The Health Pulse

MERI BECKWITH: Basically, trial infrastructure hasn't kept pace with the demands of modern clinical trials. So to show improvement over a standard, you need, effectively, ever more data and ever more robust data. And so clinical trial designs have become more complex.

But the weird thing that's happened is that like per unit of data generated, the costs have increased. So complexity is increasing, and the cost to generate each of those data points is increasing, which is pretty messed up.

ALEX MAIERSPERGER: Today's guest went through a clinical trial himself. Then he founded a clinical trial company. Here what Meri Beckwith, co-founder and co-CEO of Lindus, has to say about why they're an accountable research organization, not just a contract one.

Your company, Lindus, you've called the anti-CRO model. Why does there need to be an anti-Crow, and what does that mean?

MERI BECKWITH: Yeah. Sure. And actually, we've recently updated that to accountable research organization. I guess what we're getting at there is I've seen in my own experience the CRO model, particularly let down smaller, more innovative companies where CROs typically build by inputs, by hours, or by number of units input into a study.

And that means they don't necessarily have an incentive to keep the study moving. And in fact, what some CRO executives shared with me that led me to found Lindus was that their business model relies on there being overruns and change orders that inflate the scope of the study.

And so their incentives are just misaligned with sponsors, with pharma and biotech companies and with patients. And so what we mean by the accountable research organization model, is our business model is very simple. We get paid a fixed success fee every time the study hits one of its milestones, and those fees are pre-agreed. So if it takes us two months or 20 months to hit full enrollment for a study, we're charging the same amount.

And so our incentives are fully aligned then with our sponsors and our patients to do whatever it takes to keep the study moving. So yeah, that's what we mean by the model.

ALEX MAIERSPERGER: That's the accountable side. I love that. On the health side of the realm, there's, obviously, value-based care. And so this sounds like paying for outcomes versus paying for pure services.

MERI BECKWITH: Yeah, that's a pretty good analogy.

ALEX MAIERSPERGER: You've said that participating in clinical trials yourself really opened your eyes to how inefficient the system was. And so that means you've gone through this and experienced it. It's not just I saw this as a need, and here's what I'm building. It's I've experienced this. I've lived this. What surprised you the most when you experienced trials from that patient side, and what motivated you to do that in the first place?

MERI BECKWITH: Yeah, good question. I think the biggest surprise was just how antiquated the experience felt. So I remember thinking the value of a participant to this trial is in the tens of thousands of dollars, and yet, they're making it really hard.

It's really hard to book a screening call, visit with a nurse. I'm getting conflicting information about when I should attend the site. There's very little flexibility. And yeah, there's a mismatch of different systems. Data is being captured on paper. The website that I'm supposed to sign up on doesn't really work.

And I remember thinking, look, if this were any other category of company, if you have a customer who's value to the company, lifetime value is worth $10,000, let's say, you go to extraordinary lengths to make sure those customers are well cared for. And I think, yeah, just that contrast really opened my eyes to how much opportunity there was. If we could improve patient experience, we could improve enrollment rate, improve retention of patients on studies, and all of those cascade up to much better outcomes for biotech and pharma companies.

And yeah, that's of what we've seen in the journey with Lindus as well. So yeah.

ALEX MAIERSPERGER: In some of these conversations, I have heard that the patient recruitment is a challenge, and so you're not filling up the studies. And then there's, obviously, all sorts of implications of that. So the cost side, the speed side, how fast you're able to get drugs to market side, what you've just said is the human complexity side or just the experience side that you have. Can you, maybe without sharing every personal detail, if you had a side effect or those type of things, how do you actually go about that?

Some people say it's like patient willingness. It's like, I'm healthy. I don't want to be participating in a clinical trial. I don't want to put something in my body that I don't know what's going to happen to it. Or even if I have a medical condition, it's, I don't want to take this experimental drug. Just give me what you know and what's been researched.

And that seems to be fairly controversial. Some people say we do have a ton of willing people. We're just bad at getting to them at the right time. Can you give me a little bit of a process? If I want to take part in a study, do I just talk to my doctor about that? I fill out a website form? Where do I go?

MERI BECKWITH: Yeah, so the traditional research model, which is still employed in 90 plus percent or 95% of clinical trials today is everything is kind of handed off to the trial site. And these are, sometimes they're dedicated research sites. Other times, they're ordinary specialists or health providers that do research on the side.

