Kaveh Safavi 00:06
I think if you look at the historical trend for healthcare costs, and I mean, I’m a big fan of the work that William Baumol wrote about years ago. And this whole idea of the so called cost disease, which is if you look at the relationship between healthcare costs and GDP in every country, since we’ve been tracking in the 1960s, costs grow in every country, one to 2% faster than GDP, and it doesn’t matter how they’re organized, or how much they can, high spend low span doesn’t matter. Public Private, doesn’t matter growth in health care costs, the NHS is the same as it is in the US, even though they only pay half of what we spend spent. Okay. Why is that? Well, because the cost of human labor, which is the way we all do health care grows at the same rate of the economy of those people that they’re in and then scientific innovation, and an aging population give you the 1% a 2% flux.
Gregg Masters 00:52
PopHealth Week is brought to you by Health Innovation Media. Health Innovation Media brings your brand narrative alive via original or value added digitally curated content for omni channel distribution and engagement. Connect with us at www.pop upstudio.productions. And welcome everyone. I’m Gregg Masters, Managing Director of Health Innovation Media and the producer co-host of PopHealth Week. Joining me in the virtual studios my partner, colleague and lead co hosts Fred Goldstein, President of Accountable Health LLC. On today’s show our guest making an encore appearences Kaveh Safavi, MD JD, senior Managing Director at Accenture, where he is responsible for leading developing and driving a growth strategy that differentiates Accenture’s offerings for providers, health insurers and public and private health systems across the globe. A seasoned executive Dr. Safavi brings more than three decades of leadership experience to Accenture health. Prior to joining Accenture in 2011, Dr. Safavi led Cisco’s global healthcare practice. Before that he was Chief Medical Officer at Thomson Reuters health business, vice president of medical affairs at United Healthcare and held leadership roles at HealtSpring and Humana. So Fred, with that very brief introduction over to you help us catch up with Dr. Safavi. And his work at Accenture health.
Fred Goldstein 02:26
Thanks so much, Greg and Kaveh. Welcome to pop Health Week. Hey, great to be with you again. It’s fantastic. It’s been too long Kaveh. I think the last time we spoke was probably at HIMSS, I believe, a couple years ago. And it’s always fascinating because you come up with these incredible reports coming out about what’s happening in digital health technology. And you’ve got this latest one out the Accenture Digital Health Technology, vision 2021. What’s it look like?
Kaveh Safavi 02:48
Well,we do a tech vision every year as a company across all industries. And then we think about it in the context of healthcare. And so what you’re seeing this year represents a continuous evolution of our thinking. There are five broad themes, five broad themes across all industries. And we could talk a little bit about what they mean context of healthcare, but I think they are particularly interesting. And I think they really do represent the dramatic impact of digital health and digital technologies are having on how healthcare is going to be experienced by everybody. So I’m happy to take it at whatever level of detail you’d like.
Fred Goldstein 03:22
Yeah, so let’s start in with a little bit. I noticed that obviously COVID hit us last year. So maybe it’s an overarching principle. How did you see that impact this?
Kaveh Safavi 03:32
Well, COVID, I think serve as a catalyst as opposed to introducing a new problem to the equation, all society moving healthcare was on a digital journey. And we’ve laid out that step in all our previous conversations. I think that the nature of COVID did a couple of things to businesses in the first immediate short run, it was a shock to the system that forced businesses to answer the question, am I resilient enough to withstand a dramatic change to my business? And in many cases, the answer was no, or not quite. And what that caused was a recognition that these businesses have to have a much more fundamental digital chassis in order to have the resilience to respond to to what happened when COVID caused society to shut down. And I think in particular, in healthcare, it demonstrated because of the constant opening, closing, opening closing phenomenon, that the nature of resilience and elasticity, which is interesting, because we’ve been talking about the criticality of cloud computing as an example. And one of the reasons for the criticality of cloud computing is it gives businesses elasticity and it was always in the context of opportunity and competitiveness, never in the context of survival. COVID said, Forget that now it’s in the context of survival. So that was thing one. The other thing that COVID did was an introduced a new case for distance in healthcare. It used to be that we thought about distance primarily using the language of preferred experience and access, but we actually thought about it in the context of contagion. There are situations where it is simply not right for people to be in the room with each other. And so our ability to provide services and separate caregivers from patients or dangerous situations is now going to be built into our healthcare system. That kind of platform by its nature has to be a digital platform. So COVID introduced two new forces that accelerate what had always been the path of healthcare technology.
