In this fireside chat at the 2019 World Medical Innovation Forum (WMIF), Sree Chaguturu, MD, Chief Population Health Officer at Partners HealthCare, sits down with Seema Verma, Administrator of the Centers for Medicare & Medicaid Services (CMS). Their discussion delves into the agency's initiatives to harness innovation in healthcare through technology, focusing on efforts to promote interoperability, empower patients with data, and leverage artificial intelligence to improve healthcare outcomes while reducing costs.
Thank you so much, Administrator Verma for the opportunity to talk to you about your thoughts on artificial intelligence, data and technology. So before we get into the specifics around artificial intelligence, I wanted to ask you a little bit about your thoughts on technology and healthcare at large. So CMS has helped spur significant technology investments, and now we need to see the return on investment on technology. You recently said that the American healthcare industry must be powered by technology, not burdened by it. What is CMS doing to unleash innovation for new technologies?
Verma: Well, it's great to be here today. If I sort of give you some context in terms of how we're looking at technology and innovation at CMS, one of the things that we think about every single day is the cost of healthcare. Our actuaries are telling us that by 2020 we'll be spending almost 20% of every dollar spent in the United States on healthcare, which is simply unsustainable. And the solution is not to ration care, but to think about ways that we can deliver care in a more efficient way that is still providing high value, good quality, but just at a lower cost. So when we think about technology and innovation, our thought is we want to lean into innovation and technology and use that to help find ways to make sure that our healthcare is more efficient.
And so one of the things that really troubles me in terms of how we go forward with technology and innovation and harness that to solve some of our most difficult problems, there's been a lot of discussion about government run healthcare around Medicare for all, and I think that those proposals are the greatest threat to innovation in our country. In healthcare. The reality is government has often been a barrier to innovation. And so one of the things that we're working on at CMS is how do we reverse course and how do we make sure that we are not being a barrier to innovation? And we're looking at that across many different spheres. So the first one that we've been working on is our initiative around patients over paperwork, which is just the regulatory burden that CMS creates with our 11,000 pages of regs that we put out every year.
And we know that when providers are busy complying with regulations, which often are needed, but a lot of times they're excessive and they're not really adding to improving patient outcomes. So we've been focusing on that. The other area is trying to figure out innovation in terms of how we pay for healthcare and moving away from a for service system to paying providers based on value on the quality that they produce and doing that in a lower cost way. And we've been thinking about how regulations get in the way of value-based care, which is why we've been focused on Stark and addressing some of the barriers in Stark. The other area is just in terms of innovation, in terms of devices and new treatments. And the reality is that the government has been slow to respond to new technology. One of the frustrations I have is that we'll get a lot of letters from members of Congress and our beneficiaries saying, this service or this particular device is paid for in the private industry.
And when I come to Medicare, I don't have access to it. And that's because of the government policies. It's things that are set up in the law that don't allow us to respond quickly to new innovation and technology. We hear a lot from innovators that they can get their devices or a drug approved by FDA and they can figure out that process, but then there's sort of the abyss between FDA approval and getting CMS to code for the new treatment or device and getting coverage for it as well. So one of the things that we're trying to work on is how can we expedite that? How can we make sure that when breakthrough technologies are approved by FDA, that we can make sure that our Medicare beneficiaries have access to the latest treatment? One area that we have is something called new technology add-on payment.
And it's essentially when new devices or new treatments come out, we actually do pay an add-on, but we require innovators to show what we call substantial clinical improvement. But the problem is if people aren't using your particular product, then it's very hard to generate that data. So we're thinking about how we use the tools that we have within the regulatory framework to try to ease that process to make sure that we are supporting innovation and encouraging investment in the types of technology that is going to help our healthcare system be more efficient in the future.
Thank you. It's very helpful to hear about your framework of how the government can help unleash innovation. And so the topic of our conversation for the remainder of today is around artificial intelligence. And so moving on to artificial intelligence, the CMS has been communicating its time for disruption in healthcare and CMS wants to lead it. And you had a very exciting announcement a few weeks ago and I was wondering if you could tell us a little bit more about the CMS Artificial Intelligence Health Outcomes Challenge?
