IQVIA Holdings Inc. (IQV) Earnings Call Transcript & Summary

Hi everyone, and welcome to today's live broadcast, the digital transformation of RBQM. My name is Andy Studna, Editor of Applied Clinical Trials, and I will be your moderator for today's event. We are pleased to bring you this webcast presented by Applied Clinical Trials and sponsored by IQVIA. I would now like to take a moment to share a statement from our sponsor. IQVIA RDS is your flexible clinical development partner, leveraging the power of IQVIA connected intelligence across your entire study to deliver life-changing therapies faster. Visit us at www.iqvia.com/rds to learn more. And I would now like to share a few important announcements before we begin. This webcast is designed to be interactive, and we encourage you to ask questions during the event. You could submit questions by typing them in the Q&A box, which can be found at the bottom of the video player. You can enlarge the slide window by clicking on the small icon in the bottom right corner of the media player. The slides will advance automatically during the event. And finally, if you have any technical problems viewing or hearing this presentation, please click on the question mark help widget in the top right of your presentation window. I would now like to introduce our speakers for today. We are pleased to be joined today by Jon Hill and Debjit Chakrabarti. Jon is Associate Director of Centralized Monitoring and Debjit is a Senior Director, Medical Data review, both are with IQVIA. So with that, thank you for joining us today. And I will now hand it over to Debjit to get us started.

