From 3M Health Information Systems
Podcast Episode Transcript: Patient care from home: Achieving optimal outcomes
Gordon Moore: Welcome to the Inside Angle Podcast. This is your host, Dr. Gordon Moore and with me today is Dr. Juggy Jagannathan who is an expert in many things and a broad thinker. And I think really a renaissance man in terms of thinking about information and health care and how it comes together.
Today, we’re going to have a conversation based on a talk he gave recently that looks into how we think about understanding the motivations of healthcare monitoring strategies that can improve results in a care delivery. And I want to dive into that with him and hear what it is he spoke about recently. Welcome.
Dr. Juggy Jagannathan: Thanks, Gordon. I appreciate you talking with me again.
Dr. Moore: Tell me, what was the talk you gave recently and what was the context of that?
Dr. Jagannathan: I had given a general talk on AI in healthcare at our tech forum and the moderator said, “How about a follow on talk?” and I scratched my head and I said, “It would be interesting to see how we could take care of patients in their home setting.” And the motivation goes something like this, and this was echoed by others as well.
If you end up in a hospital, it is a failure of ambulatory care; if you end up in the ambulatory care, it’s a failure of home care, and if your home care is not good, it’s a failure of community care.
I was wrestling with this whole notion and then I said, “Okay. What is the best way to take care of patients in the home? Obviously, that is going to be the most cost-effective way of taking care of a patient, which will drive the overall cost down as well as drive the satisfaction for the patients.”
That was the genesis and then, I dived deep into it and looked at what drives it; what is the data that you need to drive this process forward and what is the process that can actually be impactful in taking care of the patients at home.
Dr. Moore: Let’s take as a starting point that we understand the motivations behind this. It’s a net good thing to take care of people in the home. You said that one of the first things is to consider what data would be useful to do that. Why don’t we start there?
Dr. Jagannathan: I look at it and I see three broad categories of data that have a major impact on this whole effort. The first category of data is the data that you collect about the patient at the hospitals and the ambulatory clinics; basically, what is recorded in their EHR. This is the trove of data which is currently being used for what is generally referred to as population management strategies. The way they use it is they look at the diagnostic codes and the chronic conditions and the risk-stratify the population and assign different resources. You assign more resources to take care of the higher risk population, the ones which are more sick.
That is one trove of data. Another trove of data is what I would broadly label as environmental data. Environmental data is many things. It could be climate data—the level of air pollution, the status of your water—is the water polluted or not? Do they live near a coal fire plant, etc?
And then, there is a set of data which have been in the news fairly extensively recently, is the social determinants of health; do the people have access to good food; good health services; access to primary care? Is the place riddled with crime or violence; is there good employment in the environment; what is the health literacy; high school graduation?
There are numerous types of data which are geographically based, but they impact the health status of the population who live in that kind of bubble or environment.
Then, there is this data which is streaming in from any time there is an infectious outbreak. The emergency department has access to those kinds of things, and you want to swoop in and take preventative action because there is an emerging problem.
Other source of environmental data is social media. There is a lot of information through social media about, particularly when it comes to mental health and the like. This is the second trove of data which is underutilized or not utilized at the moment to address the patients’ or the population needs.
The third trove of data and this is a very rich source of data which is coming into play at the moment, is the data from the home itself. There are wearables which a person can wear from head to toe and there is a lot of information coming in through this wearable data as well as from different things in the house like your toaster, your fridge. What does a fridge say about what you are eating? These all are just coming down the pike and they’re not being used at the moment.
I look at the data that is available for providing a rich picture of the context in which a person lives, breathes, works, and carries on his daily work. So, that’s the context I’m looking at as the data troves.
Dr. Moore: I hear about Apple and other companies, and obviously a huge number of healthcare startups that are out there developing devices that capture streaming data that go up to a cloud and become useful. Is anybody actually looking to see that these data are meaningful and that we can actually improve lives? Because I have the sense that there’s just a fire hose of data coming at me as a doctor, but I think a lot of it is spurious and unimportant. It’s hard for me to parse the meaning and value of these things.
Dr. Jagannathan: That’s an interesting—in fact, you hit the nail on the head, I guess. Just this week, there is a variety of studies being initiated as well as reported on based on Apple devices. Stanford Medicine just reported a big study involving more than 400,000 patients with whom they managed to recruit in a period of eight months, which is incredible, on the heart data reported by Apple watches.
