Using Tech to Lower Healthcare Costs and Carbon Health's Strategy: In Conversation With Eren Bali
Nikhil and I recently had an interesting discussion with Carbon Health CEO Eren Bali on Carbon Health’s journey from software to clinics, tech strategy and future plans.
You can listen to our conversation on Spotify or Apple Podcasts.
Eren described Carbon’s journey from a tool to share complex cases to a new EHR to actually running clinics including how they designed an EHR from scratch and thought about selling software vs building clinics. He also shared how Carbon uses tech for a better consumer experience while enabling a cost structure that works for all payer types. The episode included some interesting tidbits on why Carbon stopped their popup trailer clinics, how they partner with hospitals, and their goal to be a last mile partner for digital health startups. Plus we hit on how Eren thinks about international expansion and his goal for “Carbon to become the largest healthcare provider in the world, not just the US.”
Check out a full transcript of our conversation below:
Today on Vital Signs we're excited to have Eren Bali, the founder and CEO of Carbon Health. Eren started a hugely successful education company previously in Udemy that went public and then went into healthcare to fix one of the other hardest industries around. You've built a massive primary care company, Carbon was reportedly valued most recently over $3 billion. Thanks so much for coming on.
Thank you very much. I want to highlight that reported valuation was more at 2021 prices, so just want to put it in context.
I feel like you might be the only person who's started in an industry that's more convoluted and harder than healthcare. So maybe that's the secret to building a big business: do the harder one first, then come to healthcare and see how it goes from there.
Education is tough, but our business at Udemy wasn't very hard. We had online video courses that people could buy. We did it in a way that allowed us to scale without the friction that would exist if you had to sell to small school systems. That’s totally different. Similarly, at Carbon I was hoping I'd picked a part of the market which is more of a consumer decision. But I think it is still an order of magnitude more complicated.
To kick off, it would be fascinating to hear your journey and what made you take the leap from Udemy to Carbon?
My interest started in 2013, when my mom had this completely unexplained stroke. And there's no sad story. She has recovered, just to highlight. But I had to take two, three months away from my Udemy CEO job. I went back to Turkey. My sister had arranged 12, 13 different specialists to look at what was happening. My mom was at home so she could not travel, but we would go from doctor to doctor with a couple thousand pages of documents. We had lab results, MRIs, PET scans, and my sister nicely categorized everything. She had folders and printed out the MRIs and highlighted specific sections. She's a physician, so she was able to provide a lot of context before other specialists looked at it.
Eventually one of the physicians (I think the neurologist) said, "The lab results look like sarcoidosis," which is an autoimmune disease, but normal sarcoidosis happens in the respiratory system. She had this theory that maybe it's happening in the neural system right now. She had heard of neuro sarcoidosis, but had never seen a case herself. She had read a PubMed article once.
When she applied the treatment, it worked perfectly so my mom recovered. Again, no sad story there. But meanwhile, I'm kind of useless, right? I was just tagging along. My naïve question was: If all the doctors are looking at these thousand-page documents in the same way, writing the same notes, and my sister has to prepare, why doesn't the EHR already look like that? Why doesn't it actually map to how providers look at medical records?
My sister shared some presentations from case conferences with me where they would present a complex case to other physicians. That presentation made a ton of sense. The 65-year-old with this problem; here's the findings. It was time oriented. It wasn't just the useless part of the EHR; it was the important medical records in a nice presentation. I'm a decent designer, and I made a couple sketches about how I would build the EHR if somebody asked me to. I made a couple sketch files and showed them to my sister. She said it would be a dream if things worked like that versus the reality, which was very far from there.
How true to the Carbon EHR today is it?
