Healthcare Analytics: Right Brain Advice in a Left Brain World


Dale Sanders:                As I was reading this title and this abstract, I was going, “You know, this is going to be interesting. I wonder who’s giving this.” Then, I thought, “Oh, it’s me. I better live up to the billing here.” Hopefully everyone, I hope it lives up to the abstract in the title, that’s definitely my sincere attempt. We’ll see how that goes. Let’s dive in here. Ironically, I’m going to start a right brain discussion with a very left brain assertion, a bunch of if and, and then statements. Here’s the assertion today though, friends.

Dale Sanders:                That if healthcare analytics is fundamentally about decision support and changing or maintaining human behavior based on those decisions and healthcare data quality, other than imaging is overall very poor, given the criticality of the mission, the complexity of decision making, that is there’s a gap between the data we do have and a data we should have to support the criticality of healthcare. Changing human behavior with poor data is incredibly difficult, if not worse than difficult. Poor data if presented as good entrenches the wrong behavior and we’ll talk more about that later. Then, my final assertion is, that healthcare analytics success must be accompanied by an inordinately heavy emphasis on human engagement to overcome this shortcomings in data. That’s the fundamental assertion of today’s discussion.

Dale Sanders:                Here’s the agenda. We’ll talk about the state of healthcare data and analytics, kind of stating my case that our data is pretty poor. I’ll go over an academic reading list for the human side of analytics talking about all the factors that affect human behavior and decision making and that sort of thing and then, I’ll share the non-academic side of things, which is lessons learned from my career in the field. In the Air Force and Space and Defense sectors, at Intermountain, Northwestern, Cayman Islands and more recently the last years at Health Catalyst. Let’s talk about healthcare data and analytics in a brief state of affairs.

Dale Sanders:                I came across this paper that I wrote almost by accident. I’ve forgotten about it. We published it in 2002, in what was in the journal for healthcare information management. That journal no longer exist actually but the title of the paper was healthcare analytics standing on the brink of a revolution. This was coming to the end of my tenure at health catalyst … or I mean at Intermountain. I left Intermountain in 2005. If you read the article, I have some pretty lofty aspirations about what analytics was about to do. There’s a highlighted quote there in the middle of the page on the right. This is, process improvement and behavioral change are ingrained in every aspect of an analytic culture from the CEO down.

Dale Sanders:                In a reporting environment, these issues are less visible and more often than not in afterthought. What I was trying to say there, that there’s reporting and then there’s analytics. Analytics is really about behavioral change based around data and insights but I have to say at 2019, 17 years later, we are still not on the brink of the revolution that I thought we would be and so we’ve got to figure out why friends, for the sake of all of us and our families, patients, and families that depend on better healthcare. We’ve got to accelerate this and hopefully that’s what I’m going to help with here today.

Dale Sanders:                We’ll talk a little bit about the technology, almost nothing. What I really want to talk about is the behavioral and human side of this while we build out better technology and the digitization of the patient. There’s a cartoon that I drew in one form or another back in my Northwestern days. It amounted to our strategic data acquisition roadmap. If you look at the patient at the center of this, if we’re going to truly understand that patient, we have to round out all of the data content represented in those bubbles around the cartoon. For the most part, we’re stuck in the lower left quadrant in healthcare encounter data and claims data.

Dale Sanders:                We’re making a little bit of progress in biometrics, patient reported outcomes, playing around the consumer data, socioeconomics, genomics, epigenetics and microbiome, we’re just barely getting off the ground but we have a long way to go, friend. What I’m suggesting here is that we can’t rest on our laurels that we have keep pursuing the strategic acquisition of data about patients in order to better provide both population, health, as well as precision medicine. I’m not sure if this graph adequately captures what I’m trying to communicate here but in essence it is the quality of healthcare data, relative to the ease of human behavioral change and my assertion here is that, with poor data quality, affecting human behavioral change is extremely difficult.

Dale Sanders:                At some point, you probably hit an asymptote around human behavioral change. It doesn’t matter how good the data quality is but I think most of us would argue that where we are now with data versus where we should be given what we’re trying to do with the mission is too big a gap. As data and analytics and informatics professionals, we have to campaign to raise our performance on that axis. This is what I’d like to do. I’d like to kick off 175 billion dollar 10-year mission to digitize the patient, tongue in cheek, not the billing process which is what we’ve done so far. To me, this feels like our generation’s moon shot, is digitizing the patient.

Dale Sanders:                Frankly, I’m less interested in going back to the moon and Mars oddly given my air force background than I am about better understanding humanity and the earth that we belong to. People like John Rogers at the Center for bio-integrated electronics at Northwestern are the early pioneers. They are the Wright Brothers of digitization of the patient. Those are the folks that we need to support, fund, and collaborate with, to push a greater understanding of the patient at the center of the data ecosystem. There’s a very good article published by IEEE about the creation of digital twins that came out in 2017 I believe, a short read, worth reading. If I were the benevolent dictator in charge of all decisions like this, this would certainly be my generation’s aspiration.

Dale Sanders:                By the way that 175 billion is relative adjustment to the Apollo budget. I think the Apollo budget was somewhere around 25 billion and so this is similar kind of numbers, 10 years, that kind of investment, that kind of commitment to a better understanding of humanity and if you really peel this back, what we’re trying to do is optimize the potential of humanity. I worry that if we don’t optimize healthcare and if we don’t optimize human health into three axis of mental, physical, spiritual, that we won’t live up to our potential and possibilities as the species. This is bigger than just healthcare. This is about the potential that we have given our existence on this wonderful planet.

Dale Sanders:                If I were still a practicing C-level in the trenches of health analytics, this is the kind of conversation that I would have with my physicians. I’d say, look, here’s the truth. We know that all of these quality measures are a giant pain in the neck. We know they are burning you out and stripping you of your sense of mission, autonomy and purpose. We also know that most of these measures are clinically irrelevant. It’s a spiral of frustration. You are forced to collect irrelevant data, then we measure your performance according to that irrelevant data. Going forward, we will minimize all of that burden including the pressure we put on you and we commit to giving you the data you want, when you want it and how you want it so that you can practice to the top of your profession.

Dale Sanders:                I think quite often, we are reluctant to acknowledge what physicians know instinctively, that the data that a lot of us process and interact with is a poor version of the truth and not really what physicians need to provide the best care possible. Going with the softer side of data and acknowledging these kinds of issues, being transparent about these kind of issues and truthful, goes a long way towards building a collaborative environment and also acknowledging that it’s okay that physicians be burned out and thereby, relieving some of the pressure on them. The state of healthcare data, I would argue that significant amounts of it is qualitative, not quantitative.