Thank you, Andy. We will move with the next slide. So in today's session, we kind of aim to cover the impact of digital transformation is having on RBQM methodology, utilizing novel clinical technologies like mobility platforms and AI-enabled technologies, patient-generated data and the convergence with delivery models like decentralized trials. We'll briefly touch base on the general overview of RBQM framework, the evolving digital landscape of RBQM, how the AI and the ML models are transforming and how we operate in that space and briefly touch base on the generative AI and its potential impact for RBQM. And as all of us are aware that RBQM has evolved significantly in the last few years with increasing adoption of centralization, remote aspects and AI-driven analytics in the clinical monitoring space. So we'll move to the first slide. So -- this is, again, a kind of an overarching umbrella for RBQM framework. RBQM is kind of a collaborative approach towards end-to-end clinical trial oversight by focusing on things that matters to the steady outcome and the patient safety. And today, I think we are focusing on the shift from this RBM to RBQM throughout the trial life cycle, starting from protocol design by applying the quality by design principles. As a part of evidence-driven design at IQVIA, we help sponsors to lower the risk for their clinical trials by evaluating the protocol quality prior to the study startup with data-informed protocol assessment, design analytics and leveraging competitive intelligence. It's all about ensuring that we collect the right data with the right data sources and at the right time, for the entire monitoring of the trials. This entire power design setup initiates with identification of critical to quality factors in the protocol design, effective protocol risk assessment and ongoing management of risks using quality tolerance limits. And again, the regulatory and the GCP control forms the basis of the fundamental pillar for all this ongoing review and the assessments. So at IQVIA's RBQM framework utilizes domain expertise, strong processes along with tools and technology to effectively drive these strategies. Our data surveillance is near real time and based on study-specific algorithms, we can shift our focus more from data cleaning to data reliability aspect. And if we are collecting data centrally through systems and you don't have to do as much of source data review. And at the maintenance and the conduct phase of the trial, the centralized roles will continuously performing the ongoing data reviews, track the key risk indicators and detect early signals, thereby allowing us to see the high-risk sites at an early stage. And finally, the discipline around inspection readiness to eTMF is an important parameter, ensuring the documents are up to date. So with this context, we'll move to the next slide, which basically outlines the 3 key pillars of our RBQM strategies, risk assessment, data collection and data flow. And again, a strong cross-functional alignment upfront always leads to a more streamlined delivery downstream. So it's very important that everyone across the multiple functions know what data they're collecting, what are the scope and the components of the review and how the monitoring across the data is done. So all these help in developing these strategies effectively. So establishing the right data strategies and processes during the protocol design phase evaluating the vendor capabilities helps in maximizing the operational efficiencies and also in driving continuous data flow for a seamless review. Having all the systems that are central, where the data goes directly or they can be accessed directly saves a lot of time and effort. Hence, this entire shift in data collection from EDC to all the eSource options like connected devices, supports in direct data capture and thereby improving the overall data quality and reducing the site burden. So this entire framework also enables to see the risk around the data in a more holistic and comprehensive manner to make a faster decision and thereby, a connected downstream process always facilitates operational impact on timely database locks and the submission activities. So the next -- we'll move to the next slide. And this is, again, a bit of illustration around the key benefits of our overall RBQM strategy and some of them are reducing site burden due to the ease of this data reconciliation, centralization, et cetera, improvement in study start-up time lines, data-driven automations through multiple innovative solutions and thereby providing insightful data that ultimately optimizes the on-site monitoring and the CRA time. Transforming this data into insights with real holistic view of the trial ultimately reduces the overall trial risk and eventually consistent higher data quality and integrity facilitating enhanced analysis and better decision-making. So overall, the entire RBQM framework enables our sponsors to get their products to patients faster by focusing on the data that are more critical. So this slide is more about this entire shift that we are seeing in this digital shift emerging and also kind of endorsed by the regulatory responses in terms of risk-based approaches being adopted. So FDA guidance specifically encourages greater uses of centralized monitoring methods where appropriate, whereas EMA kind of explicitly or exclusively stated that there is over reliance in the industry in terms of retrospective document checking, et cetera, but though they are important, but they're not sufficient to ensure quality of the clinical study. And the revision ICH E6 R2 and the draft revision R3 stresses on risk strategies and quality management process for all trials with technology being fit for purpose to allow accurate reporting, interpretation and verification of the trial information. And the revision 3 in the draft phase also focuses on the data governance aspect and the oversight by the sponsor on appropriate management of the data integrity, traceability element and the overall security. So with this context about the general overview of the RBQM framework, the key elements and the regulatory landscape we'll briefly touch base on the next aspect of the presentation, which focused more on the evolving delivery models. So as we see this entire evolution of the role of the clinical monitor or CRA, this basically is coming from a result of the shift in the industry has accepted around RBQM, that helps to mine the data better. Some of the key areas that are driving this change include around innovation and technology. Data access enables analytics to allow more directed monitoring efforts as well as greater automation of administrative tasks. Remote access is also enabling a greater centralization while maintaining the data quality and it also adds additional advantage of driving greater sustainability within the industry. Targeting the site needs means monitors can remain flexible to remote or on-site, thereby reducing the site burden. And with this high demand and limited monitoring resources, this entire shift is solving the core key challenges. So with this entire evolving role of the CRAs and the refocusing of the monitoring efforts on critical process review, source data review, the CRAs will be able to focus their efforts on higher value areas just namely the recruitment aspect, areas of the risk identified through the entire centralized and the remote data review processes and ensuring the site quality through the source data review or the critical process review. So this is, again, just kind of an illustration about how the integrated monitoring model, maintaining the site engagement. So with this refinement of this entire traditional risk-based model to an integrated model, we are balancing to drive this site engagement without increasing the site burden and enabling better oversight and benefits. By merging data, tech and analytics not only drives the monitoring efficiencies, but it also allows the sponsor oversight of the trials, greater insights from the data collected, seamless process flow and overall risk management. It improves the overall speed of the clinical development cycle. And this entire integrated monitoring process includes all the centralized monitoring elements combined with the monitoring task and all of that plugged with analytics and technology, along with sponsor oversight, thereby providing efficient risk management, site and patient compliance along with data quality. So this slide -- we'll move to the next one. So this is, again, IQVIA's model, which is, again, scaled across multiple countries across all major operational time zones, leveraging digitization, mobility platforms, automation and AI/ML to maintain the site relationship management. So we know that CRAs carry higher site loads, and we really want to free our CRAs from many routine tasks and give them back time to support enrollment, manage the overall quality and kind of build the site relationships. As a part of this entire framework CRAs partner with centralized monitors sitting in the local time zone with local capabilities to ensure the study continuity. So overall, this allowed us to dial down the volume of on-site time and travel, supporting efficiency and the sustainability aspect, thereby, utilizing remote capabilities in monitoring and site management. And we have been receiving a lot of positive site feedback on this engagement model. So these are some learnings and stats based on our experience, but just to say that we have been able to recruit close to around 37% more subjects per site along with reducing some of the other key alerts like missing data query, query aging through these approaches. So we we'll move to the next one. So this is, again, the big move towards 0% SDV may sound like a big step. But if you look at SDV and SDR, essentially defined as separate activities, where SDR has much of the monitoring value. We all know that RBM trials already operate at an optimal SDV percentage and which is acceptable and the expansion of this or the integration of this AMR to EDC or sources remove the need for SDV. So whenever it comes to SDR, it has high monitoring value and with a strong impact, the review becomes much more crucial to ensure all the procedures are completed per protocol and applicable regulatory requirements by the CRAs in adherence to the guidance. So this is the next emergence of the digital area, I'll briefly touch base on that. And then transition it to Jon for the -- some of the digital landscape space. But as we know that pandemic kind of acted as a catalyst for some of these digital capabilities and driving this traditional model as we can see on the left, to current by digital capabilities. So impact to patients, sites and sponsors created the need for much greater insight to trial oversight, through trial digitization. This all starts from protocol simplification to adopting novel trial design, remote site monitoring, digital mode of data collection and the patient engagement through hybridization of trials. For example, patients may not need to come for all visits to site as the data can be collected centrally through decentralized approach. So overall, with the digitization and the risk-based quality management approach, helps in reducing the trial risk, improving the data quality, integrity and patient safety through early signal detection capabilities and with the application of advanced analytics. So thank you all. This concludes my section. Jon, I'll transition to you for the remainder of the digital tech space.

