What Comes Next for Direct-to-Consumer Genetics?
By Aaron Krol
July 16, 2015 “Mr. Obama’s regulators are chilling innovation and investment,” fumed the Wall Street Journal. “The FDA bureaucrats think they know better than you how to handle your genetic information,” Ronald Bailey protested on the website of the libertarian magazine Reason. At Slate, Gary Marchant punched out every adjective on the soda lid: “a shortsighted, heavy-handed, double-standard act of paternalism… a death sentence for direct-to-consumer genetic testing… unjustified and unwarranted.”
More than a year and a half after the FDA told 23andMe to take its genetic health reports off the market, Elizabeth Mansfield, the agency’s Director of Personalized Medicine, still sounds frustrated with the indignation the decision met with. “What a lot of people don’t understand,” she tells Bio-IT World, “is that FDA did not ban direct-to-consumer testing when it sent 23andMe a warning letter. It said, 23andMe, you in particular are out of compliance, and we’ve tried to work with you and it hasn’t worked out, so please stop selling your tests until you get into compliance.
“And that does not reflect any dissatisfactions with the model or anything like that.”
This February, the FDA made good on that assertion, giving 23andMe the go-ahead to test customers for a single factor: whether their children are at risk for inheriting a rare genetic disease called Bloom syndrome. The FDA-cleared test uses the same spit kit and genotyping panel 23andMe has always offered, albeit with more stringent labels and controls than customers may be used to. Mansfield thinks the process has cleared the air, and gives some assurance the FDA is not out to quash consumer genetics.
“I believe that [23andMe] have a much better understanding of what we’re looking for in terms of a regulatory submission,” she says. “We can move forward more easily with them.”
It’s an important precedent for the genetic testing industry, laying a straightforward path for a host of future tests. It’s also a far cry from the reports on 254 different health conditions that 23andMe once advertised.
A Path Forward
At the time the FDA’s warning letter went out in November of 2013, 23andMe was the last major ― certainly the last reputable ― company letting people buy genetic health tests online, without going through a doctor or involving their health insurance. A large crop of “direct-to-consumer” companies like Navigenics and deCODEme, formed in the mid-2000s amid the optimism of the recently completed Human Genome Project, had all long since folded or changed their sales models. When 23andMe gave up the health reports that had once told customers their genetic risks for conditions from Alzheimer’s to Zellweger Syndrome, it seemed to critics of FDA policy that the last bulwark for personal genetics had fallen.
Some of the doomsayers of late 2013 are probably surprised to see how 23andMe has thrived since. The company has crossed its long-anticipated threshold of one million customers, on the back of overseas sales and still-popular tests here at home for ancestry and non-medical genetic traits. 23andMe has penned lucrative deals with major pharma companies, started its own drug discovery business, and recently been valued at a market cap of over $1 billion, making it Silicon Valley’s first “unicorn” mega-company in the field of biotechnology.
The clearance of its Bloom syndrome test, however, has been the only tangible sign of progress on restoring 23andMe’s original direct-to-consumer business. The Bloom test doesn’t mean customers have access to any more information than before ― the company is holding out for a more comprehensive panel of tests before releasing this one ― but it does lend some spine to 23andMe’s optimism about its prospects.
23andMe is not speaking to the press about regulatory plans, citing ongoing discussions with the FDA about additional health tests. But a spokesperson for the company did tell Bio-IT World by email that they now expect to sell a larger direct-to-consumer product in the near future. “[W]e believe there is a regulatory path forward for direct-to-consumer genetic testing,” the spokesperson wrote. “Now that the parameters have been established for future submissions, we will have a health-related product that includes carrier status test reports on the market in the U.S. later this year.”
That’s a swift turnaround for a company whose ability to sell any health product was up in the air as recently as the start of this year. Carrier testing is an interesting and relevant part of genetic medicine, and a sound floor to build a comprehensive product on ― as long as it’s not a ceiling on what genetic tests can get by in a regulated market.
Proof of Principle
Bloom syndrome is a serious condition. People affected by the disorder suffer from stunted growth, frequent rashes related to sun exposure, and most importantly, a frighteningly high risk of getting cancer at an early age.
It’s also extraordinarily rare, with fewer than a thousand recorded cases. There’s a sound commercial reason 23andMe isn’t hustling to get its solitary FDA-cleared test onto pharmacy shelves.
