and we were doing many of the, the bills and implementations of many other clinical systems. And so I was in the world of data before we really focused on the data.

We were more focused on the systems and the technology pieces of it. And, um, and there was a point when, um, when I really started to really focus in, on the data and the actual secondary uses of it. And I think it was really around the time when we were still using the term business intelligence. If you remember that.

And then we kinda, we, we shifted the term to be clinical intelligence because we were focused on that clinical information systems. And so, um, and that really was honestly my, um, my foray into, uh, into this world, a real little data and analytics or real evidence or big data analytics. So what term, whatever term you want to use, um, to really, uh, describe the work that we’re doing right now.

And so I come into the world of pharma, not as a conventional or traditional uh, pharma guy. Um, but as someone who cut his teeth in the provider world, who spent a significant part of his career in nonprofits who spent, um, a good part of his career in consulting, you know, helping organizations do, um, you know, exactly a lot of the work that we’re doing now.

Um, I was fortunate that I literally spent my entire career in the world of informatics and big data. And it wasn’t, I wouldn’t say it was necessarily strategically planned. You know, some of it was, um, volunteers, the purposes was serendipitous and I ended up in it’s just something that came natural to me.

Uh, there was always a intellectual curiosity on how we can do things. Better and how we can ultimately, uh, disrupt the way that we currently treat patients and ultimately transform the system for the, for the betterment of patients. And, um, you know, I actually came into pharma, uh, with the crazy notion that I could actually disrupt, um, and influence it at least a little bit, as far as the, um, incorporation of big data analytics and how we do much of our clinical research, our clinical development, and ultimately the commercialization of, um, of, uh, of therapies for patients. And so that’s, that’s always been my aspiration, you know, it’s, it’s, it’s all centered around transformation and disruption in healthcare, but from many different perspectives.

Harry Glorikian: Well, first thing is any, a head hunter out there listening to this. Remember, he’s looking to be the head of a healthcare system at some point. So keep that in mind. Um,

Chris Boone: I mean, it’s like, that’s always been a, I mean, a goal, but you know, I mean, right now I’m, I’m, I’m really, um, really happy with the work that we do.

I think, you know, I would say at this point, um, I, um, haven’t really, really embraced this idea of what the possibilities could be the art of the possible when it comes to big data and analytics. And I don’t mean that in a, um, uh, you know, in a fortuitous or some kind of, uh, uh, you know, uh, uh, theoretical manner.

I really mean it in the sense of what I, I just, I like to use the term innovative pragmatically or innovate practically. And, you know, I hear a lot of folks that are talking about these possibilities. They get enamored with use, um, very grandiose ideas and concepts that could never really effectively be implemented in an organization.

And I pride myself on developing very, um, uh, you know, uh, very, very, uh, Uh, very good. I would like to say aspirational visions, um, but also very detailed and elaborate strategic plans to achieve it. And I can also leave these, the part of operationalizing. So, um, yeah, I mean, I think of my career at some point, if I ended up back in the hospital health system, the hospital is great.

If I don’t, that’s fine too. Um, I just love leave it leading a transformation and disruption, and then, and doing it all for the patient. So that’s good for me.

Harry Glorikian: Well, there’s a lot of transformation and disruption that’s for sure. But you know, how do you feel like succeeding along this, this journey that you’re on this goal of, of making these changes?

I mean, you know, you’ve got real world evidence experience on the regulatory and clinical decision-making side. How, how are you guys implementing or approaching analytical methods, you know, from the data side, from are using, you know, new techniques from either artificial intelligence or machine learning and what you’re doing and, and how does that generate business value and then impact to the process and the patient.

Chris Boone: Oh, man, that is a great question. And I think it’s a great question there. Um, many organizations have struggled with, you know, many of the big pharma’s or even smaller pharma’s have, uh, invested heavily in the big data analytics capabilities. Are these real world data real evidence capabilities. I mean, you know, you’re, you’re essentially hiring data scientists, um, who don’t come cheap.

Um, you’re investing, uh, know significant amounts of money and the techno technologies to support it. Um, and also it, it’s no cheap adventure and investment. And I think the question of ROI is always one that comes up. Um, you know, as far as far as, how do you essentially, uh, create business impacted, generate economic or business value for the organization and how do you measure that?

