This transcription is provided by artificial intelligence. We believe in technology but understand that even the smartest robots can sometimes get speech recognition wrong.

Welcome to This Week in Health IT Influence, where we talk to the people who are influencing health IT. My name is Bill Russell, healthcare CIO coach and creator of This Week in Health IT. A set of podcast videos. and collaboration events dedicated to developing the next generation of health leaders. I want to do a special shout out to our show sponsor Sirius Healthcare.

Sirius stepped up to sponsor the shows we've been producing in an effort to, to capture and share the experience stories and wisdom in the industry during this pandemic. I am extremely grateful for them and I'm grateful for their commitment. to the, uh, the mission of this week in health IT to develop the next generation of health IT leaders.

Today's show is about data. And when I think about data, I think of Star Trek first. The second thing I think of is Health Catalyst. And then further on, I think of Dale Sanders. If you're following Dale on social media, you know that he has been extremely busy lately, and I'm excited to have gotten the opportunity to sit down with him this past week.

And I'm looking forward to sharing that show with you right now. Today's conversation is with Dale Sanders, Health Catalyst CTO. Uh, good afternoon, Dale. Welcome back to the show. Hi, Bill. Thanks. Good to be here. Well, thanks for taking some time to meet me. You, uh, you really have a great background for this.

You have a military background, you have an IT background, and you work with one of the premier... Data companies and health care. So I'm really looking forward to the to the conversation And I don't want to minimize the military background because I've seen a bunch of back and forth with you I had a conversation today with Drexel Ford and You know, the military really spends a lot of time preparing for these kinds of situations, don't they?

Yeah, I mean, that's the nature of the military, especially as an officer, you're constantly preparing for disaster. That's what you do for battlefield conditions, and hoping that you never actually have to engage, but they are, the exercises are non stop, and the training is very formal, both the the procedural, as well as the mental training for these kind of things, so a couple of times I've reflected the last couple of weeks about this really is sort of an interesting convergence of my military background with data and decision support and battlefield management and then, you know, healthcare data and epidemiology so in a, in a kind of a morbid sense it's intellectually interesting and fascinating for me.

At this time in history, but it's sad too, I... I certainly feel bad for our country and I feel especially bad for the hotspots like New York City. Yeah, and I've been talking to some of the organizations in these affected areas and they have a lot of questions. And I'm hoping that we can delve into some of these things.

I've been, I'm trying to think if there's anyone more prolific on LinkedIn than myself. Uh, and you would clearly be way, way above me. You're somebody who I, I strive for. And if people aren't following you on LinkedIn, uh, you're constantly bringing articles that I normally would not read, uh, to the forefront.

And I'm, I'm just learning a ton about, about public health, about, uh, data, uh, about organizations that have been really focused in on, um, Elevating the game of, of, uh, of analytics and data science around this. So it's, it's been good. So let's just get right to the questions. Um, again, I've sourced some of these questions.

Dr. Benzema was kind enough to send me some questions. Uh, Charles Boise was kind enough, you know, both of those gentlemen. And so the questions are going to be a little harder than they usually are because Because I, I cheated a little bit, but let's just start with, you know, what are three analytic insights that all or nearly all health systems can access today with the capabilities that they have in place that can help them to allocate the limited resource, resources that we have more effectively?

Oh my gosh, across the board. What could they access? The, you know, uh, fundamentally there are two issues right now. that the entire country is struggling to answer. And that is the consumption rate of PPE in wartime environment, essentially. And then predicting the actual infection rate, and not just the infection rate, but the, the severity of illness rate in their catchment areas.

Um, and so it's, this analytic environment is actually quite complicated right now because We don't have a clear picture of the three identities of patient types that we should be focused on right now. And those three identities are the super high risk members of the community who, if they are infected, Will likely suffer a fatality.

Um, you know, initially it came out that hypertensive patients, patients over 60 and uh, patients with COPD were clearly at risk. But now that's starting to get a little uncertain too. The, the age and the demographics are smearing a bit. There's some debate about whether immuno-compromised patients are at risk or not.

