AI In Clinical Trials: Hope Or Hype? w/ Tom Doyle, Chief Technology Officer, Medidata

AI promises to revolutionize clinical trials and reshape regulatory oversight. But is the pharma industry ready? And can the FDA keep pace with the technology?

In this episode of the HealthBiz Podcast, host David Williams is joined by Tom Doyle, Chief Technology Officer of Medidata, to discuss the promises and potential pitfalls of AI in clinical trials.


🎙️⚕️ABOUT TOM DOYLE
Tom Doyle heads product development at Medidata, a leader in clinical research technology and a life sciences arm of Dassault Systèmes. In this role, Tom leads Medidata’s work in developing industry-leading solutions for patients, sites, sponsors, and CROs, improving and accelerating the design, execution, and oversight of clinical trials – essential in bringing new and novel treatments for patients.

Tom joined Medidata in 2019, bringing 20 years of global experience in medtech and data science. Before coming to Medidata, Tom held leadership roles at Janssen and at Boehringer Ingelheim, championing technology to drive new experiences and better insights. He’s passionate about the power of innovation to transform clinical research towards better outcomes and better experiences for patients.

Tom has a BMath in computer science from the University of Waterloo in Ontario, Canada.


🎙️⚕️ABOUT HEALTH BIZ PODCAST
HealthBiz is a CareTalk podcast that delivers in-depth interviews on healthcare business, technology, and policy with entrepreneurs and CEOs. Host David E. Williams — president of the healthcare strategy consulting boutique Health Business Group — is also a board member, investor in private healthcare companies, and author of the Health Business Blog. Known for his strategic insights and sharp humor, David offers a refreshing break from the usual healthcare industry BS.

GET IN TOUCH
Follow CareTalk on LinkedIn
Become a CareTalk sponsor
Guest appearance requests
Visit us on the web
Subscribe to the CareTalk Newsletter

#healthcare #healthcarepodcast #healthcarebusiness #healthcarepolicy #healthinsurance #healthcaretechnology #caretalk #fda #ai #clinicaltrials #clinicalresearch

Transcript

David: 0:00

Artificial intelligence promises to revolutionize clinical trials and reshape regulatory oversight. But is the pharma industry ready and can FDA keep pace with the technology? Hi everyone. I'm David Williams, president of Strategy consulting firm, health Business Group, and host of the Health Biz Podcast, where I interview top healthcare leaders about their lives and careers. My guest today is Tom Doyle, he's Chief Technology Officer at Medidata, a leader in digital transformation for clinical research. We're gonna discuss clinical trials today, the evolving landscape of clinical trials and the ground. Breaking role of ai. Do you like this show? I hope so, and I hope you'll subscribe and leave a review. Tom, welcome to the Health Biz Podcast. Thanks so much. Great to be here. Hey, I've heard of a lot of different days. I'd never heard a clinical trials day. What is that and why is it significant?

Tom: 0:56

Clinical trials day is an opportunity for us to reflect on the history of clinical research. Uh, we remember back to the very first sort of gold standard clinical trial, um, where we demonstrated, or where it was demonstrated that you could have a positive effect on scurvy. Um, it was a very simple design trial, but nevertheless it sparked an entire. Multi-decade, really multi century approach and, and research that has yielded some of the most important medicines that, um, people take today. Uh, it's a chance for us to think about how we continue to evolve and challenge ourselves in this very fast paced and, um, innovative and demanding environment, but also how we remember all the people who participate in clinical trials, all of the study teams that work so hard. To ensure the safety of patients while accelerating new novel treatments for patients, and also the patients themselves who give a lot to be a part of clinical trials, not only for the hope that it might bring them in their care journey that they're currently on, but the hope for many others who are suffering from, from similar, uh, sim, similar ailments.

David: 1:55

Now, as I recall, I think, uh, the insight about, uh, what she could do to stop scurvy was a strategic advantage for the British Navy. Is that right?

Tom: 2:03

It was, it was, uh, the impact that citrus fruit actually had on, on scurvy.

David: 2:09

Nice. All right. So, um, you know, during the, the COVID to 19 pandemic, there was a lot of attention given to clinical trials and, and trials also, uh, had a lot of responsibility for being able to control, uh, the pandemic ultimately. But a lot of lasting changes came out of it, uh, from remote trials and, you know, other elements of that. What sort of changes have you seen? And if we look pre pandemic to post pandemic in the clinical trial landscape.