I don’t, I think that, uh, many have been struggling with that, but I think for, um, from my vantage point for me, I actually, when I came to Pfizer, initially I was. Um, my charge was to really establish a real data analytics center of excellence for the organization. And really, I think, you know, when I came into the organization, I had my own view on what, what that really meant.

Um, but then after assessing and evaluating the organization and figuring out where they were on the, on the maturation continuum, it dawned on me that really where we were with our, our real agreement at that time was being a catalyst for change. Sure we had many of the, uh, very traditional, uh, responsibilities of what you would see in a kind of an office of a chief aid officer, right?

I mean, where we have, you know, data strategy, data management, data governance, uh, types of responsibilities. Um, but then we have this added layer where we, um, we’re really the experts and this whole notion of real whole data, quote unquote for the enterprise and, you know, our kind of, uh, misfortune and some regards was that we were ahead of the regulatory impetus for the use of real world data and real, real evidence slightly.

Um, at the time, you know, you had, um, you know, uh, Scott Gottlieb had uh, had, was just, he had just taken over and taken off FDA conditioner. He was, um, he was, it was a part of, a lot of his, uh, his rhetoric, his public speeches around the significance of railroad evidence when he pushed it. And that became, uh, it started with, uh, Rob Taylor, who was, he was the proceeding FDA commissioner.

And so you saw a lot of these things. And then of course we had 21st century cures, um, which was another, uh, obviously. Uh, regulatory or regulation that pushed it here in the us. Uh, there’s obviously been things that have happened in the NMA. So, you know, I think that when it came to the use of big data analytics and the use of real world evidence, as it currently stands, there were multiple, uh, drivers for what we’re saying.

Right. I mean, we had, um, you know, when you think about it, just from what we just spoke about, we talked about regular regulatory incentives and we talked about technology and you mentioned AI and machine learning and these other things. But there’s also been these market drivers that have, uh, served as somewhat incentives to, to, to, to invoke the change.

And I think a lot of that was the rising costs for clinical development. You know what we’re seeing in clinical trials? Um, obviously there’s a significant amounts of voluminous amounts of health data that are being generated right now. There is, uh, um, this push and this emphasis on, uh, creating much more of a value based system.

And then of course, you’ve just got this natural pace of scientific innovation, which we see as a current, quite a bit with the biotechs. So the real question is that now that you’re investing in all of these big data and analytic capabilities, and you’re investing in the people and with this expertise in AI and machine learning and so on and so forth.

And how do you really translate that into business value to your point? Right. Well, um, for me, when I’m leading right now as a group, that’s focused on, um, the name of the group is global medical epidemiology and the data analysis it sits and the office of the chief medical officer at Pfizer. I’m the global chief medical officer at Pfizer.

And really what we’re really aiming to do is accelerate. Um, you know, essentially establish a faster data path to show like effectiveness in Pfizer therapies for the regulators. So we’re almost trying to create alternative, uh, ways to generate evidence that are acceptable to regulators that demonstrate the clinical and medical effectiveness.

Uh, our therapies. Right. And, um,

Harry Glorikian: So if I could just jump in there. So, so how, what are the data sources that you guys are using? In other words, I, I automatically different sensors, different devices, different applications, uh, you know, monitoring systems. These are the things that sort of like flashing in my head.

Um, How are you guys accomplishing the goal that you’ve set out?

Chris Boone: Well, I will say that if I think that what we’re doing now is, is, uh, we’ll say from, for farmers pretty cutting edge, I think what we’ve used, um, and, you know, recently is obviously clinical data from EHR data or lab test results. And, you know, and as a diagnoses procedure, You know, that type of data, um, you know, a lot of things that basically your EHR, um, we also use claims data quite a bit.

