Episode 37: It’s time to take your medicine – and Charles Gellman of HiDO Health has AI robots to help you remember

Definitively Speaking is a definitive healthcare podcast series recorded and produced in Framingham, Massachusetts. To learn more about healthcare commercial intelligence please visit us at definitivehc.com. Hello and welcome to another episode of Definitively Speaking the podcast where we have data-driven conversations on the current state of healthcare. I'm Justin Steinman, chief marketing Officer at Definitive Healthcare and your host for this podcast. More than 131 million people take a prescription drug every day in the United States according to the Health Policy Institute at Georgetown University. That's more than two thirds of the adult population. According to the Center for Disease Control more than 1 billion, that's billion with a B, prescriptions are written each year and 48.6% of the US population has taken a prescription drug in the last 30 days. 24% of the US population has taken three or more prescription drugs in the past 30 days. That is a lot of drugs. According to the Department of Health and Human Services, the US Health System spent 603 billion again billion with a B on prescription drugs in 2021. That's a lot of money. But what if people don't take those prescriptions? Have you ever forgotten to take a pill prescribed by your doctor? Maybe you're rushing outta the house in the morning and you forgot, or you fall asleep on the couch watching your favorite football team and you miss that nighttime dose. Or more seriously, maybe you have a medical condition like Alzheimer's where you struggle to remember to take your medicine. We all know that for prescription drugs to work, you have to take them but sometimes it's just difficult to remember or you can't and that's where today's guest comes in. Charles Gellman is the CEO and founder of HiDO Health. Charles and the HiDO team are developing AI assisted robotics to assist patients with medication compliance. HiDO automates the medication experience and removes much of the patient burden to remember the right medication, the right frequency, the right dosage, and validates the right person through remote video record. Charles is also one of the foremost public speakers on the topic of AI in healthcare, so I think we'll have a lot to talk about today. Charles, welcome to Definitively Speaking.

Thank you Justin for the introduction. I'm super excited to be on the podcast today and talk about these subjects.

Yeah, really looking forward to. So let, let's kick this off. Where'd you get the idea for HiDO Health?

Great question. I never thought we were gonna go down this path and it's been a bit of a rollercoaster ride but my background's in data science and I was always intrigued on what would the blueprint for health look like? And in order to understand the blueprint health you need to have a series of data points. And those data points really culminated from my experience when I was a pharmaceutical representative at Johnson and Johnson. And the primary inputs are medications and in order to understand if people are taking or missing those medications it ultimately impacts your blood work, it impacts the outcomes, hospitalizations, readmissions, emergency room visits. And once I started to go down that train of trying to find that data, I quickly found out that it didn't exist. And so that's why HiDO came into existence.

So lemme get this straight. You're a data scientist who became a pharmacy rep and then became an entrepreneur.

That's correct. Yes, I've managed teams throughout the country. Pharmacists, nurses, doctors have a clinical informatics background, have been on the commercializing of sales. And then yes, off to entrepreneurship.

That's quite a journey. How's it going so far? How's HiDO Health doing?

It's going fantastic. I mean, it's been quite a journey over the last five years. They liken startups as much of a rollercoaster ride and I equate it to jumping on that rollercoaster where you have those highs and lows but we're building the tracks as we're moving forward. So it's been quite a journey so far. We're excited to share a lot of the clinical data through some of the research efforts and the impact of HiDO on patients and what that looks like for health plans patients and the overall system impact to patient populations across the country.

So you've been at it five years and you just mentioned that you've got some clinical data. So how many patients do you have using this and what are some of those results?

Yeah, so we've gone live with the Medicare Advantage plan in Northern California specifically on heart failure patients. And this was a study with about 30 patients entered into the clinical program. And what we have done by simply knowing whether or not patients are taking or missing the medications is that we've reduced hospitalizations by 80% and the cost of care for these patients by 67%. Simply, from people taking their medications as prescribed, on time, the right dosage, the right frequency, and the right count.

Yeah. You know, I'm glad you brought that up. Take your medications as prescribed. 'Cause when I was researching for this podcast and preparing, I found multiple, and I'm talking like dozens of analyses that claim approximately 50% of Americans do not take their medications as prescribed. That blew my mind. Why is medication adherence such a big problem?

Well, I think it, there's a dichotomy that happens when you have that provider patient interaction and you know oftentimes providers are pressed and they have a small window of time to describe a very complex regimen to somebody that either is ailing from a chronic disease or may not be in the right mindset or head space to understand those complicated directions and instructions. So you have a bit of miscommunication that occurs and ultimately the patient's solely dependent for their health at home. And if they're having a problem comprehending that interaction, then it leads to failure.