Dale Sanders:                I’ll go into a quantitative data here shortly but some of the examples of quantitative data that we have are lab results, some of the measurements that we see in diagnostic imaging, vital signs, dosing, genomics, anything that you can assign a discrete number to, a number not a diagnosis code and not clinical notes. Those are qualitative data sources. Clinical notes are … they need some work frankly, right? There’s lots of studies that indicate that our clinical notes are reflection of billing and malpractice motives not about patient care and handoff but they do represent the potential to be very valuable source of data but it’s not quantitative, it’s qualitative.

Dale Sanders:                It’s analog. We tend to describe diagnosis codes and ICDs as being discrete data as if discrete data is quantitative but that is not the same thing. Discrete data is not necessarily quantitative data. Quantitative data is the world of numbers and math and speaking of that, and acknowledging our shortcomings, I think it’s important for us as an industry and as a professional body of science to acknowledge that our mathematical models in healthcare are inadequate given the importance of our mission. Now, is that the fault of healthcare or is that the fault of the mathematicians? Probably a little bit of both.

Dale Sanders:                The maturity of any body of science and engineering is directly proportional to the mathematic models which describe that body’s predictability and reproducibility. That’s my assertion, inspired by the writings of Thomas Kuhn by the way, a well-known philosopher and scientist. Observation, measurement and data collection precede the development of mathematical models. It’s Newton sitting under the tree with the apple falling on his head that helps him observe and start to think and model gravity. We must acknowledge the overall lack of data and maturity of mathematical models in healthcare and adjust our approach to decision making according to the reality of our current state.

Dale Sanders:                I’m not saying that we should dismiss the data that we have. We just have to adjust the way we interact with data, the way we communicate uncertainty to one another. The way we communicate uncertainty to patients within the acknowledged shortcomings of the data and models that we have. This is my depiction of that progression, many of you have probably seen this before in other lectures and webinars. From left to right, this is how I see the progression of mathematical models as it relates to the predictability and the reproducibility of a body of science with the Newtonian physics being on the far right, jokingly the Facebook Ads Feed being on the far left.

Dale Sanders:                The math associated with human psychology and economics is pretty abysmal in terms of predictability and reproducibility. Now, we are just barely ahead of that and yet still slight behind biology because if you really think about healthcare, it’s the combination of biology and psychology for the most part. Human behavior and the biology and human body coming together is what constitutes healthcare. On an individual sort of cellular, genetic layer, biochemistry layer, we understand the human body reasonably well. What we don’t understand is how that relates to health and how human behavior affects that biology. I would say if not for randomized clinical trials, which have problems of their own with predictability and reproducibility, our mathematical maturity would score even lower.

Dale Sanders:                Simply acknowledging that as an industry and saying, okay, what are we going to do? How are we going to inspire mathematicians, physicist and others to participate in the modeling of healthcare so that we move to the right on this diagram. In order to build those models, we have to collect more data, is my fundamental assertion, going back to physics, electrical, computer, aeronautical, aerospace engineering. That’s all about collecting data than building models to describe what you’ve observed in that data. Data is fundamental to moving to the right here. Now, we get credit from randomized clinical trials but if you want to read a really well-written book by a very capable investigative journalist from NPR, Richard Harris, this is a sobering assessment of published research.

Dale Sanders:                It talks about the inequality of mouse models as it relate to human models, a field that I knew very little about which is cell mis-identity and contamination. I had no clue about that. There’s a couple of chapters on it. I think we’ve all heard about the debate and the danger of P-values and then of course the culture of research that tends to inhibit open data sharing to allowing scrutiny and reproducibility. Reading this book was motivating to me about how important it is that we move aggressively towards the digitization of the patient as a country and also start adopting and changing our human behavior relative to the data we collect.

Dale Sanders:                This is a metaphor that I used to describe the situation that we’re in. This is my family on the left at Yellowstone last fall. That’s a real picture of my analog life but I would argue that this is healthcare’s digital view of my life. It’s a highly pixelated image. On top of this poor data, we have the unusual complexities of human behavior. Tribalism, like I’ve never seen in my life, in today’s society. The social norms of the tribe being an enormous influence on behavior, there’s the truth according to data. There’s the truth according to perception and going back to tribalism and the whole debate about fake news.

Dale Sanders:                There’s homo economicus and evolutionary biology that is … what amounts to modern day survival of the fittest. How do we express our genetic propagation in today’s world when physical survival is not quite what it was to Neanderthals. It’s more now about the economics, the driver behavior, as a proxy for genetic propagation. Then, of course, there’s the behavioral economics side of this which is the irrational side of homo economicus. The economy and our economics stature and environment affects our behavior. Our tribalism affects our behavior, it also affects our truth of … or perception of the truth and it turns out that social norms of the tribe tend to be the most dominant influence on our behavior.

Dale Sanders:                With analytics, if you go charging head long into a violation of the tribe and the social norms of that tribe, you’re going to elicit an immune response from the participants. They’re going to reject your analytics. They’re going to refute your analytics. They’re going to layer their perception of the truth on your data and you’re going to stall with progression. I credit a friend at Northwestern, a psychiatrist, researcher at Northwestern that first enlightened me to this issue of tribalism, all the way down to belonging … sense of belonging to teams and professional groups. I thought about this in my sense. I almost have no tribal affiliation. The closest thing I have to a tribal affiliation is my air force identity and that’s pretty distant.

Dale Sanders:                I’ve been a bit clueless to tribal identity, until I would say the last 10 or 15 years and I wish I would have known more about my violations with analytics, of people’s sense of belonging to their tribe. It’s critically important. When you engage with analytics with an individual, you need to understand their personal identity and the social norms of the tribe or tribes that they belong to. Really important and then work with them within that context and realizing that if you step on the identities of that tribe, you’re likely not going to progress with your mission and your dreams and hopes of analytics.

Dale Sanders:                Okay, so let’s talk about a reading list here, our decision making, human behavior, all of those things. I had an undergrad student in a lecture ask me a few years ago, what’s the most valuable class you’ve ever taken? I was pleased to realize I didn’t have to give it a second thought and it was an epistemology philosophy class called What is Truth, taught by Dr. Paul Pixler, in tiny little Fort Lewis College in Durango, Colorado. That book and that class and Dr. Pixler, taught me how to think. They taught me how to peel away the false layers of belief until reaching fundamental truths but I also said, it confused the hell out of me and I’ve never been quite sure of anything since.

