Google's Getting Into Health Insurance? Verily!

Good Morning!

Some quick items from the news before the main event:

  • Health insurance giant Cigna has agreed to buy Express Scripts, one of the last remaining standalone pharmacy benefit management companies. While the deal is drawing comparisons to the CVS/Aetna merger, I don’t think the similarities extend very far. CVS/Aetna has much loftier goals for creating a new kind of provider network, while Cigna/Express Scripts seems to be a much more vanilla bid for market leverage.

  • Amazon, JP Morgan, and Berkshire Hathaway have begun the search for a CEO for their healthcare joint-venture. The revelation that Andy Slavitt and Todd Park have been approached makes me think this new company will be leaning further towards creating revolution in healthcare than simply reducing costs for the three companies. The next step is for these three to choose a name for their new venture so healthcare writers can stop having to write the names of all three companies every time they reference it.

Before we begin, a quick reminder that sharing this newsletter is not only possible, but entirely recommended. Studies have shown that sharing this newsletter can improve general well-being and even give you a warm feeling inside.

Enjoy!


Christina Farr of CNBC reported this week that Verily, the life sciences Google spinoff and Alphabet subsidiary, was getting serious about entering the health insurance business.

So, what does Verily actually do? How is Google’s secretive spinoff planning to enter the health insurance market and shake things up?

Starting at the beginning…

What is Verily’s business model?

In this case, the apple hasn’t fallen far from the tree. Verily is in the business of information. While they bill themselves as an R&D company or a life sciences firm, they’ve been built to see the world through the lens of data. This makes sense when you were born at Google, where the stated mission is to organize all the world’s information.

Eric Schmidt, former executive chairman of Alphabet, stated in his keynote address this week at HIMSS (the health IT conference) that “healthcare is becoming an information science.” You can agree with that or not, but it makes it clear how Alphabet, and Verily, are approaching the problems of healthcare.

Their most trumpeted initiative, Project Baseline, aims to gather longitudinal health data from 10,000 volunteers over the course of four years. Volunteers are asked to do a battery of annual tests, have their DNA sequenced, give various liquid samples, and even wear a special-for-this-study data gathering wristwatch at all times. This will produce an unparalleled data set of clean, verified clinical-grade data that can be studied, cross-referenced, and scrutinized like never before.

The company pitches the project as purely investigational. We know what sick bodies look like as we measure them in hospitals, but what sort of signals does a healthy body emit? What does a body look like before it becomes ill? What are the subtle signals of impeding disease that we aren’t seeing with our currently incomplete data? Verily intends to find out.

But Verily is a business first. While they keep their activities quiet, it’s been reported that they’ve already begun shopping the Project Baseline data to pharmaceutical companies. They’ll also most certainly be using this data internally to inform other activities, perhaps using it to build predictive machine learning models for managing the health of a population.

In the age of machine learning and artificial intelligence, data is more valuable than ever before. A major problem in healthcare is a lack of reliable, clean, and verified data. Verily, with their strong corporate family backing and deep pockets, is able to go to great lengths to break through the lack of solid information by gathering their own. The result will be an extremely powerful data set that Verily owns outright and can use to their own benefit, or sell access to.

From a P&L perspective, Project Baseline is essentially a huge fixed cost. They’ll spend a fortune (I’ve seen one analyst guess $100M) to gather this data. But the beauty of the data business is that the data can be sold, and re-sold, and analyzed, and sold again. The upfront capital investment is steep, but the variable cost of selling the product is near zero.

But Verily makes devices too. One of their first deals was to help manufacture some of the hardware components of a forthcoming glucose monitor from Dexcom. The deal was a classic R&D structure, with an upfront payment and a percentage of sales as an ongoing royalty to Verily. But I wouldn’t be at all surprised if the contract also agreed to Dexcom relaying anonymized health data from their glucose monitors back to Verily. As they showed by creating a special data gathering wristwatch for the baseline study, Verily will happily use their hardware engineering expertise to build devices that bring them more data.

Now They Want To Be In The Insurance Business

For every company in the business of producing insights, the ultimate temptation lies in using those insights to build their own business rather than selling them to the highest bidder. Verily is no different.

The rumors circulating have Verily working with established health insurers to bid on population health management contracts. In taking these contracts, insurers bet on themselves to be able to improve the overall health of a population. If they succeed in doing so, they get a share of the overall health expenditure savings they helped produce. Verily, with their growing trove of information, is betting they can directly apply what they’re learning in Project Baseline, through a collaboration with 3M that began in 2016, and their other projects, to actually improve health in a living population.

It’s important to note that the rumor is that Verily is partnering with an existing insurer, not taking on the risk themselves. This sort of arrangement is in line with their previous partnership-focused revenue model, and also allows them to avoid the regulatory headaches of operating an insurance company. They’ll be free to focus entirely on finding innovative ways to manage populations to better health.

Is this hubris, or can Verily actually have an effect?

But will it work?

IBM, the first tech company to make a play in healthcare, promised big things from their favorite artificial child Watson. And they weren’t without their successes.

In a blog post, IBM Watson’s William Kassler describes a study performed way back in 2013 where automation and data science was applied to a population of diabetic patients. The result was an overall reduction in blood glucose levels. This is exactly the kind of target Verily will be hunting for when they start managing their own population, as a reduction in A1C levels means a reduction in complications like eye and kidney problems, and thus a reduction in the overall expenditure for the population.

So how did they do it? This is the interesting part.

Their efforts can be summed up as improved patient scheduling. By applying advanced software to the problem of ensuring that patients weren’t falling through the cracks, the Northeast Georgia Physicians Group was able to see when patients hadn’t seen a doctor in a while. Combining this data with flags in patient records indicating which ones were less likely to be capable of managing their condition on their own, the software would then automatically remind certain patients to visit the doctor.

Without entering the exam room, without asking a physician to change a single thing about the way she practices medicine, and without changing the practice workflow at all, this group was able to lower overall A1C levels in their diabetic population by 1.3 percentage points on average.

When we talk about the promise of Artificial Intelligence in healthcare, we think of clinical decision support systems or machine learning models that can detect previously undetectable abnormalities in EKG data. But as we saw in Georgia, there are fruit hanging much lower than that.

With that in mind, it’s hard to imagine Verily will have trouble finding a way to be an effective population health manager. With their new and unparalleled data set, and the might of Google’s machine learning operation at their backs, Verily could very well have a strong effect if they focus on the right targets.