Why Healthcare Innovation Still Struggles: Lessons in Transformation with James Whitfill, MD, MBA, Chief Transformation Officer at Honor Health

<p>hello everyone this is Cole from the American Journal of healthc care strategy joined by a very special guest Dr James whitfill Dr whitfill please introduce yourself well Cole thank you it's great to be here um as you mentioned my name is Jim Jim whitfill I'm a a physician by training I'm currently the chief transformation officer at a health system down in Phoenix called honor Health that transformation officer role is kind of a unique one and I've got really three big areas that I work on there number one I lead digital strategy for the organization digital strategies broad for us includes everything from our marketing arm uh our it arms informatics AI analytics project management process Improvement sort broad range of things I also uh work with our Innovation functions so we're fortunate enough to have everything from some corporate Venture Capital work to a partnership with CDW where we have a physical lab where we test out new technologies we also look for new Innovative workflows and do some work with General Catalyst in terms of investing in and innovating New pieces and then lastly I lead a lot of our early stage strategic Partnerships we're we're a smaller organization about 15,000 employees so for us Partnerships are how we are able to scale to be able to deliver better health care for our [Music] community really appreciate you coming on the podcast I'm a big big fan of Honor Health I've met with quite a few of the previous administrative fellows there I also know quite a few people at General catalyst so it's a great organization that's doing a lot of interesting work right kind of punching far above its weight in terms of employee size uh and so I was really excited to have you on and just to give the audience some context you're a physician you also have an MBA and then how long have you been in this industry at the intersection of informatics Innovation and Medicine oh that's that's an awkward question but I actually finished my fellowship in clinical informatics so originally trained in Internal Medicine then I did a fellowship in clinical informatics back in the day where I was focused on a lot of predictive algorithms that was back in 99 so I've really been in the space for over 25 years and part of that time I've spent actually in the Imaging world just because in the early knots and teens Imaging was so much more advanced on the informatic side I've been able to kind of come back to my home with population health and internal medicine and integrated Healthcare Delivery Systems for about the past 10 years so I mean you were doing these these algorithms back when you know nobody really knew what that was right or at least the general public would not have been as familiar with that and then here we are when you know everybody is starting to learn how this works or trying to and so that's a really unique area that I think puts you at a good Advantage but we still are facing these challenges right in your area of Arizona what are some of the biggest challenges that the population in your community is facing so colge really good question what's unique about Maricopa County is we're one of the fastest growing counties in the country and so there is so much influx of new people coming here that from a health care perspective we're not growing at the same rate if you draw two lines the slope of the growth curve of people is higher than the slope of new healthc care providers what that means is um that despite the fact that we're trying to move you know more and more care into ambulatory and virtual settings kind of out of the bricks and mortar space there's just so many people moving in we continue to see a demand for more and more Healthcare resources both sort of inperson and virtual and um it's getting hard to deliver access I me we we find people are waiting you know months oftentimes to get care even myself like it's it's got to the point where and so we are so almost overwhelmed with that I mean it's a it's a it's a nice problem to have and I have colleagues in other parts of the country where the population is shrinking and that could be very challenging but for us it's really how do we figure out given the resources that we have how do we deliver the right care for the community in a way that's Equitable and fair so the first question I want to ask is let's say we do nothing and we don't do any Innovation we do no transformation we just let it right out what does that actually look like because some people will say oh this is just the natural population curves there's nothing to worry about and others uh like myself are saying this is a problem we need to use to innovate so what would happen to the community if we just didn't do anything the way I think about it is I worry a lot about the state of health care in the United States in our community to be frank probably across the globe I mean this is this is not a unique problem the reality is let's but let's look at the United States and this is true in our community we have incredibly erratic outcomes so for some disease States like if you have cancer or a broken hip for example our health system does incredibly well with responding to that we get outcomes that are best in the world but if you have a chronic illness like asthma or diabetes or you're a woman uh who's pregnant about to get worth our outcomes are incredibly disappointing so how can a system sometimes produce the best outcomes in the world and other times produce outcomes that are just just incredibly disappointing so if we just keep with the status quo we end up with this environment where we're spending a fair amount of money as a country almost 20% of our GDP and and it's not that we're not getting anything for it we just don't get a very reliable or predictable piece the flip side of that also is I you know I hope none of your listeners have ever been sick but if you ever are sick navigating the health system is hard and I'm a physician I want it to be I mean I have a lot of expertise I want it to be easy it is just it's confusing it's hard and you're having to do that when you're really scared because you're worried or when I'm sick I'm scared about what's going on with my body and I have fear running through my mind and then facing all of this complexity so I think if we just accept the status quo we end up with a system with erratic results incredibly hard for people to navigate when they they need our help and our support at their most vulnerable time yeah so that that is not going to work then right because then we end up in in a bit of a problem which is where we're seeing a lot of things now right because then that like you're saying it's it's too erratic and and then of course that has a bad effect on the