COVID-19 changed the world in a lot of ways. It changed the way that we interact with our doctors and it changed the way that doctors make decisions. Angelo is joined today by Jim Shalaby, a friend of his, a clinical informaticist, and a pharmacist.
Angelo begins this episode with a few questions about the changes caused by COVID-19, specifically around the patient data gathering, such as blood pressure. With telemedicine practice, how reliable is the data, who is legally responsible for the accuracy of the data gathered, and how exactly do clinical decision support (CDS) tools adjust with this new change in a traditional clinician workflow?
Angelo explores more on the topic of IoT devices and the data brought into medical decisions. Again, how accurate is the data from these IoT devices, such as Fitbit scales, that a clinician can diagnose and treat from?
Jim brings up some of the challenges that came with telemedicine such as workflow within a clinic. If the clinician seeing a patient wants the dietitian to speak with the patient, it is more of a challenge to coordinate than being within a few feet of each other. The other challenge relates to security policy and considerations patients need to agree to regard their personal privacy. To get into a virtual visit with a clinician, a patient has to follow the security protocol that provides a barrier for some elderly and disabled patients. Lastly, the challenge of all this data a patient could be collecting in their IoT devices is, how do you move that data into the EHR or in some format a CDS tool could ingest?
With the use of CDS, machine learning, and AI, the future is ripe for opportunity.
Further Reading
What is CDS - Health Gov IT
ResearchGate Publication on IoT in Health Care
Privacy-Preserving Single Decision Tree
Jim Shalaby on Twitter and LinkedIn
Angelo: COVID-19 changed the world in a lot of ways, it changed the way that we interact with our doctors and it changed the way that doctors make decisions. I'm joined today by Jim Shal aby, a friend of mine and a clinical informaticist and pharmacist.
I'm your host, Angelo Kastroulis and this is Counting Sand.
Jim describes himself in Jim, I'm speaking for you a little bit here, as a clinical informaticist because you seek to find ways to apply different standards and different ways of being able to reach inter-operability and practical workflow. And you've been doing that for 28 years. I've had the pleasure of working with you for many years, and we've got to work on many interesting projects together.
And the first thing we're going to talk about today are the ways in which COVID-19 has changed the way we interact with our doctors, doctors have always had a need to monitor patients. Physicians are able to say, collect their blood pressure before COVID 19 and would call us into the office. They would measure our blood pressure and they'd be able to make decisions based on the reading.
We're going to talk about decisions today too, but one of the reasons why they were weary of using that data that I would collect myself is that it's hard to know if you can trust the data. Am I educated enough to know how to use the device? Am I measuring it properly? Am I giving you the right reading? Can you trust what I told you? Those kinds of things, but there's also a second factor that comes into play.
Jim, can you talk about that a little bit?
Jim: Number two, is I'm responsible for legally. If a patient gives me a blood pressure values every week, you know, from their device. And it's saying that they are critically. And, you know, I have to account for that. I have to explain it and it's a lot more work that, and then I need to do, if I just had them come in and get their blood pressure measured.
Um, so, so there was a bit of resistance both from, from those perspectives. Give me more data. I have to do more work. And I can't trust the data because it came from the patient, but. With, um, with COVID it kind of forced the issue, right? So, now you know, clinicians, whether it's the physicians or the nurses following up with a patients, they basically, had to figure out creative ways to be able to get important information out of a patient and, and factoring in that patients can be optimistic when they report their values.
You know, I, my blood pressure is great. I'm at one 20 over 80, but you know, they're taking 15 readings and they waited until it was the lowest reading and they report that or they help hold her arm up and take the blood pressure in this position. And, you know, the blood pressure is really low
with tele-health and tele-visits things changed, I think clinicians and patients learned. Um, more effectively, but, having, the ability to video conference and be able to tell a patient, okay, show me how you measure your blood pressure or, show me your glucometer so I can see what your readings were.
Um, simple things like that really kind of. Less than that anxiety on the clinical side as to whether they're getting reliable data or whether they're going to be legally responsible for, for monitoring that data on a regular basis, it wasn't as bad as they thought.
Angelo: Yes. So remote work or remote telemedicine was one of those things that we've always had a little bit of resistance to. Cause we were worried about having a disconnect from our clinician. But what you're saying is that we learned a little bit how to use this to kind of technology to kind of improve the way that we do healthcare.