All right. Thanks, Debjit. And so as Debjit just alluded to, I will walk us through a little bit more specific around how we have started to evolve within the space looking at digital technologies and where that's really taking RBQM, and some of the views we have around this. But before we really kick off into those, let's set the stage a little bit more around where the landscape has been evolving. Our industry is overall continually evolving, which isn't a big surprise, right, probably with most industries, but what factors are currently at play. Some of them are old. Some of them are new, but let's break them down a little bit just to kind of be on an equal playing field here. So as we look at the evolution towards the digital landscape, this has been driven by several key areas. One most prominently is the growing use of AI and ML models, which are being leveraged to support things such as streamlined decision-making and processes. And if you think that just greater than -- well, sorry, according to some industry reports. I'm not going to say across the board, the greater than 75% of general consumers, and it might be higher than this are actually interacting with some form of AI-driven technology in their day-to-day life, whether that be their cell phones or within e-commerce platforms. And on top of that, there's been a rapid expansion of generative AI. It's almost impossible for anyone to be interacting in the marketplace and not have heard about this or have been impacted in some way. So there's definitely a lot of change within that space specifically. But also looking a little bit more narrowly at risk-based quality management as this is becoming more of the industry standard and with recent regulations really pushing towards this model overall, specifically after the post-pandemic acceptance of remote monitoring and approaches similar to with remote capability. A lot of that is definitely pushing the envelope when it comes to this evolution. Additionally, looking at virtual and decentralized trials, home nursing, the emergence of wearables, all these things have helped accelerate over the past 5 years, probably longer if you think about some of the technology that's been in the space. For example, we're probably all familiar with Telehealth visits of some sort, whether you're a patient in a trial or not, it's becoming more and more of a general approach when it comes to having rapid access to healthcare. And then there's also the continued expansion of technologies, more at the site level, if you think looking at eSource, eConsent, eISF, EMR and more. So those are just some of the factors that we're seeing that are really driving this push, and it's being seen across different areas within the industry, not just centered around any 1 specific group and so as we look at IQVIA specifically, how we're looking to evolve to these external market factors because, obviously, there are things that any organization needs to be doing to stay ahead of and to be leading within the space. IQVIA specifically, we're looking at the evolution towards more of a digital clinical operating model. And so that's a journey, it's not new to us. It's something that's been long in the making, but I'll walk you through a couple of the foundational evolutions that have brought us to where we are now and some of the thinking around what this really means to be in the digital operating model. If you follow from the left to the right, so we'll start there first looking at where we were over a decade ago looking to define what it meant to deliver risk-based strategies. So really the onset of RBQM helping shape the industry thinking around and the overall adoption that we've seen now with RBQM. And this has had the benefit of allowing us to focus on analytics and automation that help to transform that traditional view of monitoring and really transition that approach, we're building greater efficiency, scalability, all while maintaining high quality. And if you move to the middle section here, one of the areas when it comes to moving more towards digital strategies that has resulted in several challenges is really how you start to converge that RBQM, decentralized trial thinking, there obviously needs to be more focus on greater integrations. There needs to be data flow set up. There needs to be the monitoring capability within the DCT framework and technologies that are available all without overwhelming the sites to be able to implement. And this really to successfully drive forward in this space, requires reskilling the organization to think in new ways of working, looking at driving greater patient centricity, looking at opening opportunities for patients and sites to become increasingly managed remotely, which is not a small feat and definitely one that requires a lot of change management around. And then looking forward to the future where we anticipate the industry heading, we're looking at more of a fully digital trial. Just think you have your direct-to-patient focus, decision automation, remote patient management all being ushered in with a new age of monitoring. So definitely a lot on the horizon that we're looking at. So as we move on to some of the practical applications of how we're actually driving this, we'll look at the evolution of our AI/ML or advanced analytics models, so some of the practical ways of integrating digitization and AI/ML models, specifically within centralized monitoring applications. I'll just kind of walk you through this a little bit to set an even stage. At the core, the fully integrated data that is coming in through these trials, we have such a wealth of data where if you look at decades previously in this industry, we just didn't have the ability to tap into the wealth of data that was being generated. That enables automation and enhances insights that we can provide through advanced analytics. It also drives numerous downstream systems, including our