FDA to use AI to track clinical trials in real time
Add Axios as your preferred source to
see more of our stories on Google.
Illustration: Gabriella Turrisi/Axios
The Food and Drug Administration on Tuesday kicked off an effort to use AI and other data science tools to monitor clinical trials in real time and cut down review times for new drugs.
Why it matters: Agency leaders portrayed the move as a major step toward keeping U.S. biomedical research competitive against countries like China.
Driving the news: The agency announced the launch of two "proof of concept" real-time trials to view safety and efficacy signals for an AstraZeneca drug used to treat lymphoma and an Amgen drug for small cell lung carcinoma.
- It also solicited public comments for a broader pilot program for real-time trials that could launch this summer. It will be aligned with an AI risk management framework developed by the National Institute of Standards and Technology.
- The process could circumvent what officials call a bottleneck in drug development, in which results from trial sites are reported to manufacturers, who then analyze and submit data to the FDA.
What they're saying: "Today is a milestone day for us to challenge the assumption that it takes 10 to 12 years for a new drug to come to market," FDA commissioner Marty Makary told reporters.
- On average, 45% of the time between a Phase 1 clinical trial and submission of an application to the FDA is "dead time" spent on paperwork and other administrative tasks, he said.
- The new initiative seeks to cut down on that time to move the process faster "without cutting any corners on safety," he said.
FDA chief AI officer Jeremy Walsh told reporters "there is opportunity to shave off" as much as "20, 30, 40% of an overall clinical trial time."
- Officials said they've validated signals for AstraZeneca's trial through Paradigm Health, confirming the feasibility of real-time signal sharing.
- The agency is seeking public input on questions like which clinical trial issues might benefit most from the application of AI, and if priority should be given to specific uses for AI, such as patient recruitment and safety monitoring.