Honestly, the actual chart is pretty much the same idea still, from 2013. My original idea was pretty far from Carbon Health today. I was thinking first, what if you made a tool to allow providers to make these kinds of cases, not for EHR purposes, but to share with each other and create a rich PDF of complex patient conditions? So if the next person's mom has a similar problem, the doctor could search on a highly structured complex case database — kind of like PubMed but structured data oriented. That was the original thinking. I met my co-founder who started coding up some data structures and UIs to build those inputs. It looked interesting, but we had zero idea about whether this was a viable business, or if people would pay for it, or if doctors would use it.
It became pretty obvious that this idea that providers would spend time to create more complex cases like that was unrealistic. The thought switched to, "Okay, this has to be the primary EHR. This cannot be a complimentary thing they do on the side for the good of humanity. It has to be in the same place." Because that was the feedback we were getting: "This is cool, but this should be on my actual EHR. This is how it should work in the first place." So before pre-launch we switched the product to, “Let's make an EHR.”
My crazy idea was: What if we built an EHR from scratch, but instead of being an EHR vendor, we create a marketplace model. Every patient has one account and the providers join. If you have five different providers using the same EHR as the same patient, you actually have one account, and all your medical records are immediately the same across all providers. Imagine you have thousands of customers in the EHR, but you put them in the exact same EHR instance, like the same Epic instance, versus the traditional EHR, where they create disjointed worlds for each provider.
If you've been to 20 different hospitals all using Epic, you have 20 disconnected records. So my take was a marketplace-style approach to EHR. This would also have a lot of network effects because of the patient acquisition cycle. If you're any physician, you go on Carbon's platform and all of a sudden there are 5, 10 million users, and they can just book appointments with you. That would be very helpful.
I think we had some traction with that idea, but there were two issues: One is we did not have any actual patients, so the network effect was very abstract. This was very forward thinking for a provider. And the EHR was not robust enough to be used by thousands of practices. Building the EHR from scratch is really difficult. There's this massive surface area, and the feedback you get from customization will make you a really bad EHR.
The way we designed the EHR was interesting. We did not know how to build an EHR, and I did not want our EHR to look like all the other shitty EHRs, which is 100% of them. The way I managed this is we opened a small primary care clinic inside the office, and I made it forbidden to look at other EHRs. Literally, I'm not joking. It was like, “You'll be fired if you send me a secret screenshot of how Epic does e-prescription.”
So we brought like a hundred of our friends to say, “Okay, we are going to build a better provider.” We built a nice patient app. The patient app piece was honestly easy — that was kind of obvious. You want everything on your phone. That's pretty much the patient spec. And then the provider side, we started pretty simply. We only had a small number of visits so it wasn't a big problem. First we made our own unique approach to designing the EHR. What we ended up building was wildly different than how other EHRs look and feel and work. As an example, we had no idea how billing works, so we just didn’t build it.
You might be the first EHR ever to start with no billing. I feel like that's where they all started.
One of our biggest secrets was that felt like the most complex part. We just ignored that for a year. And the way we ignored it was that we just did not bill anybody. We collected the insurance but we did not actually bill claims and we really focused on patient provider experience. That's why the whole EHR worked and felt very different. We figured out billing probably three, four years into the company.
The secret to lowering healthcare costs the whole time was just forgetting to build billing modules.
Exactly.
Who would have thought?
It's pretty simple, honestly. That's how it would work if the government was paying for everything, right? So we focused on onboarding, scheduling, intake, getting more structured data. It's also a mobile EHR. We didn't originally anticipate this, but our providers use smartphones with the EHR all the time. Because when they're sitting and looking at a patient's case, they don't even want to open their iPads or notebooks, they just open it in their phones.
So the EHR ended up being so different from everything else. It's mobile first, real-time collaboration, very structured. You can write open-ended text but it's hard to write open-ended text on Carbon. The EHR pushes a lot of structured information. We invest in a ton of integrations with all major labs. Imaging is natively supported, prescription databases. So if a traditional EHR is 10, 20% structured data, 80-90% unstructured, like notes, we are the opposite. We are like 90% structured, maybe 10% unstructured data.
