This week I am talking to Tom Stanis (@TomStanis), CEO of StoryHealth (@StoryHealthAI), a company that combines virtual care, remote monitoring, and AI to bridge the data gap between health system specialists, primary care doctors, and patients to extend care for patients with severe conditions with their first focus around heart failure, as they return and manage their condition at home.
Tom shares his origin story and the chance discovery of cancer in his kidney following a bad road traffic accident that essentially saved his life. This changed the trajectory of his life and work moving away from helping people click on ads and buy apps to building Verily, one of the first alphabet companies where the focus was on improving healthcare. Following success with Diabetic retinopathy, he set his mind to bringing proven scientific advances faster into general use.
In the case of Congestive heart failure – only 1% of patients are receiving guideline therapy. The reasons behind this are many but much has to do with the intermittent nature of interactions that happen between the clinicians and the patient extending the timeframe and the overwhelming nature of large amounts of data and inputs that are being generated for each and every clinical decision point.
Listen in to hear him describe their approach to gathering the data and helping provide toolsets that physicians can understand that will implement the existing guidelines using automation and rapid feedback mechanisms accelerating the process of adjusting and finding the best path for patient care as quickly and as economically as possible
XXXX
Listen live at 4:00 AM, 12:00 Noon or 8:00 PM ET, Monday through Friday for the next week at HealthcareNOW Radio. After that, you can listen on demand (See podcast information below.) Join the conversation on Twitter at #TheIncrementalist.
Listen along on HealthcareNowRadio or on SoundCloud
Raw Transcript
Nick van Terheyden
And today, I’m delighted to be joined by Tom Stannis. He is the CEO of story health. Tom, thanks for joining me today.
Tom Stanis
Glad to be here, Nick. Thanks for having me.
Nick van Terheyden
So as I do with all of my guests, I think, important to get a little bit of the backstory, you’ve certainly got an interesting one with some exciting elements. Tell us a little bit about how you got here and your journey to this point, that story health, if you would,
Tom Stanis
yeah, happy to my background is not actually in medicine. I actually started off as an engineer, I worked in the early days of Google building the systems there. And then about 10 years ago, I was riding my bike in Northern California where I live and I got hit by a car I have, I have no memory of any of this. Because I had a concussion, I ended up waking up in the emergency department at Stanford. And they said, Well, we’re gonna put you into a CT scan to see whether you have any broken bones, one of the CT scan came out and said, Well, good news is you don’t have any broken bones. However, there’s this mass in your kidney. And it turns out that I had stage one kidney cancer had no idea and there’s no way of knowing when it’s stage one, there’s no symptoms. By the time you do have symptoms that typically stage four, and the chance of survival changes dramatically. Stage 190 5% chance of survival, stage four less than 5% chance of survival. So I like to say pretty directly that getting hit by a car saved my life. But it also made me step back and think about am I really spending my life when the most important things to me is helping people click on ads and buy apps. Is that really what I want to spend my energies on? And I realized no, what I wanted to do is actually give people more the experience I had where some technology just dramatically changes the where you’re heading with your health. So I came back to Google met a bunch of people. And we ended up creating Verily the first bet and alphabet really focused on healthcare. I was one of the co founders there and I ran all the software projects, everything from AI for detecting diabetic retinopathy, one of the leading causes of blindness in the world, one of the first AIS and medical devices that was actually out there. And the other project that I really am particularly proud of was on duo which I was, which was my baby, from the beginning, a fairly unusual was a diabetes program that really is still going strong. And it’s really changed the way that we care for patients that that need continuous care. While I was working on that, I really fell in love with the idea of care delivery and closing the gap. We worked on all these great science projects. But there was a huge gap between what we were able to prove in the science lab and what was actually happening in real world medicine. So I started story health really to close that gap.
Nick van Terheyden
Well, first of all, I’m thank you for sharing that story. I’m glad that you did find that and I would 100% agree that, you know, your accident really did because the stats that you describe are exactly accurate. That’s one of the cancers that we know, it’s so often late to the party in terms of showing itself much like pancreatic cancer. And you know, that’s great news. I do have an ask if you could maybe work on the other hundreds of 1000s of engineers and have them have that same realization. I think a lot of people would really appreciate it. Yes. But in the meantime, let’s talk about how that sort of pivot and change. So you had some formative experiences and, you know, certainly verily has gone on to progress through I think, you know, to be fair, it’s not been straightforward journey, because as Google and verily discovered, as you know, some of the other big players, oh, let’s do healthcare. This is easy. It’s not right, right. Can you share a little bit of your experiences and learning you obviously had some success with the diabetic retinopathy? And I’m imagining that was image and image analysis, and the on juro? Tell us a little bit about that, if you would, and any learnings that you had through that experience?