Got it. So we were talking about shocking stats and this was a fascinating research for me to kinda work through this. According to pilz.com, great website name, a medication non-adherence leads to more than 125,000 preventable deaths each year. And about 300 billion in avoidable healthcare costs. Most people don't even realize it, but the odds of dying from your own medication non-adherence are about 10 times greater than dying of homicide or about 30 times greater for someone who's over the age of 50, how can you help? Is this a problem that HiDO can solve? Who else is trying to solve this problem 'cause this seems massive to me?

Yeah, this is a monumental issue that often people have been looking at for the last 15, 20 years and no one's really understood the underlying issues to solve it. So when we first started HiDO, we thought that this was already resolved and we looked at the different technologies in the market and most of them are just simply not patient centric. So the question really is how do you eliminate all the challenges and hurdles for the patient to be successful at home? And that is the culmination of what HiDO does. It simply automates the process to eliminate medication errors from happening at home and ensure patients are taking their medications as prescribed. And that's very, very unique to HiDO.

And so how does that work? What's unique about it?

So let's take you Justin. You walk up, you're taking your medication, it unlocks with face ID, so it's the right person, the right place. Then it dispenses the right medication at the right time frequency and dosage. So we have patented RFID technology in each cap that when you put the bottle inside the unit, it auto recognizes all that information so the patient doesn't have to remember that conversation that happened with their provider. All of that's done for them. And then they simply take the medication in front of the unit, which is done through video consumption recordings to ensure that these medications are taken effectively and on time.

So, sounds expensive, right? You got RFID caps, I gotta put a machine in my home. Who's paying for all this?

Well actually the health plans are paying for it and the patients pay nothing. So what's really expensive is the cost of hospitalizations, readmissions, and emergency room visits. So this is essentially a 10th of the cost. We can take care of patients at a fraction of the cost, so they can stay out of the hospital in the ER and they can live and stabilize their chronic disease at home without being negatively impacted.

Got it. So most people though, get their prescriptions from CVS, Walgreens, Amazon Pharmacy, PillPack. Are you trying to disintermediate them? Do you work with them, you go around them?

Great question. So I foresee a future where you have full integration to mail order pharmacy, where you can get your prescription medications coming to you auto programmed and then put inside the unit. So we will never displace the retail pharmacy because you have those acute medications. But when you start thinking about chronic disease medications and the way that they're set through 90 day fills, I believe that system could be altered to support patients better and impact their future stay on those conditions.

So just to be provocative, I can, why wouldn't Amazon just do this? I mean, they've got billions and billions of dollars they don't need you and they seem to have got more AI and Alexa and Echo and all that other stuff. Why can't they just go and do this?

Yeah, I think that there's a variety of companies that are poised for a potential acquisition or licensing the patented technology. And I think that there are companies that could potentially also play in this space successfully. But this is more than just pushing out a pill bottle in a pill container. This is a complete rethinking of the design in ecosystem of how you deliver care to patients. So it's more than just drop shipping a pill bottle and just hoping somebody takes it. There's integration with the health plan. There's some type of case manager or population health that monitors the data to support those patients and ultimately make sure that they're getting the best medications for the best outcomes possible.

Yeah, I'm glad you brought up compliance. 'Cause I wanna know, like who's monitoring this? So you say you're recording it are you reporting back to a case manager at the health plan? Are you reporting back to the doctor the person's loved ones, who's checking in on all this?

Yeah, so there's a variety of different techniques to elicit those positive habits. So some of the simplest way to change behavior because this is ultimately a a behavior modification tool is to elicit responses from their loved ones. They're friends. So if let's say Justin, you miss your medications 'cause you have congestive heart failure you're taking a specific medication, well then maybe your wife is notified because you missed your medications and she wants to make sure that you stay at the hospital. So we have a decision tree based upon the health plan or the provider network and then it's pushed to them via a clinical dashboard. They can monitor these high risk patients and ensure that they're taking their medications as prescribed.

So there's both a clinical dashboard that the provider uses as well as a kind of an in the moment push to someone who's gonna help get me take my medicine, right?

Yes, yes. Because you need to think collectively as a whole. It's not just about providers prodding patients to do the right thing but patients also have to take accountability for their own health and their behaviors whether they're good or bad. And if we can deviate from where they currently stand which is about half patients not taking their medications as intended to a more defined schedule, a regimen then ultimately you can smooth out their ability to control their disease states.

