Podcasts > Ep. 155 - How AI Can Democratize Healthcare
Ep. 155
How AI Can Democratize Healthcare
Robin Farmanfarmaian,
Monday, December 05, 2022

In this episode, we talked with Robin Farmanfarmaian. Robin is a professional speaker and healthcare entrepreneur. She is also the author of "How AI can democratize healthcare?"

In this talk, we discussed AI's role in the shifting landscape of healthcare delivery towards telemedicine and healthcare. We also explored the interfaces between traditional healthcare products and service providers with the new generation of digital diagnostics and therapeutics companies.

Key Questions:

  • How do digital therapeutics work with the technology company and traditional healthcare providers?
  • What is the current landscape of the ability of machine learning to make predictions accurate in healthcare products and services?
  • How are the integrations across other applications work to enrich medical data sets?
  • How does digital diagnostics impact the traditional healthcare industry and therapeutics companies, especially in hospitals and pharmacies?


Erik: Robin, thank you for joining us on the podcast today.

Robin: I am so excited to talk to you. Thanks for having me.

Erik: Yeah, I'm really excited for this conversation as well. I was mentioning earlier, you're a new profile for us, right? So, we usually talk to the founder of a startup. Obviously, you've started a lot of companies or worked with a lot of startups before. But you also have, I would say, a very dense profile. I think that's a good way to say that. Then an author, a professional speaker, an angel investor, board advisor.

Really looking forward to this conversation, and really looking forward to digging into the topic as well, this topic of how AI can democratize healthcare. Maybe as a starting point, I always like to understand a little bit more how you came to devote a large portion of your life to this. What was it about the topic that led you to focus your energy here?

Robin: Well, as you know, I'm an early-stage entrepreneur. I've worked on everything, from things like curing cancer to sleep apnea and medical devices, and even virtual reality for radiologists. But really, the entire reason I got into healthcare is because I am also a severe chronic disease patient. So, I've had 43 hospitalizations, 6 major surgeries, and 3 organs removed, 4 misdiagnoses at the age of 26, which was about seven years after I'd had my entire large intestine surgically removed.

My doctors were telling me I was cured, but I wasn't. I was in extreme pain. So, they kept upping my opiate dose until eventually I was on 80 milligrams a day. I hated it. So, I went to my doctor. I said, "I need off of this." They said, "We don't do that. We only increase your opioids. You're going to have to find a new doctor." I realized how crazy the healthcare system was at that point. So, I went home, dropped my own methadone dose. I ended up getting a whole new team of doctors that worked with me as a team. I got diagnosed correctly with Crohn's, put on a medication called Remicade. I went into remission overnight.

So, I have so much experience both on the healthcare side — where we're trying that work with startups — but also on the chronic disease side. I've realized, in the past five years, especially, we're seeing a massive shift in healthcare to the home. Now, what will make this incredibly actionable is the artificial intelligence component. Because we can have things like remote patient monitoring 24/7, giving off a huge amount of data. But you do need that software component to be able to track things. So, really, the AI component is necessary in this successful shift in care to the home.

Erik: Okay. That's fascinating. It really illustrates also the unique qualities of this industry where on the one hand, it's the most personal possible industry. It's dealing with your body. On the other hand, it's completely opaque. You go in and you say something doesn't feel right. Somebody tells you, "Here's some reports. Here's some diagrams. Here's the diagnostics." How do you interpret that as an individual, as the person that has the issue?

So, it takes a lot of willpower in your case to push it through. But we would like that that wasn't the circumstance. Maybe a good place to start, let's try to start at a high level. If we look at the landscape of the application of AI and maybe other digital technologies, it's a very complicated landscape. So, how would you segment this? How would you break that down to help us think through it?

Robin: That's the beauty of the book that we just wrote. It's that it really breaks it down into chapters. a chapter on remote patient monitoring, which includes both physiologic monitoring as well as remote therapeutic monitoring, which is reimbursed by Medicare. Then there's a chapter, of course, on digital therapeutics which are apps that are used typically in conjunction with things like therapy and medication. A lot of them are based on cognitive behavior therapy, which is a gold standard when you're talking about behavior change.

