Episode Transcript
[00:00:01] Speaker A: You're listening to the Doc Lounge Podcast. This is a place for candid conversations with the healthcare industry's top physicians, executives and thought leaders.
This podcast is made possible by Pacific Companies, your trusted advisor in physician recruitment.
[00:00:20] Speaker B: Hi, everyone, I'm Stacey Doyle, Senior Director of Marketing at Pacific Companies, and welcome back to the Doc Lounge Podcast. In today's Ask the Expert episode, we're diving into a conversation that's at the cutting edge of healthcare innovation. Our guest is David Norris, a visionary CEO, investor and the driving force behind AI transformation at Affineon Health. David has a unique ability to merge technical innovation with healthcare efficiency, helping providers streamline their workflows and most importantly, reduce burnout. With a following of over 13,000 professionals on LinkedIn, his insights have influenced healthcare leaders across the industry. Today, we'll explore how AI is reshaping patient care, strategies for reducing physician burnout, and how technology can help providers reclaim their time and focus on what matters most, caring for patients. Get ready for an insightful conversation that could change the way you think about the future of healthcare. David, welcome to the Doc Lounge podcast.
[00:01:16] Speaker C: Thank you so much. Very nice to be here.
[00:01:19] Speaker D: So tell us, give us a little bit of background. What made you, you know, inspire you to start, if any, on health.
[00:01:26] Speaker C: You know, I've been in technology for a long time and I believe that providers are probably in one of the most difficult and challenging situations. As a technologist and someone that loves to apply it to solve real world business problems, I can't think of anything better than to help providers who are on the front lines and trying to address all kinds of needs that patients have. I think it's a great place for technology to make a real difference.
[00:01:55] Speaker D: That is very inspirational. Tell us a little bit about what are some of the challenges in healthcare that you are hoping to address.
[00:02:04] Speaker C: You know, the leading challenges that I see with providers is that they are just overloaded. You know, if you've tried to get an appointment with a doctor lately, you know, it's often challenging. So they are in this difficult situation of trying to see more and more patients, trying to help those patients with more and more complex situations. And the burnout factor with providers is just huge. They many, many times will just get tired of doing it and will go on to pursue other professions. And at this point, by 2030, there's a projected shortage of about 100,000 doctors. And if you think it's bad today, just imagine that scenario. So I think that's the number one problem that we're trying to address is give Providers some help, let them do what they got into medicine to do, which is to help patients rather than sitting behind a computer and hopefully encourage more providers to get into the business.
[00:03:04] Speaker D: Great insights there. And a topic that we discuss a lot on the podcast about just the shortage of physicians and how that is just going to become and continue to increase, and what are some ways and strategies to help that and like you're saying, maybe offset some physicians from retiring early. Just solutions around that. So I'm really excited to have you on and you've really been at the forefront of transforming healthcare through AI, you know, at affinity and health. So tell us a little bit, how can you know AI?
How are you utilizing it? Like what? Tell us a little bit more about what's the magic that you have at play.
[00:03:49] Speaker C: Well, you know, all of us have experienced some of the really, really exciting power of AI, and Affineon is applying that power to initially one of the most challenging areas that providers struggle with every day, which is their clinical inbox. And if you think your inbox is bad, you should see a doctor's, you know, typical PCP will see maybe 20 or 25 patients a day. And every patient, they end up ordering four or five different lab results. And then they have prescription renewals coming in. So they'll have hundreds and hundreds of things coming into their inbox every day.
And they're trying to see 20 patients. So they're squeezing this in wherever they can, literally getting up early, trying to work through it between patient meetings, trying to work through it, and then often staying up late at night. And Infineon has applied AI to help, much like you would do with a human. In the old days, providers would sometimes hire a nurse to help triage their inbox. Today, budgets won't really allow that to happen, but if you had that capability, it'd be great. I think all of us would love that. So what we've done is created an AI agent that can triage the inbox for the provider and really help them in focusing their time on those clinically more difficult cases where they need to really think about, let's say, a set of lab results that come back for our patient. They're trying to figure out what's wrong with the patient. They can actually spend more clinical time on that and less on the normal lab results or the clinically less significant results. So our AI agent can hand them back about five hours a week, which for PCP is a significant amount of time.
