Ram Yalamanchili (00:03.1)

Hey there, how are you?

Deb (00:04.246)

Hey, I'm good. How are you doing?

Ram Yalamanchili (00:05.904)

Good, good. I'm happy to have you here and start off this conversation more from a site and operations perspective. So I'm excited about hearing that from you. So to begin, why don't we start with an introduction. I'd love to hear about your background and your operational experience more from a site perspective.

Deb (00:29.07)

Yeah, of course. So first happy to be here. Thanks for asking me to come on. I am a study nurse by training. So I have a nursing background and like probably most other coordinators just kind of fell into my lap. They didn't know much about research and I landed a job doing nighttime research nursing on a protocol. So safety and efficacy assessments for inpatient clinic and then PKs, lab draws, anything else that was needed. then the painful process of

Ram Yalamanchili (00:56.636)

Okay.

Deb (00:57.934)

copying everything into like a triplicate carbon copy, CRF, very old school. And that was my first foray into research. Loved it. So built my own company with a physician and then expanded that into multi-site, multi-state before leaving and going to a large network where I ran operations for them to help streamline and create efficiencies among multiple sites within the network before heading to Tilda to...

help fix the problems that I had seen in the 20 years in research and try to solve some of the issues. And then currently I am Senior VP of Operations at Monroe Biomedical that is also now in process of becoming their own network. So just kind of a full circle of what I've done.

Ram Yalamanchili (01:43.018)

Awesome. And I think what's unique about this particular podcast is previously I've mostly spoken to sponsors, industry leaders, as well as investigators who are

either principal investigators or sub-investigators running their own trials, right? I think this is the first times when I'm actually to get into the site operations and leaders who are talking about or have spent more time on the operational aspect of managing sites and research in both individual sites and networks. So that's exciting just coming in from that perspective. So to begin with, right, I think I'd love to hear a little bit more about the day-to-day of how you see site operations and

Deb (01:58.382)

Right.

Ram Yalamanchili (02:24.411)

from a vantage point of not an investigator, but more from like a site manager, site director. So how do you think about it and how you contrast it from that other experience?

Deb (02:34.766)

So at a site, regardless of whether it's embedded or freestanding, networked, they all have very similar concepts. It is a coordinator trying to keep as many balls in the air as possible and not miss anything. So it is a balance of patient care, which is their first priority. Also of making sure they meet all the regulatory requirements and then all of the paperwork that goes along with it and the redundancy of task. And then troubleshooting along the way.

And while it used to just be troubleshooting protocol, it's turned into new issues with troubleshooting technology, both at the site level and at the patient level that they have to take home and they have to work through. So it is a lot of juggling of efforts while also still being a caretaker and being a patient caretaker at that, because they're all in the medical field of some sort.

Ram Yalamanchili (03:05.691)

Yeah, and what would the vantage point look like? I guess what is the difference between how a

investigator looks at site operations versus say from a coordinator's perspective, right? Because I think there's different perspectives here, which they're both on the same mission, obviously, but clearly they're.

Deb (03:46.806)

Absolutely. Yeah, I think we have the same end goal, just like sponsors and CROs and sites, same end goal, but the process and what our journeys to get to that goal look very different. So I'm going to speak more in generality rather than every investigator and every coordinator, because all the sites function differently. But by and large, I would say the coordinator is the one or the site manager that is truly running the protocol at the site.

They absolutely need the investigator and the physician to weigh in on medical decisions. But as far as day-to-day operations, how patients are scheduled, the data entry that all the other tangential things from data entry to regulatory, that almost never falls to an investigator. The investigator is needed to complete reg docs. And that generally occurs with, if they're paper, a little sticky on the side and a coordinator literally directing the investigator's hand pen and paper to sign in the right place.

Ram Yalamanchili (04:24.409)

.

