Leveraging Digital Technologies to Improve Patient Health - Health Talks

Leveraging Digital Technologies to Improve Patient Health - Health Talks

Show Video

[MUSIC] I'm David Brenner, the Vice-Chancellor of Health Sciences at UC San Diego. It's my pleasure to welcome you to one of our health talks. The title of today's talk is Leveraging Digital Technologies To Improve Patient Health. We've all noticed that there's so many amazing digital technologies available and the question I would ask our panel to address today is, how can we use this amazing technology to improve health care? How can we figure out whether a tool is useful or not useful in taking better care of patients? To address this, we have established a new center for health innovations at UC San Diego Health.

I'd like to thank specifically, Irwin Jacobs and Paul Citron who are the co-chairs of our science, technology, and Global Initiatives Committee. They really champion this. They really challenged us to take what we know about digital instruments and see how we can apply it to improving health care. I also want to take this opportunity to thank Joan and Irwin Jacobs. They provided the very generous gift to create the Joan and Irwin Jacobs chancellors endowed chair in digital health innovation, the first such endowed chair ever created. My colleague, Dr. Kevin Patrick,

will help guide us through today's meeting and introduce each panelist as they present. Dr. Patrick is a real thought leader in this field, a practicing physician, who is a professor at UC San Diego, and internationally known expert in digital health. He, in collaboration with the health system, really spearheaded the efforts to establish our new center. I would like to kick this off by asking Paul Citron to say a few words.

He's the Co-Chair of the Science, Technology, and Global Initiatives Committee. He was a leader at Medtronic's for the most advanced medical devices company, and he was Vice President, Technology Policy and Academic Relations. He, now, is an active faculty affiliate at Jacobs School of Engineering. Paul, please say a few words.

Thanks for the introduction. As the Co-Chair of the Science, Technology, and Global Initiatives Committee of the Hoff Board of Advisors, I'm pleased to have the opportunity to join today's health talk. My professional career as was touched on a moment ago began in 1972 at Medtronic incorporated in Minneapolis. Armed with a newly minted master's degree, I began as a biomedical research engineer in the cardiac pacemaker division.

I retired from the company in 2005, having been its Vice President of science and technology. This was a period when innovative technologies began a transformation of healthcare from the kindly solely practitioner stereotype to one where physicians, scientists, and engineers collaborate to find ways to restore seriously ill patients to fuller and healthier lives. We've been very successful in large measure through the leadership and vision of institutions like UC San Diego that are transforming healthcare for the better. As you will hear from today's speakers, healthcare is at an exciting inflection point. Today's speakers will share with you how digital technologies are reshaping the practice of medicine and how we as patients are the beneficiaries.

Let's consider three examples. Wearable and implanted monitors now collect important real-time physiological data and turn data into useful and actionable information for physicians. Versions of these remote monitoring systems also detect infrequent, but potentially serious events that require prompt attention. It was not long ago that ambulatory monitoring was seen by some as engineers gone wild. Today, continuous monitors are routinely used to guide and improve important syrupy scenarios such as cardiology, neurology, and diabetes. New applications are rapidly emerging.

Another example, artificial intelligence, will increasingly be applied to a collection of clinical data and treatment outcomes. These data will be used to tease out optimal treatment strategies for the patient at hand. Think about it, millions of patient outcomes from around the world will be analyzed to inform the preferred treatment strategy for each new patient. The third example is enhanced remote capabilities was served to make healthcare more inclusive, accessible, efficient, and equitable.

For instance, the waiting room experience will deemphasize the waiting component and emphasize productive patient and healthcare provider interactions. In many instances, the home will replace the traditional waiting room. I'm excited about UC San Diego's health commitment to leading the way in leveraging digital technologies to improve health care. Their physicians, scientists, and engineers are committed to keeping the patients' interests front and center in this endeavor.

I hope you will consider joining me in providing support to the Center for Health Innovation. With that let me now turn the program over to Dr. Kevin Patrick. Thanks, Paul. We're really grateful to the leadership that you and Dr. Jacobs have provided to this initiative. Thank you, David, for your ongoing support as well.

It's been really terrific. We are excited today to feature some of the initial efforts in digital health that our new Center for Health Innovation at UC San Diego Health is doing. With seed funding from both Dr. Brenner and from Patty Mason at the CEO of UC San Diego Health, we've been up and running now for several months, but this is our first public event, so you are witnessing a dawn of an era. You see here our vision and our mission.