Then we talk a lot about AI voice technology, and not just your Amazon Alexa. Though that is an incredibly important part of the healthcare system going forward, but using your voice as a vital sign in something called vocal biomarkers. So, it's the combination of these types of AI programs that are either collecting and analyzing or delivering a type of healthcare to the patient on a day-to-day basis, where most of healthcare actually happens.

Erik: Got you. Then if I think about these two industries, you have the healthcare industry, the service providers who are the ones that can get reimbursement and navigate those regulatory landscape. They're the ones that maybe are certified to provide particular services. Then you have technology providers who specialize in AI and so forth.

How does this relationship work today? If I use digital therapeutics, am I directly working with a technology provider? Or am I working with a traditional healthcare provider and the technology providers in the back end? What does that relationship look like today?

Robin: There are a couple of different business models. Frequently, you are interacting with the technology company. So, let me give you an example. AlacrityCare is remote patient monitoring for things like neutropenia, septic shock, and sepsis for cancer patients. So, they have a combination of off-the-shelf blood pressure monitors, pulse oximeters. They do blood labs in the home. They can check in, or the patient can call them 24/7 if they need to.

Now, AlacrityCare is the one who manages the entire relationship, but it's also hooked up to the oncologist. And so, a few days before you even start to feel a symptom of something like neutropenia, the software will grab it because there's parameters that are personalized to the individual around the software. Ping us if the pulse ox goes below 90 in this particular patient or something like that. It helps with that communication. What's really cool is that because the provider does look in on it, these types of things are actually reimbursed by Medicare.

Erik: Okay. Having the provider, basically, as part of this allows there to be a certified opinion that enables them reimbursement.

Robin: Yep, companies like AlacrityCare and other RPM companies, they do have physicians. These are full-on physicians and clinical oncologists. There's that ongoing monitoring component. So, that's a monthly reimbursement fee, as well. A lot of the time, I don't know what they're doing with their actual finances, specifically. But a lot of the time, it's something that they can share with the doctor. The doctor pays X amount a month to the startup or to the technology company. Then they get reimbursed by Medicare a Y amount.

Erik: Okay. Then another interesting part of this industry is that there's just a very long tail of issues that people could have. Everybody is unique and experiences symptoms in different ways. They manifest themselves in different ways. So, if I'm thinking about applications of AI, I'm often working from the industrial sector. You're working with machines. Machines are relatively predictable. They don't have odd quirks and behaviors, and go off and drink weird things that change their hormones, and so forth.

So, if we're dealing with humans and all their quirks, where are we today in the ability to use machine learning to make predictions accurately? Is there like a small set of things where you say we can do these accurately but the majority, we can't? Or are we already there where we've given enough training time with a particular patient? There's basically a long list of factors that we can predict.

Robin: We are there. We are predicting. That's what a lot of things that are in the home are doing. That remote patient monitoring is that predictive technology. Let me give you an example. So, there's a company called January AI. With three days of back data with a continuous glucose monitor like the Dexcom, in combination with food databases, patient's vital signs, and what they've eaten that day, they are able to predict in real time a patient's glucose response at the point of time that that patient is about ready to take a bite.

I know. It gives me chills to be like — wow. That's a pretty big deal because predictive glucose response. Imagine you have an apple that you eat at nine in the morning. Then the following week, you eat it at 9 PM that night. Well, you've done a lot of different exercise, sleep, fiber, protein in your diet. All of our responses to individual foods change. So, it's not like you're always going to have that same glucose response, which is why it's so incredible that they're able to do that predictive component. They're already on the market.

I'll give you another example. A company called VitalConnect, they've got a single-lead EKG monitor that hooks onto the patient's chest. I think it sticks right there. They did a study, and they were able to predict with almost 6.5 days advanced notice of a cardiac readmission. It is incredible to see a lot of this. We're getting down not to stage one but almost stage zero medicine. It's incredibly exciting. So, the next 5 to 10 years, that's just going to explode.

