September 15, 2020: Welcome to News Day. We explore digital transformation, Wexner’s partnership with One Medical, Magellan Health’s strategic relationship with Lovango, GoodRx files going public and CMS approving the first reimbursement for AI augmented medical care. Also how do you get consumers hooked into your service experience? You need to be thinking mobile applications, agility, cloud, customer system of engagement, convenience and ease. Empower the people not empower the system, right? We also tackle the challenge of a security budget and why partnerships are the best choice if you want to move fast.
News Day – Digital Foundation, Ohio State and Magellan Choose Partners
Episode 303: Transcript – September 15, 2020
This transcription is provided by artificial intelligence. We believe in technology but understand that even the smartest robots can sometimes get speech recognition wrong.
[00:00:00] Bill Russell: [00:00:00] All right before we get started, three things, clip notes is live. It’s a great way for you to stay current and keep your team current. Sign up today with an email to [email protected] We want your feedback for our 300th episode. We have a special form out there where you can provide us with feedback on the show, which really is helpful for us to make the show better.
[00:00:22] And, at the bottom of that form, you have the opportunity to sign up, to receive a free mole skin, black notebook with our logo right there. [00:00:30] Just like this one, thisweekhealth.com/ 300 is where you go to get to that form. And finally, we are active on LinkedIn and Twitter. You can follow the show or you can follow me if you’ve been following me, that select a story every weekday and started discussion on LinkedIn.
[00:00:47]When you engage, you help us to frame the discussion on the show for our thousands of listeners and subscribers, joined the discussion today. We’re looking forward to it now onto the show.
[00:01:02] [00:01:00] Welcome to This Week in Health IT. It’s Tuesday news day, where we look at the news, which will impact health IT. Today, we are going to take a look at a digital transformation article that was written by Jeffrey Moore, the, author of crossing the chasm. And, it really sets a great, framework for discussion around digital transformation.
[00:01:20] We have Ohio state doing some, partnerships with one medical. We have, Lovango doing a partnership with Magellan partnerships everywhere. So looking forward to those conversations amongst other stories. [00:01:30] This episode. And every episode since we started the COVID-19 series has been sponsored by serious healthcare.
[00:01:35] Now we’re exiting the series and Sirius has stepped up to be a weekly sponsor of the show through the end of the year. Special thanks to Sirius for supporting the show’s efforts during the crisis. If you haven’t signed up for 3xDrex, yet you are missing out text Drex, D R E X two four eight four eight four eight.
[00:01:51] To receive three texts every week with stories that will help you to stay current. It helps me to prepare for the show, and this is a service of directs to Ford, a frequent [00:02:00] contributor of the show. All right, our first story, digital transformation behind the scenes, Jeffrey Moore author of crime in the chasm.
[00:02:07] If you’re not following this guy on LinkedIn, I highly recommended what he does is he gives us frameworks for how to think about things. And in this article, he talks about digital transformation. It’s not specific to healthcare. It’s more broadly, for many industries, but what he does is he starts with a graphic and he lays out the industrial product era on the left side and the digital services era on the right side.
[00:02:30] [00:02:30] So in the, to just give you an idea of some of the back and forth here, so industrial product era demand, exceeds supply digital services, era supply exceeds demand. Industrial focus on supply chain digital focus on the customer, industrial six Sigma quality programs, digital user experience, user design.
[00:02:49] If you will, a user centered design, industrial era, internet changes, everything, digital era cloud and mobile change. Everything. Industrial B2B [00:03:00] enables me to see digital B2C enables me to be. last one. I, although there’s others, last one system of record in the industrial era gives way to in the digital era systems of engagement.
[00:03:13] And then he goes on. There’s a couple more here. The first thing he notes is that, no matter what happens, we’re building off of the industrial area. We’re not getting rid of those things. Like you wouldn’t stop focusing on your six Sigma program or your lean program. You wouldn’t stop focusing on supply chain.
[00:03:29] You’re [00:03:30] just starting to invest a little more heavily in things that are going to be customer focused, things that are going to take the friction out of the interactions with the consumer. So even as we’re talking to bringing this practical down to healthcare, even as we’re talking to our EHR providers, they respond to us.
