The FDA's RTCT Push: What It Is; Why It Matters

By John Oncea, Chief Editor, Clinical Tech Leader

On April 28, 2026, the FDA announced two concrete steps toward a fundamentally different model of clinical oversight: the successful initiation of two proof-of-concept real-time clinical trials (RTCTs), and a Request for Information (RFI) for a broader pilot program planned to launch this summer.

For decades, the rhythm of clinical drug development has followed the same basic pattern: sites collect data, sponsors clean and analyze it, and then submit it to the FDA, often months or years after the underlying events occurred. RTCT is designed to reshape elements of that pattern by allowing FDA reviewers to see pre-agreed safety signals and efficacy endpoints as trials progress, rather than waiting for periodic submissions.

This is not just a faster version of what the FDA already does. It is an attempt to redesign how evidence reaches regulators in the first place. And it is important to be clear about what the initiative is and is not: RTCT is a pilot-phase modernization effort focused on structured signal sharing, not continuous raw-data surveillance, and not a replacement for existing regulatory review frameworks.

What The FDA Actually Announced

Two pharmaceutical companies, AstraZeneca and Amgen, are already running proof-of-concept RTCT studies in oncology. AstraZeneca is conducting a Phase 2 multi-site trial, TRAVERSE, in patients with treatment-naïve mantle cell lymphoma, with participation from MD Anderson Cancer Center and the University of Pennsylvania. Amgen is running a Phase 1b study, STREAM-SCLC, in limited-stage small cell lung carcinoma, with final site selection still in process. For each trial, the FDA met with the sponsor in advance to establish agreed criteria for what signals would be reported in real time.

The FDA has confirmed it successfully received and validated signals from AstraZeneca’s trial. That validation matters: it establishes that the architecture for real-time signal sharing is not theoretical. It works in practice, at least in these early configurations.

Alongside the proof-of-concept announcement, the FDA released an RFI for a broader pilot program focused on AI-enabled optimization of early-phase trials. The RFI invites industry input on program design, implementation, evaluation metrics, and success criteria. Comments are due May 29, 2026. The agency intends to publish final pilot selection criteria in July and complete participant selection in August.

What “Validated Signals” Actually Means

One of the most important clarifications in the RTCT initiative is what “signals” actually means in practice, because the term is easy to misread. It does not mean the FDA is receiving a continuous stream of raw patient-level EHR data.

The process begins with the study protocol. The FDA, sponsor, and technical platform work together to align a clinical trial reporting schema: specific safety, efficacy, data quality, and operational events that should be detected and reported as the trial runs. The platform then ingests trial-relevant data from the healthcare setting – structured data like labs, vitals, medications, and ECOG scores, as well as unstructured data such as clinician notes or imaging reports – and maps those elements back to the agreed schema.

Kent Thoelke, CEO of Paradigm Health, the technology platform that served as the technical conduit in AstraZeneca’s proof-of-concept, describes what the FDA actually receives: “The FDA-facing signal is not a raw EHR feed or a patient-level data dump. It is a predefined signal output tied to the reporting schema, tied back to the underlying source data, and confirming that it was generated according to the agreed logic. That traceability is the key point: reviewers can understand why the signal fired, what source information supported it, and whether any clarification is needed.”

Crucially, the proof-of-concept demonstrated that the full pipeline, from initial discussions through an operational signal flow, could be stood up in less than four months in an active clinical trial environment.

Why The Proof-Of-Concept Approach Matters

The phased structure of the initiative, proof-of-concept first, then pilot, then potential broader adoption, was a deliberate design choice, and it matters more than it might appear.

Tala Fakhouri, who led development of the FDA’s AI policy before joining Parexel as a regulatory strategy executive, describes the traditional regulatory engagement model as inherently reactive. Under the current model, data flows to the agency at pre-specified milestones, formal meetings happen at defined intervals, and the agency waits between them. Real-time clinical trials replace that sequential model with something closer to a continuous collaborative loop. “The whole idea behind the proof of concept and the pilot,” she said, “is to see if that engagement can be done more continuously, with regular engagement from regulators and industry.”

