The FDA Just Blew Up 60 Years of Clinical Trial Protocol
The FDA is letting Amgen and AstraZeneca share live clinical trial data with regulators as patients enroll, not after trials end. It could cut drug development timelines by 20-40%, and it's the biggest change to how trials work in six decades.
The Biggest Change You've Never Heard Of
For six decades, clinical trials have worked like this: run the study, collect the data, package it up, ship it to the FDA, and wait. It's like filming an entire movie, mailing the reels to the studio, and hoping they like it months later. No dailies. No feedback. Just silence until the very end.
That model just got torched.
The FDA announced a pilot program with Amgen and AstraZeneca that lets regulators watch clinical trial data in real time, as it accrues, through cloud-based access. Not after the trial ends. Not in batches. As it happens. Commissioner Marty Makary called it an answer to "years" of lag between data generation and FDA review. If it works, the agency estimates it could cut overall clinical trial timelines by 20 to 40 percent.
That's not an incremental improvement. That's a fundamentally different relationship between drugmakers and the people who decide whether their drugs reach patients.
What's Actually Being Shared (and Why Oncology Goes First)
Both pilot programs focus on cancer, where speed literally saves lives.
AstraZeneca is running a Phase 2 trial called Traverse, testing its BTK inhibitor Calquence (acalabrutinib) combined with AbbVie's Venclexta and rituximab in patients with treatment-naïve mantle cell lymphoma, an aggressive blood cancer. The trial is already operational at MD Anderson and the University of Pennsylvania, and the FDA has reportedly received and validated real-time safety and efficacy signals through the platform.
Amgen is testing Imdelltra (tarlatamab), a T-cell engager that targets a protein called DLL3, in patients with small cell lung cancer. Final site selection was still in process as of the announcement, with MD Anderson and UPenn anticipated as trial locations.
Both trials use a platform built by a company called Paradigm Health, part of a system called SPIRE (Scalable Platform for Integrated Research & Evidence). Think of it as a translation layer between hospital data systems and FDA reviewers. It automates data collection and analysis, then transmits predefined safety and efficacy signals in days rather than months.
Get tomorrow's biotech intelligence before your competitors.
Join thousands of biotech professionals who start their day with our free, daily briefing.
The key word there is "predefined." The FDA isn't watching raw patient data scroll by like a stock ticker. Regulators and sponsors agree in advance on which clinical endpoints matter (fever onset, tumor shrinkage, adverse events), and the platform flags those signals automatically.
Why This Isn't Just Rolling Review With a Tech Upgrade
If you follow pharma, you might be thinking: "Doesn't the FDA already have ways to speed things up?" Fair question. Rolling review, Fast Track, Breakthrough Therapy designation: these all exist. But they work differently.
Rolling review lets companies submit completed sections of their application as they finish them, rather than packaging everything at once. It's like turning in chapters of your thesis one at a time instead of the whole thing on deadline day. The FDA still reviews static snapshots of finished work.
Accelerated Approval lets drugs onto the market based on surrogate endpoints (biomarkers that predict clinical benefit), with confirmatory trials required afterward. It shortens the path to market, but it doesn't change how data flows during the trial itself.
Real-time data sharing is a different animal entirely. The FDA doesn't wait for the sponsor to decide what to submit or when. Regulators have continuous visibility into agreed-upon endpoints as patients enroll and data accumulates. It's the difference between getting a movie's box office numbers on Monday morning versus watching ticket sales update every hour on opening weekend.
This means the FDA could theoretically make decisions mid-trial: flagging safety concerns earlier, suggesting dose adjustments, or greenlighting transitions from Phase 1 to Phase 2 without the traditional months-long gap between data lock and regulatory feedback.
The Promise: Faster Drugs, Fewer Dead Ends
FDA Chief AI Officer Jeremy Walsh framed the potential impact clearly. He suggested the approach could reduce "20, 30, 40% of overall clinical trial time" through real-time monitoring and faster go/no-go decisions.