[00:03:48] We tell them this is what’s important to us, and they respond to it as we start to move away from the system of record. And really focusing in on the system of record and start to focus in on the [00:04:00] system of engagement. We should be informing our EHR providers, that we want more tools that look like X, Y, and Z, which form our basis for engaging the patient.
[00:04:11] Cause at the end of the day, until we figure out a way to engage people in between them times, they’re sitting in front of a doctor or they’re in our acute care facility, not I’m going to be able to impact the health of the communities that we serve. So that. That is one of the ways. And actually all these other industries, they’re trying to figure out how to [00:04:30] increase the number of touch points between the time you’re say, in a car dealership.
[00:04:33] And the next time you’re in a car dealership, they want you to become a Tesla person, or they want you to become a Mercedes person because you’re interacting with them often because they’re updated. they’re actually starting to create their, Their cars, even as ways of engagement, there’s more tools on that screen.
[00:04:52] And they’re trying to figure out how to get you hooked into their overall service experience. And that’s the same thing that’s available to [00:05:00] healthcare. so some of the things, he goes through a list of things and it’s not a complete list, but a partial list. Of, some things that are distinct to the digital era.
[00:05:10] And so he goes, the first thing is it’s born in the cloud, not in the data center, right? If you’re going to be in a system of engagement, if you’re going to be creating things that are gonna touch a lot of consumers, you need to be in a place that you believe has ubiquitous access. It’s born in the cloud, it’s born on the cloud.
[00:05:27]infrastructure it’s board on the cloud economics. So it’s [00:05:30] born in the cloud. Second thing, customer system of engagement, not standard systems of record. So we need to be thinking in those terms, third thing, focus on convenience and ease of use, not security and reliability. He goes on to say, not that you would not worry about security, reliability, but he’s making the point that we need to move fast.
[00:05:49] And sometimes when you move fast, you make mistakes on the reliability and security side. And, In deference to the need to move fast, because if you don’t, you might become [00:06:00] irrelevant in your markets. he makes that point. So focus on convenience of the, your user. he then goes on to talk about consumer apps, not ISB and SAS applications, right?
[00:06:10] These are applications that are built and designed around the consumer, engage the consumer. Find out what they’re looking to, how they’re looking to engage your system and then provide them those tools, incident response, not technical support. This is an interesting point that he makes. That, we’re not always gonna know what situation and [00:06:30] where as we start to go into this consumer facing, app kind of thing, we’re going to end up in a situation where we’re not going to be able to anticipate all the scenarios that are going to be coming at us.
[00:06:41] And so we end up with more of an incident response, which we’re all very used to now with COVID and the wildfires and other things. it’s that same kind of response model that we are moving towards instead of a tech support help desk kind of model. second to last empower the people not empower the system, right?
[00:06:59] So you’re [00:07:00] not, you’re not focused as much on how do we empower our health system. You’re focused on how do we empower our. you could start with the clinicians and the care providers, but also how are you going to empower the consumers and the patients that we serve? And then finally agile not ITIL.
[00:07:18] And I know a lot of people think that’s just a buzzword, that agile is just a buzzword, really look into it. I’m not going to have time in this show to really go into it, look into it. It is important. It’s not, and it’s not that I tell goes away, [00:07:30] but agile methodologies thinking, Thinking in a way that is about speed and movement and constant iteration.
[00:07:39] And, working code is the best. A metric for, are you going to deliver, I still am running into people that are waiting to see code at the end of a project. And I say, no, Oh no stop right there. You should be seeing code all along the way. In fact, you should be seeing code within the first two weeks that you kick off any [00:08:00] software development projects, even if it’s just the login, a working login screen.
[00:08:05] So you can look at it. And you can respond to it and say, no, that’s not going to work for us. We need dual factor authentication, or we need to be integrated if you’re looking at code it’s the best way. Anyway, agile is not only in that way in software development way, but also in response to system outages and those kinds of things.
[00:08:22] Agile is just a better method than a lot of the things we have traditionally done with waterfall. So [00:08:30] he goes on to say, Hey, This is just a framework for conversation. It’s just what I think. What do you think? And I’m going to put that same thing out to you. I’m going to give you my, so what, but what I want to hear from you is what do you think I’ll probably post this article out on LinkedIn and start a discussion around it, but.
