E146 | How AI Can Tackle Disparities In Healthcare Access with Estenda Solutions' Richard Kedziora

The US is slowly moving from sick-based care to value-based care, where, you know, the healthcare system you're being incentivized to keep us out of the doctor's office. Yeah, it's good. Transition's being made, but yeah, that's the combination of the wearables, remote monitoring AI. Yeah. I have a huge impact. Welcome to AI for Farmer Growth, the podcast from pioneering Farmer Artificial Intelligence entrepreneur Dr. Andre Bates. This show aims to demystify AI for all those in biofarmer. From start-up biotech right through to big farmer, each episode focuses on all things AI and future tech related to help farmer execs to navigate through the and benefit from, that sometimes confusing, but magical world of AI-powered tools to grow and get real world results. Today's episode is how AI is bridging healthcare gaps. The integration of artificial intelligence into healthcare represents a pivotal shift towards a more inclusive, effective and personalized medical care system. Today we're going to be looking at how we can leverage AI to overcome some of the long-standing challenges in the healthcare system. My guest today is Richard Kedsiora, who's the co-founder of Estenza, which has 20. Well, you have 25 years of software product design. I'm not sure how long Estenza's been going. Can you tell us a little bit about your background and what led you guys to co-found Estenza and how you got there? Yeah, absolutely. It's crazy when I can talk about decades now. I was fortunate in grade school back in the 80s. I was my parents were able to purchase their personal computer and I played my fair share of games, but was always fascinated with the power of technology and programming. So even back then, I was playing with computers. And did intend to go to college in the early 90s and had thoughts of getting a PhD in artificial intelligence even way back when people sort of forget AI's been around for a long time now. Yeah, 50, 60, 70 years now. Yeah. Yeah. And in my case, I got that job offer as I was getting ready to graduate and applied at grad school's kind of thing. I was like, I pay more debt to regret in the world. So I went out in the world, but I have got to play with AI in its various permutations over the years and through the 90s worked my way up through the software development ranks and very quickly realized technologies is much about people and process as opposed to the technology itself. The technology is important. And then back in 2003, started Estenza Solutions, a current company that still goes strong 20 plus years later. And we have focused predominantly on the idea of digital health very much on the patient facing side. Either working with healthcare providers, corporations, large smalls and startups, helping them develop the tools and technologies that they're using out in the real world to improve patient care health and wellness is the way I've been thinking about these ways. It's not just health, but it's also wellness. There should be, I feel, a real shift even within pharma. We tend to focus on illness and treatment, but actually it would be really nice to see a shift in the world to focusing on prevention and wellness before we get to that point. Yes. So are your clients mainly hospitals? It's a mix of health systems that we're working with, combinations of doctors and PhDs. Some large, important 50 medical device companies. And then startups that are on their second round of funding kind of thing or an entrepreneur that's exited their first opportunity. And looking to give back and do something in healthcare, but doesn't quite understand what it means to get into the healthcare software development and from that perspective. But it's usually with the R&D groups developing new innovative approaches to evaluating data, looking at that data, presenting information, extracting information from them. Okay. And so what key areas in the area of digital health that you're working in, are you seeing AI show the most promise in addressing disparities? AI is quickly impacting many areas of health. I think first and foremost, the biggest immediate is sort of the back office, you know, kind of thing from that perspective. But from the patient side of things, there's still even with GNI, a lot of use cases. And one that comes to mind pretty quickly is just the idea of empathy and writing the letters and responding to patients and impact and understanding biases in communication. So there are biases in training data sets and things. But I think about as biases in communication and how you're addressing patients and populations, the patients as you're getting information out there, you know, the GNI is if you ask it, like, okay, are there any biases in this communication? How do we address them? You can make a real difference there. The other areas, sort of that forgotten AI of predictive analytics, still making a big impact in the field of what we can do. Yeah. From that perspective. And then probably a third, which we've been doing a lot in recently, is the idea of image analysis. Yeah. It's a great city direction, right? Right. One of our bigger projects is with the Indian Health Services, which was responsible for the native and our population here in the United States. Yeah. And the diabetic population, you can get diabetes written off at the EU, which is a leading cause of a preventable blindness. So if you find it, you can make an impact. Yeah. As we gather these images now over the years, we've gathered millions of these images, a human need to look at it. How do we apply the tools and the technology to those solutions, and can really help scale things, which is the area where that AI can help address the gaps. So what are the areas? So that's a really good one, the eye scanning. Are there other examples you could give of how it's helping