Erik: Okay. In a lot of these examples, it sounds like they're conducting this analysis from a relatively small number of sensors, right?

Robin: Yes.

Erik: So, it's a sensor on the heart. Maybe it's a glucose. Again, you mentioned here that there's other factors that might influence this. How long have you slept? Have you been exercising? What's your diet? So, there's all this other data that exists that could also potentially augment the algorithms. But it might exist on an Apple Watch. It might exist in an exercise app. It might exist in your sleep management app. Are you seeing integrations across the ecosystem to try to enrich the datasets? Where are we on this today?

Robin: There are some integrations happening. I look to companies like TytoCare. TytoCare is a telemedicine toolkit that's direct to consumer, about $300, sold out at Best Buy and some other retailers. It has a stethoscope, an otoscope to look in the ears, and a tongue depressor that all hook onto a video camera.

When somebody's sick at home, instead of going to urgent care with, say, a screaming toddler with an ear ache, use the TytoCare app to call a doctor over video. They partnered with things like Amwell, which is a large telemedicine company. The doctor is the one who sees the live video camera feed. It's as if the doctor is actually reaching her hands through the computer screen. Because she can say, okay, with the stethoscope, move it a little bit to the right, or move it a little bit to the left.

Now, TytoCare already integrates directly into Epic. We know, in the United States, Epic is in a very, very large number of hospital and health systems. So, the integration into the EMR is actually an incredibly important thing, especially if you want the clinician — your normal doctor — to be able to see some of the results. The integration there is so incredibly important. Because when you think about something like lab tests, I know I go to Sutter Health. I get my lab done in the home, my blood labs. My nurse typically drops them off at Labcorp. So, it's a different system than Sutter.

Well, it gets my lab test results by fax which then gets uploaded as a PDF, which just says Robin.pdf. Is this machine readable in any way, shape, or form? That's just basic labs. You go on Labcorp's app or Apple's health app. That exact same lab shows up in this beautiful graphic. I can see a trending analysis and all of that, but yet my extremely expensive GI doctor at a very large health system is not able to access that.

Erik: Yeah, maybe we can we can talk about some of the frictions in the system here. Because traditionally, healthcare, although it's a very technically advanced field, it's not known to be particularly innovative when it comes to user experience, let's say. Then you have these tech companies, which are extremely user-oriented. Because they're dealing with software, they can really do get real-time data and make adjustments on the fly to an extent.

Do you see this innovation that's happening, for example, bringing health to the home? Do you see it also impacting health in the hospital in terms of how hospitals are also rethinking their processes, rethinking their ways of working? How is this impacting the traditional health care industry?

Robin: It's all over the board, right? So, when I talk to people like the CTOs or the CIOs of giant hospital chains like HCA, Hospital Corp of America — which is, I think, the largest private one here — they are innovation-averse. They like to be the late adopters. Because they wanted to test out in the rest of the healthcare system, get the bugs out, and get it integrated. Then they bring it on in full force. So, it's going to completely depend not only on that, but what part of the country you're in.

I gave a talk recently for a player in the south. When I said, "Okay. Raise your hands in the room. How many people have heard about continuous glucose monitors, CGMs? 5 out of 600 people raised their hand. CGMs have been actually on the market for 20 years. Of course, in their current form, the Dexcom and the Freestyle Libre. Those are much more advanced models that we were seeing even 10 or 12 years ago. So, it has advanced. But this isn't brand new cutting edge anymore. It hasn't happened in the past two or three years. And yet, even a large payer didn't even know it existed.

Erik: So, you mentioned a geographic difference in terms of — you're sitting in Palo Alto. So, I suppose the know-how, it would be a bit higher. But then, we can also think in terms of demographics. I we think about the people that might be in most urgent need of improved health care, they might be the elderly who are not particularly so tech savvy. They might be poor communities that have poor access to health care facilities, or might not have great insurance and the ability to pay. If you have to pay for urgent care in a hospital, it's going to be very expensive.

