Oscar Health Insurance

Oscar uses the same messy, nonstandard data that all other insurance companies do. What was completely different was what they did with it.

A leap of faith was made by 12.2 MILLION Americans in 2017 when they signed up for health insurance: that Obamacare would survive the year, that the health exchanges wouldn't fail, and that the premiums wouldn't force their families into bankruptcy. The 2017 enrollment season wasn't only unclear for the 54,000 New Yorkers who utilized those exchanges to join Oscar, an insurance business cofounded by Joshua, Jared Kushner's younger brother. The company is geared towards millennials. Well, it was kind of depressing.

The corporation was cutting its largest network, which is located in New York City, in half, from 77 hospitals to 31 and from 40,000 doctors to 20,000, as Oscar's members discovered in July. Forced splits like that ought to have been very difficult to market.

However, Oscar managed to persuade nearly every one of the more than 50,000 members to remain. They presented a straightforward, if not very obvious, argument: Pick superior service and care over a larger directory of physicians. To sweeten the pot, they added extras like concierge care teams and smartphone-based direct scheduling. It was successful. Nobody questioned how they were going to deliver a paradigm-shifting network—one that was smaller but nevertheless superior—at the time. Furthermore, Oscar would not have been free to say. Big data was, of course, the solution.

Technically speaking, Oscar works in the insurance industry. However, its true focus is technology. The company's chief executive officer, Mario Schlosser, is a data scientist with training from Stanford University who founded Oscar by drawing conclusions from the deluge of current health care data, including insurance claims, physician directories, and electronic medical records. It was a move right out of the playbook of Silicon Valley: reduce options while improving user experience. (See Chapter 2010, Verse D8 in the Gospel of Jobs.) Oscar has a wide range of contacts. Furthermore, it isn't very broad. It is "optimized," as the Valley people say.

Oscar was using the same messy, nonstandard data that all other insurance companies use. However, an examination of their data science operation reveals that their use of it was entirely different.

Gently Launch

Oscar lacked the time and data to create a network from scratch when it originally began in 2014. It therefore leased a network from a third-party provider rather than circling New York haggling over rates. In just the first two years, Oscar lost over $100 million in New York alone. Contracting director Mike Kapok and his staff started meeting with clinics and hospitals to establish their own networks and do away with the thirty percent rental upcharge in an effort to reduce expenses. However, Kopko and other Oscar attendees soon discovered that saying no was their only negotiating tool. This is due to the fact that health providers—who have become more concentrated over the previous ten years—have complete control over pricing. Particularly at the beginning, insurers have to accept whatever is offered.

Aside from being Oscar.

The company's other goal was to create a smarter network by utilizing the data they had been gathering since 2014, so that they would not just contract with any pediatric endocrinologist but the right one. One who carried out the necessary procedures for Oscar's members, who made few mistakes, who made the appropriate referrals, and who had openings for new patients.

In hindsight, Schlosser says, "it seems obvious and reasonable, but it hasn't been done before." "What makes this unique is the innovative analytical framework—building a network around a patient's journey through the healthcare system."

In order to accomplish that, Oscar had to first turn its richest datasets—insurance claims—into a narrative that could describe a physician. Current Procedural Terminology, or CPT codes, are the uniform language used by claims to represent medical, surgical, and diagnostic services. Not improved treatment, but billing is their purpose. However, that was before Vinod Mitta obtained them.

Mitta was tasked with going through medical records and classifying doctors according to which 20,000 CPT codes they utilized starting last summer. According to Mitta, Oscar's vice president of clinical operations, "the goal was to better understand what type of physician everyone really is." "Looking at the services they offer is the best way to represent that." For example, orthopedic surgeons typically divide the body into several specialties. For example, there are specialists in knee reconstruction, shoulder reconstruction, and hip replacement; but, a jack-of-all-trades surgeon is not very common.