I imagine at this point in the business you had this question that I think a lot of digital health companies have, which is that you could probably have tried to sell that EHR to lots of other folks in the ecosystem or built up the clinics that you ended up building. How did you guys think about that path?
I think selling the EHR would have worked if it was today, because here is the thing that I think a lot of SaaS type of entrepreneurs should understand: If you're doing the traditional disruption theory and going from the low end of the market, smaller customers that Epic doesn't care about, that is how you typically disrupt industries. But in healthcare, what you call small customers, like small solo practitioners or small clinics, they're the worst customers you can have. They're very demanding, they don't have revenue, and they don't want to pay anything. You're competing with near free.
But if, for example, your early-stage customers are startups, that's a different situation. Startups have a lot of money and they don't mind paying it. That's why when you’re Zenefits or Brex, your early-stage customers are startups, which should have a ton of capital. That works pretty well. But if early-stage customers are small solo practitioners, it doesn't work well.
If you sell to startups and have a lot of capital in the bank it works. But if you want to sell to small business you're not going to be able to scale. In those markets you have to start with larger customers. That's the only way. And in an EHR world starting with a large customer is not viable. You are not going to be able to figure out what to build for them, so it's going to be an impossible bridge to gap.
But on the other hand, the actual moment I made the decision. We had our first customer, which was a visionary physician who had three urgent cares in the Bay Area named Cesar. He liked the software so much, despite the fact that it had all these holes. He said, "I'm going to switch all my clinics to this software because I believe in this." We spent another nine months building the revenue cycle and practice management, all the other “businessy” functionalities, because it wasn't a small clinic — it was a sizeable clinic group.
And we deployed it. A couple of days were chaotic, but by day three the software was stably running at the clinic. And when we did that, we replaced 14 different tools they were using with Carbon. They had an EHR, a scheduling tool, an intake tool, a video visit type appointment, multiple collection tools, a billing company. There were literally 14 different tools. I can't even count them right now. So we just replaced all of them with Carbon, and the teams loved it. Now they don't have to copy and paste your information from the EHR to the messaging system, from the messaging system to the video visit tool, the payment.
All that copy pasting creating bandaids between 14 different software packages disappeared. And the fact that we had no computers (it was all iPad-based) also felt good in the clinic. They were going from room to room with iPads in their hands. They could read what's happening with this patient. So there were all those really nice things we had to build that made a huge difference in the clinic.
One month post-launch I started seeing that almost all the Yelp and Google reviews were mentioning the application, the consumer app. That was a good signal. I'm a consumer entrepreneur — I'm coming from the direct-to-consumer world. I don't understand B2B that well, to be honest.
This explains the billing module part.
Exactly. I look at things like retention, which healthcare people almost never think about. I make my retention cohorts the way I normally like doing. I had their historical data so I could look at their retention cohorts pre-Carbon and post-Carbon. And three months later the retention rate was wildly different. There was a huge jump the moment we took over the clinics with the software. I calculated that you would increase the EBITDA margin in these clinics by 10 to 15% by just the software creating more stickiness with patients. And I said, "It does not make sense for us to own 1 or 2% of the revenues in the clinics. We are going to make a 10 to 15% EBITDA delta here."
Meanwhile, Cesar was worried that if we sold the software to all these competitors, his edge would go away. So I said we should just buy these clinics and open a lot more clinics. If we can make such a big delta in a clinic’s performance by creating a better customer experience and more stickiness, we'll dominate the market eventually. I have this very simplified view: In any market, if you have much better retention than all your competitors, you'll eventually win. As long as you can survive long enough, you are going to win the market if you have higher retention. So essentially we decided to build more of a first-party primary care play. We started urgent care, added primary care later, but it was all about the delta. The software was making such an observable delta in consumer behavior even though we did not control the clinics yet.