And there's a lot of positive benefits, right?
Jim: It doesn't substitute for face-to-face, but boy that's convenient. Um, that, that level of accessibility, of the patient to the, you know, the patient having access to their providers, certainly, they don't have to wait in a waiting room.
They don't have to find transportation to get there. They have almost immediate access to their clinician without going through the waiting room issue. the clinicians also liked it because if they had a quick question, if they needed a quick followup, they knew they could actually engage with.
Without having to play phone tag, they can schedule something and do it. So, I think it it's now become a part of healthcare. COVID has really changed the way patients. And the clinicians. And the reason I say clinicians is that I don't just include physicians. I include the whole allied staff that supports a physician, the nurse practitioner, the physician assistant, the pharmacist, the dietician , even the case manager and the social worker are using the same modalities to communicate with the patient and they've all gotten into a rhythm. It's not perfect, but they've incorporated into their daily workflows, which is really promising, it opens up new new avenues for improving health and, and really the primary thing is it allows a new way to, assess outcomes of interventions because, before.
You really didn't know what the outcome of your intervention until a year later or six months later. And you really had no idea whether the patient was taking the mat or where are they exercising or were they monitoring their blood pressures correctly or taking their weights they can do something as simple as point holder device up to the camera, they can, uh, show here's how I'm measuring my blood pressure. This is how I do it every day.
Angelo: One of the things that I find most exciting about, uh, the way the telemedicine has changed with COVID. Is that just like what you're talking about, Jim, these devices now are being incorporated into our workflow and there are tons of, uh, devices and apps and data coming from this, and that can affect.
Decision support one of my favorite topics, as you know, but, um, the thing I wanted to say about decision support is it is really hard to be able to get that decision support to the clinician at the point of care. Don't you think?
Jim: So you're now remote from the patient. Um, decision support typically has been, you know, not very good from a clinical workflow perspective, from a usability perspective, it's been more noise and value. And I think there may be a new way to look at decision support that, didn't exist before and somewhat catalyzed by, this tele-health and tele-monitoring patient engagement.
The traditional decision support where alerts have popped up at the most inconvenient times in the workflow to the physician usually, or to the pharmacist, rarely to the nurse. We, weren't taking advantage of allied staff that could actually do something about this, this decision support way upstream, if you're getting decision support about a patient and an alert pops up to a physician when they're closing their encounter after a visit. You have a drug drug interaction between this blood pressure pill and this anti-histamine change to another drug that might be too late because the patient's gone home. The physician's now documenting, they're already prescribed the drug. Now they're going to have to go back and cancel that prescription, contact the patient and start over again, basically.
But now you have a lot of helpers, who are actually doing the outreach to the patient, and they're doing things like checking rosters of patients. They have a ticker list that says today I have three congestive heart failure patients that I need to follow up on.
I have no idea how they're doing. I'm going to call each one of them and see, have they gained more than five pounds a week? And if they have that, I know they're getting fluid overloaded, and I need to prioritize, getting them on a diuretic or something, get a nurse out there. So they don't end up in the hospital next week.
Angelo: I talked about a similar idea in past episodes where maybe with machine learning, for example, we're a bit too ambitious in what we're trying to do when we could actually be providing more value, doing something a little bit less ambitious. And this is a great example instead of providing alerts, maybe to the physician.
I'm not saying that sometimes we don't need to, at the point when you're prescribing, for example, if something is going to, uh, interact with something else, you probably should know that, but. An easier and better method to implement decision support would be, um, on these lists. For example, that's pretty low hanging fruit decision support could just like you're saying, prioritize these lists and, or, or even have some predictive value so that we can make sure that, that they get called in the right order.
For example, based on, on some criteria.
But another thing I think that is important to talk about in decision support is that typically clinical decision support relies on clinical data. In other words, the data inside of my electronic health record, but there's a whole lot more data at home that I'm wearing a Fitbit, or I have a an Apple watch or my phone is a pedometer or I have a blood pressure meter or a scale.
How can those be helpful in decision support?
Jim: Okay. What if decision support can cure.
Um, what if that remote trending can show trends, not to the physician, but to the physicians helper.