Some physicians feel that the immunocompromised, uh, compromised patients are actually at lower risk. because the virus is essentially stimulating an over response. Uh, from the immune system, and that's what's, uh, you know, leading to ARDS and things. So that's one identity that we should all be trying to access, um, who's at risk and are they practicing, uh, social isolation, and if they do present to the hospital, do we have the capacity to escalate their care?

Then the second identity, uh, that's fundamental are those that are suspected with symptoms like flu and RSV, um, that might have COVID. Uh, we're not quite sure and we're not ready to test yet because we don't have enough tests to go around. But you need to start watching those folks and you, and if they present to the facility, you need to prepare two tests.

And then the third category, of course, is the confirmed COVID patients. So, those are the three identities that every healthcare system, in their analytic environment, need to start identifying and managing. The CDC, um, CDC has been largely ineffective in this response, in my opinion, and I try to give them as many graces as I can.

I know this is a confusing and chaotic environment, but they've been slow to publish the national standards that we should all be adhering to in those three registries that I just mentioned. So, uh, We've published some from Health Catalyst. We've done our best to synthesize the evidence in the industry as well as out from CDC.

And we've published those registries out to our clients, and we'd also make those available. I've published those on LinkedIn. But all, virtually all, of your analytic use cases start with a consistent and precise definition of those three patient types. Your capacity planning, your, your, um, testing strategy, uh, and when I say capacity planning I'm talking about PPE, ICUs, vents, beds, staffing, all of that.

Um, let me mention one other thing too there Bill, uh, before I forget. One of the challenges that I see everyone struggling with right now, Besides capacity planning, based on the emission rates in their catchment areas, um, is the management of PPE. And so let me just encourage everyone that's struggling with this right now to first, as an organization, because we don't have any national guidance on this, Define what you should be doing in terms of consumption on a per COVID patient per day basis.

In other words, how many, how many items of PPE, you know, hoods, glasses, gowns, bonnets, bunny suits, shoe covers, how many items of that PPE are you consuming? Or, do you think you should consume on a per COVID, per patient day basis? Does that make sense? Let me just kind of pause on that for a second. So it's, this is basic supply chain 101, right?

Model what you think is realistic, right? From a caregiver safety perspective. What you think you should be consuming, realizing that we're in war time, so we don't have unlimited resources. You're probably going to have to issue guidance that extends The, the utilization of PPE out longer than it normally would be, but, um, determine what your clinical guidance and best judgment indicate, protection of your caregivers, and also, of course, you know, transmission to other members of the hospital, but direct caregivers.

Then model that according to the inventory that you have, and then look at what your burn rate is going to be and when you plan on running out of that, and then that will drive supply chain and distribution. upstream. So that's a, uh, it's a, it's an unfortunate situation right now. Um, I, across the country in all of the leading systems that I've talked to, I don't see anyone managing the supply chain of PPE from that perspective.

What they're looking at is the consumption PPE, and then they're adjusting the, the utilization guidance according to their inventory, not according to what, um, is appropriate for caregiver safety. And if we, if we never get to a guidance based consumption model around that supply chain, um, we'll, we'll never be able to effectively protect our caregivers.

So we're letting, we're letting the, we're essentially letting the PPE that we have drive our consumption guidance. As opposed to what we think is safest for our caregivers. And then pushing those requirements back upstream. Because that's the good news. There is, there is some progress being made on the production of PPE.

Um, you know, there's, of course, there's a lack of a national strategy about how to allocate that. But if there's no models and there's no data to support the allocation, then it's going to be hard if we ever do get our arms around the distribution of PPE to know where it should go. Yeah, so I think it was yesterday or the day before CMS came out and actually started to collect that information or want to collect that information from all the health systems, you know, what you have and what's available so that they can start to allocate that effectively.

But I want to, I want to drive into the systems a little bit. Um, because to a certain extent, you're saying that the inventory is driving, uh, the use of PPE. But to, to another extent, we're, I think there's two breakdowns of systems here, right? So we're talking about registries across an entire community, and a lot of time we have registries within health systems.