Dale Sanders:                The book is no longer in print, which is unfortunate but if you can find a copy, it’s a great distillation of great philosophers talking about the pursuit and the understanding of truth and people’s various perceptions of that and how we as day to day human beings can apply that when we interact with people and when we think we’ve found the truth of our situations but it has fundamentally affected the way that I address problems and the way that I think it created a greater sense of empathy in me towards the various flavors of truth and the human perception of that. This is another favorite book of mine. Maybe my favorite sort of business related book in the modern era that I’ve ever read.

Dale Sanders:                A lot of you hear me talk about it all the time, Daniel Pink’s book about what motivates human behavior. It’s a modern take on Maslow’s hierarchy. A little simpler for me to grasp and that is every human being strives for mastery of a skill that they can feel good and proud of. For the most part, they want to be left alone. They don’t want to be micromanaged but if they have a need for help, they want to feel like they can ask for it but for the most part, they want to be left alone. They do want to feel a sense of purpose and belonging to a mission that’s larger than who they are. When we as data and analytics professionals engage with clinicians, we have to ask ourselves, what are we doing with our data, with our skills to enable greater mastery, autonomy, purpose for our clinician’s patients and administrative leadership?

Dale Sanders:                I would argue that in the US, our national analytic strategy is detracting from these basic human needs. If not, demoralizing them all together and it’s time for us to change. A couple of more books. Now, these are kind of an interesting mix. Will Durant is one of the best writers in my opinion of all time. There’s no book that I wouldn’t recommend if Will Durant was the author. This is a great fascinating look back on history and my interpretation of this book is that history doesn’t repeat itself but human nature does which then leads to repeated history. If you understand these patterns of human behavior that have repeated themselves for thousands of years, I believe it helps us understand the patterns that we see in human behavior today.

Dale Sanders:                It helps us anticipate the patterns that we’re going to see so that we can get our thinking ahead of that rather than reacting to it, we’re thinking ahead of the human patterns that we will see in our analytics and data strategies and healthcare. One of the great quotes here is that, “Therefore the laws of biology are the fundamental lessons of history. We are subject to the processes and trials of evolutions to the struggle for existence and the survival of the fittest.” That really boils down to our analytics strategies. Well, when you’re interacting with clinicians and things, in addition to mastery, autonomy, purpose, you have to ask yourself at a very biological level what are we doing that might make this human at the other side of the discussion feel better about their future and their struggle for survival at a very biological level.

Dale Sanders:                I had a great piece of advice one time from a mentor of mine TRW who said, never threaten another person’s livelihood. That was mostly at … in the job arena but what he was saying is that if you’ve threatened someone’s livelihood in their job, you’re essentially threatening their sense of survival in the biological sense. If you ever hope to change anyone, if you ever hope to interact collaboratively with anyone, you cannot threaten their sense of livelihood and survival. The next book here is a quiet book, that very few people know about. I would have endorsed it whether it was the 50th anniversary of Apollo 11 or not.

Dale Sanders:                George Mueller was the quiet unsung hero who led NASA’s human space life program. Nobody really knows about him. He came along after lots of problems and failures, early failures in the program in 1963, took over and he’s the person who organized the human space life program to success and he did it in six years. A lot of people think that we did it at the end of the decade as if we had 10 years. The reality is, George Mueller did it in six years. It’s unbelievable. If you want to read a book that describes incredibly complex organizational challenges, times unprecedented engineering challenges times political challenges times human lives at stake times hopes of the world on your shoulders, and he lived to be 97 years old.

Dale Sanders:                I can’t imagine anybody living through the stress of something like that. He lived to be 97 years old. One of the reasons I adore this man is because he was also a TRW employee and I was … I work for TRW long after he established himself but he was a lingering presence in the TRW culture so we all had a sense of worship for George. He was brilliant but humble, capable at the high levels of thinking strategy level but also at the lowest levels of electrical engineering. There are 410,000 employees and contractors involved, 20,000 companies and in those days, we didn’t have webinars and WebEx. He had to patch together things like conferencing systems to keep up to date on the Apollo program and the manned space flight program.

Dale Sanders:                I included a copy of his org chart by the way just for the sake of reference. I have to say when … and I don’t hear it much anymore thankfully but when healthcare leadership talks about their moon shot, they’re not even close. They are actually I think denigrating the significance of the accomplishments of those associated with that achievement. What we do need to do is take on a similar level of commitment. If we are really committed to the full potential of the human species, this is our moon shot. This is what we have to do in healthcare, is think bigger and higher than we ever have before.

Dale Sanders:                One of the single best references for, The Art of Decision Making, the human side of decision making is in this Harvard Business Review article when HBR was a little bit better than they are now. Frankly, they are not as good in content as they used to be. This is from … it’s a tattered, dog-eared version. I think this is from 2015. I’ve carry it around forever thinking that it had long since gone out of print but I’m surprised to find that you can buy it on Amazon. Of all the books that I’m suggesting for reading today, this is at the top of list, probably the top three references. It’s the most condensed and succinct, about the human side of decision making and making sure you don’t get trapped in all the biases that we’re subject to being trapped within.

Dale Sanders:                I think I’m almost done with the book recommendations here. If you haven’t read Nudge, it’s a great relatively easy read from a Nobel Laureate. In summary, it would say … and I’m grossly oversimplifying here but give people freedom of choice, right, back to autonomy according to Daniel Pink. Give freedom of choice but nudge, don’t shove with data towards the preferred behavior. I’ll talk a little bit more about that in my lessons learned, in my career. The book in the middle, On Human Nature is a self-help book. I’m usually kind of cynical about self-help books but this is one that I recommend reading.

Dale Sanders:                It was recommended to me by a friend whose judgment I trust and as I was reading it, I kept thinking wow … this person by the way, the author, Ed Wilson, doesn’t have a cognitive science background. I think he’s just an author and a historian essentially. He weaves really interesting stories of history in human behavior, similar to what Will Durant did in his book. He ties it back to self-improvement and what you can learn from those patterns of human behavior over the centuries. As I was reading it, I thought wow, yeah, this person is describing my shortcomings and flaws and then I also realized that it was helping me understand other people and how to interact with others. By the way, I am by far the least perfect person when it comes to interacting with other folks.