That's going down that roster, almost a thumbnail. Okay. And says, look, here are the list of patients and here's just a simple graph of their blood pressure. These three patients at the top are at the top because we've detected that their blood pressures are above 1 50 over 90. Okay.
And so we'll ask you to call those least these two CHF patients have had their Fitbit scale report back that they've had that two pounds gain over the last three days. There were five pound games over the last week. Well, now it's a whole different world because in the workflow you're not disrupting their clinical work.
You're redirecting the decision support to the physician's helper. Who's naturally plugged into that part is, and it's very open and wanting to see this. This helps them. And they also have more time than the clinician who has five minutes to see a patient and then move on to the next one. Well, what if we also added to decision support the ability to have, outreach to the patient in an automated way, perhaps they have a phone that can receive texts and those texts could be know, I'm trying to reach you. It looks like your blood pressure is elevated.
Could call me when you have a chance compared to calling each patient and leaving a message. And by the time you get to the critical patient, you're playing phone tag with them at 5:30. Okay.
And so that allows the system to support a nurse or a clinic care coordinator, to work in parallel.
Angelo: Great thought and typical decision support systems could work in that fashion, but this might even open up opportunities for machine learning. Right? Machine learning is actually really difficult to implement in a care setting, but in other settings it can be really, really helpful. What do you think Jim?
Jim: One of the challenges of a learning health system is that when you're training a system and you have a very frustrated user who only has a few minutes, they're not the best ones to be asking, to train the system.
They, they just they're too busy and their confidence level decreases because during the training process, the alerts will be a lot of times nonsensical. So they go into it with skepticism because for the last 30 years decision support, wasn't a learning system. It was a deaf system, why is it giving me this alert over and over and over again.
But if we shifted the trainings in the workflow, the ally team, the care coordinator, the nurse, the PA who are amenable to this and they have time to provide useful information. They interact with the decision support system, say, yes, I've received patient reported outcomes or patient reported data, and I've assessed it and it's relevant or I've overwritten it.
The learning system is getting useful feedback. The user is not getting frustrated. They're seeing the system improve over time because you've already given them some value. You've given them a prioritized list of instead of a random list of. To call. And then as you reach a point where if your learning system is detecting a significant decrease in the number of overrides and the number of, dismissals of alerts, it may be a tipping point to then surface alerts proactively at the point of care to the physician.
But now, it's almost like the workflow of a triage nurse or a allied staff was a pipeline. That was that I was on the quarter inch in diameter after COVID and became 10 feet in diameter. That most of the work went through those pipelines, not through face-to-face visits.
That shift kind of raised awareness and also offered that opened up. I think opera a lot of opportunities for, for direct and indirect benefits to.
Angelo: One of the other things I always thought was lacking in healthcare is that the patient is not really in the conversation. The payer has always been there and the person, the one kind of paying the bills, although you can argue the patients pay the bills and, um, providers, the one doing the care, have a conversation, and I'm just kind of a, uh, outsider listening to.
Um, do you think COVID has kind of helped to include the patient in the conversation? You know, help us be in center while the payers and providers, aren't kind of just talking in our behalf.
Jim: Yeah, I think it's helped it a lot. It's definitely helped a lot because, it was more a psychological problem in my mind for acceptance. And it was a technical problem. I mean, there definitely are technical issues, but the psychological piece was, there was a lot of resistance before.
To opening up the flood gates, so to speak, you know, they, they felt you're going to open up a flood gate of data and confusion, and I already am barely keeping up with the EMR. But because it came about it forced a kind of a psychological shift in how clinicians think, you know, it's okay.
I have no choice because I COVID is real. I have no choice, but to work through telehealth, and that's much better than not having contact with my patient. In fact, they embraced it because they said, well, when I first came out of medical school, I didn't have this.
If we had COVID back then my patients would be at risk. Now, now with this technology, I have access to my patients. It's really nice. And I have almost immediate access to my patients.
It wasn't perfect. So what's happened with tele-health and televisits and, and patient engagement is that, um, there's also some things that have to be worked out.
Angelo: Okay. So let's talk about that for a second. This isn't perfect. And it's going to introduce all kinds of new problems. In fact, solutions always, uh, introduced new problems. In fact, I call it like, uh, you pull a lever and the lever has benefits and an unintended side effects. So let's talk about that a little bit.