Today's panel is going to emphasize core principles that we're focusing on. But our vision is to become a world leader in the development and implementation of innovative health technologies to improve people's lives. The way we plan to do this is to develop a very thriving and dynamic health technology ecosystem that enables innovation at scale with impact through testing here and beyond. Today's panel is going to emphasize our core principles and one of the most important is that our center is not simply focused on technology. Rather it is based solidly on the needs of patients and their families, and to the extent possible, we are trying to do everything we can to include patients and their families in the design of what it is that we offer. You'll hear more about that today.

In addition to our panelists today, this effort is the work of many. Our center steering committee is this group of people that you see here, Nicole May and Jeffrey Pan are co-directors, they were drawn from their positions as key program managers at UC San Diego Health. Chris Longhurst is our Chief Medical Officer and Chief Digital Health Officer, that's the first position that we've had here at UCSD. Parag Agnihotri is the Chief Medical Officer for our population health services group.

Josh Glandorf is Chief Information Officer and Eli Spencer, like myself, is a physician who has been doing research in related areas to digital health for quite some time. The two areas of effort that we're going to focus on today are our two initial pilot projects in this first phase. Project 1,000, is focusing on scaling up rapidly in the face of COVID and a variety of other challenges, remote patient monitoring. The theme of it is patients at home, working with healthcare providers to make health care decisions using these new digital interventions. Again, you'll be hearing about this in a bit. The second, Paul alluded to this, is this concept of a data used for machine learning and artificial intelligence, leveraging those data from our electronic medical record to provide unique insights into the prevention treatment disease.

This really is a remarkable new era that allows us to do unprecedented things. Many of us been dreaming about this for a long time, but now we can actually do this. Without further ado, our first panelist today is Dr. Kristen Kulasa.

Kristen is a clinical professor of medicine at UC San Diego School of Medicine and Director of inpatient glycaemic control at UC San Diego Health. Kristen has been on almost all of our meetings since the inception of this project because she's got such wonderful real world experience in her world of diabetes, of what patients are confronting. Among our other efforts, she's quite active in a group called the Society of Hospital Medicine and as a mentor in that particular program. She really knows of what she talks, and so Kristen [NOISE] have at it. Great, thank you for having me. Today we're going to be talking about bridging the technology gap and patient-centered diabetes care.

In diabetes, we have loads of technology. We've got personal blood glucose meters, we've got small ones, we've got large ones. We've got ones you can talk to.

We've got ones that we'll talk to you. We've got Bluetooth meters, we've got meters that'll tell you if your reading at that very moment is above target, below target, in target. We've got continuous glucose monitors now where patients don't even need to ( - ) their finger anymore. In this particular one, you just wear it on the back of your arm, a sensor is detecting your blood sugar every five-minutes.

Rather than poking your finger, you can swipe and it'll give you your blood sugar reading and the direction of change, plus you'll be able to see the track and trend your blood sugars that it's been monitoring. We've got continuous glucose monitors that you don't even need to swipe, they just beam that information to your watch, your phone, your iPad. We've got technology to were loved ones can follow this data for you, so it works great for children and elderly. The patient can be monitoring it and a follower can be watching it as well. We've got sensors that are implantable, but you don't even have to change every ten to 14 days. We've got technology there with insulin pens will talk to your phone, talk to your continuous glucose monitor.

We've got insulin pumps with tubing, we've got insulin pumps without tubing, and we've got insulin pumps now that put all of this together with software, speak to the continuous glucose monitor and make little changes without the patient even being involved. As you can see, diabetes technology is vast and it has been able to improve patient outcomes significantly over the last many years. We've got loads of data with all of this technology. We've even got standardized views of this data in our ambulatory glucose profile. This report right here, which is an easy way to look at all the different technologies using the same language.

Basically, it's helpful for patients, we can talk about their time and range. We can look here at some particular point of time and day. That might be their weak points. In addition to this, we have lots of other reports. We can look at day modal views where you look at every single day spaghetti graph on top of each other, you can look at individual days, you can look at weeks at a time, certain time points. We've even got consensus on what does data should be to improve clinical outcomes.

