Founder Stories: Mario Schlosser, Co-Founder and CEO of Oscar Health

I don't know if I followed directions to sit in the correct chair. Somehow I feel like we should have had more like the Beatles Revolution or something in music here. Yeah, toxic.

You too.

Toxic in healthcare is not the best mix. But then again, that for a lot of people, talking about healthcare and insurance is kind of a toxic topic these days.

It's too expensive and too complicated, that's for sure.

Too expensive, too complicated. I, I was thinking this morning, um, Marty, uh, Chavez from Goldman Sachs had said, you know, if something is interesting, it gets regulated. Yeah. So healthcare is highly regulated, but when you talk about health insurance, it's probably the least sexy thing to people. So to be a startup in healthcare and a tech-oriented startup in healthcare sounds really sexy, but what does it really mean? Tell us about your business and your business model for people who aren't familiar with Oscar.

Yeah, the short story is that in 2012 when we started the company, I was going through pregnancy with my wife, first kid, now 6 years old, about to turn 7. And so we couldn't really figure out how to operate the healthcare system. It was just too complicated. It was too costly as well whenever we did then utilize services. And so my background is computer science. I thought that the insurance company really, in the way the US system works, should have lots of information data about me as an end user of healthcare and the entire system I can use and should be using on my behalf to simplify the system for me. And I really didn't think that was happening at the time. And in many ways, the insurance companies seemed to be content with getting me on as a member and then never talking to me again. And so the simple idea behind Oscar really has been and continues to be that we want to really be more the entry points to healthcare for members of Oscar. And we have, as an insurance company, a real economic model as well in making sure we get you to the right place at the right time. And the more I think health insurance in the US shifts towards individualized markets, which still isn't the case at all. Most of the insurance will come through the employer, but it's shifting, I think, towards that. The more we can push out the time horizon that we can invest in people's health for— ideally, the dream is that people join us when they're, you know, 15, 20 years old, even younger, and we can sort of take a 50, 60, 70-year time horizon. We say, okay, what should you do today? What can we incentivize you to do today? That then will keep you healthy and out of the hospital in 60, 70 years from now.

And keep you with Oscar. So—

And of course that.

Talk a little bit about what kind of markets that you've been in and what your growth has been here over the last few years. You started with just a handful of markets. Yep. As the Obamacare exchanges, the Affordable Care Act launched back in 2014 and you've continued to grow. What's your membership like now? How many markets are you in and how many markets are you eyeing?

Yep, so about 260,000 members now, about $1.4 billion in premium revenues this year. And that's in about, I believe, 12, 13 markets and cities. And we generally go city by city when we launch into new markets. The reason is that you are regulated on a state-by-state basis. The licensing process works on a state basis, but the difficulty really in rolling out a, frankly, healthcare product and therefore also health insurance products is to build the right network of physicians and hospitals. And that really is extremely local to this day. Healthcare in the US is very, very different on the West Coast, the East Coast, in terms of who takes how much risk in the value chain, in terms of who owns the delivery systems in the particular markets, in terms of who sells the insurance products. And so that's so local that we got to go city by city there. And then for next year, we have quite a bit of expansion planned again. And the biggest chunk of the business, I'd say about 85% of the business, is still in the individual markets. That's the so-called ACA, Affordable Care Act, Obamacare markets. The rest is right now in the small group market, about 20,000 members there. And we're launching next year into the Medicare Advantage markets. That's the market for 65-year-olds and above.

Which is growing 10,000 people a day.

Exactly, quite nicely. And very individualized. You know, you really join, you yourself are making the decision to join or leave an insurance company. And that's the market we want to be in, in the way that market functions.

So in a sense, you, you hope you can provide coverage in your markets from cradle to retirement.

Exactly right. That's exactly right.

So one of the things that you talk about is using technology. All the big guys talk about using technology. In fact, when you listen to their presentations these days, you'd think that they are saying, hey, we're cool, we're Silicon Valley. They don't quite take off their ties and wear jeans. But they try to sound like that. How have you actually used technology? And talk a little bit about building your own payment stack.

Yep.

And the backend to really change the game for you. How did that change the game for you?