Tom: 2:33

You know, personally, I think it's probably the first time that my family actually understood what I did for a living. That was one positive output. Um, but there has been some very lasting changes. I think one is just, uh, mentally how, uh, how clinical researchers, how both small biotechs, um, as well as. Large pharma have approached clinical trials and have approached research is with a renewed rigor. How can we go faster? You know, there is a common saying about patients are waiting and that's, that was true during the pandemic, of course, when the whole world was holding its breath, waiting for, you know, when we could come out of our houses and things like that and return to more of a sense of normalcy. But also that continues, uh, every day in lots of different diseases. Uh, oncology patients who are looking for. For new treatments and for new opportunities, new new options in their care, uh, cardiovascular patients, immunology, and on and on. Um, and really as an industry, we challenge ourselves to think, uh, really creatively how we can be faster, how we can be smarter about where we're making investments that are really gonna yield the best results for patients.

David: 3:38

When I think about, uh, moving faster in any process, but especially a really complex one, like clinical trials, often there's a focus on, on bottlenecks because that's the sort of the slowest point that's gonna hold back the overall progress. Do you think about clinical trial timelines in that way? And if so, what are some of the key bottlenecks that you see?

Tom: 3:58

I think you're, you're, you're right. And with any complex process or complex situation, it's, it's rarely the case that there's, you know, one silver bullet or one answer to all of the problems. So we look at ways, um, in partnering with, you know, with our customers and partners in the industry. I. To knock down all of the barriers that combined health for an overall better patient experience and accelerating trial timelines. Um, some of those are around patient recruitments, you know, how do we identify the right patients and reach out to more patients that will benefit from clinical trials, but who are also very reflective of the populations ultimately. That new therapies will serve, and how can we reach them in new and novel ways to engage people who otherwise might not have been involved in clinical research historically. Sometimes that means developing digital technologies that allow us, you know, to work outside of major academic centers, which have been a lot of the, the lifeblood of clinical research for so long. Um, but also it means looking into more data and AI driven approaches that help us identify who are good candidates for clinical research. The other is during the execution of the trial itself, we look, we continue to look at ways to accelerate the overall process, whether that's, you know, automated anomaly detection, for example, and getting to cleaner, better data data faster. Um, whether that's using insights and AI that help us design better trials, uh, that are more likely to. To, um, to deliver a, a good patient experience, but also a, a, a good outcome that are gonna demonstrate the safety and efficacy of a new therapy. You know, there's a lot of different at bats that we're taking essentially to, to try and shorten the overall timeframe, but also increase the, the, the positive results that we get from, from new therapies.

David: 5:40

Tom, I don't want you to take this the the wrong way, but sometimes we have people here on the podcast that, uh, go on to do and, and say crazy things after they, uh, after they finish. So one of our guests was, uh, the current FDA commissioner and what I'm referring to there is they recently announced a rapid rollout of AI and I mean rapid. Uh, so that to think it was just last month, which was may. It said that, you know, based on the results of one, uh, of their first pilot, they're going to roll out, uh, AI for scientific review everywhere and have it all completely done by the end of June. Um. What, what, I won't ask you to make a political comment, but, uh, what, what is the impact of what FDA is doing in terms of, uh, review using ai? What impact does that have potentially on trial timelines and, and regulatory interactions? I

Tom: 6:31

mean, I think it's an example of how we're truly in an age of disruption, where we are looking at AI as being a transformative technology that should challenge us to think differently about how we've historically, how we've historically done things. Um, certainly in order to take advantage of these new technologies, this new opportunity that. AI and, uh, more advanced analytics brings, it brings a renewed focus on data, um, how we are collecting and organizing that data, but also how we're being transparent in, in the usage of that data. You know, engaging patients, for example, more in the discussion around how the data that they provide in, in, in their. Context of a clinical trial can be used for sure for that trial, but also the pot potential, potential benefits that it can bring long after. You know, having those conversations upfront and early, I think will only open up new opportunities that will accelerate research. Um, similarly, we have to think differently about our infrastructure and the investments that we make. Uh, I, I think it's, uh, it shouldn't be a surprise that, you know, big sponsors, large pharma as well as biotechs. Are themselves very aggressively looking at the promise of ai. Uh, so that it's not only regulators that, um, are then applying it at the end, but we're thinking about what is that practice all the way through. Um, it should open up new opportunities for better collaboration, you know, more interim reads, better, um, working between industry and regulators so that we can continue to get what we jointly all want, which is better outcomes for patients. Um, and we all want those outcomes faster.