“It’s sort of a proof-of-principle experiment,” says Hank Greely, a bioethicist and legal scholar, and Director of the Center for Law and the Biosciences at Stanford Law School. Greely is a longtime skeptic of 23andMe’s health tests, which he regards as underpowered to give reliable information on most diseases, but even he sees the test cleared by the FDA as a pretty straightforward case.
“The thing I like about it,” Greely says, “is FDA applied the same kind of analysis they apply to over-the-counter drugs and devices. Which I think makes sense. Before you can sell a painkiller or a decongestant or some other drug over the counter, you need to prove to FDA that the average consumer would be able to understand when and how they should use it.”
The question of understanding is an important one for any medical product, but it’s an especially vexed issue in genetics. “Genetics is not widely understood, we don’t think, among the public,” says Mansfield. “At some point we have to make sure that the way the company is presenting the information about the test, how it’s interpreted and so on, can be understood by the people who are expected to use it.”
That caution is reflected in the new labeling the FDA has required for 23andMe’s Bloom test. If sold direct-to-consumer, the test would have to present users with warnings like, “this test is not intended to diagnose a disease, or tell you anything about your risk for developing a disease in the future,” and, for those who test positive, “the positive result you obtained is very likely to be incorrect due to the rarity of this variant.” 23andMe also had to complete a user comprehension study with its customers, and must make certified genetic counselors available to guide users through the quirks of genetic testing — like the implications for family members if a disease-related variant is found.
Apart from the genetic counselor, these are standard parts of the FDA’s toolkit, as Mansfield is quick to point out. “We weren’t dealing with anything in that Bloom syndrome test that we hadn’t encountered before,” she says. “We’ve done over-the-counter before, and we’ve done genetic tests before. It was just combining the two.”
Mansfield makes a convincing case that selling direct-to-consumer is not a deal breaker for the FDA. Much more relevant than your sales model is whether your product is reliable, comprehensible, and medically sound. In personal genetics, that’s where things tend to get murky.
The Big Jump
At the height of its popularity, 23andMe took a comprehensive, cradle-to-grave approach to health testing. Alongside obscure early-onset diseases like Bloom syndrome, 23andMe also gauged customers’ risks for everyday killers like diabetes and heart disease, and late-in-life specters like Alzheimer’s ― diseases we all know and worry about.
“It’s a big, big jump from the Bloom syndrome test to most of the 254 medical conditions they were releasing information on,” says Greely. “And the farther away you move from what was actually approved, the fuzzier things get.”
Most of the genetic variants thought to play a role in common diseases have only marginal effects. We all carry hundreds of variants that have been linked to small changes in risk for things like breast cancer and diabetes, often in studies that have been hard to confirm or replicate. 23andMe can try to aggregate all those variants together, but not only will those reports rest on shakier foundations than rare disease tests, they’re also unlikely to add up to more than a 10 or 20% change in total risk either way.
“I don’t know why anybody would want to know that,” says Greely. “I don’t know what you’d do with it.”
Problems of user comprehension are hugely magnified in this kind of health report. Customers are being asked to understand what a marginal change in disease risk means for their overall health; the type of evidence, from population-wide studies, that supports the effects of these low-impact mutations; how the test provider is compounding the effects of many variants together; and the difference between absolute and relative risk — the kind of statistical thinking that says a 20% change to a 10% risk adds up to 12 and not 30.
That doesn’t mean direct-to-consumer genetic tests for common diseases are off the table. Mansfield, for her part, wants to keep an open mind. “It would be wonderful if people could actually understand their risk there,” she says. “To the degree that we can figure out how to do it, and that there actually is a reasonable genetic underpinning for a disease, I think it’d be interesting.” But she also stresses that there are thorny issues in this area that the FDA’s review of the Bloom test never touched on.
“The known genetic component for common diseases is not very strong,” she says. “A lot of the risk comes from things that aren’t genetic, so how you convey the information to users is much more complex.” Among Mansfield’s worries is that a genetic test might lead users to downplay lifestyle factors like diet, exercise, and smoking, or standbys of preventative medicine like routine cancer screening.
Maybe the standard tools for over-the-counter tests can handle all those layers of uncertainty; Mansfield isn’t ruling it out. But it’s hard to imagine all the labels the FDA would require for a test like that.