From a data science perspective, the proof-of-concept was also important because it required demonstrating feasibility before the agency committed to the full operational and governance architecture that a broader pilot requires. The FDA needed to confirm not just that the technology could transmit signals, but that reviewers could receive and interpret them within an operational context.

What made that possible, according to Thoelke, was a fundamentally different working model: “This is a new model for working with the FDA, where we were building iteratively and collaboratively, rather than in a waterfall model. This collaborative approach greatly aligned all parties towards a shared goal, resulting in faster progress. I was blown away by the speed with which all parties worked to bring this vision to life.”

The three-party alignment on what to monitor was itself a critical design feature. The FDA reviewed each trial protocol and proposed signals it wanted to track; sponsors provided feedback; the technology partner weighed in on operational complexity. The result was a shared clinical trial reporting criteria document that all parties could act on simultaneously.

The FDA’s Goals – And The Competitive Pressure Behind Them

Former FDA Commissioner Marty Makary framed the initiative in stark terms: for 60 years, the agency has conducted clinical trials the same way, with key data signals sometimes taking years to reach regulators. The stated goal is to eliminate or compress the “dead time” between trial phases, the gap that currently consumes roughly 45 percent of overall drug development time without a single trial in progress.

But there is a second driver that receives less attention in the formal announcements: competitive pressure. The FDA and the current administration have both signaled concern about the migration of early-phase trial activity to other geographies, particularly Australia, New Zealand, and China, where clinical trial speed and informality of early regulatory engagement have made those markets attractive for some sponsors. Fakhouri described RTCT as the FDA’s attempt to formalize and technologically enable a similar collaborative dynamic within the U.S. regulatory system.

The FDA’s proposed 2027 budget reflects this priority directly, with explicit references to pre-IND reforms aimed at repatriating early-phase trial activity to the United States.

“Faster” Is Not The Same As “Better”

One of the most important framing questions for RTCT is what the initiative is actually designed to produce: faster evidence, better evidence, or faster access to pre-agreed evidence.

Fakhouri is precise on the distinction. “FDA is not getting raw data and making decisions on that raw data,” she said. “FDA is getting data on signals defined as pre-specified metrics that both the regulator and industry agreed on beforehand.” The signals being transmitted are transformed metrics, not raw data streams, and both parties agree in advance on what constitutes a meaningful threshold. That agreement is what makes real-time transmission actionable.

Thoelke underscored the same point from an operational standpoint, noting that what the agency has learned from the proof-of-concept is that it is “very serious about trying to move more of the data collection, cleaning, processing, and analytics steps earlier in the overall trial process.” The goal is to reduce the downstream burden of reconciliation, not to give regulators unfiltered access to raw clinical data in motion.

The practical implication is that RTCT has the potential to enable faster go/no-go decisions at pre-agreed milestones, not because the data is inherently better, but because the FDA can act on it as soon as it exists, rather than waiting for a formal submission cycle.

What Comes Next

With the RFI closing May 29, the FDA is synthesizing industry input to design the pilot program. Based on public statements, the pilot is expected to include a limited number of early-phase trials across multiple therapeutic areas, not just oncology, but potentially metabolic health and others where rapid safety and dose-adjustment decisions are critical.

Fakhouri’s advice for sponsors is direct: respond to the RFI, participate in the pilot if possible, and make sure data architecture and technology are in a place that would support participation. Whether or not a sponsor is selected for the pilot, the infrastructure decisions being made now, around data pipelines, AI governance, and signal detection, will determine who is positioned to move quickly when broader adoption follows.

For the industry, the lesson of the proof-of-concept may be less about speed than about readiness. As Thoelke put it: “Being close to the provider and their core clinical systems is key to this effort working.” The organizations that understand that principle now will be the ones best positioned when the pilot scales.

This is Article 1 of a three-part series. Article 2 examines what RTCT demands from clinical research technology and operations. Article 3 covers the legal, regulatory, and operational risks, including unresolved questions about sponsor liability, blinding integrity, and DSMB oversight.