Verma: Well, we're very excited about this. We partnered with the Arnold Foundation and the American Family Practice Physicians. Essentially, if you look at CMS, we are the nation's largest insurer between Medicare and Medicaid and the exchanges we're serving over 130 million patients, and we have a very large data set. And so we've been making efforts to make that data more available. But the idea here with artificial intelligence is could we invite innovators to use our data to provide the type of analytics and the tools so that our providers could identify potential high cost situations? So in particular, we're focusing on inpatient admissions and we're also focusing on admissions to skilled nursing facilities that are unplanned. So is there something in the data that would help providers identify patients that are at risk for being admitted? And that's one of our high cost areas. So we're starting with that. So not only to be able to do the predictive analytics, but then to also provide the tools to providers so that they can use that. We think that'll be something that we could potentially make available to all providers, but particularly those providers that are involved in value-based payment arrangements where they're at risk. So they're more interested in looking for those types of tools that can help them prevent potential costly healthcare expenditures.
It's a very exciting initiative and we're all excited about the future opportunities that are going to be unleashed through the challenge. So there was a recent speech that Bill Gates gave around artificial intelligence and he likened artificial intelligence to nuclear energy. There's both a good side and a bad side, the good side being nuclear power. And then one of the other consequences is nuclear weapons. If we think about artificial intelligence in that way, the duality of it, what are the steps that CMS is taking to ensure that we maximize the benefit of artificial intelligence to patients in the healthcare system at large? But on the flip side, to prevent those unintended consequences, what is CMS thinking about doing to prevent or protect patient privacy and security?
Verma: So the first thing that we're doing to support AI efforts is through what we call our My Healthy Data Initiative. And really this is about unleashing data. So last year we announced that for the very first time we were making data available for our Medicare Advantage patients. And so that's never been done before. That's providing data on roughly about 30, 30 some percent of our Medicare population making their claims data available for research. This year we will also make our Medicaid data set available. And again, this is a population of pregnant moms children, our disabled population aging seniors. So it's a very unique population set. And so we're making all of that data available for researchers, and we think that there's a lot of potential to analyze that data. Again, that's something where artificial intelligence can help us. So we're using that in terms of pushing forward and trying to make more available.
And we've also had a lot of efforts around making sure patients have access to their data, making it very clear to the healthcare system that people own their data, it's their data, it's their information, and they should have access to it. So a lot of our rulemaking over the last year supported that effort. So making sure the data is available. One of the things that we did was an initiative called Our Blue Button 2.0 and blue button was essentially done a long time ago and it made claims data available to patients. But you can imagine that a patient having access to an Excel spreadsheet with a bunch of claims information really didn't mean much. So last year we put the claims data and made it available through an API format, which allows the app developers to take that data and essentially have use for it, making it more understandable for the patient.
And really within less than a year, we had almost over, I think it was over 1500 app developers that are active in the sandbox. And even now we have about 15 active apps, whether it's organizing medication lists or allowing you to donate your data to research. There's just all these possibilities that are out there. So I think we made a lot of moves and I think all of the data sets that we have across CMS around quality, around claims data, we're also making effort now to figure out how can we make all of that data available in an API format because it's already public data, but just posting it on our website is one thing, but how do we make it available for innovation? One of our concerns though, and I think it's foremost on our minds, is privacy and security. We're dealing with a very vulnerable population, whether it's our senior population in Medicare, our patients in Medicaid, and when we think about policy at CMS, one of the major goals and kind of our main theme is patients first.