A Watershed Event
Meanwhile, in the vacuum left by the mass exit of direct-to-consumer companies, other kinds of genetic testing have been expanding.
In 2008, Ramji Srinivasan co-founded a company called Counsyl, one of many trying to make a go at direct-to-consumer genetics. Like many of its competitors, Counsyl quickly reinvented itself after initial conversations with the FDA. It now sells tests for carrier status in more than 100 rare diseases ― including Bloom syndrome ― on a prescription basis, often through hospitals and doctors’ offices.
To Srinivasan, direct-to-consumer testing is just a means to an end: giving more people access to genetic information about their health. That kind of information can help people make informed decisions about family planning and preventative medicine, but only to the extent that they’re aware of genetic testing and will seek out testing services. “Getting people to do preventative medicine is a very hard task,” he says. “Not that many people have heard of cystic fibrosis, or Smith-Lemli-Opitz, or impact carrier screening. We and others are working to educate people about the value of carrier screening, but more work needs to be done on that basic education front to build that market.”
Now that the FDA has published clear guidelines for direct-to-consumer carrier tests, Counsyl is in an enviable position to get back in the consumer market. The company employs more than 30 genetic counselors and has a well-developed online consultation process, which should make it relatively painless to implement the education measures the FDA has asked for. But Srinivasan says he has no intention of selling direct-to-consumer again, at least not for the foreseeable future.
“I’m not totally convinced that the market is there yet,” he says.
Genetic health companies have always faced a big challenge in pitching an opaque product like genetic testing as a lifestyle tool with mass-market appeal. “Ideas like [direct-to-consumer] that make carrier screening more widely accessible I think are a net good,” says Srinivasan. But his years adapting Counsyl’s business to the existing healthcare system have given him new ideas about extending the reach of genetics, which focus less on making tests easy to buy and more on reaching people when they most need to learn about their genetic risks.
“There’s still a huge untapped opportunity in the OB office,” he says. “There are millions of women who are going every year to the office, and getting carrier screening. We don’t necessarily need to reinvent distribution in order to reach those people.”
In fact, he adds, access to genetic health information has been steadily growing in recent years, even in the absence of any direct-to-consumer sales. Srinivasan’s favorite success story is non-invasive prenatal screening, a method of testing a fetus for chromosomal disorders months before birth, using a small blood sample from the mother. Hundreds of thousands of these genetic tests are now performed in the U.S. every year, by Counsyl and many others, a pace of adoption that has led to a sharp decline in more invasive procedures.
“The growth of non-invasive prenatal screening has been extremely exciting to watch, because that’s credible proof of momentum,” says Srinivasan. “It’s completely changed the economics of maternal fetal screening. That has been a watershed event.”
Other companies that were once in the direct-to-consumer business might see things differently. The regulatory environment is certainly more inviting now than it was two years ago. For instance, the FDA has set the precedent, during its Bloom review, that it will happily look at scientific publications and professional guidelines as evidence that a genetic variant is involved in disease, rather than demanding new clinical trials. That matches past policies for reviewing genetic tests in the hospital environment, and clears away one of the biggest expenses of approving a new product.
The FDA also decided that all genetic tests similar to 23andMe’s Bloom syndrome product, those that look for carrier status in rare, recessive diseases, should now be considered “Class II” medical devices. That classification exempts them from the highest level of premarket scrutiny ― new manufacturers will only have to prove that their products are “substantially equivalent” to the tests created by 23andMe.
“We cleared a test,” says Mansfield. “We down-classified an entire group of tests based on that. I think it’s pretty clear that we see this as an okay model.”
But she’s also content to take it slow. Every new indication for genetic testing will call for a new review to tease out its unique risks and benefits, and few will be as clear-cut or easy to understand as carrier testing. If what we want is to reach the greatest number of people with important genetic information, we would be well advised to focus on more than where and how tests can be sold.
“I think right now,” Mansfield adds, “at the stage we are at in medicine and understanding, it’s not a terrible thing for there not to be a lot of direct-to-consumer tests that are purporting to provide very important medical information.”
The FDA has not yet formally published its letter to 23andMe clearing the company’s Bloom syndrome test, and down-classifying similar “autosomal recessive carrier screening gene mutation diagnostic systems.” However, the document can be read on Scribd.