For sixty years, the architecture of clinical drug development has been governed by a fundamental constraint: data moves from patients to trial sites, from sites to sponsors, from sponsors to the FDA, and only then does the regulatory conversation begin. Each handoff adds weeks, months, sometimes years to the timeline. Each pause between phases represents what FDA Commissioner Marty Makary, speaking at a press conference on April 28, 2026, called dead time which is the administrative friction that delays potentially life-saving treatments from reaching patients who cannot afford to wait.

On that date, the FDA announced it was beginning to dismantle that architecture. The agency unveiled two proof-of-concept real-time clinical trials (RTCTs) — one from AstraZeneca targeting mantle cell lymphoma, one from Amgen targeting small cell lung carcinoma — in which safety signals and clinical endpoints are transmitted directly to FDA reviewers as the trial progresses, through a cloud-based platform built by Paradigm Health. Simultaneously, the agency published a Request for Information (RFI) seeking public comment on a broader summer pilot program, with selection criteria due in July and final selections in August 2026.

The announcement was notable both for its ambition and for its specificity. This is not a policy aspiration or a regulatory framework document. The FDA has validated actual signals from the AstraZeneca TRAVERSE trial through Paradigm Health's platform, confirming that the technical infrastructure for real-time regulatory review is not a theoretical possibility — it is operational today.

"For 60 years, we've been conducting clinical trials in the same way, where key data signals can take years to reach the FDA. The lag time can delay regulatory decisions unnecessarily and slow down the drug development timeline." — FDA Commissioner Marty Makary, April 28, 2026

The Status Quo: Why Clinical Trials Are Slow

Dead Time in the Development Cycle

Drug development is often cited as one of the most expensive and time-consuming endeavors in modern commerce. The conventional estimate — ten to twelve years from initial discovery to approval, with costs exceeding two billion dollars per approved drug — is shaped by many variables, but one underappreciated contributor is the informational bottleneck inherent in the trial structure itself.

FDA Chief AI Officer Jeremy Walsh, who began working on the real-time trials project in June 2025, has quantified the problem precisely. On average, 45% of the time between a Phase 1 clinical trial and submission of a regulatory application to the FDA is dead time — spent on data reconciliation, formatting, transmission, and review queuing rather than actual scientific evaluation. An application the FDA received in one recent case consisted of 66 million pages. Makary cited this figure not as an abstract data point but as an indictment of the workflow it represents.

The structural cause is clear: trial data is collected at clinical sites using systems that were never designed to talk to each other. It is manually reconciled, entered into separate analysis systems, formatted for regulatory submission, and only then transmitted to the FDA — where it enters another queue. By the time regulators can evaluate it, the moment may have long passed when an intervention would have changed the trial's direction.

Jennifer Litton, chief clinical research officer at MD Anderson Cancer Center — one of the sites participating in the AstraZeneca TRAVERSE trial — described the problem in visceral terms: "Right now, we collect vast amounts of complex data, often manually, and then we put them in multiple different systems that were never made to talk to each other. When we do that, we take the time away from our patients and our researchers."

The Human Cost Walsh Identified

Walsh's framing of the problem is worth dwelling on. The case for real-time trials is not primarily a case about regulatory efficiency or competitive advantage in attracting trial sponsors away from China and Australia — though those considerations are present in the FDA's thinking. It is a case about patients. "We have to consider our processes from the standpoint of a patient awaiting a potentially powerful treatment," Walsh said. Litton, who spoke as an oncologist, as the daughter of two parents who had cancer, and as someone who had been diagnosed with cancer herself, made the same point with equal force: "If you need the hope for a clinical trial, you want it right then."

What FDA Actually Announced: Two Trials, One Platform, One Pilot

TRAVERSE and STREAM-SCLC

The two proof-of-concept trials announced on April 28 represent meaningfully different oncology contexts, which is likely deliberate. AstraZeneca's TRAVERSE is a Phase 2 multi-site trial testing the combination of Calquence (acalabrutinib), AbbVie's Venclexta (venetoclax), and rituximab in patients with treatment-naive mantle cell lymphoma — a rare and aggressive B-cell cancer with historically poor outcomes. The trial sites are MD Anderson Cancer Center and the Perelman School of Medicine at the University of Pennsylvania, two of the most data-sophisticated academic medical centers in the United States.

Amgen's STREAM-SCLC is a Phase 1b trial of Imdelltra (tarlatamab), a bispecific T-cell engager targeting DLL3 and CD3, in patients with limited-stage small cell lung carcinoma. SCLC is among the most difficult-to-treat cancers, with a five-year survival rate that has changed little in decades. Final site selection for STREAM-SCLC was still in process as of the announcement.

