| | | | By Shawn Zeller, Daniel Payne, Erin Schumaker, Carmen Paun and Ruth Reader | | | | Texas' Lonestar6 is a step up. | AP Photo/Marcio Jose Sanchez | A group of health care researchers will gain access to advanced artificial intelligence systems as part of a government pilot to make the state-of-the-art tech more accessible. On Monday, the White House revealed the first round of research projects that will have access to computational time on the systems, our Mohar Chatterjee reports. President Joe Biden’s October executive order created the initiative, known as the National AI Research Resource. It launched as a two-year pilot in January. The National Science Foundation is leading the effort, partnering with other agencies, including the National Institutes of Health, and companies such as OpenAI and Microsoft. The computing systems are at the University of Illinois Urbana-Champaign, the University of Texas at Austin, the Pittsburgh Supercomputing Center and the Energy Department’s Oak Ridge and Argonne national laboratories. The health care winners: — Hamid R. Tizhoosh, a professor at the Mayo Clinic, will use the Texas Advanced Computing Center’s Lonestar6 system to study images of tissue disease. — Dr. Vadim Shteyler, a fellow at the University of California, San Francisco, will use Lonestar6 to improve the treatment of patients with sepsis by fine-tuning the use of intravenous fluids. — Dr. Kevin Schulman, a professor of medicine at Stanford University, will use the Oak Ridge National Laboratory’s Summit supercomputer to study video of patients in intensive care, aiming to improve evaluations of their condition. — Madhumita Sushil, a postdoctoral scholar in computational health science at UCSF, will use Summit to study clinical notes from the university’s health system to create an AI model for medical use. — Yuankai Huo, a professor of pathology, microbiology and immunology at Vanderbilt University, will use Summit to devise better ways of safeguarding patient data. “We all know that [companies] are putting in many, many billions of dollars into building new AI models,” said Arati Prabhakar, director of the White House Office of Science and Technology Policy. “But for the AI future that we need, we also need to make sure a much broader set of researchers have access to both the computers and the data that’s necessary.”
| | THE GOLD STANDARD OF HEALTHCARE POLICY REPORTING & INTELLIGENCE: POLITICO has more than 500 journalists delivering unrivaled reporting and illuminating the policy and regulatory landscape for those who need to know what’s next. Throughout the election and the legislative and regulatory pushes that will follow, POLITICO Pro is indispensable to those who need to make informed decisions fast. The Pro platform dives deeper into critical and quickly evolving sectors and industries, like healthcare, equipping policymakers and those who shape legislation and regulation with essential news and intelligence from the world’s best politics and policy journalists. Our newsroom is deeper, more experienced and better sourced than any other. Our healthcare reporting team—including Alice Miranda Ollstein, Megan Messerly and Robert King—is embedded with the market-moving legislative committees and agencies in Washington and across states, delivering unparalleled coverage of health policy and the healthcare industry. We bring subscribers inside the conversations that determine policy outcomes and the future of industries, providing insight that cannot be found anywhere else. Get the premier news and policy intelligence service, SUBSCRIBE TO POLITICO PRO TODAY. | | | | | | Smoky Mountains, North Carolina | Daniel Payne/POLITICO | This is where we explore the ideas and innovators shaping health care. Students figured out a 2,000-year-old math problem, using trigonometry to solve the Pythagorean Theorem. That’s the one that says the length of the sides of a right triangle are equal to a² + b² = c². 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.
| | | AI could make better triage decisions, a study found. | AP Photo | Triaging patients in the emergency room — deciding who needs care immediately and who can wait — is a job with life-or-death implications. Artificial intelligence might do a better job at making such decisions than humans according to recent research published in JAMA Network Open. In the study, researchers gave OpenAI’s GPT-4 real-world data from the University of California, San Francisco, on past emergency room patients, two at a time, and asked the tool which patients should have received care first. The system did relatively well, using clinical notes on the patients to make its judgments, and picked the correct patient in nearly 90 percent of the pairs — more often than a physician reviewer, who reviewed a subset of the 10,000 patient pairs presented to the AI system. Even so: An AI system is only as good as the data fed into it, and researchers and health system leaders still don’t fully understand how advanced AI systems reach conclusions. That means bias — introduced, potentially, by relying on clinical notes written by doctors — could go undetected. Those problems could emerge if the system was given a larger number of patients to consider. Why it matters: AI’s performance in triage and diagnostic tasks has the potential to directly alter patient care across the health system should it be implemented widely. And AI tools are already at work in emergency departments.
| | LISTEN TO POLITICO'S ENERGY PODCAST: Check out our daily five-minute brief on the latest energy and environmental politics and policy news. Don't miss out on the must-know stories, candid insights, and analysis from POLITICO's energy team. Listen today. | | | | | | Researchers say they better understand the genetic roots of PTSD. | Getty Images | Genetic research into the possible drivers of post-traumatic stress disorder is showing results. A new study of more than 1.2 million people has pinpointed 95 regions of the genome linked to PTSD — and 80 of them have never been previously identified. The researchers also identified 43 genes that seem to have a role in causing PTSD. “This is a milestone for PTSD genetics,” Karestan Koenen, study co-author and psychiatric epidemiology professor at the Harvard T.H. Chan School of Public Health, said in a statement. The findings build on existing research, which suggests PTSD has a genetic component similar to that of other mental illnesses, such as depression. But because PTSD is highly polygenic, meaning thousands of genetic variants can contribute to the condition’s development, it’s tough to predict who’s most at risk based on their genes. Researchers hope the findings might one day help scientists pinpoint genetic targets for PTSD prevention and treatment. The NIH-backed study was published in Nature Genetics last month. What’s next? PTSD is a common mental health disorder marked by intrusive thoughts and mood changes after a person experiences a traumatic event. While 8 million U.S. adults will have PTSD at some point in their lives, only 6 percent of people exposed to trauma will develop PTSD. Although PTSD rates are higher among people of African, Native American and Latin American ancestry, genetic databases historically focused on people of European ancestry. “We know that trauma and PTSD disproportionately affect under-resourced populations globally, particularly African ancestry populations,” Koenen said. “Our next steps will focus on addressing that inequity through partnerships with African scientists to make sure research in PTSD genetics benefits everyone equally.” | | Follow us on Twitter | | Follow us | | | |