David: 7:59

So it's hard to say, and I won't even try to say if, you know, FDA made an announcement last month in this direction, they're making other sorts of announcements too. So there's a lot going on. So, we'll, maybe we will have to follow up in a year or so and see where it's going. But what do you see, um, kind of right now for some of these immediate impacts of, of AI and trials? Not necessarily driven by, uh, by FDA, but you talked in any case, we know that there's always a, a, a, a use of. Analytics and other techniques. It's not that everything was just being done with pen and paper and uh, you know, rules of thumb up till now, but where are the specific aspects where AI is being used for clinical trials today?

Tom: 8:39

We're definitely a long way away from like multi-part forms and shipping around paper all over the planet. You know, thank goodness. Um, really you can think of a lot of the AI innovations in three big buckets. You know, the first, not surprisingly, are all around operations. How do we go faster, be better, higher level of quality, identify potential challenges before they occur? A lot of that, for example, is in data review using AI to surface anomalies or potential challenges in data that start to allow you to ask the right questions. Um, perhaps then working with your sites to make sure that you've got the right answers. You know, is that data telling you a story that is somewhat surprising or, or perhaps, um, uh, there's a, um, uh, an anomaly that that looks different than maybe the rest of the study that you should understand more deeply. Uh, similarly on the operation side around. Better operations decision making, you know, where should you run your trial? Um, clinical trials, uh, like healthcare broadly are, it's a very global, uh, global enterprise. Um, and a lot of decisions need to be made around, you know, what countries, what population should be included, what is that inclusion, exclusion criteria. Uh, and AI is certainly informing a lot of that. Um, this second is really around the design itself, and I think this is the really exciting piece of, of AI where we start to surface a lot of. Clinical insights that help us make good decisions, that help in the safety of the trial for patients. You know, identifying potential risks, for example, of adverse events so that we can plan for them before they occur. That generates a better overall patient experience. As an example, uh, identifying better endpoints. Um, these are all, uh, optimizations that we can make on the trial design, and we can do simulations before the trial ever starts so that we're setting ourselves up for, you know, a better overall trial experience. The third, which is really emerging now, which I think is really exciting across not only in in clinical research and across life sciences, but in all industry is this more human experience that AI is allowed to provide. Know for a long time since the invention of technology, it's been a lot about how do we help. How do we help people conform to how technology works? We trained everyone how to use, uh, web browsers and, uh, now mobile devices and the, and the like. But what we're seeing really in the, uh, the continued evolution and very fast-paced evolution of technology, including ai. Is that now the, the machine can start to create, it meets the, the user more where they are. So more conversational, for example. It's understanding or at least trying to understand more of intent. Um, and this creates entirely new ways that our users will interact with our platform. We're not just making the old thing faster, um, and maybe better and more precise, but now we're creating new ways to, for patients and for. Um, for sites for, for all of those who participate in clinical trials to interact differently with technology, and I think that's probably one of the most exciting things we'll see develop over the coming years.

David: 11:33

One of the areas that, uh, receives a lot of focus in terms of being a bottleneck as patient recruitment, but also recruitment and retention of the patient in the trial. And I'm, I'm not sure if you're how much you're involved in the recruitment side, but certainly would the, the, what you just described is gonna impact retention. Does AI have. A role there? What's, what's the role there for AI and retention? Certainly it