risk-based centralized monitoring. And so in the central monitoring world today, AI/ML and artificial intelligence specifically can benefit the risk alerting process because it helps reduce some of those manual review processes while also helping enhance the accuracy and efficiency and also the consistency with which we're delivering that. However, with that growing amount of data that's collected through clinical trials. That does lead to several new challenges that didn't previously exist. How do you make sense of this vast amount of data that's coming in that's flowing in, how do you make insightful use of and not get bogged down and logged into that data with decision paralysis. And so solving this challenge can be done by applying these analytics and AI/ML technologies or techniques to help with ingesting, monitoring the data and identifying potential issues and trends that are embedded within the data. So really opening up a lot of opportunity if you think about it when the right approaches are applied. And so some examples of how machine learning is being leveraged would include defining the most appropriate rules to identify similar records automatically. So if you think forming the foundation of fraud detection so that's one of those areas that has historically been a little challenging. How do you identify potential career patients, if you will. And so through that identification of data and being able to apply some of these machine learning algorithms, we're getting to the point where we can be more savvy with how we're actually looking at some of the industry problems that have plagued us for quite some time. Or if you think of applying dissimilarity measures to help with flagging outliers as another example. So definitely a lot that we're doing within the space. And I think it might be helpful if we jump into a specific case study, I can walk you through a little bit of how we're actually applying some of these artificial intelligence to support central monitoring. So within this specific case study, we'll go ahead and dive in, and I'll demonstrate a little bit about how the application of AI has helped enhance the accuracy and efficiency overall for our central monitoring team. So the situation, if you want to think about this one specifically is looking at the alerts for site risks. So our KRIs that are being fired and how we are managing those. The current state or previous state, if you want to think about this is really quite manual. Even though you have these KRIs that are being flagged, it still requires manual review, human decision-making to define what the appropriate actions would be to those specific risks that are identified. And so this process overall requires a significant number of man hours to be able to accomplish. And so we really looked at applying AI/ML solutions to try to see what can be done in that space to reduce some of that manual effort, but still drive the same level of quality and consistency. And so what we did is by leveraging historic data assets from our RBQM trials, we built a machine learning model that learned from the past trends and decision-making that was conducted by the team so actual decisions that are being generated in that human decision-making loop. And then applied the artificial intelligence algorithm to help the decisions -- help flag decisions for appropriate action to mitigate those risks. And so by firing from learning with historical examples, firing machine learning risk mitigation to help with that decision-making process. Now not replacing it, but to help supplement and get to those decisions quicker or provide potential options. So you get a little bit more of a quality and standardization within that process. Overall impact of what we're looking at, the model showed over 90% accuracy, which is really impressive when it comes to mimicking the decisions that were otherwise made by our central monitoring experts. And if you look at overall efficiency, the process definitely helped improve efficiency across the board about 75%, which is pretty tremendous when you think albeit 1 specific area within the central monitoring and RBQM process. It shows and demonstrates how this starts to open the door for other more innovative ways of thinking. And so as we look at the next layer of how data can be interpreted through -- it would be through signal detection is another example that I'll call out. And at the root of this is really safety, which, as we know, is paramount to any clinical study. And so when it comes to this, the early assessment of any potential safety risks is what's in the best interest of patients and really is what should be top of mind. And so by powering our RBQM processes with AI/ML in real time, it's possible to quickly gain visibility that's needed into numerous trends. You can kind of see a theme here, vast data, how do we start getting trends and insights. And so for example, identifying adverse events faster, right? So these are all things that we've looked at when it comes to AI/ML models. And so this helps to identify potential safety issues earlier while also maintaining high data quality. And then the integrity that's needed to avoid any end of study delays as well. And so we're able to do this through multiple capabilities that have been used to help with some of the monitoring trends. And if you follow through just on this bottom section here of what I'm showing you, the basic analytics so if you think more of automation tools, that's been around for quite some time. We're definitely with the wide adoption of RBQM, able to really tap more and more into these types of analytics. And then going a step further, looking at advanced and predictive analytics, so where you can start getting a little bit more fine-tuned with how you're looking at example for site health and historic trends, which for us enables more of a site tiering approach where you can say based on these advanced analytic