AI and Healthcare with Sean Cassidy

george grombacher 0:02
Sean Cassidy is the co founder and CEO of lucem Health they’re leading provider of transformational clinical AI technology and solutions. Welcome to the show, Sean.

Sean Cassidy 0:13
Thanks, George. It’s great to be here. Yeah, excited to have you on. Tell us a little about your personal lives more about your work, why you do what you do? Well, my wife and I live with two teenagers in the Charlotte, North Carolina area, we’re Midwesterners, but we decided to head south about 20 years ago.

You know, for the obvious reasons, weather and so forth. I’ve been working in enterprise software, my whole career, but I’ve been in digital health for about 20 of those years, I’ve become very, very passionate about the ability for technology to help transform how we deliver health care. And that’s the project that we’re working on now with somehow is very, very much focused on how do we deliver better care to patients and people in general?

george grombacher 1:02
I appreciate that. Is you have been in in software for 20 years, did you see AI coming? Did anybody see coming? Or is it obvious?

Sean Cassidy 1:15
That’s interesting question. I mean, obvious. I don’t know how obvious it was. We’ve been doing predictive analytics, we’ve been using regression models and other statistical techniques and healthcare for a long time. And I think if you just think about the evolution of technology generally, and think about things like analytics, specifically, it’s sort of inevitable, or it was inevitable or natural that we wind up in the place that we are today. I, what I what I don’t think we could have predicted necessarily, was how incredibly powerful and almost magical, large language models like like open AI, chi PTR, that was something I personally did not see coming. What we do is not large language models. It’s not the magical stuff. We are using mesh really machine learning algorithms to find signals in the noise of data that human beings and traditional statistical techniques cannot find it is a form of AI. But it’s not necessarily winding up on the front page of the Wall Street Journal.

george grombacher 2:23
More of a needle in the haystack.

Sean Cassidy 2:25
Exactly, exactly. That is sort of the approach. Now it turns out, and I’m sure you and I will get into it. These techniques actually are very valuable, and can be very powerful and very impactful for patients and for doctors.

george grombacher 2:39
So you’re looking around software, you’re looking around healthcare, you’re interested in helping have an impact. What is are the problems that that you said, I think we can make a dent in this?

Sean Cassidy 2:54
Well, we, you know, there are so many issues with our health care system. And I know you cover you know, those issues in your conversations here. We are motivated by a particular concern. Right now the providers have, let’s say a couple of concerns. One is that they are under extreme financial pressure, their margins have been highly compressed, they’re struggling to make money. And, and look, we need our health care providers to be healthy financially, in order for them to be able to deliver care to us. The other thing which is overlapping and operating side by side with that is an acute resource shortage, that is only going to get worse. We’re going to be short 10s of 1000s of doctors, many 10s of 1000s of nurses and technicians and others over the coming years. That’s a huge problem. And so we’re motivated by how can we help with those two highly intersected? You know, slash overlapping problems that providers face while at the same time also doing good in the world and helping patients?

george grombacher 4:09
That makes a lot of sense. The the notion that our healthcare providers need to be financially healthy, it seems really obvious, but I don’t know that I’ve ever really thought about that before.

Sean Cassidy 4:21
Yeah, I don’t think I don’t think that the general public really appreciates how difficult it is to be in the healthcare business right now. And in the United States. Of course, it is a business is a employs a huge number of people. These are well paying jobs. These are important jobs. And you know, they are there they suffer from changes in reimbursement models, whether that comes from Medicare and Medicaid or from private insurance companies who tend to follow the government’s lead in inflation has have been a problem that the cost of everything that goes into delivering health care has gone up, wages have gone up, and so on and so forth. So like a lot of businesses that have suffered in recent years, they’re no exception.

george grombacher 5:16
So how are you working to address the shortages?

Sean Cassidy 5:22
Right. So this is where the technology comes in, and where the the AI and machine learning stuff can be really interesting, quite powerful. So we ask ourselves kind of a what if question, what if we could help health care providers deliver more with the resources they already have. So imagine a clinician, a doctor and nurse, who, when they, they go in to the exam room, or when they go into the clinic, imagine if the clinical value of what they deliver is higher during the day, during the week, meaning that they are, they are able to find more diseases, make more diagnoses, they are able to be more productive, and so on, and so forth. So if we can help them, we can make their jobs easier in effect, and allow them to do more with our to do more good without doing more we sometimes say, then that can be hugely beneficial in terms of the productivity of those resources, then, of course, if you have more productive resources, then you can see a financial benefit from doing that the way we all get richer, right? It’s and we all get better off is by increasing productivity. Healthcare, again, no exception.

