Essay

*457 LOOSENING THE FDA'S DRUG CERTIFICATION MONOPOLY: IMPLICATIONS FOR TORT

law and consumer welfare

Michael I. Krauss [FNa]

Copyright © 1996 George Mason Law Journal Association; Michael I. Krauss

Table of Contents

Overview ............................................................ 458

  I.  Origins and Characteristics of the FDA's 'Certification Monopoly' ... 458

 II.  The Rationale for the FDA's Certification Monopoly: Analysis and

        Critique .......................................................... 463

III.  The Gap in the FDA's Certification Monopoly: Off-Label Prescriptions  469

IV.  Loosening the Certification Monopoly: From Off-Label Prescription to

        'Off-Label Manufacturing' ......................................... 474

      A. Critique #1: Without an FDA Right to Reject Market Decisions to

        Produce Drugs, Manufacturers Would Market 'Snake-Oil' Cures

        Harming Millions .................................................. 476

      B. Critique #2: Without Mandatory FDA Approval, Manufacturers Would

        Refuse to Produce Most Drugs for Devices Because of Their Fear of

        Products Liability ................................................ 477

      C. Critique #3: Doctors Would Never Prescribe Non-FDA Approved Drugs

        or Devices, for Fear of Malpractice Liability ..................... 478

      D. Critique #4: Private Firms Will Not Be Willing to Certify Drugs

        and Devices, Because, Unlike the FDA, They Will Be Liable for the

        Unexpected Side Effects of Drugs They Have 'Approved' ............. 479

      Conclusion .......................................................... 482

*458 OVERVIEW

The authority exercised by the Food and Drug Administration over new pharmaceuticals and devices was acquired in a rather haphazard fashion, without the benefit of any development of a theoretical basis for drug oversight. The result of this piecemeal accretion of federal powers has been a "certification monopoly" that has arguably cost thousands of American lives. Luckily, the FDA monopoly is an incomplete one: otherwise even more lives would have been sacrificed.

Building on the gap which presently exists in the FDA certification monopoly, this Essay develops and defends a proposal to loosen the monopoly, and describes the likely impact the implementation of the proposal would have on consumer welfare. The author argues that the Food and Drug Act should be modified so as to retain the positive characteristics of FDA testing while allowing market certifiers to compete with the federal agency. Tort law implications of this proposed modification of FDA powers are discussed in some detail.

I. ORIGINS AND CHARACTERISTICS OF THE FDA'S "CERTIFICATION MONOPOLY"

In the 1970s, a configuration of geopolitical factors resulted in a serious but temporary tightening of oil supplies. Several sets of sweeping federal fuel economy regulations were enacted. This expanded federal mandate led to the production and sale of relatively dangerous and unattractive automobiles, [FN1] and perhaps even to a deterioration of Americans' driving habits. [FN2] In a similar way, reactions to sensational occurrences have driven federal drug regulation. For ethical drugs, as for cars, the belief that markets (tempered by common law) can provide safe and effective products *459 has been a notable casualty of federal regulation.

The first statute regulating pharmaceutical products in the United States was the Pure Food and Drug Act of 1906 [FN3] ("1906 Act"), which established federal criminal offenses for adulteration and mislabeling of ingredients in food and drugs. The statute was apparently Congress' response to the "literary emergency" generated by Upton Sinclair's The Jungle, [FN4] a turn-of-the- century Soylent Green [FN5] portraying a world in which all medicines were deleterious, all milk tainted, and all sausages once human. Both before and after the 1906 Act's passage, ethical drugs were as freely available as newspapers (as they remain in many countries today [FN6]). No government approval was needed before drugs were manufactured or marketed, nor was a physician's prescription legally required, though doctors' advice was relied on by many and prescriptions accounted for almost one third of all drugs consumed. [FN7]

Economic analysis suggests that market forces will provide correct incentives for manufacturers to design and produce safe drugs, especially if tort remedies are available. [FN8] Patent medicines of dubious effectiveness were available on the market in 1906. It is worth noting that even at this early stage of its development, the common law provided recourse to those injured by such mislabeled or poorly manufactured drugs. [FN9] Of course, tort *460 law has little effect on insolvent tortfeasors, and, in any case, the national outcry generated by The Jungle provided an impetus for federal officials anxious to be seen "doing something" about drug "problems."

In 1912, amendments to the 1906 Act criminalized false and misleading claims of a medication's efficacy. Perhaps not surprisingly, criminal enforcement of the 1906 Act presented difficulties. "Snake-oil" producers are often "fly-by-night" and hard to locate; criminal intent is more difficult to prove than tortious negligence; etc. These factors curtailed enforcement [FN10] of the 1906 Act, even as amended.

