On Wednesday, Sept. 30, the Department of Health and Human Services announced SURPASS, an ARPA-H program to develop AI-supported tools for adaptive clinical trials, alongside three related projects focused on trial sites, research data and cancer-trial access. HHS says the initiatives are intended to reduce delays and patient burden while preserving rigorous evidence. The plans, reported by Nextgov, are not results: no evidence has been reported that SURPASS has shortened a trial.
SURPASS sits within Operation TrialBlazer, HHS’s broader clinical-research agenda, announced June 22. The distinction between the department’s goals and demonstrated results is central: the effort’s success depends on whether it can save time without weakening evidence about safety and effectiveness.
How is SURPASS meant to change clinical trials?
SURPASS aims to develop continuous, adaptive platform trials using predictive computational models, shared infrastructure, common control groups and real-time analysis. In an adaptive trial, researchers can adjust parts of a study as data accumulates.
ARPA-H describes three work areas: a simulation-based engine to design trials, an analysis engine to support adaptations as studies proceed, and an operations layer intended to automate tasks such as getting trials started and preparing data.
HHS presents the program as a response to the time and cost of developing drugs and biologics. The department says development often takes more than a decade, costs up to $2 billion on average and has a failure rate above 90%. Its announcement did not provide underlying data for those figures.
What other initiatives sit under TrialBlazer?
The September announcement paired SURPASS with three projects:
- STACK aims to expand the capacity of clinical-trial sites.
- COMMONS aims to build national research-data infrastructure governed by participant consent. ARPA-H lists targets of more than 40 million individuals and 3 million linked records; these are goals, not achieved totals.
- CINCH aims to help adults with cancer contribute data and find potentially suitable trials.
Operation TrialBlazer brings together work involving FDA, NIH, ARPA-H, the National Cancer Institute, the Office of the National Coordinator for Health Information Technology and the HHS Office of Inspector General. HHS has described plans for research tools, trial activation and enrollment, and connections between patients and studies through electronic health records. The initiatives address different potential obstacles, but whether they lead to faster enrollment or better evidence remains unproven.
What trial tools and rules already exist?
SURPASS is distinct from an FDA real-time monitoring pilot reported in April. That effort involved planned studies by AstraZeneca and Amgen, with monitoring through a direct data feed. FDA Chief AI Officer Jeremy Walsh described a potential reduction of as much as 40% in overall trial time. That was an estimate of an opportunity, not a measured result; the report did not say the studies had finished faster.
Free newsletter
Get the morning briefing
Start each day with the AI stories that matter and why — a short, free email from our newsroom.
FDA’s December 2019 guidance already covers adaptive designs for drug and biologic trials, including designs using Bayesian methods or computer simulations. It addresses trial design, conduct, reporting and information relevant to regulatory evaluation. The guidance predates SURPASS and is not an endorsement or validation of the program.
What evidence would show SURPASS makes trials faster?
Evaluations would need to compare actual trial timelines with an appropriate baseline and show where time changed: startup, enrollment, analysis or follow-up. They should also explain which adaptations were planned and how researchers assessed them, and report evidence about safety and effectiveness. Faster data analysis alone would not establish that a trial produced reliable results.
The announcement and reporting so far provide no completed SURPASS results or independent evaluation of its effects. Nextgov reported that SURPASS solution summaries are due Nov. 30, a development milestone rather than evidence of time saved. The meaningful test will be whether later evaluations document faster trials alongside the quality of the evidence they produce.