george grombacher 6:44
Which makes sense. So what are some actual use cases? How does that? Yeah. Yeah. Yeah,

Sean Cassidy 6:51
so So what we’re doing is really quite simple. There is a tremendous amount of data in healthcare. So over the past decade, plus we went through a process of digitizing health records. The government provided a lot of financial incentives to do so. So most healthcare records these days, particularly in hospitals, and clinics, the places we normally go, are digitized. We have machines, you know, modalities, right telemetry device, ISIS collecting a lot of information from patients, either in inpatient settings, increasingly in outpatient settings, think of ECGs X ray, CT images, that kind of stuff, again, digitized and available, right? So you got all this data? And it’s sitting there, right? And what is what did that data come from? Well, when we go to the doctor’s office, or we wind up, God forbid, in the emergency room, or we have a surgery or whatever, there’s this data being collected, and then presumably, we get better, we get discharged and we leave, right? And Bob’s your uncle, it’s my father, my Liverpudlian father likes to say, right, but this data is sitting there, and there’s so much value in it. And that value is hiding in plain sight. So what we’re doing is we are using very powerful, very novel machine learning algorithms to go in and not not quite mine the data, but to do very specific kinds of analyses looking for risks of certain diseases. Now, I’ll give you the simplest example, I think most of your listeners will, will it will resonate with them. We have a solution focused on on lower GI, right on colonoscopy is. So we go in to a provider’s healthcare data set. And we select patients between 45 and 75, who are overdue for their screening colonoscopy. By the way, this is millions of patients most Yes, people don’t like colonoscopy. I just had mine two weeks ago. It’s not a lot of fun. But it’s a very important thing. So, so people are overdue, and they’re avoiding that procedure, which can be a life saving thing, right? We need a certain a certain clinical observation, a lab value, not really so important, but we take that data and we run it through our solution. And what perhaps out on the other end is the top three to 5% of patients who are at highest risk for a significant finding. When they do appear for their colonoscopy. Big significant finding is cancer and adenoma, polyps, etc. And what we find when we run it through these very sophisticated algorithms is eight to 10 times more cancers and if we selected from that population at random, more than doubled, the adenoma isn’t polyps. Okay, so now let’s come back to what we were saying before. So why is that useful? Right. So we surface this list of patients, and we have clinical staff, it’s not doctors, not nurses butts off for staff, its care coordinators, they reach out to the patients in Georgia, it turns out that, you know, people will ignore those letters they get from their insurance companies saying, You got to get your colonoscopy, you got to get your colonoscopy. But if your healthcare provider calls you and says, George, we went and we looked at your records proactively, and we found something, you know, you’re naughty, you’re overdue for your colonoscopy, you really need to come in, because there’s potentially some risk there. And you don’t want colon cancer, the response rate you get to that is amazing. It’s 70, plus percent of people do end up actually presenting appearing for their colonoscopy. And again, when they do we find more stuff? Well, so let’s talk about about benefits of that. And let’s come back to the productivity question and what it does for providers. So for a patient, finding cancer early, is a game changer. It’s so much easier to treat the stage one colon cancer than a stage three or stage four colon cancer, which can be a death sentence. So that’s great for the patients. What about for the providers? Well, you’ve got gastroenterologist working in the GI suite, they just see patient after patient after patient, right. But if the yield from what they’re already doing is higher, they’re finding more cancers and more adenomas, they’re doing more good. And the whole system around them is more productive. Because, again, let’s be crass about it. But there are real dollars associated with treating cancers. Right? I mean, it costs quite a lot of money to treat a cancer, it costs more money to treat patients with adenomas, and so on. And so there is a financial benefit that accrues to providers. But it also is making the providers doing those procedures more productive. Basically, every solution that we’re delivering with AI, George operates roughly the same way, we grabbed some data, we run it through an algorithm, we’re finding patients that higher risk for a certain disease could be a cancer, it could be diabetes, it could be stroke risk, it’s a solution we just launched. And then we get those patients to present into the healthcare system. And we allow doctors to just be doctors to confirm a diagnosis if there is one. And then to put that patient on a treatment plan. So that they can avoid winding up in the ER, they can avoid dying from cancer, and so on. That’s what we do.

Seems like a no brainer.