Fred Goldstein 05:31
And I guess we probably should also point out this is this report is built on the expertise within Accenture, and then a survey you do have people throughout the healthcare system
Kaveh Safavi 05:41
Yeah, well, we actually do. So this is a global survey of all industries. So 6000 leaders across all industries and all geographies, for health, specifically, there are actually 400 leaders that are surveyed as covering 12 geographies. But the opinions represent a synthesized view that includes outside and inside points of view and expertise.
Fred Goldstein 06:02
And you talked about, as you mentioned earlier, five areas, which were this stack strategically, mirrored world, I technologist. anywhere everywhere and for me to weso starting with stack strategically, what were the areas that were specifically gotten into with that?
Kaveh Safavi 06:17
Well, the essence of stack stack strategically is the recognition and something like 98% of the respondents said that you can’t separate the business strategy from the technology strategy is that they are so there is such a deeply intertwined relationship between them that now your technology is your business strategy. And actually, what was even more interesting was the statistic that 73% of executive stuff that their technology architecture is critical to the overall success of their organization, meaning that if you don’t get it right, you can be harmed, which is a little different than using it as a competitive advantage. And so what we’re seeing now is people are starting to think about their technology investment as a strategic asset. And they’re making choices about what technology they want to use, not just as a matter of hygiene, but actually as a matter of the ability to execute their strategy. And you in a very simplistic way we saw these kinds of things played themselves out during COVID, when people had to immediately move their workforce home, or immediately provide virtual visits, if they had not made the right choices from a technology perspective, they couldn’t do that the way they needed to do. And so and it’s not a one technology problem. It’s a it’s a set of technology. So we think about this idea of making technology choices that stack on top of each other and give you the capabilities to change your business in a dramatically different way. So it’s a strategic asset.
Fred Goldstein 07:35
Yeah. And I believe that the last time we spoke, as I remember interviewing you, I believe you pointed out the fact that we spent a fortune on technology and healthcare over the last decades since the Affordable Care Act, obviously, implementing EMRs. And you talked about this loss of efficiency, as I recall, maybe it was 15%, associated with the implementation of that technology, is, are we going to get it better this time? Do you see the thinking changing perhaps as they build that stack out?
Kaveh Safavi 08:01
For sure. Because the primary driver of that loss of productivity that I described, is the act of putting data into healthcare systems and typing, that’s the primary driver of the inefficiency. And clearly the technologies available to listen to a conversation, understand that concept. And then convert that concept into structured terminology that a machine can use for future things, decision support, workflow, etc, is right now in not just in R&D, but it’s actually being moved into examples where health systems are testing, validating, refining workflow. So and all of the help not only do all that electronic health record, companies have investments in these areas, but the communication and collaboration companies also have these kinds of these technologies in flight. And the net effect of that is going to be particularly beneficial. However, that’s a single type of technology that sits on a group of technologies. And so what you’ll quickly see is the interdependency between that, and other things like your basic communication and workflow, for example. And that’s, I think, where we’re really seeing this idea of ecosystems as opposed to these individual little point solution showing,
Fred Goldstein 09:13
right, so this is the overlaying I guess, of a natural language processing system on top of a broader set of systems
Kaveh Safavi 09:20
where the particular capability I’m referring to is, is a combination of things that have been around like speech and language recognition that are amplified through artificial intelligence to improve the fidelity and the ability to understand and get the right understanding of the word so that you can actually create the right terminology that’s structured. And historically, the only way to do that was the human brain, the doctor decided how they were going to actually code a concept. And now we’re allowing the machine to learn what the doctors thinking get used to the way they talk and understand that when I use these words in this fashion, this is the term that I need to drop in and that term can then be used for other purposes.