[00:05:31] Speaker D: Wow, that's a huge time savings that you're Giving back to these doctors, that. And like you're saying, primary care is such an important physician and leads to all type of referrals and really leads people to their next step if they have any concerns with their health that they need to address. So I could see how this would really be instrumental in relieving some of that burden that they have.
So I'm assuming that this AI is able to kind of smartly recommend certain things based off of, you know, what it's reading in their inbox. Is that correct?
[00:06:10] Speaker C: Yes. Our system plugs directly into the electronic health record system that most providers use in their practice. And it's seamless. It doesn't require that you install some separate app or log into some separate system. It's all integrated right into the ehr. So the provider can rely on us to automatically handle things that are coming into their inbox, analyze them, apply a strict set of protocols where the provider can tell us what they do want us to handle or what they do not want us to handle. So it's completely under provider's control. We implement those protocols, and then we can take, for example, a lab result that comes back, analyze it, determine that we can handle it, and go ahead and notify the patient, hey, congratulations, your lab results are normal. And then close it out of the inbox so the provider gets help. I think that's the real test of AI is can you really make a practical and meaningful a bit of help for the providers? With us, we're able to eliminate between 30 and 80% of the inbox every day. And that allows the provider to go back and spend time with the patients, which is where they really want to be.
[00:07:20] Speaker D: That's right. I mean, that's why they got into, you know, practice to begin with. So I love hearing that savings that they'll be getting back, that's. That's really, really powerful. Now, tell me, you were talking about electric health record systems. Which ones will this plug into?
[00:07:39] Speaker C: It works with Most of the EHRs that are out there. And again, the real secret is to make it work in a seamless fashion. I think that's the key providers, like most of us, don't really want to change how we do things. And asking a provider to take on a new app or something they have to install can really, really be challenging. So we've designed our system so that it's completely integrated to their existing workflow. They can go in and either we will automatically handle a result for them, and that way it's removed from the inbox or if there's something that they need to actually spend some time on, we analyze the labs, we pull the entire patient history into our system, and in the process of analyzing the labs, we can highlight the things that are most important for them to take a look at. Think of it kind of like a dashboard on your car. You kind of know there's a lot of things that could go wrong in your car, but you see the dashboard, and those are the important things. We do that for things like lab results. So a provider can get through lab results faster by focusing their time on the things that are most important. And sometimes we can highlight things that aren't really very obvious because we can speed read an entire patient chart in two seconds.
Most providers would take 10, 15 minutes to read through the whole chart. So I think this is, to me, the real benefit of AI is do things to help providers. Don't do it for them, but actually help them. And that gives them the ability to spend time on the things that are most critical.
[00:09:12] Speaker D: That really, really sounds like something that could be a game changer for a lot of physicians out there. Now, tell us a little bit about.
People might be worried here about privacy laws, you know, personal information, you know, obviously hipaa. So how did you train your models? Or tell us a little bit about that so people can understand.
[00:09:33] Speaker C: Well, our system is really sophisticated. Most people don't realize how many different, for example, kind of lab results there are. So today We've processed over 2,000 different types of labs. Most of us as patients have probably gotten like a CBC or a cmp, some of the normal kind of basic tests done, but there's actually a huge range. So we've trained our system on how to handle that broad range of lab results. We typically will de. Identify results in patient information when we pull it into our system. So our system is trained on. There is a patient that has this kind of condition in this kind of lab, but it's not the specifics of that actual patient that allows us to really make an intelligent system that's trained to handle these things without disclosing all of the private information about patients. I think that's the real key with AI is you want to teach it, but you don't want to teach it all of the secrets.
[00:10:30] Speaker D: We've heard that AI may be utilized in radiology, reading of kind of X rays and things like that. So tell us a little bit about. Is that something you guys are also looking into?