Tom: 11:55

is, it is a, a challenge, both recruitment and retention continue to be, you know, one of the major contributors of, of trial delays. Um, and there can be a lot of factors that go into that. Um, certainly some parts of research have become quite crowded spaces where there's a lot of co competition for, um, a relatively few number of, of patients. Who are aware and who are open to clinical trials. Again, we can increase that pool by building trust and awareness in clinical research and grow the number of people who are willing to participate. But still in some diseases, it's, that's gonna be quite challenging. Um, the fact is also that that trials can be. Quite difficult. Um, without a support system, without the, uh, flexibility to produ to participate in research, it can be hard for patients to continue that journey through, um, through a clinical trial. And their AI is really helping in identifying, uh, burden points in a protocol, for example. So things that will create. Undue pressure on a trial participant that maybe we can think about alternatives for. Um, it can challenge designs such as, uh, presenting opportunities for more, say, mobile data capture or use of sensor technology, for example. So you're limiting the. The number of times a patient has to visit a site, um, I think we often think that, you know, trial participants live next door to hospitals or research centers, and the reality is, in many cases they don't. They may have to travel long distances. And so the more we can think about trial designs that I. Um, that are amenable to patients, the better it is. Um, long have we tried to get that insight? Certainly working collaboratively with sites, um, you know, clinicians have provided that feedback. And this is just another data point to help us get better and better, um, at designing trials that help people keep people in the trial.

David: 13:42

I, I think there's a lot of promise there. One of the things that I think about for AI is that it at least gives it a sense of empathy, um, which can be important. And even though this concept of, of burden, patient burden has been there and it's kind of assessed and, and analyzed at a certain level. It may be missing out on what the true experience is. Um, I'll just give an example. Um, for me as, uh, somebody who wanted to be a, um, participant in a trial actually for a Lyme disease vaccine, which I thought, you know, I'd had, I'd gotten Lyme disease at one point. I live in an area with a lot of Lyme and also a lot of hospitals. As it turns out, I'm Boston and on Cape Cod and they were recruiting for something great, you know, and it was kind of a rough process. But then when they asked me to be in it, and they looked at the protocol and it's like. Even though I have my own business and a lot of flexibility, there's, there's no way I could do it, forget it. Um, even though I'm motivated for that. So I think that that is a, a place where AI could at least pinpoint what's going on and probably say, you know, here's the trade off in terms of you might be able to get this level of enrollment. If you made these sort of changes, you might be able to get a lot more patients or do it faster, whatever, without compromising the, the quality. So. That really resonates for me what you described.

Tom: 14:54

Yeah, and I think the, the flip side of that is, is also true is as we're able to, you know, automate or AI can start to do more of the tasks that, for example, the, the site personnel used to do, now they have an opportunity to spend more time with patients, which, you know, I. The more clinicians you talk to, uh, the more they would wish that they could do, but they also have a lot of other tasks. The more we're able to automate, you know, coding and summarization and all of the, the, the road work and road activities that they have to do, the more time they have to spend with patients, which I think is also gonna deliver a better experience both for everyone involved in a trial or the stakeholders, but as well as patients.

David: 15:32

So I'm bullish on clinical trials and the, you know, the potential and what's already been achieved for ai. I'm also conscious that, uh, some of the concerns that people have about participating in a trial, uh, may in fact be exacerbated by the introduction of, of ai. Um, are you finding that, are there, what are the elements of AI that, uh, you know. Cause you to consider other sorts of protections that need to be in place, or at least considerations when, when discussing this with, with any kind of stakeholder in the, in the space.

Tom: 16:02

I mean, one of the things that we think a lot about is, you know, we are developing a health technology in a very unique part of, of the care journey. Um, and, and it's a place where openness and transparency is really beneficial. Uh. Especially if you're talking about, you know, patient patients or oncology patients, you know, patients with very serious and potentially, uh, rare and, and life altering diseases. Um, it's important to try and build that trust. Um, and technology can be a build big trust builder, but it, it can also be a, a barrier of trust. And so we, we think about a lot of that in the design of our software, for example. We use patient insights to design our software to make sure that we are taking the patient's point of view into account. Even though they're not customers of our software, they certainly are users The same is is true of ai. It's important that we try to build transparency. I. Around, you know, what data is used, how an outcome was determined. A suggestion, for example, was made. Of course the action was taken. Um, this is not an area where black boxes essentially are gonna work very well. Um, they will only slow the adoption, whereas the more we're able to like open up the insides and, and uh, build confidence about, uh, how the technology is operating and why it's operating, I think is one. And then you touched on the other is. The more we can think about that technology, um, being empathetic and the way that, uh, it interacts. This is I think an area that people often think about, about, about chatbots, for example. Um, early stages of them, it was very clear you were talking to a machine, and I don't think that builds confidence in healthcare, but the more that they, um, they evolve and become more and more sophisticated, the more we're, we're building that. That rapport, if you will, and that that confidence both in patients, but also in the clinicians, um, that are participating in trials that were doing the right thing.