Right, but as I mentioned earlier in my introduction though sometimes a patient can't take accountability for their health, right? Like so for example, if I, God forbid, have Alzheimer's I can't be accountable for remembering to take my pills. So now who's got that accountability?

That is a fantastic question. So think about patients with dementia, Alzheimer's they have a huge cognitive impairment. And the thinking right now is that those patients are unable to care for themselves, right? What if they could care for themselves with assisted robotics in the home? How would that impact the complete future of care if they could care for themselves?

So what's assisted robotics, what does that mean?

So assisted robotics is what we specialize at HiDO and it's helping people help themselves in the comfort of their home. So there's a variety of integrated technologies almost like a a miniaturized brain at home to ensure that they're living a healthy life. So that's the definition of what AI assisted robotics could be.

And is that built into the HiDO solution in this drug thing? Is this kinda one and the same or are they different products?

This is one of the same. So the entire system build of the HiDO technologies which stands for health information data outcomes has been built by patients for patients based upon what they wanted, their needs, their behaviors to elicit those best outcomes. So we had a patient, Gloria, who does have dementia and we had our own self-limiting beliefs that she wouldn't be capable of using this device, but we gave it a shot. We wanted to see if it was possible for dementia patients to use this device. And to our surprise, you know, she was at home and her daughter-in-law was at a movie with her husband and she was able to take her medications unassisted. And that type of relief from a caregiver standpoint is substantial 'cause there's 40 million caregivers across the United States helping support friends, family, and loved ones with varying different levels of chronic diseases. And this speaks to what the capabilities are of people that are living with this different disease states and the type of impact that can be made.

So there's a health equity angle here, right? Like how do you help level the playing field for underserved populations?

Yeah, so when we went live with a Medicare Advantage plan in Northern California and you start thinking about health equity. So different race, gender, locations, income, all the other different health disparities. What the HiDO device does is it levels the playing field no matter the level of education, the location, it's implementing an ecosystem where everybody has the right medications, at the right time, the right dosage, the right frequency, and most importantly, they have the right support ensuring they get the best outcomes per hospital. So if you knew which cocktails and medications were the most efficacious and safe based upon those different parameters, and you connect that with the blood work and the hospitalizations readmissions and ER visits, you can change dosages up or down or potentially de-prescribe medications because they're doing so well. Because for the first time ever they're controlled on a medication regimen.

So let's shift the conversation a little bit to an area where you have have some expertise artificial intelligence and machine learning in healthcare. The reality I think is that people don't understand AI just yet. That's a buzzword. But people really don't seem to grok what it means. In fact, we recently did a survey here at and we surveyed several 100s life sciences leaders and asked them on a scale of one to five what's your understanding of AI and ML and how it applies in healthcare. And an astounding 65% of these leaders of their businesses in life sciences, rank their understanding of AI and ML as a one or a two on that five point scale. So my question for you is really twofold. What's driving this low level of understanding and then what do these people need to know?

Interesting. Yeah, so the level of understanding might be from definition of what AI is and some of the height behind where they see it in the media. And where I see AI is, is really is enhancing healthcare's ability to deliver care. So you see a lot of things that are "sexy in the media and the robots taking over the world or potentially replacing people's jobs." But as I see it when you start thinking about basic AI and ML it really is a way to enhance people's productivity or assist them in gathering copious amounts of information into one single area to make better decisions. So maybe during the survey people's definition or understanding of what AI and ML is ranked now may have been misinterpreted or potentially where they see it as the capabilities is not well defined. So it'd be interesting to take a look at the survey and chat with some of the people that responded to that.

Yeah, so you know, this survey was really interesting to me and I got some more stuff that I'm gonna kind of hit you with to, so get some conversation starters here. So another piece of survey, we asked these same people on a scale of one to 10, where a one indicated no adoption of AI and ML and a 10 means it's heavily adopted everywhere inside your company. 48% of people rank themselves a one, two or three in that bottom third. I feel like maybe we're just like in the very top of the first inning and we're talking a lot about AI, but it's not really adopted yet. Would you agree with that?

I think we're in the early stages. I mean, I had my past life I specialized in advanced clinical surveillance technologies within the hospital setting and that was essentially putting together different algorithms and decision trees to make better decisions at point of care within the hospital. And it took a while to get that information and adoption throughout the market. So I think we're very early in the stages of understanding the capabilities of AI and enhancing the performance of care. So I think that's in line with where we currently exist within the medical marketplaces.

What gets you most excited about AI and healthcare?

