The Dangers of Do-It-Yourself DNA Testing

This article originally appeared on Health Beat.

Recently, Time magazine listed the retail DNA test as its best invention of 2008 (thanks to Kevin M.D. for the tip). The best?  Maybe one of the most worrisome.

Time specifically highlights the do-it-yourself DNA testing kit from 23andMe, a California-based corporation named after the 23 pairs of chromosomes in each human cell.  The company sells $399 DNA kits that consist of a test tube in which you spit and send to the company's lab. There, over the next 4-6 weeks, researchers extract DNA from your saliva and map your genome, putting the results online. You can access the results through the web and navigate a guide to your genes that estimates "[genetic] predisposition for more than 90 traits and conditions ranging from baldness to blindness." 

Admittedly, this sounds pretty cool. As Time gushes, "in the past, only élite researchers had access to their genetic fingerprints, but now personal genotyping is available to anyone who orders the service online..." But look closer at the commoditization of DNA testing and the novelty wears off pretty quickly.

By pinpointing specific genes associated with certain diseases, a 23andMe gene read-out can inform a user of his or her susceptibility to those conditions. It turns out this is a lot less useful than it might seem. For example, Time reports that one test showed that the husband of 23andMe's founder has a rare mutation that gives him an estimated 20 percent to 80 percent chance of getting Parkinson's disease. The couple's child, due later this year, has a 50 percent chance of inheriting this mutation, and thus his dad's risk of Parkinson's.


At this point, the parents-to-be have to worry that their kid will have a mutation associated with an incurable disease. If he has it, they also have to fret that he has anywhere from a one in five to a four in five chance of actually contracting the disease. Really, how helpful are these numbers? That's a big range of probabilities. I wager it doesn't feel terribly good to be tracking the genetic lottery of your son's health, disease by disease.  In fact, I imagine that it's downright harrowing.

Dr. Alan Guttmacher, acting director of the National Human Genome Research Institute of the National Institutes of Health, agrees. In September, he told the New York Times that "[DNA testing] can be neat and fun, but it can also have deep psychological implications" because it can profoundly influence the way we view ourselves, our loved ones, and our relationship to the world. As Guttmacher told Time, "a little knowledge is a dangerous thing."

Here Guttmacher isn't just talking about the strange helplessness of knowing the ever-so-approximate probability of your child getting sick. He's also speaking to the fact that DNA tests themselves only provide a little knowledge -- just one small piece of the complex puzzle that is our health. Unfortunately, DNA tests often promise much more than this. One company, Navigenics, is actually dedicated to reading your DNA and diagnosing you with a set of medical risk factors that you then discuss with an appointed "genetic counselor." The idea is that genetic tests reveal some sort of fundamental physiological truth; a complete and comprehensive assessment of our health.

It's true that some conditions, like cystic fibrosis and Huntington's disease, have been scientifically proven to be associated with particular genetic mutations. But many other conditions have not been shown to have a genetic origin -- particularly when that gene is detected without an intimate understanding of environmental factors surrounding a patient, as it is the case when researchers on the other side of the country analyze your spit. 

In light of this fact, the Genetics and Public Policy Center, a project of Johns Hopkins University and Pew Charitable Trust, warns that many consumers "might have difficulty distinguishing between tests widely used and accepted by medical professionals...and those whose validity is unproven in the scientific literature." Customers will see their genetic print-out, with risk assessments for particular illnesses tagged on each gene. But they won't have a sense of how the DNA testing company calculated that number -- i.e. how much it reflects established medical research or a best guess from a company trying to convince you of its product's predictive possibilities.

Unfortunately, the latter seems to be more plausible. In 2006, the Government Accountability Office (GAO) purchased 14 DNA tests from four different websites and sent in samples. The office found that "the results from all the tests GAO purchased mislead consumers by making predictions that are medically unproven and so ambiguous that they do not provide meaningful information to consumers."

From GAO's saliva samples, the companies sent back risk predictions for conditions like diabetes and osteoporosis with little qualification, even though "scientists have very limited understanding about the functional significance of any particular gene, how it interacts with other genes, and the role of environmental factors in causing disease." In other words, the tests were spitting out numbers and warnings even though the genetic causality of these conditions "cannot be medically proven."

Further, many other results were all but "meaningless. For example, [the companies reported that] many people 'may' be 'at increased risk' for developing heart disease." But this is true for pretty much everyone, "so such an ambiguous statement could relate to any human that submitted DNA." The laughable superficiality of the companies' test results carried over into lifestyle information that GAO provided: when the office told a company that the patient from whom the sample derived smoked, the DNA company recommended that they stop smoking. When the patient reported that he had quit, the company "gave recommendations to continue to avoid smoking." Gee, thanks -- is that really worth $400?

