DANIEL CARPENTER, JEREMY GREENE, AND SUSAN MOFFITT
STANDARDIZATION OF products is a ubiquitous and ineluctable characteristic of modern capitalism. This activity—the development of and adherence to common guidelines for the production of goods or delivery of services—might be undertaken by private actors, the state, or social organizations, or even by combinations of these. In function, standardization deposits criteria of judgment for commodities that permit consumers, producers, and other audiences to perform differentiation and commensuration among products and services. By enabling differentiation, standardization assists consumers in making quality or risk distinctions among products, distinctions that can support optimization (quality comparisons that, conditional on price differences, may be sufficient for choice), status rankings, and assessments of risk. By enabling commensuration (developing a common metric of performance), standardization can assist producers and consumers alike in “showing up” to the same marketplace, one in which products are similar enough on other characteristics as to enable the functioning of an informative price system and indeed of innovation (Greene 2011; Greene and Kesselheim 2011).
Yet standardization does much more than ascribe difference and similarity to products. It also serves as a basis for the exclusion of some products from a market. In other words, standardization functions not only within markets but also across them. It can help to define boundaries between “commodity spaces,” some of which legitimately qualify (socially, legally, culturally) as a market, others of which do not. Standardization in this sense is linked to the regulation of marketplaces, as regulation often relies upon standardization. In the licensing of attorneys, accountants, physicians, and other professionals, the state and public-private associations decide which kinds of practices and credentials qualify for “membership” in the set of producers offering services.1 Yet licensing entities perform this work by relying explicitly on standards of “accountancy” and legal and medical practice (and “malpractice”).
In the realm of pharmaceuticals, standardization and screening are linked through the power of the state to limit production and market entry to legitimized and approved compounds. Students of FDA commonly trace these functions to two pieces of legislation. The Federal Food, Drug, and Cosmetic Act of 1938 created a premarket approval clearance requirement based upon a safety standard, leaving to the federal government the authority to determine market entry. Then in 1962, in the midst of the global health tragedy of thalidomide, Congress passed the Kefauver-Harris Amendments, which together converted the preclearance requirement to a positive approval standard and added an efficacy criterion to the safety hurdle of the 1938 legislation. Although FDA had been using efficacy judgments as dispositive standards in drug approvals since the 1950s, the Kefauver-Harris Amendments shifted the continuum of burden of proof toward drug companies and greatly regimented and standardized the process of drug development (Carpenter 2010; Carpenter and Sin 2007; Jackson 1970; Marks 1996).
Yet in significant ways, it was another institutional shift—in the Drug Efficacy Study Initiative (DESI)—that decisively reshaped modern drug regulation. In re-crafting drug regulation, it also refashioned modern pharmaceutical science, modern administrative procedure, and even the structure of the biopharmaceutical industry. DESI remade all these domains by participating in their standardization. DESI was a combination of the Drug Efficacy Study (DES) and its implementation. It was brought about by a congressional mandate in 1962, when “efficacy” considerations were officially added to the safety considerations under which FDA had been reviewing new drugs. Noting that from 1938 to 1962 FDA had been reviewing drugs officially on “safety” considerations only, Congress required FDA to review all drugs approved under the 1938–1962 “safety regime” and assess them for efficacy.2 Those that failed the exercise would be withdrawn. A large-scale “efficacy” review of more than 4,000 medications ensued.
This joint exercise in standardization and screening “unmade” a set of previously stable drug markets and brought into being a set of markets that had not previously existed. To be sure, DESI did many other things. It brought far-reaching changes in law, pharmacology, and politics. Legally, it established “summary judgment” as an administrative technique whereby federal agencies could issue broad rulings on a range of products—withdrawing them or ending their economic life as commodities—without having to take each and every case to an administrative law judge (Carpenter 2010:356). Methodologically, it was in DESI more than in the post-thalidomide intensity of 1962 and 1963 that the randomized, double-blind, placebo-controlled trial was established, as the problem of evaluating “old drugs” became a template for revising the standards for evaluating “new drugs.” Politically, DESI brought a rupture between some of the older-line pharmaceutical companies like Abbott, Parke-David, and Upjohn, and the newly academic industrials like Merck and Pfizer (Carpenter 2010).
Our purpose in reviewing the development of the Drug Efficacy Study and its implementation is threefold: (1) to demonstrate how the process of standardization emerged in FDA drug regulation and contributed to FDA’s power; (2) to highlight the ways in which DESI and FDA’s process of standardization evoke long-standing debate on the impact of product regulation; and (3) to suggest potential durable legacies DESI may have on pharmaceutical markets and public health.