[00:08:47] But, generally my, so what on this is systems I’m engaged. So systems of record we’re used to those, our EHR is fundamentally a system of record. Now we’re moving to systems of engagement. this is going to be built on API APIs and [00:09:00] mobile applications, and some web applications, but more and more, it’s going to be mobile applications.
[00:09:04] That’s how we need to be thinking. But that’s going to give ways to systems of intelligence and for systems of intelligence. What we’re going to need is clean data that we can train computers, that we can train systems, and we need to be start thinking about that clean data. Do we have clean enough data to train the, The computers, quite frankly, the technology to start to recognize patterns and those kinds of things.
[00:09:27] So that’s the first story. That’s my, so what I’d love to hear [00:09:30] your, so what on this, it’s a Jeffrey Morris, the author it’s out on LinkedIn, digital transformation behind the scenes. Loved it. I loved it. Hope you, hope you like it as well. Let’s see. Next article. All right. As you can hear, I have a lot of construction going on behind me still.
[00:09:49] So you’re going to keep hearing that our next article is Ohio state. Wexner has partnered with one medical and I found this. It will be interesting. I think this is a, this is one of those [00:10:00] things that we are going to see over and over again in this digital era. And that is partnerships in, in pursuit of moving faster.
[00:10:09]It’s not that Ohio state doesn’t have the resources to build out the capabilities that one medical brings into their market. It would just take them multiple years to do it right. So they partner with one medical. If you’re not familiar with one medical, they have a, let’s just call it a clinic platform.
[00:10:25]So they can roll out clinics that are high that have a, very, [00:10:30] digital, a frictionless, experience. For the consumer sort of, you make the appointment, you walk in and recognizes your, there you go straight through to the room. again, you’re looking at large screen T large screen monitors that they’re walking you through things.
[00:10:47] They have a very, let’s say 21st century kind of experience built up around their clinic visit. And so what Ohio state did is I’ll just read, this is [00:11:00] actually a. a press release from Ohio state. So it’s going to be a little fluffy, but, the partnership will provide central Ohio residents access to one Medical’s modernized primary care model, which supports seamless access to Ohio state’s network of highly ranked specialists, ambulatory facilities and hospitals, one Medical’s membership-based model combined 24 by seven on demand access to telehealth services, parents with convenient in-person care.
[00:11:25] In addition to direct. To consumer membership we’re in [00:11:30] 7,000 employers have sponsored memberships on behalf of their workforce. Together. The two organizations will aim to achieve greater clinical integration and deliver exceptional value to consumers and employers. And then it has a quote from their CEO, about transforming the health system.
[00:11:47] Actually, this is worth reading. Let me, as an academic health center, we are focused on transforming from a health system to a comprehensive health platform to provide innovative care and. Delivers [00:12:00] unparalleled experiences for all those who look to us for care across Ohio and nationally. So thinking about that, there’s two, some people, I might say there’s two buzz words in there, but I don’t think they’re buzzwords.
[00:12:10]I think it would have frames up is how Ohio state is thinking about their health system. They are looking at it as a health platform, which means that they have to facilitate a transaction that is frictionless. They have to facilitate a transaction where the data moves around the system. they’re [00:12:30] actually creating a platform where the user experience becomes central to how things function and how things operate.
[00:12:38] They’re starting to design first of all, around a P a platform. And around experiences. So I think that’s a important, my, so what on this quite frankly, is there’s going to be times where you need to move faster and you just partner. It’s not that Ohio state does not have the resources to do this. They’re a, a world class academic medical center and they have the financials to support it.
[00:12:59] They have [00:13:00] the university that has the, technical wherewithal to support it. So it’s not, that is not the issue here. Sometimes you just have to move fast. Partnerships are the best way to move quickly. Alright, let’s take a look at a consolidation. All right. So we’ve heard a lot about consolidation of health systems, right?
[00:13:22] Systems are consolidating. We’re still, we’re continuing to see it. And I think it’s going to continue to be a theme through the end of 2020 and [00:13:30] 2021. Actually, we might as well just say it for the next five years on this show, I’m gonna be telling you the consolidations a major theme, and it will be until we, we get to the, Saint Elmos fire of, scale, economies of scale is what we’re always searching for.