Deb (04:42.656)

If it's electronic, it is sending them and flagging stuff appropriately so that all they have to do is complete their checklist rather than do all the documents from scratch. The coordinator does that part. So they're running the day-to-day operations and telling the investigator, I need you here at 2 30 to see a patient for a physical exam for this. Whereas the coordinator knows that all like they're the conductor. They know what's going on in their entire clinic across multiple studies, multiple therapeutic indications.

Ram Yalamanchili (05:12.531)

Would you say a site can be successful without a great coordinator? I guess it's probably

impossible, right? From what you just said.

Deb (05:21.55)

From my perspective, it is impossible. I would rather have a great coordinator than a great investigator. I need a physician who is intelligent, smart, ethical, but they don't truly have to have much research working knowledge the coordinator does.

Ram Yalamanchili (05:37.742)

Yeah, and that's an interesting vantage point because even from a site feasibility and how the industry is set up right now, I think the weightage probably is not exactly how you describe it. That'd be fair from like a CRS sponsor perspective.

Deb (05:50.7)

I think it's fair. think the more CROs and sponsors you talk to though, I do think that there is an underlying knowledge that I need to have the physician sign this, but I need to talk to the coordinator to be able to get the physician to sign this or to talk to them about the logistics of the trial. So I think there's some underlying knowledge there, but yes, of the, like just the,

Ram Yalamanchili (06:09.389)

Yeah.

Deb (06:18.986)

the leadership of trials on paper goes to investigators.

Ram Yalamanchili (06:23.255)

Yeah, no, that makes sense. think, I guess from another point of view, I also see the critical resource required to run really good research is a coordinator, is a strong coordinator and a coordinating team at the site. And that kind of comes into the next point, which is what are some of the operational burdens or issues which you find on a day-to-day basis, especially when you're managing a large network of sites? you know, again,

talking more from like a perspective of the operational issues, right? Which the coordinators are most involved in.

Deb (06:55.352)

Right. So I think from an operational standpoint, know, one of the biggest concerns at the site level is attrition, because then it is constantly having to retrain and bring on new staff. And it takes a while to assimilate them into research, even if they're medically based and into specific protocols. So that is very time consuming for a site. having our trials are not getting easier. They're getting more complicated, which makes the training process even harder.

So from that perspective, I think having the consistency of somebody who is not overburdened, running ragged so that you can retain your staff, I think that that's super helpful. But I also think the amount of work that now goes into running a trial, maintaining the trial is overwhelming to coordinators at large. Like there is a redundancy and a replication of so many different areas.

Ram Yalamanchili (07:38.52)

you

Deb (07:54.124)

where we just have to repeat and copy and repeat. If we maintain one log, we maintain 10 logs. And if the site's using an electronic system, all of that data already exists. It's just different views of the same information with different log titles, which we need for regulatory compliance, but it's already there. So having to maintain that for every single trial is burdensome to the sites.

Ram Yalamanchili (08:20.786)

So since you've been in this field for quite some time, right, what's changed or would you say over the past many years, especially on the site side, have some kind of meaningful operational efficiencies come together or would you say there's still more to go?

Deb (08:38.49)

I think there's more to go. think there will always be more to go. I don't think we're ever going to arrive. We can always make things better. I do think that the sites have changed a bit with the entrance of networks emerging and them all forming together and being bought out. They're no longer, I mean, a lot of the sites are no longer physician owned. They're, you know, PE owned. And so I think that changes the way that a site is managed, but from a network standpoint.

I also think having a network with multiple locations, trying to create the efficiencies when you have multiple sites so that we aren't replicating the exact same process. For instance, if you have a network and you have the same study running at five different locations, you already have cut down on probably a solid 40 hour work week to create source for every single individual site. That is probably one of the most cumbersome tasks at a site level is source creation and editing and.

Ram Yalamanchili (09:35.16)

And that sort of brings me to an interesting point, which is we are seeing consolidation in the space. And usually that's a bet on bringing efficiency in a consolidated manner, right?

