Home About RSS

Law and Biosciences Blog
Law and Biosciences Blog

The FDA and Genetic Tests: A Modest Proposal to Require Prescriptions

Next week, on July 19 and 20, 2010, the FDA will hold a public meeting to consider what to do about laboratory developed tests (LDTs), particularly those that look for genetic information. I believe that, for now, the FDA should classify health-related genetic tests as “restricted devices” and allow their use only with a physician’s prescription.

It worries me that genetic tests are currently almost universally sold without having been proven either safe or effective, but what to do about this is complicated. Many medically well accepted tests involving diseases too rare to support substantial clinical trials, sometimes scientifically and usually economically. On the other hand, some tests are sold with absolutely no scientific support.

The FDA could easily take a simple and somewhat useful step by requiring a doctor’s prescription for any genetic test. The Federal Food, Drug, and Cosmetics Act give the FDA jurisdiction over devices that are “intended for use in the diagnosis of disease or other conditions, or in the cure, mitigation, treatment, or prevention of disease, in human or other animals.” The FDA has the power to classify a covered device as “restricted.” The FDA website explains

    “Under the provision of Section 520(e) of the [1976 Medical Device] Amendments, the FDA is authorized to restrict the sale, distribution, or use of a device if there cannot otherwise be reasonable assurance of its safety and effectiveness. A restricted device can only be sold on oral or written authorization by a licensed practitioner or under conditions specified by regulation. Devices such as cardiac pacemakers and heart valves, for example, require a practitioner’s authorization.”


The hundreds, or thousands, of well-established tests for genetic diseases would continue to be used as they are today – after a request from a physician. Less established tests could still be purchased, but only if a physician, in a legitimate doctor-patient relationship, ordered it. The existence of such a doctor-patient relationship brings with it the possibility of malpractice liability, which should deter some reckless prescribing and should lead physicians, in appropriate cases, to provide the necessary counseling about the test results.

The main effect of this proposal would be to eliminate direct-to-consumer sales of genetic tests with health consequences. Consumer genomics companies, whether specializing in one particular condition and test or offering a broad range of information based on SNP chips, could only fill requests that came with a doctor’s prescription. (This is similar to the New York State statute on clinical tests, which has blocked most consumer genomics companies from operating in that state.) The firms could still market genetic tests without a prescription for non-health-related uses – genealogy, paternity, earwax type – but not for cancer risk, Alzheimer’s disease risk, or other medical uses. The prescription requirement would make genetic tests for health information more expensive. To the several hundred dollars now charged for a DTC test, the consumer (now patient) would have to add the cost of a doctor visit or two.

This is not a perfect solution. A doctor’s prescription, even if provided in the course of a legitimate doctor-patient relationship, does not guarantee that the doctor has the knowledge, ability, or time to help the patient make a good decision, about whether to be tested and about the meaning of any test results. But the chances that the patient will get that information from a physician are much greater than the chances of being well informed only through the Internet. (And a prescription requirement certainly does not rule out the possibility of additional information being provided to doctor and patient by the testing companies, through the Internet or otherwise.)

There will be other details to be worked out. It may be that professionals other than physicians should be allowed to prescribe a genetic test – clinical geneticists or perhaps genetic counselors. And there will be some legal complications. What tests are “health-related?” When is a test “intended for use” for health purposes? But the FDA already has experience with these kinds of boundary problems. More puzzling is the possibility of a firm providing simply the SNP chip (or, eventually, whole genome sequencing) results, without any interpretative information other than a reference to OMIM or other on-line resources. Would that be “intended for use” for the covered purposes?

But this also need not be a final solution. Those questions can be answered later. For the time being, this proposal would give at least some increased protection against the users of genetic tests making poorly informed and possibly damaging decisions based on those tests.

The Genome Libertarians, who vow that the government will have to pry their genetic information from their cold, dead hands, will be outraged. The 0.1 percent (or less) of the population that can be expected to understand and deal well with their genetic information without professional intermediation may have a legitimate grievance. (The other 0.9 percent of the population that believes, falsely, that they can understand their own genetic information will have a less legitimate grievance.) But the 99 percent of the population that is not made up of genetic Masters of the Universe will be better off – which seems to me a good regulatory result.

