The Robot Doctor with Robin Berzin (Parsley Health) and Aaron Patzer (Vital Software) | Disrupt SF (Day 3)
7:47PM Sep 10, 2018
primary care doctor
We're now going to have Sarah Buhr from TechCrunch, and Robin berzin from parsley health and Aaron Patzer from vital software to talk about what happens when an algorithm is control of whether or not you live or die. Please welcome on stage. Thank you.
Just to kick this off, I want to point out that Aaron actually started out in the financial space with mint.com, if those of you are familiar with mint.com. So I want to kick it off to you first, and asks, Why did you switch from the financial sector into the healthcare sector, what happened there
yet, so I actually have a lot of family in medicine. I have sister in law in emergency medicine, my brother in law's in emergency medicine. My sister is a professor of epidemiology. And in some sense, when I learned what they were doing and what kind of software they had to use, day to day, it was mind blowing. So I've done vital does software for emergency rooms. So I've spent quite a lot of time in the emergency room observing things. And for those of us in the valley, it's quite shocking. There is no search, you can't, you can search by date of birth, but you can't say, show me all of my patients last week, or show me all of my patients with high cholesterol, none of that exists. There's no Elastic Search, there's no loose,
you have to be your own advocate in fact, when you come into the doctor's office,
all of the software is desktop software. It's it's windows 98 software, literally windows 98 software. And the stat that blew my mind was doctor spends two hours in electronic health records for every hour that they spend with patients, plus another two hours at home catching up, if you could flip that equation and make the software better and safer and more usable. And they spent two hours with patients and an hour in the software, you sort of double the number of doctors that you have in the world. And that would be a very good thing.
So what is it that you do differently?
we'll leave aside the AI part until later in the discussion. But everything that you would imagine in good consumer software, which is search chat, push notifications in the emergency room today, many places, and these are reputable places like I shouldn't name names, they're using walkie talkies, they're using World War Two technology.
And it's it's just shocking. It's shocking how much slips between the cracks because of that there are I think 400,000 patient deaths a year be the third leading cause of death due to basically preventable medical error. And a decent portion of that is because frankly, the software sucks. We make it usable.
Okay. And, and Robin, you have a slightly different approach you you were in medicine. I mean, you are an internal medicine doctor. Yes. And you started parsley health to what to do what
we're rebuilding the primary health care system.
Okay, well, health care is broken on a number of levels. And I appreciate what Aaron's trying to do. My last company essentially tried to get people away from the walkie talkies into group chat and has gone on to do quite well in terms of helping doctors just communicate with each other in a modern consumer grade way. So that was a great experience. But I started parsley health because I recognize that primary care was completely broken today. 70% of disease is lifestyle driven, modifiable and often reversible.
90% of your health is socially determined, meaning determined by what you eat, your relationships, how you move, how you manage, stress, your environment, and yet both the software that we have to operate on in medicine with which I'm quite familiar with how bad it is, as well as the approach to medicine is broken. We ignore that 90% and instead, we focus on medical interventions, prescriptions, referrals, procedures, and that's not getting at the root cause of what is driving diabetes, heart disease, gastrointestinal problems, autoimmune disease, mental health and behavioral health hormone imbalances, even infertility.
And so parsley health combines and brings together nutrition prevention wellness, along with cutting edge diagnostic testing and technology into your routine, day to day primary care. We can't have wellness and primary care be separate. In a world where 70% of disease and 86% of healthcare costs are again, chronic and lifestyle driven.
So when Aaron says you go to your doctor, and they don't even know if you have high cholesterol, how do you you you keep records on how do you change that?
So we start initially just by asking the right questions. And we've had to build our own technology platform, because EMR is electronic medical records are claims billing software, that's what they do. They bill insurance for money, and then they have kind of patient care and tacked on to them, which is why they're really inefficient. It's why they're not asking the right questions. It's why doctors hate them. And they take up all this time, it's why patients can't really use them very easily, because they weren't built for patient care. Parsley is 100% direct to consumer, which means that we work for you, not your insurance company. So all the technology we've built built, and all of the data that we collect is focused on looking at you in a 360 view. So that means we ask the right questions, everything from how you were born, the medications that you've taken, what you're eating, how you manage stress, are you crushing wine, and Ambien, or are you meditating and doing yoga, those two things will lead to very different health outcomes,
we look at far more in depth biomarker data than a regular doctor. So we're looking at genomics, hormones, toxins, microbiome, all on a routine basis. And so and then, of course, we look at vital signs and basic biomarkers, things like cholesterol that a regular doctor would look at, although in a more in depth way. And we take this data on you that's wider and deeper data set than anyone in health care has. We also have a number of other types of engagement and behavioral data, and we create this view of you, that allows our doctors to interact with you in a completely different way. In addition to that, our first visit is 75 minutes. That's five x what somebody spends with their primary care doctor in a year, the average person spends 15 to 19 minutes with a PCP and a year, what are you going to accomplish in that time. And so we spend more time with you, we've resurrected the longitudinal relationship with your primary care doctor, which we know lowers costs and improve outcomes, but has like God away in primary care,
we take this more holistic approach, looking at 100% of your health instead of just 10%. And then we use the data infrastructure that we've built to help us understand you and work with you over time so that we can help you thrive as opposed to just managing your disease.
