Cross-post from PointOfLaw.
Richard A. Epstein is a professor of law at NYU Law School, a Senior Fellow at the Hoover Institution, a Senior Lecturer at the University of Chicago and a visiting scholar with the Manhattan Institute’s Center for Legal Policy. His forthcoming book is “The Classical Liberal Constitution,” from Harvard University Press.
On November 22, 2013, the Food and Drug Administration flexed its regulatory muscle by sending a warning letter to a genetic-testing company that goes under the stylish name of 23andme. The object of FDA scorn was a diagnostic kit that the tech company, backed by among others Google and Johnson & Johnson, sold to customers for $99. The kit contained an all-purpose saliva-based test that could give customers information about some 240 genetic traits, which relate to a wide range of genetic traits and disease conditions. The FDA warning letter chastised 23andme in no uncertain terms for being noncooperative and nonresponsive over a five-year period in supplying information that the FDA wanted to evaluate its product as a Type III device under the Medical Devices Act.
Legal Regulation of 23andme
There is no doubt that the FDA is on solid legal ground. This case is not like the processes involved in Regenerative Sciences, LLC v. United States, where the FDA asserted that physicians’ use of certain stem-cell procedures for joint disease involved the use of a drug that required FDA approval before it could be approved for use. In an earlier essay for the Manhattan Institute, I argued that this classification was in fact both legally incorrect and socially mischievous. In this case, the legal arguments are not available to 23andme because the current definition of “medical devices” covers not only those devices intended for use on the human body, but also those used for the diagnosis of disease. The Type III classification means that this device has to receive premarket approval from the FDA, which in turn requires that it be shown to be safe and effective for its intended use. Getting approval under this standard is arduous business, because any such approval must be for each of the tests separately. 240 tests thus require that number of approvals. The costs are prohibitive, and the delay enormous.
The FDA Warning Letter is significant both for what it says and for what it does not say. On the former, it details all the various steps that the FDA has taken in order to help shepherd 23andme through the FDA’s processes, including the types of warning that the products should contain, and the various modifications that could be introduced in order to mitigate the risks of its use. It then notes that 23andme has done little to take advantage of the assistance offered to it. Indeed, worse, it has simply gone about its business selling the kits, without so much as a bow in the direction of the FDA.
The FDA Warning Letter and Its Possible Substitute
The FDA then gives a list of the dangers that could follow from this action, which consists essentially of false positives and false negatives with respect to certain of its key components. The gist of this indictment is contained in the following paragraph:
For instance, if the BRCA-related risk assessment for breast or ovarian cancer reports a false positive, 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. Assessments for drug responses carry the risks that patients relying on such tests may begin to self-manage their treatments through dose changes or even abandon certain therapies depending on the outcome of the assessment. For example, false genotype results for your warfarin drug response test could have significant unreasonable risk of illness, injury, or death to the patient due to thrombosis or bleeding events that occur from treatment with a drug at a dose that does not provide the appropriately calibrated anticoagulant effect. These risks are typically mitigated by International Normalized Ratio (INR) management under a physician’s care. The risk of serious injury or death is known to be high when patients are either non-compliant or not properly dosed; combined with the risk that a direct-to-consumer test result may be used by a patient to self-manage, serious concerns are raised if test results are not adequately understood by patients or if incorrect test results are reported.
The gist of the situation is that more information is a bad result because it could result in overconfident actions by patients who have the temerity to treat themselves without physician assistance, without ever asking whether some of these kits were sold to customers on the recommendation of their physicians. It is of course perfectly proper for any risk analysis to address the downside of certain course of action. But it is wholly incorrect for any risk analysis to ignore the benefits that could be derived from the 23andme kit. It takes little imagination to rewrite this letter to accentuate the positive. The FDA might try using my edited version:
For instance, if the BRCA-related risk assessment for breast or ovarian cancer reports a true positive, it could lead a patient to undergo much needed prophylactic surgery, chemoprevention, intensive screening, or other morbidity-reducing actions, while a true negative could result in the knowledge that no actual risk that may exist. Assessments for drug responses carry the benefit that patients relying on such tests may turn to a physician to control dose changes or even undertake certain therapies depending on the outcome of the assessment. For example, true genotype results for your warfarin drug response test could alleviate significant risks of illness, injury, or death to the patient due to thrombosis or bleeding events that occur by using doses appropriately calibrated anticoagulant effect. These risks are typically mitigated by International Normalized Ratio (INR) management under a physician’s care. The risk of serious injury or death is known to be reduced when patients are well treated and properly dosed; combined with the benefit that a direct-to-consumer test result may be used by a patient to seek medical advice for proper treatment and explanation of the underlying disease.
The revised letter is every bit as true, if not more so, than the FDA’s, rather different picture. In fact, a balanced appraisal, which the FDA never adopts, requires a blend of the two letters, by letting consumers know that the use of these tests could lead to bad results in some cases and good results in others. But in making this assessment it is critical to keep the dynamic element in mind. The forces of competition and the desire to increase sales will drive those companies who are in the market to improve their overall results. We can expect therefore over time that the correct results will be used. It is also the case that the FDA or anyone else could remind their customers that it is best to seek medical treatment for dangerous conditions because self-medication is a dangerous business–as if that were not known far and wide already.
Dealing With Two Kinds of Error
In light of these competing scenarios, the FDA should therefore have asked the question of the relative proportions of good and bad outcomes from the test. In dealing with this question, there is no reason whatsoever for it to stop with its list of hypothetical bad outcomes, without inquiring about the frequency and severity of bad outcomes. Some information on that question is of course available in this case, given that the kits in question have been marketed for over 5 years to about 475,000 customers, and it would be useful to know just how those patients fared over that time. Perhaps the FDA could devote some of its resources in this direction. That outcome is, however, not likely. In dealing with this issue, the FDA always wants to put the burden of proof on companies to show that their products are safe and effective, and will never look at field results that the companies present that tend to support that result. The proper procedure, however, is always to work in the opposite direction. The FDA should have to show by clear and convincing evidence that 23andme leads to the dangerous results that the FDA claims by surveying customers of the firm.
The reason for this proposed reallocation of the burden of proof is what it is in all these cases. Once the FDA keeps a product off the market, downstream actions by individual consumers and their physicians cannot undo its mistakes. But if the product is let on the market, patients and physicians need not use it if its risks outweigh its benefits. It would be a far more searing indictment of 23andme if customer complaints of false positives led physicians to discourage patients for relying on these treatments. But thus far, the physician complaints have all come at a distance from experts in the field who double down on the FDA warnings. Let them warn away, so long as they do not ban products, which given standard patterns of usage may do far more good than harm. All too often the FDA acts as though consumers are gullible and uninformed. It is time that Congress recognized that much of the danger to patient comes from the FDA. Its vaunted mission is “Protecting and Promoting Your Health.” I for one wish it would quit trying to protect mine and that of other people who would like to control their own lives.