I. REGULATING IN RETROSPECT
The Drug Efficacy Study was a by-product of a congressional attempt to regulate a population of “grandfathered” drugs. The Federal Food, Drug, and Cosmetic Act of 1938 prohibited the introduction into interstate commerce of any “new drug” whose safety had not been established by examination of the “Secretary,” namely the administrative apparatus of FDA. Thousands of new drug applications were submitted to the agency in the period from 1938 to 1962, and thousands were approved. Although it was widely acknowledged that efficacy considerations were also being used by FDA reviewers in the Bureau of Medicine (later called the Bureau of Drugs), there is little doubt that safety was both the principal consideration and the variable most observable in tests of animal and clinical pharmacology.
As a result, when Congress passed the Kefauver-Harris Drug Amendments of 1962 and added “effectiveness” officially to the criteria of new drug review, a large number of drugs for which no efficacy review had ever officially been conducted remained on the market. In the rulemaking that followed a year later—the Investigational New Drug rules of 1963—FDA elaborated an architecture of phased clinical experiment that would restructure clinical experiment and R&D worldwide in the ensuing decades. Yet it was not until the late 1960s that FDA would finally turn systematically to meet its statutory charge to examine the pre-1962 approved drugs. Commissioner James Goddard, knowing that his agency was burdened by the launch of a much more comprehensive new drug review system, decided that his agency could not address the problem using its existing administrative resources alone and contracted with the National Academy of Sciences and the National Research Council to conduct the review.
The NAS-NRC effort resulted in the convening of a Policy Advisory Committee of 27 members and 21 separate panels of therapeutically specific expertise. The executive director was Duke C. Trexler of the National Academy of Sciences. The taxonomy of expertise and review reflected much of the prevailing understanding of categorization in therapeutics at the time (see Table 16.1; Trexler 1966:12).
TABLE 16.1
Desi Panel Organization by Therapeutic Category
| 1. Allergy |
| 2. Anesthesia |
| 3. Anti-Infective (5 panels) |
| 4. Anti-Neoplastic |
| 5. Anti-Parasitic |
| 6. Cardiovascular |
| 7. Dentistry |
| 8. Dermatologic (2 panels) |
| 9. Diagnostic |
| 10. Endocrine Disturbances |
| 11. Fluid and Electrolyte Balance |
| 12. Gastroenterologic (2 panels) |
| 13. Hematologic |
| 14. Metabolic |
| 15. Neurologic |
| 16. Ophthalmologic |
| 17. Psychiatric |
| 18. Relief of Pain |
| 19. Reproductive System |
| 20. Respiratory Disturbances |
| 21. Rheumatic Diseases |
Steps to establish the panels began in 1966.3 In his March 31, 1966 memorandum to the director of the NAS-NRC Division of Medical Sciences, Dr. Keith Cannan, FDA Commissioner Dr. James L. Goddard sought NAS-NRC help to assess the efficacy of drugs approved between 1938 and 1962 (Rettig, Earley, and Merrill 1992:49–50). In his memorandum to Dr. Cannan (1966), Commissioner Goddard noted the scope and likely downstream importance of the DESI task. He wrote, “Although this is a one time task requiring evaluation of material somewhat different from that now obtained in current drug approval procedures, its long range significance exceeds that of all other drug activity pursued by the Food and Drug Administration.”
FDA’s public justification for contracting with the NAS-NRC was twofold: capacity and legitimacy. In its April 1966 press release, FDA stated that it sought NAS-NRC reviews because “The FDA itself does not have sufficient medical personnel to carry out a project of this scope” and because the NAS-NRC would be able to “tap the top medical and other scientific talent of the Nation” (FDA 1966). The task required reviewing and determining which drugs should be removed from the market based on inadequate evidence of efficacy. The scope and significance of the effort depended both on developing sufficient manpower and on securing legitimacy to make difficult and binding regulatory decisions. In his March 31, 1966 memo to Dr. Cannan, Commissioner Goddard continued, “Recommendations from the most expert sources are essential if this Administration is to suppress flagrant claims, eliminate worthless products and at the same time protect the physician’s therapeutic resources” (Goddard 1966).
