“Silicon Valley is built on impatience and doing things rapidly,” says Anne Wojcicki, the CEO of direct-to-consumer genetic testing company 23andMe. “But what I had to learn is that not everything comes overnight. Sometimes we just have to wait and, if you focus on the positives, that can be just as rewarding.”
Wojcicki should know. Today, two years after the Food and Drug Administration ordered 23andMe (in a highly publicized letter) to stop selling its genetic health tests, the company is back in the genetic health analysis market with an FDA-approved test for 36 autosomal recessive disorders–a class of illnesses that can be passed down to children when both their parents are carriers of the same genes. Diseases that the company can now test for include sickle cell anemia and cystic fibrosis. And last week, the company announced that it had raised $115 million in Series E funding, setting its valuation at $1.1 billion.
It took a lot of hard work–and some internal growth–for the company to stage this comeback. On the regulatory front, 23andMe had to learn how to prove to the FDA that its testing processes and data were accurate; they accomplished this in part by hiring an experienced regulation chief to guide the process for the company. The company is also introducing a new web design today that attempts to simplify complex information about genomics for the average consumer—part of 23andMe’s efforts to successfully satisfy the FDA’s requirements that consumers are equipped to accurately comprehend the meaning of their test results.
And the FDA has been helping things along, too, deregulating the entire category of autosomal recessive disorder carrier screening earlier this year—which means that 23andMe didn’t have to apply separately for approval to test for all the genetic disorders included in the category, but instead has to be prepared for FDA regulatory audits after the tests go to market.
But this doesn’t mean the road ahead is all clear for 23andMe. Today’s announcement that some 23andMe tests are back on the market comes with a higher price tag for new customers of $199 per test, up from $99 before the FDA issued its infamous 2013 letter to the company. 23andMe still can’t market all of the genetic tests it used to sell–the company’s genetic screenings for many diseases, including breast cancer and Alzheimer’s, have not been cleared by the FDA, though 23andMe is working with the regulator on case-by-case authorization for more categories beyond autosomal recessive disorders. And 23andMe probably can’t afford to make more mistakes (like mishandling its relationship with the FDA or being unable to prove its studies follow the medical industry’s accepted clinical study procedures) in the future–thanks to a recent controversial report in the Wall Street Journal about the accuracy of direct-to-consumer blood testing startup Theranos‘s data, health startups are under a great deal of scrutiny not only by regulators, but by the scientific community and the public at large.
Despite these challenges, today’s news is a big step for 23andMe–and for the direct-to-consumer health care industry as a whole.
I’m sitting in front of a cheerful, candy-colored box that has just arrived in the mail from 23andMe. The instructions inside are straightforward: I need to spit into the test tube provided and send it back to a lab in California for my genome to be sequenced. But as simple as the task is, the enormity of what I’m doing is just dawning on me.
Take, for instance, the fact that I’m about to contribute to one of the largest medical experiments ever conducted, the outcome of which could transform health care. 23andMe gathers millions of DNA samples in its database from customers–80% of whom, like me, share their data for research. The company’s scientists are beginning to deduce some patterns from this information: The company says it has already identified a few insights about Parkinson’s, asthma, and psychiatric disorders, providing some new knowledge to researchers searching for cures.
But 23andMe isn’t just in the business of collecting big data; it is also selling consumers deeply intimate information about themselves and their families. When my report comes back in a month, I will open a Pandora’s box: I will learn about my ancestry, the possible deadly diseases that haven’t yet manifested in my body, and the conditions I might pass on to my unborn children. Sitting alone in my room, untrained in medicine or genetics, will I know what to do if the news is devastating? Am I equipped to handle the anxiety the results might cause me? These thoughts swirl around in my head for several minutes. Then I spit in the tube, reseal the box, and put it in the mail.
