When genetic tests are more misleading than helpful
Published: Thursday, December 19, 2013 at 3:00 a.m.
Last Modified: Wednesday, December 18, 2013 at 5:19 p.m.
Cutting edge companies often walk a tightrope between regulators trying to keep their technologies under control and marketers trying to push them out to consumers as fast as possible.
That's where a Silicon Valley company named 23andMe is today. The Mountain View firm has been hawking genetic tests for you to take at home. You spit into a receptacle and ship your saliva back to the company so it can analyze your DNA for a mere $99. Eventually you get a readout detailing your genetic susceptibility to hundreds of diseases.
At least, that used to be the case. At the end of November, regulators at the Food and Drug Administration instructed the firm to shut down its genetic analysis service, declaring — with plenty of justification — that the company's marketing claims were running well beyond what was valid, or legal. The service fell within the legal definition of a medical diagnostic device, the agency said, and demanded data demonstrating the tests' technical validity. It warned 23andMe to ratchet back its marketing claims for the service, which it said requires agency approval before it can be sold to consumers — “as FDA has explained to you on numerous occasions,” the agency said.
The company will still send customers their raw genetic data and an analysis of what it says about their “ancestry,” but it has stopped issuing disease-risk analyses until it gets right with the FDA. Customers who ordered their tests on or after the date of the FDA's warning letter, Nov. 22, can get a refund.
Personal genetic testing has been building toward a craze for some time, as evidence mounts that certain genes or mutations can affect individuals' health profiles. Consumer interest often spikes with news events: Anne Wojcicki, the founder and CEO of 23andMe, says inquiries poured into her office in May, after actress Angelina Jolie disclosed that she'd undergone a precautionary double mastectomy upon learning she carried a gene that predisposed her to breast cancer.
23andMe — the firm's name derives from the number of chromosome pairs in the human cell — has become the best-known company in the field for a couple of reasons. One is the unique pizazz of its corporate pedigree. It's backed by Google, whose co-founder Sergey Brin is Wojcicki's husband. (They separated earlier this year.) Another reason is its aggressive national advertising, for which it budgeted $5 million for 2013 alone. But that was before the FDA swooped down; the TV ads have been taken off the company's YouTube channel.
The problem with the service provided by 23andMe and its competitors — and the root of the FDA's concern — is that raw genetic information is very hard for a lay person to interpret. “The technology is just not ready for prime time,” says David B. Agus, a University of Southern California cancer specialist who co-founded Navigenics to market professional genetic testing services.
Agus says test interpretation should be done by physicians or trained genetics counselors. “Medicine is lots of shades of gray, so you need discussion with someone who's trained in the field,” he says. Agus says he has no financial interest in Life Technologies, which acquired Navigenics in 2012, though he sometimes offers the company unpaid advice.
The reports issued by 23andMe prior to the FDA letter typically included pages of qualifications of the genetic results for every health issue, followed by a broad disclaimer stating that the information “is intended for research and educational purposes only, and is not for diagnostic use.” That unnerves medical experts, who think many customers will stop reading when they reach the headline figure about their risk of acquiring a disease or condition compared with that of the general population. The firm's marketing can lead customers to assume they actually are getting a diagnosis.
“Any reasonable interpretation would assume this is news you can use, until you get to the disclaimer,” says David Hunter of the Harvard School of Public Health, a critic of direct-to-consumer marketing of genetic testing. “The problem is we are notoriously poor at estimating risk and communicating relative and absolute risk.” That may be especially true in genetics, where the effect of a single gene or mutation can be vanishingly incremental but the revelation that you carry it can strike like the tolling of a bell.
Genetic probabilities are commonly misinterpreted. In a fawning recent profile of Wojcicki, for example, the magazine Fast Company reported that Google's Brin changed his lifestyle after a 23andMe test revealed that he carries a genetic mutation associated with a higher risk of Parkinson's disease — “30 percent to 75 percent, compared to 1 percent for the general population,” the magazine reported.
That sounds like a horrific near-certainty that Brin will succumb. But it actually means that if Parkinson's affects 1 percent of the population, Brin's risk is 1.3 percent to 1.75 percent — based on his genetic profile alone, and leaving aside the role of environmental factors in Parkinson's. Nor do the raw figures explain that the incidence of Parkinson's rises sharply with age, with as little as 4 percent of all cases appearing before the age of 50.
Such nuances are likely to escape many customers, as is the fact that the predictive power of genetic data is “a moving target,” as Hunter describes it. Other than a few diseases or conditions, we simply don't know a lot about the effect of specific genes or mutations on risk. “Someone who's told they're at high risk now might plausibly be told they're at lower risk at some time in the future,” he says. “We're still in the early stage.”
Another concern of the FDA is the technical accuracy of the tests themselves — that is, are the companies' claims for the presence and absence of specific genes accurate? Much is at stake. If a test inaccurately identifies the BRCA gene found in Jolie's work-up, the FDA warns, “it could lead a patient to undergo prophylactic surgery, chemoprevention, intensive screening, or other morbidity-inducing actions, while a false negative could result in a failure to recognize an actual risk that may exist.”
23andMe claims to hold its analytical labs to the highest standards, but it hasn't made that case to the FDA's satisfaction despite the agency's repeated demands for validating data.
Judging from the sharp language in the FDA's most recent warning letter, it certainly appears that 23andMe has been shining the agency on. The FDA says it's been talking and meeting with company executives for more than four years, asking for specific data and issuing directives on how to rewrite its marketing material so it's not misleading to customers.
The agency says it hasn't heard a word from 23andMe since May — not even a response to a preliminary warning it issued in June, after the firm started marketing its test kits through Amazon.com. Meanwhile, 23andMe launched new marketing campaigns, including television ads, proving it was determined to bull ahead no matter what.
The company hasn't said much about its fraught history with the FDA, except for a Nov. 26 statement by Wojcicki acknowledging that “we are behind schedule with our responses” to the agency. That's one way of putting it: a weaselly way.
There's a role for genetic testing, perhaps even direct-marketed testing. But so far, 23andMe has flunked an important test of its credibility. It could have acknowledged the public's interest in reasonable oversight and complied with the FDA's regulations like a mature, responsible company. Instead it has bobbed and weaved like a bucket shop. That's not a great start for a new industry.
Michael Hiltzik is a columnist for the Los Angeles Times.
All rights reserved. This copyrighted material may not be re-published without permission. Links are encouraged.