Uh, we don't have registries across an entire community. Uh, and then the second thing is around supply chain management systems. Uh, you know, supply, if I remember correctly, some of the supply chain management things treat things like masks as consumables. So it's, you know, you look at it as a quantity of a box, and a box could have 25 in them, and, you know, we never really managed those that closely, and I'm not sure the supply chain systems, is there a certain, uh, I mean, is there something that kicks in from a crisis standpoint that you say, all right, these systems aren't going to work, we have to figure out how to do something else?

Uh, and what would you do around the registry? Uh, let's focus in on that one, because that's probably the most important one. So, how do you, how do you all of a sudden, if it wasn't in place, get registries going across your, uh, your market and your region across multiple entities? Yeah, well, that's, you're not going to be able to do it in the community like a traditional registry, that's for sure.

Um, but what should have happened and what still can happen to catch up to this, Take the guidelines that are emerging in the industry. And again, I'll post those on LinkedIn if folks don't have access to that yet. And put those registries into your analytics environment and then make sure that your EHR data collection is lining up with that, realizing that EHR data collection right now is one of the last things that people really want to do.

And so you've got to build your registries out and the attributes of that registry within the reality of the chaotic environment we have, no one has time to, you know, click through a hundred attributes to precisely define a patient type, which, by the way, if we follow the CDC PUI form, there's a snowball's chance that that form is going to be filled out right now.

There's a snowball's chance. It's a paper based form. It's cumbersome. It doesn't reflect it. You know, the workflow of the patient, you know, some of the data is kind of available on the presentation of the patient. Some of the data won't be available for a couple of months. You won't know the outcome. So that CDCPUI form is essentially useless right now.

It's a, it's absolutely a public health perspective of this disease. What we need is situational awareness, which goes back to crisis management and military, you know, the the the concept that I'm gonna opine on in later writings and things is called intelligence preparation of the battlefield and it's a common concept and practice in the military to think deliberately how you're gonna lay down the data environment in a battlefield and to give the commanding officers the information they need to adjust and adapt.

And that's what we don't have right now. So get those, those registries have to be in place with a minimum viable data set, minimum MVDS. That's another thing that CDC should have done and I would love to see. And maybe I should try to do this actually, I've been thinking about it more. Someone needs to convene the EHR vendors, the major EHR vendors, the four big names, and, um, consolidate an MVDS strategy, make sure that that's all being collected as quickly as we can out in the field, and then feed that into the analytics platform.

to drive all these use cases, um, but don't go down the CVCPUI form. Right, and you simplified that pretty significantly. Let me move on here. So, uh, let's talk about the screening criteria. Um, have the current screening criteria for who should be sent to COVID 19 testing, testing been validated in the community, in community use, and what are the opportunities to continually Revalidate and adjust the screening criteria.

Um, if by screening criteria you mean patients that are high risk or patients that are presenting mild symptoms? Right. Um, yeah, I think that, I think the, the screening criteria for preliminary symptoms is pretty well established right now, right? With fever, um, some GI issues, but it's primarily fever, body aches, dry cough, some GI issues.

sense of, um, smell and, and, uh, hearing loss, or taste rather than not hearing. So there's, there's decent screening criteria, and if you pass those criteria, then, um, there's a pretty good chance right now that you have the disease. Uh, the question is, you know, what do you do with those folks, right? And the lessons that we learned and are learning from Wuhan and other places is sending those folks back home to their family ended up being one of the the primary points of transmissibility, right?

We sent those mildly symptomatic folks in China back home and they infected their, their, um, family. And then that's that in combination with the infections with the um, caregivers is what caused the major outbreak. By the way, I might mention before I forget it. I'm fortunate enough to have acquaintances and um, conversations with Sort of my equivalent in China, uh, uh, Why do cloud is the name of the sort of the health catalyst version of China.

And so I'm talking now with those folks and their analytics teams. They're three to four months ahead of us in terms of data analysis and pattern recognition and things like that. And so that's been an interesting learning experience to, uh, to chat with them. And I'm spinning up similar conversations with friends and colleagues in Singapore.

Uh, and be happy to report back on all of that when it starts to become more clear. That would be awesome. Uh, what's the, uh, you know, what's the most surprising insight that you've learned in the last couple of months as you've been sort of tracking this? Oh, um, I think one of the most surprising things to me, uh, Bill, is the, the wide variety of severity of illness.