In the next couple years, the retention kept getting better and better. If you look at our annual cohorts — 17, 18, 19, 20, 21, 22 — each one is visibly better than the last year's retention rate. Honestly, I still think there's a big play to give our software platform, but now we have a couple million customers so this whole network effect is probably realistic to date and the software is probably more mature. I think today our original vision would make sense. The fact that we didn’t have anything to show for it was a mismatch between ambition and the reality.
It's probably a good segue here to talk a little bit about the tech piece itself. We talk a lot about tech-enabled services in healthcare circles, and the reality is that it's pretty difficult to figure out exactly where tech increases gross margins on a business. What scales linearly cost-wise with the number of patients, and what doesn't? When you saw these EBITDA margins grow, what was the actual thing that the tech was doing to enable that? Is it just that you're reminding people to come back in? Is it increasing throughput through the clinics?
A couple key things. It’ll help to separate urgent care versus primary care because it's a little bit different. With urgent care you have a lot of fixed asset investments. The more paid visits you can do per day per provider, your margin goes up linearly. For most urgent cares the main problem is visit volume. So the immediate impact was acquisition — having online scheduling, better integration, all those things. We have an integration with ZocDoc and a lot of cool direct-to-consumer functionality, which is built into the EHR itself. It's not just some third-party tool.
So that allows us to acquire patients a lot faster and because customer experience is better, the chance that the patient is going to come back is roughly 2.8 times higher when patients download our mobile application. Our mobile application is downloaded by a record 80% of our patients, which is roughly 10 times the market average. So acquisition, retention is obvious.
The other thing is a lot harder. In urgent care or primary care, when you have 5 visits a day, you can always provide an amazing experience. But as you start getting 10, 20 visits per provider per day, the experience starts deteriorating. Because it deteriorates, you lose your ability to acquire more customers. So you start getting churn; it is no longer a good experience. The entire success of EBITDA is: Can you keep the experience consistently good even when a provider is doing 10, 20, 30, 40 visits a day?
We sometimes go to 50, 60 or 70 visits a day per provider. We have virtual providers who take over the excess capacity in the clinics. If you have 60 visits a day, some of them are going to see virtual providers. There are all these things we have to do so that as volumes increase the patient experience doesn't deteriorate.
There are simple things. We don't allow patients to message providers directly because patients ask a lot of time-consuming, unnecessary questions. It's more like a Slack-style group messaging with the patient, so that 95% of the time our support team runs the visit, and only when they mention the providers do providers look at messages. These are things we designed from scratch for very high throughput clinics. Throughput is the main thing. Doctors cost what doctors cost. If you want healthcare to cost less in this country, either we have fewer visits or doctors have to do more visits per day.
The original company that we're all familiar with in this space was One Medical. I think they ended up choosing to monetize in three ways. They charged consumers a membership fee on top of their visits. They then went into this employer business where they built either on-site or near-site clinics. And then they also had these hospital partnerships. You’re in many ways the next generation of this kind of company, how do you think about these business lines for you?
One Medical doesn't have three different ways they charge. There's only one business model, which is that they essentially charge three times more than what a doctor charges. They do this by latching on to academic health system contracts. They partner with UCSF and take UCSF's primary care contracts, which are scheduled for an academic hospital, mostly treating super high-risk elderly patients. If you take those rates, which are typically two, three times what an average doctor can charge, and you apply them to a 20-year-old tech employee who's pretty healthy, there is a good margin to be made there.
So that business is lucrative. And One Medical’s business model is pretty good in the Bay Area. The only problem is they couldn't replicate it outside the San Francisco Bay Area and maybe New York. It relies on payers accepting this massive delta between what they pay other doctors versus One Medical. So the subscription is pretty much nothing. Their entire revenue comes from the rates being so high. In our case, subscription is, again, a small percentage. For us it was more about, “I don't want to do this because it means you are making healthcare more expensive.”
