publication date: Dec. 17, 2013


By Jim Evans

There’s been much recent gnashing of teeth (at least within the narrow and rather insular world of do-it-yourself genomics) about the gall of the FDA in choosing to issue a warning letter to 23andMe, effectively shutting down marketing of their direct-to-consumer personal genomic profiling. And adding to 23andMe’s woes, a class action suit has been filed against them.

The FDA’s call was an easy one. That agency is charged with regulating drugs, complex medical tests and medical devices to ensure that they are generally safe and that there is evidence to support the claims of efficacy made on their behalf. Had they not taken action they simply would not have been doing their job. 23andMe’s advertising about their product consistently claimed (both implicitly and explicitly) that the information they provided would improve your health.

The only problem was that there is no evidence to back up these claims, forcing them instead to rely on wishful thinking and anecdotes.

In the wishful thinking category, the idea that “knowing your risk” of heart disease will lead to sustained and effective action by individuals to minimize that risk has simply not been supported by evidence. Indeed, it flies in the face of everything we have learned about how hard it is to change human behavior; the idea that genetic risk information would be a magic bullet to influence human behavior for the better never held much promise in the first place.

And in the “anecdote” category, where (highly selected) happy customers trot out a personal story about how 23andMe changed their lives, well, I have a good friend who reads her horoscope every day. Last Tuesday she had a really great day, completely in accordance with what the newspaper said that Capricorns should expect. Anecdotes are a poor means of divining reality, and this is as true for medicine as in any other realm.

The bottom line is that genomic analysis is a complex medical test with the power to help, to harm, and to confuse. The FDA was simply acting within its mandate and made a reasonable demand regarding 23andMe’s product: show us the data in support of your claims. The company couldn’t do that, so the FDA took the logical next step. When I buy a toaster I expect it to live up to the claims that have been made. Why should we expect anything less for something advertised as a medical test?

But the more interesting question, and what I think underlies most of the hand wringing by those who are offended by the FDA’s actions, is a libertarian argument: “I want it, I’m paying for it, so who is the FDA to tell me that 23andMe can’t sell it to me?”

I will freely admit to having little personal sympathy for most libertarian causes, but I concede that in some realms libertarians make strong arguments.

However, a libertarian stance in the medical arena is wholly untenable for one very simple reason—medical care is a shared endeavor. Nobody—not you, not me, and I warrant not even Bill Gates—pays for their own medical care. Rather, we all participate in some kind of shared risk pool, be it private medical insurance or a public scheme such as Medicare or Medicaid.

I’ll grant the libertarians their abstract right to incur bodily harm if they really want to get poorly understood medical tests and act upon the results. Mind you, I’m not in favor of this—I’m paternalistic enough to want to protect people from harm; indeed, it’s one reason I became a doctor. But okay, fine. In the end you do have the right to harm yourself. I get that.

The fatal flaw with the libertarian argument in medical care, however, is the financial side of the equation. When complex medical tests are misused, misinterpreted and poorly understood, they generate tremendous downstream costs in terms of both harm and financial expenditure. So I’m going to end up paying for your misapplication of a half-baked medical test.

And make no mistake, genomic tests will be misapplied if we roll them out before we have evidence to guide their proper use. Considerable harm has resulted from the broad uptake of many much simpler tests (PSA, anyone?). We shouldn’t beg for trouble by prematurely pushing genomic testing before we understand how to use it. In a similar vein, physicians routinely resist pressure from patients to prescribe antibiotics in situations where they are not indicated.

We do this both because we don’t want to subject patients to the harms of an unnecessary drug, but also because others shouldn’t have to pay for the downstream costs of such misapplication when the patient develops a resistant infection.

Ultimately, the libertarian argument fails in the realm of medicine because we all pay for each other’s medical care. We all have a vested and legitimate interest in seeing that medical tests are used wisely (incidentally, this line of reasoning also gets to why most of us feel that a well-funded FDA is important in general).

The danger of the misuse of genomic data, or for that matter any medical intervention, isn’t simply the worry that patients will occasionally “convince” medical providers to engage in dramatic, unwarranted interventions like bilateral mastectomy when an invalidated test suggests an increased risk.

The temptation by the medical profession to act upon information is well-nigh irresistible and, unfortunately, doesn’t hinge solely on patient advocacy. At a minimum, poorly validated tests lead to other downstream tests and often to invasive procedures. For this reason, it is a well-established and well-supported tenet in medicine that seeking out information that you really don’t know what to do with is a bad idea, since it causes great mischief for both the patient and for healthcare costs.

We possess all kinds of medical technologies that are very cool and offer massive amounts of information. Modern medical imaging modalities such as ultrasound, MRI and CT are almost miraculous. But their indiscriminate application is discouraged precisely because financial and physical harm result (e.g. from needless downstream testing) when any medical technology is used unwisely.

Don’t get me wrong—genomics has much to offer medicine. Frankly, few are more excited about the prospects than I. But that doesn’t mean we can afford to be seduced into ignoring hard-learned lessons about the need for evidence before we implement new tests, drugs and technologies. Parenthetically, I would also ask whether you think that market forces; including clever (and misleading) advertisement is an appropriate mechanism to drive adoption of any complex medical test? 23andMe spouts high-minded rhetoric about their fierce desire to make us all healthier, “free” our genomes, and “democratize” genetics.

They may be completely sincere in those wishes (whatever they mean). But I don’t think it’s irrelevant to point out that they are a private company and that profit motives may have a tiny bit to do with their desire to see massive uptake of their tests and thus just might color their take on the utility of their product.

The sobering fact is that it’s going to be really hard to figure out how to interpret the roughly 4 million genomic variants that define each of us. That the FDA has simply pointed out that a particular direct-to-consumer company hasn’t done so is no surprise. The lack of validity for such testing at present is amply illustrated by the fact that when the same DNA sample is analyzed by different companies, wildly different results regarding, say, the risk of prostate cancer or heart disease are reported.

It’s going to take hard work and critical scrutiny to figure out the genome. This is (appropriately) high on the agenda of the National Institutes of Health and other biomedical research funders. We and others are intensely engaged in trying to determine just how genomic analysis can contribute to better medical care, both for the sick and the healthy.

Maybe crowd-sourcing, metadata analytics, and extreme “self-quantifying” have something to offer us in figuring it all out. Awesome! We welcome you to the arena. Let’s all work together to figure out the real promise of all these exciting new genomic opportunities.

All I (and the FDA) ask is that you provide good evidence that your claims are more than just promising ideas and wishful thinking before we push such testing into the realm of medical care.

If you want to read your horoscope each morning and base decisions on that information, be my guest.

But don’t ask me to pay for it.

The author is the Bryson Distinguished Professor of Genetics and Medicine at the University of North Carolina School of Medicine, director of the Clinical Adult and Cancer Genetics Services at the University of North Carolina, and editor-in-chief of Genetics in Medicine, the journal of the American College of Medical Genetics and Genomics. 

Copyright (c) 2020 The Cancer Letter Inc.