You know, it's a perplexing disease, right? To, to have a large portion of the population completely asymptomatic compared to a portion of the population who, um, progresses, you know, to fatality in a matter of days. Um, that has been one of the most surprising and perplexing things for me, and I hope... You know, that what it does is it gives us a large population to understand the immunology and the genomics of response in those folks that are asymptomatic.

And maybe it will accelerate some treatment. Um, but I also would caution, you know, we've been, we've spent lots and lots of money on vaccines for SARS and MERS coronaviruses. And nothing's been effective. In fact, there's almost no money going into that, uh, research anymore until, until the COVID 19 situation, because there was, there was no progress being made in the vaccine.

Even the, uh, the pharmaceutical treatments were, were largely ineffective. So, it'll be really interesting to see, um, whether or not we can make progress with the coronavirus that we've never made before. So that's one surprising thing. I think the other surprising thing is, um, At the macro level, um, I'm just completely surprised at our country's inability to organize a national effort around it.

It's really come down to every municipality and every state on their own. And by the way, you know, some, some folks might interpret that as being politically motivated. I've been a political independent since I was... And, um, it has nothing to do with politics. I'm in conversations with senior officials. I had a phone call this morning with a very senior official in the federal government, a Republican.

And, um, he was, he's completely stunned at how disorganized the national effort has been. And, um, And literally his advice was every state and every municipality needs to take care of itself. So that's one thing that surprised me. I think the other thing that surprises me too is the, the um, and I, this is not a big surprise, but it confirmed what I probably knew.

In all the conversations I've had with my physician friends and clients, I've purposely asked them, have you ever had any kind of training even remotely close to this that would prepare you for this kind of disaster and, you know, battle management. And the consistent answer is no. Uh, they might have had a class in epidemiology, they might have had a class in public health, but our physicians have never had formal training in this kind of thing for the most part.

I know that there are some exceptions to that. Um, so I hope as a consequence of this event that we We step back as a country and, and provide that kind of training, not just the training, but also literally the lessons from the military about how to manage these kind of situations and the organizational structures and the roles that are played and You know, it's not rocket science.

It's been solved in other sectors before I'm gonna ask you to, to talk about laying out the, the, the data thing going, the data mechanism going forward. So if we could rewind the clock, like a year and a half, two years, and we were gonna lay out Yeah. The fundamental principles for data. Um, I, I'd, I'd love to hear, you know, what your thoughts are on that.

Um, but, you know, the, the thing I would, uh, say in terms of the, uh, the response, it's, it's really, uh. We have, we lack imagination, you know, if you were to go into a health system and say, we need to prepare for this, they would have looked at you like, come on. I mean, is this going to be on the test? You know, it's, it's like, people are like, well, why didn't we lock down New York earlier?

It's like, you know, I'm living here in Florida. People from New York are, and you see them with their New York license plates on their car. They're acting like there's no issue whatsoever. They're out at Home Depot. They're, they're just driving around. And to a certain extent, certain communities, certain cultures, it's hard to just walk in there and say, we're gonna lock you down.

We're gonna do these kinds of things. If there isn't a frame of reference, we lack imagination to say, hey, this could be worse than we think, um, it's going to be. And we're just not, you know, in China you could just sort of walk in and say, this is what we're going to do, and everyone goes, yeah, this is what we're going to do.

In the U. S. it becomes a little harder because there's just cultural challenges. I, I'm not making excuses, I'm just saying it's, it is a, it is a multifaceted complex problem. Let's go to the data aspect of this. If we could rewind two years and we had the imagination to imagine this was coming. We, we saw SARS, we saw, um, you know, swine flu, Ebola, and we, we took those markers seriously.

And we said, all right, we got to prepare for the next one. What should, from a data standpoint, data and analytics standpoint, what should we have put in place? Well, it's battlefield surveillance, right? You're looking for signals, unusual signals in the field that indicate a new risk is emerging. Um, we knew, you know, that, and this is a good part about the CDC, actually, the, the field units in CDC and other, um, organizations are routinely looking at these, uh, novel viruses crossing over.