First of all, one funny thing I always notice is that whenever another health insurance CEO comes to Oscar, they are now told by their staff to take the tie off before. It literally happens. So I think this was probably the most crucial decision we've made since the, since the very beginning. We, as an anecdote, in 2012, when we were trying to get through the licensing process for the first time in New York State, there hadn't been a newly licensed commercial greenfield health insurer in about 20 years or so. It wasn't the common thing people were doing and getting licenses for. When we were doing that for the first time, we had to piece together really an infrastructure, a claims system, a network, an eligibility vendor, and all that kind of stuff that insurers use. And so we, let's say, on the claims side, for example, the adjudication of the bills we get, what we pay, what we don't pay, and so on. We were getting visited by all the kind of big vendors in this space. You know, there are tens of billions of dollars of companies in the space that just rent out claim system technology to insurance companies. And we ended up going with a very, very small 100-person company in Cincinnati, Ohio. And the reason for that was that the woman who was the chief programmer for that company was in the pitch herself, and she had like a team of 2 or 3 other programmers or so. And so right there, I was able to talk directly with her about, okay, what if we try to do this with the claim system? What if we try to do that with the claim system? And could go directly into the source code and look at how they would, how they would sort of like change this. And we, we realized there that we needed the ability of changing quickly, you know, if something isn't working, if we try to want to try something new on the incentives and plan design sides to go into the system and change stuff quickly as opposed to go through big change request processes and pay a lot to a consulting firm to do that with a big vendor on the outside. And we then insourced the operations in that system in 2017 and have in parallel been building our own system from the ground up. And so in 2017, for example, our so-called auto adjudication rates So when the claims come in and you need to figure out how to pay them. Exactly. And the auto adjudication rate is basically the percent of claims that no human being looks at before the claim gets paid. It was about, I'd say, 80%, which is a bit below industry average, about 85%, I'd say, industry average. And it's now all the way up to 93%, 94% or so, which is quite a bit above average now. And it lets us pay claims faster and therefore deal in a better way with the people we pay claims to, hospitals, doctors, and so on, but also make changes more quickly. For example, we, another kind of sort of outside-in anecdotes, we have been, we try to benchmark ourselves on the fraud, waste, and abuse sides as to what are common patterns of claims we shouldn't be paying for. You know, some dermatologist prescribing half a million worth of kind of dermatology topical cream as a little actual case. And stuff like that. How exactly do you navigate these things and what do you deny there really? And that's best done by having a firm look at a couple of insurance companies and look at what we do. And when we see new patterns, we will implement. It takes us a couple of days to implement that. And we know from people we would then work with, another insurer would take a couple of months to even just study the issue and look at the data and then make a change there. So the ability of changing more quickly the systems is extremely important there. Um, and it lets us also now on the other end of the spectrum, the member experience sides, I think that's where I was headed.

So how does that help you to actually be able to look at the patterns in terms of the type of claims? And you know, everyone talks a lot right now about using the data in order to impact behavior and to help do more prevention.

I think the, the biggest issue oftentimes in, in in most other insurer systems and in hospital systems for sure, is that they are still extremely siloed. There oftentimes isn't even a good ontology taxonomy tying claims to drug data to, you know, conversations with the customer service staff, for example. None of these things end up in the same database oftentimes, even as simple as that. And even if they do, they often end up not being connected to the same, again, sort of organizational taxonomies and stuff like that. The specialties that your doctors would be identified by on the public search of an insurance company would be often very different than those that come into the claim system and again different than those that would be sent to the regulators and things like that. And for us, we just spent the last 7 years making sure that all comes into the same place. And so we can now, for example, when you have a free telemedicine visit, which one of the things we've done since the very beginning to our own staff, You have about 50 physicians on staff that do telemedicine visits.

You do that with Teladoc.

That's overflow. But we have about 50 doctors internally that do the majority of the cases. When we get these interactions in, then they get routed to the right doctor in that team based on how medically complex the member is and based on how, for example, we still need to submit or we need— we suspect there is a, there's an undiagnosed condition with a member that we still want to have a conversation about with a physician. And that requires us— that classification of medical complexity is something that runs pretty much in real time based on our claims, our drugs. And even when the doctors call you back, they can see, for example, what searches you've done recently in the mobile app and gives them an idea more quickly as to what you recently tried to do and look at and interact with, with the system.