As you can see, with all this technology and all this data, we've come a long way for this chronic disease. One of the things and biggest barriers that we have, is putting all of this together for the time effect or to make it very efficient appointment for the patient. Taking all this data from all the different technology and devices out there, putting it into one single report which we have. But then being able to serve it on a silver platter for the provider, to be able to utilize it during the appointment, whether it be a tele-visit or an in-person visit. Then at the click of a button, being able to integrate it in with the EHR as well.

This is what we're lacking, we're lacking some sort of platform to integrate every single piece of diabetes technology out there. We have very good reports, they can always be improved on. But really bringing all that together in an efficient manner at the time of the appointment and integrating it with the medical record would be ideal. Right now for patients who can't share their data, we have to be very creative during a tele-visit, have them hold their logbook up to the video and try to see the data that way.

Maybe perhaps showing us their meter, whatever data we can get to make the best decisions and make it as easy as possible for both the patient and the provider. But even then if we do get the data, it's hard to integrate with the EHR, so right now we have to copy and paste a lot of these reports to put them into our notes. There's no discrete data that's able to track and trend this within the EHR. In diabetes, we've had patient-centered management of diabetes for a long time. This is a chronic self-managed disease.

The patient is absolutely at the center of it, and we do for management decisions, take all characteristics and shared management with the patient. We talk about comorbidities, we talked about barriers in their social life and their finances, what's going on in their lifestyle? Really it's a negotiation every single visit of how can I help you achieve the best outcomes. It should be the same way with all of this technology and utilizing it, the patient needs to be at the center as well as the provider. How can we make this easier for the patient and bring all their devices? They're already doing so much work on a day-to-day basis to manage their diabetes for them to have to take an extra 2, 3, 4 or even five steps to connect their device with the clinic so that we can utilize that data during their appointment. It should be simple that data should flow passively. Then what about between visits? If we had a command center here looking at that data that they're working so hard to collect, and even being able to help them and reach out via phone or text if something comes up, and have some shared metrics with the patient on when they need to be alerted, really trying to break down this chronic disease into achievable targets.

This is exactly what we're doing in Project 1000, is trying to develop a pilot for diabetes care and really utilizing this remote patient monitoring for blood sugars, and creating this command center to be able to utilize with the patient. With that, I'll conclude and turn it over to Dr. Patrick for our next speaker to talk more about Project 1000. Great. Thank you, Kristen.

Our next panelist is going to be Dr. Ming Tai-Saele. She's Professor and Vice Chair of the Department of Family Medicine in our School of Medicine and Director of Outcomes Analytics at UC San Diego Health. Dr. Saele has worked with leading user-centered design firms such as IDO on patient experience improvement projects, and is among the first researchers to use video and audio recordings to study how doctors and patients allocate and share their time within clinical encounters. Her research teams have created multiple tools for enhancing patient-centered communication. She really is a world leader in this area of patient-centered research.

So Ming, please go ahead. Thank you. I will talk about how UC San Diego Health is advancing knowledge in digital health with patient-centered approaches. We want to understand what matters most to patients, family caregivers, and health care team, in taking care of patients in the context of a learning health system. We systematically gather and create evidence in forms of data, qualitative or quantitative, and we apply the most promising evidence to improve care in what we do in practice. With this virtual learning cycle, we try to serve our patients.

I'd like you to think about your last encounter with a clinician or your loved ones. Were you able to get all the things that are important to you discussed or the most important things that you want to discuss? We've heard previously from patients at another organization that they had a hard time doing that. They said, "You have to be really articulate and you start the conversation, otherwise you may forget what you wanted to talk about." From physicians, we heard that they may not know why the patient came until they went into the exam room. We set how to improve patient-physician communication in clinical encounters with funding from the Patient-Centered Outcomes Research Institute. With over 5,000 patients in three systems, we created this pre-visit questionnaire for patients to actually write out what's the most important thing for their visit.

Then physicians can import that with a smart tool in the electronic health record directly enter their progress notes. Our physicians had told us that many patients are slightly anxious for their first video visit. This is during the COVID, so having a list to review made it easier to get started. Another physician said the patient list is something that they did not bring up, so I was able to ask them about it.

This tool is now been adopted by the Epic Electronic Health Record system platform-wide for all e-checkins and we're really happy about that. We've published this study in the Journal of Medical Internet research. That's about patients in clinics before they come to clinics. We know lots of patients spend most of their time at home, so how about our patients who are at home and with chronic conditions? This is what the Project 1000. We have digital health for patients with hypertension, for patients with diabetes to monitor their blood pressure, to monitor their glucose, wirelessly use Bluetooth enabled devices that are provided to our patients without additional charge, and we have a multi-disciplinary team serving those patients.

