The ideas and innovators shaping health care
| | | | |  | | By Daniel Payne, Ruth Reader and Erin Schumaker | | | | |  Some good could yet come from the sound of coughs. | Getty Images | Health providers hope artificial intelligence can hear what humans can’t. Audio-focused AI systems, they think, can turn the sound of patients’ speech, breathing or coughing into a diagnosis. How so? AI models trained on huge amounts of audio may be able to link sound with health. Google researchers recently announced the creation of a model trained on the sound of nearly 100 million coughs to aid researchers. Salcit Technologies, an India-based company, is exploring how Google’s model could expand its current work, particularly in developing tools that could aid in the early detection of tuberculosis. What’s next? Applications of “acoustic health research” could go beyond respiratory diseases. Kintsugi, a California-based health AI startup focused on diagnosing clinical depression and anxiety, suggests AI trained on sound could help. The company’s technology looks for “voice biomarkers” — from pitch to speed of speech — to determine who may be showing signs of a behavioral health problem. “It’s not the words that people are saying,” Grace Chang, founder and CEO of the company, told Daniel about what she and her company learned in creating the model. “It’s more… how they are saying it.”
| | | | 
Shenandoah National Park, Va. | Shawn Zeller/POLITICO | | This is where we explore the ideas and innovators shaping health care. The nation is so divided that elections no longer fuel political polarization, according to a new Science Advances paper. While that shift may take the bite out of divisive campaigns, the downside is that Americans don’t "cool off" after campaigns end, the researchers found. 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 new group at the FDA is considering how to use and regulate AI. | Andrew Harnik/AP | The FDA’s Center for Drug Evaluation and Research has launched an Artificial Intelligence Council to support AI use and regulation — and to ensure the agency meets the demands of President Joe Biden’s AI executive order. The council’s task: Determine how CDER can use AI in its own work and what guardrails are needed when the technology is used to develop new drugs. The council is led by Sri Mantha, CDER’s director of strategic programs; Tala Fakhouri, associate director of data science and AI policy; and Qi Liu, associate director of innovation and partnership. Council members will come from all corners of CDER. Why it matters: CDER is seeing increased AI use among regulatory submissions in drug development and new use cases for the technology. “The scope and impact of AI use in drug development are expanding,” CDER Director Patrizia Cavazzoni wrote in a letter to staff last week.
| | | | 
Doctors are relying more than ever on AI-powered tools. | AP Photo | The Advanced Research Projects Agency for Health wants to ensure that AI-powered medical devices don't break down over time. That’s the focus of a new agency program called Performance and Reliability Evaluation for Continuous Modifications and Useability of Artificial Intelligence, or PRECISE-AI. Researchers backed with ARPA-H funding will develop methods for monitoring AI-enabled device performance to determine whether devices still work as well as they did when first developed and, if not, determine the cause of the problem and correct it. Why it matters: AI tools can degrade over time — for example, when IT infrastructure changes — and methods for monitoring and training them haven’t caught up with their rapid development. In the meantime, approval of AI tools have exploded. FDA authorization of AI medical devices, like virtual assistants and medical imaging and diagnostic tools, increased 10-fold between 2018 and 2023, according to ARPA-H. “The inability to automatically monitor and maintain an AI-enabled medical device’s performance based on real-world operations creates risks for clinicians and patients,” Berkman Sahiner, PRECISE-AI’s program manager, said in a statement. “The promise of AI-enabled tools for health care is only as strong as the relevant real-world data informing them.” What’s next? ARPA-H expects to announce this week a date for a proposers’ day, when researchers can learn more about the program. | | | | Follow us on Twitter | | | | Follow us | | | | |
No comments:
Post a Comment