U.S. Launches SURPASS to Reshape Clinical Trials Through Simulation and Real-Time Analysis

Before a new therapy can reach patients, it must demonstrate safety and efficacy and navigate time-consuming steps such as setting up trial sites, recruiting participants, and organizing data. On September 30, the U.S. Department of Health and Human Services (HHS) launched SURPASS through the Advanced Research Projects Agency for Health (ARPA-H), aiming to combine computer simulation, real-time analysis, and shared research facilities so that clinical trials can continually learn as they proceed and reduce the burden of starting over at each stage.

The trial platform envisioned by SURPASS could operate continuously, accommodate new treatment arms, and allow multiple therapies to share appropriate control groups and infrastructure. Research teams would adjust trial arrangements based on accumulating data. This design seeks to generate evidence sufficient to support decisions earlier; shorter timelines, lower costs, and fewer required participants remain goals that need to be validated.

Predictive models would be used specifically to simulate possible clinical outcomes and operational workflows before a trial begins, helping teams select study designs, with tools including digital twins. Another technical focus is continuous statistical inference: analyzing data as they are updated while maintaining the statistical reliability of conclusions. ARPA-H also hopes to automate some of the work involved in trial startup, data cleaning, and adding new treatment arms, allowing flexibility in design to be implemented in everyday research.

However, changes in trial methods still require clinical settings capable of implementing them. STACK, announced at the same time, will help more sites develop research capabilities, giving patients opportunities to participate near their homes. COMMONS aims to establish a national data framework grounded in privacy and consent mechanisms to support data use that meets regulatory requirements. The third program, CINCH, focuses on patients, helping them contribute their own real-world data, improve care coordination, and find suitable trials.

Dell Medical School at The University of Texas at Austin will serve as the primary academic partner for STACK and COMMONS and will be the initial institution providing data and testing for COMMONS. The university explained that validation will examine whether integrated data can reproduce existing research findings and reliably identify patients who may qualify for trials. This gives data sharing a concrete direction for evaluation, beyond simply bringing scattered records together.

The program has also entered the proposal solicitation stage. According to the schedule published by ARPA-H, the SURPASS solicitation number is ARPA-H-SOL-26-164. Required solution summaries must be submitted by November 30, 2026, and full proposals are due January 22, 2027; an online information session and a proposer engagement event are scheduled for October 15 and November 6, respectively. STAT described SURPASS as a five-year program seeking participation from multidisciplinary teams spanning statistics, artificial intelligence, trial design, operations, and regulation.

The announcement has not yet provided the scale of the training data for SURPASS models, their predictive performance, or completed clinical validation results. STAT also noted that the initial announcement did not disclose funding for the individual initiatives or whether participants would receive greater flexibility or resources. The next key step is to demonstrate that simulation and continuous analysis can preserve credible evidence while changing trial workflows and earn regulatory trust; the program’s launch alone cannot be taken as evidence that new drug development has already become faster.