Two drug sponsors are now sending trial endpoints and safety signals to their regulator while the studies are still running.
In late April the FDA announced it had initiated two proof-of-concept real-time clinical trials. AstraZeneca is running TRAVERSE, a Phase 2 study in treatment-naive mantle cell lymphoma. Amgen is running STREAM-SCLC, a Phase 1b study in small cell lung carcinoma. Signals reach the agency through the platform provider Paradigm Health, and the FDA states it has received and validated signals from the AstraZeneca study, establishing that the technical framework works (FDA, April 2026). Commissioner Marty Makary described regulators being able to see "in the cloud in real time exactly what is happening" (RAPS, April 29, 2026).
The agency said it would publish final selection criteria in July and complete selections for a broader pilot in August. As of July 23, neither the final criteria nor additional participants had been published (CASRAI, July 23, 2026). The window is open now, and what it asks for is not an IT capability.
The prize is cycle time, and cycle time is capital
Clinical development is the rare operation where time, not headcount, is the dominant line item. Tufts researchers put the cost of a single delay day at roughly $500,000 in lost sales for an approved product, with Phase II and III trials running at roughly $40,000 per day in direct cost (Smith, DiMasi and Getz, 2024). Those are averages across 645 launched drugs and 409 trial budgets, not a figure for any one program. Read them as an order of magnitude, not a forecast.
A continuous feed compresses two distinct delays. The first is between data capture and the sponsor seeing it. Most large sponsors have already spent a decade attacking that one. The second is between the sponsor reaching a conclusion and the regulator accepting it. No internal program can buy that. It is the one this pilot puts in play.
A real-time feed does not reward the sponsor with the fastest pipe. It rewards the sponsor whose data is stable enough to be read while it is still moving. Hikari Blue · operator note
The buffer you are giving up
Clinical data operations are built around a buffer. Data is captured, queried, reconciled, cleaned and locked. Only then does anyone outside the sponsor see it. That interval is where artifacts get resolved in private: a mis-keyed lab value, a site reporting in the wrong unit, a coding error that looks like a safety signal for two days and then does not.
Continuous reporting removes the buffer. Every artifact becomes a conversation with your regulator at the moment it appears. A signal transmitted on Tuesday and withdrawn on Thursday is not a neutral event, and no sponsor gets to run that experiment quietly.
So speed here is a governance property, not a plumbing property. The sponsors who gain time will be the ones who moved data quality control upstream to the point of capture, and who can separate a signal from an artifact in hours rather than at database lock. The rest will spend the time they gained explaining noise.
Your monitoring layer now faces a regulator
At this cadence, review stops being a periodic human task. The FDA frames the pilot as an examination of how artificial intelligence and data science can improve trial efficiency, strengthen safety monitoring, inform dose selection and support earlier go or no-go decisions (FDA, April 2026). The agency appointed a Chief AI Officer, Jeremy Walsh, who is named in the initiative. Detection and triage on the sponsor side move the same way, for the same reason.
That puts a sponsor's own monitoring models in front of a regulator, continuously. Validation, versioning, lineage and drift on the monitoring stack stop being internal quality concerns. If a model classifies an adverse event differently in September than it did in June, someone has to state why in one sentence, and be able to show which version produced which signal. This is the ordinary discipline of a governed operating layer, applied to a system that now has an external reader.
One caveat before anyone briefs a steering committee. As of late July the FDA had not published data formats, transmission cadence or the specific data standards involved. The two proof-of-concept studies were not selected through a published application process, so their design is a weak guide to what the pilot will require (CASRAI, July 23, 2026). Verify the current status directly with the agency before committing a program.
Two ways to get this wrong
Wait for the specification, and arrive with a pipeline built to a standard everyone else already implemented. Or rush a feed into place on top of a quality process that still assumes a private interval, and hand your regulator your artifacts at full speed.
The capability worth building is narrower than either. It is the ability to state, at any hour of a running study, which numbers are decision-grade and which are still settling, and to show the difference on evidence rather than on assurance.
What to examine this month
- Whether capture-to-signal latency is measured at all today. Most sponsors measure database lock, which is a different number and a later one.
- Which monitoring rules are deterministic and which are model-based, and whether you could show six months later which version produced a given signal.
- What threshold defines a reportable signal, in writing, before any feed opens. The FDA met each proof-of-concept sponsor to set criteria for reporting signals in real time (FDA, April 2026). That negotiation is the deliverable, not the pipe.
- Whether a program team can act on 48-hour-old data, and who signs when the answer changes the dose.
Track one number: the rate at which transmitted signals are later revised. Zero is not the target and is not achievable. An unmeasured rate is the problem.
The question to bring to the next executive session
If our regulator could see this study today, what would we be explaining rather than reporting?
The pilot picks its participants this month. The capability it asks for, evidence that holds up while it is still moving, decides R&D cycle time whether or not you are selected.
Sources
- U.S. Food and Drug Administration (April 2026). FDA Announces Major Steps to Implement Real-Time Clinical Trials. Primary source for the two proof-of-concept studies (AstraZeneca TRAVERSE in treatment-naive mantle cell lymphoma, Amgen STREAM-SCLC in small cell lung carcinoma), for signals being received and validated through Paradigm Health, for the meetings with each sponsor to establish real-time signal reporting criteria, for the artificial intelligence and data science objectives of the pilot, and for the July criteria and August selection timeline. fda.gov/news-events/press-announcements/fda-announces-major-steps-implement-real-time-clinical-trials
- Regulatory Affairs Professionals Society (April 29, 2026). FDA announces efforts to advance real-time clinical trials. Source for the announcement date, for the role of Paradigm Health across both studies, for the quoted remark by Commissioner Marty Makary, and for the naming of Chief AI Officer Jeremy Walsh in the initiative. raps.org/resource/fda-announces-efforts-to-advance-real-time-clinical-trials.html
- Smith, Z. P., DiMasi, J. A. and Getz, K. A. (2024). New Estimates on the Cost of a Delay Day in Drug Development. Therapeutic Innovation & Regulatory Science, 58(5), 855 to 862. Peer-reviewed source for the approximately $500,000 per delay day in lost sales, for the approximately $40,000 per day direct cost of Phase II and III trials, and for the sample of 645 launched drugs and 409 trial budgets. link.springer.com/article/10.1007/s43441-024-00667-w
- CASRAI (July 23, 2026). FDA Real-Time Clinical Trials (RTCT) Pilot 2026: Where Real-Time Trials Stand. Source for the status as of late July: final selection criteria not published, no additional participants named, technical specifications for data formats, transmission cadence and data standards not made public, and the two proof-of-concept studies not having been selected through a published application mechanism. Status source, dated. Verify current status before acting. casrai.org/news/fda-real-time-clinical-trials-rtct-pilot-2026
- Fortrea (June 9, 2026). FDA's Real Time Clinical Trials (RTCT) Initiative: Key Implications for Sponsors. Contract research organisation analysis, cited as an industry view rather than as evidence. Corroborates the shift from periodic reporting cycles to continuous regulatory visibility and the resulting demands on data traceability, signal definitions and governance. fortrea.com/insights/fda-real-time-clinical-trials-rtct-initiative
The Hikari Blue team · Austin, August 2026