We have worked with qualitative researchers to learn from our patients and care team members, they did shadowing of care teams for a day, seen how they work, how they help patients with initial setups and how they go to have home visits, and they did in-depth interview with individual patients with different conditions living in different areas with different levels of resources, age, and gender. We've learned that patients who were active in the program and take readings regularly, feel like it has had a positive impact on their health. Now it's become a part of their daily routine to take care of their health. They said that knowing that someone is monitoring their readings make them feel like someone actually cares at UC San Diego Health. One very poignant example I want to share with you, let's call this Patient N, she has hypertension and deep vein thrombosis, and she had a pulmonary embolism that was surgically removed. She was in the process of establishing care with primary care and she was referred to digital health.

Our digital health specialist made a home visit, saw that her systolic blood pressure was 200, diastolic blood pressure was 100, they suggested hypertensive crisis, she also reported chest pain, headache, dizziness, and shortness of breath with exertion. The digital health specialist was a former emergency medicine technician in EMT, so he conversed with our registered nurse on the Digital Health Team and escalated for ED referral. The patient was seen at the emergency department, stabilized and discharged home. Our digital health specialists told our researcher, our service has become essential, we have called a bunch of close calls with patients.

One patient even said I saved their life. We have some early results from our 800 some patients. Among patients with hypertension, there was an average reduction in both the systolic and diastolic blood pressure. Among patients with diabetes, there's also reduction in blood pressure and their A1C measures.

We also asked about our patients with their experience. Based on your experience with UC San Diego Health Population Health Digital Health Program, how likely are you to recommend family and friends to the program? We see that 55 percent of them responded that extremely likely, and 31 percent very likely. In total, 86 percent either extremely or very likely would recommend this program. It was my honor to share with you some early learnings from our patient-centered team-based digital health program in UC San Diego Health as a learning health system.

Thank you. I give it back to Dr. Patrick. Thank you Ming. Much appreciate it. Our next speaker is Dr. Shamim Nemati,

who's an assistant professor in our division of Biomedical Informatics, and Director of predictive analytics at the UC San Diego Health. Among many of his efforts, Dr. Nemati is a UCSD ambassador to Microsoft, and it's artificial intelligence industry innovation Coalition for Healthcare. He's also a member of the editorial board of the Journal of Critical Care Medicine.

Shamim, you've got the floor. It's great to be here. I've been asked to talk about applications of artificial intelligence and machine learning and how they help with improving patient care.

I tell you it might be fun to just go through a use case and demonstrate how this type of algorithms they come about and how they can help. Around September of 2020, I received an email from Dr. Jane Burns from the Rady Children's Hospital, telling me that they are starting to see pediatric patients coming in with persistent fever, abdominal pain, with vomiting, rash skin, and in some cases severe cases of hypotension and shock. She and her team realized that there are some similarities between these patients and another group of patients that are known as the Kawasaki disease patients. In fact, Dr. Jane Burns and her team, they spend the last 20 years studying this Kawasaki disease.

It turns out this new group of patients that they were seeing, they belong to this other diseases called multi-system inflammatory syndrome that really came about during the COVID, and roughly 98 percent of these patients they are positive for SARS-CoV-2 virus. The other two percent, they had some sort of contacts with other people who had COVID. Dr. Burns and her team, they realize that there are quite a bit of similarity between these patients. In particular, many of the laboratory values are off in both groups, but there are cases that there are also differences like in the MIS-C case, the older children's in the Kawasaki disease, they tend to be the younger demographic differences between these two groups.

But ultimately many of these patients, if they don't receive the treatments, they end up developing coronary artery dilation, and basically a lot of heart problems, they end up in the intensive care. She basically said that at the point of carrying the ED is quite hard for our front end conditions to figure out the difference between these two groups, and it would be very nice to have AI system that can distinguish between MIS-C and TD patients. Around the same time I had a very talented PhD student, Jonathan Lam approached me and he was interested in doing the rotation. I said, fantastic let's work on this timely problem. Dr. Burns and her team, because of their work uncover Kawasaki disease, they have been putting together registries over the past 20 years, in particular for Kawasaki disease, as well as Febrile children's and since the beginning of pandemic also on this MIS-C patients.