What are some of the challenges we're facing now?
Jim: Okay. Right. The thing about having a patient come into an office is that the clinician had support staff around them. So I, you know, you are homeless, you have social determinants of health. I'm not an expert at that as a clinician, but Yeah.
Let me just holler across the room to the social worker and she can work with you or he can work with you.
When you're doing your tele-health, you don't have a room to holler era, across. So the clinician also has to take on responsibilities that they typically could delegate very easily because physically they were 10 feet, five feet or one door. From the person they need to send the patient to see the dietician, see the physical therapist, uh, the social worker, you have a problem with finance. See the finance folks, they were all within physically within reach. And so it's very easy to delegate that that workflow for the physician becomes trickier actually for all the clinical staff participating remotely, it becomes a lot trickier because, um, they all have.
That allied support to be represented in a virtual office rather than they a physical clinical office. And that hasn't happened yet. It's, it's something that I think needs to be addressed. It's a gap in the workflow, but it can, it's a gap that I think can also be addressed as a process that gets refined the overall impact of telehealth and remote patient monitoring, has been pretty positive.
Angelo: What about other issues like the technical savviness of patients, or maybe losing that human connection?
Jim: So that's been kind of one of the other challenges. Because the patients who are sick as tend to not be very it savvy. And even the patients who are savvy is, are very sick.
They're, they're gonna tend to be a little bit afraid of trying to connect. You know, they want the real connection. They don't want to spend time tinkering with a computer or a phone. There's still a lot of room for improvement to simplify the engagement of patients, especially elderly patients or disabled patients.
But, that type of improvement is really, probably more around the process. Yeah, And lowering some of the barriers I used to think the barriers were, knowing how to log into zoom or knowing how to log into, you know, ethic, portal my chart. But it actually, that isn't really the barrier.
I mean, that's part of the barrier, but it's not the major part of the barrier. The, the largest part of the barrier, which got handled very quickly are all of the security hurdles that the patient also has to jump through. Yes, I am from Shelby. Yes. I do give permission to get into this video.
Yes. I acknowledged these 10 questions, even though I don't understand them just so I can spend five minutes talking to my physician or my nurse. So it also opened up the industry's eyes to, see a balance between practical implementation of privacy versus impractical or unusable implementation of privacy.
Angelo: That's an interesting point that the technology shifted a little bit, learning how to log in to your patient portal versus learning how to log into zoom. A zoom has become ubiquitous or those kinds of technologies, but still learning in how to log into your patient portal is extremely difficult and it's different for each EMR.
So I actually like that shift. I think it's been positive and better. You mentioned the privacy. I think that's really important HIPAA, which is designed to protect our privacy was not really intended to be a barrier. It was intended not to stop our care, but to protect us. And it wasn't designed for an environment like this, where we are now having to use telemedicine and move data around and help us.
And instead, now we have all these barriers put in front of us. What do you think about that?
Jim: You know, that the HIPAA was never designed to make that kind of a barrier exist, but it was interpreted and implemented in ways that created barriers, especially for tele-health. Now, it's interesting, the solutions we're racing. Clinicians and patients, are a lot smarter than we may think, you know?
I want to share with you, my data. I have no way of sending it to you through some patient portal, but they'll hold it up to the camera and here's my trend. I'm scrolling through it. And the physician just documents he'll transfer last week is that the blood pressure is within normal range or outside of normal range sometimes.
These are clues for being able to simplify, interoperability to allow easy access to that information without having those hurdles was barriers,
Angelo: We've talked about the places that technology can improve, but technology has also made a measurably good impact. We talked about the fact that these applications and devices are now proliferating, and so, patients have access to a whole lot of really good, incredible information. Now, another thing I want to talk about, um, we did mention that.
Teleconferencing technology is a positive thing because it eliminates this. Uh, I have to wait an hour to meet my doctor and spend maybe less than five minutes talking to them and then hustled through the whole process. Uh, instead now we get to have meaningful interactions, but the thing I wanted to talk about a little bit is.
