Invisible Technology and AI: The Future of Scaling Virtual Care
Kent Dicks, CEO of Life365, shares how artificial intelligence (AI), “invisible technology,” and smarter reimbursement models can overcome barriers to virtual care and remote patient monitoring, transforming health care into a more proactive, cost-effective system.
Kent Dicks: My name is Kent Dicks, and I'm CEO of Life365. This is my third company, and the second in digital health. I've worked in digital health for about 20 years. I am a major patent holder for this company and my previous company on wearable technology and artificial intelligence (AI) technology, and I was a graduate of Arizona State University in computer science.
What are some of the most significant barriers to large-scale adoption of virtual care and remote patient monitoring (RPM) technologies, and how do you address them?
Dicks: I think a major industry problem has been reimbursement. Even now that reimbursement is starting to become more prolific, alignment of risk for patients is still an issue. Providers are increasingly subject to risk models and value-based care models. This means they need to start managing populations cost-effectively and proactively. They also need to engage patients, get early insights from patients, and get data cost-effectively.
[Data collection] is a very problematic issue for most health care systems, which causes a problem in scaling. There are a lot of RPM companies out there that think about deploying scales, blood pressure monitors, pulse oximeters, glucose monitors—anything that's needed to try to get a patient reading at home and compliant. But there are issues and headwinds that are also causing problems under the value-based models. People are getting older, they're getting sicker. When I started in this industry, health care costs were approximately $2.5 trillion dollars in health care spend per year. In 2027, we are going to probably be up to $6 trillion. When I started, about 60% of people 65 or older had one or more chronic diseases. Now, it's close to 80%.
The number of care providers in the industry is really going down, too. We don’t have enough care providers for this aging population that's getting sicker and more costly. Technology is going to play a really important role going forward.
Looking back, what have been the accelerants along the way in previous industries necessary to create a trillion-dollar invention that's out there? The personal computer rocketed the applications as they were developed. The internet has rocketed applications and consumerism. The smartphone is another example. What is that thing that's going to start changing health care? The answer is AI. We can talk about AI today and come back in 2 months to find the landscape has almost completely changed.
AI is going to play an important role in bridging gaps to gather more insights into care for patients. Instead of waiting for a patient to get sick, go to the emergency room, and be admitted to the hospital, we want to start getting data in advance, run it through AI, and start doing what I call the 5 P’s: proactive, preventative, predictable, personalized, and prioritized care.
What is invisible technology, and what does that look like in practice for both clinicians and patients, particularly older adults?
Dicks: For solutions to get adopted by clinicians, it must be in their own clinical workflow and be somewhat seamless. Clinicians should be able to go into their electronic health record (EHR) and say, “I'm ordering Kent Dicks to be connected.” What they don't need is to say, “I'm putting Kent Dicks into RPM, then I need an assistant to order the devices and monitor and engage with the patient, and another person to handle billing.”
Unfortunately, many RPM companies have the approach that they're delivering technology to the doorstep of the doctor and having the doctor choose how it is deployed. It can't be that way. Part of the invisibility of invisible technology is that it must be seamless in their clinical workflow.
The other part of it is being proactive and predictable. We want to go from reactive care, where a lot of the RPM companies are, into proactive care. Part of that means moving into a class of patients who are not spending quite that much on health care dollars yet. But we can try to get that data for AI to get early insights from things that are already around the patient, almost device-less. Just like the interaction you and I are having right now. We can use the camera on this call to potentially collect some physiological or biometric data. We can use the microphone to get indications from voice biomarkers. In fact, we have a great partner, Canary Speech, who works with us at Microsoft to try to get indications of what needs a patient's attention.
The biggest thing that I am all about is getting data. Some data can be collected from lower-grade devices, or deviceless interactions, or biometrics from facial acquisition that AI can benefit from. Maybe that data doesn’t need to be shown to the clinician or patient, but they can absolutely be used in AI to look at trends and start measuring somebody against themselves to nudge them into compliance.
I think we'll see more of this kind of data coming from invisible sources, like from cameras, microphones, cost-effective disposable wearable sensor devices, or from Apple Watches. It's cost-effective and allows us to line up earlier with a patient to get them compliant.
How do you address patient and provider concerns about privacy and cybersecurity risk when collecting data?
Dicks: We personally use devices that are US Food and Drug Administration (FDA) cleared. They've already been through their own verification, validation, and risk assessment to be cleared by the FDA. We are also on platforms that are HIPAA-compliant and encrypted. We're trying to use all the security measures that we can going forward.
When we enter into a relationship with a patient, we agree to keep their data private and to use it for their care, maybe to be shared with a clinician, or even to motivate a patient to engage more in their care. We do everything we can to try to keep their data as secure as possible in multiple different ways, but nothing is 100%.
I think I was asked this question on Fox News and Fox Business years ago about security. The same concern arose when people were switching from writing checks at the grocery store to using ATM cards. There are still inherent issues, but you're trying to keep it as secure as possible through secure point-of-sale (POS) and everything else. Now, we see people who have completely gone from using ATM cards to having tap payment on their phones. We must also be super careful because scammers are trying to get this information, so it is important to protect yourself.
How can automation and hyperlocal design help to improve outcomes without contributing to clinician burnout or data overload?
Dicks: That's a big concern. I always say that doctors don't need any more data. We are going to produce more data as we start to get more connected. The thing that needs more data is AI. It needs more data, and in a cost-effective manner. For models to work, they need to have multiple different types of data sets, and that's the reason why we like working with Microsoft and their Fabric program. It's designed to be able to take in multiple levels of key data sets from labs, medical devices, behavioral health, or other sources. The AI can read it and organize it in a way that allows it to be relevant to a doctor. A doctor doesn't have to assemble the data and try to make sense of it.