findings, which sites are showing a little higher risk for example, that might need a little more follow-up, we're actually deploying our monitoring team so that we can help support those sites. And if you think a little bit more complex. So you start moving towards cognitive automation. And so the AI decision support. And an example of how we've been building that out would be within the subject level data review process, introducing robotic process automation to help support within this space. And so as we look at in addition to the AI/ML, where are we going kind of in the next frontier. Well, I alluded to this earlier, but it will be no surprise probably the most. Generative AI is really that next frontier that is emerging. And looking at the impact of artificial intelligence, it's been quite a long history, and there's been a lot of advancements that have been made. This isn't new for us to be talking about. It might be newer for the rapid rate with which the industry is starting to look at and adopt technologies. However, nothing over the course of the development of some of these AI and ML approaches over previous years has really captured the imagination quite like generative AI or large language models. We've heard a lot within the industry about this. And the shift in the AI overall has really been a result of 3 fundamental changes the way that I've broken it down. So as we've talked about looking at that wealth of digital data that is available, it back ends a question of what can be done, what additionally can we glean from this vast amount of data to help support the end goals of delivering medicine and treatments to our patients quicker. And so that's probably one of the biggest impetus. And then if you look at in addition, the processing, compute power that really the expanded ability to carry out more complex data processes and activities through the continued development and expansion within the space, that's really set the foundation for that emergence of generative AI. And so that, in addition to advances just in general with AI algorithms with our data scientists and teams that have really been focusing on these areas, all that's leading to this emergence. And so as we look at how this is impacting the industry, we can really conceptualize this into several pros and cons. As with any new technology or any new processes or approach, you want to really tease out where is the benefit and the impact versus the trade-off some of the cons. And so that's one thing that is really imperative here is we're looking at the use of generative AI, because on one hand, this opens us up to solutions like human-like interactions and interfaces, which is great. you have the ability to continually learn on these models as well. So you're feeding new information continually, which is really impactful. However, on the other hand, you open up to bias and hallucinations potentially. So there has to be the right approaches that are applied to developing these generative AI models. There's also a little bit more variability and unknown around the cost structures. Right now, its scalability, it can be rather cost inhibitive to apply generative AI models in all cases. And then also the elephant in the room, the current regulatory environment, it's continually evolving. And as regulators are looking to catch up and really carve out what the use of generative AI means and how that will be governed, there's a little bit of an uncertainty within the industry to some degree. But as we look at potential areas that are ripe for development when it comes to the emergence of generative AI. There are several within the RBQM space. And I'll call out just a few of these. I won't hit on each of them. But several examples that we can look at would be within the safety data reviews that are conducted for patients where we're capturing potential risks early on and even preventing them from occurring and by having these human-like models that are helping to identify these areas and with subtlety, bring that up to the teams for action. It's a huge benefit. Also looking at driving inspection readiness throughout every stage of the trial, within this kind of thinking more around the standardizing of document management practices. So how do you quickly resolve queries, with limited manual effort as you think of reconciliation of eISF and TMF, a very -- an ongoing task over many years, but also very manual task in human-driven task historically. So it opens up to potential human error and really looking at ways that you can integrate generative AI and more savvy implementation, it opens up a great opportunity for ensuring consistency and quality. And then that feeds in nicely to data quality as well. So identification of data variability and inconsistencies at time of event is really critical. And so being able to do that with much more precision and then also preventing the downstream impact if you were to have any data quality issues as it feeds into the ecosystem. So it's clear that we in the industry are at an exciting inflection point with the emergence of numerous technologies and innovations. There's definitely the ability for this to really shape the future of the industry and how we're going about driving and delivering quality and efficiency overall. I would encourage you guys, it's shameless plug, but we do have a white paper that was published early last year that explores a little bit more around these topics. The digital workforce specifically some thought around generative AI and how this is being shaped with thinking in the industry right now. So I will leave this here just a little bit of a plug, feel free to check this out. It's an interesting read if nothing else to help that against some of your thinking that you guys might have in this space. And so with this, we have abundant time for Q&A. So I will go ahead and hand over now to our moderator, Andy, to get us kicked off on our Q&A.