For Wojcicki, the collision of big data and intimate data is among the most fascinating aspects of genomics. “This, to me, was always the beauty of the genome,” Wojcicki says. “We’re solving a mystery for humanity, but that mystery actually trickles down to you. The only way for us to decipher the genome is to have massive amounts of data about it, but what this means is that over the next 50 years, I could tell you a new secret every week about what is contained inside your personal DNA.”
However, the long game was always to aggregate this personal information to create an enormous data set that could radically improve health care. And Wojcicki, who had been a health care investor for several years, believed that the health care system was in dire need of change; she felt too many people were taking medications that did not work for them (the original tests 23andMe marketed told users about their tolerance to different drugs) or dying prematurely because they didn’t know they had genetic predispositions to certain diseases. “I thought it was really a tragic system,” she recalls. “How can you learn about a disease when you do a study of [only] a thousand people? If one of the bottlenecks in health care is that you don’t have enough data, then why couldn’t we just use the Internet to start collecting it?”
In 2006, only three years after the human genome had been sequenced, Wojcicki cofounded 23andMe with biologist Linda Avey. The company allowed people to access, and potentially understand and benefit from, the information contained in their genes. The founders wanted genetic testing to be available to consumers without the intervention of doctors or the medical system so that individuals could have access to more data about their own health risks, and so that 23andMe could gather data en masse that might help move medical science forward. It was—and, in many ways, still is—a radical idea.
By 2008, 23andMe’s DNA test kit was hailed TIME Magazine’s Invention of the Year. The test could identify 600,000 genetic markers; estimate a user’s predisposition for more than 90 conditions, from cystic fibrosis to heart disease; evaluate how responsive a user might be to particular drugs; and even tell customers how much of their genetic material was Neanderthal. As the cost of genotyping went down–and as new rounds of funding poured into the company–23andMe was able to offer the tests at increasingly low prices to help scale their customer base and gather more anonymized research data for itself and its third-party partners. Prices for test kits and results fell from $999 at launch to $399 in 2008 to $99 in 2012. Slowly, hundreds, then thousands, of customers signed up to learn about themselves.
But not everybody was on board with 23andMe’s mission. Naysayers worried that a little knowledge about one’s genetic makeup can be a dangerous thing. Doctors did not appreciate patients coming in with genetic reports they did not understand, demanding procedures they did not need. Many pointed out that genetic factors often play only a small part in the manifestation of disease. “Throwing testing out there without appropriate genetic counseling is a kind of genetic astrology,” Arthur Caplan, a professor of biomedical ethics at NYU Medical School, says. “These companies are making it seem that your major risk is your genome,” he says, pointing out the significant role that environmental and wellness factors play in preventing disease, even many hereditary ones.
Policymakers were concerned that employers and insurers might discriminate on the basis of genetic information. The Genetic Information Nondiscrimination Act of 2008 prohibits health insurers from denying coverage or charging higher premiums based on genetic information, but what about cases of life insurance, disability insurance, and long-term care insurance? While 23andMe assures customers that all data used for research or shared with third parties is anonymized, in the case of a privacy breach, an individual’s personal account data could theoretically be used against them. Moreover, the company has made it clear that it reserves the right to hand over a customer’s DNA to law enforcement under court order. “The idea that databases are private is from the 20th century,” Caplan believes. “There are going to be accidents and hacks: The question is not if, but when.”
The FDA, for its part, kept knocking on 23andMe’s door, questioning the accuracy of the tests and whether the reports adequately educated the consumer. One problem, company executives assert, was that 23andMe was not responsibly equipped to respond to the FDA’s concerns. Astonishingly, the company says there were no legal or regulatory experts on staff who could competently parse dense FDA memos or understand exactly what the company would need to do to be compliant. “It wasn’t that we didn’t care about the FDA,” Andy Page, 23andMe’s president, insists. “We really lacked the skill set to interpret communications that were coming our way. Our regulatory process was way too slow and not as responsive as it should have been to the FDA’s feedback.”