Deb (09:36.942)

doing track changes with version control. So I think having a system in place that does have a decrease in the replication to increase the efficiency is super helpful.

Ram Yalamanchili (10:04.811)

you bring up source creation as one of them. But going back to the original discussion on the CRC or the coordinator being a big bottleneck, how does that tie in in this new model? Is there opportunities to address the problems you just mentioned, especially on the coordinator side by being a network of sites?

Deb (10:25.71)

So I think sites have two different approaches right now to how they view coordinators. One is the coordinator that wears every single hat from feasibility to study startup to helping with invoicing, patient care, recruitment, things like that. They truly run a study end to end with little to no help. The other option is

a much more siloed approach to tasks and duties where you have a coordinator that may see patients, but you have another team that does all of your data entry. You have another team that does all of your recruitment, another team that does your regulatory. So I think when you do that, there's pros and cons, of course, to both. But I think that there are efficiencies that are created because a network is doing something one time and relieving all of these other.

tangential things that have to occur for research, but are not things that have to be done inside of the clinic.

Ram Yalamanchili (11:19.986)

So given that perspective, do you think the future is networks or independent sites? Where would majority of research get down there?

Deb (11:28.918)

That is a tricky question. I think networks that are at a certain level and potentially still have the same end goal, which is the patient care and the development of new treatments and technologies. I think there's a lot of benefit to that. The concern is as the networks grow that their end goal is very different than the original physician-owned site, right?

I think when things get too large, they can get complicated and messy and things can fall through the cracks. I think there's a balance between both. I think there will always be a need for both. We are seeing a consolidation. I'd be curious in five to 10 years if we see the opposite trend. I'm not sure, but I think that some of the centralization of function can occur without a network in place, honestly.

Like whether it's with a consulting group or something else that does the centralized functionality for multiple sites, even if they're not owning them, I think that there's other options there.

Ram Yalamanchili (12:36.592)

Yeah, and I think that's an interesting point, right? We've seen these trends where you have consolidation because you want efficiency and get to these consolidated plays. I also think you have better opportunity to bring capital for innovation. So, you you're able to invest in better technology, better process, better oversight, things like that, which also then points to what are some of the interesting innovation which we can then bring in into the research.

Deb (12:56.302)

Absolutely.

Ram Yalamanchili (13:06.378)

which may not be available at a site level. At an individual site level, it may not be feasible to invest and sort of grow such capabilities. But at a network level, maybe it's a little easier because of the efficiencies and frankly, the capital and the people and the focus. So what do you think are some of those central functions which could potentially be disrupted or helped by AI or...

the rise of like this new function, right? New capability, which is coming up right now.

Deb (13:39.404)

Yeah, so I think that there's a handful of items that don't take critical thinking skills, that doesn't take a medical decision, that are involved in research, that are repetitive, redundant, that could be taken over by AI. think regulatory is one of those items. It is very clear that at the beginning of a study, the same handful of documents have to be completed in the same manner, checked for completion and accuracy and filed in a certain way.

That doesn't need to be done by a human anymore. I think that that is very simple. There's documents that come out throughout a study from an IRB. A new protocol comes out and is amended. It gets submitted to the IRB. It gets approved. The approval itself has a cascading effect of, you know, a handful of steps that have to occur that are the same every time. There's no reason AI can't take that over. Send a PSP to an investigator, sign, get a training law.

Ram Yalamanchili (14:29.205)

.

Deb (14:36.504)

for the entire team on the DOA to sign off on it and have training on what has changed, get the source updated based upon any trade. So series of steps, that doesn't change. So any of those places where there is no change, it is the exact same process. I think AI can help alleviate the burden. So that's the regulatory side of it. I think finance is another huge place where AI could come in.