4 Responses to “The FDA and Genetic Tests: A Modest Proposal to Require Prescriptions”

  1. Dirk says:

    Well, I certainly disagree with your proposal.

    “It worries me that genetic tests are currently almost universally sold without having been proven either safe or effective,”

    “safe”: Genetic tests provide information. Information, i.e., knowledge, cannot cause bodily harm.

    “effective”: Information/knowledge is not a treatment for a disease.

    Thus, both “safe” and “effective” are inappropriate terms to apply to genetic information/knowledge.

    You may instead argue that the usefulness of genetic test results is limited by technical measurement errors, and the derived genetic disease-risk predictions may not consider all relevant genetic information (not all relevant SNPs considered, copy-number variations unknown, unknown which allele is the dominant/recessive one, etc.). The ‘technical measurement errors’ can be minimized by the gene-chip developing companies and/or DTC companies themselves: direct sequencing to compare with the gene-chip results. The ‘limitations of disease-risk predictions’ are addressed by disclaimers within DTC companies’ agreements. It is very likely that serious DTC genetic testing companies will strive to incorporate the state of the art in their disease-risk predictions.

    “but only if a physician, in a legitimate doctor-patient relationship, ordered it”

    The health care system in the USA is organized to treat existing medical problems and to provide a high income/profit to physicians, hospitals, insurance companies, and those manufacturing the treatments and diagnostic tests.

    In contrast, DTC genetic tests serve to inform the individual citizen about his/her personal genetic information [free flow of information = freedom of speech = First Amendment] and may allow him/her to enact life-style changes so as to prevent/delay the on-set of a disease.

    In summary, I believe governmental regulation of DTC genetic testing companies is neither needed nor appropriate.

  2. John says:

    I agree with Dirk.

    Neither “safe” nor “effective” are applicable to genetic test results. Genetic test results are simply information.

    The information provided by a genetic test is a physical characteristic of the individual, like eye color, height, or weight. Genetic information just happens to be invisible, like blood pressure or blood sugar or the blood level of hCG (the hormone arising from pregnancy that results in a positive pregnancy test). I can look in the mirror and obtain my eye color. I can go to the pharmacy and obtain my blood pressure (both systolic and diastolic). I can purchase a glucometer and obtain my blood sugar measurement. I can purchase a pregnancy test kit and obtain my pregnancy status. All without a note from my doctor. Why should I not be able to obtain my genetic composition?

    “I believe that, for now, the FDA should classify health-related genetic tests as ‘restricted devices’.” Genetic tests are not devices. The blood pressure measuring apparatus is a device, the glucometer is a device, the pregnancy test kit is a device, and the instrument or chip that determines the genetic information is a device. But the information generated by the blood pressure apparatus, the glucometer, the pregnancy test kit, or the genetic test are not devices. Information is information.

    “Genetic tests with health consequences.” The health consequences that can result from a genetic test are not different from the health consequences that can result from a determination of blood pressure, blood glucose, or pregnancy status. All can result in worry or concern. However, all other significant consequences require the intervention of a health care practitioner who is presumably competent to provide appropriate additional information and help with decision making and who would be responsible for the intervention.

    “Many medically well accepted tests involving diseases too rare to support substantial clinical trials, sometimes scientifically and usually economically. On the other hand, some tests are sold with absolutely no scientific support.” Neither of these statements is applicable to genetic testing. Clinical trials usually are applicable to interventions; genetic tests are not interventions, they simply provide information. The scientific foundation for virtually all currently available genetic tests is strong and unquestioned. Accuracy, sensitivity, specificity, etc., data are published and set the standard for biologic testing.

    We are deemed competent to obtain and interpret our blood pressures, our blood glucose levels, and our pregnancy status – all without a note from our doctor. We are deemed competent to invest our retirement savings, to purchase real property, and to incur debt – all without a note from our financial advisor. We are deemed competent to determine our alcohol intake and, after intake, our ability to drive a motor vehicle – all without a note from anyone. Our investment decisions can result in destitution. Our decisions concerning alcohol can result in destitution, disability, and death.