That’s what we think. Yeah. You know, we are very gratified. And we’re not geniuses. We’ve been doing this a long time. And we’re definitely, you know, we’ve come through the school of hard knocks. But we are gratified that the response to the approach we’re taking, which is very simple, it’s easy to integrate into the existing care delivery workflows and system. It’s been very positive. So that feels good. We’re gonna keep doing it as long as we can do it.

george grombacher 13:01
Makes a lot of sense. In terms of privacy concerns, regulatory stuff, how is that? How has that been?

Sean Cassidy 13:12
Yeah, so great question. And something we should all be concerned about, as your listeners are probably aware of, we have a lot of rules and regulations about the use of healthcare data and the privacy and protection associated with health care data. We are just at the level of our operating model and culture and values extraordinarily careful with healthcare data, we’ve been our team we have, I don’t know how many decades on our team of working in digital health and health technology. That’s one of the many things that keeps me up at night is the idea of a data breach. So we’re extraordinarily careful with patient data, we are allowed as what’s called a business associate of a provider. So we enter into an agreement with our provider partners, to be able to use existing healthcare data to serve patients to help treat patients better. So if you think about what we’re doing, here we are, we are legitimately trying to help deliver better care, we’re trying to find patients who have cancers or diabetes or these other things. And so we are allowed to do that. So from a data perspective, there’s sort of a long standing set of norms associated with with the kinds of things we’re doing. So that’s good. on the regulatory side, the FDA has started to regulate AI and machine learning algorithms as medical devices under certain circumstances, in particular, when they’re used for diagnostic purposes. So there are tools out there, where I can run, let’s say, an ECG, you know, the output of an ECG through it and it will diagnose AFib they will say, George, you’ve got a font or you have some other condition of the heart. And doctors are allowed to rely on that information as a as a diagnosis just as If you got a diagnostic test on your blood, what we’re delivering is a clinical decision support tool that is upstream from diagnosis, we are not a diagnostic tool. And so we are not subject to at least right now to regulation, by the FDA. Things are very fluid FDA as it relates to software and technology, because obviously, we can move software so much faster than we can move, let’s say a molecule, the development of a drug. And I think the FDA is actually doing a really good job of trying to create a framework where we can all feel comfortable that these technologies are being appropriately used to deliver care, while giving innovators degrees of freedom to be able to continue to move things forward again, for the benefit of all of us.

george grombacher 15:46
Well, I’m glad to hear that. And that certainly makes sense. who are who are your clients, I imagine it just listening, it makes sense to me apologize that one of the big challenges for insurance companies, and really for employers is rising costs when somebody gets really, really, really sick. So this strikes me as this is a wonderful wellness tool to help people to stay healthy and to get in front of these things like you’ve been talking about. So I’m sure that insurance companies are probably pleased employers are pleased.

Sean Cassidy 16:22
Yeah, so insurance companies have a very strong interest and incentive in identifying diseases or detecting diseases earlier, it is less expensive to treat a stage one cancer than a stage three cancer, it is less expensive to have a patient on a therapy like a blood thinner or what are statin or whatever, rather than having them show up on their back in the emergency room, and so on. So payers absolutely care about us. You have asked Who are our customers, who are we working with? We work with providers, we’re trying to help providers deliver more proactive care, deliver better medicine and fall for these productivity and margin issues that I mentioned before. However, we are working on a set of relationships with folks like payers and others who see the value of what it is we are delivering and are willing to and interested in sponsoring the deployment of these solutions that providers and something payers have been doing for quite a while now, which is helping providers to the extent providers want their help deploy technology.

It all makes sense. Just the more hands you can get into the certainly the better. But at the end of the day, you need to get in front of people and and make your case and I’m sure that the buying decisions are not necessarily fast ones, but maybe I’m wrong.

They’re not, you know, because providers are financially stressed and have invested a lot in technology over the last decade. Plus I mentioned that before digitizing health records, a lot of those investments have not paid off the way they expected. And so they’re their IT teams are often focused on trying to get value out of investments they’ve already made. If you can come in, though, with a very clear value, proposition and ROI. And you can show that it’s simple and straightforward to deploy. We find a very receptive market for that sort of thing, even when times are tough economically, financially in the industry. And then just to your other point. Yes, health care delivery in the United States is a partnership between providers, payers, drug companies and device manufacturers, employers, you mentioned them before the government in various forms and the local and state level or at the national level. And so we should all have an interest in helping though at the tip of the spear that says helping our hardworking doctors and nurses and technicians and other others deliver better care. That’s what we’re focused.