David: 17:56

So soon before all the attention shifted to the large language models and generative ai, there was a lot of discussion about AI and the biases that it could reinforce based on the data that it was trained on. What are you seeing in terms of the impact there with, uh, biases in ai and what do you or sponsors or others do to try to mitigate those biases?

Tom: 18:21

Yeah. Bias has existed in, in analytics since the beginning of analytics and um, certainly as we've developed more sophisticated algorithms, whether they be, uh, machine learning models, now, large language models, um, it's important that we've always had an idea towards, uh. The limitations that the data that we have, um, and the insights that that data can provide. Uh, I think the, the great thing is in, in research and in clinical research, there has always been such a focus on that, you know, thinking very deeply about inclusion and exclusion criteria, for example, so who participates in a trial and what that will mean for the. Conclusions that you can make from it and what limitations you have are important. As we take that into our software development practices, it's very important that we continue to keep that focus on what really is the limitation of the, the technology and the data that we have. Um, I think it's also encouraged a lot of companies to, um, including our own, to, to look in collaborations that help, you know, build. Bigger and bigger sets of data that can start to address bias in, um, in a sort of, uh, algorithmic way, if you will. Um, but the last is, is there's still very much a high human touch in clinical research and that should continue. And that's, you know, the ultimate in balancing out potential biases. You know, we have, we have people at the end of the technology who are interacting with patients and we still rely a lot on them on, on, um, making good informed clinical decisions. I.

David: 19:51

A lot of this discussion, uh, certainly involves things that metadata is involved with, but I wanna give you an opportunity to talk about maybe some of the specific initiatives for metadata, uh, related to AI and how AI fits in. Is it just sort of the thing to talk about today? Is it gonna be, you know, fairly central, but just talk about what your, the kind of, the vision is and where some of the specific initiatives are falling within metadata.

Tom: 20:15

Yeah, it's a, it is a great question and certainly, you know, it's, the question on the front of everyone's mind today is, you know, there isn't a, a meeting that I take that goes by, that someone doesn't ask, Hey, what is your strategy around ai? What are you doing? Um, what I'm very proud about is, uh, you know, ours began, um, uh, more than 10 years ago now in beginning to invest in, in this technology, um, that began with recognizing. The value of the data assets that exist in clinical research and the importance of curating that data and really focusing on the quality of that data so that it could be used to train models and build algorithms that would help, you know, accelerate clinical trials or give us insights that were difficult to find. And so that gave us a real headstart, um, as we've continued to now incorporate that into our technology. I think a big step that we took, uh, last year is recognizing that we really needed to. Uh, accelerate that focus on AI specifically and embedding it into all of the experiences that users have, um, on our platform, but in technology generally. Um, not that the AI should sit off to the side, that something that you. Swivel chair over to ask some questions and then come back to your daily work, but rather that it becomes part of your daily work and more and more it works collaboratively with you to solve a problem. Um, and that's really what we've been focused on. You know, some of that is, is very upfront in trial design and, and startup, um, building a more agentic feel that is helping. For example, make decisions around, uh, countries and site footprint, moving towards accelerating site startup and getting the right documents and agreements around. Um, and then tracking like where they are, uh, but also then during execution that we continue to kind of focus on the anomaly detection, the, the, the, the quality of data, if you will, and accelerating the, the readiness of data for, um, for clinical interpretation. I think that's a, another big area. Uh, the third is we continue to, you know, also be inspired by our industry that has all been about, you know, innovation and, uh, to some degree moonshots. And so we keep looking for what are some of those really transformative opportunities. You know, we've had the privilege of working with some very forward-thinking companies, um, in, in rare disease, for example, that have, have embraced AI techniques to really change fundamentally the way that trials were designed using synthetic, uh. Um, control arms, for example. Uh, really removing people from the control arm of a trial so we could get more people on experimental therapy without losing the, the statistical power of that trial because we can augment it with, um, with synthetic patients. Um, we see. A lot of interest in that space about being able to explore and understand disease and the impact of, of medicine on patients, um, without always having to, you know, poke a patient so to speak, but rather to be able to simulate more and more. Um, and metadata is very proud and the partnerships that we've had to be, uh, kind of at the forefront of that. And we'll continue to, um, to look to be a good partner in that space.