And so what this means is, in the traditional model, I, as a patient might be seeing my doctor for a visit, and they might say, oh, by the way, you're eligible for this clinical trial I'm running. And then give me a flyer and then hope I sign up. Or they might just call me up from their database and say, oh, you're eligible for this trial.

And that's a challenge because the health care system is overwhelmed and is pulled in many different directions, lots of competing priorities. And so understandably, research sites often aren't the best at reaching out to patients, convincing them to take part, making it easy for them to take part. And so yeah, that's where a lot of the problems come in.

The other big problem, and you've kind of alluded to this as well, is that in this research site model, that's, obviously, a kind of self-selecting sample of the kinds of patients or research sites that could be suitable. Because not every health care practice, every geography is kind set up to do research.

And so what we've done is say, OK, let's flip the model on its head. Let's just use the sites as the place where the study visits happen. And let's try and do as much of the patient finding centrally as possible. So finding patients in the community via other parts of the health system that aren't set up to be research sites, via patient communities, via even direct advertising, all of those are really effective.

And we'll do all of the patient screening centrally and then just hand deliver those patients to sites, so they can be, consented and then take part in the study from there, so decoupling that patient enrollment, that patient recruitment part from the side part, which yeah, has been very effective.

ALEX MAIERSPERGER: We're starting to see more adaptive trials. There's some new regulatory guidance that supports some different trial design. Where do you think trial design and not just the technology side needs to change the most?

MERI BECKWITH: Yeah. I'll start by saying where I don't think it needs to change, which is the randomized controlled trial design where you have groups of patients or randomly allocated to a placebo or treatment arm. The reason this is so good is randomization is, basically, magic, because it allows you to control for all possible variables without knowing what they are and run a truly clean experiment to show that something has an effect.

So I think there will always be a place for randomized controlled trials until AI is able to simulate every aspect of human biology, which is a long way off, I think. But to directly answer your question where it needs to change, so yeah, I mean, the FDA has been on a tear recently of releasing really interesting guidance, firstly, for Bayesian clinical trials versus a Frequentist approach.

So this basically means, rather than run a big, expensive study and find out at the end whether it worked, you can basically say, OK, well, given this prior evidence I have, this suggests this treatment might work. I'm going to feed that into the statistical design of an upcoming prospective study. And basically, lowers the bar or it allows you to take that previous evidence into account.

So this, essentially, just means we can be more flexible with study designs, potentially needing to enroll fewer participants or needing to give fewer participants a placebo. And that's critical for imagine God forbid, you have terminal cancer and the only hope is a potential new oncology treatment being trialed in a clinical trial. You really don't want to get the placebo on one of those studies. And so a Bayesian approach can help minimize or eliminate the need for placebo overall.

I think something else, which is really cool is the FDA has issued guidance about combining phase 2 and phase 3 studies. So today, you have phase 1 clinical trial, phase 2, phase 3. Going through that process takes, on average, 10 years for each new approved medicine we have. That's just simply way too long.

Patients don't deserve to wait that long, and of course, that has a cost to it as well, because that drug, while it's in a trial, could, instead, be earning revenue in the market. And so, yeah, I think supporting, as you said, in an adaptive trial where you have one trial where maybe you're trying three different dosage, maybe combinations of dose and patient population in an initial cohort. And then the strongest one gets rolled out into a larger cohort. But it all happens in one combined phase 2/3 trial.

That's just an obvious win. That doesn't sacrifice any data, but allows us to potentially get these products to market a lot faster. So I think that's really interesting and something I'd love to see sponsors actually do and adopt more of.

But yeah, as a sidebar, something we see again and again in this industry is the FDA will roll out guidance. They will give people permission to do an adaptive trial, but uptake will be really low from the industry. And yeah, I think it's a shame. And not often do you see the regulator pushing private companies to adopt new technology and new approaches. But yeah.

ALEX MAIERSPERGER: I think we see that quite a bit across health care. Some of it, I wonder if it's a personality trait too, if we're just we're cautious from a physician side of thing, and rightfully so. I remember I worked in a hospital setting as a hospital administrator, and one of our orthopedic surgeon leaders got up and we did these personality tests amongst the team. And it was the more analytical side was over here. The more emotional side was over there. And it was like, everyone in the entire room was over on the analytical side of like, wait until I see the data.

MERI BECKWITH: We've done the same.

ALEX MAIERSPERGER: And so Yeah, it's just it's like a personality test. So even if the regulators say you can go faster, it's like, we'll find a way as an industry to not get there.

MERI BECKWITH: Yeah. Do no harm rule number one. So yeah, that definitely shows.