Federal drug regulation remained insubstantial until a tragedy occurred in 1937. In attempting to formulate a liquid form of sulfanilamide (a sulfa drug) one Samuel E. Massengill, a Tennessean doing business as the Massengill Company, marketed a syrup that employed diethylene glycol as its solvent. One hundred seven children died after ingesting this poisonous concoction. [FN11] Though Massengill was found liable in tort for his gross negligence, [FN12] the temptation to provide additional "protection" to voters proved irresistible for a Roosevelt administration that was enjoying a huge expansion of federal powers following the "switch in time that saved nine." [FN13] President Roosevelt signed the Food, Drug and Cosmetic Act of 1938 ("1938 Act"), which for the first time required firms to submit New Drug Applications ("NDA's") before introducing pharmaceuticals into *461 interstate commerce. Each NDA enumerated the proposed uses of a drug and documented that tests "proved" the drug safe [FN14] at recommended dosages. NDA's were automatically approved after 60 days unless the FDA determined before that period that they did not establish the drug's safety. In addition, the 1938 Act contained a seemingly innocuous provision allowing discretionary exemptions from labeling requirements. The FDA subsequently interpreted this provision to greatly expand its authority, and physicians' income, by creating a category of "ethical drugs" that could henceforth be sold only by prescription. [FN15]

FDA safety evaluations obviously caused delays in the marketing of new drugs. But these lag times were relatively short: by the end of the 1950s a record number of new drugs was being marketed with mean regulatory delays of seven months. The 1938 Act is most renowned, however, for the delay it occasioned in the introduction into the United States of the German sedative Thalidomide in 1960. While an FDA reviewer investigated reports (which had in fact been published in Europe for years) that Thalidomide caused peripheral nerve damage, [FN16] news of a different side effect, deformities in the fetuses of pregnant users of the sedative, led to its withdrawal from markets worldwide in 1961.

The 1938 Act had prevented any Thalidomide tragedies in the United States, if only indirectly. [FN17] Congress nonetheless sprang into action again to dramatically expand drug regulatory powers. In 1962, a hastily drafted Food and Drug Act ("1962 Act") required that the FDA henceforth find drugs both safe and effective before they could be marketed. [FN18] Whereas under the 1938 Act NDA's were approved unless the FDA denied them, under the 1962 Act disapproval was the default position: for an NDA to be granted the FDA had to affirmatively conclude that a new drug was safe and effective. [FN19] Furthermore, while the 1938 Act allowed drug manufacturers *462 to test the safety of their own products (subject to documenting their tests for FDA review), the 1962 Act required that all human testing of new drugs be pre-cleared by the agency. The FDA acquired, additionally, the power to pre-approve all drug advertising and labels, and to regulate "good manufacturing practices."

As others have noted, [FN20] the combined effect of the 1938 and 1962 Acts was to fundamentally alter the nature of American pharmaceutical production and distribution. Whereas the 1906 Act had essentially sought to improve market efficiency by prohibiting false claims and information, [FN21] after 1962 a central regulatory authority, rather than the market choices of suppliers, physicians, and patients, determined which drugs were desirable or undesirable.

As might be expected, one effect of the FDA's post-1962 "certification monopoly" has been a much more lengthy "drug lag." By 1967, average post-test review times for an NDA had increased from seven to thirty months. [FN22] Largely because of agency involvement in testing, total pre-authorization development time for a new drug, which had averaged from four to six years before 1962, increased to over sixteen years by 1990. [FN23] Only one out of 5,000 new drugs now completes this process successfully, at an average manufacturer's cost of $200 million. [FN24] Economic theory predicts that these substantial increases in the cost of developing a drug for the United States market affect both the number of new drugs developed and the market price of developed drugs during their patent monopoly. [FN25] Commentators have amply documented both these forecasts. [FN26] At the margin, otherwise marketable drugs will now be unprofitable *463 ex ante to a manufacturer. [FN27] Also increasingly chronicled is the net cost in lost lives of not approving effective drugs quickly, as the following section relates.

II. THE RATIONALE FOR THE FDA'S CERTIFICATION MONOPOLY: ANALYSIS AND CRITIQUE

The 1938 and 1962 amendments came during "New Deal" and "Great Society" administrations, when confidence in government as a provider and allocator of goods and services was high. Welfare economists explain a preference for government regulation over competitive markets by the existence of "market failures." In the case of drugs, the "market failure" argument is typically that profit motive will lead manufacturers to produce too many drugs too quickly, that these drugs will consequently be poorly manufactured, [FN28] designed, [FN29] or marketed, [FN30] and that consumers will suffer losses that could be avoided if manufacturers were subject to government regulation. Several "consumer watchdog" groups embrace this market failure theory wholeheartedly. Dr. Sidney Wolfe of Public Citizen, for example, recently warned Congress not to privatize any FDA functions because drug manufacturers "mak(e) decisions based more on who fills their pocketbook than what is best for the public health." [FN31]

In formal economic theory, the "market failure" commonly attributed to pharmaceutical manufacturers' profit motive is ascribed to "imperfect information." The theoretical argument goes roughly as follows:

. Private choice of goods and services maximizes global satisfaction only if consumers perfectly know what they want and flawlessly calculate whether products offered for sale will give them what they want. These conditions imply an elimination of all risk.