So in everything that we do at CMS, we try to put ourselves in the shoes of the patient. And so when it comes to data and privacy and security, that's the view that we take. How would we want our data to be used? And so with our blue button project, when we were making that data available to app developers, we had very strict guidelines about security and privacy, making sure that they were communicating with our beneficiaries about how the data was going to be used and making sure that there was plain language. We looked at regulations in other countries, we looked at regulations from different companies that are already involved in this space. I think there's a lot more work to be done in this area. It's not something that we're going to do one and done. It's something that we're going to have to be constantly vigilant about. And one of the concerns is that even our HIPAA and our privacy protections, that only extends to healthcare providers. And so once the data starts traveling outside that area of where HIPAA regulates, we need to think about privacy and security outside of the healthcare ecosystem. So I think that's something that remains as a challenge,
A lot of excitement, but also a lot of work ahead of us. And so as we think about maximizing the potential of artificial intelligence, it hinges as you have already alluded to, and making healthcare data more interoperable and accessible to the various players in the healthcare ecosystem. And I was wondering if you could just tell us a little bit more about what CMS is doing in interoperability and making data available between organizations.
Verma: Sure. So we think about interoperability. One of the concerns that we've had is that after the federal government made an investment of some 36 billion, the reality is hospitals and doctors started using EHRs, which is a tremendous stride forward. But we went from paper filing cabinets to essentially electronic silos. And the patients, their data is locked in into electronic systems. And for any of us using the healthcare system, I've had my own experience, which I've been very vocal about. My husband had a cardiac event and he was at an airport and I was at a different airport, and the paramedics are calling me and saying, is there something in his healthcare background? He's not breathing, what do you know? And at that moment, I didn't have that information. And we spent a week in the hospital and when we left, I wanted to have all of the tests that they performed as sort of a baseline to give to his doctors back home.
And essentially what I got was five sheets of paper and a CD rom. And so my kids were like, what's a CD rom? Because our computers don't even take that anymore. So we're sort of outdated and we're behind. But I tell that story because it really speaks to the problem for patients is that your data is in many, many different places. And I like to think about a vision for the future of what interoperability means, the time that a baby is born to be able to collect all of the information, all of the interactions that they're having with the healthcare system to be able to combine claims data or information from wearable technology. Imagine what that will do, not only for the patient, and you combine that with artificial intelligence, it will lead to a greater understanding of their own situation. For providers to be able to have that complete data set, I think will lead to more personalized treatment.
We're not duplicating tests, and I think just collectively, it's going to build our collective understanding of the human body and healthcare. It's going to advance treatments and cures, and we're moving already into an era of personalized medicine. And I think that will advance that. So if I think about all the things that we are doing at CMS to lower healthcare costs, to improve quality, to have a better patient experience, I think this is one that really falls across every initiative we have, whether it's trying to foster more value-based care, more coordinated care, more seamless care. So we've taken a very hard stance on interoperability, making it clear that this data belongs to patients. Just last year, we essentially overhauled our entire meaningful use program, and now we call it promoting interoperability. And so you have to do more than just use an EHR. You have to demonstrate that you are providing patient data to patients that you were sharing that data with them.
And if you're a hospital and you're not doing that, we'll penalize you. And if you are a physician, your incentive payments are tied to that. So we did that last year along with Blue Button. And the other thing that we did just very recently is we put out a proposed rule, and this would essentially require the insurers that do business with us. So that's our Medicare Advantage plans, our Medicaid managed care plans, as well as all of the insurers that are offering on the exchange. And we said to them, look, we're doing this already in Blue Button 2.0, you should be doing the same. And so that's something that we've proposed and we're seeking comment on that. And that was done in coordination with the rules from the ONC around information blocking. So if these rules go through, this could provide access to claims data for over a hundred million people.
That's great, and thank you so much for sharing that story about your own personal experience and how it's very easy to talk about and it's theoretical, but to hear that story is very close to home. So you already alluded a little bit about this in your previous comments, but I wanted to get a little bit further into ownership. And so as you talked about interoperability to raise these questions about data ownership among providers, EHR vendors, patients, et cetera, how are you thinking about resolving these ownership issues as you push forward these multiple initiatives?
Verma: It's very clear to me that the patient owns the data. It's their information, and they should have complete control over that. I just met with some of the folks today from open notes where they have really been on a crusade to make sure that notes are available to patients and they're doing research and showing that when that's available, they're actually seeing positive outcomes in terms of lowering cost and increase increasing adherence to medications. So everything that we're going to do, all of our policies will have that theme that it belongs to the patients. I think the work that we've done around information blocking is also making that very clear to the community about what information blocking is and what it's not. And I think all of that really supports the efforts around interoperability.
