From Reactive to Real-Time: How AI Is Reshaping Clinical Trial Operations

Josh Stent, CEO of Momentum Clinical Research, joins Jimmy Bechtel to share how his 19-site network in Australia and New Zealand is using AI and automation to cut administrative burden, improve data accuracy, and get more time back for patients. From centralized data reporting to automated invoicing and smarter recruitment funnels, this episode breaks down what practical AI adoption actually looks like at the site level. 

Transcript

Jimmy Bechtel 0:00

Hello, and welcome to SCRS Talks provided by the Society for Clinical Research Sites. My name is Jimmy Bechtel, and I'm the chief site success officer with the Society. Today, I'm joined by Josh Stent, who is the CEO of Momentum Clinical Research, here to share a little bit with us around how AI and integrated systems are starting to change how we approach clinical research operations. But before we get into the topic, I know everyone's probably really excited to hear from Josh on this. Josh, first, let's learn a little bit more about you, if you wouldn't mind sharing, some of your background.

Josh Stent 0:34

Yeah. Thanks, Jimmy. Josh Stent, CEO, Momentum Clinical Research. I've been working in the sector now, for, two and a half years. but prior to that, spent 10 years in, the medical field, both in, general practice, urgent care, and, dentistry, integrating businesses and transforming digitally how they work. over the last several years, I've brought that same, transformational mindset into clinical research. and at Momentum, we've built a scaled network, in New Zealand and Australia of 19 sites, where we run everything from, first in human healthy volunteer phase one trials through to large scale, phase three and phase four studies. and I think what I've realized, on the ground here, is, some of the old ways of working, can't keep up anymore. And watching the space that we operate in, in clinical trials, there's a significant amount of process and administrative burden, that we can do a lot about by thinking, thinking in new ways about how we use technology.

Speaker 1:25

That's great, Josh, and that's, exactly what we're here to talk about today. AI is, the hot topic amongst the industry right now, particularly at the research site level. so any insights we can bring to, the enterprise will be, helpful, I think, for everybody. And even in healthcare in, in general, we know that it's starting to find its way into a variety of different aspects in which we deliver care, but of course, in the clinical research arena as well. So from your perspective, where do you see the most meaningful applications of AI in our, clinical research operations?

Josh Stent 1:58

Yeah. I think, Jimmy, to touch on your point there, I think that, what we're seeing is, a lot of ideas about how AI can be used. So I wanna just try and touch on today a few, practical applications of what we're doing at Momentum and give everyone a bit of a point of view. So- If we think about, it being discussed in healthcare and the most meaningful applications, the era of reactive spreadsheet-based clinical trials operations is over. I think if we want to bring therapies to market at speed, we need to recognize that the modern protocols, the complexity of those, we can't afford to be waiting for things like data queries to come in or milestones to be missed to figure out that something is broken in the process. So what I think we're seeing is a major shift from that retrospective observational to a proactive real-time management of the data. and that transition is not just about bolting on new software, it's about transforming, in my mind, the whole ecosystem. So rather than having siloed, fragmented database, we move to, AI-driven environments that are resolving bottlenecks, to ensure that they don't impact timelines So if I think about what we're doing at Momentum, a couple of examples of that would be, we've centralized all of the data into a kind of enterprise-level reporting data lake. And what that means is o- over the course of an evening, for example, we can, run a script, some interpretive information through, say, the source data, in order to understand where there are errors in that data and present those errors back to the team. So when they come in first thing in the morning, they can see any deviations to the protocol that might have occurred, anything that's been missed that specifically needs to be tidied up. That way we're not waiting for a monitor to come and see us, then correct things, historically. What that means for our sponsors is that they get, much more predictability and confidence in the data. and it means that ultimately they're gonna be able to get their products faster to market. It means that our team can spend more of their time, rather than, dealing with monitor queries, which takes a significant amount of time, and I'm sure will resonate with this audience. we can spend the time seeing participants and working on that next study.

Jimmy Bechtel 3:50

and Josh, that's really a common theme, and that's what this is all about, and something that we constantly find ourselves continuing to go back to and really need to focus on. And it's not AI replacement, it's AI augmentation. How can we augment and, enhance the work that we're doing to free up our time from doing the remedial, redundant tasks, like you said, like monitoring and what goes along with that, so that the sites and the site teams and the patients can do what they need to do together to advance medicines and get things done a lot quicker. So I think that's, really the main focus here is what applications- When we're looking at these, we ask ourselves, can go to do that so that it frees my time up to, be more attentive and more present with my patients.