Fred Goldstein 09:59
Got it so the other you talked about was this mirror of mirrored world? And what do you mean by that, or what is essentially mean by mirror world?
Kaveh Safavi 10:08
So mirrored world is a recognition of the fact that increasingly, there’s so much data about what we do that people are creating simulated environments, and in a simulated environment, they can test and learn. And, for example, this is already happening in scientific discovery, where people are simulating in silica phenomenon for drug discovery as an example, and even devices. But increasingly, people are moving toward using a mirrored world, if you will, or replicated environment to do things like test experiences, or test workloads. So for example, there are several pediatric hospitals now that are basically creating the opportunity for children, we’re going to have a procedure endoscopy, a surgery to through the use of things like virtual reality and other things, understand what it will feel like to be admitted to the hospital to go to the operating room, the recovery room, all of that is designed to give them a simulated experience to prepare them. There’s also people looking at this from an operating perspective, I want to change your workflow, I want to change something about how this team is working, I can go in and simulate the environment, make the change, and then adapt the other pieces that go around it. What’s critical, though, is this is not a software problem, it’s a data problem, a mirrored world isn’t about software its about the data, the data is what are the pixels that create the picture. And because everything is putting data off, we’re now able to bring them together and create this kind of a picture. And, you know, in manufacturing plants and unsafe industries, people have already been doing this. But it requires us to think about more than just clinical data in a record. If you just let’s just go to a hospital environment, we do a mirror to do this, what you’re really doing is taking the stuff that’s coming off of all the devices that are emitting information today that’s basically being dumped. So that you can you can actually use it to replicate and create an environmental context. And then you can learn or test something in there as an example.
Fred Goldstein 11:59
So that brings up an interesting question that we’ve seen now coming up, both from the data acquisition side, as well as from the algorithm side, which is bet, you know, bias in the data or bias in the algorithm. And, and you’ve talked some about AI, how is that being dealt with in do you see that ultimately getting solved in an effective manner?
Kaveh Safavi 12:19
Well, the bias is generally there are two reasons for bias one is an incomplete sample. And so the more complete the sample is, the better that is, the second kind of bias is more of a reflection of the lived experiences that we have today are simply biased, and therefore the computer will perfectly perpetuate that that’s a little different, right. And that requires more of an active manipulation of the information to solve for that problem, the published data on the fact that patient care had been representing, let’s say, racial disparities. So doctors were making different recommendations based on the race of the patient, that wasn’t the standard of care, all the data in the world will just perpetuate that you have to choose to go in and manipulate that and make a difference. So I think that we’re, you know, the completeness, maybe we can solve for that. And we will the intrinsic bias and the way information is created, that’s a different kind of problem that has to be solved. I think one of the things that’s happening, though in health is recognizing that historically, the universe of data that we were focused on was clinical. And increasingly, we’re recognizing that that’s not the case, for example, people who are working on understanding patient well being at home by some versions of sensing or ambient sensing, they don’t really need clinical data, what they really need is environmental data like about the patient and their current where they were in the room they’ve been in what position, all those kinds of things are useful. The record isn’t the point. And when we know, we know how this is true for all of us. So my point being that the universe of data that we care about is much bigger than just diagnosis, treatment, CPT, ICD Nine, Snomed, it’s a lot more than that.
Fred Goldstein 13:52
And so, you know, as you as you brought that up, I was thinking myself, okay, we have all this clinical data set, we’re now bringing in the social determinants of health data set, and you’re now talking about essentially a third data set of real experience with
Kaveh Safavi 14:06
environmental context.
Fred Goldstein 14:07
Yeah, exactly.
Kaveh Safavi 14:08
And then there’s actually another data set. Right. And that’s preference,
Fred Goldstein 14:11
right,
Kaveh Safavi 14:11
sitting in the back of all of our minds is what our wishes are, which isn’t any of those data sets, but is captured. So retailers have been very much focused on trying to infer preference through behavior, right, and then making that part of it. Well, that’s just as true for healthcare. And so now there are people trying to figure out how one can infer preference, whether it’s patient preference or clinician preference through a set of other activities. So we can’t forget that.