economy because if people are unhealthy they can't work and if people can't work then everything eventually falls apart and we we stop running and so the the problem is though is that we've been and you and me were're discussing this we've been trying to fix these things for many years now right around you know the time I was born was when population Health was starting to get popular right and so you know here I am and population Health still has not been you know value based care still has not been fully implemented uh in the way that it was originally envisioned and so what are you know what is your stance are you optimistic with what things are going in the future are you pessimistic I mean you've been in the industry a while you know what do you what do you feel like uh the future holds well you Cole you ask really preent questions and that's it's sensitive so if you if if you indulge me I'll I'll share a little story when I was um when I was getting my MBA I happened to have just an amazing Professor gentleman by the name of Robert Burns and and he was somebody who was able to kind of go through from a health policy perspective perspective from 1960 to today all of the things the challenges that the health system you know experiences but more importantly he went through all of the solutions that we've been trying over and over again Primary Care First value based care population Health new digital tools and he by the way these are the things that I've spent my whole career implementing and trying to make work and he just kept showing over and over again how these things that we've tried you know we've oftentimes tried them as a country two three or four times and we always think it's going to be different the next time we try it and we're not getting those different results and I found this to be like incredibly depressing because I thought my first of all like I might as well just throw out my whole career because this was all what I was working on but I went to him like okay what's the answer what's the Silver Bullet what what do we need what do I need to do as a physician what do I need to do as a as a digital leader to help find us have a path forward and he he really left me with two haunting but uh Concepts that I've I've really kind of drawn to Heart number one he said Jim you know we really need to stay close to that relationship between the clinician and the patient do everything you can to make that better and don't get distracted with all of this other stuff right because that tends to be distracting and at least if you're gonna do something Innovative and we need you to do Innovative things we need you to find new things don't just repeat don't have your Innovation be the thing that's been tried three times didn't work try to find new things and so Cole that really left me on a almost kind of evangelist mission to be able to say to all of my peers we can't give up we've got to have hope but let's not fool ourselves into thinking the answer is just recycling some of the innovations that we tried that didn't work we really have to push ourselves to think in very different ways that is super hard but if you think about what the United States is best at we're best at sort of trying to figure out really really hard problems usually only after we've tried everything else to avoid it that's a good way of putting it and and it is true if you look you know historically the United States has has been kind of a comical innovator right we innovate so much but only at the point where we're really you know held with our backs against the wall I I have a problem though with with one of the things you mentioned and it's it's that we need to innovate about that relationship with the patient we need to innovate with the actual patient care but here's where I have an issue is that I'm not a clinician a lot of my other colleagues are not clinicians I mean I was a medical assistant a while ago but that's the extent of my exper experience right so when it comes to innovating in patient care how do we create an environment for Innovation as individuals who are not clinicians like yourself you know you are clinician so you have that experience how could we do better as administrators to create that environment as non-ins it's a provocative question I guess a couple of things that come to mind the first is one of the things that's hard about Innovation and Healthcare is the fact that many entities are running on incredibly tight margins right so if you look at um at a lot of Health System not all but many Health Systems either have like a 1% operating margin or a negative margin that creates challenges because just natively we don't want to innovate like it's just because the risk of failure is so high right we worry about closure right and so I think setting up safe spaces safe from both a patient care perspective and safe from a like from a financial perspective how do we create spaces where people are able to interact in this sort of in a different healthare space and digital potentially could unlock a lot of that and I actually think that some of the biggest revolutions around technology and AI if they're going to come by the way I'm a little bit skeptical about it but if it's going to come could come from outside of the health system in sort of an in in another space that's safe but we need to do it in a way that people are informed right so patients don't feel like they're being experimented on they have a right to sort of contribute into that so I guess what I would think Cole is how do we create spaces where patients and clinicians and administrators are willing to come together in a in a place that's got some relaxation of the regulatory constraints and some of the financial constraints and and it's in a fully informed way like this has to be transparent it has to be a place where there's a lot of trust we don't want to be deceiving anybody but I think there's a group of patients out there that are also hungry for something really different and I know that's vague but that's what comes to mind if I think about how do we create sort of some sort of incubation engine some sort of innovation engine maybe outside of what we've done before it's good advice it's good advice you mentioned though these external right these external vendors or external um areas of course they have you know VC funding uh you know they're not really risking their communities if they go to business because a lot of them are tech companies things like that and so um great area of opportunity for technological innovation but you said that you were also skeptical on the implementation of that you've been involved in Radiology which was one of the most promising field for AI pretty much you know kind of ever in healthcare what does the current state look like and what is the cause of your skepticism this episode of the