Uli Chettipally: 00:00 If you can imagine fee for service system is like physician being on a treadmill and somebody else is controlling the buttons on how fast the treadmill has to run.
Intro: 00:13 From Spaugh Dameron Tenny, it's the Prosperous Doc podcast. Real stories, real inspiration, real growth, a show for doctors who are ready to improve their overall wellness in every aspect of life. Now here's your host Shane Tenny.
Shane Tenny: 00:31 All right, welcome back to the Prosperous Doc podcast. Today we're talking healthcare and artificial intelligence. 50% of what happens in a typical US clinical practice is unnecessary, ineffective or dangerous. It's not my opinion. In fact, it's the opinion of my guest today, Dr. Uli Chettipally. He has worked for over 25 years in emergency medicine with Kaiser Permanente in California and is really a pioneer on this topic of artificial intelligence and healthcare. He's the founder of the Society of Physician Entrepreneurs in San Francisco, the president of InnovatorMD and overall, I think you'll hear, as we talk about his story is just passionate about using data and AI to support value based model for healthcare. One where healthcare systems better predict who's going to get sick, prevent them from getting sick and provide better treatments. In fact, last year in the midst of all of the different things that he juggles, he also found time to publish his first book, Punish the Machine: The Promise of Artificial Intelligence and Healthcare. Dr. Uli Chettipally, thanks for being with us today.
Uli Chettipally: 01:46 Thank you Shane, for having me on your podcast. It's a real pleasure to be chatting with you today.
Shane Tenny: 01:51 Well excellent. Well, maybe we'll start with the opening statement that I gave or the opening statement, I guess, that you made based on your experience in medicine. You've said that you're convinced half of what goes on in a medical practice is either a waste of time or ineffective or dangerous. What do you mean?
Uli Chettipally: 02:10 Well, let me start with a story. I had this statement on a slide deck. It's part of several slides that I wanted to present to my CEO at my last company and I wanted to present this saying that we need this project that I was proposing, which is a research project to figure out which things are essential, which things are ineffective and which things are dangerous in our practice using technology. This slide was in the middle of the presentation and so I was going through 10, 15 slides. And when I saw the slide, I was a little worried, what will my CEO think looking at this? 50% of what happens is useless? Are you you crazy? Or something like that. And so what happened was I went through the slide and went to the next slide and he said, "Stop, can you go back to the previous slide?"
Uli Chettipally: 03:15 I was really scared he's going to fire me or something. And he took out his cellphone, took a picture of that slide and said, "Continue." He didn't say anything about it. But he just took a picture of that slide and put it in his files, I guess. But yeah, that is what I think is happening in the current healthcare system. And the reason for that is probably the business model of healthcare. The business model is designed for physicians or providers, hospitals, labs, radiology, whatever services, to do more for patients, to prescribe more medicines, to do more surgeries, to put people in the hospital, more beds. It is not designed to look at the outcomes of all of these actions.
Uli Chettipally: 04:13 Now, for example, if a physician is practicing in a fee for service system, they are paid based on the number of visits that the patients make. Now, the question is, does that visit actually make the patient healthier? Is that surgery that the patient has, what are the outcomes of that? If you're prescribing a medicine, does it really work? Or is there a better option? And so those are the things that we don't normally look at. And that's why, by the way, that number is not from my own mind. It is based on research that the current system does a lot of things that are not necessary and sometimes dangerous.
Shane Tenny: 04:56 And you've spent more than two decades working in emergency medicine with Kaiser, which is, I believe described as a value based healthcare system, as opposed to fee for service like you were just touching on. Explain a little bit of the difference between kind of the traditional mass market fee for service medicine and a value based model and what the difference makes.
Uli Chettipally: 05:17 Sure. A simple way to describe value based care is the business model is you make money when the patient is healthy. In a fee for service system, you make money when the patient is sick. And so they're exactly the opposite of each other. And so let's say in value based care, you make money when the patient is healthy, that means that you will figure out ways to keep the patient healthy. You'll figure out ways to keep medications, not prescribing unnecessary medication, not doing unnecessary surgery, not providing unnecessary hospitalization. You want to find out about problems that might happen in the future and kind of predict and so change the focus to preventative medicine.