Fred Goldstein 14:35
Absolutely. I’ve always wanted to have that parrot on my shoulder that talked to me exactly the way I to get me to do something. Now, once they understand me well enough, my parrots going to be different than your parrot I assume. So you also talked about I the technologists in this area of now sort of getting this data out to other individuals to be able to manipulate and use it in a sense.
Kaveh Safavi 14:55
So this concept is really we call this democratization of technology and it’s fundamentally based on things like no code, low code technology that’s coming out or the ubiquitous presence of natural language processing. So you don’t have to be a technologist to really manipulate technology to help you. And, you know, we started to the evolution of democratization started a few years ago with analytics, where we started to do things like well, let’s make decision support available to the business leaders. So instead of asking your decision support people to run a report, I could go in and configure report. And you know, we see tools that are out there that you know, CRM and sales management world has been built like that, where the leaders play around with the data. But increasingly, some of these other technologies are no code technologies are low code are allowing business leaders to effectively change the technology that they use for their workflow through talking to the machine. But they do have to understand technology. So what we’re saying is that we’re not trying to make everybody a software programmer, but it’s because software is your co worker, in many cases, your ability to interact with software is becoming so easy, and you actually have to have that capability. So I’ll give you an example. We call this thing technology quotient. Of course, our company spends its time helping people figure out how to use technology to solve problems. And we have, we actually insist that all 500,000 of our employees certify at the end to actually take an exam in seven different technology domains, meaning more than just the superficial knowledge. And we The reason for that is, we believe technology, literacy is too critical to just be the domain of a few people now that that plays itself out in different ways. But my point simply being that in like in the manufacturing world, as an example, there’s already a term called a cobots, co worker robot. And what we also know is for a human, to interact with a technology robot co worker, there are a whole set of skills of interacting with that machine that are different than the normal skills of you up to humans interacting, that they have to learn all of this is about this whole idea that the ubiquitous nature of technology, and recognizing it as our coworker has a set of literacy and a set of capabilities that go with it. And that’s how we’re going to have to think about the world.
Fred Goldstein 17:04
So you, as an organization have done this, I can’t imagine the hours and cost of implementing that type of a training program out to that your group, how is a hospital going to look at that and say, or a health care system and say, well, we’re going to work that into our everything else, obviously, to create this idea that many people talk about of data liquidity, getting it out, right, one way they can use it.
Gregg Masters 17:27
And if you’re just tuning in to PopHealth Week, our guest is Kaveh Safavi, MD, JD, Senior Managing Director at Accenture, a multinational company that provides consulting and processing services, for more information, go to www.accenture.com and do follow their work on Twitter, @DrKavehSafavi and @AccentureHealth, respectively,
Kaveh Safavi 17:56
what’s gonna happen when they can’t see how they can survive without it. I mean, there’s no business case until you unless you understand the fact that it’s fundamentally you can’t afford to be without it. And then all of a sudden, you change the metrics on what you do on thewhy statement, I think you’re starting to see this in parts of businesses increasingly, I mean, we are seeing this in a number of organizations now, where they’re saying that our digital capabilities need to be present in all departments. So instead of having a Digital Center of Excellence, where everyone goes there, everyone has to have a version of this in their operating because it just comes too fast. You know, you can’t just send a request in or go hire a developer, you’ve got to be able to figure out how to take no code. This actually happened during COVID. There were some there are some no code scheduling technologies out there for workforce, that that we’ve what we began to see were line staff that were essentially building the scheduling capabilities for them to deploy their own people. Because the capability existed, if they had to wait for software to be developed the pandemic, the vaccine would have been developed by the time they had happened, right. But the technology made it so that line workers could go there and say I need the logistical capability that helps you figure out where people go with this valuable demand. And I can act and I know it’s too hard for it’s too much for humans to compute but machines can do it. And because this technology is accessible, I know the business concepts, I can actually create the technology that I can then start to use right away. And I don’t have to go hire a third party to do this stuff for me. And that’s the kind of thing we’re beginning to see.