Tom Stanis
Yeah, well, I think most one of the challenges that a lot of engineers have when they enter healthcare, as you said, of these, you know, lots of people like myself that were interested is, we come in a little fabric with our approach, right? We need to actually learn the hard way, what it is to truly validate devices to prove that we can do what we say we can do. And so I think eventually, you’re seeing that that just takes a while to prove out, right. So big pipeline running there, and you’re starting to see the fruits first fruition of that coming to market. But But medicines, the stakes are high, and we have to be very careful that we’re actually helping patients. There’s been, you know, long history of scientific ideas. That sounded great, but then they didn’t prove out right. So we want to make sure that we don’t do that.
Nick van Terheyden
So that’s absolutely true. And, you know, that’s very much part of the scientific discovery process. We we go down blind alleys. Yes. You know, one of the, one of my favorite sort of sayings is, you know, the more mistakes you make, the more you opportunity you have to learn, but you have to learn from them as part of that process. But I think, you know, one of the things that really, we see a lot in medicine is we make progress, but that progress is not widely distributed. And that’s not always about equity. You know, that’s one of the certainly one of the significant problems, but even excluding that, we failed to get to the point of wider distribution. Tell us a little bit about that, if you would,
Tom Stanis
yeah. So I this is something we see in our day to day work all the time. You mentioned access is a big issue. Certainly, that’s that’s certainly part of the problem, even beyond access, though, is how was care delivered. And in most cases, if you have a continuing condition, you’re going to see a specialist at a clinic visit maybe once every three months. And just seeing a specialist every three months is simply not enough time, especially if it’s only for 15 minutes, you’re just not going to get much time you have so much to talk about. And so we see this time and time again, lots of registry, studies have shown that patients don’t get guideline directed therapies. And these visits. The one that I like to talk about, and you’ll probably bring this up later is is heart failure. We know that less than 1% of patients are actually on guideline directed therapy for heart failure, even though it is the biggest killer in our country, bigger than cancer and AI is just as serious. So that is an opportunity for us to think about a different way of caring for patients. And why is that? Why don’t more people actually get the right therapy, a lot of it comes down to if you’re only doing doing a visit, and you have to make lots of small changes as people adapt to therapy, it’s gonna take years to get there, and then in those years, are going to be hospitalized, unfortunately, and then we have to almost start all over again. So that that is a real challenge. And then the physicians and the MPs that are doing this work, they really don’t have much idea of what’s going on between visits. So they don’t know how to optimize your care, when you’re home, maybe you’re having side effects. Having bradycardia, low heart rate, there’s no way for them to know these things. So that’s, that’s a real challenge. And then finally, as you mentioned, patients run into all sorts of barriers with the system themselves. Like they almost have to be experts in how to navigate because if you show up your pharmacy, and you want and one of these new medications has been prescribed, a lot of times the pharmacists will tell you well, there’s a $900 copay for that medication. Right. And what do patients do they walk out, they don’t get the prescription filled out, even though there was probably a path there, but it’s really just too hard for patients.
Nick van Terheyden
Yeah, you highlight a number of challenges for the patient journey, you know, including obviously, the cost element and the cost shifting that’s taken place. Before we get into that, maybe you can help people understand that that 1% Because I think even to me, that’s a little bit shocking. I knew it was bad. Yeah. But I have to say I didn’t think it was quite that bad. Yeah, that’s an astoundingly low percentage of individuals that are following best practice. If we just improve that, if we doubled it, well, let’s not get it. But if we just improve that and focused on that alone, yes, I’m imagining a you know, given it’s one of the top three killers, early intervention and so forth, lots of opportunity to not only save lives, but improve health and wellness. What is going on there? And what is what is it that people are missing? And why?