He got to start it with building a model. Around the same time, Dr. Burns established a collaboration across the US and started collecting data sets from many cases of MIS-C throughout the country. What we had access to was very simple information about the patients as well as their age, and their laboratory measurements on the vital signs.

This particular data-set had diagnosis of Kawasaki and as well as MIS-C, as defined by the Center for Disease Control and American Heart Association, and we use that as what we call the gold standard. We had examples of patients and they had their diagnosis. Next, what Jonathan did, he said, "Let's build a two-stage algorithm". In the first stage, basically the algorithm distinguishes between MIS-C and not MIS-C patients.

Patients come to the emergency department, they collected seventy laboratory measurements, as well as the five clinical signs and the algorithm is able to classify patients into two different buckets. The first one being MIS-C, the second one we're not MIS-C. Then among the non MIS-C patients, there is a separate classifier that says, is this patient in Febrile control patient or is the Kawasaki disease patients? The reason we came up with this two-stage approach was because really clinicians wanted to know about this rare and unfamiliar case and they didn't want to miss the diagnosis of MIS-C, so that's the first stage followed by the second stage. The particular algorithm that Jonathan use is something called Neural Network.

But before getting into that, let's just walk through how the traditional models work. The traditional model often rely on laboratory values and clinical signs, and what they do is they assign certain risk factors through each of these clinical features. In this case, a risk factor is smaller than one, it means it reduces the risks for that particular disease and risk factor that is greater than one, it means that increases the risk.

You can imagine that a simple approaches to say, how many of these risk factors the patient has? We are going to add them up based on the make a decision. You could have something like that for MIS-C, something like that for KD, and that's like a simplest type of machine-learning model that you can have. What these types of models and remits is the interactions about the factors.

If a patient comes in very young age and has rash and also the platelet counts are abnormal, we are now looking at interaction among multiple risk factors. That's where some of these more advanced machine-learning models like deep learning techniques come into play. Just a little bit of background about that. It turns out the deep learning methods they were initially motivated by studies of human visual system.

Human beings are just absolutely amazing at recognizing patterns. If you have a pet, whether the pet is close by or far distance, whether it is early in the morning or late at night, you can recognize your pet and that's what your visual system does. But the way the visual system is able to accomplish that is through the sensors that you have in the eyes, in particular the photo-receptors, but if you look at these sensors, they are pretty simple. You have receptors for different colors and then receptors that they fire when an object moves from the left to right or up to down, but what they do is essentially they detect very simple features of the visual scene. Sometime around 1968 or 1960 's, Hubel and Wiesel who were electrophysiologist, they started looking to the human visual pathway and they realized that from the time that light arrives at the eyes, it goes into several layers of processing.

In particular, these layers are known as V1, V2, V3 areas in the brain. The next thing they did was they essentially poked in some electrodes into different areas and they asked the question that neurons in these areas, what did they respond to? They realize that in the V1 area, the neurons were responding into fairly simple patterns. By the time they got into upper stream, into other areas, they noticed that the neurons were firing or responding to more and more complex patterns. But the fascinating thing is that this ability to recognize complex patterns, it emerges from his hierarchical nature of the brain where it's combining very simple features at the lower level. Simpler, such as the contrast, the level of lightness, etc. Inspired by the human visual system, computer scientists started to put together what we call artificial neural networks that are capable of identifying complex patterns in the data.

One example could be a patient with hypothermia whose immune system compromise and has elevated white blood count, together these risk factors that might be indicative of infection. Again, inspired by this, Jonathan, what he did was he said, "Let's grab these risk factors that we have and pass it into what we call an artificial neural network". This neural networks, they have multiple layers of processing, we have the simple risk factors as input to these models and then there are these intermediate nodes that they can pay attention to certain aspects of the input data. Depending on how the weights in the network are defined, certain nodes they can focus on, let's say rash and platelet count and the age, another node might focus on myocyte and white blood count, etc.

Then these risk factors, they get combined together. This is an example of a machine-learning algorithm that is able to look at clinical signs and certain nerves, combine them together in a efficient way to make a diagnosis for MIS-C. Then what we did was we said, among the other patients, why don't we pass it to a second algorithm that characterizes Kawasaki disease. Jonathan was able to build this model and he was able to externally validated on data from another 16 hospitals, all within a period of a few months. Then our clinicians we're super excited, they said, can we have a tool, at the point of care in the emergency departments such that we can use it to detect these cases of MIS-C? What he did was he basically put together what we call the KIDMATCH calculator that you can enter the signs and symptoms and then you can click a button and it produces a risk score for MIS-C, but also it displays the top contributing factors to the risk, so that the clinicians, they know what is happening inside of the AI system and what factors is paying attention to.