ALEX MAIERSPERGER: You said something sort of jokingly of like, I don't think AI is there yet or won't be fully-- like, we're not going to run full AI trials with no humans involved. But I think there was a word there that was yet.

Where do you see AI having the biggest impact, and where do humans fit in the loop? If we can get to-- I've seen numbers thrown out there of 80% of clinical trial work could eventually be automated, or some of the pre-data prep and some of the post looking at things, where does I sort of fit in this picture? And then, I guess, where do humans fit in the picture in the future?

MERI BECKWITH: OK, yeah, a bunch of different places. If you think about what is a clinical trial, it's a structured experiment that needs to follow a very rigorous protocol to keep the data fair and clean. And then that data needs to be monitored and interpreted.

These are all tasks that fundamentally AI is, well suited to. I'll give you some concrete examples, so it's not airy fairy, but today we use AI heavily for document generation. So typically, a sponsor of biotech pharma company might give us an 80-page clinical trial protocol that sort describes the experiment to be run. But that is still a far cry from a complete clinical trial that's ready to launch.

And so there are a lot of other ancillary documents you need to produce. So patient safety plans, monitoring plans, data management plans. So we use AI to generate first drafts of those plans, working to a specific template, and then have our human experts review and add to them. And so I guess there's a theme there of we're not using AI to completely replace humans, but to augment them.

Further down the chain, we use AI to monitor. We're starting to use AI to monitor clinical trial data. So again, clinical trials are generating millions of data points. Traditionally, humans monitor those by hand. You might have doctors looking at case report forms, or-- I don't know, data from a site and trying to spot, oh, is this a potential adverse event, like a negative medical reaction to a drug that we need to follow up on?

AI is just has superhuman capabilities there, because it can look across far more data points than a human and can stitch together insights from seemingly disparate things. And so that's been really valuable. And what it's also starting to spot is operational issues that humans would definitely miss.

So for instance, indicators that a site might be entering fraudulent data, or a patient might be entering fraudulent data, or that a piece of equipment at a site was miscalibrated, that kind of thing. So yeah, we're just, obviously, in the foothills of what AI can do.

ALEX MAIERSPERGER: You talked about your own personal experience with clinical trial, and then you talked about some of the opportunities on the site side. One of the things you mentioned was the physician experience of, if I'm going to go into a clinical trial, sometimes it's up to the physician or to them to be able to say, hey, there's this clinical trial that you're eligible for.

How about their experience on that physician or the site experience? Where is that breaking down? Where can it be improved? Are you doing anything to say-- I can imagine a physician, obviously, has limited time with each individual patient, limited time to understand all the studies, all the opportunities that are out there. How do you get to them to be able to say, hey, this clinical trial is available for you?

MERI BECKWITH: Yeah, absolutely. So yeah, it's a really good point. And similar to the patient experience, the better experience a physician has, the better the trial goes, basically. A few really basic things is just providing them simple tech that works.

So traditionally, on a clinical trial, physicians might be capturing information on paper. They might have multiple systems that aren't integrated that each have a different login, which obviously makes it harder to engage. We have our own technology platform. It's part of how we're able to run clinical trials more efficiently. It combines functionality of these are industry terms and very acronym heavy, but EDC, CTMS, and database. And so that's already one login to use.

But on the patient recruitment side, something we do is, it's all about taking as much burden off the physician and off the health care system as possible. So rather than rely on physicians to remember when a patient is eligible or to manually search through their records, we use AI to take trial eligibility criteria, turn it into search codes that can just be run automatically against a set of patient electronic medical records and surface patients that are eligible. And then reach out to those patients with approved messaging. And then our platform kind of picks up from there of educating the patient about a trial, and so on.

So all of that replaces manual work that a physician or a coordinator at a research site would potentially do. And so that's typically where one of the bottlenecks is for patient recruitment.

ALEX MAIERSPERGER: Patient centricity. Maybe this is a hype versus reality or a how far we've come type question. Patient centricity has been a buzzword for, maybe, decades, a decade plus. So we've talked about it for years. You hear it on main stages at conferences. Has anything meaningfully changed on the trial side, or is it still talk?

MERI BECKWITH: I think it's hard to say. I would say, yeah, it's tautological, but the reason it's still a kind of hype buzzword is because it's just not living up to the hype. The reality isn't living up to the hype.

Yeah, I mean, I was at a conference yesterday where the topic came up a lot, and everyone agreed it was more hype than substance. Why is this happening?