All of these people basically reported on the—they’re basically wearing Apple watches and any time they get an alert on irregular beats or AFib, they basically were notified. About half-a-percent of these people are notified and then, there was a follow-on study and about 30 percent of them were judged to actually have genuinely an AFib. This is kind of tipping the iceberg as to what is possible.
But the authors of this study note that the fact that they noticed an AFib in a 35-year-old, they don’t really know what to do with the data at the moment because normally the way you treat an AFib is giving blood thinners, but you don’t want to give blood thinners to all kinds of people. So, herein lays the problem. It is actually new data that is coming in and physicians and others don’t have the science to back up as to what exactly to do with this thing. That’s the future wave.
In fact, Apple actually released three new studies that they are initiating, one on how well hearing is impacted based on the noise level of the environment so they’re continuously monitoring the hearing. Women’s health is the other area and, of course, actively monitoring and the like to figure out how it translates to good health outcomes.
It’s just not Apple; the Fitbit and the others, they’re all in this space and they’re all working to improve the population health, so to speak.
Dr. Moore: One note of caution I want to throw in there. I apologize for being a little bit of a wet blanket with this, but there’s a rich history in healthcare of finding a new way to identify something going on and then, realizing over time that that early identification doesn’t actually lead to improved outcomes; all it does is scare a lot of people, create a lot of cost, and leave a lot of damage in its wake.
Some examples are early detection of lung cancer with chest X-rays. In the past, that has had no impact on survival rates, longevity, but a lot of people got a lot of X-rays. The kinds of things that we do when we discover a new way of seeing something then, has to go through very careful studies to answer questions like, “Does early detection improve outcomes?”
There’s a whole framework for these sort of screening healthy people that the U.S. Preventive Services Task Force puts out to say, really, you need to clear seven different hurdles before we can say this is a valid and useful test. There are very few of them that actually clear those hurdles. There are a lot of tests that have been ditched over time because they actually cause more harm than good. That’s my concern about some of these.
When I think about the number of people with AFib, how many of those would actually have a complication from taking warfarin as a blood thinner? That’s not a negligible risk for people.
Dr. Jagannathan: Absolutely. We are talking about AFib, but when you talk about genetic screening, that releases a whole new can of worms. You’re at risk for XYZ and a million other things and if you don’t know what you would do once you get the test results, is it good to have the test or not have the test? This is going to be the dilemma for a lot of people. The research has to be done and all the seven hurdles, which you talked about—I don’t know what those hurdles are. But it really makes sense to go at it systematically. There is no substitute for research, and I think what you’re seeing is the availability of this data and now, they are looking at how can you systematically go about it to improve the population health.
Dr. Moore: One of the things you mentioned that is really interesting to me is the ability of Apple to enroll 400,000 people in a study in a very short period of time. That’s just orders of magnitude greater and faster than population studies in the past and my guess is at a much lower cost than a typical enrollment. That’s fascinating and maybe opens the door to accelerating research in a way that can be very valuable.
Dr. Jagannathan: Yeah, they’ve reduced the cost of enrolling to almost nothing. You just download an app and this particular study is basically a virtual study. It’s a completely new way of doing studies; it comes with its own problems, but it’s an interesting new approach. It’ a brave new world.
Dr. Moore: Now we have a lot of startups developing devices; we have large companies weighing in like Apple and they’re being thoughtful in doing research about what do they find when they capture these data and what’s the crossover from the data to actual health impact for people. That’s good. Where do we go with that? What do we do? We started with a discussion of how you care for people in their home.
Dr. Jagannathan: I would like to highlight a particular study, which just came out this week. This was in North Carolina and it turned out that one county in North Carolina had the highest healthcare utilization in the whole of US and researchers dived in to figure out what the hell is going on.
They found out that for the entire county they had four physicians taking care of the entire population. Whenever somebody falls sick in this community, they did the only thing feasible for them, which is land up in the emergency department by calling an ambulance. You can imagine the ambulance is the Uber for this population to take them to emergency department to take care of whatever problems they had.
These researchers obviously went through this and it’s a heavily medicated population. They basically said, “Telehealth is the way to address this population. Make healthcare accessible. Four physicians is not going to cut it. Make the access available through tele visits and telehealth, particularly, access to specialists and the like.”
Telehealth is, of course, is being approved by Medicaid, Medicare and a lot of health plans around the US now, pretty much everyone, is supporting the notion of televisits. It’s not just for the Medicaid population, but for also more affluent population who can just visit the doctor from an app from their work or from their home. That’s one major role I see for taking care of almost 60-70 percent of common ailments that people face—could be handled without a visit to their clinic or even if an ED of some kind.