That's great. And we've talked a lot about artificial intelligence and how it can improve providers and patients' lives. I was wondering if you could talk about internally looking at CMS and how CMS is thinking about using artificial intelligence to prove its own operations. Sure.
So one of the things that I do at CMS is I meet with our data analytics team with our actuaries on a quarterly basis, and we look at utilization and I will see some increases in a particular area. And I'll say to the team, and this has been going on for a year now, and I said, well, why are we spending more money? What happened? And I was like, is this the flu? Do we have more inpatient admissions? Was this the flu this last quarter? And it really started there where they said, no, we check that we actually didn't see an increase in admissions. And they said, what we're seeing is an increase in outpatient admissions. Well, that's normal because sort of where the industry is going. But the bottom line is we've gone through for several months now trying to identify why we're seeing some trends and we don't have a clear direction.
We are having difficulty understanding the data. So it's very clear to me that if we're going to be ensuring and protecting and sustaining the Medicare and the Medicaid programs, we need to understand what the drivers are of healthcare costs so that we can try to lower them. So that's one area that we're looking at. The other area is program integrity. So we do a lot of work around fraud and abuse, and unfortunately one of the things that we've done is we've really created a lot of burden for the healthcare system. We sort of create an entire system around the two or three bad apples, and we require providers that are, if we see any information, we'll go in there, we'll ask them for their medical records. And there's just a lot of work that goes on for those providers, and it really does create a lot of burden for them.
There's a lot of appeals, and so it really creates a long, burdensome process. And so we've been thinking about how we can use artificial intelligence to audit records and do it that way? And so that there's not a need for providers necessarily to take active steps. We can look at the electronic record and we can look at their billing and we can identify the bad actors that way without providers having to take a lot of active steps, I think it can enhance our ability to be more effective around fraud and abuse. So that's one area that we're focused on. The other area is around measurement. So as we have a lot of discussions about quality and value, the idea with values, we want to pay for care that is high quality care and low cost, but we've sort of gone overboard in how we define quality, and there are so many metrics.
And now anybody that comes into my office wanting to propose a new quality measure, I'm like, whoa, you have to show how it's not going to create a burden for the system. And so again, that's where I think that AI could play a role in this is that we can identify certain trends or certain measures that if a provider succeeded on that particular measure, I call super measures where we can look at their success on a particular measure around cost and quality together. And then if they succeed on that particular measure, it tells us more about the provider. Also, I think AI could help us identify what those measures are and also help extract that data so that it's less burdensome for providers to report that.
Well, it's exciting to hear about the way that you're thinking about artificial intelligence, improving program operations and potentially improving quality measurement, which is very exciting. So we're going to move a little bit towards questions from the audience, but before we move to that, I was wondering if you had any closing thoughts you wanted to leave the audience with?
Verma: I think that we're at sort of a turning point I think in healthcare that we're at an amazing opportunity or crossroads where we have all of this information that's now readily available and we have the tools to move forward. And I think what we need to do is get government out of the way and make sure that we are supporting innovators to do the things that they do best. We want to make sure that we continue to have the best healthcare system in the world. And if you look at other countries that are moving to more socialized healthcare, they don't have the type of innovation that we have. So I think we all have to sort of recognize the important role that the United States plays in bringing innovation to the entire world and recognize that it is the private healthcare system that we have in place that is actually spurring that innovation and be very careful and concerned about those initiatives that would try to put more government in control, and that would really thwart innovation moving forward.
That's great. And we might need somebody to come and get this app because the questions aren't showing up here. But while this is being fixed, I did want to ask you a question around Blue Button 2.0. It's a very exciting initiative, and you've had the opportunity to see a number of innovations coming out of Blue Button 2.0. Are there some that are the most exciting for you that you've seen? And are there big categories of innovation that you haven't seen as much investment in and would like to spur innovators to do more in those areas?