This was a very nice example. It all came together in a period of a few months during the pandemic and our clinicians have been using this providing feedback, we have been improving the user interface. Moving forward, while this particular use case involves what we call a web-calculator, where clinicians have to enter the individual numbers. Now at UC San Diego, we have the capability to bring data in real-time from a electronic health record system, which gets its data from the laboratory, pharmacy, other sources, as well as the different devices, whether it is Linklater's, IP infusion pumps, dialysis devices or bedside monitors. We are able to bring all of those into our computational environment.

More recently we have been working on bringing also imaging data or variable sensors. Some of the advanced sensors that are bio-engineering collaborators they're developing such as this tiny little sensor for measurement of blood lactate on a minute-by-minute basis. The idea is that the data gets harmonized in one place and then we can use it to make predictions. In fact, just over the past two years working with Dr. Chris Long and his team, we were able to deploy to new algorithms, one of them is for early recognition of sepsis in hospitalized patients, the other one is for prediction of need for mechanical ventilation. Then working with Dr. Tai-Seale,

we just had a proposal submitted that says, why don't we also bring together data from things like air quality in a pollen levels, weather patterns, as well as other smart devices that look at compliance? The idea here is with the use of AI, we would be able to perform data enrichment and do multi-modal analysis. I'm going to stop here. Thank you for listening. Thank you, Shamim. Much appreciate it. Our final panelist today is going to be Dr. Jeff Schaffer.

Schaffer, he's a member of our UC San Diego Patient Advisory Council. When I asked him additional information that I could relate, he said, "Well, I speak fluent Italian and I cook like I live in Tuscany," so that's a bit of background on Jeff. But again, Jeff was nominated by Ming to talk about this patient advisory council work and so maybe they can spend a couple of minutes talking.

Ming will tee this up a little bit with Jeff to get a sense to the group of what his role has been. So Ming. Great. Thank you very much, Kevin. It's my great pleasure. I'm really excited to be sharing this time with Dr. Schaffer,

who is a member of the first patient advisory council in our population health services organization where the digital 1,000 program is. Dr. Schaffer, could you please share your experience with our audience on how you got involved in the patient advisory council? Certainly. Thank you, Dr. Tai-Seale. I just wanted to thank you and the rest of the team for the honor of being a part of this health talk. I think after listening to the talks, I think I would be or I am a good example of showing the instance of patient and how the digital health technology help the patient and improve their situation. For my example, it was dealing with hypertension.

What happened was is I did go to a primary care, I did discover that I suffered from hypertension. I guess it was due to the COVID and to the time because this happened recently and I never had that issue. I was directed by the primary care to the popular Population Health System, which I figured, I don't know what that is and they actually contacted me. I must say, going to the doctor's office, and I'm being a doctor myself in the veterinary field, I found that I suffered from a white coat syndrome, which I felt I never even heard of that.

Meanwhile, going to a doctor, they even increased my hypertensive measurements. When I spoke to them, I hate the population health system, they said, "Well, why don't we work with you from home and try to regulate this?" That's exactly what happened. The group or the team of this population was absolutely amazing because I've never taken medication my entire life, so they said, "Let's try to regulate it because it was extremely high with medication. I worked with them for months, which was quite interesting because I was having bad reactions to the medication. I was having effects and it wasn't coming down.

But because of the interaction and because of the relationship with this population health system, we got it under control, and I'm actually quite happy of my numbers now. They did send me the system that you see here. The system that I am addicted to because I now measure it and I just find it's extremely helpful, and that we've also discovered the proper medication after certain changes of ACE inhibitors and diuretics. Now I got to the correct medication. I must say, it's been a fantastic experience.

I would be available for any patient also who wanted me to advise them on how great it could be. Thank you so much, Dr. Schaffer. Really appreciate your insights.

Dr. Schaffer has been an invaluable member for our patient advisory council. So thank you again for joining us. My pleasure.

First, I want to thank everyone, particularly Mr. Schaffer for participating in this, because it's really very, very helpful at the end of the day to understand how it actually works. Some things look great on paper and they don't work in practice. Let me ask Ming the first question.

