Industry Watcher’s Digest

Buzzworthy developments of the past few days.

• An HHS agency is developing ways to keep aging AI-enabled medical devices up to snuff for the long haul. The Advanced Research Projects Agency for Health, or “ARPA-H,” says it will conduct the work through its Precise-AI program. Precise here stands for Performance and Reliability Evaluation for Continuous Modifications and Useability of Artificial Intelligence. The program will come up with mechanisms for not just detecting problems but automatically fixing them too. It will also create alerts of impending trouble for clinicians, AI developers, hospital administrators and regulators. Details here. 
   
• Generative AI can drown its users in a firehose of unhelpful information. Tim O’Connell, MD, offers the vivid word picture by way of cheering the advance of “medical grade” AI in MedCity News. He also calls for starving black-box tools by feeding only transparent algorithms. “What clinicians want and need is information at the point of care that is accurate, concise and verifiable,” writes O’Connell, a healthcare AI entrepreneur and clinical informatics academician. “Medical AI has the ability now to meet these requirements while safeguarding patient data, helping to improve outcomes and reducing clinician burnout.” Read the rest. 
   
• Patients need to know not just how AI is being used but also why. At times the latter might include leveling with them in a spirit of unprecedented transparency. Maybe even saying something along the lines of: “Because AI will help us improve not only your care but also our efficiency, productivity and financial health.” Columbia University bioethicist David Hoffman, JD, comes close to suggesting as much in a video interview with Marianne Kolbasuk McGee of BankInfoSecurity. “When we talk about AI influence on medical devices and treatment,” Hoffman adds, “we haven’t yet defined what we can reliably represent to patients as safe. And so we can’t make a judgment about whether that [AI’s] safety is sufficient to justify whatever positive effect [it] can produce.” View the video.
   
• Africa is bursting with biodiversity. A native researcher is using AI to screen the vast greeneries for potential remedies. “Traditional drug discovery involves understanding the structure of proteins in the human body or in a pathogen and painstakingly identifying molecules that can bind with or interact with them in desirable ways,” explains the scientist, Fidele Ntie-Kang, an associate professor of pharmaceutical chemistry at the University of Buea in Cameroon. “Machine learning and AI allow us to dramatically short circuit this process.” Learn more in a Q&A published by Gavi, the Vaccine Alliance. 
   
• Aidoc makes the list. So does Zebra Medical Vision. There are only three more in Analytics Insight’s picks for the top five healthcare companies using AI for diagnosis. Take three guesses and then click here to see how many you got right. 
   
• Startup founder Tina John of Avira Health took home top honors in an innovation challenge. Avira helps postpartum mothers start their motherhood journeys with know-how and confidence. John was awarded $10,000 by the event sponsor, CharmHealth. Categories included hardware accelerated AI, generative AI and tech-led innovations. More here. 
   
• The Australian government has found generative AI ‘worse than humans in every way’ at summarizing documents. Worse yet, far from reducing burnout, it “might actually create additional work for people.” The coverage is from an Aussie independent outlet called Crikey. Where has it been all my life?  
   
• Recent research in the news: 
   
  • Mayo Clinic: Pregnancy-related heart failure is under-detected; AI-enabled stethoscope helped doctors diagnose twice as many cases
     
  • Weill Cornell Medicine: Machine learning helps identify rheumatoid arthritis subtypes
     
  • Boston University School of Medicine: Artificial intelligence will play an increasing role in scientific publications
     
  • American Society of Plastic Surgeons: Machine learning predicts which patients will continue taking opioids after hand surgery
     
• AI funding news of note:
   
  • Magic secures $320M funding, partners with Google Cloud for AI supercomputers
     
  • £1.8M funding to extend London’s healthcare AI technologies and capabilities
     
• From AIin.Healthcare’s news partners:
   
  • Health Imaging: PHOTO GALLERY: Examples of FDA-cleared AI in radiology
     
  • Cardiovascular Business: AI for TAVR, robot-guided echo and more: Key research at ESC Congress 2024
     
  • Radiology Business: Radiology data sharing vendor Enlitic to acquire rival for $5M
     
  • Health Imaging: AI spots 25% of interval breast cancers missed by radiologists