. A risk-free universe does not exist, of course, and is especially absent where disease is concerned. Consumers of drugs, in fact, have doubts both about what ails them and about the effects of proposed remedies for that ailment.

.*464 . Producers of a remedy could provide useful information on the latter subject to consumers, but informational asymmetry encourages false or positively biased reports of drug efficacy.

. Furthermore, information is a "public good": once it is originated, it can be reproduced or broadcast. Since others can cheaply copy costly-to- produce information, the argument goes, the private market is likely to produce insufficient quantities of data about pharmaceuticals. Nor is it likely that disinterested third parties would provide correct evaluations of a drug's safety and effectiveness.

Thus, concludes the argument, government must fill the gap created by informational market failure. Government can do this by testing and certifying drugs, and by requiring exceptional quality before authorizing drug marketing.

There are many difficulties with the market failure rationale for the government's drug certification monopoly. Some can be usefully sketched now. First, the market failure argument explains little because it explains too much. Virtually all consumer goods and services are purchased by buyers with incomplete information. How long will that pair of panty hose last, and will its chemical formulation provoke allergic reactions? Will this champagne bottle explode in my face, scarring me for life? Is the doctor operating on me this morning recovering from last night's party? Will the law school I just applied to have an excellent reputation when I graduate three years from now? On all these issues, should I trust the information I receive from the not- disinterested department store, bottler, surgeon or admissions office? Consumer information is never perfect, and ignorance (i.e., risk) therefore inevitably influences all choices relevant to our well- being. But few consumers are confident enough in government's ability to discover and process information to argue for the replacement of market allocation and consumer choice by regulatory agencies merely because information is imperfect. [FN32]

Moreover, the major premise of the market failure argument, that information is a "public good" underproduced by markets and therefore efficiently provided only by government, is flawed. Competitive forces privatize information much more readily than is commonly believed by welfare economists. [FN33] Competitors with better products profit by pointing *465 out imperfections in their rivals' merchandise. Testing publications, from Consumers' Reports to Car and Driver, find markets for information, ranging from open evaluation to formal certification of products. [FN34] The value of these publications (and for that matter, the value of pharmaceutical manufacturers' products) derives in part from the reputation for quality they acquire. In fact, the "goodwill" value of brand names, trademarks, and other intellectual property, which both product manufacturers and consumer publications ferociously protect, [FN35] is largely a function of this reputational effect. Commercial activity (whether by automobile manufacturers or drug producers) consists not of one-shot events but rather of a series of repeat dealings; only myopic producers generate false product information if it is likely that the misinformation will become known. Admittedly, fly-by-night operators by definition don't feel reputational risk: but that is precisely why the market values their products less than it does those of a Proctor & Gamble or a Merck. That is why the used car sold through classified ads costs less than the same car sold by a reputable dealer. In the absence of any federal regulation, market prices for products already compensate ex ante for informational imperfections.

Of course, reputational markets are not perfect. Most notably, "agency problems" incite misbehavior. Employees may, for example, deceive drug (or car, or pantyhose, or surgery, or law school) "purchasers," because by the time the misinformation is discovered the employee will have comfortably retired, having personally reaped a profit while the corporation is left with the long-term loss. While agency problems do exist, in government [FN36] as well as in private industry, it is important to note that tort and contract law create powerful incentives for private firms [FN37] to monitor employee behavior. Thus, state common law instills confidence in information without the need for heavy-handed federal regulations.

*466 For example, incorrect information dispensed by manufacturers typically constitutes a "breach of express warranty," giving rise to strict liability suits by consumers injured as a consequence. [FN38] Poor manufacturing techniques, shoddy design work, inadequate warnings or careless preparation (whether by a too-hasty marketer of a drug with side-effects or by a partying surgeon) will result in tort or products liability. This liability is designed to fully compensate wrongfully harmed victims, thereby forcing wrongdoers to internalize the cost of their misbehavior. Some believe that tort law under-compensates plaintiffs, [FN39] while others feel that tort recovery is too generous. [FN40] Depending on who is right, tort should be made more plaintiff- or defendant-friendly. But that is a question of tort law, not of government regulation. [FN41] If the amount manufacturers (or service providers) must pay injured parties in tort suits is greater than the long term gain they derive from deceiving their customers, [FN42] the flow of information will be improved. [FN43] This will occur precisely in cases where it is appropriate to modify the manufacturer's (or service provider's) behavior. [FN44] Tort and products liability law can and do result in increased informational output from manufacturer to consumer precisely in those instances where such output might otherwise be insufficient. [FN45]