Next, Mitta presented his clusters to a group of doctors in focus groups. Subspecialty titles occasionally matched the ways in which doctors organized their practice. Not always, though. A ninety percent agreement in CPT codes indicated that ear, nose, and throat specialists and sleep physicians may be combined; however, a tonsillectomy performed to remove a malignant tumor is not the same as a snore-preventing snip. Mitta took these realizations and rearranged the information. He then returned to the physicians. Then back to the information, and once more to the physicians.

A unique taxonomy of healthcare providers was the end product: 400 buckets based on the actions of physicians rather than the titles displayed on their doors. The vice president of network construction at Oscar, Nick Rebar, explains, "We boiled the ocean of health care billing into something clinically relevant." He claims that starting with doctors rather than merely using blind data mining produced a richer model.

Working with Data

Its members were aware of what they were getting into, in a sense: fewer hospitals and physicians. However, Oscar never made it clear how exactly it had assembled that network. It was unsure if it would succeed.

However, after five months of its massive trial, Oscar has cause for optimism. With its extensive leased network, 2% of premium expenses were paid outside of the network in 2016.Oscar's out-of-network expenses have decreased to 0.6 percent this year. Additionally, its market share in New York has remained stable at about 20% of patients. It's too soon to declare it a victory, according to medical researchers: The great majority of participants in any programmed, not just Oscar's, are in good health, and frightening medical occurrences necessitating visits to experts are uncommon. So how will the network fare as time goes on and Oscar's members are dealt an uncertain deck of health cards?

The networks-building teams at Oscar don't claim to have the solution. After all, they are data scientists. Additionally, they're trying to ensure that it ages nicely. Oscar created a tool that scans thousands of fields every day to bring up to current the most recent information about the doctors in the network, such as which ones are no longer accepting new patients and which ones have had a spike in complaints. It also takes into account member needs that do not fit into Oscar's model, such as the requirement for a pelvic floor physical therapist. The team responsible for signing contracts sets to work, bringing in new providers to cover the gaps—a process that takes roughly eight weeks for an individual and closer to sixteen for a hospital. As a result, even when it is changing, the network maintains functional stability.

That should, in theory, turn off members because research indicates that seeing the doctors of their choice is what counts most to patients. However, Oscar's membership in New York has also shown that it is open to change. Before moving her firm to an Oscar employer plan this year, immigration attorney Dana Raimondo, located in Brooklyn, was covered by an Oscar individual plan for two years. "They get that you might not understand all the ins and outs of health insurance," she claims. "But Oscar was so good about explaining those things, it just feels more human and personalized."

The reason for this could be the company's ability to foster a sense of brand loyalty or the fact that Oscar's membership is somewhat younger than that of most insurance pools. Although the business declined to comment on precise figures, Backchannel said in January that the majority of its members are between the ages of 26 and 35.

For Oscar, those are both excellent things. However, this can cause issues for its rivals and their clients. According to Simon Haeder, a political scientist at the University of West Virginia, "the argument could be made that specifically advertising one of the narrowest networks out there is attracting healthier people disproportionately." Should Oscar manage to endure long enough to serve as a model for more intelligent, smaller networks, then more might follow suit. It can result in a market segmentation that could be disastrous. "In the worst scenario, that would result in the well-known death cycle wherein progressively healthier individuals leave big network insurers until the sickest individuals remain,” claims Haeder.

However, according to Oscar Schlosser, that is not at all how the optimized network would function in practice. Its precision and depth, he claims, are its virtues. "For a patient who is less well-off and requires care from physicians in a few highly specialized fields, that is more appealing and beneficial, not less."

Certain safeguards were in place to prevent this kind of thing from happening thanks to the Affordable Care Act. Providers who cover a higher percentage of low-risk individuals are required by the risk adjustment programmed to transfer money to higher-risk policies. Oscar was forced to pay up $31 million last year; the business claims it is delighted to pay this amount. However, the ACA is already being considered for repeal, so it's likely that the programmed won't last forever. The market will then determine who wins and who loses. Oscar will have outlasted the historic health care law that allowed it to enter the insurance industry if it survives that long. It will also be grateful to its data.