Dale Sanders:                I can be blunt, I can jump to conclusions so I have all sorts of flaws here, friends. I don’t want anyone to come across believing that I believe in myself here. I can be like a bull in a China closet when it comes to changing things but I keep trying to improve and that’s what these books are all about. The last book on this page from Tali Sharot, a brilliant cognitive scientist. I believe she’s at the University of London. She’ll be speaking at our health analytics summit this year by the way. I was really excited that she agreed to speak. She comes at the human side of data driven behavioral change from a scientific perspective that I think is better than any other book I’ve read.

Dale Sanders:                I highly recommend this book and it kind of confirms common sense and that is data is persuasive when it fits your world view, right? If you present data to someone today, if it supports their view, wow, they’re all over it. If you present data insensitively, how did I say that? Can anybody help? In other words, if you violate a person’s sense of tribalism and survival with the data that you present, guess what, it’s going to entrench the opinion that you’re hoping to change. Boy, do I see evidence of that like I’ve never seen in my life in today’s society. If any of my Facebook friends are listening today, we had quite the spirited debate about the recent tweets from President Trump towards the four members of the squad in Congress.

Dale Sanders:                It was quite an interesting debate to see both sides of that opinion emerge and there’s no budging opinions with data in that context. My advice to everyone is give up. Stop convincing and start connecting. Okay, another great reference here that has had a profound impact on my career in life and this is one that I actively day to day, think about and practice. The RAND Corporation wrote this paper. I think it was in the mid 80s. It’s about the time I was exposed it. This is actually my flavor of that paper so it’s not exactly the same but it was inspired by the RAND Corporation. What they essentially said is that, if you’re going to be an effective change negotiator … they were actually talking more about treaties at that time.

Dale Sanders:                There are three qualities that you have to exude to the persons and people on the other side of the negotiating table. That is you have to be believable, relatable and credible. You have to be honest, sincere, trustworthy and if you … if the people don’t know you well-enough, to know that, they have to see evidence of that in your background. Believability is fundamental. You cannot compromise on that. Relatability is the question, is this person similar to me? Can this person empathize directly with my situation and role. This is where tribal affiliation starts to take place. All right. Can you affiliate with the sense of identity and tribes that I belong to, whatever they might be.

Dale Sanders:                You have to find common ground for relatability to be an effective change negotiator with analytics. It doesn’t matter if it comes down to shared sports teams. Maybe the same hobbies but at some human level, you have to find the ability to relate to the other person in order to have an effective relationship from an analytics and data perspective. Then, the last criteria is the credibility factor and that is, even though there is a large overlap in the Venn diagram in relatability, hopefully, there’s a large relatability overlap, there has to be something about you that’s different enough and respected enough by the person on the other side of the table that they appreciate you and they welcome your dialogue, they welcome your expertise, they welcome your involvement in their lives.

Dale Sanders:                This is something that … I would say, we’re not great at this in Health Catalyst culture but I would say, there are elements of this in our Health Catalyst culture and I’m probably the biggest flag bearer, but being very deliberate down to the assignment of client, executives and account representatives down to these kind of criteria. Making sure that we’re putting the right people in the right place for their success as well as the other folks on the other side of the table. Okay, now, let’s breeze through lessons learned from the field. My career started in 1983, in the Air Force as a command control communications intelligence officer, specializing in nuclear warfare.

Dale Sanders:                There were interesting parallels, believe it or not. Like healthcare, there are strong personalities. It’s life and time critical, there’s this interesting mix of subjective and objective information that you have to synthesize and make these critical decisions in very tight, compressed timeframes. What’s not like healthcare was the military’s insistence on more data, more intel before making high risk decisions. I’d like to think that we were better at this in healthcare than we are. I mean, there’s literally an insistence among military leadership and there is great impatience, if they don’t have access to the right data at the right time, to make high risk decisions.

Dale Sanders:                That military insistence on better data and better intel actually drives the technology acquisition strategy of the military, right, so the military leadership doesn’t sit passively back and wait for the industry to provide this for them. The military leadership proactively procures the technology to satisfy their need for more data and more intel. I don’t see that same sense of insistence and drive from our clinicians and administrative leadership. What I see is more passive acceptance of mediocrity when it comes to data, as if they can’t do anything about it but the reality is you can’t force through free market capitalism the progression of better data and analytics into the healthcare industry if you want to.

Dale Sanders:                I will also say that not like healthcare is that … frankly, digitizing aircraft, satellites and ICBM is easier than digitizing humans. I get it but for the same reason, that we took on the challenge of the human space flight program, we have to take on the digitization of the human ecosystem. By the way, these are actual photographs of the work that I did in the air force. There’s a Looking Glass aircraft at the top. There’s a DSP Reconnaissance Satellite in the middle and Minuteman ICBM below. That’s the world of digital engineering, digital ecosystem that I grew up in and I came with that mindset in sort of that expectation when I transitioned into healthcare.

Dale Sanders:                That’s essentially what I’ve been striving to do ever since I joined in 1995, ’96. If you want some interesting reading about this general topic, Google aircraft health maintenance systems and prognostic and health maintenance in the aircraft industry. There’s a lot of great articles out there and you can actually see how in the 1990s, the aircraft industry and airline industry made a very purposeful decision to become more data driven and more digital. Today’s Boeing 777 has a 11,600 reportable faults, 270 prognostic alerts. They collect more than a terabyte of data per flight. All of that feeds into an aircraft health maintenance system.

Dale Sanders:                Think of that as the electronic health record for the aircraft. By the way, there’s a level of analytics on top of it and decision support called prognostics and health maintenance. Ironically, same three letter acronym as Population Health Management. The concepts apply, now, we need to think about the digitization of health care more like engineers and less like clinicians at the moment, right? There are some pilots involved in these aircraft health maintenance systems but for the most part, it’s the engineers who are trying to maintain optimum health of the aircraft and optimum length of life. It’s the same exact thing that we’re trying to do with patients.

Dale Sanders:                Quality of life and length of life. That’s exactly what they’re trying to do with an aircraft. We need to start hiring more physicist, more aeronautical engineers, more electrical engineers and bring that influence and that mentality into healthcare. I’m going to make kind of a tangential suggestion here back to my air force days. The air force was suffering a lot of failed contracts a couple of decades ago. They stepped back and said, look, our traditional approach to contracting and treating negotiations are not working so they funded some research at the University of Tennessee and University of Tennessee created a framework called Vested Contracting and Negotiations.