strategy of Health was sponsored by modality Global advisors modality Global advisors optimizes Hospital Revenue enhances patient experience and delivers proven results visit modality Global advisors.com to learn more so let's think about two things and again I I want to be very careful when I start this I'm not not looking to trash anybody um it is it is really hard to see the future and if any of us were I mean maybe Warren Buffett's really good at it but if anybody else was really good at it like we would probably have a giant Financial Empire so if you think about the adoption of electronic Healthcare records you know go back to like 2010 there was a real hope that if we could digitize the American Health Care System it would unlock massive amounts of productivity there were good reasons for that in almost every other industry when you invest billions ions of dollars in it people get more productive we've seen that over and over in manufacturing Finance banking real estate all of these things we had this strange thing though happen we invested billions of dollars in electronic Healthcare records in the United States there's a lot of good that's come from it I don't want to I don't want to trash that however productivity on most clinicians went down I mean it's this fascinating totally opposite output that we didn't expect to have happen another example would be in around 2016 2017 when the convolutional neural networks were coming online and their ability to do image uh recognition just it seemed to be obvious and there were some folks making predictions that we would just not need Radiologists to interpret images anymore because these neural networks within a few years should be able to do that if you fast forward to people who know the space today we have a massive shortage of Radiologists through for a variety of reasons part of it is because our image volume went up and we didn't train up more folks but we also know that the number of people training in Radiology went down you know in 2017 because of sort of these forecasts that well we don't need this field anymore so it just it's it underscores for me uh creates a humility of how hard it is to predict and so I like to think about that in the the Medieval Times you know Physicians people like me were sure that there were humors that got out of balance in the body and if we would just bleed people enough um you know we could sort of reestablish things and people were absolutely sure that that was the truth because that's how they had been taught it wasn't until we had sort of the Scientific Revolution where we began to have this idea of I have a hypothesis I gather some data to test the hypothesis if it meets some statistical criteria then I think the hypothesis is true or I think it's not and I'm really encouraging people not to say Ai and Technology will never work or it'll always work but we've really got to stay tightly grounded in the literature uh to show where is it making a difference where is it not making a difference because we can spend a lot of money really quickly on a lot of this Tech and if it's not what I say at my own organization is technology in general and AI in particular needs to do one of four things it either needs to make our clinical outcomes better it needs to make our clinicians experience better it needs to make the patient experience better or it just makes us more efficient so that we can do the same amount of work with less resources or more work with the same amount of resources if it can't do one of those four things we shouldn't be adopting it now we can we can test it in small you know examples to make sure that it's able to yield those things but everybody's so excited sometimes around technology they forget that we need one of those outcomes to happen it's kind of like that difference you know between kind of like rationalism and empiricism where you know just because makes sense rationally right like we look at how in factories cameras now do all of the the checking of components that are are coming down the conveyor belt you no longer have to have people checking for defects even in very small items and rationally it would make sense that that would work in you know X-rays and even when we do some of these small tests on radiological Imaging it does work but when you release it to the wider population it starts having these flaws I still don't understand exactly why though that's happening and and you are kind of as you said the the expert in AI right now for your organization you're having to do a lot of research why are we not you know like just for example Elon Musk said we're going to have self-driving cars by 2021 uh similarly a statement was made in in the early uh 201s about how we're going to have no more Radiologists right but why why are we here where we have that big of a need I mean you mentioned the there are shortages and Recruitment and and of course that's been an ongoing problem but why isn't AI working to that level yet because it makes sense with all the training data all those years of it getting better and better we would be able to do it so why aren't we there yet so I think first of all this probably the most important question for us to answer so I'm going to give you I'm going to give you an answer but I myself am very humble by the fact that I don't think we as a species fully understand this yet um the first thing is that I that I had a really smart uh friend of mine tell me that we oftentimes in healthc care try to compare ourselves to manufacturing or production um workflows in terms of technology and whatnot but we are not a production line we're a repair shop and then when you're a repair shop every single like you think of like what a body shop does like every car that comes in has had a different accident I mean there's some similarities but you have to approach each one with sort of a unique sort of combination of diagnosis and fixing and you multiply that by the complexity of the human body like it turns out as as somebody once said who knew Healthcare was so complex right it is very um like it is amazingly complex and for everybody that uses the example like why can't we be like Southwest Airlines which only flies seven 37s only has one seat of course that may not be working out for them either these days but you know I think about like uh there's something you know there's tens of thousands of icd10 diagnoses there's thousands of allergies that somebody can have there's tens of thousands of medications somebody can be on just take those three things and think about all the different combinations that you can have even with the power of cloud computing like that is way more complex so that's that's one big thing that the second piece is that we've just we haven't learned in our ability to um to make predictions like and in fact we don't we don't even believe that