Love it. Well, Shawn, thank you so much for coming on. Thanks. Certainly for all your work. Where can people learn more about you? How can they connect with lucem health? Well,

please come find us at Lucent health.com, we’ve got some good material out there. We are publishing the white papers and other things on AI on early disease detection. So if people are interested in learning more, we’d love to see you there. And we we talked to everybody. So come chat with us and we’ll share more about what we’re up to. Excellent.

Well, if you enjoyed as much as I did, showing your appreciation and share today’s show with a friend who also appreciates good ideas go to lucem health.com. That’s Luc e m health.com. And check out everything that we’ve been talking about today and so much more. Thanks again, Shawn.

It’s a pleasure, George. Thank

you. Till next time, remember, do your part by doing your best

Transcribed by https://otter.ai

george grombacher 16:00
So if I want my iPhone, and my Tesla and my Bitcoin to work, we need to get the metal out of the ground.

Pierre Leveille 16:07
Absolutely. Without it, we cannot do it.

george grombacher 16:13
Why? Why is there a Why has production been going down.

Pierre Leveille 16:21
Because the large mines that are producing most of the copper in the world, the grades are going down slowly they’re going there, they’re arriving near the end of life. So and of life of mines in general means less production. And in the past, at least 15 years, the exploration expenditure for copper were pretty low, because the price of copper was low. And when the price is low, companies are tending to not invest more so much in exploration, which is what we see today. It’s it’s, it’s not the way to look at it. Because nobody 15 years ago was able to predict that there would be a so massive shortage, or it’s so massive demand coming. But in the past five years, or let’s say since the since 10 years, we have seen that more and more coming. And then the by the time you react start exploring and there’s more money than then ever that is putting in put it in expression at the moment for copper at least. And what we see is that the it takes time, it could take up to 2025 years between the time you find a deposit that it gets in production. So but but the year the time is counted. So it’s it’s very important to so you will see company reopening old mines, what it will push also, which is not bad, it will force to two, it will force to find a it will force to find ways of recalibrating customer, you know the metals, that will be more and more important.

george grombacher 18:07
So finding, okay, so for lack of a better term recycling metals that are just sitting around somewhere extremely important. Yeah. And then going and going back to historic minds that maybe for lack of technology, or just lack of will or reasons, but maybe now because there’s such a demand, there’s an appetite to go back to those.

Pierre Leveille 18:33
Yes, but there will be a lot of failures into that for many reasons. But the ones that will be in that will resume mining it’s just going to be a short term temporary solution. No it’s it’s not going to be you need to find deposit that will that will operate 50 years you know at least it’s 25 to 50 years at least and an old mind that you do in production in general it’s less than 10 years.

george grombacher 19:03
Got it. Oh there we go. Up here. People are ready for your difference making tip What do you have for them

Pierre Leveille 19:14
You mean an investment or

george grombacher 19:17
whatever you’re into, you’ve got so much life experience with raising a family and doing business all over the world and having your kids go to school in Africa so a tip on copper or whatever you’re into.

Pierre Leveille 19:34
But there’s two things I like to see and I was telling my children many times and I always said you know don’t focus on what will bring you specifically money don’t think of Getting Rich. Think of doing what you what you like, what you feel your your your your your, you know you have been born to do so use your most you skills, do what you like, do what you wet well, and good things will happen to you. And I can see them grow in their life. And I can tell you that this is what happens. And sometimes you have setback like I had recently. But if we do things properly, if we do things that we like, and we liked that project, we were very passionate about that project, not only me, all my team, and if we do things properly, if we do things correctly, good things will happen. And we will probably get the project back had to go forward or we will find another big project that will be the launch of a new era. So that’s my most important tip in life. Do what you like, do it with your best scale and do it well and good things will happen.

george grombacher 20:49
Pierre Leveille 21:03
Thank you. I was happy to be with you to today.

george grombacher 21:06
Damn, tell us the websites and where where people can connect and find you.

Pierre Leveille 21:13
The it’s Deep South resources.com. So pretty simple.

george grombacher 21:18
Perfect. Well, if you enjoyed this as much as I did show up here your appreciation and share today’s show with a friend who also appreciate good ideas, go to deep south resources, calm and learn all about what they’re working on and track their progress.

Pierre Leveille 21:32
Thanks. Thanks, have a nice day.

george grombacher 21:36
And until next time, keep fighting the good fight. We’re all in this together.