I think the most exciting aspect is what if we can create a blueprint for health? What if we can alter the future of care, not just now but also future generations? So as I see it, it's not just enhancing and assisting care, but what if we ultimately could change the trajectory of empiric treatment where providers are basing their decisions on medical training and clinical expertise to the forefront of analyzing all the most current and relevant information based upon outcomes and clinical studies and put that decision making within hands reach to care for people the way that the best institutions care for people. And there's not so much variability. So I think that there's a tremendous amount of promise for AI and the delivery of care in the future.

Interesting, interesting. So again, I'm looking at this life sciences study which is literally hot off the presses. We're about to publish this in the next couple weeks. And so in fact, by the time this podcast gets us it might actually even be published. So some of the stuff that jumped out at me when I was looking at the survey, we asked people what their top five use cases would be. We gave them a choice of like 10 and they could pick multiple choice pick one. 66% of people said this is the top one. 66% said they used it to mine unstructured data. 60% in the number two spot said it was supporting earlier disease diagnosis and coming in third with a 49% vote was identifying patients for clinical trials. And then coming in fourth was accelerating drug development and discovery. Now again, I don't think that we have a lot of the people you're talking more about care delivery and these are people more developing drugs. Do these results surprise you at all?

No, actually, I believe that we can ultimately potentially accelerate drug delivery substantially. And you're able to do that if you know some of the basic parameters of care. So if you're able to select smartly who those patients are based upon disease state, the types of medications, the actual outreach, and consistency of provider communications or connection, I think that there's a tremendous amount of value that could be aligned. So that's not shocking. And the other early disease onset detection, I think all of those different parameters based upon the large dataset that that would all make sense.

What worries you about AI?

Well, you've seen in the news the different types of models of what could happen if there was nefarious actions. So I think it could be used in a couple different ways. You know, one, I talk about the good outcomes and what are the best outcomes based upon the correct variabilities and inputs but it could also be used in different directions. And there's been studies, I think there's a a couple of folks that were playing around with AI of creating different medications the antithesis of what good would be, and then creating those molecules and they shut down that. So I think that there's a variety of different ways and there needs to be a way to ensure that you're getting the correct information out there and that it hasn't been compromise in any such ways.

Yeah. You know, so our survey this is the last survey results that'll hit you with here and then we can kind of move on a little bit. So we picked, our survey came back into the top three results were 76% of people number one concerned was privacy and security, 67% was data quality and 54% came in and a third spot with ethical concerns. I really do think the privacy and security thing is good although I'll have to say having seen the most recent "Mission Impossible" movie, I'm also terrified of what an AI could do as it tries to take over the world.

Yeah, with any type of new technology that comes out, remember when the internet first hit, right? Everybody was scared about that, then Y2K. You always have these moments in time where people are scared of new technologies that ultimately they benefit all of humanity. So it just depends upon, how people intend to use it. And I think, the majority of folks have good intent and then you always have those bad characters and then that's why you build those checks and balances to plan accordingly.

So where are we going tomorrow? What are you most excited about when it comes to AI?

I think tomorrow is very exciting. If you start thinking about the potential impact to AI I think we can rapidly accelerate drug discovery and patient recruitment. There's companies that specialize in computational data, drug discovery. There's other companies that have advancements to assist and enhance the delivery of care and productivity. So I think the future AI is exceptionally bright when you start thinking about what it enables, and enables to give people time back to their day to focus on their most important discoveries. So I'm very excited about what the potential is.

Awesome. Well, I'm excited too. Charles, thanks for your time today. This has really been great. Thank you.

Thank you so much Justin. I appreciate it.

And for all our listeners out there, thank you for listening to Definitively Speaking, a Definitive Healthcare podcast. Please join me next time for a conversation with Lesley Solomon Venture chair at Redesign Health. Lesley has a very cool job. She gets to look around the healthcare ecosystem, find the biggest problems, and then start companies to solve them. She's a professional company starter, if you will. In fact, redesign Health's whole business model is to find big problems, create a company, hire a leadership team, and do it all over again. I can't wait to talk with Lesley about how all this works and what are the most interesting companies that she started. I do hope you'll join us. If you like what you've heard today, please remember to rate, review, and subscribe to the show on Apple Podcasts, Google Podcasts, Spotify or wherever you get your podcasts. To learn more about healthcare commercial intelligence can support your business, please follow us on Twitter or X at @DefinitiveHC or visit us at definitivehc.com. Until next time, take care. Please stay healthy. And remember, as your mom or your doctor would say, "Remember to take your medicine."