I came up with a very bad analogy for this—so I apologize in advance—but my dad was very into cars. He would see a car and notice it started sputtering and replace the spark plugs. If that didn't help, he’d replace the fuel filter. And that if that didn't help, he’d replace the spark plug wires. I'm not comparing doctors to my dad, but right now in health care there are some similarities. Personally, if I go on to a cholesterol medication, my doctor will start me on lisinopril or another medication that doesn’t work. They try their best, but it’s the same process with antidepressants and SSRIs. It's not always science.
Now, we're using pharmacogenomics, and I think that's going to be awesome when we start looking at your body type and your genome. But we're not there yet to be able to do that. In my dad's case, you can't try to do it yourself. You must take it to a mechanic. They must put it on an analyzer, but at least it tells you the code of where the issue is. I am not a doctor, but I think that being able to have insight pointing to where the issue is and allowing them to do their own assessment and diagnosis would be incredibly helpful.
This requires multiple different data sets. EHR is a great place with a lot of data, but no doctor on this earth is going to be able to process all that data on you or me and try to make a diagnosis within 5 minutes. But the EHR data, genomic data, and the claims data are all great in an AI world for learning about Kent Dicks. Maybe every time Kent’s blood pressure gets above 180, or his oxygen gets below 85, or he gains more than 4% of his body weight in a couple of days, he's labeled a congestive heart failure patient. Maybe he's going to be heading towards the emergency room. That is great insight, but then you need to have observational data about what's happening with Kent today. That data comes in from all these devices—a cost-effective solution.
With that observational data, an AI system can watch 20 000 patients at once and notice when one needs help because it can compare the data coming in against learning data. It would be incredibly helpful for a doctor to be able to prioritize from that perspective.
Technology is here to stay. It's going to play an important role. I always say when you're in a skid, you must turn into it and embrace it. From that standpoint, I think the technologists and the health care providers need to be aligned, because AI is here to stay, and it's going to be what helps patients going forward.
With the US Department of Health and Human Services (HHS) requesting insights on digital health for Medicare, what policy or design recommendations would you prioritize to support the national expansion of virtual care?
Dicks: I know the American Medical Association (AMA) and the Centers for Medicare and Medicaid Services (CMS) have made recommendations for this. We didn't get our first set of reimbursements until around 2018, and they wanted to stop very conservatively because they were worried about waste, fraud, and abuse. Not many people adopted it because they were worried about a backlash from Medicare and Medicaid.
And then we got into [the COVID-19 pandemic]. [The pandemic] allowed the readings to go down from 16 days of readings to 2 days of readings. We saw bigger adoption in telehealth and also in RPM. The rumor is that our recommendation this fall will be that the readings go down to 2 days a month for compliance. I think that will increase adoption significantly. It is my opinion that, probably by 2027 or 2028, we'll see the ability for doctors to allow their EHRs and AI systems to bill on behalf of keeping patients compliant.
A doctor may be brought into anomalies along with AI, and maybe those are under different codes, but there should be codes out there for a doctor allowing their system to monitor their population of 2000 or 3000 patients they're seeing regularly to nudge them and keep them in compliance. I think that's a big thing that needs to be done, especially with physician and nurse shortages. I think that AI will be billable at a significantly reduced rate and will go after a much larger population of patients to nudge them into compliance.
I also think they really need to straighten out the licensure issues so that patients can be seen by a care professional across all 50 states and territories. Now there are multistate complex compacts for licensure, but I think that it needs to be more at the federal level.
How do you envision the role of virtual care and RPM evolving in the near future? What breakthroughs do you believe are needed to truly transform health care delivery at scale?
Dicks: I'm a data guy, so I think data and AI systems are important. We have what I call “wearables, hearables, and seeables,” but I think we’re increasingly going into what I call “disappearables”. These are woven into your day-to-day fabric and cloth—more disposable patches, more connectivity directly to the cloud, and not necessarily through an intermediate device. But I think we're already starting to see an industry preference [for avoiding the need to] send a Bluetooth scale or blood pressure device, because they can be problematic in pairing and become unpaired. Instead, cellular can be built into individual devices and have the data shipped directly to the cloud.
Disposable sensors, AI, and data are all important. Those are all potential changes I see coming. This industry is now being driven by the consumer and by the payer. More than anybody else, we as consumers are starting to decide on who we see and how we see them based on the dollars we spend and who we spend them with.
From that perspective, my choice is to sit behind a computer—or the 2 or 3 screens that I have in front of me—and communicate with my provider regularly. I am compliant enough that if I have a lab 2 blocks away, I can go there and get my blood tested regularly if my doctor orders it. I prefer not to be on medications, but my supplements and medications all come through mail order when I need them. I think the system must be more heuristic and learn when to start shipping those out automatically. It's like shipping my food to my door through HelloFresh or Blue Apron. It needs to be more intuitive to my needs.
We're in a very dynamic time with the government, too. My approach is that sometimes, through chaos comes opportunity. Things do need to be shaken up to be able to be reassembled in a different way and make room for new processes and technology—even if I may not necessarily agree with how that is being done. I think we’re going to see a lot of changes over the next couple of years. Health care tends to move very slowly, but the accelerants are an aging population, increased costs, and AI. AI is going to give consumers the power to say what they want quickly. If health systems, providers, and payers are not willing to move as fast with the consumer, then the consumer will find another way to get what they want.