David: 23:15

You know, when Medidata was a startup, um. You, you really changed the industry, the whole electronic data capture, and a lot of folks were kept, you know, flatfooted and there's a lot of very rapid growth for, for quite some time. I'm wondering if you look now at, at ai, which is arguably even more of a fundamental, I. Disruption, although it probably required electronic data capture to actually to, to, to do anything with it, but it's potentially at least as big as as EDC. How is an established company position compared with, you know, startups or startups in a position to go and kind of reinvent things and, and, and leave others flatfooted? Or is there an advantage for incumbents or some kind of combination?

Tom: 23:57

You know, it's, it's a, it's a great, uh, question and one that we could probably spend hours talking about. I love this saying, you know, on the shoulders of giants and, um, I benefit from the 26 years of, you know, metadata investment and progress before me. That has put us in a position to do some of the super exciting things we're doing right now, but it's not a given then that, that will yield another 25, 26 years of, of success. It's still to us to be hungry to continue to. Push ourselves to think in the way that a startup does out of necessity. And I think that's a, a, a thing to really focus on. Um, one reason why startups tend to be really fast innovators is, you know, they take a lot of risks. They try a lot of things. Um, we, we talk more about the successes in startups than we do necessarily about the ones that didn't pan out. Um, and there is sometimes a belief that. Large established companies, uh, big pharma players maybe can't do that. Um, and I, I don't think that that's true. I, I've had the benefit of working for some, um, personally, but also now, uh, working collaboratively with many who I would say are very innovative and are carving out parts of their organization to think like startups as well, so they can be like much more aggressive. Um, but we will continue to see this is a space that there's a lot of startups that are gonna come out with really exciting tech. Um, that we as a technology, a health technology company, will look to embrace and how can we be, uh, a part of that, but we're also a, a part of the more established part of the industry that is gonna make investments that deliver a lot of that at scale. So I guess the shorter answer is, I think it's an all of the above, but it really starts with making sure that everyone is, continues to be like hungry like they were when they all first started. Whether that was a. A company, a, a pharma company that went after their, their first molecule and their first approval, um, and now they're onto their hundredth, um, or a technology startup that, in our case, started in electronic data capture and thought there was a better way to, to run clinical trials, um, to where we are today. So I'm really excited for that. I, I think we are up for the challenge. Um, if anything, we're, we're energized by that. And just from the, the discussions that I have with people around our industry, I think that's generally held is that we look at this for the transformative opportunity it is. Um, not that there is any one person who has like cornered the market on, uh, on success in that.

David: 26:19

Tom, my last question is one that I ask all my guests, uh, which is whether you have a book that you'd like to recommend to our audience, something you've read recently or less recently, and doesn't have to be about business or ai.

Tom: 26:32

You know, I, I laugh because I, I would say the, the book that I read most right now is Dr. Seuss's 1, 2, 3. Um, having two children now under two or under three years old. Um, but it, I am a very avid reader, and now I've, I've shifted more to, to, to podcasts and, um, and there is a very, uh, active community now of, of, of podcasts. So I can listen to that on my way into the office. Um, for example, whether that's on ai, on technology,

David: 26:56

uh, or on current events. We can make it very, uh, self-reinforcing here. You could recommend the, uh, you know, health Biz and care talk podcasts as the ones to it to do I my favorite, uh, at that era, at that age, I dunno if you have, is, uh, Fox and Socks. So that's, I read that this

Tom: 27:10

morning.

David: 27:11

Yeah, that's a good one. How fast can you read it? We could have a little contest. I dunno if I could do it like I could in the old days. You know, I'll tell

Tom: 27:16

you a, uh, a little inside story about, um, about, um, metadata next. We were talking about that just before we started. Yes. So big presentation or big uh, com, uh, customer event that we do. I actually last year read myself, uh, and recited it over and over Fox and Socks, because it's almost like a tongue twister. So it's a, yeah. Great way to get into the motion of, of getting on stage and talking about exciting. That's good.

David: 27:37

I think it goes, uh, Fox and Socks. Our game is done, sir. Thank you for a lot of fun, sir. That's right. In any case, my guest today has been Tom Doyle, chief Technology Officer at Metadata. We've been talking about clinical Trials day and AI and a whole lot more. If you like this podcast, please subscribe. Tom, thanks so much for joining. Thank you.