    Mr. Greeley, please explain why the genetic test is so dangerous. I just don’t understand.

  3. Hank Greely says:

    If information can be neither safe nor effective – if it is powerless – then why bother to write? That part of your post is just silly. Of course information can be used safely or unsafely, effectively or ineffectively. It is not safe or effective apart from how it is used, but the same is true of plutonium.

    You may have a plausible, though very broad, First Amendment argument, but note how much of the structure of regulation is would invalidate – particularly all of the laws against unauthorized practice of medicine and all of the FDA’s regulations.

    As a consequentialist matter, is the world better off if all genetic tests have to go through trained professionals or not? I think the former captures more benefits and avoids more harms than the latter, by increasing the chances that customers/patients will use genetic information well. Is your disagreement with that assessment of the costs and benefits or with the consequentialist framework?

  4. Hank Greely says:

    My earlier reply was to Dirk; I think my comment and John’s crossed in cyberspace. Much of what I said to Dirk’s comment applies to John’s, but here’s a bit more.

    First, the FDA made a decision that home pregnancy tests, home HIV testing, and home blood pressure monitors were safe and effctiveness to be used without a doctor’s prescription. It was similar to the process the FDA went through in deciding that claritin was safe and effective enough to be sold over the counter. It has not made that assessment with respect to genetic tests.

    Second, under the Federal Food, Drug, and Cosmetics Act, a test IS a device. It doesn’t matter whether it is a mechanical product. The test results are not a device, but the test that produces them are.

    Devices often go through clinical trials if the FDA concludes they require a PMA – a “premarket approval.” There’s no inherent reason why someone could not take 100 people, give a random 50 of them a genetic test and the others not, and then look for various clinical outcomes down the line. (Hard to do that as a blinded test, I’ll matter.)

    I would say the “DNA dating” test has (almost) no scientific support. The vast majority of the DTC tests aren’t (close to) frauds, but neither are they as strong as, say, testing for sickle cell or cystic fibrosis mutations. The scientific basis for the SNP-chip risk assessments used by many of the consumer genomics companies is neither “strong” nor “unquestioned.” Note that the companies will even disagree about which SNPs are strong to use and how they should be interpreted.

    And, yes, our government system does respond differently to different risks. Yet your investment decisions, and your borrowing decisions, are affected by substantial government regulation, as are your retirement planning decisions. Driving is dangerous, but we do regulate driving. And, for reasons both utilitarian and historical (and perhaps in part guild protection/income enhancement for the medical system), we heavily regulate health care.

    As to why the genetic tests can be dangerous,consider two scenarios.

    First, a women gets tested for 3 “common” mutations in BRCA 1 and 2 (“common” is relative to other BRCA 1 and 2 mutations; in fact only about 1 woman in several thousand will have one of these mutations). She is told that she does not have the mutations. She decides as a result that she is safe from breast cancer risks and so no longer gets mammograms. She develops breast cancer that is spotted too late and dies. Or she finds that she is at higher risk, she gets mammograms regularly, but she doesn’t know/realize that she is also at higher ovarian cancer risk and she does nothing about that and dies of ovarian cancer. In both of these cases it is her misuse of the genetic test information that causes harm, but it is misuse that consultation with a professional might well have prevented.

    Second, consider a 55 year old person who is told that he has one APOE4 allele and so is a double to triple the normal risk of Alzheimer’s disease. He has been bothered lately by a sense that his memory is not as sharp as it used to be. He obsesses over this test information, concludes he is in the early stages of AD, and commits suicide. A conversation with a doctor should have prevented that overreaction.

    How common are these scenarios? Not very common, I certainly hope, and they may be non-existent. But they aren’t implausible. What’s the countervailing gain in allowing DTC testing without professional intemediation? It’s a little cheaper, meaning more people will get it and the companies will make more money. How do you balance those outcomes?

Leave a Reply