Josh Stent 4:41

Yeah. I think I remember speaking to, Rick, at, Realtime, who a number of people will know, when I first started in the sector, and trying to get an understanding of the utilization of our team's time, and the time that it spent seeing participants, versus the time spent, inputting data or doing other administrative tasks. And, when we start to understand and report on some of that through these tools that we're using, we can see that sites are spending, 20, 25% of their time seeing participants, and the rest is in processing. and how do we identify and break down those processes, to create more time, that ultimately creates more value?

Speaker 5:15

Exactly. I couldn't agree more. and when we then take this into r- recruitment, speaking to patients, we know that based on data, anecdotally and, non, subjectively and objectively, it remains the biggest, if not one of the biggest challenges in clinical trials. Josh, maybe talk a little bit about how you think AI can help sites and networks improve recruitment from maybe identifying eligible patients more effectively to potentially affecting and streamlining really the pre-screening and outreach process.

Josh Stent 5:48

Yeah. I think it might be quite good, Jimmy, to break this into two parts. So the first part is, if you think about the top of the marketing funnel. So one of the things that we... identified early on is that, when you start to work in the sector, having come from somewhere else, what you realize is that there's just an incredible amount of work that goes into solving problems for future patients, and that the process of being involved in a clinical study for a participant is, in fact, very safe, but so many people don't know about it. And part of the reason for that is 'cause we historically have spent so much time from a recruitment point of view targeting the people that have been involved in a study before or that have been on a database or mining database to do this, rather than also spending time on the top of the funnel, which is creating awareness for people to get involved in studies. So the first thing related to AI is, as a consequence of being able to use, AI tools- We have been able to significantly open the funnel from an advertising point of view to get people to put, say, one or two bits of information into a social platform that enables us to then make significantly higher volume of calls to get people into the database. So, for example, you might respond to an ad with two or three bits of information. That ad then pulls into our practice relationship management system. We're then able to quite quickly surface up a series of questions that you can go and answer related to your healthcare if you're new to us, rather than asking you to fill out a huge form online where you traditionally would get bored and not get involved. Then as we run that through the system, we're able to appropriately segment who we're going to call first by lifting the people that have answered the most amount of questions or look like participants that have been involved in studies before to the top of the list. Because what we also know is that when it comes to healthcare and when you're dealing with older populations, people still want to speak to a human, right? We can't remove that from the process. But what we can do is trigger the right calls to be at the surface to make sure that we're optimizing our recruiter's time to make the appropriate calls first and therefore being able to convert people into studies faster. So the other thing that we've been able to do with the database specifically is start to segment and break down the database into therapeutic areas. Then we can engage with those participants in the database by serving up really great content that talks to issues that they might be having or new research that's out related to their diabetes, for example. So the participants stay engaged and warm. So when we have a study that comes up that serves those particular participants, we're able to quite quickly automate a process that sends them some information and gets them involved in the study faster. So, you know, there's a couple of examples there. Ultimately, what this does is both create further awareness by getting people into the funnel, but then also enabling us to get to the right people faster so we can recruit and, yeah, keep our sponsors happy.

Speaker 8:52

That's great, Josh. And again, goes to speak to your point here. The message that we're trying to get across is this enhancement concept, this augmentation, this removal of redundant things. But it also goes, especially when you talk about your recruitment work and things you were trying to do, it also goes to speak to tailoring the message and the work and what's being done and how you're approaching recruitment to the patients and the specific patient population. It helps narrow in and maybe take some of the guesswork out of how do I bring what I'm trying to bring to these patients most effectively in a way that's going to be received by them, and how do I reach them most effectively? and it incorporates some of that more strategic, approach around patient recruitment and identification into the recruitment process. eh, thus, again, alleviating the work that might go into trying to formulate some of that.

Josh Stent 9:48

I think that's right, Jimmy. And if you think about, a lot of this stuff is a new, more broadly, it's new to our sector. And I think because of some of the limitations of, the CTMS products as an example being the core basis for clinical trials and not customer relationship marketing, we're able to accelerate some of those principles that you would find in a lot of other sectors and bring them to our own sector in order to engage and qualify people more quickly, to get them onto those studies.

Speaker 10:14

That's exactly right. That's what it's all about. That's what we're here to do. and any way we can do that more effectively and/or more safely, and with patients in mind, I think we're all, we're all on board with that. I wanna talk a little bit about workforce here too, Josh. clinical research, we know is so process driven. and teams often carry a lot of the administrative burden at the site level. So let's talk a little bit about how AI tools can empower staff then and increase productivity without losing that human element, the expertise that we know a lot of these trials really depend on.