Uli Chettipally: 06:11 For example, screening for cancers, immunizations, controlling diabetes, high blood pressure, high cholesterol, so that the patient will not have a heart attack in the future or a stroke in the future. Those are the kinds of things I'm talking about when I talk about value based care is taking the full risk for the patient so that you get a set amount of money to be able to take care of the patient, whatever may happen to that patient. It is in the provider's best interest to figure out how to keep that patient healthy and avoid the things that will make them sick.
Shane Tenny: 06:54 Elaborate a little bit, for the vast majority of physicians in our country, certainly the majority that we come in contact with are on a fee for service model, RVU based compensation, production base. You run the test, you do the surgery, you do the procedure, there's a revenue value assigned to that. Can you maybe go one layer deeper and the concept makes sense, getting paid to keep them healthy. How do you actually measure that? Because of course, when people are healthy, they're not coming in to see you.
Uli Chettipally: 07:21 Yes. That brings us to the next level, which is what are the things that keep patients healthy? Is a visit to the physician a good way to keep people healthy? Or can you do a phone call? Can you do a video visit? Those things were highlighted during this COVID season. What happened was that a lot of these clinics, they could not see patients in the clinic and they could not do elective procedures, elective surgeries on these patients. And so a lot of practices lost big time money because they were not able to do it. Whereas physicians in value based care, they did fine because they were getting paid anyway. They were able to figure out the same thing. How do you keep a person healthy? Do they necessarily have to come to the hospital? Can you do that on video? Can you do that on telephone? Can you send somebody to their home?
Uli Chettipally: 08:23 Although it's not reimbursed in a fee for service system, in a value based care, primary care, it's probably better to send somebody to their home to check their refrigerator, make sure they have food, to make sure that they have transportation. They have these other things which will affect their health outcomes. It's a totally different way of looking at it.
Shane Tenny: 08:45 And so within a system like Kaiser's is the hospital system that a patient would be directed toward or a doctor's office, so that sort thing, are you receiving almost a membership fee for this individual? And in some ways we get to keep the fee whether they come in or not.
Uli Chettipally: 09:04 Exactly.
Shane Tenny: 09:04 And so if they come in and we have to run x-rays well, then that's a cost to us. And if we come in and those sorts of things, it allows there to be just a different upside down way in some respects of cost control.
Uli Chettipally: 09:15 Exactly. That's exactly what it means. You get a fixed amount per member per month and you manage all the things that are needed for that patient. And so when you order an x-ray or a CT, you have to critically think, how is this test going to help me keep the patient healthy? And so you cannot order unnecessary things. Also, your thinking will be more longterm. Okay, if I don't order this test today, how will it affect the long term health of this patient? If I don't control this diabetes or high blood pressure, what will happen in the long term? Is the cost going to be more in the long term if I avoid this expense now? Or is it better to put more effort now, more expense now so that the long term the patient will be healthier?
Uli Chettipally: 10:13 And so it's a totally different mindset and which actually helps patients also understand why they need certain things and why they don't need certain things. In fee for service, you don't necessarily have to know the patient's outcomes one year from now, two years from now, 10 years from now. Whereas in value based care, you definitely need to know that so that you do the right thing for the patient.
Shane Tenny: 10:40 When you articulate the difference in philosophy and mechanics, it seems so logical. Is value based medicine spreading? And if not, what are the headwinds either just within the system or within regulation or within Medicare or whatever?
Uli Chettipally: 10:56 It is spreading, but much slower than what one would expect. One of the big problems is that fee for service generates more revenue, creates more expense and so there are more jobs, more profits, bigger organizations, bigger systems. And what it does is it kind of creates this too big of an industry to fail. And so to justify its existence, they have to perpetuate the current system. And also there's the question of, oh, if we decrease, if we become more efficient, that means we'll have less jobs. That means people will lose jobs. If you do less tests, that means that there'll be less investment in the technology or in the operating rooms or in the people. And so those are some very difficult questions. Is the system existing to benefit the system? Or is the system existing to benefit the patient? And that becomes a bigger philosophical question.
Shane Tenny: 12:03 And where in your journey, your career, your experience as a practitioner, did your interest in data and artificial intelligence, where does that enter the picture here?