FDA officially contracted with the NAS-NRC in June 1966, and the NAS-NRC convened its Policy Advisory Committee in July 1966 to establish procedures for the efficacy review process. The Policy Advisory Committee subsequently established 30 review committees (see Table 16.2) comprised of six members each and assigned drugs to committees for review (Rettig, Earley, and Merrill 1992:49–52). The NAS-NRC, not FDA, selected panel members and organized panels roughly along the basis of therapeutic class. Based on evidence from trials submitted by firms, medical literature, and “the experience and judgment of panel members,” panels assessed and reported drugs as effective, probably effective, possibly effective, or ineffective (Shorter 2008).4
TABLE 16.2
Chairmen of Thirty Desi Panels, June 1968
| Allergy: Dr. Bram Rose |
Anesthesia: Dr. Emmanuel Papper |
| Anti-Infective I: Dr. Heinz Eichenwald |
Anti-Infective II: Dr. William L. Hewitt |
| Anti-Infective III: Dr. William M. M. Kirby |
Anti-Infective IV: Dr. Calvin M. Kunin |
| Anti-Infective V: Dr. William B. Tucker |
Anti-Neoplastic: Dr. Emil Frei III |
| Anti-Parasitic: Dr. Harold Brown |
Cardiovascular I: Dr. Edward Freis |
| Cardiovascular II: Dr. Sol Sherry |
Dentistry: Dr. Edward Zegarelli |
| Dermatology I: Dr. Harvey Blank |
Dermatology II: Dr. Adolph Rostenberg, Jr. |
| Dermatology III: Dr. John A. Haserick |
Diagnostic Agents: Dr. Gilbert M. Mudge |
| Endocrine Disturbances: Dr. George Thorn |
Fluid-Electrolyte Balance: Dr. Maurice Strauss |
| Gastroenterology I: Dr. John T. Sessions |
Gastroenterology II: Dr. Henry J. Tumen |
| Hematology: Dr. William H. Crosby |
Metabolic: Dr. Don H. Nelson |
| Neurological: Dr. Melvis Yahr |
Ophthalmology: Dr. Irving H. Leopold |
| Psychiatric: Dr. Daniel X. Freedman |
Relief of Pain: Dr. Louis C. Lasagna |
| Reproductive System: Dr. Albert Segaloff |
Respiratory Disturbances: Dr. Sol Katz |
| Rheumatic: Dr. Charles Ragan |
Surgery: Dr. George D. Zuidema |
The subsequent NAS-NRC review process occurred behind tightly closed doors. The committees’ reviews were kept confidential, meetings were not publicly announced, and meeting locations were not publicly revealed (Bryan and Stern 1970:15).5 This degree of secrecy was consistent with FDA practice at the time—deliberations were kept private until regulatory decisions were issued, partly, FDA argued, to prevent market speculation and to encourage uninhibited deliberation.6 Commissioner Goddard recalled that panel chairmen were publicly announced but not panel members in part to “protect panel members from potential pressure from industry” (Goddard 1969).7 FDA also justified keeping the NAS-NRC reports confidential until the agency published orders in the Federal Register by arguing that releasing the reports earlier could induce panic, especially since FDA did not always adhere to NAS-NRC recommendations.8 Once FDA released information and orders on drug efficacy, the reports from the NAS-NRC also became publicly available (Edwards 1970).
The scope of the panels’ work was stunning both in terms of the vast swath of drugs reviewed and the outcome of the panels’ reviews. By the close of the decade, the NAS-NRC panels produced 2,824 reports on 4,349 different drug formulations from 237 firms. Commenting on the scope of DESI panels’ work, FDA Commissioner Charles C. Edwards noted, “At least 10,000 man hours were spent in formal decision-making” (Edwards 1970). Approximately two-thirds of the submitted drugs were evaluated by one NAS-NRC panel. The remaining one-third of the drugs were reviewed by multiple panels, ranging from two to fifteen different panels depending in part on whether the drugs were single-entity or combination drugs and in part on the number and nature of the efficacy claims made about the drugs (Edwards 1970). Estimates suggest that each panel reviewed, on average, 150 drugs (Rettig, Earley, and Merrill 1992:51). Commissioner Edwards noted that these reviews covered 80 percent “of the drugs in use” in 1970, that 10 percent of the reviews resulted in a panel determination that a drug was “ineffective,” and that many more were deemed only “possibly effective” (Edwards 1970).9
Each panel had six members plus a chair (see Table 16.2), with the idea that the members would be chosen as representative of the expertise then existing in particular domains of clinical pharmacology and treatment. Each approved new drug application (NDA) was reviewed by the appropriate panel, and the NAS-NRC kept a separate database of its review proceedings in which the NDAs were assigned “log numbers.” While the panels and FDA itself were criticized for proceeding too slowly, the Academy issued quarterly and annual reports on the drug-by-drug review.