You also have concierges who work with people one-on-one. So how much of this is AI, how much of this in involves real people having to do that, and having real people, does that add to the overhead and making it affordable?

I think the algorithms right now are mostly in our stack, are mostly oftentimes in very pragmatic places in the claim system. For example, I'd say 4 or 5 percentage points of that increase in the auto adjudication rates came from simply better matching of names of a doctor on a claim versus doctors we think we have in the network, you know, because that's sort of where the healthcare system is just bizarre. There isn't a great, widely followed identifier of physicians. I mean, there's the NPI, for example, but even that isn't oftentimes super tightly on the claim submission side. And so therefore we have to do a fair amount of actually literally name matching, information matching to make sure we pay the claim to the right person. That's where oftentimes the sort of AI is in a sense, simple machine learning, simple kind of fuzzy logic and so matching in there. On the member experience sides, we try to serve up signals and alerts to the humans in the loop. And so for example, we do a lot of work on provider data and making sure we understand what doctors do and what specialties they have and what they do best and so on. But if you have a bit more of a complicated ask about who you want to go and see for a particular reason, we have an escalation team sitting behind the concierge teams that then goes in and looks manually at our own data and finds you the right physician. But it's, again, built on top of a very good taxonomy and a good data collection efforts. But it just makes still more sense to then say, OK, let's really get this right if you want to go to a particular physician for surgery or something like that.

Yeah, something that you've been doing, you've been working in specific markets and you find the right partners. How long before Oscar can scale? Is becoming a national insurer in your sights, or do you really want to focus on doing it really well in certain areas?

I think we want to do it well in certain areas. I look, I mean, I think we want to take the product to as many people as possible, of course, over the years. But again, as I mentioned before, healthcare is just unbelievably local. In Southern California, physician groups that take risk in our quasi-insurance companies run the show for the most part. On the East Coast, nobody takes risk, and if anything, it's coming more from the hospital leadership, from the hospitals, to go a bit more into that risk-taking market. So given that it's so local, the patterns change in every city we're in. For us, in a market like New York City, Maybe 5 or 10% of the members we have are on government subsidies in the individual markets. So they make— the majority makes more than $50,000 a year as individuals. In a market like San Antonio, about 85% are on government subsidies.

So is New York a more profitable market than San Antonio?

No, interestingly enough, not.

Because prices here are so much higher?

Partly because of that and partly, frankly, because the ACA has a pretty smart design around how risk adjustment, risk transfer works. You get compensated as an insurance company for managing a member with a certain disease pattern who you can get healthier than the other guys in the markets. That's what this risk adjustment transfer formula, for the most part, does pretty well. And I think, yeah.

You mentioned the ACA. The other big issue Now we've moved beyond repeal and replace, but the other big issue is, as you're going into the Medicare market, is whether or not there's going to be rebate reform. And the Trump administration has said they will. You don't have a PBM. How does that regulatory change— these are the discounts that the people who do the pharmacy benefits pass on to insurers— how do you handle rebates? How will the changes in the policy impact you? And how is that shaping the way you look at your business?

Yeah, the rebate issue hasn't really been an issue for us because any drug costs we have go straight into the premium. And therefore, essentially, if we have a lower cost than we thought in pricing our plans and we utilize drugs, we will have a lower price next year for the market. We will put a lower premium to the market next year. We generally have a medical, so-called medical loss ratio. That's all medical costs divided by premiums, includes drugs. About 30% of our costs are drugs. It's a really high chunk of that, of that cost. Our medical loss ratio this year is in the kind of like low to mid-80s, and that tends to be pretty similar because all the markets we're in. And if we are in a market where it goes into the 70s, for example, we would go back in the next year in that same market and reduce the price. And so I think that's one of the dynamics you see in the individual market, which you just don't see in the employer markets. If you are a large employer in the employer markets, you will not switch insurance companies on an annual basis because you'd have to tell your 10,000 employees something different is happening. In the individual market, there's a lot of price competition, and you really— you saw last year for the first time, despite the fact that the administration is still the administration, tweeting is still the tweeting, And, you know, the uncertainty, still the uncertainty that rates actually came down. And that's a pretty remarkable fact. And that is—

and insurers like you actually expanded in markets.