For each trial, FDA met with the sponsor to pre-establish the specific safety signals and clinical endpoints that would be reported in real time — a critical design step that Walsh described as a prerequisite for making real-time review tractable rather than overwhelming. Without this pre-specification, real-time access to trial data risks creating exactly the information overload Walsh warned against.

Paradigm Health's SPIRE Platform

The technical infrastructure underpinning the RTCT initiative is Paradigm Health's Study Conduct platform, part of its broader Scalable Platform for Integrated Research and Evidence (SPIRE) end-to-end clinical trial infrastructure. Paradigm Health, incubated by ARCH Venture Partners, has engineered its platform specifically for this FDA collaboration, working with the agency since early 2026 to define reporting and validation protocols and ensure software interoperability.

The platform works by ingesting data directly from electronic health records and other structured and unstructured sources at trial sites, algorithmically evaluating FDA-defined data points and reporting criteria in real time, and transmitting only the critical signals needed for regulatory determinations to both the trial sponsor and the FDA. The resulting data flow is traceable, auditable, and designed to protect patient privacy — while minimizing the transfer of unnecessary datasets that have historically contributed to the 66-million-page problem.

Kent Thoelke, Founder and CEO of Paradigm Health, described the ambition succinctly: clinical trial data can be analyzed for key signals in near real time and shared with trial sponsors and the FDA in days, rather than months. That is the core value proposition — not incremental improvement in an existing workflow, but a qualitative change in the regulatory conversation that AI-enabled data infrastructure makes possible.

The RFI and the Summer Pilot

Beyond the two proof-of-concept trials, FDA released an RFI seeking input on the design and implementation of a broader RTCT pilot program, which the agency intends to launch in summer 2026. The agency has set a May 29, 2026 deadline for public comments, plans to disseminate final selection criteria in July, and expects to complete pilot selections in August. The pilot program will be aligned with the AI risk management framework developed by the National Institute of Standards and Technology, embedding the initiative within the broader governance architecture for AI in regulated settings.

The agency also asked the public to weigh in on a separate proposed pilot to work with companies using AI to enhance safety monitoring and dose selection, identify safety signals, and improve patient recruitment in clinical trials more broadly — suggesting that real-time data review is one application within a larger agenda of AI-enabled trial modernization.

The Deeper Implications: What This Changes

From Phase-Gated to Continuous Development

The most far-reaching implication of the RTCT initiative is not the efficiency gain within any single trial. It is the pathway it creates toward continuous clinical development — eliminating the structured pauses between trial phases that currently add months or years to the development timeline.

Under the current paradigm, clinical development proceeds in discrete phases — Phase 1, 2, 3 — each run as a separate study under a separate protocol. The hiatus between phases, during which trial data is assembled, submitted, reviewed, and an agency decision rendered, can stretch to twelve to eighteen months. The FDA's explicit goal is to use real-time data access to reduce or eliminate this hiatus, enabling what Makary called continuous trials that run across all phases of clinical development without the interruptions that currently define the process.

This is a structural reimagining of drug development, not a process improvement. It has potential implications for how trials are designed, how protocols are written, how sponsor-agency relationships are structured, and — most significantly — how quickly efficacy and safety signals can translate into regulatory decisions.

Safety Monitoring Gets Faster

Real-time access to safety data is not only a speed advantage — it is a safety advantage. When an FDA reviewer can see in real time that a patient in a clinical trial has developed a fever, been hospitalized, or shown a tumor response, the agency's ability to respond is qualitatively different from a system in which that information arrives weeks or months later as part of a formal submission package.

Walsh noted that this capability transforms not just when the FDA sees data, but how it thinks about decisions. "We're reimagining what information we need and when we need it in order to make a decision," he said. For the FDA, the shift represents a move from episodic review of accumulated data to continuous surveillance of trial progress — a model more analogous to how clinicians monitor patients in intensive care than to how regulators have historically reviewed drug applications.

Gesund.ai perspective:

This is precisely the operating model that clinical AI built for real-world evidence generation and post-market surveillance has been anticipating. Real-time trial monitoring is the regulatory counterpart to real-world AI-based safety monitoring. The data architectures, governance frameworks, and validation methodologies that health AI companies have been developing for continuous patient monitoring translate directly to this new regulatory context.

The Competitive Dynamics for Trial Location

FDA Commissioner Makary has been explicit about one of the ancillary motivations for the RTCT initiative: making the United States a more attractive location for early-phase clinical trials. Currently, many pharmaceutical and biotech companies initiate their first-in-human trials in Australia, the United Kingdom, or other jurisdictions with more streamlined regulatory entry points and faster trial startup times. The FDA is aware that this trend has long-term implications for the U.S. clinical research ecosystem and for the availability of cutting-edge trials to American patients.

