Industry Watcher’s Digest

Buzzworthy developments of the past few days.

• For a good while now, AI eye exams for diabetic retinopathy have held steady as a model of AI usefulness in clinical settings. Here’s another reason to applaud the use case and hope it inspires other equally valuable applications. In a multisite clinic system serving minority residents of Southern California, the exams help patients who won’t take time off of work for normal doctor appointments unless they have no other choice. The AI exams are much quicker and offer that choice. “Ninety percent of our patients are blue-collar,” explains the director of operations for the clinics. “They don’t eat if they don’t work.” The director calls the technology “a godsend.” KFF Health News has the story.
   
• Remember the 2023 class-action suit accusing Humana of using AI to deny care to Medicare Advantage patients? The insurer is asking a federal court to spike it. “Allowing this case to continue would require the Court to apply 23 states’ standards, and risk outcomes that conflict with the federal government’s Medicare rules,” Humana writes in a motion filed last week. “This is exactly why the Medicare Act has such a vast preemption provision.” McKnight’s Long-Term Care News is on it.
   
• If not properly trained, AI can lead to bias and discrimination. Just as bad if not worse, AI chatbots can generate medical advice that is misleading or false. Happily, those are the worst of the negatives. In a consumer-friendly primer on the technology, Mayo Clinic Press acknowledges the few (already well-known) downers while spending many more words on the good that healthcare AI can do—things like detecting imperceptible conditions, anticipating risk of disease years in advance and matching hopeful patients with clinical trials. Read the item.
   
• Besides, AI bias is avoidable. When it pops up, it can usually be traced to an avoidable combination of technical errors and human decisions. Marshall Chin, MD, MPH, a professor of healthcare ethics at the University of Chicago, breaks it down for Information Week. “This is something that we have control over,” Chin says. “It’s not just a technical thing that is inevitable.” Other experts echo that observation and describe bias-busting steps to take. Article here.
   
• The CDC opened its Office of Public Health Data Surveillance and Technology last April. One year on, the Federal News Network looks at how far OPHDST has come. One major thrust right now is figuring out how to apply Gen AI for parsing public health data to beneficial effect. In this endeavor the office is looking to partner with academia and industry to “understand the full spectrum of [available] tools and how they could be used,” OPHDST director Jennifer Layden, MD, PhD, tells the network.
   
• In introducing a couple dozen healthcare-oriented AI tools last week, Nvidia markedly advanced its long-term designs on the medical sector. “In healthcare, digital biology and generative AI are helping to reinvent drug discovery, surgery, medical imaging and wearable devices,” Nvidia CFO Colette Kress said at the company’s spring GPU technology conference (aka “Nvidia GTC”). “We have built deep domain expertise in healthcare over the past decade.” CNBC coverage here.
   
• AI database supplier Activeloop has raised a fresh $11M in Series A funding. This round nudges the company’s total funding to right around $20 million. Activeloop says it will use the broadened backing to, for starters, refine its Deep Lake technology. This allows the company’s enterprise clients to readily connect unstructured multimedia data with machine learning models. CEO Davit Buniatyan says Deep Lake lets enterprises “create AI applications that are more accurate, boost AI engineering teams’ productivity by up to 5x, and cost up to 75% less than market offerings.” Announcement here.
   
• In other funding news of note:
   
  • Tennr puts fax machines back in vogue for healthcare organizations using AI, as it secures $18m from a16z
     
  • Binghamton University set to receive $5 million landmark gift to pursue AI research and development
     
  • Treatment.com AI Funded by the National Institute of Health to Develop a Culturally Sensitive AI Approach to Health Equity
     
• Research news roundup:
   
  • Stanford Medicine-designed AI tools tackle soft tissue sarcomas, identify new treatment strategies
     
  • Analysis of social media language using AI models predicts depression severity for white Americans, but not Black Americans
     
  • Novel research leverages AI to identify dogs at higher risk for cancer
     
• From AIin.Healthcare’s news partners:
   
  • Health Imaging: New AI model predicts cancer risk based on breast asymmetry
     
  • Cardiovascular Business: New data registry will track how AI-powered plaque assessments impact cardiologist decisions
     
  • Radiology Business: Large language models excel at simplifying radiology reports