Josh Stent 10:50

Yeah. I think, Jimmy, the thing that you pointed out there, which is critical, was, the greatest asset that we have in our industry is our clinical professionals. the investigators, the nurses, the coordinators that, are able to do the interpretive work when they've got a patient in front of them that no machine is going to be able to process. And yet, we we routinely subject them to this huge administrative burden that not only drains their time, but also, can compromise, our ability to get more participants in, or their overall experience as they're left with that burden. really what we've been considering and what I think needs to happen is it's about rethinking the division of labor, if you like, between, our incredible clinical teams, and the machines. And, what I see as being critical here, is that when it comes to, productivity and the processing of information, this isn't necessarily a job for AI. but more a job for, what's called robotic process automation or people might have heard the term RPA. if you think about what AI is doing, it's very interpretive. it's taking something and interpreting the information and saying, "This is what you might do with that information." Versus robotic process automation, which is literally taking a task and being trained on how to do that task in the same way that a human does. Ultimately, we wanna preserve the human element to, make sure that is being interpretive, doing quality, checks, and real-time auditing, but we can significantly speed up the process. And so an example of some of these things that we've been putting into practice every day would be, setting up study templates. we've got 19 sites at Momentum. If we've got 5 to 10 studies on a site, we can effectively set up the study templates by, getting, RPA to push those study templates out, so they're the same across every site. we can then do the work of, loading up the contracts and then at the end of each day, kinda checking against a contract to make sure that the invoicing has been done appropriately. we might have, hundreds of open studies on at any one time, so to expect a, coordinator, or multiple coordinators to be able to kinda go in and understand from an invoicing point of view how each one of those things needs to be processed, versus being able to do that in an automated way, then get, AI to do the interpretive work to understand what's happened through the course of the day and spin up differences, which means that, we can significantly reduce, the invoicing burden, but also the accuracy of invoicing for our sponsors. another good example is the work that we're doing, on this with, setting up pre-screening templates would be another example where we're able to, get the RPA to do the templates across, across multiple sites rather than having to get somebody to go in and process, multiple spread- spreadsheets. The other one is the data entry into EDC. this takes such a significant amount of time. and while we, don't wanna be doing anything interpretive when we're moving information from one place to another, the most basic thing is being able to move what is source data that somebody's put into the source, into an EDC, which is simply a task of moving data from A to B rather than something specifically interpretive. So we don't necessarily want AI making those interpretive decisions. we want, a human to, be able to, be in control of that. but for our sponsors, ultimately what it results in is, a significant reduction in human error. and it means for patients and our participants that are involved in the studies, they get much more face-to-face time with our professionals who can, spend time taking care of them, helping them to learn about their disease, rather than our staff staring at screens all day.

Speaker 14:09

Those are interesting examples, Josh. And, what I heard from that, and I guess the message I would help our audience understand is first taking the time to identify, okay, what are the capabilities of AI, which are understandably valuable to know, but also separate from then how do I apply that? Where do I apply AI and automation and empowerment versus where do I need to remain in the loop? And you used that word interpretive, and I think that's where we need to continue to rely on our staff is that, again, as you mentioned, that interpretation part of it. allow AI or some other AI-based tool to do the work and do the execution and, the remedial task associated with it, but when it comes to interpreting what the outputs are, that's where we need to keep the, quote unquote, "human in the loop," is a common term in this space, to an- analyze and understand and apply the information that we're able to gain from those tasks as examples that you mentioned.

Josh Stent 15:11

Yeah, that's exactly right, Jimmy. I think where, where people go wrong is, first trying to identify and understand the problems that you're trying to solve. And then think about the right application to solve that problem. and then overlay the human element, which is AI's, anything, or RPA or any of these tools, is only as good as its briefing. And then only as good as your ability to question, everything about the response that you get to ensure that, the data is accurate and the process is correct. What we did was just go through an exercise of understanding- where are the manual hours that have been applied that could be better spent somewhere else? How do we think about ranking and stacking those hours? And then how do we get the appropriate teams together to think about what the solution might be? Because the other critical part is then starting to prioritize, what are you gonna do first, in order to get the most gains? but also what are you gonna do first that might not be as impactful until you get your head around how to do some of this stuff, which means that if you get it wrong the first time, it has a minimal impact. So those would be a couple of just follow-up points.

Speaker 16:11

Great points. I couldn't agree more. I wanna talk a little bit about integration here, because, at its core, whether it's AI or some other tool, integrating these into our day-to-day is an important part of the success of these tools, but oftentimes a major issue and a major roadblock. We see these systems, especially when sites and networks begin to grow, become more and more fragmented. So why is that integration so critical for networks that want to continue to scale while maintaining a degree of consistency and high quality?