Dale Sanders:                It focuses on outcomes, not transactions, focuses on the what, not the how. Pricing models are incentivized to optimize both parties. Insight and resolution of issues versus oversight governance structures, focuses on open book financial discussions. I’ve tried several times when I was a CIO in healthcare, I could actually invoke these kinds of contracting requirements because I had the autonomy to do that and it’s effective. I would highly suggest that all of you think about embracing Vested Contracting and Negotiations including the way we contract with physicians. Including the way we contract with physicians.

Dale Sanders:                I think our physician contracting is a vestige of a bygone era, dominated more by lawyers and I think we need to be more humane than we have been and ironically, this is coming from the military, it’s a common sense approach to contracting and negotiating. Okay, lessons learned at Intermountain. It was very interesting, right? I had no clue when I joined Intermountain of what an IDN was. It was slow for me to realize that this economic balance between hospitals, clinics and insurance was a very healthy tension. A single CEO managing the economic motives of all three of those missions balances very effectively care, quality at the lowest cost possible. That economic balance actually makes it easier culturally to achieve behavioral change, because the economic incentives, everybody is kind of lined up.

Dale Sanders:                More importantly, that gets overlooked I think sometimes is that the Utah’s state symbol is the beehive and it’s not about honey. It’s about working together selflessly for the common good. It’s an embedded part of Utah’s culture. The Intermountain’s culture is very communal. Reducing clinical variation fit that culture. Everyone would line up around the philosophies of Brent James and David Burton to reduce clinical variation but outside of IDN’s … or outside of Intermountain’s culture and outside of Intermountain’s economic model, it’s really hard to find sticky, long term commitment to clinical variation reduction.

Dale Sanders:                I just got back from a conference in … an international conference in Canada and I can’t remember, I think it was either eight or 12 different countries who are represented and every single country is struggling with the same thing, that is the adherence to what we call evidence based care. Well, what I would suggest is, based on what I just discussed about the poor state of data and research, we don’t really know what evidence based care is. If we really don’t have a substantial understanding of evidence based care, we don’t have mathematical models, we don’t have data to inform that, you can’t expect physicians to line up behind evidence based care that they feel instinctively violates what they see day in and day out.

Dale Sanders:                There again, you have to address the human side of this within the context of the data that we have access to. Lessons learned at Northwestern. My first executive level meeting was a disaster. Northwestern recruited me to build a data warehouse, they created a CIO position for me. What they really wanted was analytics and data warehouse. I went in to my first meeting, all full of myself, all … I’m the big new hero on the campus. First meeting with the deans, council and I came in with an Intermountain centric message for analytics and it didn’t fly. It was a Hindenburg. Dr. Drew Landsberg, the dean pulled me aside afterwards.

Dale Sanders:                He’s a great guy, a very, very humane, soft-hearted but very direct leader and he said, you got to understand Dale, you just told those physician researchers that they should reduce their variation in care but they came here to experiment with controlled variation. You better come back with a different message. It was like a lightning bolt realization. I mean, literally, he didn’t have to explain it to me. In that conversation, I realized exactly what he was saying. The team, me, Eric Just, Mike Doyle and others that worked together there, we pivoted our strategy towards research and addressing the needs and the motives of the research side of Northwestern and we became data heroes.

Dale Sanders:                It was one of the most satisfying periods in my life and career. The other thing that I learned at Northwestern was that you can make the best out of data with good process. Just a framework that everyone should think about in analytics is the PICO Framework, right? Asking the right questions. The patient, how would you describe a similar group of patients to the one you’re interacting with, that’s a patient cohort. What are the interventions that you’re considering and how are you going to measure those interventions, do you have the data to measure the interventions and what do you want to do to this patient and this group of patients?

Dale Sanders:                Then, who are you going to compare the intervention to see if it’s effective or not. This is kind of experimental design but it doesn’t have to be that formal. It can just be a thinking and framework thought process. Then, finally, how are you going to measure the outcomes. As analysts, this has nothing to do with SQL logic, really. This has everything to do with asking the right questions of the humans in the room, the patients and the clinicians and then building your analytics and data strategy around it and just following this framework, will help make the most out of the poor data quality that we have while we improve the quality of data over time.

Dale Sanders:                I want to call out, an interesting thing that happened to us at Northwestern. When I landed there, both the Cerner and Epic systems were pretty poorly underutilized. This was in 2005. We scratched our heads and we say, wow, we’ve spent a lot of money on this. It’s not being utilized. I think we better publish some core principles of clinical documentation and we can purposely constrained it to one page. We didn’t make measurement the core message. What we did is made common sense the core message so if you look at … that’s the actual document on the right hand side. If you look at that, it’s not loaded with measures. It’s loaded with common sense that almost anyone can agree to.

Dale Sanders:                Oddly, I had a conversation in 2007 with folks that would later play a large part in ONC and the development of meaningful use. They saw what we were doing, by the time I talked to them, by the way we had built a dashboard, we were giving feedback, nudging type feedbacks to the clinicians about their use of Epic and Cerner. We weren’t shoving anything down their throats. We’re just letting them know this is how you’re using the product that you paid for, relative to your peers and relative to what we think are good benchmarks. You do what you want to with it, giving them the sense of autonomy to adjust on their own. We had a dashboard, we have this document. This actually became the seeds of what ended up becoming the weeds of meaningful use and I’m struck by the way history twisted that the wrong way.

Dale Sanders:                Starting in 2009, I would say that meaningful use became meaningless use and that’s when physician burnout started to emerge in its peak as we’ve adopted the HRs and continued to roll out the concepts of meaningful use even if we call it something else. It drained the analytic energy out of the industry and we are obligated, friends, as taxpayers, patients, members of community to unwind the legacy of meaningful use as quickly as we created it and start over. It’s not just for patient care, it’s for the efficiency of the industry, it’s for the benefit of clinicians who are burdened and burning out as a consequence.

Dale Sanders:                Okay, Lessons learned in the Cayman Islands. That was an interesting experience for me. I loved it. I’m endlessly grateful for going there. It was a population health laboratory in a very learning … heavily learning environment for me. At the end, I had the nudge, not push to embrace analytics in the Cayman Islands. There was some interesting reasons behind that. Not the least of which was the autonomy of government-employed physicians in a small labor pool, made behavioral change quite difficult, right? We couldn’t go out and hire new folks that weren’t willing to collaborate and if they weren’t willing to collaborate, there wasn’t a lot we can do from a human resources perspective. It tested our diplomatic skills to their fullest.

