Alphabet’s Google division is, fundamentally, in the business of selling data. That is a useful thing to keep in mind when Alphabet’s Verily comes calling for your medical data. But Google is also inarguably useful; this is why, despite knowing that my every move is being tracked by the company, I still make use of Google search, Gmail, and Google Docs, among its other myriad services. Verily’s Project Baseline is, in some sense, the health equivalent of those kinds of services — it has the potential to greatly expand our knowledge about what human health looks like. Not incidentally, the project will be of service to Verily as well.
Researchers will collect genetic data, blood samples, medical images, and other information.
In 2014, Verily — then a division of Google X — announced the Baseline Project, a collaboration with Duke University and Stanford University to try to get a sense of what a “normal” human looks like. Today, the group announced it will begin enrolling 10,000 healthy people, following a pilot in about 200 people that began in 2014. Over the course of four years, researchers will collect genetic data, blood samples, medical images, and other information from the study participants.
That “other information” might include environmental data, as well as responses to phone surveys, and data from sensors in the Study Watch, a sensor-packed smartwatch announced last week. The studies are starting in the San Francisco Bay Area and North Carolina, though the scientists behind the effort hope to expand the areas surveyed. And because the program is meant to be nationally-representative, recruitment may be a little slow. When it’s over, there will be a database of anonymized data that plenty of researchers — including those from the pharmaceutical industry — will have access to.
This style of study isn’t unprecedented; in fact, it’s been a feature of medical discovery for quite some time. The most famous example is the Framingham Heart Study, which began in 1948 with about 5,000 patients. At the time, doctors didn’t know much about heart attack and stroke, except that they were common and often deadly. So the Framingham study was devised in order to follow people ages 30 to 62 from the town of Framingham, Massachusetts for years and see if there were clues to those ailments. In 1948, and every two years afterwards, the study participants checked in. A second generation was added in 1971, and a third in 2002.
Framingham has provided clues to most major cardiovascular risk factors
Over the course of decades, Framingham has provided clues to most major cardiovascular risk factors: high blood pressure, high cholesterol, smoking, obesity, diabetes, and a sedentary lifestyle, among others. Framingham alone wasn’t enough to identify all these contributors, of course — but it told other scientists where to look. Something like 1,200 articles have been published in academic health journals over the last 50 years on Framingham alone.
Project Baseline is twice the size of the original Framingham study population and is attempting more comprehensive measurement. And unlike Framingham, which was funded primarily by the National Heart, Lung and Blood Institute, this study is funded by Verily. Government spending on science has stagnated over the last decade, and Framingham has been among its casualties; the study lost 40 percent of its funding in 2013 as a result of the budget sequester. In fact, it would be a lot harder to get a publicly-funded study like Framingham off the ground today — both because it’s expensive and because it’s hard to predict what studies like this will find. Verily’s Project Baseline, then, is a mightily ambitious piece of basic science, and one that could prove useful.
Advisory board member Adrian Hernandez, a cardiologist and professor of medicine at Duke, says the Project Baseline group is “aiming to build an early discovery platform.” It’s possible subtle changes occur in some areas — biomarkers, behavior, anything really — before a disease takes hold, Hernandez notes. Discovering what those changes are may lead to earlier and better treatment. Beyond those broad brushstrokes, it’s a bit difficult to say what a study like this is for until well after the fact, points out Stanford’s Sam Gambhir, who also sits on the advisory board. With a cohort study like this one — or like Framingham — it’s impossible to know the medical impact until well afterward.
It is, however, possible to take an educated guess at what Verily gains by running the study. The point of a publicly-traded company is to make money for its shareholders, and Verily is owned by Alphabet, which is a publicly-traded company. So Verily stands to make money — the question is how.
Verily, according to Alphabet’s investor documents, sells R&D services and licenses. This is worth keeping in mind; the data generated by Project Baseline will be shared with “qualified researchers.” Duke and Stanford are the two obvious places where the data will be shared, and the researchers there are likely to use Project Baseline data just as they’d use other data. When I asked about the possibility for pharmaceutical companies to access this data, Jessica Mega, the chief research officer at Verily, got squirrely. Here’s her initial reply: “There's a scientific executive committee that will review every request and the composition includes individuals with a leadership from Duke and Stanford, so it would need to be in line with the overall mission of the study.”
Many companies sell their large databases of customer information
I asked again, if someone who worked at a company like Pfizer would be able to gain access to the data, and Mega said, “Yes, as long as the intent is try to improve medical discovery.” Verily doesn’t yet figured out what “access” to the data will look like, I was told, but “the philosophy of the study is to make this information broadly available to qualified researchers.” Verily declined to comment on the possibility of fees for accessing its data.
Many companies sell their large databases of customers’ information for discovery research. 23andMe, for instance, sells de-identified data from its genetic tests to researchers (such as Stanford), drug companies (Genentech, a subsidiary of Roche), and other entities (including the Michael J Fox Foundation, which does Parkinson’s research). In fact, the data is 23andMe’s moneymaker, not the tests it sells. Ancestry LLC, which also sells genetic test kits, does the same. It stands to reason Project Baseline could create similar revenue opportunities for Verily.
There is also the matter of the Study Watch, which provides Verily with a number of opportunities. The most obvious is a consumer version of the Study Watch, though Mega says the company doesn’t currently plan to make the watch commercially available. The watch could also be used as a platform other researchers license in order to gather fairly continuous data. As it happens, the Study Watch is being used in the Personalized Parkinson’s Project — another Verily study, taking place in the Netherlands. That trial and Project Baseline may serve as trial runs for future research uses.
Those two studies may also provide some sense of how reliable the data from the watch is, as well. (The Study Watch wasn’t used in the Project Baseline initial pilot run of about 200 patients.) Most wearables don’t require FDA approval, as long as they are marketed as “wellness devices.” But any kind of specific medical claim would probably require the regulators’ sign-off. Verily declined to comment on any regulatory plans for the watch.
At this stage, it’s too early to speculate about what public good might come from Project Baseline, though it does seem likely there will be advantages to the work. It’s even possible that Project Baseline will be as useful to research as Framingham was 50 years ago. It may provide — as Framingham did before it — clues to the roots of common ailments. There may even be new drugs developed out of it. One thing is nearly certain, though: Project Baseline is meant to benefit Verily, too.