these underserved or remote even populations? I think it's in making it available to more and more people. I think of it less than imaging, but in mental health and speaking. Yes, speaking. Those are the two areas where we're talking about the gaps in health care and the opportunity and how AI can help and make a difference. There's that mental health crisis, the obesity crisis here, if not globally, and there's just not enough professionals to take care of all of those people that need help. So we've been in this area of digital therapeutics for a while now. And if there's 300,000 digital health apps out there, there's only a couple of these, a couple thousand of these digital therapeutic apps. And the difference being, I think of it, if you're a pharmacologist and everything, is a medication that is going through clinical trials and FDA review. In a lot of cases, there are prescribed digital therapeutics. A couple dozen of those can make a real difference in helping people in the real world and think a lot of those engagements with the patient are on the phone. We're all carrying around our smartphones every day. Yeah, it does seem like a lot of healthcare. Well, it kind of the house and the phone, I think, of the two areas that I see the most, I've mentioned this on the podcast before, but one of my friends is an architect and she's building homes now that have AI and the bathroom mirror that basically can detect microchanges in the face or body that are indicative of particular conditions. And the toilets do urine and feces analysis each day and it all reads on to a dashboard. So they're very expensive homes right now, but that's going to come down. So I see it there and in the phone, it's true. The phone is a big health data for a start, isn't it? It's got so much data in it now. As we talk about gaps in understanding that data and think you go to your practitioner, whether Dr. Nurse, Educator, kind of thing, and trying to, I wear one of the rings that are out in the market and people, tons of Apple watches that might be all the way through. Like the ure ring, yeah. Yeah, if you take that to your provider and be like, hey, help me understand this data. They only have a very limited short amount of time with you. Yeah, of course. So the more the AIs can be used to. To putting it. Yeah. to put that data and help people make better decisions outside of that hospital, healthcare practitioner encounter the better off or other than me. And that leads into another point, which is the personalization of patient care to ensure better outcomes. So that's another big area that if they've got their own personal data that's constantly being collected on their phone or their ring or in their house, can you speak about the personalization of the healthcare with AI? - That's another huge area where that potential exists. From understanding, it's an overwhelming amount of data to interpret these days. It's, you know, that's about my 20 year journey. It's, you know, early on, even pre-EMR days, it was like, "Okay, how do we access data? What data do we have available?" - Absolutely. - Now there's so much data. How do you figure out what is important? What is relevance? - Yeah. - And then you do that pattern detection, you know, calling back to sort of old school, A, you know, for out in AI, you really understand what's going on with that data. And then provide advice to that patient in language they understand. So it's a huge component that the GNI is, we develop systems and, you know, work with people that are developing the education materials. And, you know, you really need to be cautious about the language level that you're writing at. And then as you translate materials, that, you know, it's translated appropriately, the AI systems out there can do this very well, not. Still always, you know, double check, cross check, you know, I'm going to take it for, you know, exactly as it's produced, but, you know, it does have issues. But that, again, that scalability in addressing, you know, individuality has, you know, huge capabilities today and in the future. - And what about, you mentioned back office, actually. And I do remember I had an interesting podcast with the CEO of a mental health facility. And he said they were using it for streamlining a lot of the admin sort of functions. So what kind of back office are you seeing? I mean, clinical workflows, back office, what kind of things are you seeing the most for its use? - Yeah, the one that's really coming on strong now, that we're seeing more and more, and you're hearing more and more about is the idea of ambient listening. - Yeah. - So EMRs, you know, talk to 10 doctors and 10 of them, we're gonna say, I don't like my EMR. They weren't designed, you know, with that doctor and the physician encounter in mind. They were much more defined for collecting data for building purposes, honestly. - Yeah. - And, you know, it's expanded. So it's like, how many times can you click and add these notes and the physician, you know, healthcare practitioners, the nurses, they want to look at the patient, they want to interact with the patient and listen to them and not be face down, you know, in their laptop. - Tell you. - So we've seen where, you know, Shrives come into the office, but now the technology exists and it's being rolled out where the computer, the AI can listen in the room and translate that into, you know, medical notes. - And, notably, I've heard, you know, this story, which I like where it's the patient is talking about their legs are restless at night. And so they take a, you know, Shada Whiskey, you know, to help, you know, calm their nerves down and go to sleep, kind of thing. And what's interesting is the AI translates that to patient self-medicates. So much more medical way. - No, like I would do. - Understanding that, you know. - Yeah. - In my own personal interactions, you know, I would ask it