The panels and the DES organization as a whole decided to assign pre-1962-approved NDAs into one of four categories,10 arrayed in publications and below in decreasing order of their perceived value and effectiveness:
Category A: Effective
Category B: Probably Effective
Category C: Possibly Effective
Category D: Ineffective
These categories, invented on the fly by an organization of experts who convened behind closed doors with little publicity, would be implemented as public policy. Most of the initial attention around the DES categories, when the listings were first announced in the Federal Register, focused on those drugs assigned to Category D—and specifically scheduled for removal from the marketplace.
II. UNMAKING THE MARKETS FOR SAFE DRUGS
FDA’s subsequent decisions to declare some drugs ineffective and to order them off the market were essentially binding. FDA brought the force of rulemaking to its withdrawal decisions by publishing orders in the Federal Register instead of using a formal adjudication hearing for each drug (Carpenter 2010:348).11 Receiving a designation of ineffective meant withdrawing the product from the market, with some classes of drugs incurring the lion’s share of the withdrawals. Some estimates suggest, for instance, that between 45 and 50 percent of psychopharmacologic drugs commonly used at the time were deemed ineffective and removed from the market (Shorter 2008). Fixed-combination antibiotics, such as Panalba, comprised a second class of drugs heavily represented in the category of drugs deemed “ineffective.” Firms such as Upjohn sued FDA for having received an “ineffective” designation and for the drug’s subsequent removal from the market, challenging the legality of the DESI process and ultimate decision. Critics alleged that the process, by forgoing a full administrative hearing for each drug adjudged and removed, denied firms their due process rights and disregarded clinicians’ expert judgment (Carpenter 2010:329–30).12 However, FDA’s decisions to order drugs from the market prevailed, which we discuss in more detail later.13
Table 16.3 summarizes the withdrawals by selected categories of the Anatomical Therapeutic Chemical (ATC) Classification system used by the World Health Organization. The table helps to clarify the clear concentration of withdrawals in selected therapeutic categories, namely antibiotics, central nervous system drugs (where mental health drugs such as tranquilizers and reserpine-based therapies are classified), vitamin and mineral supplements on the prescription market, and respiratory agents. Yet those pharmacologic classes spared from DESI’s enforcement are also notable, namely in various categories of oncology, beta-blockers in cardiology, many dermatologic therapies outside of wound and ulcer treatment, and antidiabetic and laxative drugs.
The stark differences across therapeutic categories bespeak both the variable perceptions of therapeutic efficacy in these cases, as well as the variable development of standards from which the NAS-NRC panels could draw in making efficacy determinations. The NAS-NRC panels did not conduct any new research but rather summarized samples of previous studies. Hence both the amount of research that had been conducted and the character of that research—in particular, whether it satisfied the “adequate and well-controlled trials” requirement of the 1962 law—were pivotal in shaping whether or not drugs were removed from legitimized pharmacological and medical practice in the United States. In related research, we have examined the possibility that these drug removals were associated with reduced mortality in the associated diseases, in part by removing subpar therapies from these areas, in part by creating “market space” (a sort of regulatory Schumpeterian creative destruction) for better therapies to emerge (Carpenter et al. unpublished). In other words, removing therapies from the market that failed to meet efficacy standards (perhaps destroying some firms’ market share) may have created room for alternative, innovative therapies to take their place.14
TABLE 16.3
Partial Summary of Desi Withdrawals by ATC Therapeutic Code
| ATC Code (Level 1 and Level 2) |
Categorization |
DESI Withdrawals Initiated, 1968 and After |
| A (Alimentary Tract and Metabolism) |
|
|
| A01 |
Stomatological Preparations |
17 |
| A04 |
Antiemetics and Antinauseants |
0 |
| A06 |
Laxatives |
0 |
| A10 |