Yeah, I think, again, this industry moves very slowly. And the experience I've had as a patient is testament to that, very simple technology. And just thinking about the process from a patient perspective could have massively improved the experience for patients and therefore, the enrollment rate retention and so on.

Why doesn't it happen? I think, yeah, there's a lot of industry inertia. I've of seen it from the sponsor side where there was a study where the sponsor was failing to enroll. Actually, sorry. Well, this is another study that I was approached about enrolling in recently.

It's a study. It's fairly heavy duty. It would involve six hospital visits in six months, multiple endoscopes, not pleasant. I'd have to take time off work. And I said, OK, look, I just don't think I have time to do this. Are you at least paying for participation, or can you fund travel?

And they're like, no. And I'm like, OK, well, are you wondering why you're struggling to enroll patients? Why aren't you paying for travel? And it's like, oh, the sponsor doesn't want to. And I'm like, well, they're worried about coercion. I'm like, well, not paying patients for this much time I would also argue that's unethical, as well as crippling your enrollment.

So yeah, I think there's a bunch of things there. But yeah, I think just the usual industry inertia is probably the biggest culprit.

ALEX MAIERSPERGER: You've worked inside companies. You've invested in companies. Leading a company now, what's been harder than you expected, building a company in this space?

MERI BECKWITH: Yeah. I mean, I think it comes down to people. And obviously, as companies go, we're, obviously, a people-heavy model, rightfully so. But yeah, making sure when you're building an organization, you're doing it thoughtfully. You're thinking about how all the different parts of that organization work, how different people fit together.

And hiring. Hiring is probably the one magic bullet. So if you hire the correct people for every role, everything will just get magically easier. But that's basically impossible to do. And so how do you make sure that you're even interviewing the right people for the role? And then how does your process make sure that you're assessing them correctly.

And then once they're in the role, giving people the support, mentorship, guidance and ownership they need. So yeah, that's definitely the hardest thing. And yeah, it just never gets any easier, unfortunately.

ALEX MAIERSPERGER: Not to steal you away from your current company, but if you were dropped into the other side on the pharma company side tomorrow, what's the first thing you'd change about how they run trials? Is there any advice that you have right now being able to work with them? What advice do you have for the listeners that are designing and running them?

MERI BECKWITH: So Yes. So I think-- where to start? There's a lot of things. So I think starting really high level, taking a much more pragmatic view of clinical programs. I think biotech and to-- well, some extent pharma-- have a tendency to gold plate trial designs.

So you get protocols that are designed by committee where multiple stakeholders each have a pet project or piece of data they want captured, and you end up with this monstrous protocol that doesn't do anything, let alone the most important thing of whatever showing efficacy. So I think just being really ruthless about clinical and pragmatic about clinical trial design.

Number two, embracing some of the more adaptive trial designs, particularly at something like phase 2, where you should just be able to get much more signal from an adaptive design where you're simultaneously comparing multiple patient cohorts or multiple doses. I don't think that's done enough.

It's pretty crazy if you think about most other products. They go through hundreds, if not thousands of iterations before reaching end users, whereas the drugs you put in our bodies maybe go through three or four substantial iterations in terms of dose or formulation, from whatever, phase 1 to approval. And so yeah. Yeah, I could go on, but I'll stop there.

ALEX MAIERSPERGER: That gives some real actionable items though, for those that are leading the design side or leading the trial side. So really helpful. Yeah you mentioned the 10-year timeline and maybe longer, in some cases, of drug discovery and then development and getting through the process.

If clinical trials do run dramatically, faster and more efficiently-- and I'd love to hear that sort of time frame-- are we at 10 years now and we get it down to 8? Or is it we get it down to months, something? What would that mean for innovation in medicine at large and for patients waiting on new treatments?

MERI BECKWITH: Yeah, so I mean, firstly, how fast can it go. I mean, COVID showed us, Operation Warp Speed, vaccines, two years. And those were trials run as much data was generated as a trial that was run in a non-pandemic setting. It's just that the phases were condensed. They were run in parallel, and people were motivated to move quickly, firstly because of the pressure of COVID. And secondly, I think, frankly, the competitive pressure.

A lot of times, pharma doesn't have that. There aren't that many competitors vying for the same indication. So yeah. But, realistically, yeah, I mean, I see no reason why you can't shrink that down to just a few years. But I think it'll come gradually.

So we know from our track record that we can meaningfully shrink startup and enrollment periods in some cases by years. So if you apply that across a whole program, that's shaving a couple, maybe two or three years off the program. But yeah, I think we can go much further.