Um, but I think really the direction that things are headed now is that you want to see more incorporation of your molecular profiling data. Um, do you want to see more data from mobile health or, um, uh, wearable devices, as you just mentioned? You also want to see, um, the incorporation of data from environmental factors, you know, that affect, um, you know, uh, you know, everybody is making this push towards the notion that your environment is a better indicator of your health than your genetic profiles. So we, we obviously want to be find better ways to analyze that data and what stories and what insights you can draw from that. Um, we, uh, we’ve leaned heavily in the past on patient reported data, obviously as part of randomized clinical trials, you’re doing a lot of that when you’re creating these um, uh, patient diaries of patient reported outcomes that you use for there. And then, uh, um, and then we, we, we, um, you know, and for my group specifically, uh, we rely heavily on the literature. You know, what? We are able to get a better understanding of what the disease burden is, the clinical characteristics, the prevalence and incidence rates, um, you know, kind of the standard of care for many of these, uh, clinical disease areas that we’re focused on. So you can see, um, there is a, is a wide array of, uh, of data types and data sources that we use. I think that those are, will only expand as the world gets smarter with how you can effectively analyze those data types. Um, but you know, I, I don’t, I don’t know if we’re quite there yet.

Harry Glorikian: So, you know, you’re talking about a lot of different sources types of data where it’s coming from. Some is streaming. Some in some is episodic. You know, what are some of the roadblocks or speed bumps that you’ve encountered along the way that you can, you know, share pearls of wisdom with, with people listening.

Chris Boone: Oh man. It’s um, that is there, you know, I, I like to describe it as there are, uh, obviously some regulatory issues. Um, there are some ethical barriers. Um, you know, there are some even financial barriers and ultimately cultural barriers to how we think about sharing data, linking data, and ultimately using data.

Um, you know, what, what we’re seeing right now, if you think about it, there’s, there’s a lot of discussion around. Um, how data should be shared and whether, uh, you know, you even heard, some people say that, uh, patients consumption, uh, should be, um, you know, they should be rewarded are, are, are essentially paid right for their, for their data.

Um, we are there’s discussions around that. There’s obviously tremendous costs with procuring data because you gotta remember for us, we don’t own, we don’t have an EHR. We’re not a provider. So we have to, uh, construct these um, very, uh, contractual relationships to actually have access to the data, to do the things that we need to do that are outside of an RCT.

Um, and then I think there’s, there’s like these, I don’t think that there’s a general consensus around the principles and practices concerning, um, uh, data access or, or data ownership and even, uh, the control of individual patient data. But a lot of that at the, at the root of it is still very much centered on this, this notion that there isn’t a trust across the various sectors from whether it be from providers to pharma, you know, pharmacy, to payers, the payers define, uh, payers to whomever. Right? I mean, so we need to see the level of trust, which I think is probably the single greatest barrier to progress in the space, um, across the entire industry. So,

Harry Glorikian: Yeah well, and you know, I mean, it’s always. Interesting. Right, because there’s a profit motive versus a. What, what providers may consider, you know, managing the patient. It is a very scary landscape, especially when you see what’s happening with Facebook and those, they don’t necessarily help the conversation when you’re talking about data sharing and different, um, dynamics.

But, so how do you think about utilizing technology and data to sort of improve things like clinical trial? I mean, We’ve, we’ve heard a lot about things like virtual trials or remote trials and then hybrid trials. And it sounds like you’re getting real-world data from different technologies or different data streams, you know, how are you thinking about incorporating that into. How it would speed up or streamline your process? Well,

Chris Boone: I think that if you start, even earlier than that, let’s just say before you even get to the point of a clinical trial, because I think the default response is that, okay, we have this new medicine, this new molecule let’s do a clinical trial. And what I, what I would like to see is the, the kind of hierarchy of evidence, if you will, um, incorporate or, or really truly embrace the other means that, which you can generate evidence that’s acceptable to a regulator. And so what I mean by that is that when you think about early clinical discovery, which is, we put up by replace critical role in here, what we’re trying to understand, like the disease epidemiology, right? Which includes like, What’s the current standard of care around this disease.

Um, what’s the disease burden, what’s the unmet need. And those are some fundamental questions that we’d like to answer very early on in the discovery phase. And then when you get to the clinical  trial stage, you want to ask yourself, okay. Does it make most sense to do a randomized clinical trial or does it make more sense to do what I would deem as a big data or real relevance trial?

Right. And in a real world evidence trial, that’s where you start to get into all the different models and types that you just mentioned. I think the only one that I didn’t hear you saying was like pragmatic clinical trial, uh, which incorporates all of those things. Right. It’s a different manifestation of it.