Ultimately, in the words of Dr. Muin Khoury, director of the National Office of Public Health Genomics at the Centers for Disease Control and Prevention, "the uncertainty [of medicine] is too great," to view DNA testing as a sort of medical crystal ball. Even within the context of our genes, the possibilities are endless. To its credit, Time points out that "many diseases stem from several different genes and are triggered by environmental factors. Since less than a tenth of our 20,000 genes have been correlated with any condition, it's impossible to nail down exactly what component is genetic."

In fact, even when doctors do know that there's a genetic component to a given condition, they're not always sure which genes to look at. For example, in 2006 the Boston Globe noted that "there are hundreds of mutations in two well-known breast cancer genes, BRCA1 and BRCA2, for which reliable commercial tests exist. A woman could be told that she didn't have the common mutations but might still be at high risk from less common mutations or a different gene altogether..." Translation: even though we know there's a genetic component to breast cancer, it's very difficult to pinpoint which gene is the problem -- particularly if the only way of communicating with a patient about the issue is watered-down risk probability.

Meanwhile, the vagueness of DNA test results works in the favor of testing companies. If they keep things simple and superficial, they can make cross-promotion easier. In the GAO study, for example, the DNA test results were synched with expanded product offers such as dietary supplements, which had only a tangential relationship to the patients' test results.

This sort of aggressive marketing is direct-to-consumer medicine at its most profitable. Companies often want to convince patients that they have a certain condition and then sell them on the cure. In prescription drugs, this "disease mongering" has usually been about listing symptoms to get people scared. But DNA testing kicks things to another level: convince people that they are actually hard-wired to contract a particular disease, and your cure becomes that much harder to resist.

It's no wonder that experts at Johns Hopkins are worried that "advertisements may...underemphasize the uncertainty of genetic testing results, or exaggerate the risk and severity of a condition for which testing is available, thus increasing consumer anxiety and promoting unnecessary testing." Given what we've seen in direct-to-consumer medicine up until now, this is a very reasonable fear.

Another plausible concern is that DNA tests, in their superficiality and over-simplification of medicine, will be routinely misinterpreted by patients. Time cites the case of Nate Guy, a 19-year-old in Warrenton, Va., who "was relieved that though his uncle had died of prostate cancer, his own risk for the disease was about average," according to his 23andMe test. This sounds uplifting until you realize that, by the age of 70, the vast majority of men have prostate cancer. Almost all of them will die with prostate cancer, not from it. (Something else will kill them before this very common, but usually slow-growing, cancer catches up with them.)  An "average risk" of prostate cancer means you'll probably get prostate cancer and live with it for years, just as do nearly all older men.

Presumably, Guy doesn't know this. One gets the sense that he thought his uncle died of prostate cancer because he died with prostate cancer, and that this fact meant that his uncle had been uniquely susceptible to the disease. Now Nate finds out he has an average level of vulnerability and thinks that he won't get prostate cancer. Statistically speaking, none of this is probably true -- but this is the sort of reasoning that happens when patients are confronted with misleading, sparse data about their health, devoid of a broader medical context.

One can imagine that, had Nate been disheartened with the results of his test, he would have similarly embraced the definitiveness of the results and undergone unnecessary prostate screenings throughout his life -- screenings which have never been shown to actually improve survival rates. Either way, Nate's taking the wrong message from his genome. Indeed, the likelihood that the patient will go for more screenings -- just to be safe -- combined with the fact that people are paying $400 a pop for a test which vaguely suggests whether they may or may not contract a disease makes DNA tasting a profoundly cost-ineffective health care option.

Genomics is a field that's new and exciting; scientists will and should pursue it. But it's probably not something you should try at home. From what we've seen so far, do-it-yourself DNA testing risks exacerbating many of our most pressing health care problems: the deceitfulness of for-profit medicine, the dangers of direct-to-consumer health care, the glut of wasteful, potentially harmful, screenings, and the general misconception that -- if our gizmos are fancy enough -- we can all live forever.

Understand the importance of honest news ?

So do we.

The past year has been the most arduous of our lives. The Covid-19 pandemic continues to be catastrophic not only to our health - mental and physical - but also to the stability of millions of people. For all of us independent news organizations, it’s no exception.

We’ve covered everything thrown at us this past year and will continue to do so with your support. We’ve always understood the importance of calling out corruption, regardless of political affiliation.

We need your support in this difficult time. Every reader contribution, no matter the amount, makes a difference in allowing our newsroom to bring you the stories that matter, at a time when being informed is more important than ever. Invest with us.

Make a one-time contribution to Alternet All Access, or click here to become a subscriber. Thank you.

Click to donate by check.

DonateDonate by credit card
Donate by Paypal

Don't Sit on the Sidelines of History. Join Alternet All Access and Go Ad-Free. Support Honest Journalism.