Finally, while the market failure argument supposes (incorrectly, as seen immediately above) that private corporations' incentives to produce information are insufficient, it also assumes that government incentives to inform consumers are not distorted. But this benign view of government has been challenged by political theorists ever since the Federalist Papers. Scholars who monitor civil servants know that "agency problems" are notoriously likely to reduce bureaucrats' efficiency. [FN46] "Public Choice" analysis teaches that informational asymmetries lead government agencies to be overly cautious in approving new drugs. The loved ones of those who have died because a non- approved (or non-developed) drug they never heard of might have saved them had it been developed and marketed are *467 typically unaware of these possible causal links. On the other hand, every crippled victim of a defective drug is visible to journalists, politicians, lawyers, and judges. Every new drug is, therefore, potentially another Thalidomide to bureaucrats. [FN47] As Sam Kazman has written:

From the FDA commissioner to the bureau heads to the individual NDA reviewers, the message is clear: if you approve a drug with unanticipated side effects, both you and the agency will face the heat of newspaper headlines, television coverage and congressional hearings. On the other hand, if the FDA insists on more and more data from a manufacturer, and finally approves a drug which should have been on the market months or years before, there is no such price to pay. Drug lag's victims and their families will hardly be complaining, because they won't know what hit them. . . . They only know that there is nothing their doctors can do for them. From the standpoint of . . . politics, they are invisible. [FN48]

The distortion resulting from FDA regulation is systematic. "Type 1" certification error is when a drug is approved that causes more harm than good, while "Type 2" certification error occurs when a drug is not approved, but that drug would have caused more good than harm. The above analysis indicates that market forces, tempered by contract and tort remedies, tend to minimize the sum of these two errors: anything less would reduce drug manufacturers' profits. Government regulation, in contrast, is acutely sensitive to Type 1 error but relatively unfeeling to Type 2 error.

Public choice theory has been borne out by practice, as is revealed by recent illustrations of the informational failure resulting from the FDA's certification monopoly:

. The cardiopump is standard equipment in ambulances in many countries, but it is not available in the United States. Initial tests in the United States in 1992 showed that, though not universally effective, use of a cardiopump led to a survival rate thirty-five percent greater among unconscious heart attack victims. Nevertheless, in 1993 the FDA ordered that use of the pump cease because unconscious patients could not give "informed consent" before the device was "tested" on them. The cardiopump may help many people: one estimate is that its use could save up to 7,000 Americans each year. [FN49]

FDA regulations require that the clinical studies necessary to obtain approval of a new drug or medical device be "randomized" and "concurrent," i.e., new devices or drugs must be blind-tested against placebos or FDA- approved products. This poses ethical problems if the new device or drug is so superior to existing technology that the "placebo" half of the double-blind pair is virtually condemned to death. However, the FDA tolerates this state of affairs and tends to intervene only when real victims are publicly identified. For example, the agency ordered hospitals to stop using specialized baby ventilators, which are irreplaceable in saving very sick infants because they provide uniquely tiny breaths of air, because hospitals refused to "blind-test" them and thereby condemn fifty percent of air-deprived infants. Dr. Martin Kessler of Georgetown University Medical *469 Center estimates that scores of babies died as a result. Subsequent to the FDA decision, protests from doctors who pointed to named infants' deaths were aired on the television show "20/20." Only then did the agency again allow use of the ventilators. [FN55]

The pre-market approval process for new medical devices, required since 1976, [FN56] averaged 840 days during the first six months of fiscal 1994, compared to 420 days in 1990. If manufacturers wish to improve devices already on the market before 1976, they must file a notification (or "510(k)") application with the agency. [FN57] Although the FDA is legally obliged to rule on 510(k) applications within ninety days, in fact the average review time apparently increased from ninety-eight days in 1990 to 213 days in 1994. [FN58] These expensive delays, obviously a product of the agency's zealous desire to control for Type 1 error, result, at the margin, in decisions not to introduce new or improved devices.

These are only some illustrations of FDA conduct; many others could have been provided. [FN59] They are not mere technical failings that could be addressed through more careful agency oversight. [FN60] Rather, they are the inevitable result of the institutional incentives of public regulators. The more complete the government's monopoly over drug and device certification, the more pernicious are these results. Fortunately, however, the FDA's monopoly has never been complete.

III. THE GAP IN THE FDA'S CERTIFICATION MONOPOLY: OFF-LABEL PRESCRIPTIONS

The FDA's certification monopoly for drugs and new devices has hastened the death of thousands of Americans. But, as discussed in Part I supra, the use of ethical drugs and devices in America pre-dates mandatory pre-certification rules. This chronology has left "gaps" in the FDA's *470 certification monopoly: pockets of market-oriented behavior that have not yet been stifled by the agency.