Um, but I do think what we’re, so where you see me thinking about things, is that okay if we do do a randomized clinical trial, the needs are different. This is how we can use wearable evidence to optimize that clinical trial by identifying. Um, where are these patients actually are? Right. Um, you know, where is a, where are the clinical trial and RCT most likely to be successful with the recruitment and retention of patients just based on where they reside.

It also would help with understanding what the feasibility of doing that clinical trial is life, because you’re not, obviously there are a significant number of clinical trials that are delayed, um, just because of, you know, whether it be patient recruitment issues or size selection, challenges, you know, you name it.

Um, but then if you go this other route that’s kind of the adaptive trials or real world trials angle that you were alluding to them, what exactly does that look like? And I think that, um, that’s the area that’s like. Um, that we’re beginning, we’re in the nascent stages of exploring. I mean, we’ve been doing randomized clinical trials for what five decades or so that’s been around for a while, but this whole area of real-world trials and big data trials presents a mess opportunity, but we’re going to have to put forth the investment and be a bit more exploratory to see how it works.

And I also really feel like that’s what, uh, the FDA and even the EMA. They’re encouraging us to do now the scary part about this is that there isn’t a lot of prescriptive guidance around how and what you should do. Um, and of course, no one wants to, uh, uh, you know, take that risk and have, you know, it’d be a dismal failure and without any complete, uh, without any guidance at all.

Right. So I think that where the regulators are saying, if they’re learning too, and they want us to try to admit it and see how they feel about it, and then they, as they learn, they’ll give. Um, more, um, specific or prescriptive guidance around how we should, how, how we should think about these things in the future.

But, um, it’s, uh, you know, there’s a lot of gray area in there right now. Um, and it does require a level of, um, exploration, um, to really figure it out. And that’s, that’s where we are right now.

Harry Glorikian: Oh yeah. I mean, it’s funny because I, you know, I feel for the regulators, I mean, with the with the rate of change on the technology side,

Chris Boone: Yeah. It’s like they can’t keep up, you know, that’s the thing, it makes it, it makes it impossible because the regulator thank you. But with the pace of, of scientific and technology innovation, I mean, it’s just impossible. Right. And so, you know, so for them, I don’t fault them for where they are right now. I really don’t.

I mean, it’s just a tough situation for us to be in because it’s such a regulated industry. Right. And so, uh, So you’re relying on them, but they can’t, you know, um, they’re in a tough situation as well.

Harry Glorikian: So yeah, no, I mean, it’s funny because I used to think about it. You know, a lot of it was driven by scientific change and now it’s not just scientific, but technological and data sciences, and it’s hard enough for us to hire the people we need to hire, let alone somebody at a regulatory agency hiring and which, which sort of brings me to the next question.

So. It sounds like, and, and not knowing the structure of your group, you have to have unique talent sets to pull off what you’re trying to pull off. And these people aren’t falling off trees. So how do you, how do you find them or how do you entice them to, you know, not go to Facebook? But to come to Pfizer.

Chris Boone: That’s interesting. Cause that’s, again, that’s a great question. And it’s been one of the more sobering realities of, um, of where we are when it talks about the, um, um, the availability of the abundance or lack thereof when it comes to data science, you know, and what I’ve come to find out. This is, this is my, you know, this is my personal perspective.

People can take from them what they want, but you know, I think at one point several years ago, Everyone was fine to have to be more like Facebook or these other companies in Silicon valley and hire all the data scientist that they could. What I quickly come to realize is that you can hire data data scientists, anyone, but they can have very limited to no value.

Right because the reality is the key is to find those data scientists who have domain expertise, where they can really interpret their all analysis to help inform decision making. That’s ultimately what we want to get to. Right. And so, um, and so really the key is to, um, I find is to really develop a hybrid approach where you’re using, you know, these very classical data scientists who would just.

Expert modelers, or they’re a machine learning scientist or whatever their background is. Right. And you partner them with folks who really understand the farm of business. Right. And what you’re hoping is that you start to get some level of, uh, organic growth. Um, and there’s a bit of cross-training where you’re the science of the data.

Sciences is learning the business. And the business wants to learn more about data science and, um, and you can really start to move the needle there where I think what I quickly realized is that you can’t find, there are no true unicorns in this situation. Every team that you have should be built with the notion that it’s going to be an interdisciplinary team where you’re going to have experts in it.

