HHS unveils efforts to speed up clinical trials with AI

Kevin Carter/Getty Images

The department announced a broader multi-agency effort to support U.S. clinical trial infrastructure this past summer.

The Health and Human Services Department announced Tuesday a program and three projects aiming to speed up clinical trials with artificial intelligence, advanced computational models, real-time analysis and other technologies.

The Simulation-augmented, Real-time Platform Adaptive Seamless Trials — or SURPASS — program seeks to create AI-powered platforms that continuously and adaptively support clinical trial designs, according to an HHS press release. The platform will enable “earlier, smarter decision making that will help speed effective drugs to patients,” according to the release. 

The department’s Advanced Research Projects Agency for Health, or ARPA-H, will host two informational sessions before a SURPASS solution summary is due on Nov. 30.

Companies interested in the SURPASS program will have to focus on three technical areas. 

The first is building a phaseless design engine that simulates clinical trials so that researchers can create faster trials with fewer patients while complying with regulations, according to an HHS release and a SAM.gov listing. 

The second is developing a continuous inference engine that allows researchers to conduct real-time or on-demand analysis of clinical trials, supporting trial adaptations. 

The third is creating an agentic AI operations layer that will automate trial startup activities, speed up data collection and support other procedures.

“SURPASS is developing the technology and biostatistical frameworks needed to analyze trial data as it accumulates, allowing teams [to] learn in real time, adapt when needed, and generate stronger evidence faster with less operational burden,” Daria Fedyukina, SURPASS program manager, said in the release.

HHS also unveiled STACK, COMMONS and CINCH, three complementary ARPA-H initiatives meant to transform U.S. clinical trials.

STACK will create more domestic clinical trial sites and boost trial enrollment by using AI “to accelerate clinical site activation and convert research-naïve sites into clinical trial sites,” according to the HHS release. Evidence Health is the prime STACK awardee and could receive up to $41.18 million, according to an ARPA-H listing.

COMMONS will build “the connective tissue for nationwide data infrastructure” by creating “architecture for consent at national scale and enable access to regulatory-grade data,” according to the HHS release. The project targets over 40 million individuals and 3 million linked records, according to an ARPA-H listing. Evidence Health is the prime COMMONS awardee and could receive up to $49.9 million, the listing says.

University of Texas at Austin Dell Medical School is partnering with Evidence Health on STACK and COMMONS. The medical school is preparing sites for clinical trial research and developing a national research data resource, according to a press release.

CINCH aims to help cancer patients contribute their own data to speed up the identification of clinical trials that may be appropriate for them, according to the HHS release and an ARPA-H listing. CINCH is meant to operate as a patient-initiated way for adults to better understand their conditions. Courage Health is the prime CINCH awardee and could receive up to $8.95 million, the listing says.

The HHS initiatives build on other Trump administration efforts to speed up clinical trials.

HHS announced Operation TrialBlazer in June, a department-wide effort to support clinical trials through pilot programs, public-private partnerships and patient connections. 

The Food and Drug Administration announced in April that it is using AI and cloud computing to monitor clinical trial data in real-time under a pilot program. An FDA official estimated that the effort could cut overall clinical trial times by as much as 40% without compromising patient safety.

HHS says the clinical development of drugs and biologics often takes over a decade, costs an average of up to $2 billion and fails more than 90% of the time.