Fred Goldstein 19:27
So it’s becoming much easier for these groups to implement these types of services and systems.
Kaveh Safavi 19:32
Yeah
Fred Goldstein 19:33
In 2020, I noticed that AI was one of the top five in your your categories, and it’s dropped, has it dropped because it’s sort of become mainstream and it’s just now considered in there or is it dropped because it’s been a little too overhyped?
Kaveh Safavi 19:47
No, no, it’s the former and it’s not it’s not hyped its ubiquitous because the the issue is that as a transformative technology, that’s why you don’t see language like cloud anymore, or AI. These are fundamental capabilities. The question is how are you going to feel it? Right? So things like mirrored world is only possible because the software can learn and change itself without explicit programming, as an example, of the democratization presence of no code or low code technology is, in fact, because the technology is smart enough that it allows a non coder to manipulate the machine and to manipulate the technology in a way they can get a result thats useful. So it’s much more about the pervasive and ubiquitous nature of it than it is about the fact that it’s a discrete thing that is either there or not there.
Fred Goldstein 20:34
How much of this? You know, we’re seeing incredible impacts on operational efficiencies, things like that, you know, how much of this is going to be operational? versus clinical?
Kaveh Safavi 20:46
Yeah, well, I think that the truth of the matter is that it’s easier to see the gains in the non clinical areas first, for a variety of reasons. The first as you don’t have to do clinical validation to do that. So big issue, right. Also, in many cases, you don’t have to change everybody in the healthcare systems lives to do it. Many of these, many of these operating tests are owned by smaller groups, right? So it’s one thing to try to change the way every doctor practices medicine. So another thing to go out and change a function, whether it’s a call center, or a revenue or supply chain, or whatever it might be. So the bottom line of it is, is that I think we’re going to see the non clinical have a greater manifestation, clinical will certainly gain its benefits in different ways from this, there’s no question like the mirrored world is probably going to deliver a lot of results in clinical R&D and training, maybe even more than operational just because because in that case, you don’t want to experiment on human beings. So the ability to experiment and simulate without actually having to do something to a living thing has tremendous value. You just have to be able to replicate the environment. And that’s requires us to have data from the biological perspective. But I think, I think that’s how we see this thing play itself out is both and depending on the technology, more favors one or the other, just depends on which problem we want to solve.
Fred Goldstein 22:01
So the big questions come up for years and big argument is, and I’m asking you, obviously a physician, an expert in this area, is replacement versus additive for physicians, Are there areas where this technology may in fact, replace?
Kaveh Safavi 22:15
Well, so even the biggest issue for the jobs people do is fundamentally about the nature of the judgments that they make. And let’s call it the non routine tasks. So if you look at routine tasks, whether they’re physical or cognitive technology has already replaced not just in health, but outside of health, a large percentage of non routine tasks or routine tasks, most of the jobs that have been created in the last decade are what we would call non routine, whether they’re physical or cognitive AI, that, for example, starts to hive away at simple non routine tasks, but it leaves all of the other non routine tasks behind. So it’s really a capacity creation story, not an elimination story. Now, people have looked at healthcare, specifically, both clinical and non clinical, and asked the question about how much of the tasks that are performed, are routine enough that they’re within the reach of basic automation and artificial intelligence as we have it today. I think that the estimates that I’ve seen land in about 30%, for clinical kinds of jobs, and more like 40%, for non clinical, so we don’t see 100% in a lot of things. And I would argue that if 100% was the issue, they’ve already been eliminated by and replaced by technology. And so the truth is that this is much more about expanding capacity. And this is the thing to remember in healthcare is we have a shortage of labor globally, that’s getting bigger, the demand is growing faster than the supply, there is no amount of supply creation that we can ever do to close that gap. So frankly, if we don’t figure out how to extend our resources, with technology, we’re never going to solve this problem. Similarly, from an affordability perspective, we know that if adding a skilled clinician was the only way to solve a problem that creates an affordability issue, we’ve got to figure out lower ways of serving populations, particularly ones with limited resources, we have to figure out how to combine human and machine. So this is in my mind completely about tasks, and not at all about jobs.