All right. Great. Thank you, Jon, and thank you, Debjit as well for such a great presentation. [Operator Instructions] So, we will just give it a minute to get our first question ready. And it is going to be, would you recommend having a 0% SDV approach if you have sites with only 1 patient Also, what if primary endpoint data is captured on paper, for example, ECOA and the only way to confirm the correct entry is through SDV.

All right, Debjit, I'll let you take a stab at that one.

Okay. So I think -- Andy, can you repeat the question for me once?

Of course, yes. So here's the question. Would you recommend having a 0% SDV approach if you have sites with only 1 patient. Also, what if primary endpoint data is captured on paper, for example, ECOA and the only way to confirm correct entries through SDV.

Yes. I think the general move that we are recommending is obviously moving towards 0% SDV because if you are collecting data centrally through connected devices, et cetera. So there is not much for you to do SDV and the focus more on the critical review, which is source data review. So which would be applicable for depending on the -- irrespective of the large number of studies. But again, the fundamental aspect here is that if we collect data centrally through e-source options, then there is not much for source data review -- source data verification. It's more emphasis on source data review. And the other element, I think, is about examples of the like primary and secondary endpoints related to informed consent or the subject eligibility related to the primary and secondary end points can be collected. The data in the predominantly, I think we collect that in the EDC. But again, if there are dependency, we can definitely connect through the eSource options.