Dale Sanders:                The culture was also caught between the US and UK influences of healthcare which made analytics kind of confusing, which analytic strategies do we follow. For the most part, we built our own and I want to emphasize that culturally, I’ve never been treated better than I was treated in the Cayman Islands. I love the people. I still have a fond desire to go back there sometime. There was always this sense though and I think there’s truth to it and it’s okay that there was the American Guy with my own agenda and I’d only be around for a short time, all we had to do is kind of wait the guy out, that being me. That’s a normal human reaction to an outsider. I have no problem with that, being the case nor observing it.

Dale Sanders:                At Health Catalyst, I’ve observed a lot of interesting things across dozens and dozens of clients in the industry and here’s my observation, depicted in the cartoon on the left. The head-scratcher. At Intermountain, our skills were developed and rooted there but they play well in some places but not in others. I mentioned earlier that culturally, Intermountain were unique. Economically, there aren’t very many IDNs. There aren’t many beehive cultures like Intermountain. The clinical variation reduction message that we have strongly advocated has taken a bit of a back seat in the industry to an emphasis on the quality of care analytics tied to reimbursements and preventive care.

Dale Sanders:                Essentially, a lot of the MIPS, MACRA, Meaningful Use, et cetera, all of those difference in incentives now that are now tied to quality analytics, I’m not sure if it’s good or bad but that’s just reality. We had to adjust the way we engage with our clients so that we don’t assume that their culture and their economic models can accommodate the head-scratching cartoon on the left of the diagram there. My advice to analytics practitioner is that you have to adjust your analytic strategy to the cultural economic drivers of the client. Each of those market segments, each of those client segments has slightly different behavioral economic drivers.

Dale Sanders:                I’m not going to go through all these details. The slides will be available for reference. A quick scan of that will show you that an IDN has different economic drivers than a large physician practice or direct to employer healthcare. Certainly an academic medical center has different motives, economically and culturally than a community hospital. You have to adjust your analytic strategy to line up with the economic and cultural drivers of the organization. Okay, wrapping up. Find your data empathy, especially for physicians right now. They need us and we need them. Find your data empathy, engage with them in a way that progresses the industry rather than pulling it back which I’m afraid for the most part we’re doing right now.

Dale Sanders:                Be truthful and acknowledge the limitations of our existing data and analytics then just adjust your decisions accordingly. I’m not saying that we should grind to a halt. I’m just saying that we should use our data within the context of its quality and thoroughness and then adjust our decisions and the way we communicate degrees of uncertainty to one another and to patients. Give physicians some form of analytic hope. Give them some data that they want and need. Find it. If you don’t have it build a strategy for it but give physicians some form of analytic and data hope. Look for and adapt to the social norms of the tribe that you’re involved with or you’re dealing with and void those head-on collisions.

Dale Sanders:                Don’t tread on the tribe, if you’re trying to affect change. Nudge, don’t push. Connect, don’t convince and let’s all campaign for digitizing the patient. Rather than going to the moon, rather than going to Mars, let’s digitize humanity and let’s realize the full potential of the human species that will then enable us to go to other planets. All right. Thank you. All done.

Sarah Stokes:                Great.

Dale Sanders:                Thanks friend. Let’s see we’ve got a couple of other slides, don’t we?

Sarah Stokes:                Yeah, we have just a couple of wrap up poll questions here before we jump into that Q and A. Do you have availability to stay over for a few minutes?

Dale Sanders:                I can stay over, yes.

Sarah Stokes:                Okay, right. Okay, so before we move into the Q and A, we have a few giveaways for complimentary healthcare analytics on the registrations. This is an annual event with more than a thousand provider and pair attendees occurring this year, September 10th to 12 in Salt Lake City, Utah. The event will feature brilliant keynote speakers from the healthcare industry and beyond. This slide gives you a glimpse of some of the folks that are confirmed for this year’s event. Let me go ahead and launch our first poll question here, which is if you know that you’re able to attend and are interested in being considered for complimentary passes for a team of three to attend the Healthcare Analytics Summit, please answer this poll question.

Sarah Stokes:                I’ll give you just a few moments there to get your votes in, because we are nearing the top of the hour. Okay, I’m going to go ahead and close that. One, two, three. You’re going to have to act fast on these. Okay, and then one more giveaway here. If you know that you’re able to attend and are interested in being considered for a complimentary individual pass to attend the Healthcare Analytics Summit, please answer this poll question. Again, you will have access to today’s slide after the fact, I know Dale gave a ton of great reading recommendations so you will be able to follow up on all those after the fact and we will also be posting a recording of this session.

Sarah Stokes:                Okay, I’m going to ahead and close that poll and then move on to our final question which we will leave this open as we go into the Q and A. Well, today’s webinar was focused on the importance of utilizing those right brain skills in the left brain world of Healthcare Analytics. Some of you may want to know more about Health Catalyst products or professional services. If you would like to learn more, please answer this poll question. I’m going to go ahead and leave that open and then, Dale I’ll pull up the questions here for you. Now, a lot of these are really detailed so …

Dale Sanders:                Okay, great.

Sarah Stokes:                You’re welcome to click through on my computer too if that’s helpful to you.

Dale Sanders:                All right. I might ask you to do it. Maybe I’ll do it on your screen. Okay. Let’s see here. Let me go through these, try to triage these a little bit.

Sarah Stokes:                You could use the mouse right here.

Dale Sanders:                Thank you.

Sarah Stokes:                It would be easier for you.

Dale Sanders:                Let’s see, let’s read this one here. My sales experience in startups with a clinical research organization automation software for phase one, so it drives humans to do something, mostly what’s in it for me or how does it relieve my pain, right? The only way to get to that point is for a human to allowed to be vulnerable and share their pain or the gain outweighs the risk. I see this is people willing to provide information into an exchange or allow data to be seen, shared from a policy perspective or patient not knowing how it benefits them. So many syncs of data but how to make people allowed to trust and want to connect and share, if the motivation is enough, then it would supersede the need to control.

Dale Sanders:                Yeah. That’s a deep thought Suzanne. I agree friend. I don’t know really how to respond to that other than I agree with it. I think it’s very thoughtful and eloquent. There’s a question here, isn’t this a condemnation of the STEM movement? Well, I didn’t mean for that to be the case and I referenced a lot of engineers and physicist, I’ve been in healthcare and health metrics for 25 years, but my undergraduate is in history and political science. The humanity students have always known this. It’s the engineers and MBAs are those who don’t. Well, I think there is some truth to that David. I mean, I grew up … my undergrad is in chemistry, biology, double degree but I went to a liberal arts school and philosophy minor was part of what I did.