[00:10:45] Speaker C: You know, it's a great question. So we do not read images. We rely on the radiology and their technologies to do that. What we do is actually on the other side, which is if you've ever seen a radiology report back, they can be extremely difficult to understand. So we help by taking those reports and working with the provider to understand how do we communicate to that, to the patient in a way that makes sense and that they can understand. So taking a complicated note from the radiologist that a provider might understand and then crafting that into a message to the patient today, that's what providers end up having to do. And it's often a lot of work to try to write out a report that the patient can understand. So we do that for them. We also do that for obviously lab results where a labor result could be actually quite complicated coming back and the patient just wants to know, hey, am I okay? What do I need to worry about? And so we're able to translate that medical terminology into a consumer friendly terminology. Another kind of important part of our system is getting it to the patient quickly. I don't know if you've ever had the experience, but sometimes you may go to the doctor and the follow up can take some time. And if you get some lab results back and maybe you get to see them at the same time as the provider and something looks abnormal, you might be wondering, what do I do? And the provider might not get back to you for a week or two. And often that causes the patient to send the provider a note saying, hey, I saw this is abnormal, what do I do? And of course that then adds more to their inbox, so they just get further and further behind.
We're able to process those lab results and give them an answer in a matter of seconds, which for the patients is a better patient experience. And for the provider, it helps to head off those unnecessary questions that would otherwise just add to their burden.
[00:12:42] Speaker D: Yeah, I love that. I mean, I think we all can relate to that when you get the results and then because it's real time now when a lot of these apps. But then you don't really have obviously the interpretation and obviously the medical guidance and all the knowledge that the physician has. So I think that would be really, really game changing for both physicians and patients. Now tell me, do the physicians then have the chance so they're reviewing everything before it's sent to a patient? Obviously it's streamlining the process and making that communication a lot more consumer friendly. But what kind of approvals are there?
[00:13:20] Speaker C: Yeah, in our system, the provider is in complete control always. So they define the rules or the protocols that allow AI to handle some of these things automatically and they pre approve all of the messages that are going to be sent to patients. So there's nothing that happens without their approval and their sign off. And I think this is super important because, for example, if we're processing a lab result that came back in as normal, that doesn't mean that it's all normal. There can be values that moved significantly that are still within the normal range, but they moved in a short period of time. And that can be an indication of something important. So in our protocols we can define anything that's clinically significant that should really have oversight by the provider and we can direct it to the provider for review rather than in the old days, they used to write programs that would use very simplistic rules for sending things out and often they could make mistakes. So we take the approach that the provider is always in control. They define exactly what should happen and by doing that, we don't have to worry about the fact that we're doing something that would be clinically bad for the patient. Patient safety is top of our list.
[00:14:37] Speaker D: Love that. Now tell me, obviously some physicians are concerned about AI replacing the human connection in medicine. How do you see AI support rather than replacing providers in their roles?
[00:14:51] Speaker C: It's a great question. You know, I spend a lot of time with providers. I'm very pro provider and I think what I've seen with most providers is that they just love to have better tools that help them to be able to do a great job and they don't really care too much about what's behind those tools. Take an mri. Most of us don't know what's inside of an mri, right? We just know it works. There's a lot of AI inside of MRIs. Providers don't care. They want to make sure that the tool works well. So with our solution, and I think with most others, they really look at the output and the outcome from the tool and if it is reliable and you can depend on the results of it. We have seen nothing but positive reaction from providers. We so far haven't had any providers that, you know, I don't want AI because I don't trust it or anything like that. It's more about does it work. And so in our product we have taken extreme measures to ensure that our AI delivers consistent, high quality results. That includes having a human in the loop layer where if we get a lab result, as an example that we've never seen before, it's automatically routed to a human team that can review. It can ensure that our system is trained appropriately on how to process that kind of result, and then it can go to the provider. So we put the trust of the provider at a high level because we think that's really all providers care about is does it work reliably?