That this proliferation of new technology and this new way of communicating is actually helping us because we're eliminating bias in the system. Because if, if the clinician is only making decisions on data, that's found in your health record, that's biased data because obviously someone had to decide what goes in the record and you're missing out on all this other data that is extremely valid.
Uh, don't you think Jim.
Jim: Yeah. Yeah, I think you're right. I think you're absolutely right. It's also, again, the psychological impact has been interesting too, because on the provider side, as well as on the patient side, I now have to make time to be able to communicate, remotely with you, but, in a way they almost seem more relaxed.
It's not 10 people waiting at my door, waiting for me to sign off on orders and labs and. They may be doing that virtually on my computer, but they're not doing that physically, which adds another level of stress. The clinician's a little more relaxed when they're talking to the patient.
It's, it's interesting. Cause we all think about the patient needing to be relaxed talking to the clinician, but there's just as much anxiety and other way around, it's a clinician trying to try to appear as their patient and listening to the patient. When they know in the back of their minds, they have a to-do list that's long with, people waiting for them to respond to things.
You know, we're missing something in interoperability. The other interesting thing is looking at data, patient visits and comparing, patient reported outcomes to, clinically, computed observational outcomes.
And that's a really interesting area that has never, never been available until COVID .
Because outcomes almost where all is computed through. A bias perspective and that's just a clinical, electronic health record perspective because that's where all the data resided
it's a lot of that data comes in through interviewing the patient, talking to them or, or collecting information, but it's still, it's still interpreted through, a somebody's under pressure, usually not with a whole lot of time and trying to say, yes, the patient's pain is country.
But as the patient's pain really controlled after hip surgery, however, if you look at that patient reported outcomes, now with telehealth there's giving feedback through their patient portal or through responses to text surveys .
Those types of, patient engagement, apps and patient engagement technologies really started to, Expand, during COVID , and be used. I mean, they've always been around, but the willingness of clinicians to use these apps more because they now didn't have direct access to the patient, really kind of increased with COVID.
Angelo: Jim. I really feel that advanced technology like machine learning and artificial intelligence really has a place in, uh, in improving our healthcare. Uh, what do you think.
Jim: You can do all the prescriptions you want. You can do all the calls you want, but if you can't verify that the patient is actually taking active, a part in improving their. Including something as simple as swallowing that pill every day, then you'll never reach outcomes, you know? And so the barriers to improvement are frequently not, you know, the complex clinical, intervention.
They're, they're really just human factors. So I think that's an opportunity for it to really move in the right direction. But, I think what would be really interesting. Is to have, better monitoring and comparison or contrast of what patient engagement is doing from outcomes perspective, compared to, what our traditional way of assessing outcomes and then adjusting, almost like adding handicaps in a golf game, or in a sailboat race observing the differences between those outcomes and adjusting, to a real outcome.
And from a learning health perspective and from a machine learning perspective, that kind of information could be extremely useful, instead of, analyzing this manually certainly is not practical cause a huge amount of data.
But applying, machine learning algorithms with significant supervision, to see what kind of differences and what kind of patterns we're detecting, can really impact those chronic conditions very quickly.
We're doing some work right now with, symptom management for cancer. That's, that's another one that's really amenable to remote monitoring because with or without COVID, a lot of cancer patients at some point are not very mobile. They’re very sick.
They have to be monitored remotely and to be able to collect their symptoms be able to identify opportunities for helping them control their pain or their nausea, or the fact that they now have a fever. And that's a big Sentinel flag for potential infection after chemo and, being able to early detect and do something with that using technology can, can have a huge impact on outcomes.
Angelo: And that is an area that's really ripe for further study and further improvement COVID certainly has helped us learn many lessons but I think also the modernization of our society is making clinical decision support extremely interesting. And I think we're going to have an evolution. It's going to change a bit and it's not going to be the same kind of decision support at the point of care, there's going to be all sorts of new kinds of clinical decision support. I'm really excited about seeing where the future goes. Jim, I want to thank you today for joining us. That's all the time we have. And I look forward to chatting with you again. You're also genomics expert and I can't wait to dig into that.
And I also want to thank the listeners for joining us today.
I'm Angelo Kastroulis. And this has been Counting Sand. Please take a minute to follow rate and review the show on your favorite podcast platform so that others can find us. Thank you so much for listening.