Yeah, exactly.

And that was to do that again.

Right. And that happened because there was real price pressure on this market. People will switch if they get a better deal somewhere else. And a better deal oftentimes, I think, has more from just a lower premium more towards understanding of, okay, what do I have to pay on an ongoing basis throughout the year? And so next year we're doing some interesting things like on the drug side, you asked, a $3 drug tier where we kind of selected a couple of 40, 50 drugs or so that people have to take a fair amount and put a much cheaper copay on those and therefore can essentially tell you in the app, hey, if you're going to doctor for a particular reason, Show this to the doctor so you get one of those drugs and can save money on getting those particular drugs. And those are also then the drugs we can go back to our PBM, our pharmacy benefit manager, with and say we need to make sure that those drugs in particular we have a good deal on and that we can manage this in particular tightly in adjudicating those drugs.

So we're near the end here. I want to go rapid fire. A lot of talk about Medicare for All. Could Oscar exist in a Bernie Sanders Medicare for All?

I have said this many times, I think, and if somebody was able to build a more efficient single-payer system, I'd gladly give the company back. You know, healthcare is too important for us. I just, I really don't believe that that's the issue. I think if you look across the world at places like Switzerland, Germany, I'm from Germany originally, they're all mixes of public and private payer and in hospital as well. Those tend to work the best. The Canadian UK system do certain things well, but they're not the best from an outcome and cost perspective overall. And so therefore, I think if the compromise is Medicare Advantage for all, yes, then we could certainly exist. And I actually think that would be good for the system. It would be more innovation, more competition in the system. And I think that's, if anything, that's probably where it's trending. But let's see.

Okay, so here's the last 2 minutes. We've got the game that folks like here. Overrated, underrated, hot or not, swipe left, swipe right. So let's start with Blockchain.

Oh, definitely overrated. I have to always say, I wrote a paper in 2003 that's at a big computer science conference that talked— that was in the early days of distributed hash tables and kind of putting information in chaotic systems and stuff like that. So the technology isn't even all that new. I'm not claiming that we invented it, but it's not that new. It was around at the time. That's why I mention it. I definitely overrate it in the way we're talking about right now. Telemedicine, that's still underrated. I think it's right now mostly tele-urgent care, and we are doing a lot of work to expand it towards the entire spectrum. Stay tuned for that, for more to come from us by the end of this year. UnitedHealth, I mean, you cannot argue with unbelievable market cap success, you know. I think that's— I would not say that they've always used their their market power towards the best of healthcare costs and healthcare outcomes. Otherwise, we wouldn't have 5% healthcare cost inflation for the past 10 years.

On the other side, Clover Health.

I look at this, I think they're in the same camp as we are. So underrated. I think we all get the small guys among us. The smaller ones among us have— we are all fighting the same fight. So definitely underrated.

And finally, Haven.

Remains to be seen. I don't know. That's a good question. I think, I think last year everyone was terrified of them. Yeah, I would say underrated in the fact that they can have a lot more impact in ways that maybe we don't expect it yet. But if they just really focus, if they even just focus more on putting good research out, good parts of their data out, you know, what are their— what they've got a million people, what are their negotiated rates with their own insurance companies and hospitals? Put that out there. That will be pretty powerful. I think one of the things whenever I meet somebody from there that I tell them is they could be really powerful in simply going to the IRS and CMS and saying, hey, let employers give pretax dollars to the employees to buy whatever they want in the markets, whether it's Haven, Oscar, or anything, anything like that. That I think would be the biggest unleashing of, I think, innovation in healthcare because it would let all of us, whether it's Clover, Oscar, anybody else, have this 50-year time horizon getting members on board and keeping them. And so hopefully they're going to knock on that door and lobby for that. Whatever they're also building, I think that would be great.

We're going to leave it on that note. Thank you so much, Mario.

Thank you.