I think people have this, this kind of, uh, uh, misguided point of view that they think that you’re going to find the single individual. Who’s an expert in a domain expert with this very robust kind of a data science capability. And I’m not quite sure if there’s some cases. I mean, don’t get me wrong.

They’re out there. And some in rare cases, but they’re not as an abundant as we all think. And I don’t know how easy it is to find those folks, those data scientists off the street. Who can just come in and add instant value. Sure. They may be able to build the best model, but if no one can interpret it, then it doesn’t mean anything.

Harry Glorikian: Right.

So, you know, I, I’ve kind of evolved from that notion that you need all data scientists and things. And I really believe that. Um, you should take more of a hybrid approach of kind of growing many of these guys organically. And, um, and also, you know, obviously bringing some of that expertise in because you do need that deep expertise, but I don’t, I don’t think I need to compete with Amazon or Facebook for data scientists and another talent that you find too, is that, um, I frankly.

The data scientists won’t work. If they feel like it’s very challenging and rewarding to them. And, um, and you know, and this, this work is not for everybody. Right? And, uh, and so you have to find folks that really have a passion and give them, uh, give them the narrative, the story of how the work really matters.

Um, because what I find is that, believe it, or not many of these guys, they want to do rewarding work and whatever would however they define that for themselves is, uh, As an individual, um, uh, perspective, but you know, they really want to do rewarding work. And so you’ve got to find folks that are really passionate about what we’re trying to do.

Harry Glorikian: Yeah. I mean, this reminds me of like, when we were doing the genome at applied Biosystems, we’re like, okay, we need bioinformatics. What the hell is that? We’ll get the biology guy and get the Informatica and put them in a room and have them, you know, duke it out. And, you know, it’s only recently where we actually have, you know, you can actually get a degree in bioinformatics. Um, you know, the rest was on the job training and, I feel like with what we’re doing now, it’s, it’s almost a step wise challenge, especially with the way that technology is jumping forward from the tools that it’s providing our industry, you know, which one do you apply?

Which analytic technique, which, which training method, what chip set, what, you know, there’s just so much that comes into it. That teams keeping up with it is, is, is, is non-trivial.

Chris Boone: You know, that is, um, that’s such a great, great, great, great point. You know, I, um, so yeah, it’s your point. I actually teach and, um, I actually have a son that’s in college and, you know, and I, and I feel like I’m sharing the same advice with the students as well.

Um, you know, this whole idea and in my study in college, actually right now, he’s learning Python. Yeah. It’s part of its programming.

Harry Glorikian: Yep. Yep.

Chris Boone: Well, this would be great for the gray and I’m thinking like, okay. Yeah, that’s very current. Um, but you know, at the same time I’m thinking about what I actually tell my students, which is I  emphasize to them the needs to be more focused on the soft, you know?

Sure. I would like them to understand like the technical concepts. But man, to your point, it’s like the moment they mastered this technology or this approach, and it becomes obsolete, then what are you? What’s your alternative?

Harry Glorikian: Exactly.

Chris Boone: I’m like, uh, I just, I just think that, and then as we continuously talk about automation and what that means, especially when it comes to advanced analytics, I don’t know how.

I mean, I just, we don’t know what the jobs will look like 10 years from now. Right. And, um, and so it’s a tough situation, but I think that if, you know, one thing I know. Is that the soft skills and the need for that will never go away. Right. And, um, and so I just need, I just, I wish we can, we put so much emphasis on these, uh, you know, programming languages or these technical specific technical skillset and not enough on what it takes to work in a team and work, uh, in leadership and other things that I think were important.

Harry Glorikian: So do you see the technological implementations that you’re working as, as. Um, incremental, uh, you know, are they step wise? Are they exponential or is it, you know, in that still ramping phase where you’re still trying to figure out how to put all the pieces together?

Chris Boone: I think, you know, can I say all of the above?

I mean, Pfizer itself has been, you know, you know, several years ago they announced, we announced that it was before my time. So I’ll say that, but, um, but Pfizer announced the partnership with IBM and to do, you know, it was IBM Watson health to do the, uh, uh, the IO or immuno oncology, uh, project. There’s going to be a way to accelerate the discovery phase.