[00:13:47] If you’re not familiar with Saint Elmos fire, it’s a, I actually don’t. Really know it. I saw it from a movie back in the, back in the eighties and it’s a fire that the ships would chase thinking it was a lighthouse and [00:14:00] it turned out to be something that led them back out to sea and they never arrived at the lighthouse.
[00:14:05]that’s a sad story. Anyway, consolidation of providers into health systems increased substantially 2016 to 2018 health affairs, health affairs.org. Here’s the abstract. I’ll just share two quick pieces of it. A provider consolidation into vertically integrated health systems increased from 2016 to 2018.
[00:14:22] More than half of us physicians and 72% of hospitals we’re affiliated with one of 637 health [00:14:30] systems in 2018 for profit and church operated systems had the largest increase in system size driven in part by a large number of system, mergers and acquisitions. let’s see. A little more on the results, using national data, we found that, share of primary care physicians affiliated with vertically integrated health systems increased from 38% to 49% or 11 percentage points from 2016 to 2018 in 2018, more than half of all physicians.
[00:15:00] [00:15:00] And 72% of all hospitals were affiliated with one of 637 health systems identified in the compendium of us health systems. From the agency for health care research and quality. what’s my, so what on this, my son went on, this is, it’s going to be a little bit of a curve ball in that. I think there’s an opportunity here.
[00:15:21] I don’t think that providers that is independent physician practices are partnering with health systems [00:15:30] because they want to, I think they have to. I think they have to get access to patients through contracts. I think that’s one aspect of it, but I think the others just firmly on our doorstep, which is the cost of the technology aspects and the sophistication of the technology aspects of being in healthcare has gotten to be too excessive for the independent physician.
[00:15:52] Thinking about the security, the security budget alone is challenging. The EHR, customization. is [00:16:00] significant and let alone the EHR implementation. itself, any kind of digital tools that are going to be required as well as access to AI and some new capabilities. They’re going to find themselves just from a shoe budget standpoint, falling behind on a daily basis.
[00:16:16] And so I don’t say that to be negative. I say that to say, I think there’s a business opportunity here. Anytime you hear something like that and you go, Oh, they’re falling behind. They don’t want to do something, but they have to do something. If somebody can come in here with a [00:16:30] technology platform that gives them access to advanced analytics, a high functioning EHR, that’s customizable to the physician, practice ways to engage their consumers, to, to create again, I keep saying this, you reduce the friction of the transactions and.
[00:16:48] And I think more of them would stay independent. So there is an opportunity. There’s an opportunity for orange, for some enterprising, organization or group to come in there and say, we’re going to [00:17:00] create a platform specifically for independent physicians and independent physician practices.
[00:17:06] That’s my, so what on that curious, what you think on that? A good RX files to go public, boasting a track record of profitability. All right. So I like good RX. Our family uses good RX. I think it’s a, if you don’t know what it is, transparency for finding the lowest cost on prescriptions in a, in your local market, I think is a [00:17:30] good thing.
[00:17:30] I don’t understand why it’s not something that’s just. Provided when the physician writes the script, to be honest with you, when a physician writes a script, it should, it should have the script right there, and it should have not the clothes that you chose, this pharmacy, that’s where we’re going to fill it.
[00:17:47] And it should say, here are the three places you can get this filled and here are the prices. I think that should be standard. And actually there are some health systems that do that. That’s not an original idea with me. I know that, we’re seeing that more and more. I [00:18:00] think it was a st. Luke’s at a Bethlehem, had it.
[00:18:02] And I’m sure a couple of other Stu anyway, the other aspect of this is that, the IPR IPO market and healthcare is pretty hot right now. So past 12 months you saw the Vongo Phreesia health catalyst change healthcare, one medical go health, progeny, Amwell. Any others are missing. I can, but there’s a lot of them, right?
[00:18:23] So there’s a lot going on in this space. the thing that makes good RX distinct is odd. but they [00:18:30] have, the revenues are up and their profit is up. They actually have profit in 2019, they did 66 million in profit, which was at 50%, from 43 million. The previous year, just interesting. I think it’s a good tool.