One important breach in the monopoly results from "off-label prescriptions." Physicians may not prescribe, nor may manufacturers market, drugs or devices that have not received FDA approval. [FN61] Courts have found both doctors and pharmaceutical companies "per se negligent" in tort when they defy the FDA's power in this area. [FN62] But if a drug or device is approved by the agency as safe and effective for one purpose, no FDA regulations prevent doctors from prescribing it for any other purpose. Thus, doctors, hospitals, and researchers may discover that a drug approved, labeled, and marketed for treatment of one kind of cancer is effective at treating other diseases. They write up these discoveries in refereed medical journals and other publications. [FN63] Physicians read or hear about these other uses and then prescribe the drug for the "off-label" uses. Under common law, writing an "off-label prescription" is not "per se" negligent. Tort law will, of course, sanction independently negligent prescriptions, for instance, if a doctor has misread the journal literature or has neglected to prescribe the correct dose.

The FDA has sporadically attempted to suppress off-label prescriptions. [FN64] In 1979, the agency attempted to prevent a physician from prescribing a drug, which had been approved as a medication for lead poisoning, for a patient with arteriosclerosis. The Fifth Circuit Court of Appeals rejected the FDA's claim that the physician had "misbranded" the drug, i.e., falsely claimed that the agency had approved of its proposed use, by prescribing or by advertising the new therapy. [FN65] FDA power has been more successfully invoked against manufacturers, however. Pharmaceutical companies had routinely publicized, to doctors and to the general public, *471 results of new research on their already-approved drugs, thereby alerting physicians (the vast majority of whom do not subscribe to, say, the New England Journal of Medicine) to new ways to help patients. In 1991, however, FDA Commissioner Kessler succeeded in prohibiting this "misbranding." [FN66]

These efforts to extend the FDA's certification monopoly to off-label prescriptions have cost lives and money in several ways. For example, many insurers, following Medicare's lead, [FN67] deny or limit drug cost reimbursement for off-label use. Nearly twenty-five percent of oncologists surveyed by the Government Accounting Office in 1991 reported varying from their preferred treatment because of anticipated denial of Medicare reimbursement. [FN68] But because Medicare reimbursement for costly hospital stays is fixed, hospitalization is used to circumvent the out-patient system. Sixty-two percent of the oncologists surveyed had admitted patients to hospitals solely to ensure reimbursement for drug therapy that would be denied on an out-patient basis. [FN69] However, recent and developing case law has limited state governments', and, to a lesser degree, private health insurers', discretion to rely exclusively on FDA approved use in determining coverage. [FN70]

Another illustration of health risks created by the FDA's crackdown on "misbranding" involves what is perhaps the most widely used "over-the- counter" drug. Influential studies suggest that the risk of heart attacks for males over fifty years old is cut in half by taking one aspirin every other day. According to an article in the New England Journal of Medicine, publicity for the aspirin treatment could save from 10,000 to 100,000 lives each year. [FN71] Yet the FDA, until recently, prohibited aspirin manufacturers *472 from making any claims about heart attacks. Since aspirin is no longer patented, no company has the incentive to spend the millions of dollars for clinical trials and other tests necessary to lead the agency to authorize publicity for this "new use." [FN72] Even when drug patents are effective, unless a manufacturer is willing to submit an NDA supplement for an unapproved use, the manufacturer has little incentive to generate information on the safety and effectiveness of the new use. Nor may manufacturers even distribute reprints of peer-reviewed journals or textbooks in which off-label uses are documented, even though drug producers are an efficient clearing house for such material. [FN73]

Despite its efforts the FDA has not stemmed the tide of off-label prescriptions. Estimates vary widely, but it appears that off-label uses presently account for from twenty to sixty percent of all prescriptions currently written. [FN74] A recent study by the Government Accounting Office found off-label prescribing to be the rule for many types of therapy. [FN75] Tort law recognizes this custom, declining to impose any requirement that physicians inform patients that their medication is being prescribed for a non- FDA-approved use. [FN76]

It is interesting to note that off-label prescriptions have proliferated despite tort law's implicit bias in favor of FDA-approved uses. A physician has a functional and limited tort immunity in the sense that he is never negligent merely for prescribing a drug for an FDA-approved purpose, although he still must transmit appropriate warnings and must check for possible drug allergies, etc. [FN77] If, on the other hand, a doctor prescribes a *473 drug for "off- label" use, there is neither immunity nor per se negligence. [FN78] Common law negligence principles apply, i.e., a doctor will be liable to an injured patient if the current state of research or customary medical practice impugns the off-label use. [FN79]

Obviously, many more people would die, and the clamor about FDA- induced "drug lag" would be more intense, if off-label prescriptions were suppressed. In an important sense, the medical profession has reduced public pressure on the FDA by privately performing research and issuing recommendations that manufacturers find too costly due to FDA regulations. However, private research and recommendations only deal with already-approved drugs and devices; despite the gap in its certification monopoly, the FDA retains legal control over the approval of a drug or device for its first approved use. The proliferation of beneficial off-label prescriptions demonstrates the illogical foundation of the agency's monopoly. Why should FDA approval be a sine qua non to "first use" when, once approved, the drug can be put to any number of other non-approved uses, subject only to the general rules of tort liability?