we're bad at making predictions around how how health how technolog is can impact Healthcare so what happens people keep making predictions and they we don't even check ourselves on it the last piece that I I think is we don't understand the the interactions between AI algorithms in humans enough yet to show how we put them together um and a couple of examples and Cole you and I were talking about this you know I think a little bit around and by the way I don't play chess so if I say if what I say is stupid or or whatever please don't haunt me on social media and tell me I'm I'm an idiot I you can do that but but but I don't profess to be a chess expert but I find chess to be a really interesting model of how technology and humans work together and we had this era I didn't realize this but Alan Turing wrote one of the first chess programs back in the 50s and so there was this era from the 50s until the 90s when the chess tools would always lose to humans and and kind of by the 90s that that it didn't lose always to humans but only you know sometimes but Gary CASRO famously lost to deep blue right and that was a moment where all of a sudden AI became better than human for all there was no human that could beat uh the best human you know could not beat uh an AI but it interestingly enough led to this next era which was called Cyborg and if you're in the Chess World it's called cyborg chess so an AI plus a chess master together can beat an AI and so we entered this era that sounded great like oh well the answer is humans plus AI together can be more powerful than humans alone or AI Alone by the way it's a super comforting message it's a comforting message I hope is true for me as a doctor because I I want to stick around but what's interesting is we're starting to see some examples Radiology again is another place where we're seeing some examples where it may be that sometimes humans don't know how to use the AI or they ignore the AI or they become too reliant on the AI and they don't think critically on their own and so there's going to probably be a lot of sort of much deeper understanding that we have to do and testing to understand you know when should we have humans doing work when should we have ai doing work if they're going to do it together how do we do it together it's not just as simple it may be not as simple as the way as as like when we put two people together and create that workflow yeah I I I think it's it is a comforting message because we like the idea that we could be enhanced right we've always liked that idea where we could be you know better than possible and I think especially people who are uh you know at the terminal end of Education in their field like Physicians like researchers you know they've been through all this education um but but I I guess I have one of the questions there is why isn't that playing out right why I mean you it feels like it would at least be able to play out in terms of we've written all these medical textbooks we've written all these uh decision support guidelines right you know I you look at up toate and there's a thousand things you could search on up to date or dynamed so but that doesn't seem to be working when we plug those in with AI right you know you can't really have a uh an intern or a resident isn't as good as an attending now with AI all of a sudden right that experience doesn't translate over the AI system why is that have you seen any indication of where that's heading as well in the future well um so a couple of thoughts and I you know one of the people that I follow really closely to try to help me understand all this is Eric toppel and and Eric kind of pointed out and about three or four months ago the latest review of all the literature suggests there's only about 80 studies right now that are looking at how Ai and human clinicians are working together now 80 is better than zero but if you think about like the hundreds of thousands of papers that are out there it just shows you how new we are to the space the the AI technology particularly after the Transformer architecture sort of revolution has moved so quickly that we haven't had a chance really I think to kind of keep up from a from a um from a like a like a empirical study like from a scientific method studying perspective so I think that is that is part of it there are also tremendous complexities I think of the Watson experiment right with the Sloan ketering very expensive um and it turned out that the information that Sloan ketering had was so unique to Sloan ketering it doesn't translate so what we've learned is that it isn't it is really interesting how and this is obviously one of the sources of AI bias right you have to be so sensitive to what is the information you know that you're using to train your different AI algorithms on so that's great we've identified that so how do we have a model that can then work despite the fact that we have those insights we still haven't kind of quite cracked this code I will leave you though with a an interesting provocative piece of data that that I think is a Harbinger it's not proof yet I I give a talk I give lots of talks on AI I give them Grand rounds to I'm old so I give it to people who are my peers um and but I also recently gave my AI talk to the group of high school volunteers who were coming into honor health for the summer just as as volunteers to learn and I asked that group but a group of about 30 uh high school high school age uh folks how many of you all know somebody that is having an emotional relationship with an AI and almost 20% of them said yes that's powerful now I can go to a different audience like a group of 55y old professors and say how many of you all are even interacting with Gen of AI or would you get into self driving car and like nobody wouldn't even raise their hand so it is but what that what does that tell you it tells you that there's a group of our you know this technology is Advanced enough that for a group of people they are having personal and emotional relationships with it and that tells you there's a tremendous amount of capability there we just don't know yet how to unlock it um I don't know if that makes me optimistic or pessimistic it's a great spot to leave off on though to to give our audience something really interesting to all over thank you so much for for coming on Dr wh really interesting episode here uh I'm you know going to have to really think over a lot what you said I'm excited to rewatch it once it's it's been done edited but I hope we can have you on again in the future as well really appreciate your time well it's cole really grateful to be here you're you're a phenomenal host it's really great what you're doing thanks for your support for Honor health and for the whole field again we've got to figure out a way to make things better for our patients and our caregivers</p>