ALEX MAIERSPERGER: Did you see the headline? I think the story is actually a few years old, but it made a recent pass in the news again of the Australian guy with his dog who had cancer. The dog has cancer in the leg. The Australian guy goes to use an LLM and essentially, gets the step-by-step instructions of first sequence the DNA, then create the vaccine.

And then he found like a bespoke vaccine manufacturer that was willing to make one dose of this vaccine, and it shrunk the tumor in the dog's leg by 50% or 60% within a few month period. Anyway, it was like the story ends on the miracle of my dog wouldn't stand up, and now, my dog is jumping over fences and chasing rabbits again.

How realistic is that? Is there a time frame that that's a person's individual cancer journey, where it's I can create my own vaccine? I can create my own wonder drug?

MERI BECKWITH: Yes. So firstly, that is essentially, there's no reason that workflow doesn't work in humans today. And I think the responses to that story were really interesting. Rorschach test, where you had all the more Silicon Valley people going, oh, isn't this amazing. Look, this is coming to humans.

And you had all the life sciences people going, actually, this wasn't a properly approved drug and whatever. This is just some guy doing it in his dog. But what I think it's revealing is that the barrier to that happening is, essentially, partly regulatory. So when you're designing a medicine that goes into a population, you rightfully need to go through clinical trials to show it's safe and effective.

But if you're just doing an n of 1 experiment, I legally today, could do that on myself. And there's basically no regulation that stops me from doing that. You might struggle to find doctors who would sign off or oversee that treatment. But there's nothing stopping anyone from doing that today.

So yeah. So I'm not sure if that actually answers your question, but just some thoughts on the reaction to the story.

ALEX MAIERSPERGER: Really helpful. Yeah, and maybe worrisome, I guess. Maybe you just gave permission for people that if someone is out there looking for the legal rules right now, they're like, wait a minute--

MERI BECKWITH: Whoa. Hey. This is not legal or regulatory advice. Yeah, it goes without saying.

ALEX MAIERSPERGER: I heard permission then research. All right, it's 2036, which sounds very far away. I think we were promised in the movies that 2020s, we would have more flying cars than we do. We would have all this sort of teleportation technology.

But 2036 is not very far away. Still 10 years. Lindus health being wildly successful, what does a clinical trial look like from the patient's perspective 10 years from now?

MERI BECKWITH: So actually, picking up on that theme, I think we should see bifurcation between personalized medicine, for which the standard clinical trial process is inadequate, as demonstrated by that story where personalized medicine maybe goes down a separate regulatory pathway.

And just the notion of a 500-patient phase 3 just breaks down when you're talking about personalized medicine. But I think the other half of the bifurcation will be these larger chronic conditions, like we've seen in, obviously, metabolic disease and obesity, where the biology is pretty uniform across the population. And to do those efficiently, we need to get much more efficient about how we design and conduct research.

So what I think is really interesting here is that, if you look at a graph of pharma or life sciences, R&D productivity, so the return, the productivity of every dollar invested into life sciences R&D, that graph is basically just a massive exponential downward slant.

And it turned negative somewhere in the 2010s. So basically, since 2010, on average, every dollar invested in developing new drugs has been wasted. It's produced negative return. And that's despite the COVID vaccines. It's despite all the progress we've made in metabolic disease.

And so what that shows to me is even with these really common diseases that we are making huge strides against, we still need a research infrastructure that is just much more efficient. And I think we're showing how that's possible.

And so to actually answer your question, yeah. So I think you'll see that bifurcation, and I'd like to see Lindus be a significant part of just a different model for conducting research into common chronic conditions.

ALEX MAIERSPERGER: What you just said is fairly shocking, I think, from a public standpoint is you think and you see some of the headlines that just say we've gotten so much better at treating cancers or we've got more personalized therapies. So it feels like we're making some big headwinds on some diseases.

But is it the amount of money that's getting invested to get there has just become so exponentially bigger? Or is it, really, we're all worse at math than we used to be, or what's the--

MERI BECKWITH: Yeah. What's going on. So basically the trial infrastructure hasn't kept pace with the demands of modern clinical trials. So to show improvement over a standard, you need effectively, ever more data and ever more robust data. And so clinical trial designs have become more complex.

But the weird thing that's happened is that per unit of data generated, the costs have increased. So complexity is increasing, and the cost to generate each of those data points is increasing, which is pretty messed up when compared to 20 years ago, we have a much better understanding of human biology. We have far greater, far more powerful technology. So something that is very broken.