[00:18:42] I wish it was just commonplace. I wish there wasn’t a need for this kind of tool, but there is. So this is a good solution. I hope their IPO goes well. I, to be honest with you, I think it’s going to be. a lackluster IPO. And the reason is because, I think it’s because of that profit. I [00:19:00] wish they were reinvesting that profit and what’s next, outside of their, transparency into, medications.
[00:19:07]I’d like to see some other tools becoming along and what’s in the pipeline and quite frankly, Profits are not for whatever reason are not really valued in today’s IPO market. More of what’s valued is your investment in the future and being ahead of the game. So there’s an opportunity there.
[00:19:24]but speaking of success stories we have Lavonne go is going to [00:19:30] cocreate a new digital and clinical care experience across behavioral health continuum with Magellan health. this story is just downright fascinating to me. And I shouldn’t say story. This is a press release. I’m pulling this straight from the Magellan health, website.
[00:19:46] So here it is Magellan health. One of the nation’s leaders in behavioral health today announced this strategic relationship with Lovango. The leading applied health signals company, empowering people with chronic conditions to [00:20:00] live better and healthier lives. Making the Wago for behavioral health available the Magellan health customers and their members as adults, usual entry point to a wide range of Magellan health Vongo solutions and services.
[00:20:13] All right, so there you have it. again, this is one of those partnerships. The partnerships are about moving fast and here’s what I read into this. I don’t know anything. I’m not on the inside of this, but Magellan health probably had a problem. They had a digital problem. They didn’t have a way to engage their consumers.
[00:20:28] They didn’t have a way to grow [00:20:30] based on a digital platform. So they stepped back and said, how do we get there? Or Lovango approached them and said, look, you’re about to get your lunch handed to you by a handful of really good competitors that have digital capabilities. They were digital first.
[00:20:44] And behavioral health. Second, that’s a little harsh, but they had behavioral health capabilities, but they, I thought about it first from a digital standpoint. And there’s a handful of those companies out there we’ve even had multiple of them on the show to highlight, the digital nature [00:21:00] of behavioral health and how it works together.
[00:21:02] So Laval steps, Lavanya and says, look. We know how to engage chronic patients. We know how to gauge patients. We know how to engage health systems and, we can create, we can be your digital front door and we can help you to, I’m sure Magellan could have hired people. I’m sure they could have built it out over three years.
[00:21:19] You don’t have three years. So at some point they have to step back and go partner. We’re going to partner with a while ago. We’re going to go from being in last place in this whole, Yeah. [00:21:30] Spectrum of behavioral health companies that are digital to leapfrogging and potentially taking a lead just by partnering with the right provider so that, I don’t want to do too.
[00:21:40] So what’s about partnership. so how about this as a, so what. Which is behavioral health is a significant deal. If you don’t have a behavioral health partner that is digital first, be looking at that to try to figure that out. That is something that your health system is going to be looking for probably in the next 12 months [00:22:00] to, to bring that to bear.
[00:22:02] All right. There’s a, I don’t usually do these kinds of stories and I really haven’t vetted it and it makes me a little scared to go into it. Cause somebody is going to. Somebody is going to vet it and come back to me. But I wanted to cover it because it was really well written and done, but it’s a PhD camp out of Australia.
[00:22:23] Okay. Luke Oakton, Rainer, a radiologists PhD candidate out of, [00:22:30] Out of Australia, as we said. And what he did is he goes into this whole aspect and I know I want to share this. I want to share this because he lays an awful lot of responsibility for AI not being adopted in health systems at the feet of the CEOs.
[00:22:46] So that’s the first reason I want to go into that. The second reason I want to go in that to this is because, we need to be following reimbursement very closely because it will dictate where technology goes, especially in this AI and machine learning space. All [00:23:00] right. So he starts with, it’s widely known that the medical AI community that has, that it has yeah.
[00:23:06] Troubled marketplace for AI developers. Okay. Medical AI, community troubled for AI developers. A majority of companies have developed useful AI models, but they’ve been unable to sell them. This has led many predictions that we’re going to see a crash amongst medical, AI startups. And I’ve heard that same thing, an AI winter, if you will, to be clear, this has never been a problem with technology.
[00:23:29] I agree. I [00:23:30] think we have a lot of really good. AI solutions that are out there, but on the clinical side, it takes a long while to get these things in place because they have to be proved out. They have to be, they have to be vetted just like any other, clinical process that we’re going to be embedding, deep learning works.