Whenever you talk about downstream behaviors, that's all BS. You all know that. If you want to make healthcare cheaper, you have to charge less money. You have to provide care at lower cost. Our intention wasn't being the cheapest provider necessarily; the goal was to be the average cost provider, like 50th percentile. Let's not be any more expensive or cheaper, but with a premium customer experience. I know that's possible, because right now we are not even on average with One Medical. Carbon's average rates are 35% right now, so we are actually cheaper than the average doctor's office, with a premium experience.
We know that it's premium because a quarter of our patients are very high income. 25% of patients make more than $200,000. Rich people like Carbon because of the clinic look, consumer experience, people are friendly, the app is great. But what's really disruptive is the experience that is also accessible to people making a $40,000 household income. For us it was less of a business model choice and more about, “Can we make a decent healthcare experience accessible to maybe not 100%, but 70, 80% of the population?”
You talk a little bit about One Medical being a great business model in some capacity. They sold for what was probably not a fantastic outcome for everyone, and had been doing this for 12 years. How do you think about growing into the valuation you have especially if you’re charging Medicare rates?
One Medical's business is very lucrative in the San Francisco Bay Area. The problem is that it’s impossible to replicate in other parts of the country. Their problem is not margins. The fact that they don't disclose any regional margins or clinic margins should tell you that they’re probably very good in some places, because they disclose the average, and a lot of their clinics are in the San Francisco Bay Area. That business model relies on a high density of young, affluent customers, and in the San Francisco Bay Area, there are a lot of young, affluent customers who live next to each other. But you can't do this in Columbus, Ohio and Virginia and Alabama.
They can’t prove that they can expand this to an average American city. I think that was the entire limitation of One Medical. Carbon's business model, because it does not rely on a very high reimbursement (we are looking at average reimbursement rates, not asking for more), our clinics work in pretty much every city. It's not as lucrative as One Medical in terms of revenue per visit. But that reduction in lucrativeness in margins comes with a massive scale advantage. We could open 4,000 Carbon Health clinics in the country if we wanted and they would all generally work.
Are they all going to be freight ship containers? I'm just kidding.
That was a great idea. We shut down the popup clinic, the popup trailers, in a very unusual way. The model was working well with the popup clinics. It didn't work in an unexpected way. It was all the security that was the problem.
You got robbed too often, especially in San Francisco. Margins were decent but maintaining that security expense was killing it. We had to put actual security in that just killed the margins there.
Wow. That's surprising because a lot of people talk about, “How do we bring care closer to people?” One way you do that is through these popups and mobile clinics. These are a lot of the unexpected things that I think come from trying out models like that.
Yeah. I still believe the same thing. I think there's a really good model to bring it to some places. But they're not good for San Francisco, LA downtown type of areas. I think there's some version of it. Especially really rural markets, like islands which have no access to healthcare in Hawaii. So I think that's a good idea. We just did not have the budget to be able to figure that out. And original placements didn't work because of the security. People actually stole the trailer itself. Literally, our manager was driving down the highway and our trailers were purple so it's pretty easy to notice and he saw the trailer driving the other direction.
The Fast and the Furious franchise is really going in a different direction nowadays.
Exactly. They did the job they were supposed to do, which is in the height of the pandemic they played a massively important role for communities. And in that time period they were no brainers. But I think in urban areas, full-size clinics are better. In the rural areas we would probably bring popups back.
That makes sense. You talk about how the model can scale in tons of different geographies. Is every patient in the US a potential Carbon patient? Do you think, "Hey we're never going to go into these kinds of environments or these populations.” You bought a diabetes management company. You have primary care, urgent care. Who isn't a Carbon patient?
On the payer side, the only thing we haven't been able to support is Medicaid in California and New York. California's horrible fee-for-service Medicaid contract rates makes it impossible to be a for-profit company and make it work. The only way to make that work is when you do managed Medicaid contracts, which is what companies like Cityblock and the other companies are doing. Managed Medicaid is reasonable; that's what we are trying to do. But just fee-for-service Medicaid in California is impossible without federal reimbursement. Other than that, we have been supporting every payer we can support, which again, was part of the mission.