Verma: Sure. I think a couple of areas that I thought were particularly interesting were there were some apps that allow you to donate your data to research, and I think that's really important. We've also had some conversations with NIH about their sync for science program, and is there a way for people to donate data directly to the NIH? I thought that was important. There's also what I'd call just patient friendly tools, like there was one that would analyze the medications that you were using and give you recommendations about which drug plan a Part D plan that would be most cost effective for you based on the drugs that you were taking. But I think ultimately what we hope to get to is at CMS, we're working towards a vision of an empowered patient where they have quality data cost data, and they have their medical record at their fingertips. And so looking towards apps that can take all of their information from every single provider and aggregate it for them so that they can understand it better and allow them to give it to the next provider and translate that information so that it's usable to the provider, I think that's some of where it'll be very important to have that type of innovation. Yeah,
That's great. This has been a story that's happened. This is a question from the audience that has happened in many different industries. As we introduce a technology, it can exacerbate inequities. And so the question here is, does the digital divide pose a serious long-term problem in exacerbating health inequities? And I might add, is there anything that CMS can do to help mitigate against exacerbating those inequities?
Verma: Sure. And that's always something that we're very mindful of. I think that being said, it's not a reason not to move forward. We need to leverage technology. And I think even if we look at inequities in our healthcare system, or as we're moving to more of a digital healthcare system, one of the conversations that we've been having at CMS is focusing on rural health, that as we're looking at a lot of these tools, it requires broadband access and access to internet providers. And in some areas of the country, that's not a possibility. So that's something that the administration collectively has been focused on. How do we address rural areas and make sure that they have that connectivity? So again, not a reason to not move forward, but something that we have to be mindful of is how we make sure that we're providing that technology to every American.
In an ideal world, what would data privacy look like? Would everything be encrypted and anonymized, for example, or do you have a different vision for data
Verma: Privacy? I think for data privacy and security, it's going to mean different things in different places. I think the patient should always be in control of their data and the information and control where that goes, and they have to have that information when they're making decisions about where they're sharing their data, they need to understand who they're giving it to and what that entity is doing with their data. I think there's also just larger issues around artificial intelligence, cybersecurity about making information available. So I don't think that it's one or the other, but they have to come up together. Great.
How can CMS take a lead on getting patients to understand their healthcare and to really establish realistic expectations without appearing to ration care? You've really articulated the centrality that the patients have in controlling their data. And so what do you see as CM S's role in helping them understand what to do with that data?
Verma: I think it's one to say we describe an empowered patient. What does that really mean? It's not just having access to information, but it's information that is understandable and actionable. We have a lot of this conversation around price transparency and CMS is really pushed to have doctors provide their charge masters and make that more public. And that's just the beginning of our efforts on price transparency. But as we're having that conversation, it's making sure the information is usable. So having reams and reams of data, it doesn't make a lot of sense. And that was the problem that we had with Blue Button and the original design of that program. They just put out all this claims data doesn't mean anything. So I think that's really where we need innovation from the private sector using artificial intelligence to take that data to make it understandable and to make it actionable. If you think about being able to look at the data and say, remind people you need a preventative service or you might be at risk for these types of diseases, and making it available is only part of the story. We have to figure out how to make it actionable.
Great. And then there's a specific question that we have from the audience about how we hold EMR vendors accountable to opening up apps and interoperability? We talked about it at large across the healthcare industry, but how are we thinking about EMR vendors
Verma: Specifically? I think that's going to be a challenge. I had an opportunity to go out to Colorado a couple of weeks ago, and we met with all kinds of innovators that had new all types of technology. One was around helping doctors essentially do coding and documentation and trying to reduce the burden, and other ones that were helping with price transparency. And it was all very clear that they need to be able to create their innovation, and they have to do that in tandem with the EHR, and we have the way EHRs have been adopted in our country, there is not a whole lot of competition. There's only a few vendors and many providers have invested very heavily, and so it's not easy for them to change course. So that being said, I think it's a problem that we need to address because if we're going to move forward on innovation and turn the corner on trying to do something about our growing healthcare costs, that is part of the solution, is how can we bring innovation? And if the EMR vendors are blocking that and not allowing that to happen, then they're hurting our country and we need to address that. Right.
Comments