Tom Stanis
Yeah, so fundamentally, you have to understand a couple of details here that I think make this a little bit clearer. So there’s four different medication classes, part of that are part of that guidelines. You have your beta blockers, your ace inhibitor, or or RNA drugs, you have your MRIs. And you have this new fancy drug called stLt intuitive pairs. So lots of fancy different drugs we can use to treat disease, you can’t just prescribe it and be done would be great. If we could do that. We have to start off at a small dose of each of them and slowly increase them as the patient tolerates them. And of course, every patient is going to have a different level that they tolerate, at which point they’re going to develop low blood pressure, maybe some symptoms, maybe other things, but we need to push them as high as we possibly can. You can imagine if you were to look at all accommodations, we have to go through probably 15 different steps in the best case scenario with these patients to actually get them on the optimal therapy 15 different medication changes is just there’s not enough time there’s not a visits there’s not enough slots, enough cardiologists that there’s simply not enough with an existing clinic based care model. In order to be able to do that efficiently, you really need to think differently about how you’re going to do that. And that’s really where we come in is we’re going to do this much more quickly. We tried to do it on a two week schedule with the patients at home, and really measure what’s going on with them at home, and help them get through all these barriers so that the clinician is involved the entire time, but they’re more conducting the orchestra, rather than having to visit with these patients on every single visit and play every instrument.
Nick van Terheyden
So let’s talk about that two weeks. And what happens in that two weeks, are you talking about, you know, those four different treatment modalities working through all four of those in that period of time?
Tom Stanis
Yeah, so we typically change, you know, we’ll make one change medication every two weeks. But I mentioned that, like, there’s a lot of different side effects. So we need to watch out for the vital supplement side effects, the similar side effects. And for a lot of these drugs, we need to have labs drawn, frankly, to see whether you have kidney issues. So all those things we need to get through pretty quickly. And the only way to do that is really to orchestrate a set of home services to make it happen.
Nick van Terheyden
So as I’m listening to this, I’m thinking that, you know, some of this is flow, or workflow, ultimately sort of driving people through this process. I’m imagining some automation. But there’s another aspect to this, that I would, I would think would contribute to at least expedite the process is some of this newer home monitoring or remote patient monitoring. That’s, you know, starting to come of age, I want to say I’m, I think we’re seeing more of that and more clinically valid versions of it. Where do you think the best or the fastest opportunity is to start that process of improvement?
Tom Stanis
Yeah, so that is definitely part of it. The other part of it, though, is, as a clinician, typically you’re dealing with, frankly, more and more data, like if you were going to do this every two weeks, now I’m going to give you even more data to look at, right? Do you have more time to see more data? My guess is no. So we really need to have algorithms that can interpret that data, and really bring the clinician in when they need to be there and be able to say things are going fine when they’re going fine. Right? That’s that’s the thing that I think is really missing, in order to make this make this a problem is being able to really build the AI on top of that, that can act as the the can be there every moment of every day, to make sure that things are going well and be the true guardian angel.
Nick van Terheyden
So if if that’s the sort of critical component to this, you, you’re now playing what I would call interference in between the physician and the patient to try and reduce that data flow. In my mind, it’s never a data problem. It’s always a filter problem, you’re filtering appropriately. But that now starts to cross into clinical activity. Where are you with that process and validating that?
Tom Stanis
Yeah, so we think of it as the clinical term is clinical decision support, right? And there’s kind of two forms of this one is the way where, hey, I know exactly what the rules are. And I set them up myself as a clinician, and I can kind of ensure that they’re working right. And then there’s the kind of more magical version of clinical decision support where there’s some AI, that’s just like figuring it out behind the scenes and deciding, right, we are very much in the first camp today. Right? We we can set up systems that are explainable, that’s the only way to do this really in a way that, frankly, is safe, we will get to the point where we’ll have clinically validated AI algorithms that can do it. But I think it’s we’re still in the process of validating that part of it. And I want to be sure that it’s there. Because it’s a great opportunity to take the burden off because it’s really hard to tune these rules. Because every patient is different, like, Oh, what is it? What is the heart rate less than 50? Mean? Is that a problem? Well, sometimes it is sometimes it isn’t, right? So we have to be in the the promise of the AI or algorithms is they can actually see that difference, be able to interpret that. But we’re a little ways away from being able to truly trust that.