Fred Goldstein 24:07
So it’s about efficiencies and things and and you raise the point of costs and dollars and efficiency. Does this finally get us to where we actually can see some sort of, I’m going to say bend in trend? Because getting people to take it negative seems to be impossible. Is that going to be feasible with this technology? Or is it just something we’re going to pay lip service to?
Kaveh Safavi 24:28
I think if you look at the historical relationship, trend for healthcare costs, and I mean, I’m a big fan of the work that William Baumol wrote about years ago. And this whole idea of the so called cost disease, which is if you look at the relationship between healthcare costs and GDP in every country, since we’ve been tracking in the 1960s, costs grow in every country, one to 2% faster than GDP and it doesn’t matter how they’re organized or how much they build high spend low spend, doesn’t matter. public, private doesn’t matter. growth in health cost the NHS is the same as it is in the US, even though they only pay half of what we spend spent. Okay? Why is that? Well, because the cost of human labor, which is the way we all do health care grows at the same rate of the economy of those people that they’re in. And then scientific innovation, and an aging population give you the 1% or 2% Plus, okay, we’re not going to stop scientific innovation, we’re not going to stop aging population. We’re the only industry that has not figured out how to substitute technology for human labor and gain productivity. And this has been documented repeatedly, healthcare has lost productivity in terms of services created relative to jobs created. And that is the fundamental problem that we have no hope of solving, if we don’t think about substituting technology for human labor. So can we make the curve one to one because we can take away a little of the cost of human capital? So we give ourselves some headroom? Yeah, maybe. I mean, but the truth of the matter is that the basic progression of science and the basic progression of demographics is also a driver of trends. So I think we’re obligated to try to solve the problems we can solve because as a society, we’re going to try to cure disease, and we’re going to try to treat illness and that’s going to come at a cost.
Fred Goldstein 26:09
Absolutely. We want that progress, obviously, in technology, and science, etc. So we can have better lives for everybody moving forward.
Kaveh Safavi 26:17
And I’ll just amplify that Ed with a Fred, excuse me with another comment, which is that the reality is that I’m not sure people are worried about how much money we spend on health care as much as are they getting their money’s worth? We focus on this as if people wake up every day saying, you know what, I want to make sure we spend less, I think the bigger issue is, and you see this in every society and every poll, nobody’s really trying to shut their healthcare system down. They’re just angry that they’re not getting value for their health care system. Right.
Fred Goldstein 26:41
Right.
Kaveh Safavi 26:41
That’s a very different conversation than just spend less create austerity. We don’t see that.
Fred Goldstein 26:46
Right.
Kaveh Safavi 26:47
And I think I think this is part of the the issue here is let’s people want their money’s worth. And that’s the real demand. And that’s where some of these things kind of play themselves out is to help people get their money.
Fred Goldstein 26:57
And I know Accenture will be at HIMSS coming up,
Kaveh Safavi 27:00
yes,
Fred Goldstein 27:00
anything to plan for it, your booth things going on people to see
Kaveh Safavi 27:03
do have uh So we do have our new consumer survey research coming out, which has some interesting findings out I’ve seen it. And we also have some interesting research that we’re doing around physicians and physician attitudes that I’m looking forward to share again .
Fred Goldstein 27:16
Well, fantastic. I’m looking forward to stop by and seeing you and the others again, as COVID sort of goes post, not quite yet. But fantastic to have you again. Kaveh. It’s always a pleasure.
Kaveh Safavi 27:26
See You in August, Fred,
Fred Goldstein 27:27
we sure will, back to you Gregg,
Gregg Masters 27:29
and thank you, Fred. That is the last word for today’s broadcast. I want to thank Kaveh Safavi, MD JD Senior Managing Director at Accenture for his time and insights today. For more information on Accenture, go to www.accenture.com or follow them on Twitter via @DrKavehSafavi and @AccentureHealth respectively. And finally, if you’re enjoying our work here at PopHealth Week, please subscribe to our channel on the podcast platform of your choice and do follow us on twitter via @PopHealth Week. Bye now.