Yeah, and Aaron, I want to get to know what you think of all of that. But first, I want to address that this is not without this is kind of costly for someone. If you don't take insurance, you have to Bill your insurance later.
Correct. We don't accept insurance directly. And it's not because we don't want to take insurance. It's because the insurance world hasn't caught up with us yet, parsley cost a little less than $5 a day for an annual membership, there's no copays that includes all your doctor's visits almost four hours a year with your physician, all of your visits with a certified health coaches on staff to help you make these behavioral changes and lifestyle changes that your doctors recommended unlimited messaging, access to online to us online. So the other day might be patient wrote to me, and she was having a muscle spasm. And we wrote back and forth together on our messaging platform, a little bits are easy, simple chat, and I took care of I still see patients one day a week as part of the company, I took care of her problem remotely. And she didn't even need a visit, that level of service would normally cost 20 grand a year or more with a concierge medical service. And we're less than $5 a day. And so while Yes, it is over and above your regular insurance payment, or what you would think of as primary care, the level of service and the outcomes that you're seeing are so vastly far in a way above what you could ever get from a typical primary counted caring counter. We've saved individual members hundreds of thousands of dollars simply by reversing their gi autoimmune cardio metabolic and mental health disorders and helping them truly transform their lives.
Aaron, what do you think I want to get your thoughts on on
Robins approach, we have very different angles with which we're about approaching healthcare. On the one side, we make software for hospitals as opposed to consumer so we're attacking different ends of the market. But more than that, I feel parsley and I'm an investor in a competitor to be for full disclosure. And my problem with all of those is they are health care for wealthy people who are 20 to 45 and are already in good health. And frankly, it's not going to move the needle in my opinion, emergency software, for example, is the backstop for people who either don't have insurance, they take everyone of all economic strata.
And I think that
I don't think that concierge health is is going to affect the bigger picture to be skeptical. But actually, I'd like to talk about artificial intelligence and healthcare since that is what what I would
argue with that, you know, you can't rebuild a healthcare system by sitting there within the existing existing system and waiting desperately for it to change. I can't speak to competitors, but our members are not healthy people getting ever so slightly healthier. Our members are people with chronic diseases that have been failed by the existing healthcare system, and we are massively moving the needle saving again, hundreds of thousands of dollars. In one member alone. I had a patient who was in yes in her 20s, but she had a severe autoimmune disease vasculitis, and had been fired by the top rheumatologist and your New York City, because none of his drugs were working through end up testing invite, and identifying inflammatory triggers, we identified food sensitivities, alterations in the microbiome through approach to diet and short term supplements, fixed all of that her disease, which had given her authors in her mouth all the way down her throat and was costing millions of dollars to the healthcare system over her lifetime
can been a now in remission for three years. And so I can't speak for other services. But at parsley, the
it's not about
the existing medical system with kind of a fancy flat screen or slightly better, more consumer grade technology. It's about fundamentally changing your approach to medicine. And so while we're starting at a price point, that's a sustainable business model. And that can allow us to scale because there's a very large Tam for it, especially given you can use FSA and HSA submit to for out of network reimbursement. Our age range right now is 18 to 83 and is not 20 year olds and early 30s who are healthy and just getting extra healthy, it's people who have chronic conditions like 67% of 18 to 45 year olds in this country, according to the NIH. Aaron,
how do you use AI to take the same approach only she's saying that she, she's able to help people who have these chronic conditions, no one's been able to help them, how does your program help in that
same so a doctor, you have plenty of credibility. But yeah, definitely, let's let's talk tech. Because we're at TechCrunch. And I love the technology of it, what we're able to do is to make machine learning predictions on free form doctor's notes, which today are just sitting in a PDF somewhere, or based on what the patient comes in and actually says to a nurse or fills out in a little mobile website, or chatbot,
and make a prediction with that vital signs, historic information about whether you're going to be admitted, where they're going to be discharged, where they're going to need a CT scan, whether you're going to need an X ray, what medications you'll be on whether you're going to wait two hours, or whether you get to wait 10 hours, which depends on the patient. And the condition. What that means is that it has anybody been to the emergency room lately, it's 140 million visits. And it's a terrible experience. You sit there for three or four hours, and nobody talks to
you. It's the emergency room. It's like any, any kind of hospital visit that you're taking, I just had a baby,
so I know. Yeah, so there's, there's just no visibility. So simply giving people an expectation of what's coming next, puts them at ease,
makes them less panicky, and then
patients seriously when they say that are happening, going through.