I guess, with these digital solutions, it's easiest for somebody in Palo Alto who's very well-educated to adopt them. But there's probably actually less benefit there. How do you see these getting out and getting into the hands of the elderly, the disadvantaged communities, that might need them the most?

Robin: I'm actually seeing some of the large payers really starting to innovate a lot. Even before the pandemic, I looked at Medicare is really one of the organizations that sets the standard of care in the United States. As Medicare goes, the rest of the industry here typically follows. Because if they make something a standard of care, and one of our other major payers that are private, they're like, "Oh, we disagree that that shouldn't be a standard of care." It's hard to disagree against Medicare.

I've been watching Medicare, and Humana especially. Because Humana has a huge number of Medicare Advantage patients. They are really innovating. They released some new codes this year — CPT codes for reimbursement around a digital therapeutic using cognitive behavior therapy for conditions like opioid use disorder. So, they are now reimbursing for that, which is amazing. They've also come out with new code around remote therapeutic monitoring and respiratory health, as well as musculoskeletal system. So, that means physical therapists, or respiratory therapists, or helping patients with their asthma or their COPD. Now that does get reimbursed and covered by Medicare.

Another one that they released — I think it was just a couple of years ago. It wasn't this year. It's not as new as the other ones. It is what they call their new technology add-on payment. Companies like Viz.ai, which is an AI-imaging analysis company for LVO strokes, which is a very damaging type of stroke. If the software is used in the emergency room on a stroke patient who is suspected of an LVO, it's reimbursed at over $1,000.

So, we're seeing the big payers and other types of Payers like United and Anthem and stuff. We're seeing some interesting things going on, especially in telemedicine, really trying to build up the telemedicine side. Interestingly, Anthem — they just changed their name to Elevance Health — has released four digital therapeutics themselves to their patients for free. One around COVID care, one around Crohn's disease, one around diabetes, and one around oncology. Like we're seeing with the pharma companies, the payers are partnering with these startups. Anthem partnered with a company called Sidekick Health. Sidekick is the one who's created these digital therapeutics, and then they distribute it to their patients that they cover.

Erik: Okay. Interesting. So, that means that they're really finding value here, because they would not agree to reimburse if they didn't see the value in this. They're savvy business people. This is reducing their total cost of business.

Robin: Exactly.

Erik: There's another field of innovation, which is the pharma companies working with digital therapeutics to improve the effectiveness of pharmaceuticals. Have you seen movement here? Where do you see this going?

Robin: Absolutely. Yes, the companion apps to medications. So, I looked at Click Therapeutics. Click Therapeutics has one digital therapeutic on the market for smoking cessation. That's already being covered and all of that. But they've done two big deals with pharma. Boehringer, which is Germany, they have pledged $500 million for Click to develop an app around their medications for schizophrenia. Otsuka Pharma based in Japan, same type of structure deal. $300 million for Click to co-develop an app around major depressive disorder, which of course, Otsuka covers.

So, we are seeing pharma companies really putting — that's not an insignificant amount of money. It's not a massive amount for a big pharma, right? But it's not insignificant either.

Erik: Yeah, is that is that Boehringer Ingelheim?

Robin: Yeah, exactly.

Erik: It's interesting because this is public now. But here in China, they have a digital innovation team. The first product that they're launching is also this kind of digital therapeutics. Again, they're going fairly big. So, I think Vi has made the decision that they're going to make some big bets in this space, I guess, in multiple geographies, which is also interesting.

That's great. So, the pharma companies are moving in that direction. Do you see the innovation more coming from tech companies, and then pharma recognizing that and saying, okay, let's partner with this company? Or pharma also saying, hey, we're going to hire a bunch of data scientists and build our own apps? Where do you see the thrust of innovation coming from?

Robin: Pharma companies, especially in the United States, are very what they call opaque — the opposite of transparent. So, if they have hired a huge number of data scientists and are doing things internally, they're not going to tell us. But what I can say is that, typically, the business model around some of these pharmaceutical companies is they wait for the startup to get developed enough, or they give them some money and get it developed enough until it's a win. Then they either acquire it or give it a lot of funding.