Ram Yalamanchili (14:58.101)

So,

Deb (15:01.952)

and really be disruptive to think about a large network that has a team of 15 people on their finance team. I think that can be drastically decreased with AI. That is just input going into a computer system and then putting it into a formulated invoice, something to send off to sponsor, something that sponsors can pay even without. mean, why make sites invoice and do pass through expenses for things that we know that are going to occur? And the sites

Ram Yalamanchili (15:27.861)

That's an interesting perspective because

Deb (15:31.222)

are so overworked, we leave money on the table. It's not intentional, but we're not even getting compensated for work being done. I do think that there's ways that AI could move into that space. And actually, the reconciliation process would be so much simpler from a site and sponsor and CRO perspective for all three.

Ram Yalamanchili (15:55.27)

Ideally, if you can streamline this at the site level, that translates into streamlining on the sponsor level as well, right? Because there's the whole monitoring piece which comes in. And if you can, in an auditable way or a traceable way, say this is what has gone together or gone through to get this work done. so I guess what I'm getting at is, it's an interesting paradigm which we're talking about where if you have an AI

do some of these functions and it's also able to give you real clear reasoning on why it has done this particular job. I think the world of monitoring also changes because the way you monitor perhaps changes in this kind of world, isn't it?

Deb (16:38.286)

Absolutely. I mean, I think a monitor by and large in our current way of doing work. And again, I think we have an antiquated system that was set up based upon a system that doesn't exist anymore. And I can explain that real fast. Just if you think about

Ram Yalamanchili (16:52.275)

I'm very curious on what you mean by that because it's insightful.

Deb (16:55.276)

Yeah, think that so we all started on paper, right? Before the Internet was a thing, research was going on. Every clinic was siloed in their own physical space, documenting all of the data, you know, pen to paper. And then to get that data to the statisticians and the data managers and the safety reviewers, the only way to do that without the Internet and all these electronic systems and platforms was this is where the triplicate carbon copy came into is a little pink, yellow and white.

and somebody sat up and hand copied all of the data points that the sponsor wanted or the statistician thought was important. And then a monitor came and checked. They weren't checking for protocol compliance. They were checking for transcription errors, which occurs because we're human. And it doesn't matter whether the site reviewed it five times, the monitor came and reviewed it. There's still going to be an error in that data. So there's always this human error possibility. So then,

Ram Yalamanchili (17:32.73)

Okay.

Deb (17:51.192)

They tore off two pages and it got sent out in two separate shipments because if something happened to one of them, it would destroy both. So they had to go in two separate packages. And then they crunched data on the backend and then would send us manual by a fax machine, a query. And we had to go make changes on the carbon copy and make a change on this paper to send back to them. Like we started that process and that was our case report form. And then we moved into an electronic case report form, which is what we know as the EDC today.

And it was just a faster way to get data. Then sites adopted e-source were already entering it into an electronic system. And now you want me to reenter it into another system, just get the data out of the system. Like there has to be a way for AI to move in or I'm saying antiquated because I'm wondering what the point of an EDC system is now. I know why it existed, but I don't know why the CRF has to exist in the same capacity today. So I think we're ready for changes in our current industry.

Ram Yalamanchili (18:21.616)

Okay.

Deb (18:50.082)

But right now, while there still is this dual system of source and CRF, I think AI can come in and help with doing that. Imagine if a monitor could come to a site and monitor for protocol compliance rather than transcription error. How much better our data would be?

Ram Yalamanchili (19:07.922)

Mm-hmm. That's right. And what do you think is the percentage of time spent on things which you probably think are like, you know,

quote unquote antiquated, right, in that bucket. Because you bring up an excellent point. There are a class of queries and problems which we see being raised by monitors, and those entire class of problems are probably gonna go away because they will not exist. This is something an AI can do in a very high accuracy way, Transcription being one of them, but there's other things like correlation checks and cross-document verification. So we have certain class of problems where I can see that happening.

But again, I'm curious, we're not gonna completely change the role of a monitor. think monitoring is still important. There will still be a big role for them. But I'm kind of curious, what percentage of these type of activities are currently the effort put in from a monitoring perspective?