[00:16:23] Speaker D: I think that sounds really great that you have some type of, basically, if it hasn't been in your system before, there's a way that you're getting that oversight of the actual provider, or I'm assuming a panel of providers that are looking at that before it's then moved on into, you know, into your system and your model. So I'm ensuring that builds, you know, trust with providers, which is, like you said, is the most paramount thing that they, that they all want because they, you know, the patient provider trust is so important. So what are some of the other things that you initially did to kind of get the initial feedback? I'm about this. I think providers would love to hear about that.
[00:17:07] Speaker C: Yeah, Well, I think this is one of the hardest things in AI is we've all used ChatGPT and other AI tools and we've probably all experienced hallucinations where you ask a question and it gives you very confidently an answer back that is completely wrong. And that, I think, is the big challenge in health care is to make AI work in a reliable fashion. So in our system, we designed a capability for our system to get feedback from this human lay. And initially we reviewed everything manually and that allowed our clinicians to teach the model, this is the right way. This is the right way. And over a period of years, we were able to train our system very effectively on how to handle over 2000 different types of lab results. It's not an easy process. I mean, it takes a lot of time and effort to do that. But then at the end, you have a very, very reliable result. So you have to build a system that can know when it's seeing something it doesn't know how to handle and it can ask for help. And by doing that, then insert humans in the loop to be able to teach it how to do it. That's, I think the critical fashion here is most people think AI can just know how to do everything on its own. And in medicine, that's just not the case. There's a lot of things it doesn't know how to handle.
[00:18:26] Speaker D: It sounds like you and your team were ahead of the AI curve because you said it's. You've been training this for a couple of years. So is that how long has, have you been building this model?
[00:18:39] Speaker C: You know, we're about two years old, so we've, we've been building it from the beginning. And the thing that we did that is maybe a little bit different than most is we didn't take the, you know, the two people in a garage going build a solution. We actually engaged providers from the beginning, so we were able to work with some very, very seasoned providers that would give us very, very useful feedback. And as we built our proprietary data set, which is now huge, that allowed us to train our system on how to handle a very, very complicated set of clinical results. And I think that's another area that in AI and in healthcare, a lot of applications of AI have been relatively simple. Our particular one is very complicated because lab results are very clinically challenging. And then if you take into account the history of a patient, that can also be quite clinically complicated. That ends up with a really interesting challenge. How do you take a complicated patient and a clinically difficult lab result and figure out what to do next? So that's where we play. And that has resulted in a pretty sophisticated system. We had deep AI experience, obviously, in the company from the beginning that allowed us to build this system. But I don't think there's any shortcut. You have to do the work. You have to spend the time and build the data necessary to be able to train a system to do this. So it's. It's just not easy to do. Like some technologies are overnight successes. I think this one just takes a lot of time.
[00:20:10] Speaker D: That makes sense. I loved when you said that you were pro provider, so I would love for you to share. Do you have a story of, you know, a provider that has implemented Athenian and their results?
[00:20:25] Speaker C: I do, yeah. I spend a lot of time with a lot of providers, so picking one is always difficult. But I'll pick two just for fun. So one provider that we work with, when we first met her, she had thousands of items in her inbox and she was literally drowning. She couldn't catch up. And it's often described as like a treadmill or a hamster wheel. You know, once you get behind the daily volume, coming in will take up all your time. And you have thousands of things that you can't catch up. So she was really, really in trouble. And once we implemented the AFINEON system, we were able to help her get caught up and then stay current with a much lower level of stress. And so by implementing the system, it gave her back an ability to focus on patients rather than to Just worry about, oh, I gotta go spend 10 minutes trying to catch up, because otherwise I'm just too far behind. So that was one example. The other is a great provider that we work with who runs a great practice in Nebraska. And Dr. Schroeder is really, really excited about our solution because it was able to give him two things, gave him back some time, which he was able to then get back to some semblance of a real life, not just be spending time on lab results all the time. And he was able to expand his practice because it was then really feasible for him to spend some time thinking about hiring some other providers that could come into the practice and help him. Before that was difficult because he just had no time to think. So I think being able to improve his quality of life and at the same time being able to help him expand his business because he could now have some time to think about it was really good.