Nick van Terheyden
For those of you just joining I’m Dr. Nick the incrementalist and today I’m talking to Tom Stannis. He is the CEO of story health. We were just talking about the the process by which you start to include automation. And I think elegantly, what you’re saying is that you’re providing a tool set, yes, that essentially actions, the clinical rules that I as a clinician might use when processing these patients, but you’re automating that so that it’s really still under the guidance. As I think about that. The first question I have and you know, I always go back to a radiologist friend of mine who said If asked me one day No second longer I’m going to use it. It sounds like there’s some effort involved on the part of the clinician. Is that true and how you overcoming that? Yeah. So
Tom Stanis
I mean, there is definitely effort in terms of setting up plans and things like this. But if you think about it, the alternative is another follow up visit with his patient every month, for you know, for 10 to 12 months, that’s a lot more work. Right. So in the end of the day, I think you end up with potentially compressing a lot of time, and being able to do that very quickly. It also frees you up as a clinician. So now I have these slots available on my schedule where I was gonna have follow ups. And every cardiologist in the world is dealing with a long line of new patients waiting to see them, we just don’t have the cardiologists to deal with the volume we have. So would you rather be seeing a new patient that really needs your particular diagnostic specialty, and then be able to kind of add them to your Panel, expand your influence that way, that’s that’s the real value to clinicians is stop explaining an ACE inhibitor for 1000s Time to the 1000s patient and start doing the interesting work that only you can do.
Nick van Terheyden
So I in that model, and I agree with you to be clear, I think, you know, remove the repetitive work that tools, technology is good at process that. But there’s a kicker in the back of this tied to compensation, which in the fee for service model says, well, actually, it would be good to get the patient in because then I get income, you know, so unless we shift to this more capitated type model, that’s going to represent some challenges. But as you described, there’s this big line of people. What’s been your experience, if you’ve started to see this roll out?
Tom Stanis
Actually think there’s lots of value in a fee for service world to do this even today? Because you start to see things like actually new patients are reimbursed more than follow up visits. In second, guess what every patient that has a new patient is a chance for there to be a definitive treatment or procedure that needs to be done, a device needs to be implanted a surgery that needs to be done, right. That’s actually where a lot more of a business of healthcare comes from is finding those opportunities. And guess what, when you see more patients, you find more things, more people that need these things. So it actually opens up the fee for service world, as well as providing better care, it’s nice to I always think it was a bridge to the value based world that you’re talking about how can we treat our patients better, continue to be able to expand our business, but also be starting to lay the foundations for this value based care world that we know it’s coming in the future?
Nick van Terheyden
And in your experience? Are you seeing improvements in that interaction, particularly with patients who, you know, in some instances may say no, I want to see the physician because you know, at some point, you’re I don’t want to say limiting it, but you’re limiting it unless it’s necessary. What’s been the response to that?
Tom Stanis
Yeah, so that’s interesting. One of the things we do is, it was not just all technology, there’s a service component to this as well. So we have health coaches, and nurses that are part of our staff that work with patients and get to know them throughout this entire process. And they’re far more available than typically, the clinic staff is at a health system, because this is their job like this, what they’re focused on is getting people optimized. And what we found is that they build a real relationship. And patients think that they’re actually just getting better care, they actually don’t see a divide between our staff and the health system staff, they feel like it’s just all one, one group of people caring for them, they feel like they’re getting better service. So they love that. In addition, we are there to help them with so many other things that they run into, right? So why do people get frustrated with the health system? It’s not because it’s not because you know, the doctor did the wrong thing, that’s very rarely the case. Usually, it’s, I can’t get an appointment, or now you want me to come back and take more time off work, that I don’t have an order to come to you again, or you know, you’re, you’re late for your appointments by an hour, I have to sit in the waiting room, these are the things that patients really struggle with. And by doing it more asynchronously, they actually love that experience a lot more. And then it leads to, frankly, stickier situation for the health system. The no show rate we have for patients that come back for visits, it’s three to 6%. Whereas you see in the industry, it’s more like 10 to 20%. So by building that continuous relationship, you actually build a stronger relationships that come back to the clinic as well.
