Latest / CareTalk: Healthcare. Unfiltered. / Making Health AI More Collaborative w/ Health Universe CEO, Dan Caron | HealthBiz Brief

Transcript

Making Health AI More Collaborative w/ Health Universe CEO, Dan Caron | HealthBiz Brief · CareTalk: Healthcare. Unfiltered.

• 0:06
  Welcome to Health Biz Briefs.
• 0:08
  I'm David Williams, President of Health Business Group here with Health Universe founder and CEO Dan Karan.
• 0:13
  Dan, thanks for joining me today.
• 0:15
  Hi, David.
• 0:15
  Great to be here.
• 0:16
  So our topic today is about an open source platform for healthcare AI apps.
• 0:21
  Now, Dan, my first question for you is, what is an artificial pancreas and what do you have to do with it?
• 0:28
  That's a great question.
• 0:30
  I'm a type in diabetic and I participated in an open source group.
• 0:35
  We hacked the Omnipod insulin pump and we integrated that into a mobile phone application called loop and it automates the delivery of insulin up or down.
• 0:45
  Uh, there's tens of thousands of people that used it.
• 0:48
  And it really proved that an open source community can come together and create a safe innovation, um, in a, in a disease state.
• 0:56
  And that application, uh, became the first sort of.
• 0:59
  open source FDA approved medical application last January.
• 1:04
  My role, I was the reverse engineer.
• 1:07
  I sort of hacked the wireless communications, but there was a really great and large group of people that worked on everything from documentation to the mobile driver to everything.
• 1:19
  So it just shows that open source community can create real innovation in healthcare.
• 1:24
  Sounds like when it's like Ocean 11 type of thing.
• 1:27
  You've got all the experts.
• 1:29
  Exactly.
• 1:29
  Yeah, yeah, we had code names.
• 1:31
  Yeah, this time for good.
• 1:33
  Great.
• 1:34
  Then what is Health Universe?
• 1:37
  So Health Universe is a startup that I founded last year.
• 1:40
  It is a collaborative platform for health AI.
• 1:44
  I realized that there was no vertical aggregating health AI.
• 1:49
  And I wanted to build a platform that could bring data scientists and clinicians together and really create a sandbox environment to mature AI algorithms from prototype to stable to institutionally validated and even FDA approved.
• 2:05
  And really it's an ecosystem for researchers and clinicians to mature these really exciting technologies.
• 2:15
  How's it going so far?
• 2:16
  Oh, great.
• 2:18
  We have, we've received quite a bit of attention from our hackathon last year and we have some really well-known cancer institutions deploying.
• 2:28
  algorithms onto the platform.
• 2:30
  And it's super exciting.
• 2:32
  I wanna live in a world where the clinician anywhere can run the most advanced breast cancer detection model right from the browser, because I think we need to figure out a way to socialize these really powerful tools for the world.
• 2:48
  I've been hearing a lot lately about AI agents in general and in healthcare.
• 2:53
  Can you tell me what is an AI agent?
• 2:56
  And...
• 2:57
  What is the potential for them in particular in healthcare?
• 3:00
  Yeah, so AI agents are going to take over the world in the next few years, and hopefully they will be kind to us.
• 3:08
  But effectively what an agent is, it's an algorithm that can make decisions about the work that needs to be done.
• 3:18
  So you can imagine in healthcare, an agent might be refilling a prescription.
• 3:25
  And there may be different questions about that prescription, or there may need to be a prior authorization that needs to be done.
• 3:31
  And an agent would automate that process and sort of remove all of the decision elements that sort of hold up healthcare today.
• 3:40
  And on Health Universe, we're building medical co-pilots that remove a lot of these burdens at the moment.
• 3:47
  And they really allow for sort of a faster, more efficient future for healthcare.
• 3:53
  Got it.
• 3:53
  I was going to ask you if there's a distinction between agents and co-pilot, which is another term that I've heard.
• 3:59
  The way you described the agent is a helper.
• 4:01
  Co-pilot, I think about, I think it's used that term is used that way a lot, but I also think about co-pilot, if something happens to the pilot, they can take over.
• 4:09
  And I don't know where the agent to do that is.
• 4:11
  I'm trying to actually get my head around that distinction.
• 4:13
  Yeah, that's a great question.
• 4:15
  So co-pilots generally come from the generative AI world, the world of GPT-4.
• 4:22
  and chat GPT.
• 4:23
  So a co-pilot is something that you're, it's generally a chat interface.
• 4:28
  Whereas an agent might not be a chat interface.
• 4:31
  An agent could be just making decisions autonomously around the clock about workflow that needs to be taken care of.
• 4:40
  Whereas a co-pilot is something that a clinician would be sort of interacting with or a patient would be interacting with to get a job done.
• 4:49
  Got it.
• 4:50
  So I can imagine like a...
• 4:52
  you know, an agent might be useful for something like prior authorization, where it's like, I don't really want to be around.
• 4:57
  It's like you go deal with it.
• 4:58
  Yeah.
• 4:58
  And you come back when it's dealt with, right.
• 5:00
  And a co-pilot might be, I want to explore my health and how I might change my fitness routine or something like that.
• 5:05
  Yeah, perfect.
• 5:06
  That's exactly it.
• 5:08
  All right.
• 5:08
  Well, I'll get that.
• 5:09
  I'll sign up for that prior authorization agent, uh, when, as soon as it's, uh, as soon as it's available.
• 5:14
  Great.
• 5:15
  Well, Dan Keren, CEO of health universe.
• 5:17
  Thanks for joining me on health business briefs.
• 5:19
  If you like what you hear, check out My Health Biz podcast for a full-length interview with Dan about his life and career.
• 5:26
  Dan, thanks again.
• 5:27
  Thanks, David.