Uli Chettipally: 12:13 Well, before I started my job here at Kaiser Permanente, by the way, I retired from Kaiser Permanente last year. Before I entered the job and I was in my research training and my professor said, "Kaiser Permanente is the future that you should go join there." I was already doing research in hypertension and so he kind of understood my skills and preferences and I followed his advice. But once I came to Kaiser, it really opened up my eyes. Kaiser was one of the first big companies that actually adopted electronic health records. They spent big money. Epic was the EHR company that they chose to go with. And the implemented it and I think by 2005 or so, they became paperless. They implemented it in the outpatient clinic, in the inpatient, on the finance side. By the way, for people who don't know what Kaiser Permanente is, it's actually three different companies working together.
Uli Chettipally: 13:17 There's the hospital company, there is the physician group and the insurance company. They are all working together to form this integrated system. And so all these records from the insurance side, going to the hospital side, going to the physician side, they're all contained in one place. And imagine the power that can deliver, because it becomes really efficient and all the services are under one roof, the pharmacy, the radiology and the physician offices. And so it makes it much more integrated. And around 2005 or so we became paperless. And then I thought, wow, we are collecting tons of data. How is this data going to affect the next few years that we continue to practice? And that's when I started myself and another physician, we started our research group to tap into this data and figure out how can we make the care better for the patient?
Uli Chettipally: 14:20 And that was the big project called CREST Network. We started that and currently there is about 14 physician scientists, lot of research going on. I built a technology platform to be able to do just that, pull data from the electronic health record, analyze it and give physicians the decision support based on the data from similar patients that were seen before so that, that knowledge that accumulates over time can be translated into a software tool, which can provide that support to the physicians.
Uli Chettipally: 15:00 I'll give you an example. Let's say a patient walks into the emergency and they have chest pain. And so this person coming into the emergency, first thing we worry about is, is this a heart attack? Or is he going to have a heart attack? Is this just that initial symptoms? And so when the patient comes in, a text message goes to the physician saying that there is this person, Mr. A, who is in room 23, who is a potential candidate for this technology to be able to analyze their data and give a suggestion to the physician. And when the physician clicks a button inside the Epic EHR, our platform opens up, it run them through a couple of schemes and gives them an answer. And the answer is the chances of this person having a heart attack in the next seven days is .003% and so you don't need to do any more testing. And then the physician is able to choose whether to follow that instruction suggestion or not.
Uli Chettipally: 16:08 But what we have found is that over time, physicians started understanding, wow, this is so powerful. If it can help me make a decision whether to keep the patient in the hospital or let them go home. And so that way you're not missing any bad things, but you're also not over treating a patient or over testing a patient. And so the length of stay in the emergency department decreased, the hospital admissions decreased, the post discharge testing decreased. And then we caught a few high risk patients and actually did the right thing for them. And so that's how technology can help with the care of the patient and to get to better outcomes for the patient.
Shane Tenny: 16:55 And so the platform that you built, CREST Network, which Kaiser has adopted is not just taking the raw data that's put into EHR from their visits, but it's then overlapping that with essentially, I guess, symptom analysis and some sort of conclusion that helps guide the physician, the attending doc, who's on call or right there, doesn't have the luxury of having seen the chart from the last five years of visits, but the system can tell them based on their history, everything we have in here, here's the risks, here's the possibilities to get better more appropriate treatment, whether it's more treatment or less treatment is really the outcome.
Uli Chettipally: 17:33 Exactly, exactly. And so it's not only just the patient's data. We're also looking at 100,000 other patients that are similar to this patient and so their data. And so it becomes even more powerful as the machines become more and more powerful, they can analyze bigger and bigger sets of data. And so that's how we drive these decisions. Physicians feel great because now we have, it's like building evidence as we go along and it's great for a system like Kaiser.
Shane Tenny: 18:10 Sounds tremendous. And I'm thinking for those of us who are football fans, we now get to watch football and see in real time the statistical likelihood of making a field goal from 57 yards, that is the artificial intelligence that in some ways you're talking about here.
Uli Chettipally: 18:24 Exactly.
