Google says it can save lives, but can we ever trust it with our health data?
Google hopes that one day it might be able to save your life. As the tech giant moves deeper into healthcare, it plans to fight medical misinformation in search results, create tools to be used by thousands of doctors, and improve the accuracy of diagnosis with technologies like computer vision to read X-rays.
Those are just its aims for this year. Google’s broader health mission was outlined at a conference in San Francisco earlier this month, where its top doctor set out to show why the company may be the most ambitious of the many trying to use technology to transform healthcare.
“We have 10 companies with 1 billion users and five with 5 billion or some insane numbers,” said David Feinberg. “We would love to be one of those where we can say we saved billions of years of life this year.”
The wiry grey-haired child psychologist, who is the former chief executive of the Geisinger healthcare system in Pennsylvania, was hired last year to sort through Google’s scattershot health projects to find a focus — and, eventually, a business model. One task is to combine artificial intelligence teams working on health at Google Brain and DeepMind with those working on smart devices for health at its Nest division.
But just as Dr. Feinberg started work on his first project — a search engine for electronic medical records — a scandal erupted after the Wall Street Journal reported in November that Google engineers had access to medical records held by Ascension, the country’s second-largest healthcare system, as they built new products. As a result, the U.S. government is investigating the partnership. And Dr. Feinberg must now work out how to persuade patients to trust Google with their sensitive data.
At the conference, Dr. Feinberg protested that he was “super, super proud” of Google’s work with Ascension “despite what they say in the newspaper.” Google has said the data were protected and the partnership was legal and common — hospitals regularly work with third-party technology companies. Ascension said the “groundbreaking work” will help improve care and that Google is not allowed to use the data for marketing or research purposes.
They will not necessarily discover the new penicillin, but they could make hundreds of millions of dollars a year and scale fast
Yonatan Adiri of Healthy.io
But if Google is going to come anywhere near reaching its ambitious goals in healthcare, it needs to convince health systems like Ascension to give up their patient data and ease patients’ concerns that it might be used for less than clinical purposes.
Robert Wachter, chair of the department of medicine at the University of California, San Francisco, and author of Digital Doctor, a book about how technology is transforming medicine, says Google’s task has become harder as potential partners worry about the public reaction. “Inevitably, anyone who is thinking about doing a partnership with Google, if they weren’t thinking about the optics before this, then they were probably asleep. If they aren’t now, they are definitely asleep,” he says.
Google is well aware of the challenge it faces. “It would be a shame on us if we can’t get this technology into the real world because we don’t act in a humble way, we don’t act in a thoughtful way,” said Dr. Feinberg.
After transforming almost every other industry, the world’s largest tech companies now want to fix stretched healthcare systems, an industry worth US$8.7 trillion worldwide. As they hunt for new growth engines, Google, Amazon and Apple are convinced they can make money by cutting costs in the sector. With vast financial resources — last week the stock market value of Google’s parent company Alphabet rose above US$1 trillion for the first time — computing power and banks of AI experts, no one should underestimate their abilities.
Amazon has bought an online pharmacy and entered into a joint venture, Haven, with JPMorgan and Berkshire Hathaway, which aims to create a new kind of health insurance for the companies’ 1m employees. Apple is using its Watch to improve fitness and cardiovascular health. Microsoft is selling cloud services to hospitals and drug companies, while Facebook is offering tools for users to manage screenings.
Under Dr. Feinberg, Google is focusing on its major mission — organizing the world’s information — and its capabilities in data, AI and sensors.
Dr. Feinberg was hired to bridge the gap between healthcare and technology. Far from a traditional Googler, he was nevertheless known as an innovator. As Geisinger chief executive, he pushed for patients to have their genome sequenced and launched healthy food pharmacies to serve “prescriptions” to patients such as diabetics. He has bolstered Google Health with other industry leaders including Karen DeSalvo, a former national co-ordinator of health information technology, and Robert Califf, former commissioner of the Food and Drug Administration.
Although the company has not been so specific about its eventual plans, its vision is to build a Google Assistant for your health, according to Eric Topol, author of Deep Medicine, a book about AI in healthcare. It could combine electronic health records with data from smartphones, genomes, information from sensors such as glucose monitors and even apps that record eating habits, he says.
Seen through that lens, he says Google’s recent US$2.1 billion offer for Fitbit, the activity tracking wearable technology company, is about acquiring “one of the most exquisite data sources for a couple of billion.” He adds: “That’s nothing for Google, a company that seems to print money.”
Google’s Cloud business has partnerships with many providers beyond Ascension, including the Cleveland Clinic in Ohio and the Mayo Clinic in Minnesota, where it is opening an office specifically to serve that health system. These agreements could give it access to certain medical data.
It is also interested in sucking up more consumer health data, beyond what it knows from search and location history. As well as Fitbit, deals done by its Nest division suggest it sees the smart home as an important door into digital health, helping monitor patients outside of the hospital. In 2017, it bought Senosis Health, which builds smartphone apps that track information such as lung function by using the microphone as a flowmeter, and Knit Health, which makes a smart baby monitor that could also be used to monitor adults’ sleeping patterns.
Andrew Matzkin, a partner at Health Advances, a consulting firm focused on health tech, says both Google and Apple believe they can use new devices to get even closer to you than your smartphone — and detect conditions such as depression or heart rate variability.
There has been a lot of focus on wearables, he says. “But I think just as important and maybe even more important is ambient sensors: bedside devices, under-mattress sensors, sensors integrated into toilet seats.” Google has patented various of what it calls “non-invasive health-monitoring devices”, including a toilet seat sensor to measure heart rate and blood pressure, which some might consider invasive.
