The Pulse
HHS Launches SURPASS to Redesign Clinical Trials With AI
HHS announced SURPASS, an ARPA-H program intended to use computational models, shared infrastructure and real-time analysis to change how clinical trials are designed and run. ARPA-H lists October 12, 2026, as the webinar registration deadl
AI.info Team · Written with AI ·
The clinical development challenge
More than 90% of drug candidates fail, and clinical development often takes over a decade, according to the Department of Health and Human Services. The department says the process can cost up to $2 billion on average. On September 30, HHS announced SURPASS, an ARPA-H program designed to use computational models, real-time data analysis and automation to change how trials are designed and run.
SURPASS stands for Simulation-augmented, Real-time Platform Adaptive Seamless Trials. It is meant to combine predictive models, shared trial infrastructure, common control groups and ongoing analysis, so researchers can adjust studies as evidence accumulates rather than wait for each conventional phase to finish. The program aims to evaluate drugs and biologics faster and with fewer participants; the announcement describes goals, not results from a working system.
“Today's clinical trial system often requires too many stops, too much duplicated infrastructure, and too much time before researchers can understand whether a trial is on the right track,”
Daria Fedyukina, Ph.D., SURPASS program manager, said.
She said the program is developing tools and statistical methods for teams to analyze accumulating trial data and adapt studies as needed.
Three technical tracks shape the trial redesign
One track would use “digital twins” and other predictive models to simulate clinical and operational outcomes before a trial begins. ARPA-H calls this a phaseless design engine: a system intended to help researchers design studies with fewer patients while building evidence that could support regulatory confidence in the approach.
A second track would develop an inference engine for continuous or on-demand analysis. It is intended to allow quicker changes to a trial and reduce reliance on large conventional control groups while maintaining rigorous evidence standards. The third track would automate parts of trial operations, including study startup, adding treatment arms, collecting and cleaning data, and preparing datasets.
The ARPA-H program page frames the ambition as a question: whether clinical trials that now take more than a decade could be shortened to less than four years. That is a target for the proposed work, not a timeline HHS says the new program has already achieved. The agency has not announced trial results from SURPASS.
STACK, COMMONS and CINCH target barriers outside study design
HHS paired SURPASS with three projects aimed at problems that can keep a trial from reaching participants or using data effectively. STACK plans to use AI to speed clinical-site activation and prepare sites without prior research experience to run studies. Its stated aim is to expand trial capacity and bring research opportunities closer to patients.
COMMONS is intended to create a privacy-focused system for consent at national scale and improve access to data suitable for regulatory use. CINCH focuses on patient data and care navigation: it aims to help people contribute real-world information, coordinate care and find trials that may suit them. Together, the projects address site capacity, fragmented data and the effort involved in connecting patients with research.
Those are planned functions, not evidence that the projects have already expanded enrollment or improved trial access. HHS says the work is intended to generate reusable tools, standards, regulatory documents and demonstrations that could inform other sponsors, research sites and regulators.
ARPA-H sets a January 2027 proposal deadline
ARPA-H has opened a funding opportunity for SURPASS and is seeking proposals from teams spanning statistics, AI, clinical-trial design, operations and regulation. Its program page lists an informational webinar for October 15, 2026, with registration due October 12, and a Proposers’ Day event for November 6, with registration due October 28. Solution summaries are due November 30; teams must submit one and await feedback before proceeding to a full proposal.
Full proposals are due January 22, 2027, according to the ARPA-H program page. The next test is whether teams can turn the proposed models and shared infrastructure into trial designs that regulators can assess and patients can join, including at sites that have not previously conducted research.