The theoretical and practical drawbacks of the status quo make it worthwhile to elaborate on a recommendation, initially sketched by the Competitive Enterprise Institute, [FN80] which if properly developed could preserve FDA evaluative functions while saving hundreds of thousands of lives. The proposal is to terminate the FDA's "first use" monopoly by permitting "off-label manufacturing," just as "off-label prescribing" is presently allowed. This proposal would resolve the agency's incentive asymmetries and result in FDA consideration of both "Type 1" and "Type 2" certification errors. The remainder of this Essay clarifies and defends this proposal.

*474 IV. LOOSENING THE CERTIFICATION MONOPOLY: FROM OFF-LABEL PRESCRIPTION TO "OFF-LABEL MANUFACTURING"

To recap this Essay's findings, first, FDA authority has expanded piecemeal in reaction to scattered tragic occurrences of drug misuse. The result of this expansion is that the FDA has metamorphosed from a guardian against fraud in marketing to a prime determinant of the production of ethical drugs and medical devices. Second, the FDA's present role cannot be justified by economic or political theory. The agency's skewed incentives, which contrast with those of manufacturers and physicians, have had the effect of aggravating illnesses and accelerating deaths. Third, the FDA has attempted to suppress what remains of market-driven selection of ethical drug use. Despite its efforts, a significant proportion of medications are used in non-FDA-approved ways. If this private defiance of the FDA had been eliminated, the perverse effects of the agency's disincentives would have been far more tragic.

What measures can be taken to reconcile economic and political theory with FDA practice? The outright abolition of the agency, or even its reinstatement as a "consumer protection" division devoted solely to policing and prosecuting fraudulent claims about drugs, are not viable solutions. A significant segment of the public is likely convinced that the agency should retain a role in the ex ante evaluation of drug safety and efficacy. On the other hand, the phenomenon of off-label prescriptions demonstrates that the FDA has never exercised hegemony over the evaluation of therapeutic uses of medication: practitioners, controlled by common law tort, do "end-runs" around the FDA with quite desirable results.

Once this is grasped, it becomes apparent that the FDA's function would not change if practitioners could recommend non-approved drugs as easily as they can now recommend non-approved uses of drugs. Just as any doctor may today legally write "off-label" prescriptions of drugs or devices, it could be lawful to manufacture drugs and devices "off-label." The FDA would, under this proposal, remain accessible as a state-funded monitor of safety and efficacy, as well as a policeman of consumer fraud.

This proposal does not create any windfall gain for manufacturers of dangerous drugs. By analogy to physicians writing "off-label prescriptions," manufacturers producing "off-label drugs" would not benefit from tort immunity. [FN81] Producers of "off-label drugs" would be liable for design defects [FN82] (i.e., marketing a drug that should never have been put on the *475 market) in these drugs, while producers of FDA-approved drugs would be immune to liability just as are doctors who write "on-label" prescriptions today. [FN83] All producers of drugs would, of course, still be liable for manufacturing defects and failures to warn (e.g., if a batch of the drug is contaminated, or if a label incorrectly indicates the appropriate dosage, etc.) just as physicians are liable for negligently mis-prescribing approved drugs under current law. But liability for a design defect would be precluded only if the FDA has approved the drug.

My recommendation is emphatically not a proposal to "emasculate" the FDA. In fact, in at least two areas it enhances the value of FDA certification. As just stated, it provides "on-label" manufacturers with a design defect "shield" they presently lack, courtesy of agency approval. [FN84] Second, while physicians are presently not compelled by statute or common law to inform patients that they are writing "off-label" prescriptions, [FN85] this proposal would assuage nervous consumers and increase FDA visibility by requiring doctors to obtain "informed consent" from patients if an ethical drug the doctor recommends is not approved by the agency.

*476 What effect would this modification of FDA authority have on the allocation and distribution of ethical drugs? The remainder of this Essay explores this question Socratically, by answering a series of hypothetical criticisms dealing with the interface between tort liability and decisions regarding drug production and use.

A. Critique #1: Without an FDA Right to Reject Market Decisions to Produce Drugs, Manufacturers Would Market "Snake-Oil" Cures Harming Millions

Readers who have followed the analysis in Part II of this Essay supra will recognize that this critique is unfounded. Manufacturers interested in maintaining a presence in the market and in enhancing their brand image have little incentive to market "snake oil." Excellent quality automobiles, croissants, and knives are produced despite the absence of mandatory pre- sale government approval of each design, even though defective versions of any of these products are extremely detrimental to our health.

