Digital Healthcare and Healthy Aging | Consultant, Clinician, Researcher, Product Manager.
Great insights from David Bell, Ph.D. on the challenges of #AI in #healthcare documentation. The complexity and nuance unique to each healthcare market and the fact healthcare typically has few rote tasks but a lot of paperwork, makes it critical that #AItools be developed by experts in the industry with deep understanding of the most appropriate problems for AI enhanced tools to solve.
#AI in healthcare News alert! A Texas court has provided guidance on #AIApplications in charting. As we in #HealthTech and #healthcare grapple with implementing AI-based solutions, it often feels like we are flying blind in terms of #AIRegulation. So it's helpful to now have some guidance from the State of Texas, which to me boils down to three main principles: 1) No harm: Your system can't put hallucinations into the patient chart. 2) Accuracy: you can't make false statements about the accuracy of your system in your marketing; and 3) Disclosure: you must disclose the limitations of your solution. Here's my commentary on this guidance, as someone who spends all day running #homeHealth agencies and developing #AISolutions for the #HH market: A couple of college kids in a dorm room can make an AI product demo that is mind-blowing. They can get 80% of the way to a perfect solution in a weekend. The problem is that getting the other 20% to work takes 99% of the effort. That's what the Texas case deals with: the other 20%. Things like eliminating hallucinations and measuring how well your system works in edge cases fall into that other 20% category. The problem is that it is hard for developers (not to mention customers and investors) to figure out in advance how much work will be required to get from 80% to "good enough" for a particular application. In most healthcare applications, 80% isn't nearly good enough. #Medtech #Founders need to make decisions about how to approach their market. For us at GrandCare Health Services and io Health we take a stepwise approach. I get fired up about using AI to solve problems for #HomeHealthClinicians and #HomeHealth companies today. That includes avoiding disproportionate moral, business, regulatory or legal risk. I'm in my happy place working on getting better care to more patients at a lower societal cost. #GrandCare focuses on delivering excellent care to patients and #io focuses on delivering near-term, hard-money #ROI and #ROV for customers like GrandCare. I'm going to #HLTH and I'd love to see you there. DM me if you're going. I'd love to hear your thoughts on all of this. Thanks to Keith Grunig, MBA for bringing this news to my attention! Here's the press release: https://lnkd.in/gphukA9C And here's the associated court order: https://lnkd.in/gCk8Z2bZ #HealthcareAI #HomeHealth #HomeCare #MedTech #HomeHealthDocumenation #AI #iohealth Jon Warner, Jomo Kenneth Starke, Jon Higginbotham