Deb (20:08.62)

I mean, I would say well over 50 % of a monitor's activity is checking for transcription errors and making sure that the ICF log matches a screening and enrollment log, which matches the screen fail list, which matches. So all of these things are compared against each other to make sure that they're all matching. It's all the same data. When we use an electronic system, it could be one data point that shows up in six different places. And there is no transcription error then.

Ram Yalamanchili (20:33.842)

Okay.

Deb (20:37.598)

I think that it's not just going to create all this additional time for a monitor to review for protocol compliance, which is, I would say, one of the most important aspects of their job. I think it's also just gonna decrease, monitors will still be important, we still need them in clinical research, but I think it will decrease the amount of time. A lot of this can be done at home, it's not gonna be a plane trip.

to your site every single week to double check data on site. There's clearly things on site that will always be there. Drug is on site. They have to check that. Like there is an element of on-site presence that is needed, but not to the extent that we have it.

Ram Yalamanchili (21:19.131)

Another question I have is, who do you think is going to be the first mover in adopting AI for any of these functions we just spoke about, right? Would it be the sites or would it be the sponsors? And I ask that because right now we always have this middle layer of a CRO. you know, obviously incentives have to align. And I think when I hear you say how this all works, obviously the site is incentivized to reduce the burden on their coordinators.

sponsors incentivize to reduce the amount of time and effort and money spent on some of these activities also. So from that perspective, have you got a thesis on where are we gonna see proliferation first or more obviously?

Deb (22:03.598)

So it's an interesting question. I haven't spent much time thinking about who is going to be more of your earlier adopter of this technology. think, and I don't say this lightly, I think the sites have less to lose from a financial standpoint to adopt. And only because we're not paying millions of dollars to run a trial and a sponsor is forking over money. And if this doesn't work, if something breaks in the system,

Like they're on the hook to repeat it. And the site isn't, we're out a piece of our pie, but not the entire pie, if that makes sense. That being said, sites generally can't afford all the new technology. Like you mentioned before, like we don't have the infrastructure and the finances to get that buy-in at that level. I think sponsors and CROs do, but I think it's going to start in smaller companies first at a pharma and CRO level rather than the large ones.

Ram Yalamanchili (23:01.504)

Yeah, and I think you bring up an interesting point. It's not so much only on value, it's also risk.

You have this lens of risk which plays probably a bigger role in terms of adoption. And clearly that sponsor has more to risk in this case than potentially a site, from what you're saying. Interesting. So we didn't talk about a few other things in this conversation from an AI perspective. I'm going back to some of the things you just mentioned around regulatory, for example.

I see the back and forth every time a study is getting started up or a site is being started up, ongoing maintenance, close out, there's tremendous amounts of documentation being shipped back and forth. And I think almost everything happens through email right now. And there's multiple systems involved. We constantly hear about multiple login problem, multi-system multi-login problem within the space. So if there is a potential to...

use AI to reduce the burden there and also sort of think of it as as simple as joining any network, right? Like joining LinkedIn or Facebook or whatever. It's a fairly straightforward process to kind of come in and start getting value out of these type of networks. And I almost imagine, is there a world where you can imagine a proliferation of number of sites who can participate in research because the burden's now reduced?

You're not entirely reliant on that one coordinator to do all these functions. You have these AIs or, know, even centralized consulting companies. I think you mentioned several other options where the site doesn't need to these functions. And at the same time, they get a higher quality output because they're reliant on technology rather than a, you know, a variable quality of a person, for example. Does that mean that we're entering a world where there's probably like...

Ram Yalamanchili (24:55.907)

more opportunity for research in more places or is there other gates which are in place which maybe, you it's not entirely the full picture.