Dr. Moore: The more you describe the North Carolina—I haven’t read the study, but I can imagine that the four physicians is probably too few for the population in that county. Not know if that’s true or not, let’s posit that it is and, therefore, the only way to deal with surges in demands is to say, “Go to the ED,” which is a very expensive way of addressing something, which could be like an upper respiratory infection or a rash or a medication reaction. Because it may be difficult to attract more physicians to that locale, telehealth is a logical way of adding resources without forcing physicians to move to places they might not want to live.
Dr. Jagannathan: Exactly. Yeah, that becomes one of the—this is the classic social determinants of health; right? Access to primary care. What is the access to primary care? If you don’t have access to primary care, what do you do? You’re left with very little options, so you land up in the ED. That’s one of the things we need to guard against.
Certainly, that is one of the ways in which you can—one of the primary ways in which you can take care of the people at home. There are other ways, but that’s the major one, which is going to have a significant impact on the cost of being able to take care of the patients.
Dr. Moore: When I think about telehealth, though, it’s just digitizing the face-to-face interaction so that I don’t have to be physically face-to-face, but there are other ways I’m thinking of beginning to automate an interaction. What other tools are out there for maybe therapeutics or managing?
Dr. Jagannathan: Let me at least rephrase. It’s videoconference call; that’s one thing, but you can incorporate additional devices to provide more rich interaction. Suddenly, the camera can zoom in with any dermatological conditions or any kind of rash or anything like that they could probably remotely view.
And there are tools out there, which may not be available in homes, but if the televisit is facilitated through some intermediary location like a Walmart or something like that, it’s possible that they will have remote stethoscopes and, of course, you can take the temperatures and the like.
There is a category of tools and devices that will make the tele visits richer, but if you go beyond telehealth, then there is a variety of vehicles available with devices and these will fall under the classification of wellness coach or the devices can—you can have a personal coach assigned to you, which is not necessarily a telehealth visit with your physician, but might be a visit with your wellness coach who will pace you through different things that you are doing.
It could be a virtual assistant giving you guidance on what you should—helping you explaining some things. There is a range of tools out there, which are available for everybody with the new set of tools. Some of the wearables are incredible. I didn’t know that there was a sock—smart socks, which will tell you about your blood circulation and warn you to go talk to your physician because you might not be enough blood circulation in your socks.
Dr. Moore: That’s a new one for me.
Dr. Jagannathan: The number of devices which are coming here, head to toe, is just incredible.
Dr. Moore: Wow. What about wellness?
Dr. Jagannathan: The amount of wellness apps out there. There is a significant number of wellness apps; right? Of course, everybody who has an Apple watch knows about all these tracking of your number of steps you take; what’s your heartrate is throughout the day. I wear one and it will tell me, “You haven’t met your exercise goal today; you haven’t met your calorie workout goal for today.”
Those things help you keep going and there are things that track your sleep. The number of sleep apps are just unreal. They’ll monitor your sleep; they’ll tell you how well you’re sleeping. I saw this crazy wearable where it will actually monitor your EEG and whenever you’re in your rem sleep, which is when you’re supposed to be dreaming, it will flash lights and they claim that it makes you have more vivid dreams. I don’t know if that’s true or not, but that was interesting.
But the more practical aspect of monitoring your rem sleep is you can set your alarm and say, “Wake me up within this period.” You give a time, okay, let’s say six o’clock and you give a 20-minute bandwidth band, and in that 20 or 30-minute band, it will figure out when is the time you’re least in deep sleep—you’re not in deep sleep and you’re in a rem sleep and then, it will wake you up at that time. You wake up when you are not in your deep sleep; you’re woken up when you’re in your rem sleep or light sleep state.
Dr. Moore: Now, I’m split because I was going to say—you mention that there are so many of these apps. Part of the problem I have is trying to figure out which one are just noise, and which one are useful. There’s also a long history of wellness apps that have fallen by the side. One of the problems in a lot of the studies is there’s a selection bias. People who choose to use these apps tend to be active and relatively healthy anyway. When you compare them to the people who don’t use them, they look better, but that’s just who they were and not necessarily because of the app. One of those selection bias issues that has to be managed when you study it.
