Josh Murphy
Aaron Kamauu
Wendy Turenne
We’re moving toward a world where you can do more continuous evidence generation and break down siloes.”
As regulators around the world grow more comfortable with real-world data (RWD) and AI-enabled evidence generation, pharma companies are rethinking how, and when, they build evidence. What was once a post-launch activity is increasingly becoming a continuous process that begins early in development and extends throughout a product’s life cycle.
At Panalgo’s annual Data to Insights conference, Josh Murphy, SVP of RWD and Strategic Partnerships at Norstella—Panalgo’s parent company—spoke with Wendy Turenne, Principal at RWW Advisory and Aaron Kamauu, CEO of Navidence, about shifts in the regulatory landscape, the importance of integrated evidence generation, and where global agencies are aligning on the use of AI and RWD.
Q: Having data increases chances of regulatory approval, but it’s not guaranteed. What should pharma companies be doing in early regulatory submission discussions to prepare? What’s your advice on how to be part of evidence generation and regulatory submission?
Aaron Kamauu: For integrated evidence generation planning, you need to broaden the perspective of stakeholders downstream. We’re moving toward a world where you can do more continuous evidence generation and break down siloes. It’s evidence at launch, not after launch.
Wendy Turenne: Have your infrastructure ready. You’re building evidence, learning from it, then, as you go further, even after launch, it’s still a flywheel. Using data to understand the care setting you’re launching into will help you understand the evidence that will be needed.
Q: Are you seeing consistency across global regulatory bodies?
AK: Some agencies are more open and public. Generally speaking, it’s consistent in terms of the value of RWD in submission packages. There’s a combined FDA/EMA list of published best practices around the use of AI, so that shows that there’s collaboration between the agencies in aligning on concepts of interest. I expect that to turn into more formal guidance.
WT: There’s a fair amount of consistency across guidelines. We’re scientists. We want transparency. You have to use data that’s collected in the setting that’s relevant to where care is going to be delivered. There’s real consistency in the foundations, fundamentals, and core regulations. Variability comes in on the willingness to act and make decisions by the regulators. The guidance feels similar, but the biggest variability comes in the regulators accepting the evidence or making decisions based on it.
Q: Will approval in one country, like China, speed up approval in the U.S. and Europe?
WT: Well, certainly in Japan, they understand that there’s innovation happening in the U.S. They have put really specific pathways, like bridging studies, in place and have been thoughtful and intentional about opening up use of RWE.
AK: We’re asking these agencies to judge whether a new innovation is going to be available in their market. When you leverage RWD as a supplement or augmentation to replace things like random clinical trials, they want assurances that it’s the same kind of patient population and the same outcome measures; the operational definitions matter a lot. There’s still an expectation that a meaningful percentage of the study population is coming from the country where you’re getting that approval.
Q: The guiding principles that were published by the FDA in January 2026, the “10 AI Commandments,” if you will, are “human-centric by design,” “risk-based approach,” “adhere to standards,” “clear context of use,” “multi-disciplinary expertise,” “data governance and documentation,” “model design and development practices,” “risk-based performance and assessment,” “life cycle management,” and “clear, essential information.” Are any of these a surprise to you?
WT: Isn’t this what we’ve been doing with scientific evidence and analytics this whole time? We know what good looks like and how to validate things. AI is still based on good science and high-quality, complete data.
AK: When I first saw the list, I thought, this should be obvious, so it’s a good reminder about applying good science to advancing technologies. They highlight having a specific use, and when I think about research, I think about starting a research study with a research question, the target population, the target audience, and the intended use by the audience. It’s the same thing with AI: What am I using it for? And what does good look like?
Contact us to find out how Ella AI and IHD can power your evidence generation strategies, faster.