In rebuttal, one might contend that government pre-approval of drugs is more important to the public than pre-certification of the wares of the neighborhood baker or cutler. But this contention fails to recognize that all certifications tend to become valuable over time. It may well be the FDA's structural obsession with avoiding Type 1 headlines that instills misplaced public confidence. If the public retains this confidence following implementation of this Essay's proposal, manufacturers will make the investment needed to obtain FDA approval, for they could then recoup the investment by exacting a premium in the marketplace. Over time, market valuation of FDA approval would stabilize (formally stated, a demand function for the agency's services would emerge). The label "FDA inspected and approved," if truly esteemed by the public, would add more than enough value to a drug or device to ensure the agency's survival even in the absence of a statutory monopoly. [FN86] The government could then engage in informed cost-*477 benefit analysis of the FDA's policies, including its fixation with Type 1 errors.

B. Critique #2: Without Mandatory FDA Approval, Manufacturers Would Refuse to Produce Most Drugs or Devices Because of Their Fear of Products Liability

Both economic theory and present-day practice suggest that fear of products liability does not stop manufacturers from producing goods. Manufacturers produce motorcycles and ladders despite the absence of pre- market government approval. They are held liable when their product is defectively designed or manufactured. [FN87] Conversely, FDA approval does not presently immunize manufacturers of pharmaceuticals (as distinguished from Class III medical devices [FN88]) from products liability. [FN89] This Essay's proposed immunization of manufacturers of FDA approved drugs (against design defect liability [FN90] only) might therefore increase availability of ethical drugs. [FN91] At the very least, it is obvious that the elimination of the agency's certification monopoly would not remove ethical drugs from the market.

As discussed in Part I supra, manufacturers now decline to develop certain drugs because pre-approval expenses render such development unprofitable. To accept Critique #2 is to imply that manufacturers will refuse to produce drugs with or without agency oversight. This is a non- sequitur. Clearly, the elimination of the agency's certification monopoly would not remove ethical drugs from the market.

*478 C. Critique #3: Doctors Would Never Prescribe Non-FDA Approved Drugs or Devices, for Fear of Malpractice Liability

The prevailing "gap" in the FDA's certification monopoly dulls this criticism. Fear of malpractice liability has not precluded off-label prescriptions, even though off-label prescriptions lack the tort protection that derives from FDA use-approval. [FN92] Doctors are clearly willing to be held to common law negligence standards [FN93] when prescribing drugs for off- label uses. They would arguably agree to be held to the same standards when prescribing off-label drugs. [FN94] Indeed, common law tort has evolved to the point that a physician's decision not to prescribe a drug for a widely accredited off-label use could well constitute malpractice. [FN95] In this way, again, the private sector has mitigated the effects of the FDA's disdain for "Type 2" error. Legislation permitting prescription of off-label drugs would further alleviate this problem. Note that these new drugs would likely fill very urgent medical needs: the relative tort protection provided by FDA certification means that physicians would likely need good reasons [FN96] for utilizing non-FDA approved drugs or devices if there exists FDA-certified alternative treatments.

One might wonder how doctors would learn of "off-label manufactured drugs." Currently, medical journals, which, incidentally, can be admitted into evidence at malpractice trials, describe off-label uses of FDA-approved drugs. The procedure for introducing non-approved drugs would likely be very similar. Journals and other disseminators of knowledge about the safety and efficacy of drug usage would find it in their interest to also produce information on new pharmaceuticals. The same reputational pressures that encourage the New England Journal of Medicine to refrain from publishing unscientific drivel on new drug uses will promote accuracy on the part of private reviewers of new drugs. [FN97]

*479 Private firms already provide such informational services in fields where the government has not seized a legal monopoly. Corporations regularly evaluate and certify design and production parameters, or "standards," [FN98] for goods and services as diverse as bricks and law schools. Reliance on these evaluations and certifications is an important element of malpractice defense for, in the above examples, architects and guidance counselors, even though the certifications lack mandatory legal authority. [FN99] Unlike government agencies such as the FDA, private firms cannot be politically constrained to accredit, or to refuse to accredit, new products or designs. [FN100] Furthermore, rival certifying firms might arise if a sufficient number of knowledgeable parties react negatively to an existing agency's inefficiency or corruption. [FN101] Competing agencies exist in many fields, [FN102] some of them well known to physicians. [FN103] But no firm may compete with a government monopolist.

D. Critique #4: Private Firms Will Not Be Willing to Certify Drugs and Devices, Because, Unlike the FDA, [FN104] They Will Be Liable for the Unexpected Side Effects of Drugs They Have "Approved"

This is another baseless critique. It begs two important questions. First, for reasonably foreseeable side effects (for example, teratogenic side effects that were discernable through testing), this critique's underlying assumption *480 is that such calamities will occur frequently as a result of lax private drug certification, i.e., that market incentives will not induce private certifying agencies to discover these problems. This Essay challenges that assumption. Markets, supplemented by tort law, provide strong incentives for certifying agencies to do their jobs well. Second, for unknowable future side effects, it is not clear why a non-negligent certifying agency should be liable for results it could not have reasonably prevented. Manufacturers of ethical drugs are not liable for such side effects today, [FN105] and there is no reason why a certifying agency should assume presently nonexistent liability in their place. [FN106]