This is not just in the world of digital health. This is how pharmaceutical is now innovating in small molecules and actual medications. Because there's so much coming out of academia. There's so much going on with AI for drug discovery, that you may as well let 100 different startups get to, say, human trials and then acquire them for a huge amount of money. But it's still a fraction of the price that they would have paid had they had 100 different projects internally. You know what I mean? So, it's just hedging their bets.

Erik: Yeah, it makes sense. 9 out of 10 startups will fail you. Choose the one that succeeded, and then you avoid that R&D cost of going through the process. Startups can innovate a lot more cost effectively than a large corporate.

Robin: Exactly.

Erik: So, if we think about some of the barriers that we might be looking at here, we'd be remiss if we didn't cover regulations. Obviously, this is a heavily, heavily regulated market. It sounds like some of these products are coming to market fairly quickly. So, does that mean — what does it take? If I'm building a digital therapeutics app, and I expect to be reimbursed, do I have to go through a formal FDA process? What does that process look like to get the product out into market?

Robin: Absolutely, yes. If the software is a medical device like a normal, which is a digital health product, typically — a digital therapeutic would count as that — it's less than a year. It's a much faster process than, say, a medical device that's a physical medical device. Of course, it is significantly less than a pharmaceutical.

So, yes, you can do it probably in under a year if you've gotten all your clinical trial data. If you're going to be reimbursed — you already know this, right? Because Medicare's already reimbursing for cognitive behavior therapy-based digital therapeutics. As long as you clear the FDA, you can just use that code and slide right in there just like another type of aspirin. You could just slide right into something. You know what I mean? That's not the hard part if the codes exist.

The FDA clearance, you do need to go through it for a doctor to be able to prescribe the app and then get that reimbursement. You do need FDA clearance. The times you wouldn't necessarily need FDA clearance is if you're for entertainment purposes only. You're not going to be going through the healthcare system, and it's not typically a reimbursable event.

Erik: Okay. If we had a very high level break down the solutions into two buckets, one bucket would be diagnostics, where you're not influencing the patient. You're collecting information. The second would be therapeutics where that could be an AR device that's helping somebody with PSTD or some other thing where you're actually influencing the mind of the patient to some extent. Is there a difference in the process there, or are they both treated within that same track?

Robin: I've not actually done the clinical trials myself. Whatever company I work for, we typically outsource that. There are companies in the United States that are experts at the IRB, at setting up the clinical trial, and taking care of all that stuff for you in a more cost-effective way than you can do it yourself. But from my understanding, it's going to depend if it's a 510k, if it is a De Novo — which means there's no device or software on the market yet that is a predicate. Meaning, it's not very similar — or it's a PMA. The PMA is the hardest thing to get through of this software class, say, a Class 2.

When you're dealing with a De Novo, it is going to be harder than just a plain old 510k. But companies like Achille did come under a De Novo. They were able to get emergency use authorization because they cleared the FDA during the pandemic. So, they got to market much more quickly. Achille has a video game for patients between the ages of 8 and 12 who have ADHD. So, the patient sits there. They play a fun game for 30 minutes a day. Then the software locks, and you can't use it again until the next day. So, that was a De Novo. It got that emergency use authorization. So, the De Novo slowed it down. But the emergency use fast tracked it to market.

Erik: Interesting. It sounds like there's always kind of this strategic decision. Do you want to be a first mover? Do you want to be a fast follower? It sounds like the path is a bit easier for the fast follower. That's always the case.

Robin: In the United States, yes. If you have something, be that fast follower. Absolutely. Because once you get to market, then it's all really about the marketing and sales. Your IP is probably locked up, and their company's IP is locked up. So, you're making sure there's no overlap.

Yeah, I would always want to be a fast follower. Problem is, most of the time I end up being that first. I worked in genetic testing in 2005, where we used four snips to determine your cardiac risk. Now they use thousands. I'm always at that very early stage. I worked on now cryopreservation and organ transplants with tissue engineering, 2011 or 12. Now the industry has moved dramatically in the past 10 years. So, I ended up getting at that very, very early, early a lot of the time, which is fun. But be that fast follower most of the time, if you can.

