How Oak Street Won By Owning the Provider Stack
And actual ways we can use language models in medicine
A few weeks ago, CVS bought Oak Street Health for $9.5B. Yup, that’s billion with a capital B. This adds to the ever growing list of health tech acquisitions by larger companies trying to get a piece of the provider side of healthcare. However, Oak Street stands out because it isn’t strictly a provider side play but an interesting combination of a unique care and revenue model combined with an integrated tech stack.
The Oak Street model isn’t revolutionary by any means. The company operates on a capitation framework with the payors that it partners with. We’ve seen this before with many other provider groups but they key ingredients Oak Street had were customized technology in the form of their own EMR, a well-structured care team where the providers were employed by Oak Street, and lots of time (11 years in the making). On top of this, they were focused mostly on the Medicare and Medicare Advantage market and could tailor their care to that specific age group.
Instead of a fee-for-service model where they would have to focus on volume of care, a capitation framework incentivized them to ensure that downstream costs of their patients (ED visits and hospitalizations) were kept to a minimum. In plain terms, this means that they have to take good care of their patients and prevent bad things from happening (which sounds like the kind of thing a doctor’s office should probably do in the first place).
Given that Oak Street did so well, it begs the question of whether owning the full provider stack is important to bend the cost curve and “make it” as a health tech company. Other companies, such as Pearl Health, are choosing to forgo owning the provider side and are instead focused on bringing analytics and data to primary care providers. The thought is that having access to these tools will allow doctors to better manage their patients and prevent those downstream costs on their own.
Pearl Health has an interesting model where they take on financial risk for patients but don’t actually employ the providers they are working with. Instead they negotiate contracts with providers which vary in terms of upside and downside risk. The big benefit of this is that Pearl can be a pure software company and not have to deal with the logistics of actually running a clinic. This removes countless roadblocks because delivering care is actually not a simple problem (and definitely a much more different one than running a software company). The downside of this, however, is that they are not in charge of a provider’s entire panel which means that the provider isn’t fully engaged with or incentivized to strictly follow Pearl’s recommendations. By not owning the provider stack, you get into a position where you have to find a middle ground with physicians to try and get to the outcomes you think they should be focusing on. It’s kind of like when McKinsey comes to your company to give you recommendations on a specific project but your head is still focused on the 20 other meetings you have to attend to that are outside the scope of what they are talking about.
Owning the provider stack solves this problem because you own the top level goals and objectives but also the players downstream who are going to make those goals happen. Combine that with focusing on a specific population (MA patients in Oak Street’s case) and you can create and tailor an experience to actually manage health for a population of patients effectively.
Of course, the reason tech companies don’t always want to go down this route is because owning the provider stack is expensive, cumbersome, and complex. Being a good software company is hard enough. Trying to be a software company and healthcare provider at the same time is no walk in the park.
Adding to the complexity is that healthcare delivery varies greatly by location - what you do in one state doesn’t always carry over to other states in terms of ancillary services, community organizations, and other points of the healthcare system your patients may interact with. Startups don’t like this because it inherently goes against their ability to quickly scale one solution across various markets.
Oak Street has shown that owning the provider stack can work in primary care if done right. We’re seeing more and more companies like this pop up, trying to get their foot in the door and get a piece of the primary care pie.
However, another application of this that we are beginning to see is owning the provider stack in specialty services. Companies like Thyme Care are using a similar model but focusing on specific chronic conditions (in Thyme’s case it’s cancer). This only works because there are specific diseases and conditions that tend to become the primary issue for specific patients and why they interact with healthcare system in the first place. Other areas where this kind of model could work are neurology and GI diseases - where patients are engaging with the healthcare system repeatedly for one specific reason (i.e. conditions like multiple sclerosis or Parkinson’s for neurology and Crohn’s or Ulcerative Colitis for GI). We’re also seeing this in serious mental illness for the Medicaid population through companies like First Hand. The economics of this will of course vary based on the population that you are focusing on but niching down and delivering quality care in a capitation model for a specific population seems like a recipe for success if a company can figure out how to effectively combine technology with healthcare delivery.
Oak Street’s acquisition has definitely proved that building a provider health tech company that can use technology to bend the cost curve is definitely possible. Given this, we’ll likely see more companies adopt similar models and improve upon with newer tech - especially in a post-pandemic era where telehealth adoption has become mainstream.
If you’re interested to learn more about Oak Street and how future companies could be even better, definitely give this article by Brandon Ballinger a read.