Shane Tenny: 18:25 It's real thinking and it gives us the probability of, is this an MI or is it just indigestion? Based on history build et cetera, of similar type patients. This CREST Network, is it expanding beyond Kaiser? Are there similar type or competing type algorithms being employed, Epic and other systems or what's the future? I guess is what I'm wondering.
Uli Chettipally: 18:46 Yeah. The future is that as we get better and better technologies, the artificial intelligence and the machine learning and as we get access to larger and larger data sets, so these systems will become even more intelligent. Right now, the CREST Network is basically within Kaiser, but it does partner with other organizations, especially academic organizations that really love this kind of stuff, but they don't have access to the data, access to the quality data that we have come to use. And also the kind of platform that this helps and enable that kind of care to be delivered. This is probably the largest such system in the industry right now, but I'm sure there's nothing fancy about it. It's just somebody has to take it and do it. The technology itself is pretty basic. It's nothing fancy. You don't need a lot of money or effort to do this.
Shane Tenny: 19:49 And so are there other systems around the country that are looking to Kaiser or asking or exploring similar type programs for themselves?
Uli Chettipally: 20:01 Yeah. In general, Kaiser is at the forefront of this movement in value based care. By the way, they have been in business for more than 70 years with the same business model and only recently they have been getting the recognition that they deserve. And so other systems are trying to figure out how to do this. There is the ACOs, accountable care organizations where hospitals or doctors groups are partnering with insurance companies and trying to figure this out. One of the biggest challenges is that culture shift, that mind shift that needs to happen is that every time a physician sees a patient. That's why it is so hard to replicate because either you have to start from scratch or you have to do a total 180 degree turn in what you've been doing so far in a fee for service system. And that's why it is challenging for others to figure out.
Uli Chettipally: 21:00 But there are some startups and newer companies that are trying to do that in primary care, especially like ChenMed And they're trying to address the high utilizers within Medicare and Medi-Cal. They've been very successful, I would say. And there are new companies that are coming up like that, where they're changing the paradigm and starting from scratch with this newer business model.
Shane Tenny: 21:27 Well, it's fascinating. We're going to take a quick break and when we come back, I want to ask you about some of the myths that I think folks have when they hear artificial intelligence. We'll be right back.
Shane Tenny: 21:43 When it comes to retirement planning, we've been told by magazines, websites and podcasts that the most important ingredient to solve is knowing our number. That is the total amount of savings and investments that will be required to maintain your standard of living once you stop working full time. In reality though, there are actually three important I's when it comes to retirement planning. The first is certainly knowing how to replace your income once you retire. But the other two I's are just as important. Where will you find your identity? And where will you find your influence? These are the two aspects that make a working career so fulfilling, identity and influence. The intellectual stimulation, the relational connections, the sense of being needed and making a difference every day. The opportunity to talk about projects and progress when you're with friends or neighbors or at a party. Identity and influence are the things that make life fulfilling.
Shane Tenny: 22:53 As you plan for retirement, make sure to identify ways to replace all of the components that have been meaningful in your work, income, identity and influence. These ingredients together will help ensure that you have the prosperous retirement that you've longed for. This financial wellness tip is brought to you by the great financial advisors at Spaugh Dameron Tenny, helping doctors and dentists for over 60 years, make smart financial decisions. You can learn more at sdtplanning.com. To learn more about retirement planning, you can access the free retirement readiness quiz brought to you by Spaugh Dameron Tenny. The link is in the show notes below.
Shane Tenny: 23:44 All right, Dr. Chettipally, we're talking about your background in medicine. We're talking about the innovation you've brought to data and to artificial intelligence. You wrote a book this year called, Punish the Machine. Tell us a little bit about the concept, what led you to write the book.
Uli Chettipally: 23:59 Sure. When I was doing this research within Kaiser and I was looking at the data that is coming out and the results, it was so amazing that we can implement something like this. And then I was thinking, what if we had really high power algorithms, like how Google or Netflix or Apple have and implement that in healthcare? And that's when I started thinking about it and researching more. And that's when I wrote the book. The book is designed mostly for common people to be able to understand what it is, why it is important and why it is the future. The big trends that are happening right now are that healthcare costs are going up. Skyrocketing. Physicians are burning out. They are. If you can imagine fee for service system is like physician being on a treadmill and somebody else is controlling the buttons on how fast the treadmill has to run.
