Anne Wojcicki, chief executive of 23andMe, delivered a keynote presentation at South By Southwest Interactive on Sunday afternoon that, in many ways, amounted to a defense of the company as a provider of health information.
In November, the Food and Drug Administration demanded the personal genetics company immediately stop marketing its “Saliva Collection Kit and Personal Genome Service,” saying it had yet to demonstrate the accuracy of those tests. In the time since, the Mountain View, Calif., firm has been restricted to selling ancestry reports and raw genetic data.
In an onstage Q&A following her presentation in Austin, Texas, with Re/code’s Kara Swisher, Wojcicki acknowledged the agency’s decision was a blow to the company.
“Having the FDA come and shut that down has had a significant impact on the business,” she said.
During the keynote, titled “The Future of Genetics in Our Everyday Lives,” Wojcicki stressed that 23andMe is no longer providing health information.
But she went on to list a series of earlier examples where customers benefited directly from the genetic information provided by the company’s $99 test. Among them: a woman who finally realized her daughter was suffering from celiac disease; a father who realized his son was fructose intolerant; a doctor who now takes daily aspirin after realizing he had mutations that put him at higher risk for blood clots; and a woman who was able to tell her doctors just before emergency surgery that she had a pseudocholinesterase deficiency, which made her sensitive to certain anesthetic drugs.
Wojcicki argued that the traditional healthcare system is economically incentivized to treat disease rather than to encourage prevention. And she suggested that the regulatory environment needs to catch up with the genomics revolution sweeping medical research and personalized medicine.
“Everyone has the right to get access to their genetic information and to understand it,” she said. “It’s your data, it’s all about you.”
During the Q&A, she added: “The FDA serves a real purpose: To protect public health. So I think there’s a question about all this genetic information coming out: What’s the right way to regulate it?”
When directly asked if the regulatory guidelines should change, she said: “There’s a lot of interest in changing the guidelines to become more accommodating with full genetic data.”
But Wojcicki’s presentation largely sidestepped the two chief complaints of the FDA: That 23andMe hasn’t proven its tests are accurate and that it ignored the agency’s repeated requests for information.
The strongly-worded FDA letter noted:
FDA is concerned about the public health consequences of inaccurate results from the PGS device; the main purpose of compliance with FDA’s regulatory requirements is to ensure that the tests work.
… Months after you submitted your 510(k)s and more than 5 years after you began marketing, you still had not completed some of the studies and had not even started other studies necessary to support a marketing submission for the PGS. It is now eleven months later, and you have yet to provide FDA with any new information about these tests. You have not worked with us toward de novo classification, did not provide the additional information we requested necessary to complete review of your 510(k)s, and FDA has not received any communication from 23andMe since May.
Reliability is more than a theoretical concern. A series of reports have found that results from the specific type of genetic test in question can provide variable or erroneous results, either in the genetic data itself or the conclusions drawn from it. One customer submitted to two 23andMe tests and found “85 SNPs were called differently.”
At this point, the company doesn’t sequence the full genome, nor even the complete exome (the part of the genome that codes for genes). It just looks at SNPs. These single-nucleotide polymorphisms are the roughly 10 million base pairs among some 3 billion in the human genome that are different in an individual. In other words, it’s only about 0.33 percent of the entire genome.
“The SNP chip method that 23andMe uses was never very good at providing useful genetic information,” wrote Hank Greely, a Stanford law professor focused on legal and ethical issues in biosciences, in a piece for VentureBeat. “Its advantage has been its low cost.”
The other problem is interpretation of those results. Some genetic diseases are clear cut. Three copies of Chromosome 21 instead of two means Down syndrome.
But in many other instances genetic variations merely increase the likelihood of diseases. An estimated 45 percent to 65 percent of women with the BRCA1 or BRCA2 mutation will develop breast cancer by the age of 70, according to the National Cancer Institute. (Others put it as high as 80 percent.) And inheriting a single copy of the ApoE4 gene boosts the risk of Alzheimer’s by about threefold, according to an NIH study.
The rest of the explanation for why a person with those mutations will or won’t actually develop diseases lies in some little-understood combination of environmental and lifestyle factors, or the complex interplay of other genes.
In other words, the results are not conclusive; the human code is rarely as clear cut as the computer variety.
So the method by which this information is imparted to consumers becomes critically important. Being told you’re triple the risk for Alzheimer’s sounds scary, but the average lifetime risk at age 65 for a man is 9.1 percent, so that person is still below 30 percent. In other words, he has better than even odds that he won’t get the disease.
So the question for the FDA, 23andMe and other companies working in this area becomes whether SNP results and links to some studies are an appropriate way of delivering this kind of news, or whether a trained medical professional who can put those numbers in an appropriate context should be involved.
“You might freak out, say ‘to hell with my 401(k), I’m taking early retirement,'” Greely told me. “But you’re more likely than not not to get Alzheimer’s.”
But he’s even more concerned that negative results could give people a dangerous false sense of security. For instance, if a woman doesn’t have a BRCA1 or 2 mutation, she might decide to skip her annual mammogram.
“You may be overly reassured by a negative result,” he said. “Not having a BRCA1 or 2 mutation means you are not at 55 to 80 percent risk of breast cancer, but it lowers your overall risk from 12 percent to about 11.5 percent.”
In other words, barely better than the general population.
During the Q&A, Wojcicki acknowledged that there’s a fair debate over what information from medical literature is relied upon and delivered to customers. But she insisted the SNPs test they use is “a gold standard,” with 99.99 percent reproducibility.
She added that the company is interested in exploring other technologies for genetic testing as they become standardized and reliable. Indeed, the company has already conducted a pilot program for sequencing the entire exome.
“The potential for 23andMe is really to massively change the way we are approaching our own health in terms of the prevention side and really change the way that we’re doing research,” she said. “I hope in 10 years we can have a society where people are proactively managing their health and getting information that helps them avoid disease.”
This article originally appeared on Recode.net.