More fundamentally, this Essay's critique of the "informational market failure" argument emphasizes that profits can be earned from certification even if there is a risk of liability. Some "certifiers," such as Underwriters' Laboratories, charge manufacturers directly for their services. Others, such as Consumers' Union (publisher of Consumers' Reports), earn money solely from the public sale of evaluative information. Both types of companies devote considerable resources to protecting their intellectual property, which enables them to privatize the information they produce. [FN107] Still other companies, such as Hearst Corp., publisher of Good Housekeeping (whose "Seal of Approval" is affixed to products) earn money from subscribers (purchasers of their evaluative information) and from manufacturers of the products they certify. Although payment of fees by manufacturers might blur the distinction between publicity and certification, protection of *481 trademark value and tort liability [FN108] furnish incentives to accurately assess products.

Insurance companies would likely have a comparative advantage in competing as quality certifiers. Liability insurance companies already have a vital incentive to accurately evaluate products whose quality is likely to affect their exposure. One of America's most venerable private certifiers, Underwriters' Laboratories ("UL"), is in fact a cooperative venture of insurance companies that presumably lowers total risk, thereby increasing its members' profit, [FN109] by assuring that only certified products are produced or used by insureds. [FN110] Since UL approval incurs the venture's liability to consumers injured by a defectively designed product that never should have been certified, [FN111] and since insurance companies are quintessentially solvent, UL approval can be seen as a kind of surety bond paid for by manufacturers, assuring purchasers [FN112] that a product design is effective. [FN113]

The UL product evaluation and certification model holds particular relevance for the ethical drug industry under this Essay's proposal to allow off-label manufacturing. Drug manufacturers could offer to indemnify (as if they had issued a surety bond) any physician found liable for prescribing a non-FDA- approved drug they produced. This is analogous to Ford Motor Company contractually "holding harmless" independent dealers for liability incurred as a result of selling a defective new (non-government-approved) Ford. In both cases, of course, manufacturers either self-insure or purchase *482 their own insurance against concomitant risks. [FN114] Drug manufacturers deciding to forego FDA certification could, by risking their capital ("self- insurance") or by purchasing liability insurance, produce quality guarantees for products in which they already have an interest and an expertise. Poor certification practices for ethical drugs would directly cut into the profits of these "insurers," who in fact are serving as a bond for the quality of their pharmaceuticals. [FN115] Good certification practices, accomplished through the establishment of reputable independent agencies (such as Underwriters' Laboratories), maximize insurers' and manufacturers' profits. Both tort liability and self-interested monitoring by insurers will hold certifying agencies in check.

CONCLUSION

The FDA's legal control over medicine has never been all-encompassing. The monopoly that the agency does possess was acquired episodically and accidentally. FDA power over drug certification is not justified by economic or political realities, and while its exercise has prevented some injuries it has also had deadly consequences by delaying or even preventing the introduction of new drugs. However, a gap in the FDA's power, resulting from the ad hoc process through which its jurisdiction was granted, has mitigated its potentially disastrous effects. This current process contains the kernel of a safer and more effective drug regulation policy. This Essay argues that the FDA's authority should be modified by allowing doctors to prescribe non-approved drugs and devices just as they today promote non-approved uses of drugs. Indeed, the FDA itself has recently presented a (very diluted) version of this proposal [FN116] in an effort to calm *483 hostile critics on both sides of the aisle [FN117] in Congress.

Although the relaxation of the agency's monopoly that this Essay proposes would not substantially alter the FDA's everyday work, it would have far- reaching and beneficial effects. It might lead to the establishment of a drug certification industry composed of different types of competitive firms. Non- profits (such as Consumers' Union), for-profits (such as Hearst Publications) and insurers' cooperatives (such as Underwriters' Laboratories) could be expected to enter the drug certification industry. To compete successfully, these firms would have to convince producers, consumers, and doctors that their evaluation processes are high quality. Manufacturers and physicians would make and use privately certified products under these new ground rules, as efficient risk bearers would assume risks of liability via insurance and bonding mechanisms.

Of course, no private evaluation industry will be viable if consumers balk at non-FDA certified drugs. Since this proposal does not eliminate the agency, consumers who remain troubled by non-FDA-approved products will be able to insist on FDA certification. If consumers remain unconvinced that the market has provided efficient substitutes for the agency, drug manufacturers will continue to submit their products to the FDA for pre-market approval. On the other hand, if the FDA becomes perceived as an inefficient agency focused exclusively on the elimination of Type 1 testing error (at the cost of an inefficiently high number of Type 2 errors), the agency's "market share" of certifications will decline. Sick people will increasingly opt to consume privately certified ethical drugs. If this happens, an FDA that wishes to prosper sans its certification monopoly can be expected to put patients' best interests at heart by balancing Type 1 and Type 2 errors instead of responding to bureaucratic pressures at the cost of thousands of lives. At that point consumers will finally benefit from an FDA which puts their interests first.

END OF DOCUMENT