Writing in Forbes, Matthew Herper bluntly described 23andMe’s approach as “the single dumbest regulatory strategy I have seen in 13 years of covering the Food and Drug Administration.” Indeed, a glimpse into any thriving health care company should reveal a robust infrastructure to help manage its relationship with the FDA. Health care “companies that have been successful in the past built regulatory affairs departments; those that didn’t failed,” Daniel Carpenter, a professor of government at Harvard who is an expert on the FDA, tells Fast Company. “It wasn’t just about proving scientific, chemical or biomolecular claims, it was also interfacing with the legal structures that our country has for the development of these products.”
23andMe, however, was staffed with tech executives who had no experience with the FDA. “We didn’t understand the expectations of the FDA,” Page says. “In hindsight, I wish we had.” The company’s experience serves as a cautionary tale to the mushrooming tech startups trying to innovate in health care. It is now abundantly clear to companies competing in this space that it is crucial to develop strong legal and regulatory infrastructures early on.
Whether 23andMe was really incompetent when it came to regulation, or suffering from Silicon Valley hubris, the fact is that in May 2013, it stopped responding to the FDA at all. So it shouldn’t have come as a surprise that by November, the company received a harshly worded letter from the FDA ordering it to stop marketing genetic health tests because the company had failed to document the accuracy of its test results to the FDA’s satisfaction, and because the regulatory body was concerned consumers might misunderstand information in their 23andMe test result reports. But people inside the company insist they were, indeed, surprised by the decision. In a split second, 23andMe could only market its kits to tell users about their ancestry.
Jennifer Dooren, a spokesperson for the FDA, says the drastic measure was necessary because 23andMe had not provided the agency with adequate evidence to support the validity of its tests. “The company was making unsubstantiated medical claims,” Dooren explains in an email. Without the right kind of data and documentation, it was impossible for the FDA to tell if 23andMe’s science was reliable. For instance, she points out that there are thousands of possible mutations across the BRCA genes that elevate a person’s risk of breast cancer, yet it seemed that 23andMe was only analyzing a small proportion of those mutations to assess a customer’s risk. In another example, she explained that it was unclear how 23andMe was identifying a person’s response to drugs like the blood thinners warfarin and clopidogrel.
While 23andMe was presenting its kit as a source of interesting information–much like a Fitbit offers useful nuggets of data about your sleep cycles–the FDA classified 23andMe’s genetic health testing kit as a medical device because it was intended to diagnose, prevent, or treat disease. “Results of such tests may lead consumers to make major medical decisions that may be irreversible without input from a qualified health care professional,” Dooren says. As our editor-in-chief Bob Safian once wrote, “The FDA is tasked with protecting consumers, even from themselves.”
But it was not all doom and gloom: the FDA was also laying out a pathway for 23andMe to become compliant. The FDA has very stringent requirements for collecting data and presenting documents to prove that a test is valid. According to Dooren, the agency had provided 23andMe with feedback about how to conduct studies to prove a test’s accuracy and discussed ways of managing some of the risk that a consumer might misinterpret or misuse the information in their reports. But 23andMe stopped responding to the FDA’s attempts to resolve these issues and even launched a national television campaign marketing the unapproved tests. The next logical step for the federal agency was to order that these tests be taken off the market.
In those dark days, Wojcicki and Page searched for advisers who could help them determine what to do next. They came across Kathy Hibbs, a lawyer who had spent her career in the biomedical industry. Hibbs happened to be in Washington, D.C., when the FDA letter was released to the public, talking to policymakers and health care executives in her role as general counsel to Genomic Health, a company that diagnoses early stage cancer. “It was a warning letter that everybody read,” she says. “I don’t know anyone in the health care space who wasn’t aware of it.”
