When a patient in an early-phase drug trial experiences a serious side effect today, weeks or even months can pass before that information reaches the FDA through the standard cycle of batched sponsor submissions and retrospective reviews. The agency wants to close that gap. In April 2026, the FDA announced it is moving toward real-time tracking of clinical trial data, with two proof-of-concept studies already running and a broader pilot program planned for summer 2026.
The shift could reshape how regulators spot dangerous reactions and unexpected benefits during the earliest, riskiest stages of drug development, the period when patients are most exposed to unknowns.
Two trials are already testing the concept
The FDA named two active studies as its initial proving ground. The first is AstraZeneca’s TrAVeRse study (NCT05951959), which is evaluating a three-drug combination of acalabrutinib, venetoclax, and rituximab in patients with previously untreated mantle cell lymphoma. The second is Amgen’s STREAM-SC trial. Both are designed to test whether continuous data feeds between trial sponsors and the FDA can replace the slower rhythm of periodic submissions.
Unlike conventional trials, where regulators typically see results only at pre-scheduled milestones, these studies funnel safety and efficacy data to the agency as it is collected. The technical infrastructure powering that exchange was built by Paradigm Health, which announced its collaboration with the FDA in early 2026. The company described itself as the conduit for real-time data exchange, handling reporting protocols, validation, and the flagging of what the agency calls “key regulatory events,” a category that covers adverse reactions and other safety-relevant milestones.
The FDA’s existing Safety Reporting Portal at HHS serves as part of the broader digital backbone, providing a centralized node where sponsors already submit electronic safety reports. Integrating that portal with the new real-time feeds is a core piece of the engineering challenge.
A summer pilot and an AI framework
On April 29, 2026, the FDA published a Request for Information in the Federal Register (docket FDA-2026-N-4390, 91 FR 23100) soliciting industry input ahead of a wider pilot. The program’s formal title, “AI-Enabled Optimization of Early-Phase Clinical Trials,” signals the agency’s intent to use machine learning and automated analytics to process safety and efficacy signals as they arrive, not after a trial phase wraps up.
Separately, the FDA released draft guidance titled “Considerations for the Use of Artificial Intelligence To Support Regulatory Decision-Making for Drug and Biological Products.” The document lays out a risk-based framework covering when and how AI tools can generate information that feeds into decisions about drug safety, effectiveness, and quality. It includes definitions, context-of-use categories, and expectations for documentation and validation. Because it remains a draft open to public comment, its provisions could still shift before they become binding.
What this means in practice
In concrete terms, real-time tracking means a trial site recording a serious adverse event on a Monday could have that data visible to FDA reviewers the same week, rather than waiting for the next scheduled data submission, which under current norms might be months away. If the system works as designed, regulators could issue protocol amendments, partial clinical holds, or patient safety notifications far sooner than the existing process allows.
For patients, the promise is straightforward: faster detection of harm and, potentially, faster confirmation that a therapy is working. For drug developers, the incentive is a tighter feedback loop that could reduce the time and cost of early-phase trials by catching problems before they compound.
Open questions the FDA has not yet answered
The public record is strong on intent but thin on operational detail. Several gaps stand out.
First, the Amgen STREAM-SC trial lacks publicly available registry details comparable to AstraZeneca’s TrAVeRse study. Specific endpoints, enrollment figures, and timelines have not been disclosed, making it hard to assess whether both proof-of-concept efforts are at similar stages.
Second, the FDA has not published any early results or progress reports from either trial. The Request for Information is a call for input, not a status update, so there is no official assessment yet of whether real-time feeds have actually produced actionable safety signals faster than conventional reporting. The speed advantage remains a design goal, not a demonstrated outcome.
Third, the two initial sponsors are among the world’s largest pharmaceutical companies, each with deep data infrastructure and large regulatory affairs teams. The Request for Information does not appear to include targeted provisions for smaller biotech firms that may lack the technical capacity to build real-time data pipelines. If participation demands the kind of engineering Paradigm Health has provided, the cost of entry could effectively limit the pilot to well-resourced companies.
Fourth, the FDA has not detailed whether sponsors will need to use specific data formats or standardized ontologies. Without interoperability standards, AI tools may struggle to compare safety signals across trials or drug classes, potentially limiting the benefits of continuous monitoring.
Why the draft AI guidance matters for the pilot’s future
The draft AI guidance is not a side document. It will define the validation and documentation standards that any AI tool must meet before it can support a regulatory decision about a drug. If the final version tightens those requirements, some tools currently in development may not qualify. If it loosens them, the door opens wider but so do concerns about algorithmic bias and error.
Patient advocates, academic researchers, and industry groups are all expected to weigh in during the comment period. The eventual shape of the guidance will determine not just which AI tools are eligible but how much trust regulators place in automated safety signals versus traditional human review.
For now, the evidence supports a clear but measured reading: the FDA has committed to experimenting with real-time clinical trial oversight, has enlisted major pharmaceutical sponsors and a technology partner to test the concept, and has laid the regulatory groundwork through a formal rulemaking process. Whether the infrastructure can scale beyond a handful of large-company trials, whether AI tools will improve signal detection without introducing new errors, and whether patients will ultimately see earlier, safer access to new therapies are questions the summer 2026 pilot is designed to begin answering.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