Drugs Used in Diabetes |
0 |
| A11 or A12 |
Vitamins or Mineral Supplements |
18 |
| A15 |
Appetite Stimulants |
0 |
| C (Cardiovascular System) |
|
|
| C01 |
Cardiac Therapy |
7 |
| C02 |
Antihypertensives |
1 |
| C07 |
Beta Blocking Agents |
0 |
| D (Dermatologicals) |
|
|
| D03 |
Preparations for Treatment of Wounds and Ulcers |
13 |
| D07 |
Corticosteroids, Dermatological Preparations |
2 |
| D10 |
Anti-Acne Preparations |
0 |
| J (Antiinfectives for Systemic Use) |
|
|
| J01 |
Antibacterials for Systemic Use |
127 |
| J02 |
Antimyotics for Systemic Use |
7 |
| J07 |
Vaccines |
0 |
| L (Antineoplastic and Immunomodulating Agents) |
|
|
| L01 |
Antineoplastic Agents |
0 |
| L03 |
Immunostimulants |
0 |
| M (Musculo-Skeletal System) |
|
|
| M01 |
Antiinflammatory and Anti-Rheumatic Products |
12 |
| M03 |
Muscle Relaxants |
21 |
| N (Nervous System) |
|
|
| N05 |
Psyscholeptics |
22 |
| N06 |
Psychoanaleptics |
25 |
| R (Respiratory Agents) |
|
|
| R01 or R02 |
Nasal or Throat Preparations |
30 |
| R03 |
Drugs for Obstructive Airway Diseases |
10 |
| R05 |
Cough and Cold Preparations |
3 |
| V (Various) |
|
|
| V09 |
Diagnostic Radiopharmaceuticals |
2 |
| V10 |
Therapeutic Radiopharmaceuticals |
0 |
The ultimate legal authority for removing drugs from the marketplace was not in dispute—FDA had the authority to declare invalid the new drug application of any drug and any company. Yet the central legal debate concerned the administrative efficacy with which the administration could do so. The Upjohn Company, headquartered in Kalamazoo, Michigan, in particular contested the process by which drugs (and its fixed-combination antibiotic Panalba) were removed by summary judgment and publication of removal lists in the Federal Register, demanding instead that each and every drug withdrawn proceed through a separate administrative hearing. Still, the removals proceeded apace in the 1970s, especially after the U.S. Supreme Court legitimized FDA’s summary judgment procedure in a critical set of cases called the Hynson quartet: Weinberger v. Hynson, Westcott & Dunning; USV Pharmaceutical Corp v. Weinberger; Weinberger v. Bentex Pharmaceuticals, Inc.; and Ciba Corp v. Weinberger. In Weinberger v. Hynson, Westcott & Dunning (1973), the court found that firms’ right to a hearing was conditional on firms’ submission of “substantial evidence” of effectiveness. In terms of FDA standards established by 1970, substantial evidence of effectiveness meant producing two well-controlled studies that supported a drug’s efficacy claim (Carpenter 2010:356). Justice William O. Douglas’s opinion, writing for the Court in the Hynson case, affirmed that the 1962 amendments and subsequent FDA rules “express well-established principles of scientific investigation, in their reduction of the ‘substantial evidence’ standard to detailed guidelines for the protection of the public.” In other words, the “well-controlled studies” standard FDA promulgated in 1970 was deemed “appropriate.” The Hynson case marked the Court’s affirmation of FDA’s capacity to make efficacy judgments about drugs. It also paved the way for FDA’s subsequent removal of over-the-counter drugs, 300 of which FDA found “ineffective” (Carpenter 2010:356–57).
III. MAKING NEW MARKETS FOR OLD DRUGS
For those drugs found ineffective and then removed from the market, the DESI program was the first step in a bitterly contested dismantling of vast segments of the pharmaceutical market. For those drugs found to be effective, however, DESI added robust academic and federal claims of therapeutic value to which any subsequent manufacturer could refer when marketing their own generically named product.15 This was especially relevant as several pre-1962 DESI drugs were already off patent, and many more were going off patent each year.16
Several members of the Drug Efficacy Study were concerned that a DES Category A rating of “effective” applied to a drug by generic name might inappropriately be read as a federal certification that all versions of the same chemical compound, whether produced by brand-name manufacturers or small “fly-by-night” houses, were the same. Several of the members of the NAS-NRC committee, most notably the pharmacologist Alfred Gilman, had significant concerns that the DES judgments really applied only to the brand-name version of the drug evaluated and not to every pill, capsule, or table that might contain the same active ingredient.