Consider what that means in practice. The average drug takes 10 to 12 years to go from discovery to approval. A 30% reduction would shave three to four years off that timeline. For cancer patients running out of options, those aren't abstract numbers.
The pilot also hints at something bigger: continuous trials. Instead of rigid, sequential phases with long pauses between them, drugs could potentially flow from Phase 1 into Phase 2 seamlessly, with regulatory oversight baked into the process rather than bolted on afterward. Makary has described this as the long-term vision.
The Risks Nobody Wants to Talk About
Of course, watching data in real time introduces problems that batch submissions never had.
Statistical landmines. Legal analysts Mark Tobolowsky and James Valentine flagged concerns about Type 1 and Type 2 errors (false positives and false negatives, respectively). When you look at data repeatedly as it accumulates, you increase the chance of seeing patterns that aren't real. This is the same reason you shouldn't check your retirement account daily; normal fluctuations look like catastrophes when you watch them in real time. Without robust statistical analysis plans, premature decisions could kill promising drugs or advance dangerous ones.
Privacy questions. The FDA says it's only accessing aggregated signals, not raw patient data. But continuous data streams flowing through third-party cloud platforms create surface area for breaches. The agency requires data to be "traceable, auditable, and privacy-protecting," though specifics on encryption and audit protocols remain thin.
Regulatory creep. If the FDA can see everything as it happens, does that change the dynamic between sponsor and regulator? Could real-time visibility create pressure to intervene too early, or bias reviewers who've been watching a trial unfold for months before seeing the final package? These are open questions the pilot needs to answer.
Commentators like Fred Ledley have noted it's wise that the FDA introduced this "only as a pilot program" to "evaluate this concept in practice, not in theory."
The Bigger Picture: AI Meets Regulatory Science
The real-time data pilot isn't happening in isolation. The FDA simultaneously issued a Request for Information (due May 29, 2026) seeking public feedback on a broader pilot launching this summer. That program will focus on AI-enabled optimization of early-phase clinical trials, covering safety monitoring, dose optimization, safety signal identification, and patient recruitment.
The through-line is clear: the FDA under Makary is betting that technology can solve problems the agency has wrestled with for decades. Early-phase trials, characterized by high uncertainty and tiny patient populations, are where the most time gets wasted. If AI can help identify the right dose faster, flag safety problems sooner, and recruit patients more efficiently, the downstream effects compound across the entire development timeline.
Paradigm Health's platform is the first proof of concept. Whether it scales depends on how well the AstraZeneca and Amgen pilots perform, and whether the broader industry trusts the system enough to participate.
What Happens Next
The summer 2026 expanded pilot will be the real test. The current proof-of-concept trials are small, focused on oncology, and limited to pre-agreed endpoints. Scaling this to larger trials, more therapeutic areas, and more complex endpoints will surface challenges the pilots can't predict.
But if it works? The implications are enormous. Drug development's biggest bottleneck has never been science alone; it's been the glacial pace at which information moves between the people generating data and the people making decisions about it. Real-time sharing attacks that bottleneck directly.
For patients, this could mean treatments arriving years earlier. For companies, it could mean fewer expensive late-stage failures (because problems get caught sooner) and faster revenue from successful drugs. For the FDA, it could mean a more active, engaged role in drug development rather than serving as a distant gatekeeper who only weighs in at the end.
Sixty years is a long time to do anything one way. The clinical trial system was overdue for a rethink. Whether this particular rethink works remains to be seen, but the ambition is undeniable: turning a one-way mailbox into a two-way conversation, happening in real time.
Deals & M&A
4 min read
BioNTech Built a Vaccine Empire. Now It's Tearing It Down.
BioNTech is shuttering four manufacturing sites and cutting 1,860 jobs as it dismantles its COVID vaccine empire. The company is betting everything on becoming an oncology powerhouse by 2030, armed with €16.8 billion in cash and 15 Phase III cancer trials.