There are a lot of related questions. But people were wondering, what's the limitations of digital health? What you don't want to do by digital health and what works better by digital health, in your opinion? In my opinion, I think of what works well is it's something that provides the necessary information that could influence either personal health decisions or inform the clinical team in changing clinical decisions. The information is valuable to affect choices, whether it's personal lifestyle habits or clinical decisions, like in Dr. Schaffer's case, what medicine for his hypertension is most effective. When we have digital health remote monitoring tools that provide that kind of data, that's very helpful.

I think for things when it's not helpful, I think if the devices are not well calibrated or if it's excessive amount of data, and if the system is not built. If the infrastructure is not there to support the patient, then it may be premature to use those. People could get anxious.

They may not know what the data is telling them. I know there are lots of experts who can answer this question probably much better than I can. [LAUGHTER] [OVERLAPPING] Jeff, please go ahead.

I just want to point from the patient perspective, I think one of the limitations is compliance. Here is you send them a machine, you tell them to do it and they don't do it. What I found super fantastic from this population that was building the relationship with the caregiver who followed up with me, and it was an amazing relationship that my compliance of taking my test, I started getting addicted to it. I was learning more about myself doing the hypertensive tests.

I saw when it was high and then it was emotional. It was attached to my emotions, what's happening in my life, so it was fascinating. It was an extreme learning experience as a patient.

But the compliance issue, I think would be difficult. But because you guys at UCSD and the team really cared about me, I complied. I have another really quick questions for Jeffrey and Ming. People want to know how they can get on your patient advisory council. Wonderful. Send me a message, we'll be very happy to invite you.

We have currently seven members who are patients or family caregivers, and we meet quarterly and it's been a wonderful experience. We've had two meetings and more will come so we can always have more contributors. Thank you. I have some questions for Kristen. People want to know how you integrate research with clinical care and in particular, if the type of data generating you're using real-time or as an analyzed and use later? Absolutely. We do both. We use the data in real-time for clinical care and management decisions at the time of the patient appointment. We also have the ability to utilize that data for research retrospectively.

As we were building our Clinical Research Center and really trying to be able to have patients be able to enroll in a lot of these trials going forward. It requires a lot of infrastructure that we're working on building, making sure we have consensus at the time so we can reach out to people in real-time who might be interested. But absolutely from a larger data perspective, we can use the data for retrospective research, looking at outcomes and things like that. I don't know who can answer this. Maybe Kevin. [LAUGHTER] Maybe not.

There are questions about what insurance covers and it doesn't cover. We have these technologies. We're doing a project, but what is anything that we know is covered by insurance yet or not? I can speak to that. Good. Thank you.

In diabetes, we deal with this on a daily basis. A hundred percent, insurance will cover a glucose meter. Now, whether that is the glucose meter that interacts with our system flawlessly with one step and having the Bluetooth ability for that passive flow of data, versus the meter that requires 16 steps to connect and get that data, that's up to the insurance company. This is one of the issues that we're dealing with in Project 1000, where we're using this one particular meter that we have to provide to the patients.

But in trying to design and to be able to scale this for our entire patient population, and to be able to utilize the meter that is covered by insurance, we're running into these little technical difficulties and the nuances between different devices. That being said, we also have codes to where we can bill for that interpretation of this data between visits, so we do get some reimbursement to help support a population health type team to be able to look at that data between visits. I would say it's definitely headed in the right direction, but still quite a bit of ways to go.

That was an excellent summary and I just want to say something about new technologies, new clinic technologies. The industry is required to, first of all, get FDA approval and that's a very rigorous process. But on a reimbursement side, we have to produce credible data of the value of that new technology, and that value is measured in a number of ways, but it's not like buying a TV set where if you like the way it looks, you go and buy it. We have to, in industry, create real hard and fast data that the technology delivers benefits to the patient that's measurable. I have another question for Paulson. People want to know whether or not companies will develop a whole slew of different devices, with one company developing multiple devices, or everything be one-off like the glucose monitor, which is one area of specialization, the blood pressure is another area? First of all, technology is a big field.

Most companies will develop a division or at least a multi-personnel function to focus on an application and you have to recognize that virtually every technology that you've been exposed to in today's presentation when you look at the first use and you look at it five years later, it has gone through many iterations and new models and more sophistication. Companies tend to form periods of acute specialization on an application by application basis. I don't know if that answers your question. No, that's exactly right. Thank you so much.