Uh, uh, you know, I’ve lifecycle right here now. So that, that was, uh, I mean, that was a full punch, right? And to what that really been. I think that everyone, all parties involved as it, as well as the industry. Um, we’re very excited around what the possibilities could be with AI. And, um, I don’t know if we, I don’t know if people, I don’t know if we were able to really articulate use cases specifically because the technology is just felt so new right to this whole, uh, clinical research clinical development space.

Now you fast forward to where we are now. And the whole industry has had some lessons learned from others and from maybe their own experiences. And now they’re approaching things a little differently. So I think, um, when you say step wise, so on one hand, sure. That was an all-in plunge. Right. You know, I don’t know if this is all part of the hype cycle or what, right.

And then there’s a point when you get to where you’re like, okay, I’ve tried that. Um, I may or may not have seen all that I wanted to see. Um, so let me be a bit more step wise. I would say strategic. Uh, an opportunistic about the use cases and, and a little more deliberate in how we’re going to execute on those.

So we were not going to boil the ocean, so to speak. Um, we’re going to identify specific use cases that we know we can really move the needle with this technology.

Harry Glorikian: So-

Chris Boone: And I think, uh, um, where we are now is in that phase of exploration where we’re saying, where can we invest that we know will have a significant impact?

And have value and, um, and that’s where we want to focus. Right. Um, it doesn’t have to be the entire pipeline or the entire life cycle. It can be specific stages where we think we would have the maximum benefit. And so, so I think that’s where, so it’s like, you know, that’s why I said all of the above, because on one hand it was all in.

Then on the other hand, you’re kind of thinking, okay, let’s be a bit more strategic and opportunistic. And then you also got this level in some areas where there’s still some level of skepticism around what can this technology really do. And so it’s like the proof of concept phase, you know? And so this depends on where you are in your organization.

Harry Glorikian: Yeah. So, uh, you know, without divulging confidential information, um, you know, do you have areas that you see. These capabilities sort of moving the needle.

Chris Boone: I didn’t want to show the more obvious things. Um, you know, when you think about my world specifically, which is kind of the robot data, real world evidence world, um, you know, there there’s been extensive, uh, conversations.

And dialogues around using NLP for Instructure or, or the we’ll say provider notes, so to speak, especially when you start to get into these areas of like different types of rare disease and cancer tumor types. Right. Um, and so, uh, so using NLP, for example, to really. Understanding examined, um, you know, voluminous amounts of, um, um, uh, provider notes or data, um, would be a huge, huge win.

Right. Whereas I think, you know, maybe the case five years ago would have been, well, I want to use. All of the different AI techniques to understand all data elements within the HR. Um, now it’s probably a bit more focused and the question is a bit, uh, it’s better formulated. And so thus you get a better use of the technology.

So I think that’s like one use case. I think there’s other opportunities to do. You know, as I mentioned earlier, one of the data sources we use is around. Um, the vast amounts of literature. Um, and I think that’s where, uh, the IBM Watson health, uh, technology really came into play, that it was very, it was able to adjust, um, you know, um, uh, you know, massive amounts of, of literature and data and analyze it in, uh, you know, much faster than a human could actually actually do it and, and produce, uh, some direction and things like that.

So I think. There are opportunities to get, even when it comes down to literature searches, which there’s so much literature out there that in and of itself could be a huge use case, especially from a group that’s focused on the understanding of disease. Which is what we do. And so, uh, so I think those are kind of like two, uh, very, um, targeted use cases when I think of it and how it could be used.

Harry Glorikian: Well, I, I, you know, I look forward to, uh, sitting down with you in person or meeting you and, you know, at the conference coming up in, in a short while here, hopefully we’ll get a chance to hear more about what, what you’re doing at Pfizer and yeah. Interact more. There’s this, this is a small world and it’s always good to be in touch with people who are at the forefront of, of what’s happening in the field.

Absolutely. Thank you very much. Thank you. And that’s it for this special series of AI machine learning and analytics. If you enjoyed Moneyball medicine, please head over to iTunes, to subscribe, rate, and leave a review. It is greatly appreciated. Hope you join us next time until then farewell.