Dale Sanders:                I took art. It’s been invaluable to me and I think if I had any success in life it’s because I have had the ability to touch that right part of my brain. I’m actually saying, what I think we need to do is bring more soft skills into the engineering world, into the left brain world, is kind of the whole point of the webinar today. Again, if you go back to the diagram of mathematical maturity models, some of my clinical friends will be irritated with this but you don’t have to have great math skills to become a physician or a clinician. The appreciation for … you have to have good stats I would say, statistics but not so much. The kind of algorithms and math you get exposed to in chemistry, physics and the engineering disciplines. I think we need more of that.

Dale Sanders:                Okay, how do you see pharmacy play in this realm of right and left brain as we have prescription data yet, pharmacist can also play … it’s gigantic. I mean, that’s … Pharmacy should be empowered to do a lot more to aid and advise patients, submit orders, refill prescriptions. I mean, almost every country, I would say, come to think of it now, the commonwealth countries in particular but even beyond the commonwealth, allow pharmacist to practice a lot more at a higher level than what we do in the US. We need to let go of the patriarchal approach that kind of puts physicians at the center of every single decision and just let it go and leverage the skills and the knowledge and the human interaction that pharmacist are so good at.

Dale Sanders:                Yeah, totally great point Dee, thank you. When our data conflicts with the subjective qualitative information, we receive, sometimes we are measuring invalid and unreliable metrics, varying quantitative and qualitative data even in patient’s family, clinician narratives, this links emotion to data. Interesting point. Data that is not believed maybe due to the presence of conflicting experience, objective incoming data. That said, much of this is well-stated in novel and healthcare. Associate Russell Ackoff’s DIKW Pyramid, yeah and wallowing in data level without proceeding is evident in health care. Well, Rebecca that’s very thoughtful. I’d like to hear more about your thoughts in this regard. This is very thoughtful. Connect with me in LinkedIn if you don’t mind, Rebecca Hancock.

Dale Sanders:                Let’s see, a question here. What type examples of data from healthcare do you have or can you show like from HEDIS, Carepoint and other type of systems that have information, what does the future of data might look like? Well, I’m not quite sure Edmond that I fully understand your question. Sarah, do you … can you help, am I missing something there? I’m sorry Edmond, I don’t fully understand your question friend.

Sarah Stokes:                Edmond if you’re still on the line, you can just resubmit, rephrase it a little bit.

Dale Sanders:                Yeah, that would be great.

Sarah Stokes:                Then, we’ll circle back to it.

Dale Sanders:                Thank you. I mean, I’ll answer the last part of that, what is it the future of data might look like? It’s going to be more generated by the patient. In fact, what we’re seeing in the old days, all of the data that we’re collecting as a society, all was collected in the data center. Well, what’s happening now is that 70, 80, 90% of the data that’s being collected in the world, is being collected at the edges of human’s experience. It’s not in the data center anymore. Absolutely, bio-integrated sensors and patient generated data is the future. The volume of data, it’s going to go up exponentially to the edges of the organization and society.

Dale Sanders:                Let’s see. Hi, John Kenegi. The evidence presented suggest the solutions to digitizing healthcare lay outside the social norms and capabilities of the current tribe of digitizers. Let me think about that. Is it outside the current tribe of digitizers? I’m not so sure John because I think ultimately, we all belong to the patient tribe. I mean, I think that’s one thing that we can all rally behind. Now, that may … Quite often physicians don’t see themselves as patients but their family certainly are and I think the tribe that will drive digitization has to be two tribes. It has to be patients who insist on better care.

Dale Sanders:                Then going back to the military analogy, it has to be the leadership in healthcare that demands and insist on better intelligence data to support their mission. That’s the thing that the military leadership does very well. They insist on better data and they procure and they drive the development of technology that leads to better decisions. Whereas in health care, physicians and administrators sits sort of passively back and buy mediocre products, spend billions of dollars on those and they don’t insist on something better. Gigantic misstep in our country’s evolution of healthcare when we incentivized all of that money from high tech on existing products that didn’t progress the digitization of the patient in any substantial form.

Dale Sanders:                Okay, let’s see here. Lawrence Epstein asks, would not banding the patient by segmentation of expected need have more success than digitized individuals? Well, I would say Lawrence that we can’t accurately band the patients by expected need if we don’t have more data about them. It’s a smeary image. It goes back to that highly pixelated image of me and my family. Now, we can take … and we can make an attempt to segmentation based on the data we have but segmentations are not very precise, right? I kind of live in this world for years. The patient types and patient segmentation that we have is pretty imprecise. It’s grossly imprecise actually.

Dale Sanders:                We do have some hope ahead though. AI, machine learning, pattern recognition algorithms are allowing us to see patterns and data and segmentation of patients who we’ve never been able to see before and you’ll see more of that emerging in the industry and you’ll see more of that emerging from Health Catalyst specifically. Let’s see, Varmi Marla asks, you started this talk discussing data sharing between companies to deliver consistent and quality care. Did I do that? Did I started by talking about data sharing? I’m kind of sleepy. You maybe have an idea. What are some incentives that can be provided to providers to share data, particularly those not in academic centers?

Dale Sanders:                I recently read that this is happening in the automotive companies, especially with autonomous driving tech. This will probably, partly because the auto companies are now competing with tech companies for market share for cars. What are some thoughts to deliver similar incentive to accomplish data sharing in healthcare? Well, I think, my short answer to this complex topic Varmi is that the patient has to be at the center of the data sharing ecosystem. Right now, there is literally no incentive whatsoever for healthcare providers, insurance companies, et cetera to share data with one another. In fact there’s disincentives economically.

Dale Sanders:                What I hope is that overtime is the gravity of data, the center of gravity for data has to shift to the patient and as it shifts to the patient, then … my computer is about ready to restart here. Then, as we shift to that gravity to the patient then I think it’s going to be up to them to decide how to share it rather than the perverted way that we have right now which is health care systems deciding what should be shared in insurance companies and things.

Sarah Stokes:                We are nine minutes past, just want to call that out so you’re welcome to go on.

Dale Sanders:                Let me just see here. Andy Canter. Andy, I’m going to … I got to get to you here, I knew you were going to put me on the spot I’m sure so I’ll give you a chance to make me squirm here. From Andy Canter. Perhaps, the question about the form of the data. Do we really argue about the quality or the content of a data element or about what data element means. Well, that too, right, friend. The ability to interpret the data is as important as just recording it. We can map the genome but what does that actually mean in terms of the patient? Actually between genes, proteins, microbiome, behavior, everything, can we capture data the way that allows for capturing all those interactions to be able to transform the data based on context, et cetera.