It's was really helpful. Shamim, I have a question for you. Some of my colleagues are worried about being replaced by artificial intelligence, particularly in the areas of radiologists and pathologists, do you have any words of wisdom for them? Well, we just had a grant funded.

It's focused on machine physician symbiosis. I think clinicians are at the bedside and they have access to certain information about patients that currently AI systems, they just don't have. We often talked about nursing intuition or clinical physician intuition. I wouldn't worry about that. I think if anything, the current AI systems, they can really benefit from receiving constant feedback, from clinicians and improving their performance. Although in my field on herpetology, they find that in the clinical trials, if they read the slides by AI, they're much more reproducible than if they're read by individual pathologists.

[NOISE] I think AI systems, they are more consistent, but I think there is this issue of data integration that clinicians they just have access to so much more information about patients that AI systems they don't, as of now. If I could add to that [OVERLAPPING] I was thinking another question. Go ahead, Kevin. Well, it's the challenge and we all know on this meeting, people don't have one condition. They don't have one problem. They often have a set of concerns, set of issues that have to be dealt with, and the integration, you've heard me talking about this, the home visit that establishes certain contextual issues. I think that pivot or that shift, many of the things we're talking about today can free up simpler things by virtual technologies to allow that deeper and richer understanding to occur.

Somebody has hypertension and they had diabetes and they've, just lost a loved one so there are also depressed. A variety of these things the deeper you dig into them, the more opportunity there is to, I think, provide better care. These things can be adjuncts to this better provision of care with luck that will happen. It's funny, you should say that. The next question was, are there apps for depression and anxiety? Maybe Kevin can answer that.

Well, absolutely. Yes. I mean, mental health, the whole field of mental health is one that's exploding and I think COVID advanced that because of remote care. But I would again turn it around and say, there's lots out there, but are they proven effective? Are they proven over the long haul? Many, problems like depression are long-haul, long-term problems, and one is required also to take an adjunctive medication along with them.

Again, we are at the dawn of this and I think there are some, and we will be discovering more of them in our center for innovation. There are different researchers. I'm happy to go offline and introduce folks [inaudible] a young researcher here at UCSD is doing some phenomenal work in interface technologies that people are using to then discover things that might not otherwise be either found in the first place or managed over time. Thank you. Ming, several people

are concerned about the balance between technology and hands-on patient care that for example, either sitting, talking to the patient, picking up visual cues, things like that they're afraid will be lost in a digital world. Do you want to comment on that? Sure. I think it's really important to have that trusting relationship between the patient and the physician or the health care provider [NOISE] and it's really hard to replace that one-on-one in that safe space. However with COVID, we've seen there are a lot of real-world limitations, constraints on the ability of a healthcare system, to continue to provide that out of concern for patient health and for healthcare professionals health. I think the digital, the telehealth as an augmentation to that in-person face to face real-time communication is a tool that could support the patient and meet the patient's care when perhaps the first best is not available, but it's much better than the alternative of not having care. We have recent paper published by our faculty at UC San Diego that noted that patients with cancer, during the COVID period, they're coming at a later stage.

It's really important that we have additional mechanisms for patients to continue to receive care given the constraints because of the pandemic or because of other issues [OVERLAPPING] Maybe Jeff wants to comment how it's worked for him as a patient. Remember when you see a doctor, their time is valuable, your time is valuable, and therefore having the data and the information in front of the doctor and he's already studied, and he's gone over the history actually saves the time. He knows what's happening and therefore, you can really get into more important face-to-face information, doctor-patient relationship. I actually found, I totally understand that because you want to have a good doctor-patient relationship, but this actually saves the time and it makes that time even more beneficial, to both parties. It's actually is the savings grace the health technology. It is a balance, but it really works.

Thank you, Jeff. There's some concerns about security, about all this data going all over the place and who gets to see it and does it the company who's generating the data, the device generating the data actually gets to keep the data, or is it just available for the physician in charge? Well, on the UCSD site, we have a HIPAA compliant environment it's a security environment and the data doesn't really leave the institution. Under the device's side the key term there is consent and to the degree that patients they understand what they are consenting to. Their data might go to different places.

It's very, important to read the fine lines. Thank you all. This was really fun. I really appreciate it. Bye-bye now.

[MUSIC]

2022-03-01 16:03

Show Video

Other news