Uli Chettipally: 25:00 At some point you see 20 patients a day, okay, you get paid a certain amount. What if they decrease the rate per patient so that you'll get less money for seeing the same number of patients? Do you have to see more patients? That means the treadmill gets faster and faster and faster because somebody else is controlling the buttons. That's why physicians are burning out. The second big reason is that they feel that whatever they are doing, it's not making a big dent. And when they came into the practice of medicine, they had very high goals, which is okay, I'm going to make a difference. I'm going to help these people. I'm going to stamp out disease, but that's not happening.
Uli Chettipally: 25:44 They are just doing these mechanical things to get paid and they're not seeing the outcomes. The third reason is they're also doing some actions, some activities which are not directly connected to patient outcomes, which is entering data into the electronic health record. And physicians hate it, but they have to do it because otherwise they won't get paid in a fee for service system. Those needs keep growing, the more and more and more data that you're entering. That's how physicians are suffering.
Uli Chettipally: 26:20 Let's look at the patient side. Patient outcomes are not any better. In fact, some of the outcomes have gotten worse. As you might have seen, some of the people are living shorter. Certain sections of the population. Hypertension is under less control right now than it was say 10 years ago or five years ago. The outcomes are going bad and so we need a solution. We need something fundamentally different. You might think that, oh, we have this technology, we can find out outcomes, how it is affecting. But the current system, current fee for service system, there is no incentive to get good outcomes. There is no incentive to even look at outcomes. If you have five different drugs you can give, why should you give one drug over the other? Or there are five different tests or are there are two different operations which operation will you do? Well, you'll pick the one that probably will get you more reimbursement or a device or things. The selection of treatments, the selection of tests, all those are dependent on the reimbursement.
Shane Tenny: 27:37 I was going to say, you can kind of see the clash in the culture as insurance carriers or systems try to implement results based measurements or patient satisfaction scores. And then just the conflict that creates within the providers in that traditional system. If the data, if more systems, more insurance carriers explore pulling the data together to impact outcomes, I think sometimes there's just this innate fear that, oh, the more we adopt machines to do the thinking for us the less need there is for the providers. Is that connection legitimate? Is there a risk to physician jobs or physicians' incomes to implement the sort of analysis that you've created?
Uli Chettipally: 28:21 If you look at the physician, I think the physician's job is probably the last person to be affected by technology. I think most of the jobs are either the people who are sitting at desks in insurance companies, or much lower level job. Physician's job is very highly skilled and we will always need physicians. Even if you have the best technology, you will always need physicians. It's like you have GPS, but you still drive the car. Even if you have a self driving car, you still sit at the wheel. And so you need somebody, a human being to sit at the wheel. Does that mean that we should not use GPS? No. Because when you use GPS, you become more efficient and the outcomes are better. You reach your destination faster. That's the analogy I take.
Uli Chettipally: 29:14 The reason for knowing the outcomes is because you can prevent bad outcomes. You can decrease the cost of care, overall cost of care and you can increase the health of patients. And that was the main impetus for me to write the book is that the current system is punishing the doctors and not helping the patients at all. And we are trying to protect the technology. We are using the eighties and nineties technology, the EHRs. Why are we protecting the technology? Let's make the technology work harder. Let's punish the machine so that the doctors can be helped. Don't punish the doctors, punish the machine so that the patients can be safe and better health outcomes can be achieved through this technology.
Shane Tenny: 30:07 When you talk with docs or administrators or executives within other systems, there's always the headwind of change. And the cost of change. Is the cost of change that you propose, is it really a financial one or is the cost more cultural, effort, learning, knowledge, resistance? How do you balance those things? What's the argument in favor?
Uli Chettipally: 30:35 I would say physicians are not averse to change. The only reason, I think the main driver is the threat to income and threat to their way of life. If you tell a physician, "Hey, you can see 25% less patients today and I'll pay you the same." They'll be happy. Anybody will be happy. Oh, all you have to do is, you have to look at this data. That's all you have to do. Actually COVID was a great example where suddenly the visits to the clinics dropped, which means if you're a fee for service provider, your income dropped. And the reason why they adopted telemedicine was because now telemedicine will pay the same as a visit. Telemedicine has been there for a long time, but it was not adopted. It was not adopted not because physicians are averse to change. It was not adopted because the financial incentives were not aligned. How quickly we went from zero to 80%.