Deb (25:04.814)

I do think that the potential exists when you start to alleviate site burden, that smaller clinics and embedded models with physicians who want to do research, but it is such an endeavor to get started. I think it opens the door for them, which is kind of a fascinating place to be because they are the practices that have the patients. So another huge issue, which

is not something that I think AI is gonna tackle right away, is patient recruitment and finding the patients for these studies. It's easy to enroll, you know, GLP-1 studies right now, of course, but to enroll something that is in a smaller therapeutic indication field that is more of an outlier or a rare disease, like those are hard and you have to go find the doctors that are treating those patients. And if you could give them the means to run clinical trials out of their office with the assistance of AI and the centralization of functionality.

I think that would be huge for clinical research.

Ram Yalamanchili (26:03.117)

That makes sense. And that brings, I think, different set of value where traditional CROs have played, right? I think if you are proliferating across more sites, maybe monitoring just takes a different approach now. It's not, like you said, there's a 50 %...

overhead and all the workflow that potentially going through. But now maybe you have to shift towards monitoring in different ways across more sites than fewer sites. potentially the costs all kind of like equate out. You're not really like losing anything in the business. It's more like you're expanding the opportunity while bringing into the model. Interesting. So I want to close on another topic. I do hear quite a bit in terms of

know, sites missing things and that leads to cascading set of problems, which then you would have to alleviate. And so in some ways, the industry currently tries to avoid as much as they can any obvious issues which arise. But if you do end up getting into some kind of a missed data or errors data or some kind of a regulatory issue, there's probably a cascading effect of work which needs to happen to kind of get back into compliance, right?

Can we talk about that second part? Because I understand the aspect of trying to avoid as many issues, but from a risk lens, you still have to think about what happens if we hit an issue and how do we rectify it? And what is the operational overhead involved from that perspective for a site?

Deb (27:39.64)

Significant. So just kind of restate that, even though we are trying to adopt and implement technology and platforms that could help ease the burden and try to decrease the error margin, we still need to be ready and prepared to handle when errors do occur. Is that correct? And so what does that look like at the site level, the burden for that component? Yes. mean, so like logistical standpoint,

Ram Yalamanchili (28:02.391)

Exactly.

Deb (28:06.574)

First of all, that could mean the halting of your study activity. So all of your revenue potentially could stop because a lot of sponsors are gonna say, hey, stop enrollment until you fix X, Y, and Z. So first of all, you've already just shut down your revenue stream. You've also taken away the potential for patients that were planning on coming in. So it's not just finances, it's also patient care that's been affected. You have to figure out how to remedy the mistakes that have already been made. And then you have to come up with plans to put into place that meet the approval of the IR.

be the sponsor, your CAPAs, so that you can figure out how to prevent, like, why did this happen, and how do I prevent it from occurring again? And then once that has occurred and you get all of your paperwork documentation retraining of your team in, then you can potentially have your site reopened. So there's long-term effects from that, and then there's very short-term effects. But I do think the overhead is extensive every time something like that happens. So of course, we all want to do our best to.

decrease and mitigate so we're not in that position.

Ram Yalamanchili (29:08.772)

Right. And I think in that lens, that's

an interesting situation because there's very few businesses I can think of where a complete halt is like one of the outcomes of, you know, maybe a coordinator not performing their responsibility accurately. And that sort of gives you a binary outcome in terms of what your business looks like into the going future, right? So that, I guess going back to the risk lens,

It's not that there's no risk on the site. This is an interesting risk which site networks especially probably will have to tackle with because you have more scale, but at same point, you're also talking about sort of binary events which can affect the entire business in terms of keep going or not, right? I see. All right. Well, no, this is great. Thanks for spending the time, Deb. I really enjoyed it. I we covered a lot of interesting topics today. But yeah, thank you for your time.

Deb (30:03.509)

Thank

Ram Yalamanchili (30:09.785)

and

Deb (30:10.318)

Appreciate it. Thank you so much for having me.

Ram Yalamanchili (30:12.16)

Yeah, thank you.