[00:22:10] Speaker D: Love both of those stories. Sounds like they really, really caused a great positive impact to the practice. So I want to have you any of our listeners. They may be physicians or they may be healthcare system executives. If they wanted to start using affinity in health, how do they do that? And is there a way that.
[00:22:33] Speaker B: Do you have a trial?
[00:22:34] Speaker D: How does this work?
[00:22:35] Speaker C: Yeah, we encourage customers to try our product. Once you try it, you love it. It's one of those things that within a week or two of using it, you'll find it indispensable and you won't want to live without it. Imagine getting 30 or 40% less volume in your inbox every day and having better customer satisfaction with your patients because they're getting responses faster. That's a win all the way around. So it's definitely a solution that we encourage people to try. Our website's www.afenion do and you can go there to get more information or send an email into infoffinion.com we encourage people to try it. We encourage them to get their providers in the practice to try it. That's the best way to really get a sense for it.
[00:23:24] Speaker D: I would definitely recommend doing that because we use ChatGPT and once you try it, like you're saying you can't not utilize it anymore. So it sounds like a great way for, you know, provider who wants to save more time, get some of their time back for things like they want to enjoy and really give more time for their patients. This is something that they should all look into trying. So I really appreciate you sharing this with us. Tell us, what do you see? I mean, Thinking more towards the future, what do you think in the next five years? What are going to be these big things that physicians should be preparing for from a technology standpoint?
[00:24:02] Speaker C: Well, I think technology is just at this incredibly exciting time. We've all seen it and what you've had is just a little taste of what's coming. But I think for providers, one of the most important things is the relationship with the patients. There are often times when a provider could help a patient potentially prevent illness or prevent it from getting worse, but often they don't know about it because the patients are reluctant to just phone their provider up or send them a message. They know they're busy. I think technology is going to enable a whole new level of relationship between the patients and the providers that will start gradually and will eventually get to the point where providers can be more involved in the lives and in the health part rather than the sickness part of healthcare. I think that will have the most profound impact on healthcare in general. And it's not rocket science. My pup, I take my pup to the vet and I get a call back the next day and three days later and a week later on. How's Wolfie doing? How's that medicine working?
How many times have you had that happen in healthcare? How many times has a doctor called you out of the blue for no reason? It just doesn't happen. And it's not because they don't want to. It's because they have a panel of three or four thousand patients and they just don't have time. So I think technology over the next three to five years is going to change that so that now providers can be more involved, can be in more of the health rather than the sick side of it. And I think that could make a massive difference for all of us.
[00:25:38] Speaker D: Well, we're lucky to have you on today. So I really appreciate your time sharing these insights and really sharing how AI is going to hopefully improve the work life balance for a lot of our.
[00:25:53] Speaker B: Providers out there that are so overburdened.
[00:25:55] Speaker D: Right now and experience burnout and, you know, the desire to, you know, wanted to even retire sooner. So I think this is an exciting development from a tech standpoint. And again, just go ahead and tell us again how to get a hold of you and your team.
[00:26:12] Speaker C: Absolutely, yeah. It's a real pleasure to talk about this topic. It's something I'm obviously passionate about. So we can be reached at www.affineon.com or if you want to send an email infoffenion.com and my email address is david.norrisffenion.com so feel free to reach out. We love to talk to providers and care teams and we'd love to help.
[00:26:34] Speaker D: Thank you so much David. Appreciate your time today.
[00:26:38] Speaker C: Thank you very much. Bye bye.
[00:26:40] Speaker A: Thank you to all of our listeners. If you would like to be notified when new episodes air, make sure to hit that subscribe button. And a big thank you to Pacific Companies. Without you guys, this podcast would not be possible. If you would like to be a guest, Please go to www.pacificcompanies.com.
[00:26:58] Speaker D: Thank.