Dale Sanders:                Our analytics aren’t going to make sense without it. Yeah, I agree, friend. I mean, I kind of come to that from two different ways. I’ve never met data that I couldn’t eventually make sense of. A lot of times, going back to the military, you would bay the payloads in sensors for example and you didn’t always know exactly what kind of insights or what you might do with the data but boy, you sure were glad you had it when you did and then your ability to exploit and understand the data tended to evolve after your realization of its collection. I would not say that we have to have a firm understanding of how we might interpret the data before we engage in the collection of data.

Dale Sanders:                I think sometimes the understanding follows the collection but yeah, always insightful there friend, thank you. What do you think, maybe until quarter past, three more.

Sarah Stokes:                Yeah. That’s fine.

Dale Sanders:                okay. I can stay for whatever. How many people do we still have on?

Sarah Stokes:                173.

Dale Sanders:                173, okay. Working at an IDN, my experience has shown that digitizing is only one step towards this revolution, data sharing. Yeah, yeah. We get back to it don’t we here Bill. Bill Hickman asking. Data sharing, even between silos in an IDN is lacking currently. I agree with that. How can we move that aspect forward to help analytics advance care, especially in light of the CMS future push towards population health management. Well, there is absolutely nothing within the scope of data sharing in an IDN that prevents the CEO of that IDN to force data sharing if it’s not happening and it’s a shame that it doesn’t.

Dale Sanders:                I worked on a giant very exciting, very interesting project in Alberta a few years ago. It’s still kind of going on and it was fundamentally about creating a data resource for Alberta as a province to inform precision health and population health and we brought together … I think we had nine focus groups, 85 some participants on those focus groups and it was all about data sharing and crossing those human barriers, again, back to the soft skills and most of the time that I spent in that endeavor was motivating people, getting people to believe in a purpose greater than themselves, getting them to believe that sharing data is as important as sharing organs and blood.

Dale Sanders:                Here’s the interesting thing too. We had a large constituency of patients informing us. They were one of the most prominent. In fact, they co-led the project, our patient advocacy group and they were stunned, the patients on that panel were stunned that we in healthcare in the US and in Canada were not sharing their data more effectively for the betterment of their healthcare and their lower cost. They were stunned to find out that we weren’t already doing this. All of this hesitancy that we have around privacy and all that kind of thing is to me, it doesn’t hold up because patients expect that we should be leveraging this data to their benefit already.

Dale Sanders:                Of course protecting is privacy which we can do so just keep pushing Bill. It should go all the way up to your CEO. Let’s see here, there’s a question from … As chief clinical officer at a behavioral health system, I’m trying to mobilize a culture to use clinical decision support. My strategy is to use a small number of metrics and use the data to tell a relevant story to our clinicians about what they do and the results of their efforts. Any thoughts on how to nudge a culture that is self-proclaimed data averse? Well, I would have stay … first of all friend, behavioral health going back to my progression of mathematical maturity, it’s over on the far left, if you’re really kind of talking about psychology and psychiatry.

Dale Sanders:                Decision support in your environment is probably the most complicated that we have in healthcare so God bless you for taking it on. The way that I approach decision support at the clinical level, embedded in the EHR is I actually look at it in three loops of decision support and the analytics associated with that. It’s … you want to make decisions, clinical decisions at the population level, hundreds of thousands of patients affected by the decisions and the analytics at that level. Community health, population health is kind of the theme. The next level down and the next loop is the protocol level. Now, you’re segmenting the population and you’re saying in general, we want to treat patients of this segmentation this way, generally speaking.

Dale Sanders:                That’s kind of where evidence based care comes in, protocols, and that kind of thing. Populations at the top, protocols in the middle. You’re dealing with tens of thousands of patients, maybe in that group. Your mean time to improvement there is measured in months and maybe hopefully even weeks if you can turn your protocols around fast enough. Then, the last is the area that you’re working in which is the hardest which is clinical decision support. That’s at the patient level so it’s populations, protocols and patients and that’s where controlled variability and variation from the protocol is justified in the pursuit of personalized precision care for the patient.

Dale Sanders:                My experience has been, if you engage with clinicians around those three close loops using the same data to inform those loops, different types of analytics obviously, different types of algorithms, that’s one way to engage clinicians in a healthy discussion. The other is I invoke a study that Kim Komoto at University of Utah and Intermountain LDS published in BMJ a number of years ago that showed if we give clinicians the substantiating data to change their decisions at the point of care, versus a conference room somewhere, they’re 15 times more likely to change that decision if you give it to them in their workflow.

Dale Sanders:                They’re 15 times more likely to follow the best practices protocol if you share that decision support with them in their workflow. That’s the good news if you can do it. The bad news is quite frankly, is electronic health records and their software APIs make clinical decision support very complicated to implement. Most of those user interfaces were not designed in JavaScript, in HTML5, in all the tool and R and Python, right? That’s not native to the HR software architectures and quite frankly, that software architecture of EHR holds back our ability to inject clinical decision support. You’ve got a big challenge.

Dale Sanders:                I don’t know if I’ve helped you there but let me know if I can help any further, reach out to me on LinkedIn maybe. Okay. I better take one more question here and this one is a good one. You have … they’re all good. You have not referenced AI. Does that impact analytics? Yeah, for sure, friend. I embed AI within the broader context of analytics so it’s a form … it’s a subspecialty of analytics. It has its own unique complications but I actually believe that AI has the ability to engage better behavior … I’ll say behavioral change more effectively than what I would call the rules based declarative programming approach that we have traditionally followed in analytics.

Dale Sanders:                Moving from SQL to sort of the fuzzy world of AI, the pattern recognition world of AI is actually more in lined with human cognition than the left brain rules based approach that we followed in health care analytics up to this point. AI has a potential to push us further along with behavioral change. Even though there’s a lot of debate right now about the validity of the models and the training and the testability, we’re going to get past all that and eventually, the output of those AI models is going to resonate with the human brain better than the rules based approach that we followed so far. Okay. Well, I don’t think I can talk anymore. There’s one more.

Sarah Stokes:                You did have a note on a friend on one.

Dale Sanders:                Atul. Well, thank you Atul. Good to hear your … I appreciate you joining friend. It would be really boring if no one joined or ask me questions so thank you. The audience makes the show not me.