Dr. Jagannathan: Absolutely, and this was actually mentioned in the study with the Apple watch. They were basically saying that the mean income of people who have these Apple watches are something like 80K, and a mean income of an android watch is in the 60K region. So, when you start doing these kinds of studies, it’s going to put you in a different population than the ones which involve the social determinants of health and the Medicaid population. You may need a completely different set of strategies there than the ones who go around wearing Apple watches.
Dr. Moore: Exactly. Then, when you mention don’t wake me up in the deep sleep; that sounds attractive. That’s just an awful feeling when the alarm goes off during a deep sleep.
Dr. Jagannathan: I believe that’s a useful app. I’m tempted to try it myself.
Dr. Moore: All of a sudden—I’ve been poo-pooing them for a long time and now, I’m thinking, “That’s not a bad idea.” So, what do you do with all this information? How do you put together healthcare delivery in a way that takes advantage of this?
Dr. Jagannathan: First, off the bat, I mentioned that we have three troves of data and the only data that we currently have is the thing, which is recorded in the EHR. The next trove of data, which I think people are trying to attack is the SDOH data, particularly because there is a lot of Medicaid population out there and there is a genuine desire to bring them into a population health value-based care environment and track it.
We now have some ICD-10 codes related to the social determinants of health and once they start coding those kinds of things when these people show up in clinics and hospitals, perhaps you’ll have a better picture of the social status and the social determinants of health recorded in the EHR and might form the basis.
There are other ways of getting this data and some companies have been successful in accessing this data. Suddenly, the social services have a record of who they are taking care of and just by virtually combining the social services and physician services in one environment, in one dataset, you suddenly have access to the data, which is much richer than before. In a number of places that’s been happening, the combining the social services and, for instance, PA has announced that they will try to coordinate these physician services and social services in some cohesive way. That’s going to be good for the Medicaid population of that state.
Dr. Moore: You mean P-A, as in Pennsylvania?
Dr. Jagannathan: Yes, P-A, as in Pennsylvania. That’s one trove of data and the wearable data, suddenly, a lot of it can flow back into an EHR and in a limited way, some of it’s already flowing in. People who opt to provide the data—how many steps they take in a day—whatever. That whole slew of that dataset can go into an EHR and some of it is going through Google APIs and Apple APIs and they are already aggressively promoting their APIs. I saw some study somewhere that EPIC has managed to collect a bunch of this information.
It’s still at an information stage and it’s not something actionable until you start standardizing the data format and running analytics on it and use that to do some prediction about the health status of the individual and determine whether you can actually make any actionable intervention.
We are still in the phase of data collection at the moment. Sporadically, some of these are coming in, but it’s still not something any physician would want to see. It has to be more fed into some machine running other than to figure out what in the world is happening. We are still in the data collection phase, leave alone the analysis phase and intervention phase.
Dr. Moore: Yeah, I’d say that makes sense to me. As I think about some of the things that we have the capacity now to be able to say, from lots of well-validated studies, we can see that people with very high risk—you had earlier mentioned that risk stratification is one of the things that’s important in terms of understanding data.
So, using current, even administrative datasets, we can say, “This group of people exhibits attributes of extraordinary risk of being hospitalized or ending up in the emergency department or having bad outcomes.” Therefore, we can think about high touch interventions: telehealth with care manager outreach and things that you mentioned before so these tools become an enabling strategy. I think there are good studies that validate that approach.
In the meantime, we can then step back and say, as we have the capacity to risk adjust populations and look at outcomes like the rate at which people end up hospitalized for potentially preventable reasons.
You had mentioned pollution, so we can see when there’s more pollution, do we see more children with asthma being admitted to the the emergency department of the hospital. We can begin to see if there’s an app or if there’s an approach that sets up alerts, does it change that rate? Does it actually have an impact? Because what we want is better outcomes, but we have to be careful not just to add complexity and cost to an overly costly system and overly complex system that we have now.
Dr. Jagannathan: That’s a good point. Sometimes, the preventive strategies that you’re looking at may not necessarily be cost-effective in the short term; but if it helps the population to be healthier and happier, that is an end goal itself and it’s going to improve their work productivity; employers are going to be happier. The payer may not necessarily see some of the benefits sometime in the short run, but in the long run, it’s going to make a difference. I remember just seeing a study on that front as well. Thus, focusing on a population health really, does it really help? The answer is “yes,” but it will take some time.
Dr. Moore: Well, Juggy, thank you so much for your time today.
Dr. Jagannathan: Thank you. Enjoyed it.
Dr. Moore: I loved the good conversation and so, I look forward to more.
Dr. Jagannathan: Thank you.