[00:23:47] And there are lots of ways that it can be applied usefully in medicine. It wasn’t alignment problem. The people who procure medical technology, the CEO’s are motivated by business needs. Not. But not by how useful the [00:24:00] model is. So he lays this entire thing at the CEO’s feet. Actually, as I’m reading this again, you should just read system because I’m, I don’t think the CIO is, are really selecting an AI platform for clinical decision making and that kind of stuff and thrusting it on the system.
[00:24:17] I think it’s the system that’s making that decision. And when the system makes that decision, he’s making the case that the business needs Trump, the efficacy of the. of the clinical results. And I would also [00:24:30] make the case on this, that the clinical results, are, far from vetted at this point.
[00:24:35] And so we are at the early stages of some of these getting funding and being vetted. All right. So he goes to the strongest business. Incentive is money earning more and spending less and providing a models can help, can help here. Can help here, obviously, whenever there’s an incentive, it drives behavior and that’s really true for health systems.
[00:24:56] He actually shares a story. And again, I haven’t looked [00:25:00] at this story in detail, but somebody, one of my listeners who’s smarter than me. Who’s looked at this, I’m talking to you, David Ben Smith. he goes on to tell this story and, And it’s around radiology. So reimbursement is how medicine incentives actually, help people central payer, whether a government or insurance company decides what medical management is cost effective to improve health.
[00:25:22] When a test or treatment is reimbursed, then healthcare providers get paid and use it. All of a sudden CIO is our real it systems are really excited. Pay some money to a [00:25:30] company and get much more money back for using a product. Does it work? Work well? It turns out it does work. This is an example. I’ve spoken about mammography CAD before an old form of AI intended to assist in detecting breast cancer.
[00:25:44] This became popular in the two thousands. When CMS decided to reimburse cat aided mammography tests, a provider would get about $10 more. If they used CAD than if they did a standard reading within a decade, almost every screening [00:26:00] mammogram in America is red with CAD assistance. But you say maybe if they just used it because it was amazing.
[00:26:07] Nope. It didn’t work again. These are his words and I would love one of my listeners to help me with a story. In fact, nobody else uses it. I’ve never found an exact number, but CAD use outside of the USA is particularly, non-existent why? Cause it doesn’t work. And you don’t get paid for it. Just think about that.
[00:26:28] Medicare has spent hundreds of millions, if [00:26:30] not billions, on a technology, which didn’t work driving widespread use financial incentives are powerful and dangerous things. So financial incentives, a big deal. absolutely. So in the context of this, he talks about the company that’s getting this thousand dollar, Per patient reimbursement, vis AI with, a product called content, C O N T a capital CT contact.
[00:26:56] A CT. So what they do is, let’s see, visit [00:27:00] AI claims by reducing the time specialists to review the CT scan of possible blockages. They prevent long delays during which time more and more brain cells are dying from lack of blood. They have published few papers on the topic here in here, and had, provided a fair bit more to CMS to justify this claim.
[00:27:18] So CMS is ready to, support their, to reimburse based on this AI and, the way that they got CMS to reimburse it, they had, [00:27:30] and again, this has been a while in the making it’s 20 1840 page document. there’s also a 2000 page document, from CMS that you can read as well. I’m not going to read that.
[00:27:41]but here’s the, the main things that. Vis AI put in front of them. They showed several things faster time to notification of the clot busting specialist, faster time to transfer from peripheral hospitals to the central hospital, where the relevant procedure can be performed and faster time to clot busting procedure.
[00:28:00] [00:28:00] These things alone. He goes on to say these things alone are interesting, but rely purely on the existing knowledge that delays lead to worst brain injuries. As the saying goes and stroke time is brain, but vis AI didn’t stop there. They actually did the thing I always harp on about. They showed outcomes, improved, modified ranking score at discharge improved NIH stroke score at day five and improved mrs.
[00:28:26] At day 90. he went on to state that [00:28:30] once this came out, not everyone was real happy on social media and other things. I, I, again, I shared this story for two reasons. One is it’s interesting that he laid the entire thing at the feet of the CIO. I don’t think that is, I think that is a miss, calculation on his part, but the second is.
