The advance of artificial intelligence tools presents a growing signal to noise challenge for postmarket surveillance as the US Food and Drug Administration and industry try to sort through the reams of information reviewed by computers.
Drug sponsors and regulators must determine how to sort the potential adverse events that will be gathered, as well as which require more exploration and action.
“Many companies right now are looking at the social media channels that they own. Are complaints showing up? Are adverse events showing up?” Coleen Glessner, executive VP quality and ethics and compliance at Emergent BioSolutions, Inc. said 15 May at the Food and Drug Law Institute Annual Meeting. “With these tools that opens a whole host of opportunities to look far more broadly, to look at every social media channel … It really gives you power to search far more broadly than just the channels that you are on. But there’s a tremendous risk there.”
Ideally, there would be tools that help not just find the data but also make sense of it, she said.
“Imagine the volumes of data that are going to come in,” Glessner said. “Imagine the types of things that get posted … So imagine how do you monitor and oversee that. That gets right to how do you not have a signal-to-noise ratio problem?”
Social media monitoring has long presented a challenge for industry given the quantity of information sponsors may be required to report to regulators and the slow-to-come agency guidance on the topic that has generally not kept pace with advances in technology.
FDA expects manufacturers to report adverse events documented on sites it sponsors, not third-party sites, but which sites a firm “sponsors” remains somewhat ambiguous. (Also see "Social Media “Revolution” Expected In Adverse Events Signal Detection" - Pink Sheet, 17 February, 2014.)
Some pharma companies already utilize AI and machine learning for post-market surveillance. Genentech, Inc. recently told the FDA that companies are using AI/ML to increase compliance, reduce manual review, and improve the quality of individual case safety reports before the information is entered into the safety database. (Also see "Artificial Intelligence: Industry Wants FDA To Boost Access To Databases, Clarify Use In Assessing Drug Efficacy" - Pink Sheet, 7 September, 2023.)
AI And FDA On Senate Radar
AI seems to be on the agenda of almost everyone in Washington these days. (Also see "Artificial Intelligence: US FDA Plans Guidances On Algorithm Bias, Product Development" - Pink Sheet, 16 March, 2024.)
On Capitol Hill, Republican Health, Education, Labor and Pensions committee leadership is looking at the FDA’s regulatory frameworks to determine if updates are needed to encompass AI, Barrett Tenbarge, ranking member Sen. Bill Cassidy’s general counsel for the committee, said during the FDLI conference. The committee also wants to ensure the FDA has the tools it needs to regulate AI, already a concern for the agency. (Also see "Cost Of Reviewing Drugs With AI Is ‘Prohibitive’ For US FDA" - Pink Sheet, 7 May, 2024.)
“We are looking at this from the standpoint of there is tremendous promise and potential,” Tenbarge said. “How can we make sure that specific to FDA in particular that the agency is well positioned to be able to weigh the risks and benefits of products that use AI, which is a core task that FDA is not unfamiliar with.”
While the committee does not seem anywhere close to legislation, Tenbarge indicated it wants to be proactive to ensure the regulatory framework for AI is fit for purpose “before there is some sort of acute issue that prompts Congress to act.”
There’s no “pen to paper on a particular bill right now,” Tenbarge said, adding that this is “a long-term process.”
One current obstacle is bipartisanship and Cassidy’s relationship with HELP Committee Chair Bernie Sanders, I-VT.
“As you all know on the Hill that it takes two to tango,” Tenbarge said. “So my boss has not been shy about his interest in talking about some of the things like AI and sort of his frustration with how the chair of our committee has decided to spend the committee’s time and resources.”
Senate AI Roadmap Says Little About FDA
Others on Capitol Hill are working in a bipartisan fashion on AI. However, that group is not hyper-focused on the FDA.
The Senate AI Working Group, which includes Majority Leader Chuck Schumer, D-NY, issued a 15 May roadmap for AI policy, which said little about the FDA and its needs.
“Ensure that the Department of Health and Human Services (HHS), including the Food and Drug Administration (FDA) and the Office of the National Coordinator for Health Information Technology, has the proper tools to weigh the benefits and risks of AI-enabled products so that it can provide a predictable regulatory structure for product developers,” the report stated.
This article was reported and previously published by Citeline’s Pink Sheet newsletter.
