FDA pilots real-time clinical trials

Clinical trials usually run in bursts. Sites enroll patients, data managers clean records, sponsors package updates, and regulators see the results later. The FDA is trying to break that rhythm. On April 29, the agency said it has already started two proof-of-concept real-time clinical trials, with AstraZeneca and Amgen feeding safety and efficacy signals to FDA reviewers as the studies run. (fda.gov) ### What is “real-time” here? It does not mean the FDA is peeking at every patient chart second by second. Basically, it means trial data and endpoint signals move through cloud systems to the agency continuously or near-continuously instead of waiting for the usual scheduled package submissions. The FDA framed this as a new review model for early-phase studies, especially where quick safety reads and go/no-go calls matter. (fda.gov) ### What exactly launched? Two things. First, the FDA said two proof-of-concept trials are already underway. Second, it released a Request for Information on a broader pilot program that it plans to launch this summer. The agency wants feedback on how cloud platforms and AI-enabled tools could improve efficiency, speed, and decision quality in early clinical development. (fda.gov) ### Which trials are in the first wave? AstraZeneca and Amgen are the named sponsors. AstraZeneca’s study is a Phase 2 trial in previously untreated mantle cell lymphoma. Amgen’s is an early-stage lung carcinoma trial, described in coverage as Phase 1b. Paradigm Health said both are using its Study Conduct platform, and that sites in the AstraZeneca study include MD Anderson and the University of Pennsylvania. (abcnews.com) ### Has the FDA actually received live data yet? Yes — and that is the part that makes this more than a concept slide. The FDA said the trials have been successfully initiated, and outside coverage added that the agency has already received and validated signals from the AstraZeneca study. That matters because the hard part is not announcing a pilot. The hard part is ge(abcnews.com)fda.gov) ### Why does the agency want this? Because the current handoff model is slow. Reviewers often see important trends only after data has been compiled, checked, and formally submitted. FDA Commissioner Marty Makary pitched real-time trials as a way to compress development timelines, and FDA AI chief Jeremy Walsh said the approach could cut 20%, 30%, or ev(fda.gov)ial phases and regulatory feedback. (biospace.com) ### Where does AI fit in? Mostly in triage and pattern-finding, not in replacing trial science. The FDA’s RFI asks how AI-enabled technologies could help reviewers process incoming data faster and improve decision quality. In practice, that likely means flagging anomalies, surfacing safety signals, and helping reviewers navigate messy operational data. But the agency is still in pilot mode, so the exact AI workflow is not locked yet. (fda.gov) ### What is the catch? Live data feeds create new operational burdens. Sponsors need streaming ingestion, provenance, audit trails, and tightly governed analytics so everyone can trust what the FDA is seeing. A real-time pipe is only useful if the underlying data is standardized, traceable, and interpretable. Otherwise you just move confusion faster. That is why this could matter as much for clinical data infrastructure vendors as for drug developers. (clinicalresearchnewsonline.com) ### So what changes if this works? Early oncology trials could become less stop-and-start. Regulators could spot problems sooner, sponsors could adjust faster, and promising programs might move between phases with less waiting. The bigger shift is cultural — the FDA would act less like a recipient of finished trial packages and more like a live participant in the evidence flow. That is a very different operating model for drug development. (fda.gov) The bottom line is simple. The FDA did not just announce an idea. It started two real trials, named AstraZeneca and Amgen as the first sponsors, and asked industry how to scale the model this summer. If the pilot holds up, clinical development starts looking less like quarterly reporting and more like a monitored data stream. (fda.gov)