By demonstrating that the U.S. regulatory environment can match or exceed the efficiency of peer jurisdictions — while maintaining its gold-standard review rigor — the RTCT initiative is partly a competitive signal to industry. If FDA can review real-time signals from a U.S.-based trial as rapidly as a sponsor could initiate and run a trial in Australia, the calculus for trial location changes.

The Gesund.ai Perspective: What This Means for Health AI Companies

Validation Infrastructure Becomes a Competitive Moat

The RTCT initiative places a premium on something that the health AI industry has often treated as secondary to model performance: data infrastructure that is auditable, interoperable, and designed for real-time regulatory transmission. Paradigm Health's selection as the technical platform for the FDA's proof-of-concept trials is instructive. Paradigm was not chosen because it had the most sophisticated AI model for safety signal detection. It was chosen because it had built an infrastructure — the SPIRE platform — capable of ingesting EHR data directly, evaluating pre-specified criteria algorithmically, and transmitting only the relevant signals in a format that FDA reviewers could act on, with a traceable audit trail protecting patient privacy.

For health AI companies operating in or adjacent to clinical trials, this signals a clear direction: the companies that will capture the value of real-time trial modernization are the ones that invest in validated, interoperable, regulatory-grade data infrastructure, not just in the accuracy of their algorithms. Model performance is table stakes. Infrastructure trustworthiness is the differentiator.

Pre-specification and Human Oversight Remain Non-Negotiable

The FDA's approach to real-time trials is deliberately bounded. Walsh and his team spent months before the announcement working with AstraZeneca and Amgen to pre-determine the specific safety signals and endpoints that would be transmitted in real time. This pre-specification is not an administrative formality — it is the mechanism that prevents real-time access from becoming information overload. It reflects the same principle that effective clinical AI governance applies to algorithmic decision support: the system should surface the pre-defined signals it was designed to detect, not generate open-ended outputs that require unstructured human interpretation.

This is a meaningful constraint for AI companies positioning their tools as real-time trial monitoring solutions. Systems that generate alerts, flags, or signals outside a pre-specified, regulator-approved framework will not fit this model. The FDA's framework rewards narrow, validated, pre-agreed AI outputs — not general-purpose AI analysis applied to trial data in real time.

The Decentralized Trial Opportunity

The RTCT initiative has natural synergies with the accelerating adoption of decentralized clinical trials (DCTs) — trials in which patients are enrolled and data collected remotely, reducing the reliance on physical site visits. Real-time data transmission from EHRs and connected devices is central to both paradigms. Companies building AI-powered infrastructure for decentralized trials are building toward the same technical requirements that the RTCT framework demands: direct data capture from patient records, algorithmic evaluation of pre-specified criteria, and real-time transmission to sponsors and regulators.

The FDA's explicit interest in AI tools that improve patient recruitment in clinical trials is also relevant here. Recruitment failure — the inability to enroll sufficient patients within trial timelines — is one of the most common causes of trial delays and failures. AI-powered patient identification, matching, and outreach tools, built on compliant access to real-world data, address a problem that the RTCT framework does not itself solve. Companies combining real-time data infrastructure with intelligent recruitment capabilities will be well positioned as the pilot program expands.

The Regulatory Relationship Is Changing

Perhaps the most underappreciated implication of the RTCT initiative is what it signals about the evolving relationship between FDA and trial sponsors. Real-time data sharing requires unprecedented transparency — sponsors cannot control the information environment in the way they have historically done through curated submissions. FDA reviewers will see safety signals and efficacy data as they emerge, not in the shaped narrative of a formal submission package.

This changes the nature of the regulatory conversation. Companies must be prepared for a dynamic, ongoing dialogue with the FDA during trial execution rather than a structured exchange at defined regulatory milestones. For companies with genuinely clean safety profiles and strong efficacy signals, this transparency is a feature. For those whose trials reveal unexpected signals, the implications require careful thought. Either way, the era of the 66-million-page submission as the primary mode of FDA communication may be ending.

"Real-time trials have been talked about for years. We demonstrated that it is not only possible, but also potentially transformative for the clinical trials ecosystem." — Jeremy Walsh, FDA Chief AI Officer

Open Questions and Challenges

Data Standardization Across Sites and Systems

The TRAVERSE trial is being conducted at two of the most technically sophisticated academic medical centers in the United States — MD Anderson and Penn Medicine. Their EHR systems, data governance frameworks, and IT infrastructure are not representative of the broader clinical trial site ecosystem, which includes community hospitals, regional cancer centers, and outpatient clinics with varying levels of digital maturity. As the RTCT initiative scales from proof-of-concept to a broader pilot, the FDA and its technology partners will need to address data standardization across a much more heterogeneous set of sites.