It is not yet clear how Dr. Feinberg’s work will fit with health projects in other areas of Alphabet, Google’s parent company, which includes Verily, a life sciences division, and Calico, a unit researching ways to slow ageing.
Verily’s work is split between ambitious moonshots — such as robotic surgery or implantable devices — and software, some of which is already on the market. Verily products like Onduo, a joint venture with Sanofi to create a diabetes management platform, might move closer to Google Health.
Yonatan Adiri, an Israeli digital health entrepreneur who leads Healthy.io and worked with Geisinger, says Dr. Feinberg’s appointment shows Google coming back down to earth. He says at first Google tried to be “ultra sophisticated”, launching an effort in the U.S. to transform electronic medical records in 2008, and then it experimented with “very bold” and “very Googly” projects like Calico and Verily.
Dr. Feinberg is at the heart of Google’s next act. “They will not necessarily discover the new penicillin, but they could make hundreds of millions of dollars a year and scale fast,” he says. “Feinberg is the right man for that job.”
Google’s biggest advantage in healthcare — its capacity to combine data sources and to learn from them — is also what worries its critics the most.
Christine Lemke, co-founder and president of Evidation, a health measurement platform that provides data analytics to tech and pharma companies, says one of Google’s key advantages is it can use unconventional sources of data to understand health.
“Because of search, they’re already the largest provider of healthcare information and they can leverage that position to incorporate other non-traditional data to drive even more personalized care,” she says. “But the challenge, as we saw with the Ascension story, is one of trust.”
Dr. Feinberg recognizes that people are wary of Google potentially combining consumer data with medical data — and says it does not. Yet in San Francisco he suggested that mixing different types of data can be medically advantageous. Talking about Google’s work to improve the accuracy of mammograms by using medical records and genetic information, he said: “There is incredible power in the ability not only with the data, but combining the different data sets.”
Already, hospital systems may be wondering whether working with Google is worth the potential negative publicity
Google escaped the severe backlash over privacy that hit Facebook. But its brand is still associated with building deep and detailed profiles of users. Already, hospital systems may be wondering whether working with Google is worth the potential negative publicity.
Bill Evans, managing director at Rock Health, a venture capital firm focused on digital health, says he is concerned about the “spillover” to the broader industry. “What we worry about is the potential chilling effect that perceptions of a major brand has on emerging and highly ethical innovators,” he says.
Google had mentioned the deal with Ascension on an earnings call — but many patients whose data were being collected only found out about it from the media.
The scandal echoes the controversy when DeepMind, a London-based unit of Google, partnered with a National Health Service trust. The Royal Free trust was reprimanded by the UK information commissioner in 2017 for allowing them access to information on 1.6 million patients without explicit permission. DeepMind last year transferred control of its health division to Google.
“In some ways, after watching DeepMind in the NHS . . . it is completely predictable that a large data-sharing arrangement with a healthcare organization was going to receive a lot of public scrutiny,” says Dr. Wachter.
This dual role of holding consumer and health data is already the subject of a lawsuit against Google and the University of Chicago. The suit claims that Google would be able to re-identify supposedly anonymous healthcare data that it accessed through a research partnership. The plaintiff is concerned that Google could connect the times of appointments from the medical records with the data it holds on people’s smartphone locations.
Google denies it has made any connection like this. A spokesperson said: “We believe our healthcare research could help save lives in the future, which is why we take privacy seriously and follow all relevant rules and regulations in our handling of health data.” Sam Smith, a privacy activist at Med Confidential, says the question is whether consumers are happy with a big company like Google, Facebook or Amazon doing everything. “The answer seems to be no,” he says. “If Google wants to make money by knowing everything about you, hosting health data is probably not worth it.”
But some in the healthcare sector disagree. Unity Stoakes, president of StartUp Health, a digital health accelerator which hosted the event where Dr. Feinberg spoke, says Google is integrated “into the fabric of our lives” with search and smartphones. “People trust these companies,” he says.
Google may actually be seen as the best of a bad bunch: in a Rock Health survey earlier this year, only 11 per cent of consumers said they were willing to share health data with tech companies. But among that 11 per cent, Google was the company they trusted the most.
Even if Google wins people’s trust — or at least their complacency — it will face other struggles in healthcare
Even if Google wins people’s trust — or at least their complacency — it will face other struggles in healthcare. Google has failed before: the 2008 attempt to transform electronic medical records stumbled because so much health information was kept on paper. It will shut down projects that do not work. Earlier this year, it paused its work on Google Fiber, the superfast broadband project.
Dr. Wachter was on Google’s advisory board in the mid-2000s. He says that back then, it failed to recognize the complexity of healthcare compared with other sectors. Now, he says they are taking a “much more mature approach.”
“Imagine a world where Google does advanced navigation tools for patients, searching for migraine [symptoms], or the best doctor, the best treatment,” he says. “Google has a major pre-existing advantage.”
However, Google is at its heart a consumer company that has struggled when selling to businesses, falling behind Amazon and Microsoft in cloud computing. Even before the Ascension deal, some potential customers were nervous about partnering with Google, because they feared their data were the product. Adiri says smaller partners such as laboratories or imaging companies were concerned that the cloud services Google was trying to sell them looked too good to be true.
“They are nervous that their data is the big value of their business and they would be giving it away,” he says.
For Matzkin, it is not surprising that Google has not figured out how to make money in healthcare because there is no “simple cookie cutter answer.”
But he adds: “One of the biggest advantages companies like Google or Apple have is their ability to take a very long perspective. They have huge ambitions but are willing to view their efforts in this space over a 10-year horizon — and possibly longer.”