Erik: Yeah, that's right. Well, that's a hard decision to make. If you're passionate about something, you're going to do it. You're not going to wait and say, "I'll wait for somebody else to pick it up." That's the entrepreneur’s dilemma.

But that then raises another point, which is this concept of red oceans, fierce competition, and then blue oceans where you can differentiate. If it's a space — often with digital apps, it's like, well, somebody can reverse engineer it. They can change the code. It's often hard to patent things. Then the question comes, how do you defend a great new product? Aside from just being really quick on sales and building up your sales force, how do you see companies building in differentiation and trying to maybe not become monopolies but defend market space to an extent?

Robin: It's going to depend a lot on what you are doing. But if you want to look at developed companies in this space, look no further than Omada versus Livongo. They both started around the same time. They're probably both over 10 years old now. They are disease platforms for things like hypertension. They work with employers. The employers are the ones who pay them, and then the employees are the ones who benefit from these platforms for things like hypertension and diabetes. At that point, they're head-to-head, really competing for the same number of patients.

But the diabetes market and the hypertensive market is gigantic. 1 out of every 3 Americans is hypertensive. So, while they are direct competitors, there are still a lot of a lot of employees to go around. But once you get to that point, it is. It's all about marketing, marketing, marketing.

Erik: Yeah, I guess the market is big enough that it can accommodate 8 or 10 players even, potentially. So, it sounds about picking your particular customer segments, and then saturating those sales channels and building the strong relationships. Okay.

Robin: Yep, like anything else. That's exactly it. If you're not going to win through pure patent, like intellectual property, then you're going to have to win with marketing. Apple is where they are today because in part, they have incredible marketing and brand messaging and brand awareness. Absolutely incredible. Think about all the different computers on the market now. Why do you pick one over the other?

Erik: Yeah, absolutely. You could pretty much go and build an Apple. If you have $10 million, you could start manufacturing them. You're just not going to be competitive from a brand perspective. There’re obviously big economies of scale. I guess, those factor in here as well. The more patients you cover, you can spread your costs across them.

Robin: Exactly.

Erik: Let's see. Anything else we haven't touched here on the barrier side? Anything else that you see companies struggling with as they're bringing solutions to market — whether it's on the customer adoption side, the regulatory side, the competitive side? What do you see as the problems to overcome in this space?

Robin: As an entrepreneur in healthcare technology, yes, yes, yes. Everything is hard — just creating this brand-new market that most people don't even know about, or this brand-new product that people didn't even know they needed, getting it through the regulatory process. If you're not going to outsource that to experts, that's a whole new type of thing that you have to learn. Every step of the way is difficult. It's amazingly fun, but it is absolutely difficult. If you don't want to constantly be climbing uphill in a windstorm, don't go into entrepreneurship.

Erik: Yeah, it takes a certain personality, right? You have to love pain a little bit to jump into that. Well, let me ask you. I know you have your hands in a number of different projects right now. Are you involved as an executive in a startup right now?

Robin: No. Typically, now I just do advisory board or board of directors. Because what I can bring to a startup, it doesn't need me to be there 24/7 like my head in a computer. I do things like the strategy and the messaging for all the marketing. I know how to tell a story and get that into shape for people. Business development and sales. I cannot go — I make sure I network at a certain event where I'm giving a keynote, and I will pick up amazing investors or partners for more than one of my startups.

So, the only thing that I care about is that I don't overlap. So, if I'm working in VR radiology, I'm probably not going to then work with a radiology reading room that goes and builds them in hospitals. I'm not going to work with a competitor or something like that. But otherwise, it's just part time on a lot of these. Because I can do my job very quickly because of the way I've set it up.

Erik: Yeah, well, that's a smart way to structure this. You have a portfolio, and you get to spend your time having powerful engagements with people, and then helping your companies when there's a good fit. Let's maybe go through those quickly if we can. What are some of the companies that you're working with today?