So like on a call to avoid right now. Yeah,
and so it definitely helps with patient safety, you know, just giving them just a mobile website that they can say, I need help right now I'm sitting in the waiting room have been here for four hours. Well, Robbie, my
art argue that, you know, she's seeing patients more longer, more frequently more available. So she might be solving problems before they even get there. It's you
preventative medicine as opposed to, you know, when the when you're in the emergency room, it's maybe it's too late. But most ER visits
in this country are actually for chronic preventable conditions. Because of the way our healthcare system is structured. I'm having worked in an ER myself and training unfortunate lately, we've short changed primary care so much people don't have access. So they're using the ER, often as their primary care doctors, we all have this image on TV of people getting like wheeled into the ER, you know, having a stroke, and they're completely, you know, unconscious, and it's an absolute immediate emergency. Unfortunately, that's what an ER should be used for. The vast, vast majority of ER visits are actually for chronic lifestyle driven diseases of our time. But some might
argue that what your approach does is it is for the wealthy people, it's people who don't need to have insurance who can pay you. So only wealthy people are being able to access something that you're offering. Whereas, you know, it's poor people that by and large, maybe have to go to the emergency room, and they don't have that remembers cab
insurance. And they're not all wealthy. Again, there are people with chronic issues, who are looking for a better solution for their health. And the exciting thing for us is because we have built this data architecture and are building applications on it. And we're building a brain around the type of health that we're providing, which is completely unique in the market. Over time, we will continue to work with the large insurance companies and self insured employers and even smaller employers, which is already happening. It's not that you start with this current price point, and you stay there forever. But you start there as a starting point so that you can build something new and awesome and amazing, and then expand it to everyone else. And that's a pretty clear
roadmap for us anymore to say. So yeah, I do.
I'm just curious as to sort of what the tech stack looks like, and how tech enables it, because I'm an engineer, and I just love that stuff. Yeah, well, thanks for
being the moderator. We have an incredible technology stack, and an amazing engineering team. And best in class design team we've built in patient portal and an ability to track your outcomes over time, we've built the first ever symptom tracker that's relevant to an under 65 population, which companies like Google are interested in using, we are building end to end technology that not only creates incredible efficiencies, and patient care and adult experience, actually, Aaron,
I do want to ask you on on your approach towards medicine, I do want to know, how are you solving. So I've heard all about like this AI, this idea that you are going to implement this technology within hospitals with emergency rooms, but I hear that pitch. So many times, it's a one off one off one of to each hospital, you have to sell the product to each one. And it's not really actually solving the problem. So how do you
take the approach? Yeah, so most startups I find in health care, at least
who are going after hospitals are not attacking. What I would say is the core problem, which is the electronic health records yourself, you're absolutely right, that they are pieces of billing. So there's an eye chart, there's epic,
that's the biggest one
that you've learned or an epic are the two large eyes are about 10,020 thousand employees, about 10 billion in revenue each. But there's actually five or six other players who are making, you know, sort of billion yours better. Like, what,
what's what about it is better?
Hey, you know, it actually reminds me a lot of the sort of mint versus Quicken. Like one of them is simply usable, beautiful, requires no training we actually do we put nurses on the software and give them zero instructions and say, we usually give them a goofy scenario. Like, you know, Bugs Bunny has been shot by Elmer funded, go triage them, and see if you can do it. Or we'll do that with just normal people, and see if they can use the software. So it's like consumer software. But for hospitals, so you eliminate a three month training period, it's in the cloud, it's got all the features, you'd think it doesn't need it to be
more beautiful, it seems like it needs to be more functional.