[00:28:46] That, if I were a CIO, I’d be keeping an eye on me. Bereavements reimbursements are going to drive what I’m going to need to have in place. And I’m not sure we’re ready for this level of AI. I’m not sure we have the data to support it. I’m not sure we have the systems to support it. I’m not sure we have the skills and [00:29:00] capabilities to support it.
[00:29:00] And if the, if they start reimbursing it, our systems are going to be knocking on our door saying, Hey, we need to participate in this. This is a thousand bucks per patient. Let’s get in front of this. So anyway, thought I put that’s my, so what on that is get ready for some of this stuff. And I think your EHR provider will help in some way, but in other ways, you’re going to have to figure out what gaps are gonna exist and fill those gaps.
[00:29:26] Let’s see, here’s a quick one. Multiple [00:29:30] workers fired at George, after George Floyd’s medical records, improperly accessed. And, I just, I share that to say, not all attacks are external and we have to remain vigilant internally. it’s a politically charged environment. there’s a lot of incentive.
[00:29:47] There’s a lot of incentive financially, too. If you had George Floyd’s medical records, probably sell those to fill in the blank and get that information out there. hope, hopefully people know by now that every click in the EHR is [00:30:00] monitored. but for those who don’t, who happened to be listening to this show, every click in the EHR is monitored.
[00:30:06] And when I say every click, every click, we know what you’re doing in the EHR and we have to, it’s just part of what we do. let’s see, there’s another story on, AI, discrimination and really AI bias. it’s a pretty it’s a hefty article. I don’t, I’m not going to go into it in detail because I spent a little time on, some of the others, but it is a Harvard business [00:30:30] review.
[00:30:30]I don’t know if it’s an article. Andrew Bert, August 28th, 2020 is our artificial intelligence fair. Great. And he goes on to talk about that. We have a lot of frameworks for determining if AI is fair and we, all we have to do is look at the laws that are already in place. We go look at the equal credit opportunity act, civil rights act, fair housing act, equal employment opportunity commission.
[00:30:53] We can look at those things as frameworks and structures, where we get into trouble. It’s when we use things as proxies, I’m [00:31:00] really summarizing significantly here. when we, use things as proxies, let me see, It occurs seemingly it occurs when it’s seemingly neutral variable, like level of home ownership, acts as a proxy for a protected variable like race.
[00:31:13]what makes avoiding disparate impact so difficult in practice is that it’s often extremely challenging to truly remove all proxies for protected class. And so that’s what we’re trying to do. let’s see. It’s sometimes not even clear what the most fair decision really is. And one study Google, AI [00:31:30] researchers, the seemingly beneficial approach of giving disadvantaged groups, easier access to loans had the unintended effect of reducing these, this group’s credit scores.
[00:31:39] Overall easier access to loans actually increased the number of defaults within that group. There are five, therefore lowering their collective scores over time. to me, this was just a, it was, it was interesting. He, and he does go on to say it’s really complex, right? Despite all these complexities, however, existing legal standards can provide a good baseline [00:32:00] for organizations seeking to combat unfairness in their AI.
[00:32:04]but we have to know what the algorithms are. he goes, yeah, I have to say first regulated companies must clearly document all the ways they’ve attempted to minimize and therefore to measure disparate impact on their models. They must, in other words, carefully monitor and document all attempts to reduce algorithmic unfairness.
[00:32:21] And a second thing is regulated organizations must generate clear good faith justification for using the models. They [00:32:30] eventually deploy in fair methods existed. If a fair methods existed, they would also have also met the same objectives, liability can, and soup. Okay. So again, this is a, this is actually, it’s not a long article.
[00:32:45] It’s just a half the article in terms of. Of what it’s talking about. Here’s my, so what on this, we used to have it governance. Then we went to data governance. Then we went to, fill in the blank governance. I think the next thing we’re going to have is we’re going to have our rhythm covered it.
[00:32:58] So we’re going to have what we’ll call [00:33:00] it. Something else. We’ll call it our AI whatever group or, our, whatever, what we’re going to be looking at in that group is anywhere we’ve introduced where computers are making decisions. And we’re going to be evaluating. Is that decision right wrong?