When 23andMe’s leadership reached out to her, she was more than willing to talk the company through the unfolding crisis. “As somebody who wasn’t directly working at the company, what I see is that there were communications going back and forth between 23andMe and the FDA, but they they weren’t connecting on what the words actually meant,” Hibbs says. “Internally, people didn’t understand what they were expected to do. It was a small company at the time and complying with the FDA was an overwhelming data challenge.”
Page and Wojcicki wanted to hire Hibbs on the spot to build a regulatory team at the company. Hibbs wasn’t looking for a career change, but it soon became obvious to her that shepherding 23andMe through this tumultuous period wasn’t just about getting the company’s test kits back on the market; it was about setting precedents for a a burgeoning industry. “The regulation of 23andMe is one of the most fascinating challenges that could exist,” Hibbs says. “It was clear to me that 23andMe was going to end up helping to set policies that would be far-reaching in terms of consumer tech and health care.”
Six months after the warning letter, Hibbs became the chief legal and regulatory officer at 23andMe. As someone who has spent her entire career working within the FDA’s regulatory framework, Hibbs has been able to help the company compile data and present it in an acceptable format to satisfy the FDA that 23andMe tests are reliable. Her expertise has been crucial to helping the company rebuild its relationship with the regulator.
Despite all the turbulence in the months after the FDA issued its letter, the 23andMe customer base was still growing, even though users could only purchase the ancestry kits. At the time of the warning letter, 23andMe had half a million customers; a year and a half later, it announced its millionth customer. Eighty percent of these customers consented to share their DNA, which meant 23andMe could continue to work on its research, even if it couldn’t market genomic health tests and reports. “This trend that we anticipated from the beginning was suddenly happening,” Page says. “People really wanted access to their genetic information. That provided fuel for us to keep on going, despite not having a clear sense of when we would be back on the market.”
Page and Wojcicki began to see an opportunity not simply to get back in the FDA’s good graces and return their health tests to market, but to learn from their mistakes and use those insights to build a better health care company (and make more money–the company is not yet profitable). “When we received the warning letter, it was definitely an awakening that people are watching,” Page says. “We need to make sure that we are overprepared for what might come.” The company hired new talent to create and run new branches of the company and upgrade existing services while also helping to satisfy FDA requirements: designers who could provide a more comprehensive and clear genome education experience for the consumer, regulatory masterminds to get the company’s documentation in order, business development managers who could strategically sell genetic data to third parties, and therapeutics experts who could develop drugs.
These new executives were tasked with furthering the mission of the company, but many were also supposed to help create new streams of income. For instance, Emily Conley, a former Stanford neuroscientist who now runs 23andMe’s business development, is responsible for sharing data with industry partners like Pfizer and Genentech. The recently launched therapeutics division of the company will provide new opportunities for her team to collaborate with companies in advanced stages of drug development. “The bet that 23andMe is placing is that when you incorporate human genetic data into the drug development and discovery process, your rates of success go up,” Conley says. “We are hoping to move from a model of nine out of 10 drugs failing when you put them in humans to a better ratio.”
23andMe is also confronting a design challenge, since the FDA has made it clear that if the company is to proceed with selling health tests, it must ensure that consumers are fully aware of what they are getting themselves into and that they understand what their test results mean. Since 23andMe markets its tests direct to consumers, there’s no doctor or genetics counselor in the room to explain results to users when they read their test reports–most customers rely primarily on the 23andMe website to help them interpret their results. “One of the positives of having to be off market for two years is that it gave us time to rebuild the entire site,” Wojcicki says. “We asked ourselves what we had learned over the last seven years and what it was that we had always wanted to improve on the experience–and just do it.”
The company brought in Scott Andress from Yahoo to be the director of product and Brad Kittredge from Jawbone to be the VP of product. “There’s no corollary for our product on the market,” Andress says. “Just about every consumer has not had a genetic test before. We have to acclimate users with how this whole system works from the very earliest touch points we have with them. We set out to take something complex and make it simple without losing any accuracy.”