Nick van Terheyden
So as you’re gathering all of this data and you know, leaning on the clinician expertise you’re building what I imagine is a data set even sort of proof points that ultimately might sort of move and shift things to a more automated model that would obtain FDA clearance. What’s your sort of timeline where do you see that sort of coming through?
Tom Stanis
Yeah, so that’s definitely the path that we’re on. I think that we’re starting a big RCT right now to kind of prove out all of that stuff. The data from that will really help us towards our FDA submissions in the in the future, we will have to, of course, validate the specific algorithms. So this will just be the first part of it. But I think it’s a couple of years out, before we start to see really FDA cleared devices for doing a lot of this more automated.
Nick van Terheyden
And to be clear, for the listeners, when you say FDA devices, you’re talking digital therapeutics, it’s not really a device in in, yes, you know, like a wearable or something. So great opportunity for that. But at this point narrower, I think for congestive heart failure is your sort of primary target, do you have other targets? Well, so
Tom Stanis
the it’s funny, the pattern repeats itself, you see this across all kinds of disease things, such as hypertension, same problem actually a little bit simpler, but the effects are more dramatic in terms of the under treatment of patients, and the sequelae of that. In cancer, we see this a lot actually in between when people are on chemotherapy, a lot of times they will go home from their therapy, they’ll have lots of side effects from that. It’s very famous, the kind of the side effects people have from chemotherapy. There’s actually lots of drugs to treat that. But they’re 50% of time, they’re not actually used according to guidelines, all over the place you see in medicine, opportunities to really improve this kind of therapeutic inertia. If we do it outside of a clinic visit and when you do it more connectedly. So this is really just the tip of the iceberg.
Nick van Terheyden
So one of the prevailing things that I hear repeatedly around AI is it’s not generalizable. I mean, that seems like Tokyo for AI. It sounds a little bit like some of this is generalizable, maybe there are modules or elements of this that you can apply in other specialties. Maybe validate is is that true? Are you approaching a more generalizable sort of technology?
Tom Stanis
Yeah, so we think of it as the story health platform. And that’s what we built from the beginning. Because we definitely saw that that all these components of making sure that patients are, you know, going through the treatment correctly, that they’re able to make it make progress, that the symptoms are being reasonably managed, that they’re able to get to the lab. And if they’re not able to get to the lab, why not? All those things are very generalizable and usable across these different conditions. So I feel like though, that you do need to build those disease specific modules to really understand the expertise, and not just say, Oh, just use it for whatever. Like, I do think it is important to have expertise in each of the different treatment areas also as protocol modules.
Nick van Terheyden
So looking back, what would you say were the sort of inflection points to, you know, real progress in this? And what are you excited about for the future?
Tom Stanis
Yeah, so one of the big inflection points was this, this registry trial done by some people at UCLA, Duke and a bunch of other centers that really highlighted the problem. That will happen about three years ago, and it was wonderful to see scientific community really focus on, hey, what’s the gap? And why is it happening. So that’s been wonderful to see the whole cardiology community be linking arms around this problem. So I think that was a big thing. The second thing really was the digital health revolution that’s happened in the last year or two COVID really have there’s one silver lining is that we’ve thought about new models of care, because we had to. And that’s been very exciting for all of us. And I think that’s going to be the next big that leads to the next milestone, which is, we’re gonna get to the point where the current fee for service world of is starting to crumble, frankly, a lot of the health systems are really struggling, they’re losing money. So we’re starting to see that be a big impetus to to say, we can’t just do things the way we used to, we actually have to work with more partners and do things differently. And I think it’s the big milestone that’s coming next.
Nick van Terheyden
Yeah, I think the I’ve said this, any number of times on this show, you know, that it’s almost an, it’s painful to say, the silver lining to the pandemic, but it really has been, it sort of opened people’s eyes to the potential allowing for an approach that I think historically was resisted for a variety of, you know, maybe good reasons, and in many cases, not good reasons, you know, somewhat dogma. But indeed, you know, changing the way that we approach this so that we start to deliver and, you know, in that instance, you now get to the inequity, you deliver better care and obviously, price points economics all dropped. So, I think exciting times, you know, delighted to hear about it, unfortunately, as we do each week, we’ve run out of time just remains for me to thank you, Tom, for joining me on the show. Thanks very much.
Tom Stanis
Thank you, Nick, for having me. It was wonderful. Discussion thanks