In terms of the services, what we’re trying to do is take over 70 to 80% of all of your interactions with the healthcare system. This means we want to take the lowest complexity but highest frequency part of healthcare, which is things like urgent care, primary care, mental health, some women’s health, and current care management. That's what we want to do. I don't think patients think of a difference between urgent care and primary care. One Medical is called a primary care company, but 80, 90% of the visits there are what you would call urgent care visits.
I think that separation is abstract and unnecessary. When you need care, we want to handle it most of the time, and only 20 to 30% of the time we want to pass you to the hospital. It's not just services; we also have extras. We are putting ultrasounds in the clinics. We have EKGs, we have ER physicians on call. We are trying to maximize what we can do in the physical box. We go to the hospital as infrequently as humanly possible. That is the design of the services.
I know you guys have a few partnerships in place. You have one with hims, you have one with John Muir. If you guys are trying to take 80% of interactions, what is the point of those partnerships? What do they actually look like? Are you guys referrals to these hospitals? Is hims referring people to you?
There are two wildly different partnerships. Hospitals, the intention is 20%. The 20% of the time you have to go to the hospital is one of our biggest issues. It's really hard for us to place patients with specialists. So a lot of people think about specialist discovery and cost measurement and steerage, but the reality is that most of the time it's not possible to find any specialist who's going to see you in a timely fashion. There are a couple components to the hospital partnerships, and we have six of them right now. Stanford, Froedtert etc. We do a deeper care coordination integration. We are EHR integrated with their Epic in the hospital, which allows the patient records to be passed more easily. And then we have operational work to pick which providers are available and all this data they want.
Also, the hospitals have some requirements for these patients. They require us to know which patients to send them or not send them. It's not about sending them more patients, otherwise they wouldn't partner with us. We are trying to send them fewer patients. We send them only the patients that are relevant to their services. Stanford is already 97% full, or some ridiculously high utilization, so they don't want patients who can be seen in an outpatient clinic. They only want to see if you are a match with their complex services. Those hospitals are typically the only providers who can do it anyways. I think a good trend in general is specialty providers pushing their eligibility requirements to the provider, so providers know who to send to which specialist.
That's the main thing — this whole care coordination agreement. But we have some other interesting partnership opportunities with them. We're trying to de-crowd their ERs by design. This is one thing that a lot of healthcare nerds miss. People say every ER visit costs $2,500 on average. If you reduce ER utilization, that means lower total cost of care. But the reality is a lot of ER visits do not have to be ER visits. The ones that we can carve out are not necessarily the most expensive ER visits. They are the least expensive ER visits. The ones which will have to be in the ER are the most expensive ones. Hospitals don't really lose money when you remove the least complex ER visits. It's marginal cost of care reduction, and mostly much better customer experience, and it de-crowds the hospitals without lower acuity patients.
Hims, just to be clear, is a completely different thing. We have two different partnerships. The other type of partnership we have been doing is a better example. There are all these digital health startups. They have built nice services but they often need a physical clinic to send their patients to. And we are the only company with both a physical location network and good software integration, so they send us the patient, we track them, we track who sends which patient, we manage them. So we are trying to become the last mile for digital health startups. I think that is going to be a very big thing we do, especially as we get closer to a nationwide network.
All the virtual telemedicine companies got to hand off at some point, right?
Exactly. We want to be that easy handoff point. We are not trying to take over virtual from them, but if you just need some lab results from the patient, and then let's say you want to also do a vaccine while they're in the clinic, send it to us, tell us why you are sending it and we'll do exactly what we are supposed to and then push the records back to you automatically. That's the vision.
Let’s talk about your clinic acquisition strategy. I feel like in the last few years the acquisition strategy was pretty popular, especially with people raising at tech multiples and buying physical assets at lower multiples. How did you think about the financing piece of this, when you're now doing brick and mortar clinics that might actually generate cash but then you also have an R&D component?