Shane Tenny: 31:46 Is to align the financial systems, I think your point is great. I'm smiling because I often will tell folks I say, "If you want to see the impetus, follow the money."
Uli Chettipally: 31:57 Exactly.
Shane Tenny: 31:58 To align the financial incentives, is that the role of the healthcare system to initiate that? Is that the role of the insurance providers? Blue Cross, United Healthcare, the Cignas? Where does that start? Who needs to play in that sandbox to get that done?
Uli Chettipally: 32:12 It's the role of the payers. Whoever is paying, they have to see what they're getting for their money. And is there a better way of doing it? Some payers took it on themselves. For example, Amazon and those guys, they wanted to create their own system. There's some struggles there. Medicare wants to do it, but then there's always this political pushback. And so the companies that can do that are the payers. There is a payer system that wants to give their clients the best healthcare outcomes. Those are the ones that have to push that change otherwise it will not happen. Follow the money. Where does the money start from? The source.
Shane Tenny: 33:00 Yeah. Well, it's a fascinating work that you've been at and it'll be fascinating to see what the next decade holds. Before we wrap up, I want to ask you about your latest venture with InnovatorMD. Can you talk a little bit about that?
Uli Chettipally: 33:17 Sure. After I finished my tenure at Kaiser Permanente last year, I started this company. What I've noticed is that a lot of the innovation that actually works is coming from physicians and the people on the front lines. They are actually seeing the effect of the current system, what is not working. And also they also see the potential of what could work. And so I thought, nobody's talking about what physicians are doing on the innovation side. They're always, everybody's complaining that physicians don't want to change. They are this group of people that hate change, that are old fashioned, that are against new innovations, but that's not true. And so I wanted to highlight that. I wanted to help physicians elevate their and also their projects, the work that they're doing. And that's why I started InnovatorMD. And our total focus is on physicians who are doing innovation.
Uli Chettipally: 34:20 And so we do monthly meet ups every third Thursday and everybody's welcome to join. You don't have to be a physician. We do courses for physicians on teaching some of the stuff about money, about devices, about business processes, about investing and all, how do you raise the money for your startup? The third thing we do is we do these big conferences where we bring in a big group of physicians who can talk about what they're doing in their work, so that that'll inspire others to pursue some of those ideas. And so we hold a conference. It's called InnovatorMD Global Summit and the next one is in January so I would encourage people to come join that. They can go to the website, it's called innovatormd.com and you can see all these events and sign up.
Shane Tenny: 35:11 And is the conference, I wanted you to highlight that because I know it's coming up in a couple months, by the way, we'll put a link to it in the show notes. If you are listening to this and can't jot it down or driving in the car or something. The conference in January is going to be virtual or going to be in person?
Uli Chettipally: 35:26 It's going to be virtual. We do have a space booked, but it's definitely going to be virtual, I think. Unless things open up and it becomes COVID safe, but we can't tell right now.
Shane Tenny: 35:43 But a great opportunity. I know just from the conversations we've had for those entrepreneurial physicians, those of you with ideas always floating through your head, the side gig work, this is the chance to rub shoulders with others who are wired similarly and get ideas and brainstorm. And so, as I said, we'll link to that in the show notes.
Uli Chettipally: 35:59 Great.
Shane Tenny: 35:59 Dr. Chettipally, thanks so much for being with us today. Thank you for listening, for joining us on today's Prosperous Doc podcast. As always, we've got more episodes queued up, ready to roll out every Monday for you. Please subscribe. If you do, you'll get for the first heads up to our new episodes. Of course you can find us on most social media channels and welcome your comments, reviews and feedback either through iTunes or Google Play or you can always email me directly, email@example.com. Thanks. We'll see you back here next time.
Outro: 36:34 This episode of the Prosperous Doc podcast is over, but you're not alone on your journey. Spaugh Dameron Tenny has been helping physicians and dentists prosper through financial planning for over 60 years. To connect with us, visit sdtplanning.com today and take your financial wellness to new levels. Join us on the next episode of the Prosperous Doc podcast.