Okay. Great. Thank you, Debjit. And moving on to our next question. How does RBQM help increase patient recruitment?

It's a good question. So I can take a stab here overall. I would say from a model approach, RBQM, when it comes to patient recruitment, you can have several different views. So I guess, the more natural way to think about this is if you have more limited on-site interaction with monitors, face-to-face interaction potentially that impacts in a negative way recruitment. Some of what we've seen, though, with more of the higher touch site models that Debjit was talking about when it comes to RBQM, at least within IQVIA, we're still able to maintain that high presence and high touch when it comes to sites and working through recruitment and recruitment challenges that might come up to keep the study front of mind to also support the overall patient recruitment process. So it's one that I think it comes down to how the model is structured within any organization and then also that focus on the recruitment piece. Because we know in different therapeutic areas, there might be different challenges that come up with patient recruitment. So you go into each of those scenarios knowing how you need to be able to amend and be flexible when it comes to recruitment strategies. That would be my kind of initial thought around how RBQM maybe impacts or what it looks like when it comes to patient recruitment. Debjit, anything you want to add to that specifically?

No, I'm good, Jon, with that. I think you have covered that I think the evolving delivery models definitely addresses that aspect of the element.

Okay. Great. Thank you both. So now, we're going to move on to the next question. This one is a longer form. So here you go. Is it recommended or what guidance to perform centralized monitoring on a Phase I study, say, 10 patients in one site versus a Phase II study, 20 to 50 patients in 10 to 20 sites versus a Phase III study over 200 patients in over 50 sites. I agree that centralized monitoring is highly beneficial for Phase III studies but what about Phase II and Phase I studies when there is less data sites and patients.

Yes. I think Jon, feel free to add. But I think in general, I think the way the centralized monitoring strategy is that, again, it goes back to the risk assessment, right? So we make sure that we coordinate with all the cross-functional functions and definitely define for each function, what kind of mitigations are you taking care through centralized monitoring, again, irrespective of the number of patients and all, it's predominantly based on your critical factors, critical to quality factors, what are the mitigations in a centralized monitoring plan, what are the mitigations in a site monitoring plan? And what are the mitigations in a data management plan? So in that way, we are making sure that all the mitigations associated with critical factors from the protocol are acted upon across the various cross-functional team and there is no redundancy in terms of picking up those outliers. So I think that, again, goes back to risk management, then it is integrated to the operation plans and finally, to the downstream delivery. So that's the overarching concept in terms of deploying centralized monitoring for any trials in our role. Jon, anything that you want to add more on to that?

No, I think you well summarized there.

Okay. Great. Now on to our next question. When we say adoption of skill sets in evolving delivery models, is IQVIA looking at specific competencies for centralized staff?

Yes. I think this, again, I can take it, and then Jon, feel free to add. I think yes, exactly we are seeing that the delivery models have been kind of evolving. So some of the key areas that we are looking at making sure that the collaboration skills plus the risk-based mindset, it's very important to kind of deploy such models. And in general, I think analytical mindset, critical thinking capabilities because those are key attributes for us to kind of identify the risk identification and the assessment in this space. So definitely, and I think we are also kind of deploying a lot of capability programs to ensure that we kind of in-house develop such competencies.

And I would add, it's a space where there has been sizable change, right, as RBQM models came on to the scene and then as they continue to evolve with technology. So we do see that there's a lot of appetite for innovation and hunger to learn and more of that change being seen as something that is a normal or a regular part of the role, and that's really what helps make these teams successful as well as that willingness to adapt to change and to look for better ways of completing some of the central monitoring processes. So I think that fits in nicely with that critical thinking and the analytics mindset, but really having a focus and desire for innovation.

All right. Thank you both. Next up, how are you measuring the success of RBQM implementation?