The design team attempted to do this through a combination of words, videos, white space, and repetition. They worked with doctors, geneticists, genetic counselors, designers, and regulatory experts to determine how to most effectively convey information to the user. Ultimately, they created an online report that first allows the user to see their results without any distractions, then provides educational materials users can explore at their own pace; the hope is that this approach will help ensure that users are not overwhelmed with too much information at once. At the very least, the new site is an improvement over the old site simply because it’s less cluttered and easier to look at; users will likely feel less overwhelmed with the amount of information presented to them on a single page.
There are times when the reports may convey devastating information about someone’s health. In the future, for instance, the FDA might approve 23andMe’s BRCA screening that determines a person’s risk of getting breast cancer. In these cases, Kittredge says the goal is to be direct with the consumer, but also to express empathy. Most importantly, the site should make it very clear what the appropriate follow up steps should be. “It’s important to remember that if someone got that result, it’s because they sought this test out and wanted this information,” he says. “Our job is to help them understand that information and take the right action on it.”
In order to meet FDA requirements, the design team had to prove that the reports provided on the website would be comprehensible to any American consumer, regardless of their background or education level. “The FDA did not prescribe that process,” Kittredge explains. “We had to define that protocol and propose it to the FDA to see if it would be acceptable evidence that we could deliver this information directly to consumers.” In the end, the design team went to multiple sites around the country to show reports to 667 randomly recruited subjects who represented a cross-section of the population. Ninety percent of these subjects had to (and did) demonstrate that they could retain key concepts from the report in order for the FDA to give 23andMe approval to proceed. The protocols that 23andMe’s design team developed will set the standard for all other direct-to-consumer genetic testing companies that are seeking FDA approval.
By February of this year, 23andMe had managed to convince the FDA that it was equipped to offer consumers data about their status as carriers. It took several months for the agency to post its full decision summary, and now 23andMe is finally marketing a testing service that will offer both ancestry information as well as genetic data about autosomal recessive disorders that you might be capable of passing on to your children.
It’s early days for just one category of tests, and we’ve yet to see how this new chapter plays out for 23andMe–but this is still an important step forward for the industry: By working with companies like 23andMe over the past few years, the FDA is creating a path for approving other health-related genetic tests that consumers can purchase directly, without the intervention or consultation of a physician. “The FDA appreciates that many consumers would like to be informed about their genomes, and their genetic risk for development of future disease,” says Jennifer Doolen, the FDA spokesperson. “The FDA has worked and will continue to work closely with any company marketing direct-to-consumer genetic tests without stifling innovation.” However, she says that not all health-related tests are the same; some pose relatively low risk, but others require proof that they are safe, effective, and include safeguards that will mitigate risks to the consumer.
Arthur Caplan, the NYU professor of biomedical ethics, points out that even with FDA clearance, there are important ethical quandaries to work out about privacy and the consumer’s ability to process their genetic risks without counseling. “Remember that the FDA’s job is not to police new technologies,” he says. “It’s just to prove that they are safe and that they work. In the U.S., we have left it to the market work to out ethical questions: That’s not a great solution unless we act in a responsible manner.” That’s why Caplan believes consumers and medical professionals should keep pushing back against companies like 23andMe when it comes to ethical questions.
Still, with a regulatory system in place, 23andMe is in a better position to provide evidence to the FDA that it can legitimately offer other health-related genetic tests directly to consumers. As Wojcicki looks back at the FDA’s correspondence with the company, she believes that the agency has shown a remarkable willingness to embrace 23andMe’s mission, although it was hard for her to see this back in the days when the two organizations were not communicating smoothly with one another. Despite the pain, she says, the FDA’s approval has been worth the wait.
“Being the first FDA authorized direct-to-consumer genetic test out there is revolutionary,” Wojcicki says. “I think a lot about the consumer health movement, and I believe this is an important step forward. This idea that you can take control of your own health and challenge your health care is a new idea. I think that’s what we’re waking up in a lot of people.”