[00:33:17] Is that decision fair, unfair. we’re going to have to create this level of governance where we’re looking at the algorithms. So first that’s going to require us to understand the algorithms and we shouldn’t be buying software that we don’t understand the algorithms or [00:33:30] building software that we don’t understand the algorithms.
[00:33:32]but we’re gonna have to collect that information, bring it before that governance group is going to have to be able to look at it because quite frankly, you’re going to have to document it anyway. At some point, you’re going to get an audit, I don’t know, from who, but you’ll get an audit from somebody on your fair practices of using that data in the care of patients.
[00:33:48]So that’s going to be an important metric. So I think that’s the, so what the, so what is, do you have an algorithm? Governments? That’s an awful name. You come up with a better name, but you get the [00:34:00] idea w the group that’s going to be looking at how computers make decisions with how chat bots interact with, with your patients, how we made those decisions.
[00:34:08] You don’t want the technology group making that by themselves. Actually, you don’t want the clinicians making those decisions by themselves. You probably want ethicists. Ethicist ethics, ethicists. That’s right. Ethics. at the table talking about, yeah, you probably want diversity and a representation at the table to talk about.
[00:34:28]again, the bias that’s [00:34:30] readily available or readily transparent. Okay. That’s enough of that. But I think the last story I’m going to leave you with. Halakah is moving. He’s having some fun. I don’t know that it’s just John. I, Mayo has been doing some great work, for a while here, but, but this has John’s fingerprints on it.
[00:34:48] So I’ll share it with you, miter partners with Mayo clinic nuance on common data elements. So if you remember the conversation we had a little while back. With, with a nice Chopra, he talked [00:35:00] about the fact that we, the cures act gives us an opportunity to self govern, to identify the data elements and to bring those standards forward as health systems.
[00:35:12] And if we don’t, what’s going to happen is the government’s just going to keep setting them. They’re going to set the floor. They’re going to keep raising the floor unless we do this ourselves. And he gave the example of the energy sector, which did this on their own. And then the banking sector, which did this on their own.
[00:35:27]So those two industries that do this on [00:35:30] their own, but healthcare has yet to really produce any of these data standards that have gotten traction across multiple health systems. in walks this, this, partnership, which I think is a huge step in this direction, the McLean Virginia based nonprofit organization announced a partnership with Mayo clinic to conduct research and development on common data elements for oncology cardiology, and COVID-19.
[00:35:54] The collaborative research will further the development of platforms for intelligence automation, including [00:36:00] M code and M car common data standards for oncology cardiology to improve quality and coordination of patient care. there’s a lot more to this story it’s worth picking up. I got it from healthcare innovation group, HC innovation group, doc.
[00:36:14]I don’t know who the author is. Author is David rats. We’ve covered some of his stories before September 8th, 2020. I share this to say. John is a very partnership oriented type guy. if some of this stuff is of interest to you, I think there’s an [00:36:30] opportunity to expand the, the data sets use across the industry.
[00:36:35] And as we do that will become the de facto standard and eventually the, the standard for, the use of. data sets across oncology cardiology, and COVID-19, which as would just be a huge benefit for, combating this pandemic and addressing some of the, extremely challenging, conditions that present themselves in oncology and [00:37:00] cardiology.
[00:37:00] So I think that’s enough for today. I have another 15 stories, so you’ll have to come back next Tuesday and I keep checking back on the, on the LinkedIn feed. I’ll drop a story a day, get this discussion started and, just to see what you guys think that’s all for this week, don’t forget to sign up for clip notes at this weekhealth.
[00:37:20] Or this week in health it.com special. Thanks to our channel sponsors, VMware, StarBridge Advisors, Galen Healthcare, Health Lyrics, Sirius [00:37:30] Healthcare, Pro Talent Advisors, HealthNXT and our newest channel sponsor McAfee for choosing to invest in developing are the next generation of health leaders. The show is a production of This Week in Health IT. For more great content. Check out the website this weekhealth.com or the YouTube channel. if you want to support the show best way to do that, share it with the peer, send them an email, let them know that you’re listening to the show and you’re getting a lot out of it. Please check back every Tuesday, Wednesday, and Friday for more episodes. Thanks for listening. That’s all [00:38:00] for now.