Yeah, in terms it is more functional in the sense that it includes all of the predictions about what's happening next. And what that means is, if you know that somebody needs a procedure three hours before you otherwise would have made that decision, because a machine told yourself, then you get them out of the emergency room three hours faster. that's a that's a time and money savings. But more importantly, it gets them back home or into a proper hospital bed if they need it.
What's the feedback you've had from doctors and nurses? Because I, I feel like you know, I was someone that didn't really even have to the doctor that often I had never been in a hospital before it had a baby. And then when I when I was in there, I noticed that people weren't looking at the chart. They weren't communicating with each other. The handoff from shift to shift it was I had to be my own advocate. So how do you solve that problem? Yeah, so let me tell
you a little bit more about the problem. So I was in the emergency room, and I was working with a technician who is operating the CAT scan, and his screen was lit up read with all these people waiting. I said, should you even be talking to me right now, it looks like there's a big backlog. And he said, Well, I couldn't put an IV and somebody so and I couldn't find a doctor on the walkie talkie. They weren't answering their walkie talkie. So a little icon that was about eight pixels appeared in a note somewhere I took somebody an hour and a half to find that because there are no push notification so it's I know it maybe it sounds crazy other than the artificial intelligence in some way we don't have any additional functionality we just do what health records were supposed to do well that's it honestly that's the secret to good software just do it well few clicks, good user experience and add a couple features that
you know improve the workflow and improve the amount of time that people spend in the emergency
room Robin you're a doctor he's what he's talking about what that have helped you when you were in internal medicine it within a hospital or another service
you know, in without all the last company did did much of what he's talking about in for banal for ears for the general gentleman floors. Yeah, I mean, improving the UI and bringing
medical technology up to consumer grade is definitely helpful and important. Again, I think the bigger issues of the emergency rooms tend to be symbol overload understaffing so no matter how many push notifications you get, it doesn't necessarily get a doctor there faster. And so I think there's a number of complicated issues around
in healthcare. You know, healthcare is a people business powered by technology. And a lot of people want it to be a technology business powered by people. And I just think that technology is incredibly important to the future of healthcare. But understanding how those two things interplay is really important. And I think AI is super interesting for healthcare, you know, with AI. However, ultimately, it's only as good as the data sets that are input into it, right. And in healthcare, we have a really big problem with the data sets, because most of the outcomes data we have in healthcare is claims data. So that's really a measure of utilization. It's not a measure of like, how healthy people are. And so what's really exciting about personally, and what's part of the really interesting value of the company is that, as I mentioned, we're collecting a data set that and asking questions and looking at biomarkers that I know that no other service in our category or other primary care services are looking at. And I'm really excited about that, because we're also trending those data sets over time. And that will allow us to build some of that predictive technology that's really, really helpful. I'm not, again, focused on the emergency space. But in the primary care space, you know, 99% of your health is not happening in the doctor's office, certainly not in the hospital, it's where you're living your life, I think there's a really big promise in that.
And it sounds like there, you mentioned, the getting the right data. That's one of the big holes that we're missing right now, where are some other things that we're missing that would make it much more succinct
so that that data access actually is a tremendous problem. So it took us literally one year to get effectively proprietary information for a major city from a major research university hospital, I think I had, you know, meetings with 19 C suite executives, and to get that data into clean it up, it's incredibly rare. So I often say that in healthcare, you know, I had an operation inside my inside my nose five years ago, and I have zero records for that I went to Jiffy Lube in Florida. And they knew that I had gotten an oil change in California, you know, three months prior. Jiffy Lube has better records on my oil changes than we have in terms of connectivity between hospitals. And so that data is sitting
siloed. And there's not been a lot of incentives towards interoperability, although CMS and health and human services have been working on that for a while, at least in the last administration, and continue to do so
one of the big principles of parsley is that you as the patient Own your data, and so we make everything available to you online, you can apply that you can download it. And we in addition to giving an incredible patient care and helping people transform their health for the rest of their lives, we're teaching them to own all of that data to store it to be able to use it to that they're empowered to be the CEOs of their own health, it's never going to work to have no matter how good the systems are, even if we get to greater standards of interoperability. So we're going to work to have this data live in all these different systems, you should be ultimately that repository of your health data. And I just hope that everything that's being built here at TechCrunch in the health space, keep that in mind. You guys have
been fabulous on stage. In fact, this has been such a great discussion and in 10 minutes are going to have a QA across the across the hall. I think that way I felt this way follow them. If so, if you want follow up questions as to what they're building please go to that. q amp. A. Thank you so much for being on stage with us today. Thank you.