I think acquisitions for us is a little bit different. There's an acquisition playbook that PE firms know better and Optum. They just consolidate different hospitals, different clinics, and then that is an acquisition arbitrage game that companies do. But in our case, a majority of our growth has been organic buildouts. Acquisitions is mostly for two reasons: One is we have seen a lot of providers who were retiring or were burnt out from the business side and they just wanted to pass their teams to what they considered good hands. And we were, in their mind, the good hands. They did not want to sell to PE firms. We were going to take care of their team. A nice retirement option for providers was one piece of it.
One thing I was worried about was, if you open 2,000 locations, that likely means one or two thousand clinics go out of business. So if we’re exclusively growing by doing our buildouts, we're going to push a lot of small businesses out of business. And often you see really good teams — we would hire them. The cost was not very high so it made sense. The other thing is that issues contracting is a massive friction for us. When we go to new markets for sometimes 12 to 18 months, we do not have contracts with everybody. That's a huge overhang on the business.
One hack is when you go to a new state, you buy a small clinic group and it comes with all the insurance contracts, so we don't have to spend 12 to 18 months playing catch up. For us it was mostly about expanding contract coverage in new states and occasionally there were teams we really, really, liked and we wanted them to be part of Carbon Health. Our acquisitions were mostly a co-hire type.
The one difference for us is, even if you think about the acquisition arbitrage type of game, the multiples go up when you're a larger group. If you just integrate 500 clinics into one entity, it is objectively worth more. And when we acquire clinics, unlike other acquisitions, we integrate acquired clinics in six weeks from signing to close. Because the way we integrate is we stop doing everything we’re doing, we put our software in, and boom, that’s it. That six is mostly staff training. Retrain the staff on our clinics because we do more in clinics than traditional clinics. It's really rewiring their brain to be more patient-centric and then catching up the skill gap.
We did a good job productizing the medical record migration piece. That is pretty easy; that happens pretty much overnight. When you take these existing clinics and put in this unified software platform, you could argue that if we bought 2000 clinics and spent a couple billion dollars, it would be objectively worth more than the cost of individual clinics.
One thing I want to hit on before we move to the quickfire round is that you grew up in Turkey. I think you mentioned some of the context in which you were first exposed to healthcare, from an international perspective. Do you think the Carbon Health model would work internationally? Is it something that might be in the cards?
It will 100% work outside the US. My pitch has always been that we want to become the largest healthcare provider in the world, not just the US. We had a lot more interesting options where large health systems or other companies would want to use our software. The software piece of it is very expensive and the US might be one of the few countries where you can subsidize the R&D investment domestically. Ideally, our R&D investment could be applied internationally.
Turkey probably is not the perfect market for this, but there are a lot of other markets, like the UK, Germany, New Zealand and Australia. There are a ton of countries where this is a no-brainer fit. We were going to start the international expansion in 2025, but with the current capital market environment, we had to extend the timeline a little bit. Probably put two, three years on top of that for me. But eventually Carbon Health is going to be a global healthcare provider.
We like to end our podcast interviews with a quickfire round. I want to kick this one off with one we ask all of our guests, which is: What's one of the things that's often talked about in digital health circles that's over-hyped and one thing that you think is correctly hyped?
Value-based care, even though I'm a big fan of the concept, is massively over hyped. This whole concept of, “I’m going to do this preventive thing.” For every one example of that being accurate, I hear 99 examples which have no chance of being even remotely accurate. A lot of prevention is not a bad thing, but the reality versus what people are pitching or assuming is wildly different.
At the end of the day people just say, "Oh our healthcare is just for fixing people. We should do more preventive care." That sounds pretty cool, but how are you going to prevent a viral infection? How are you going to prevent people from falling and breaking their legs? Again, not a bad thing, we believe in it, but it is 10X less critical than people think it is.
And then what's appropriately hyped? Or under-hyped even.