Josh Stent 16:45

Yeah. I think about the biggest barriers to scaling networks, as you've said it, it's really the disconnected, fragmented technology stacks that create variability and risk. But also in the short term, create a significant amount of manual work that can create, confusion for sites that are coming together on an integrated, world. I think if you want to achieve speed and quality at scale, you've ultimately gotta reduce the number of control points where processes fail or where manual intervention is required. So the way... again, it's not too dissimilar to the point that we talked about just now, Jimmy. the way to do that, I think is to really map the entire value chain from start to finish, right? So if you think about where does a clinical trial start? It starts at, at feasibility, or it starts at a lead from a CRO or a sponsor direct to talk about a particular protocol. It moves through, feasibility through to site site selection to site initiation, and then it gets handed off to an operational team, where you go through the regulatory process of getting things start up. you- Set up a study template, you, create a source, before you start to brief a recruitment team potentially on how they're gonna start filling the study, right the way through to, some of the things that we've touched on, in terms of what you're doing at a site. And so I think by understanding, that journey for a participant and for the site, you can understand, where you as an organization, lack visibility or control at a particular step or lose sight of information. So one of the things that we identified was that the handoff from the, BD feasibility team to start up, we lost sight of a lot of that information, and therefore it might slow down a process. so being able to kinda create a system that talks to the sales function, which enables you to manage and maintain control of that, that visibility, it means that you can be faster and brief your teams more effectively, while also keeping your sponsors engaged. ultimately, if I think about kind of wrapping all of that, what it does is, it provides you with the ability to understand the things that you need to influence first, but also, manage your sponsors more effectively through that journey.

Speaker 18:46

Thanks, Josh. it's all about taking the time to map, like you said, map it all out, chart your course, and see where the systems are going to make their way into your operations, understanding how that's going to play out, and then what, really what the long game is, for all of that work. To, to taking a second and assessing the whole picture rather than a system-by-system solution type of process sounds like something that, anyone looking to venture into this, space would definitely need to do. But looking ahead, Josh, how do you see AI and these integrated systems, that concept we just talked about, working together to improve a lot of the key parts of the trial li- lifecycle for both sites and for sponsors?

Josh Stent 19:28

Yeah. I think ultimately, like we've talked about, the future of clinical trials, isn't doing things a little bit faster. It's fundamentally thinking about how we redesign the life cycle to, become more connected, continuously optimize and, get, get these products to market faster, for our sponsors and for the patients that are out there waiting for them. and when you combine some of the AI-driven work with, some seamless system integration, ultimately you're able to unlock, a true end-to-end view of feasibility, startup, and the continuous monitoring that happens, on a daily basis. and the way that you ultimately do that, which is our entire sector is built on, is, is data, right? And so when you think about the value chain, and being able to understand that value chain, being able to pull all of that data into a single place, that enables you to significantly reduce recruitment times, understand, who's likely to perform better on a study than otherwise, having visibility of, startup times for your sponsors, ultimately is the thing that's going to significantly change, the way that we bring products to market. And if you think about it from our customer's point of view, what are they looking at? And some interesting stuff recently about, what is a CRO looking to do? they're looking to, spend less time with CRAs out at sites monitoring. they're looking to, run their own tools over specific sites to understand who's recruiting and performing better, and therefore should be awarded specific studies. And at the highest level, what the sponsors are trying to do is free up their own capital, by spending less days trying to get product, to market, and therefore being able to reinvest that into some other great new therapies, that a patient out there is waiting for. if I think about some examples again of what we're doing here at Momentum to deliver that, we're sucking all of the data from the various systems into a single data lake, and then being able to query that data lake on a frequent basis to be able to pull reports that we can provide to our sponsors, or our customers, or being able to look at participant groups that might be appropriate for a particular study. so I think ultimately what this all leads to, Jimmy, for our sponsors, it's gonna radically accelerate timelines, and yield higher quality data. for our patients, it's gonna mean a smoother, supportive journey from screening all the way through to study completion. we're gonna for our teams and our s- our staff, we're gonna be able to spend more time in front of those patients, providing the care and educating them on their particular disease, while spending less time on all that, terrible administrative burden. So I think ultimately we're really redefining how drugs are gonna be brought to market.

Speaker 21:51

I couldn't agree more, Josh, and I think that's a really excellent place for us to end our conversation here today. the opportunities are there, and it's really exciting. you described a lot of, win-win-win scenarios here that we can unlock through the power of, tools that are based in artificial intelligence. And it's really exciting to see not only what's being done, but also, again, what is possible using some of these tools. So thank you for spending some time shedding some light onto not only what you're doing, but what you're seeing and what you're hearing in the industry, and where we're headed with these really great tools. So again, thank you for being here with us today.

Josh Stent 22:25

Awesome, Jimmy. Thanks. Appreciate the time and, yeah, look forward to catching up again soon.

Speaker 22:28

That sounds great. And for those that are listening, we look forward to continuing these discussions, highlighting other important pieces of information and shedding some light onto the future of artificial intelligence and its application into our industry. for now, thanks for listening, for tuning in, and until next time.