Robin: I mentioned AlacrityCare. That's one of my favorites. The VR radiology room is called Luxsonic. They're already working with the Canadian Space Agency. It's incredible because you can have, say, eight physicians all at the same time in VR, in the same kind of room in virtual space. The radiologist's images — whether it's a CAT scan, or an MRI, or whatever — are in 3d. Every single person in the room can manipulate them and turn them.

So, if you're looking at a heart — say, oncology is a big use case — tumor boards are very difficult to put together that you need a lot of doctors all at the same time to be in the same space, or at least on the phone. Instead, doing it in virtual reality means: a, it's easier to schedule and b, it's much better because all of the radiologists and all the doctors on the cancer team are looking at the exact same images at the same time. It's like being in a movie theater. It could be like, "Okay. So, here's the heart."

Erik: Interesting. I can imagine. Before, you have somebody from Switzerland, and you have somebody in Canada in there on the phone basically trying to use words to describe the situation and maybe looking at pictures on a screen. Now you can manipulate and collaborate. So, that's great. I can imagine, if we think about the developing — the experts in the world are so consolidated in a small number of countries, right? So, if you're in Africa or you're in India, you have to get on a plain. That's not going to happen in most cases if you really want to talk to an expert, right?

Robin: Exactly.

Erik: Okay. Great. Let me ask also, then, what are the areas where you would like to be involved in? What are the things that you're really excited about, you'd love to find a company that you could work with in that space?

Robin: Vocal biomarkers, we haven't talked about those yet. A vocal biomarker is the ability to use a patient's voice as an actual vital sign. So, let me give you some examples. Sonde Health — what they have done is they have amassed a database of unidentified 4,000 unique features to the human voice. That means it's measuring 4,000 things that are going to be up or down or changed over time. So, being able to monitor someone's voice for neurodegeneration, depression, or bipolar anxiety and stress, insomnia or sleepiness, these are some of the applications that have already either hit the market or are close to markets.

Now, another really interesting use case for a vocal biomarker is very simple and very elegant. The ability to count the number of times someone coughs. Think about that for a second. It's such low-hanging fruit, but it's such an important thing. Imagine you've got a patient who's had bronchitis for a period of time. They go into their doctor, and the doctor says, "How's the new medication? Is it working?" Well, I think so, doc. "Well, how many times do you think you coughed yesterday?" I don't know. 700, 400. 30. We do not know. Nobody ever just sits there and counts, and then writes down the exact number of times they cough.

Well, a company called Hyfe AI — and there are a number of other companies coming out on the market now — they have a little microphone that clips to your shirt. It tracks the number of times you cough, and it records it so that the physician later on could go back into the software and listen to the types of coughs.

Now, the next step in that scenario — there's a company in Germany that I know that is doing this — is that they are categorizing the cough. Was that a whooping cough? Was that a wheezing cough or a productive cough? Once you know what are those categories, you know exactly how to treat it. It can help you understand or lead to a diagnosis. A whooping cough? It's probably a whooping disease.

Erik: Fascinating. I've been coming off a cold for a couple of weeks now. So, I've also paid a lot of attention to my coughs. Definitely, you can tell the progression of the disease as you pay attention. Interesting. Some of these things are so intuitive, because you also get the feeling. You get the feeling, hey, somebody doesn't sound quite right. My grandfather, he's taken a bit longer than he usually would to reply, or he's saying the same thing twice in a row. These little things that intuitively you pick up, but then to actually run the analysis, a human can't do that. You just aren't tracking.

So, let me ask then. Is that a device that you would snap onto your collar or something? It's just kind of listening throughout the day? How does that work? Because you want to have as much data as possible there.

Robin: It depends on which company you're talking about. But Hyfe AI, specifically, they have a little microphone that clips to the shirt. The company in Germany, they actually have two microphones, small microphones, that you put in the room. So, you're tracking them either at night or while they're home.

