VICE | Jan. 30, 2014 — Geneticists, lawyers, students and industry experts packed into a conference room on Stanford’s campus this week to discuss the ethics behind the FDA’s decision to effectively shut down the health assessment element of the world’s leading direct-to-consumer genome analyzer, 23andMe.
“We’re here to talk about why the FDA dropped the anvil on 23andMe,” said Stanford Law School professor Hank Greely, a mustachioed man in a Cosby sweater who led the session. “Rules for consumer genomics have been up in the air for some time. Some of us have been calling for regulations for about a decade now.”
The promise of 23andMe is informing participants whether they have a mutation that makes them more susceptible to a certain disease. The company mails customers “spit kits” and then analyzes people’s DNA via their mailed-in saliva sample.
In November, the FDA issued 23andMe a warning to stop marketing its $99 in-home DNA collection spit kits, the most common means of consumer genome analysis in the US. The agency wants to ensure that the kits and tests performed in 23andMe’s lab are scientific and accurate in assessing disease risks. It’s unclear how, exactly, that process is unfolding.
Presumably, the FDA has asked 23andMe for proof of the validity of its tests, but the agency hasn’t said, so we don’t really know. Meanwhile, 23andMe has continually dragged its heels in responding to the FDA’s requests, as Motherboard’s Grace Wyler detailed last month.
“The FDA is working on something—guidance or clearance,” Greely said, “but it’s not really clear what.”
The 23andMe kerfluffle is of particular interest to academics at Stanford, which was ranked No. 1 among universities in genomics by US News & World Report. The outcome will likely shape the parameters of a field that is emerging as a cornerstone of the personalized health care movement.
Stanford anthropologist Sandra Soo-Jin Lee is gathering data on a sample of 23andMe’s estimated 500,000 genotyped customers. So far, she and her team have interviewed 80 people who traded some of their spit for a glimpse into their future health predicaments. At this point, Lee has more questions than answers, but her research is turning up a few commonalities among people who seek knowledge of their health profiles.
“From what we gather from the consumers there is a strong push for the right to information,” Lee said. People tell her, “‘This is my data and I should be able to have access to it. Even if I can’t gleam meaning from it I should be able to use it how I want.'”
“People thought they would do things differently going forward, but it’s not necessarily clinical planning,” Lee said. “It’s life planning.”
At the heart of the Stanford roundtable was whether there is harm in people being given health information that isn’t 100 percent certain and, consequently, whether the benefits of releasing genome analyses to people without medical savvy outweigh the costs of them misunderstanding that information.
Greely summed it up this way: “To what extent do you hold back people who can make safe use of the information from some of the people who can’t?”
When a woman’s analysis shows that she has at low risk for breast cancer, Greely said, she could be inclined to think, “Thank God I no longer have to take those damn mammograms.” But that’s not an accurate read of the data because having a lower than average risk doesn’t mean a person’s risk isn’t significant.
What is unclear, and what the FDA hopefully will shed some light on, is the validity of the information 23andMe and its competitors are selling.
Greely and two of his understudies have submitted an article on the matter to the Journal of the American Medical Association.
“Basically, our piece says that it’s a good idea to begin this process,” Greely told the conference room. “There are good reasons to doubt the clinical validity of the data that 23andMe gave out. Not that they’re fraudulent or bad … but [the test results] vary a lot depending on who is doing the interpretation.”
During an experiment in 2009 led by an epidemiologist at the University Medical Center Rotterdam, three different genome sequencing companies analyzed identical DNA samples and returned significantly divergent conclusions. In one instance, one company determined a customer to be “high risk” for type 2 diabetes while another company determined that same customer to be “low risk.”
It’s those kinds of contradictions that are giving people on the frontier of consumer genomics pause.
The pressure on the FDA to draw up guidelines for these tests is increasing as the cost of sequencing drops lower and lower. In 2001, sequencing cost north of $100 million per genome, according to the National Human Genome Research Institute. In 2013, it cost less than $10,000. The march toward low-cost sequencing technology is commonly called “the $1,000 genome.”
Actually, we may be there already. A company called Illumina claims to have already broken through the $1,000 threshold with its new line of “ultra-high-throughput” sequencers. (Illumina won approval from the FDA for the new clinical sequencing technology three days before the agency cracked down on 23andMe.)
That’s all the more reason to tee up the discussion of how deeply we want people to peer into their health histories, and their health future.
Toward the end of the Stanford discussion, Greely summarized the wider context of the consumer genomics debate thusly: “It’s a libertarian versus a paternalistic approach to health care,” he said. “Health has traditionally been paternalistic—we don’t really trust patients to always know what’s best for them. The libertarian view thinks we should move beyond that.”