And so, um, you know, if you're going to follow the military metaphor, those, those human intelligence sources that are out in the field looking at this, Could start sending signals, right, collecting intelligence, reporting that back up to a structure that's nationally organizing all of those signals and looking around and saying, okay, are we starting to see, for instance, an unusual spike in flu like symptoms?

That we can't quite explain. So, um, interestingly, a lot of folks and a lot of physician friends of mine have all said, you know, we had unusual symptoms of flu, um, in December, January, maybe even as early as November. But now everyone's wondering if that was actually COVID in mild to, uh, to asymptomatic patients.

Um, and then some of those, of course, progressing into ARDS, but were they just, I think, largely diagnosed as flu cases? So we're actually looking at our data right now to see if there were unusual spikes in flu like symptoms. in November, December that would have signaled that something was happening. So that's the early warning system that you want to put in place.

Alright, so let's assume we see the early warning. Now, now where do we go with the data? Well, like with any kind of early warning, you have to decide, are we seeing false positives or false negatives, right? And we tend, in healthcare, we tend to lean towards false positives, right? In other words, As we dial up sensitivity and specificity, and we're finding a balance between those kind of things, generally speaking, what we do in healthcare is we dial up sensitivity, and we prefer to err on the side of false positive.

So, um, in these kind of situations, you can't do that. Because if you assume that everyone has the disease, you're going to over treat and you're going to consume all of your supplies. It's like, it's like, um, scrambling your fighters every time your radar picks up a crow. That's literally the, the, the military metaphor.

You cannot scramble fighters every time... If you dial up the sensitivity of your radar, every time a crow flies by, you're going to scramble fighters. You're going to soon run out of fighters and pilots and, you know, fuel. And, and so, what we should have done, looking for those symptoms, unusual collection out in the field, listening to what clinicians are saying, and this has to be an ongoing thing.

You can't, you know, in peacetime, the military is an intelligence network that's constantly mining for threats. that you don't know about. So you don't want to spin up this network of threat intelligence. when you think there's a threat, because that, by that time, it's probably too late. You have to have it constantly in the field, monitoring, looking for patterns that might indicate a threat.

Then you have to forward that up to a national center that's sifting through that to look for patterns across the world and across the country to make sense of it. Um, and one of the first things that we, we should have done is, is reacted to the early warning signs in Wuhan. And there were plenty of epidemiologists Who kept warning us, as soon as this emerged in Wuhan, that this could be a significant problem, and we did not listen to that.

And I put myself in that. I made a couple of posts on social media saying, you know, let's not forget that, um, you know, there's 18, 000 deaths in, from normal influenza this year, so let's not overreact. Um, So we should have paid more attention. We should have, we should have started ramping up registries at that time.

Hopefully there'll be another round of that. I can't imagine that, as a country, we wouldn't invest in that now. Yeah, I, you know, I, I really appreciate your time. Um, You know, there's, there's part of me, actually, you know what? You were kind enough to give me your time, and Health Catalyst is kind enough to, uh, to give me your time.

I really want to close this out with, what is Health Catalyst doing for your clients, uh, right now? Are there any new approaches or things that you're doing, uh, around this that, uh, that you can share with us? Yeah, we're doing a number of things. Of course, our teams, you know, typically we have analytics engineers, that are associated with our technology at client sites.

And so, those folks spun, the analytics engineers spun up very quickly. Um, one of the early applications that, that we deployed was a patient tracking application so that, um, we could, we could watch the progression of a COVID patient through the healthcare system, um, both by location as well as by, um, the people that they were touching and interacting with.

So we, we got a sense for who had been exposed, um, at our client sites that might not have had PPE involved. So that was one of the first things we did. Of course, there was all sorts of ad hoc analytics, um, all around the, um, the management and supply of ICUs and ventilators, um, we've got a capacity planning tool and modeling tool that we, uh, leverage based upon the epidemiology work at UPenn.

So they were great. They put, uh, An open source, uh, GitHub repository out there for their model. We borrowed that model, we enhanced it, and we added a few features, we contributed our code back to them. So that capacity planning model is out there now, and I encourage, um, I can send a link out to you for that by the way.