about, you know, okay, what can I eat for dinner tonight? - Mm-hmm. - That's like, okay, I try to eat a lot of vegetables, but still need to wash my proteins who tend to eat, you know, chicken a cup times a week. And it responds, it's like, okay, with your fux-soterian, you know, lifestyle. It's like, okay, it made that translation between, you know, what I was actually saying to, you know, an actual lifestyle. - Yeah. - Interesting. - That's interesting, because that's also something that I pushed quite a bit for pharma sales reps, you know, for the CRM systems, because when we're given, you know, we do a lot of AI build projects and AI analysis. And when we're given CRM data, you know, the problem with that often is that the reps haven't filled it in because, you know, they're rushing from doctor to the, just on a few spotty notes and that's it. Whereas actually, as you're saying, that listening ambient listening could actually fill that in and actually fill it in, potentially a lot better than the reps. Certainly in many reps, I see the data that's about doing it. - It can do that, it can do that translation, recording the audio. But think of where it can go, you know, tomorrow. So now it's listening to that information. It's aware of what's in the medical record system. It can start prompting questions. - Yeah. - How it prompt to the physician to then, you know, decide if the questions are worth asking, to then engage, you know, better engage the patient. And then provide, you know, feedback around that. You know, you think of the your average physician and I saw recently, you know, medical information research doubles every five years. - Yeah. - We, we, as people, humans can't keep up with that amount of information, but a computer can. - Yeah. - So particularly in cancer, so much new information is coming out, new research is coming out. In cancer is such an individualized, personalized, challenge for people. It's like, okay, how can you be aware of, okay, there was this one other study done, you know, across the country with a patient that looked just like me. - Yeah. - Add similar conditions, similar age, similar, you know, manifestation, you know, presentation of, you know, my cancer. Okay, that worked for them. Let's try this over here. - So physically impossible for the doctors, for example, to keep up with that. And I can't remember when it was, but it must have been at least 15 years ago, where we were doing a project with AI in oncology in with medical literature. And it was at that time, and as you're saying, it's gone up, increased since then. But at that time, there was 5,000 new articles every week published in oncology. And that's, you know, every week, no human can read 5,000 articles in a week. So yeah, it's, I mean, it's a godsend, if you've got the systems right. - Right. - And what about research, 'cause you mentioned research earlier on, what ways are you seeing it, contributing to advancements in medical research? - I like doing my personal situation, but I really like asking questions and brainstorming, trying to figure out what that next level is. - Yeah. - Because actually a project we've submitted to the SBIR here under their, their programs kind of thing, around the idea of smell testing. And smell testing and the loss of your smell, something that has been underappreciated until COVID hit the world kind of thing, and you started losing your sense of smell of COVID. It's like, okay, now there's a lot more people starting to pay attention to it. And one of the things the NIH does is have specific repositories for different sets of information out there. And there isn't one just yet for smell data, all factory testing, and a lot of ways of testing that smell data. And so we're gonna set up a repository of this information, and then use that AI to allow people to query and interact with that data, understand and discover new patterns, new trends in that information. And what we hope to do is I'd be able to aggregate data across all sorts of studies, different tests, normalize that data. And then instead of having to understand the R programming language or SQL and having that extensive background to do that, you can just enact your language prompt and just say, hey, tell me about this relationship. So it really opens up the idea of what you can do in terms of figuring out the next steps. - Yeah, absolutely. - You think it's like, wow, which is very powerful. - And what about, so you're in the US, so there is a lot of data in the US. I remember we were doing a project where there was, I think I can't remember the HR data or claims data, but it covered, it was like 80% of lives covered, but there's still that extra 20% missing, isn't there? So what kind of strategies have you seen, speaking of healthcare gaps in place so that we can kind of access data that is harder to get though that we can actually get data on these groups that aren't actually in the majority of the mainstream databases? - Yeah, you know, I come with, and a lot of things that the technology had on it, I think of what is one of the biggest challenges in AI today, and it's access to that data, understanding that data. So I wish I hadn't easy answered to that question, but that is definitely a challenge. It's, I started, you know, at the beginning talking about people and process as being important, and that is a key aspect of this. There was just a recent research report that women are still underrepresented in clinical trials. - I know. - Surprise, but you know, this is what the research was demonstrating. And as you develop these AI systems and you're trying to get access to that data, there's bias in the training data, and that is one of the huge challenges today in these AI systems. So if we help develop one AI system at one particular hospital, one health system, that might have a bias related to their patient population. So now you might go through an