And to me, it's pretty clear that it's this business model that rewards activity over outcomes. You can see in the lack of technology adoption across the space. I gave some examples about the FDA pressuring the industry to adopt different trial designs. They've also been pressuring the industry to adopt more technology in how clinical trials are monitored, for instance. So I think all those factors are playing a role.

But yeah, it is. It is really scary. So that negative trend means that if nothing else changes, people will just stop investing in new medicine, and we'll just have flat or declining life expectancy and life quality, which I don't think is something that no one wants. So yeah.

ALEX MAIERSPERGER: This is almost speed round-ish, which we're not quite there yet, but putting you on the spot for, again, hype versus reality. And maybe it's like optimism versus pessimism, where do you fall in the camp of like-- and maybe it's over 20 or 30-year time horizon, so a little bit further future, do incentives stay in place and we keep plugging along at the pace we are and try to squeeze a little bit more efficiency and things out.

Or is there a world that you really tangibly see where we cure a bunch of forms of cancer and where we have individualized drugs and readily available vaccines, and just that amazing future? How close or how far apart is that split?

MERI BECKWITH: Yeah. No, I think we're the precipice. And there's a notion that you'll see people referred to called the century of biology, which we believe in. If you think about it as a pipeline where the top of the pipeline is our understanding of biology, that is just exploding. Thanks to AI.

We also have never had so many modalities, such an exquisite understanding of what targets are drugable. And further down the pipeline, you get to the clinical research phase, where we need to see if this stuff actually works. That's where the bottleneck is. All of these problems are really surmountable.

I think with the amount of money and the amount of excitement being poured into the top of the pipeline, you will just see a pressure and a demand the bottom to unblock. I also think you'll start to see more competition in pharma. You've already seen this in the metabolic health space.

But one of the reasons pharma is so inefficient is because they just don't, frankly, suffer from much competition. Each pharma company is almost like slightly glib, but it's a collection of monopolies where they may only have one or two other competitors in the space who can do the clinical development and the distribution.

But that's changing. So we've shown that clinical development, the clinical trials part is getting easier. It's getting cheaper. And the distribution part is getting easier too. Telemedicine, just the ease of marketing direct to patients is all changing. So yeah.

ALEX MAIERSPERGER: We see aging populations all over the world. We've talked to a few senior experts or experts on seniors. And so part of the challenge that they talk about is brain and heart. It almost came down to lifespan and longevity, essentially, comes down to how long your brain can function adequately without Alzheimer's and some of the other challenges that come along real advanced populations, and then how long your heart can beat successfully.

So is there an area where we see major investments in those two categories mostly? Is it cancer? Are there personalized chronic conditions? Do you have any thoughts on just the overall investment over the next few years?

MERI BECKWITH: I think, yeah, there's been a lot of excitement in the longevity field, but it's not really lived up to the promise. And I think because no one can really agree on what constitutes a longevity drug. Just because the only real biomarker for longevity is healthspan and lifespan.

But I think, ultimately, the GLP-1s kind of are longevity medicine. And so I think we'll start to see a fusion of the longevity movement and the traditional world of biopharma.

ALEX MAIERSPERGER: AI's promise in the CRO space or the accountable research organization space, is it like, currently agents? Are there going to be some augmented employees? Is it new technology that you're pinpointing exactly places? Is it an overall underlying platform AI everywhere?

MERI BECKWITH: So yeah, I think, it's definitely augmentation, not outright replacement. Very hard to predict. I mean, right now, it's largely speeding up existing processes, and it's starting to allow us to do things that we just couldn't do before, like with the monitoring example.

Pretty hard to predict, and I haven't thought that deeply about what it looks like. I do just think that if you really break down what happens in a clinical trial, I mean, basically, everything except directly dosing a patient, being there physically with them should be automatable via AI.

That doesn't mean it makes sense to. But if you play that forward, I reckon about 70%, 80% of the kind of manual work that goes into a clinical trial will be automated. So don't know exactly how we get there, but we think we should be the first to figure it out.

ALEX MAIERSPERGER: Well, patient centricity has been a buzzword for a long time, but from this conversation, it sounds like you're truly putting the patient first. And in some cases, you've been this patient. And so being able to say, here's what we can do to create a healthier future and truly faster is super exciting. So, Mari, thank you so much for being on the Health Pulse Podcast today.

MERI BECKWITH: Yeah, Thanks so much for having me. Great to chat.

ALEX MAIERSPERGER: If you want to learn more about clinical trials or you want to join us as a guest, please email us The Health Pulse Podcast at sas.com. We'll see you next time.