Tech-enabled healthcare works. Tech-enabled healthcare providers are hot right now and I think that trend is going to continue. I really like this. Companies are taking specific verticals of care. I think it's appropriately hyped. I would say mental health is appropriately hyped. Not lucrative mental health benefits for white collar tech employees, but serious mental health, I think is appropriately hyped. But the version that raised all the capital is over hyped.
You had to deal with a lot of crazy policies that make up US healthcare. If we gave you a magic wand and you could make one policy change to US healthcare to improve health and lower costs, what would you do?
I would make it a requirement for all payers and providers to publish their insurance contracts. There are all these laws about how hospitals are required to publish their listing prices, which are completely BS numbers. The thing that you want for transparency is actual contract rates, which we know. By the way, it's not even hard. I can send a patient to Sutter Health and look at the bill and say what their contract rates are. It is pseudo-public information that should be forced to be published.
If you are a provider, and you are going to argue that you should be getting paid three times, four times more than other providers, you can argue that. But you should argue that publicly when all the numbers are out. Because right now there are a lot of providers providing fairly affordable care and the money they are saving by providing affordable care is just subsidizing people who provide extremely expensive care.
You're newish to healthcare but have now been doing it for a little bit. If you were talking to someone who's starting in healthcare now, what's one thing you wish someone had told you when you were just getting started?
When we hire new employees, a lot of our employees are tech employees. They come from technology companies. And they can't imagine how bad existing healthcare infrastructure is and how bad the market is. When they want to work in healthcare, they just work on really acute problems. We hire people, and they think the way to solve the billing issues is just better machine learning algorithms, predictions. They might go through really advanced stuff because they cannot internalize how bad the basics are.
It took me several years to figure out that the basics are so bad. Just solving that is a huge leap from where we are. Once we figure out the basics, maybe then they will become the bottleneck. But otherwise it's hard to understand how bad payers are, how bad billing works, how bad their infrastructure is, how uninterpretable it is. They cannot imagine how bad it actually is.
When I was getting started in healthcare, the thing that was always surprising to me is that the deeper you go trying to understand why something is the way it is, you hit this impasse where you're like, "This logically doesn't make any sense, right?" Even for the ideological thing that you're trying to achieve, this is not doing that. Like, why does that exist? And then there are so many examples that you're just like, "Oh, it's like a historical quirk. Oh, it's like it was just done this way and now everyone just does it." And it's rampant. We had wage freezes in World War II that led to getting healthcare through your employer.
There are random historical facts, which I don't know, several years later caused the healthcare system rules. And partially because nobody went ahead and changed them. By the way, on the policy side, I think something that would really help fix the US economy, healthcare-wise, is if we make it so that plan coverage is completely individual — as in, you pick your plan and let the company continue to pay a certain amount of dollars per month in your plan. But the plans need to follow the patient. Employers cannot have exclusive selection. Individual coverage should be the default. And then if employers are allowed to do pretax healthcare coverage, which is a historical artifact from the second world war... they should just put money into your existing plan.
Well this has been a ton of fun. It’s refreshing to talk to someone with such wide ambitions. Rarely do we hear someone say, "Well, we'll start in the US, but we're going to go international." So I think the magnitude of what you're going after is fascinating. Where can folks go to learn more about you and Carbon Health, if they're intrigued by our conversation?
If there are startups or other healthcare providers, our doors are open. We are very open to partnership, collaboration. I want to believe that I'm ambitious and realistic at the same time. We are realistic by realizing that we have to build clinical services and we have to do software. It's just going to be step by step. But also, there’s something I heard from Marc Andreessen: "Disruption always takes a lot longer than you think it's going to take. But it also becomes more impactful than you think." I believe that we are being realistic, but also ambitious. I would say to your audience that if there's something we can do to help their business, their thing, we are open to doing it. I think our software should be mostly for internal needs. I think there's a worldview to license that to other providers, including startups. Hopefully that view will be something we can talk about more publicly sometime soon.