FDA review of process and be approved as good, then you take it to another practice, another health system. You still have the test of there to make sure that it works for that population and be aware of those biases. So I think to your question of getting to that information from the underrepresented populations here in the US around the world is very much an educational marketing component to that. A lot of people sort of use HIPAA as a defense for not sharing, I can't share this information because it's HIPAA. Well HIPAA was not designed to silo the data. You can use it, you have to be aware of privacy rules and regulations and use it appropriately with GDPR over in the EU. It's like if you do get that permission from the patient just looks how it's being used, you can use this information. And speaking of privacy, what kind of measures are in place that you're seeing that to ensure the privacy and security of patient data because it is the most hacked kind of data. I mean, I can't remember the exact numbers. I looked at it for another podcast on security a few months ago and I mean it was like 101 a few years ago, it was 145 million people's healthcare data had been hacked and then I think it was around 333 million or something like that in 2023. So it's, you know, there was that big case, I think it was 2017 where in the UK lots of hospitals were held ransom because their data was the target and operations, everything had to stop because they couldn't get information on the patient. So huge areas. So yeah, what kind of things are you seeing in the hospitals today to avoid that kind of situation? Yeah, it's technology is important, but it is so much more about people and process. Most of those hacking attempts and stuff like that, they're very much driven by inside actors or just receiving one of the nefarious email and someone clicking on it. Yeah, they weren't very aware, less person into the back door. So technology is a strong component of it. But training, you know, the people and having a process is in place to protect that information is paramount. You know, on the dark web, a patient medical record with some identifying information is worth more than the credit card because then you can go out and you can buy the drugs and the medications that you then resell and so yeah, it's unfortunately that medical record information is very valuable, let alone you get out in the real world and you know, I don't want the world knowing I have PTSD. You know, we have a lot of protections in place around that and it not being used nefariously, but you can, you know, go down the rabbit hole of okay, the insurance company finds out that I have or even not even insurance company, you know, nowadays, you know, is I hire people kind of thing? What happens if I find out someone has attention deficit disorder? Yeah. You hire them, you know, and that is very much an extreme, you know, situation, but that's, you know, where that potential can go is that information is stolen and available. Yeah, and there's that dichotomy between data access and confidentiality, you know, that kind of, how do you balance that? Yeah, it's, it's, we always look at these projects as we help people with this and develop systems and it's about risk management. That's, yeah, we drill that in every day, you know, so it's like, how do you eliminate or mitigate the risk of this information getting out there? And how do you segment that control of the information? You just don't have everything in one database. So it's like if it's back attack, then all of the information is out there. You know, how do you segment, you know, control and access to that information, only make it, you know, available to those that have the need for an authorized by patient. Hi, my friends work through a company where they basically, every time someone logs in, it's a new thing. And she said, she always says zero trust, zero trust, that's a mantra, zero trust. So where do you see the future in terms of what kind of AI innovations do you anticipate are going to have the greatest impact on bridging healthcare gaps? Yeah, we talked a little bit about it. And I think it is that area of combining wearable information, some aspects of that remote home monitoring, so your healthcare practitioners can be aware of that information and how AI can use that data and make a difference. You know, you as an individual, you get sick, you go to the doctor, you're there for seven to ten minutes, you know, 15 if you're lucky, so much of what you do outside of the four walls of the house, all the doctor like, makes a difference in your life. Yeah. So as, you know, I have trouble sleeping last night, okay, what's going on, you know, it's like, or you might potentially be getting sick, you know, so take it easy, you know, the next kind of day, make sure you, you're hydrating and eating better kind of thing to help mitigate, you know, getting sick. I think that's where the technology, it's in that personalized medicine understanding the data is where we're going to see the biggest difference. The US is slowly moving from sick-based care to value-based care, where, you know, the healthcare system is being incentivized to keep us out of the doctor's office. Yeah, that's good. The transition is being made, but yeah, that's the combination of the wearables or no monitoring AI. Yeah. And I have a huge impact. Fantastic. I'm very much for your time today, I really appreciate it and such an important and growing need for this area. So thank you. Thank you. Thanks for listening to this episode of AI for Farmer Growth. If you have received huge value from this show, or maybe this episode has highlighted how you can use AI in your company, then we would love for you to support the show by leaving us a five-star rating. Please make sure you hit the subscribe or follow button on your podcast app now to make sure you never miss an episode.