FDA’s views on AI are evolving

Presented by UnitedHealth Group: The ideas and innovators shaping health care

Sep 26, 2024 View in browser   By Ruth Reader, Carmen Paun, Daniel Payne and Erin Schumaker Presented by  THE REGULATORS Tala Fakhorui | Courtesy of the Food and Drug Administration The Food and Drug Administration plans to release new guidance on the use of artificial intelligence in pharmaceutical research. The new recommendations, slated before the end of the year, follow a consolidation of three AI working groups into one council that will guide the agency’s AI policy work. The guidance is based in part on responses the agency received to a 2023 discussion paper on the topic. The paper received more than 800 comments from 65 organizations, said Tala Fakhouri, the FDA’s associate director for data science and AI policy. Between 2016 and 2022, the agency reviewed more than 500 drug submissions in which AI was used. Fakhouri spoke with Ruth about how the agency is handling the new technology. The interview has been edited for length and clarity. What are some examples of how companies are using AI in drug development? AI can be used in a clinical trial to stratify patients for different dosages or to identify which patients require inpatient monitoring after they take the drug. There’s also a lot of AI that’s being used to increase the efficiency of drug trials so that you don’t have to recruit a lot of patients to take a placebo. Then there’s a lot of work in advanced manufacturing, identifying safety signals after a drug has been marketed and it’s out in the population. How is the FDA adjusting? Our evidentiary standards are the same. We sometimes get asked these questions when it comes to emerging technologies: “Are FDA standards going to change? Are you potentially going to dilute the regulatory standards?” Absolutely not. But when AI is used, we can’t just review the output. We also need to learn about how you built your model. How did you train it? What data did you use to train it? Is the data fit for the purpose? What are challenges to evaluating AI? We already know that we need to ask questions about the data. A lot of that data sits either in silos within the same organization or across organizations. So the one thing that I think collectively the ecosystem needs to think about is: How do we increase the pool of fit-for-purpose data to train models better?   A message from UnitedHealth Group: Learn how UnitedHealth Group is using artificial intelligence to make health care work smarter, for everyone. Learn More.   WELCOME TO FUTURE PULSE Ferry Point Park, Bronx, N.Y. | Jeffrey Diaz This is where we explore the ideas and innovators shaping health care. Blind women trained in finding lumps are examining women’s breasts in parts of India where mammography equipment is scarce, NPR reports. Share any thoughts, news, tips and feedback with Carmen Paun at cpaun@politico.com, Daniel Payne at dpayne@politico.com, Ruth Reader at rreader@politico.com, or Erin Schumaker at eschumaker@politico.com. Send tips securely through SecureDrop, Signal, Telegram or WhatsApp.    A message from UnitedHealth Group:   WASHINGTON WATCH Paul's got Democratic support for his bill to tighten oversight of gain-of-function research. | Getty Images A Senate committee’s bipartisan vote Wednesday suggests that concerns about risky research Sen. Rand Paul (R-Ky.) first raised during the height of the pandemic in heated arguments with then-National Institutes of Health official Dr. Anthony Fauci are now widely shared on Capitol Hill. How’s that? The Homeland Security and Governmental Affairs Committee approved Paul’s Risky Research Review Act by a vote of 8-1. Only California Democrat Laphonza Butler objected. The measure would strip the NIH of authority to fund “risky research” on pathogens capable of sparking a pandemic and give that power to a new Life Sciences Research Security Board comprising nine members appointed by the president. “It’ll be that extra layer of protection to try to make sure something like a lab leak that could kill half of our population doesn't happen,” Paul said. Why it matters: In July 2021, Paul accused Fauci of lying to the committee about NIH funding of so-called gain-of-function research that can make pathogens more dangerous while studying them. Paul has said he believes research conducted by an NIH grantee at China’s Wuhan Institute of Virology led to the Covid pandemic. Fauci said at the time that the NIH didn’t fund such research and that Paul didn’t know what he was talking about. Fauci declined to comment on the Wednesday committee vote. The legislation would override a recent White House policy, due to take effect next May, which keeps decisions on risky research within existing agencies. That policy will require researchers who propose to increase the transmissibility of a pathogen or make it more virulent to submit a risk-mitigation plan to the agency funding the work, which will weigh the benefits against the risks in deciding whether to fund it. What’s next? The committee’s approval of Paul’s bill doesn’t guarantee a Senate vote, and the House hasn’t passed a companion measure. It's up to Majority Leader Chuck Schumer to schedule a vote.   A message from UnitedHealth Group: UnitedHealth Group has been deploying artificial intelligence (AI) and machine learning (ML) applications at-scale for years. Now, improvements in data and processing power are enabling more impactful uses of AI/ML, ushering in new possibilities for patients, providers, and the overall health care system. Learn More.   TECH MAZE Doctors aren't happy about the fees electronic health records firms are charging. | Jon Cherry/Getty Images The nation’s leading professional society for doctors wants HHS to investigate the fees that electronic health records firms charge to send patient records to other providers or insurers. In a letter sent Wednesday to Micky Tripathi, HHS’ national coordinator for health information technology, American Medical Association CEO James L. Madara said doctors incur high costs for transmitting patient data, and he wants the Department of Health and Human Services to determine whether the fees violate its information-blocking rules. Tripathi’s division set rules earlier this year to enforce a provision in the 21st Century Cures Act of 2016 barring providers from refusing to share patient data when it’s what the patient wants. HHS’ information-blocking rules also prohibit charging excessive fees to access electronic health record data. The fees make it difficult for doctors to serve patients, AMA President Bruce A. Scott told POLITICO. “Many physicians can’t connect to even a nearby hospital because EHR developers want to nickel-and-dime them to make interoperability a reality.” Why it matters: So far, HHS has received more than 1,100 complaints of information blocking but hasn’t issued any penalties. Bill Cassidy (R-La.), ranking member of the Senate Health, Education, Labor and Pensions Committee, sent a letter to Tripathi and HHS Inspector General Christi Grimm in June, asking them to explain why their agencies haven’t enforced information-blocking rules. Neither agency has sent a formal response.   Follow us on Twitter Carmen Paun @carmenpaun Daniel Payne @_daniel_payne Ruth Reader @RuthReader Erin Schumaker @erinlschumaker   Follow us

To change your alert settings, please log in at https://login.politico.com/?redirect=https%3A%2F%2Fwww.politico.com/settings This email was sent to edwardlorilla1986.paxforex@blogger.com by: POLITICO, LLC 1000 Wilson Blvd. Arlington, VA, 22209, USA Unsubscribe | Privacy Policy | Terms of Service