Real-Time Clinical Trial Oversight and the Infrastructure Gap: Q&A with Raj Indupuri, eClinical Solutions

The FDA's move toward real-time clinical trial data review is less a technology challenge than an infrastructure one—and for most sponsors still operating on fragmented, siloed systems, the gap between where the industry is and where regulators are heading is significant.

To explore this further, Applied Clinical Trials spoke with Raj Indupuri, CEO and co-founder of eClinical Solutions, about what continuous oversight actually requires of clinical and data teams, how RBQM needs to evolve at the enterprise level, and what it means to build AI that sponsors can actually trust.

ACT: What does the FDA's move toward real-time clinical trial data review actually require of sponsors and CROs operationally, and how ready is the industry?

Indupuri: From our perspective it's an exciting push from the FDA, demanding this continuous review. But at the same time, I don't think the industry is as ready as it should be. The industry has been built on fragmented systems, different EDCs, different software vendors, delivering non-standardized data. To do continuous review, you have to bring all of this together, and it requires modern data infrastructure and advanced pipelines to support different users in the clinical value chain.

The other big benefit here is that you can now think about compressing timelines, which the industry has been struggling with as complexity continues to increase. You can achieve real-time oversight. But that's not possible when your data is trapped in different silos, and milestone-based review just won't work in that environment.

Sponsors cannot achieve real-time review until they invest in unified data infrastructure that harmonizes clinical data from all these fragmented sources and systems. And not only building that infrastructure, but also having a pipeline so that when data gets in, it can be ingested in real time and delivered in a highly governed way to users so they can act on it for immediate decision making.

This also requires the industry to think about implementing infrastructure and systems holistically, instead of point solutions. Think about the entire value chain. Right now, a lot of companies have so many different point solutions, and that makes this much more difficult.

ACT: How does continuous data review change the way teams need to think about data quality and governance throughout a trial, not just at key milestones?

Indupuri: The most important shift here is that quality has to be built into the entire process continuously, rather than waiting until the end. Right now it happens at certain milestones—database lock, interim analysis, submission preparation. But in a real-time oversight environment, governance and risk management become more critical and come under scrutiny across the entire life cycle of the trial.

Teams cannot rely on manual processes and accessing different siloed systems for cleaning, reviewing, and analytics. Governance needs to be more around continuous monitoring with traceability and alignment across departments—clinical operations, data management, safety, stats, analytics. In the majority of companies, these departments still work in silos where data gets copied and isn't unified.

Organizations and teams really need to think about using a modern data infrastructure that unifies with this pipeline and with intelligent products that can support governance in real time. When data gets in, capability needs to be built in so that trust is built in, and it can support this continuous review and downstream decision making. It also requires a different agility and a different mindset in terms of how you do review—rather than going back to time-based review.

ACT: Where does RBQM need to evolve to function effectively in a real-time oversight environment?

Indupuri: This is a very important aspect as we move into continuous data review. The industry has been adopting RBM for a while, but many companies have been implementing it at a study level and in a siloed way, rather than thinking enterprise-wide quality and risk management.

To move in the direction the FDA is suggesting, you need not just a unified infrastructure but an integrated approach to data review, quality, and risk management. There's a huge opportunity to move from study-by-study implementation to an enterprise level where you can bring data review, quality, and risk management together—where clinical, operational, and safety teams are all connected to the same infrastructure and the same system to continuously monitor risk.

What can really accelerate this now, which is quite exciting, is adding an intelligence layer with AI. At eClinical we are innovating and investing heavily in agents across this life cycle, where you can eliminate a lot of manual tasks and bring this integrated approach to end users in a governed, trustworthy way. That reduces cycle times, makes decision making much faster, and keeps patient safety at the center. And with this approach, you can reduce not just cycle times but also the cost and effort needed to do review and achieve the milestones you need for submission to FDA or other regulatory bodies.

ACT: As AI tools become more embedded in life sciences research, how do you ensure they're supporting rather than outpacing the governance frameworks trials depend on?

Indupuri: The pace of AI innovation over the last 12 to 18 months is something we have never seen. What used to happen over decades is now happening in months or even weeks, and that has huge potential to improve the entire value chain. But when it comes to review, quality, and risk management, one of the key challenges is that AI in general is a black box. For sponsors to adopt it, it needs to be trustworthy. There needs to be governance and explainability to give confidence to different users in the value chain so they can trust the outputs and take action.

What we believe in is bringing transparency—making AI a glass box. We want to give sponsors visibility into how patients are being supported, what context is being provided to the agents, and how outputs can be evaluated. The way we are building AI agents for customers is to give sponsors the ability to inject their own context—their sponsor-specific workflows, SOPs, protocol context, historical data, governance frameworks—so that they can trust more.

And on the output side, we want sponsors to have access to evaluate and validate what the systems are producing, and to ensure they have traceability over time so their users can trust and act on those outputs. The opportunity here is huge. It's not a question of when—it's a question of how we take advantage of it and move the needle.

ACT: How is RWE intersecting with these shifts, and what role might it play in a more continuous, data-driven regulatory environment?

Indupuri: One of the most important shifts is that the boundary between traditional clinical data and real-world data is blurring. Going back to the earlier point around continuous data review and decision making—you have data coming from clinical trials, but there's a huge opportunity to look at real-world data as well. We believe these boundaries will continue to blur, and if you have modern data infrastructure that can harmonize and bring these together, it can immensely support continuous oversight and give confidence to regulators and sponsors that the evidence is being produced from real-world data as well.

Organizations that can think about this infrastructure in a broader way—building pipelines that bring clinical trial data and real-world data together seamlessly and building a connected ecosystem—will significantly benefit. And it ultimately comes down to how you deliver that combined data to different stakeholders for efficient, governed decision making. Using AI with the glass box approach we talked about, and more importantly having a pipeline that can do this in real time, will make a huge difference and can move the needle in a significant way.