The choice to align the pilot program with NIST's AI risk management framework is relevant here. Standardized risk governance is a prerequisite for scaling AI-enabled data infrastructure across diverse institutions with different risk tolerances, regulatory environments, and technical capabilities.

Patient Privacy and Data Security

Real-time transmission of clinical trial data from EHRs to a cloud platform with FDA access raises legitimate patient privacy questions. Paradigm Health has emphasized that its platform minimizes the transfer of unnecessary datasets and maintains patient privacy protections. But as the program scales, the governance frameworks for data consent, de-identification, access controls, and breach response will need to be both technically robust and publicly defensible.

This is not a reason to slow the initiative — the current model of months-delayed paper submissions does not provide meaningfully superior privacy protection. But it is a domain where regulatory clarity and public transparency will be essential to maintaining the trust of patients, whose participation makes clinical research possible in the first place.

Equity and Geographic Access

One of the more persistent critiques of the clinical trial ecosystem is its concentration in major academic medical centers in a small number of metropolitan areas, leaving patients in rural communities and underserved populations with limited access to cutting-edge trials. The promise of decentralized, technology-enabled trials has always been partly about expanding geographic access. Whether the RTCT initiative accelerates or inadvertently reinforces site concentration will depend on the deliberate choices made in designing the broader pilot program — choices that the public comment process, open through May 29, 2026, is designed to inform.

Conclusion: The Signal, Not the Noise

The FDA's real-time clinical trial initiative is easy to read as a story about technology — about AI platforms, cloud data transmission, and algorithmic signal detection. It is those things. But it is more fundamentally a story about a regulatory institution recognizing that the information architecture it has operated under for sixty years is no longer adequate to the pace of biomedical discovery or the urgency of patient need.

The TRAVERSE and STREAM-SCLC trials are small experiments. The summer pilot that follows them will be larger but still bounded. The full realization of continuous clinical development — real-time, phase-agnostic, perpetually transparent — is likely a decade or more away. But the direction is now clearly set, and the proof of concept is established.

For health AI companies, the implications are both practical and strategic. Practical: the technical and governance requirements of real-time regulatory review are becoming a new standard for AI infrastructure in clinical settings. Strategic: the companies that help the FDA and the biopharmaceutical industry navigate this transition — by building the infrastructure, governance frameworks, and validated tools that make real-time trials trustworthy at scale — will be among the defining players of the next era of drug development.

At Gesund.ai, we build with the conviction that AI in healthcare is only as valuable as the governance and validation infrastructure that surrounds it. Real-time clinical trials are the clearest demonstration yet that this conviction is not a constraint on innovation — it is the condition for it.

For inquires: please contact us.

Bibliography

1. Lawrence, L. (2026, April 28). FDA testing speedier drug development with real-time clinical trials. STAT News.

2. U.S. Food and Drug Administration. (2026, April 28). FDA announces major steps to implement real-time clinical trials. FDA News Release. fda.gov.

3. Paradigm Health. (2026, April 29). Paradigm Health announces collaboration with the FDA to transform regulatory review of clinical trial data. BioSpace / PRNewswire. biospace.com.

4. Clinical Trials Arena. (2026, April 29). FDA launches pilot for real-time clinical trials. clinicaltrialsarena.com.

5. Fierce Biotech. (2026, April 28). FDA unveils plan for real-time review of clinical trial data, with AstraZeneca and Amgen already on board. fiercebiotech.com.

6. Pharma Phorum. (2026, April 29). Amgen, AZ will pilot FDA's real-time clinical trial plan. pharmaphorum.com.

7. Axios. (2026, April 29). FDA to use AI to track clinical trials in real time. axios.com.

8. Globe Newswire. (2026, April 28). FDA announces major steps to implement real-time clinical trials. globenewswire.com.

9. US News / HealthDay. (2026, April 29). FDA moves to real-time clinical trial patient monitoring, faster drug review. usnews.com.

10. U.S. Food and Drug Administration. (2026). Request for Information: AI-enabled optimization of early-phase clinical trials pilot program. Federal Register. federalregister.gov.

11. National Institute of Standards and Technology. (2023). AI Risk Management Framework (AI RMF 1.0). nist.gov.

12. DiMasi, J.A., Grabowski, H.G., & Hansen, R.W. (2016). Innovation in the pharmaceutical industry: New estimates of R&D costs. Journal of Health Economics, 47, 20-33.