Different companies are going to be able to leverage some of the smartphones. Because they just need to make sure, at least in this country, that it's HIPAA-compliant. Amazon Alexa speaker is HIPAA-compliant, and it's a microphone. Things like the Apple phone, of course, we access our own data on that as well. So, that can be leveraged. I have not seen someone do that yet. But it is so early in the counting coughs industry. Over the next five years, that metric will definitely explode. That's easier than something like bipolar or anxiety.

Erik: Yeah, fascinating. Robin, I've got to say I'm really envious of your job. You get to just spend your time talking to fascinating people and imagining the future of healthcare. Really, it's a fascinating career. You've scoped out for yourself.

Robin: Thank you.

Erik: I think we've covered quite a few topics, so far. Anything that we haven't touched on that you think is really important for folks to know?

Robin: No, I could go on all day on things like vocal biomarkers and digital therapeutics. But really, what it comes down to is that shift in healthcare delivery to the home. So, we're seeing a lot of the insurance companies preferring and even forcing things like IV medication into the home, urgent care into the home, even hospitals at home. That's post-acute care. That's a really big deal. Because patients now, they don't have to go into a clinic where, a, they're surrounded by infectious disease. Anytime you go to a doctor's office, there's going to be germs around. That's just part of the trade. People heal better in their own environments. They're not being woken up. They have their comfortable pillow. They have their animals, or their favorite food, or their favorite blanket. In the hospitals, it's really difficult to heal and get a good night's sleep. So, the hospital at home is a big deal one.

But really, seeing the shift and seeing companies like CVS — CVS is now Aetna. It's one of the major payers. It's a pharmacy benefit manager, which means it packages up drugs, meeting out drug packages, and decides the cost of them, and sells them to the employer of insurance companies. So, it's also a retail pharmacy. I watched to see what they are doing, because they really are now a full stack healthcare company from start to finish. So, they have a device called the CVS Symphony. It's a smart speaker, wall sensors, and a wearable care button for aging in place. The sensors on the walls measure things like temperature and air quality, and fell detection. Then the individual, the elder at home, can press the button that's either around their neck, or they wear it on the wrist, or wherever they want to wear it.

But really, seeing CVS starting to really shift into the home as well and build up some of their primary care style offerings, this shift is here to stay absolutely. It's starting to be controlled by a lot of the really, really big giant corporations that want to shift it into the home. So, I'm excited about that. But I also know that takes a huge amount of revenue away from the hospital systems. That's a big deal, because we need our hospitals for acute care, right?

Erik: Yeah, well, you mentioned earlier that you haven't had the chance to travel so much in the past few years. I guess, most of us haven't. But if you get the chance in the next couple of years to visit China, please stop by Shanghai. I can tell you, it's a fascinating ecosystem. You can imagine here in China — you mentioned specific hospitals there. In China, the tier one hospitals are going to have 100,000 people, and then completely crowded because everyone wants to go to a tier one. Then the tier two and tier three are pretty low sophistication. You just don't have the hospital ecosystem.

So, when you talk about the value — I mean, this is a government initiative in saying, "We want elderly people to retire at home. We don't want them to be in the hospital. We don't have enough hospitals. We don't want them to be in elderly homes." It's all about then health care in the home. How do we keep people healthy and happy at home? So, please join me here in Shanghai in the coming years, when you have the chance. I would love to meet you in person. I think you're just a fascinating person, Robin. Last question then. What is the best way for folks to get in touch with you if they'd like to follow up?

Robin: Sure. So, I'm the only Robin Farmanfarmaian in the entire universe. I'm on LinkedIn. You'll find my website right away, or my democratizinghealthcare.ai website, which is about the book. All three of those places are really easy to get to me.

Erik: Wonderful. Thank you, Robin.

Robin: Sure.

Contact us

Let's talk!
* Required
* Required
* Required
* Invalid email address
By submitting this form, you agree that IoT ONE may contact you with insights and marketing messaging.
No thanks, I don't want to receive any marketing emails from IoT ONE.

Thank you for your message!
We will contact you soon.