That would be great. We've got a number of different dashboards. Frankly, I'm a little disappointed in our ability to provide a dashboard sooner to this. We've got a little spun up with a couple of Projects that had a longer lead time, they kind of got away from my, um, influence, and, and by the time they were down the runway, I didn't have time to redirect it, but anyway, that distracted us from, uh, the core COVID management dashboard that has all of the things you would expect to see on there, right, about informed by the registry, right?

Um, all about logistics and supply chain, because that's, that's where battles are won or lost. It's all around this logistics and supply chain. Um, and then longer term, the next couple of weeks, we have a, we have a patient safety and syndromic surveillance application that we will deploy specifically to COVID.

We have an existing patient safety surveillance system for inpatient monitoring. We're adjusting that to address COVID management and we're also adjusting it to incorporate um, signals in the community and we'll release that in the next couple of weeks. Um, so those are, those are kind of the big things.

I don't think I'm, I don't think I'm forgetting anything off the top of my head. Of course, the registries, I mentioned that, that's been pushed out to all of our clients. Is there, I'm sorry, last question, I promise. But is there anything we're doing, so, uh, patients who recover, and there's been weird, uh, it's, it's really hard from where we sit to determine what's really happening.

But they're saying, hey, there's a risk of people actually getting, um, getting the virus again or getting sick again. Um, are, are we, are we tracking those patients once they leave the health systems and, and, uh, or are we just waiting for them to re present at this point? Um, I don't know of anyone that's tracking them proactively, Bill.

We're, of course, we can track them when they come back to the system. I don't know a healthcare system, you know, like Singapore, South Korea, China, they were definitely tracking those folks through the social media apps. I'm not aware that that we're doing anything like that. By the way, let me comment really quick on one item of testing that I still am stunned that we did not achieve as a country.

You can break the U. S. into ten different essentially CDC regions and Health and Human Services regions. We should have spun up random sampling of the populations in those 10 regions right away to get an idea of who was infected and then implemented social isolation around who was infected in the in the geography, in the context that they had.

Now our only option is, um, is sort of rationing tests or, you know, mass testing of the entire population and rationing tests according to who presents isn't going to give you a good sample and mass testing isn't going to be realistic. So I'm still surprised and I, you know, this common sense has been confirmed by a lot of my PhD epidemiology friends.

They still don't understand why we haven't implemented A national sampling testing strategy to really understand who's asymptomatic, who's mildly symptomatic, and um, and what that means to hospital capacity management. Absolutely. Well, Dale, I thank you very much for, uh, taking the time as this progresses.

I would, I would love to have you, uh, come back on, 'cause I, we are learning so much almost by the, by the hour at this point, or by the day at least. Um, and I think in two weeks we'll have had, uh, just a. A ton more information to talk about. Yeah, and if I, you know, if any of your followers, um, have questions and things, Bill, they can reach out to me or that you can uh, reach out to me if they contact you first, and I'll try to chase down answers.

Um, situation is still pretty fluid. There's not a lot of answers right now, but I'm happy to help where I can. And I'm also, one other thing, I'm, I, you know, I'm, I'm working at, um, on some national committees and things, um, and work groups, and so If I can use that venue to bring forward issues and happy to do that.

Yeah. And, um, yeah, let's, let's get those links. We'll get that done. And, and again, thank you very much for your time. I really appreciate it. And, uh, yeah, I look forward to catching up again. Okay. Thanks, Bill. That's all for this show. Special thanks to our channel sponsors, VMware, Starbridge Advisors, Galen Healthcare, Healthlyrics, and Pro Talent Advisors for choosing to invest in developing the next generation of health leaders.

If you want to support the fastest growing podcast in the health IT space, the best way to do that is to share it with a peer. Send an email, DM, whatever you do. You can also follow us on social media, uh, you know, subscribe to our YouTube channel. There's a lot of different ways you can support us, but sharing with peers is the best.

Uh, please check back often as we will be dropping many more shows, uh, until we flatten the curve across the country. Thanks for listening. That's all for now.