The problem of generic equivalence, Duke Trexler recalled later (1969), was a question that “continually intruded into the work of the panels.” When the DES final report was issued in 1969, it was accompanied by a set of comments solicited from the nearly two hundred panel members regarding their reflections on the problem of therapeutic equivalence of generic and brand-name drugs. There was a great deal of concern, but little consensus. As the DES reviews of effective drugs were completed and phased into regulatory implementation, the members of the review panels continued to voice discomfort that DES standards of efficacy would be attached to generically named drug products (Bryan and Stern 1970).17
As Gilman had predicted, after DESI, FDA found itself in an unusual bind regarding effective drugs. To require all new copies of off-patent drugs to submit an extensive new drug application (NDA) merely to market a drug that had been approved and used for at least seventeen years seemed both economically wasteful and potentially unethical. Instead, FDA proposed in 1969 that manufacturers of newly off-patent drugs need only file an abbreviated new drug application (ANDA) that showed proof of chemical equivalence according to standard pharmacopeial compendia. In other words, manufacturers needed only to show that the active chemical was present at the dose specified without any contaminants. This regulatory action simultaneously lowered barriers to market entry for potential generic manufacturers and opened up a clearly sanctioned set of DESI-effective drugs as a new market space for would-be generic manufacturers.18
Take for example, the small company Bolar, which, by the late 1960s, was trying to transform itself into a legitimate manufacturer of generic drugs. By 1969, an intensified FDA inspection described Bolar as a generic drug manufacturer whose product line consisted nearly entirely of amphetamines and vitamin tablets and a smattering of minor tranquilizers (“Report of Intensified Inspection, 4 November 1968” 1969). Only two of their drug products—both versions of the vasodilator pentaerythritol tetranitrate—had even begun the process of formal FDA approval through the ANDA process. For much of the late 1960s, FDA waged a slow campaign against Bolar products as adulterated and misbranded goods (Warner 1968).
It was doubly ironic then that when FDA and NAS-NRC released the first list of ineffective drugs published by the DESI project, one of the drugs found ineffective was pentaerythritol tetranitrate. Faced with the prospect of losing permission to market the only drug it was selling with legitimate FDA approval, Bolar’s CEO, Robert Shulman, wrote a frustrated plea to the head of the Bureau on Drugs, Henry Simmons. Would FDA please let Schulman and Bolar know which drugs the DESI review had deemed “effective”—and therefore a viable market—so that Bolar might focus its efforts at regulatory approval in more productive channels? “We are a small manufacturer of generic drugs. In view of the recent NAS-NRC studies and the Food and Drug Administration actions based on these studies, we are very interested in re-evaluating our product line. Accordingly, we are asking for a little guidance along these lines. Would it be possible to obtain a list of drugs that have been evaluated and into which category (effective, ineffective, etc.) that they fall [sic]. If we knew this information, we would be able to delete the ineffective items and replace them with effective drugs” (Shulman 1970).
The evolution of Bolar’s correspondence with FDA over the 1960s documents the means by which new generic drug firms could be recruited into new regulatory regimes—especially when led by the carrot of the new markets potentiated by the promulgation of federal standards of drug efficacy. The standardization of efficacy performed by ANDA and the NAS-NRC review was a double-edged sword—the same action that cut away profitable but DESI-ineffective items such as pentaerythritol tetranitrate could also stabilize markets for generic production of DESI-effective drugs. Further correspondence between Bolar and FDA documents the transformation of this firm from a gray-market producer of illegitimate copies to an increasingly ethical (if still occasionally intransigent) regulated producer of legitimate copies.19 By 1973, FDA inspectors could note with some satisfaction that “as a result of the NAS/NRC studies they have discontinued manufacturing all amphetamine products which formerly constituted the bulk of their business” and Bolar had shifted to a product line made up predominantly of “pending or approved ANDAs filed as the result of DESI announcements.”20
IV. CONCLUSIONS: DESI AS AN EXERCISE IN MANIFOLD POWER WITH POTENTIAL LONG-TERM IMPACT AND AS A RESEARCH TOOL
Several features of the DESI exercise merit discussion here. For one, the DES implementation linked the gatekeeping power of the regulatory state with conceptual power of the regulatory state.21 While there was appreciable contestation over differentiations among drugs that remained on the market,22 it was the possibility that a drug would be removed from its profit-making role in the therapeutic marketplace that occasioned the greatest anxiety among American and international drug companies. So, too, the decision to delegate the decision making, review, and processing of the thousands of pre-1962 approvals to the National Academy of Sciences marked a critical late-twentieth-century precedent: the practice of FDA delegating to an advisory committee or to the National Academies the advisory task, but not the ultimate disposition, of potentially controversial tasks (Moffitt 2010). These legacies of DESI persist over a half-century after the NAS-NRC panels convened.
In addition, the DES’s development of methods for effectiveness ratings relied upon developments in clinical pharmacology. The deployment of these methods, and the panels’ judgment that well-controlled or randomized clinical studies were the best demonstration of effectiveness, marked both an important moment in statutory interpretation (the “well-controlled studies” plank of the 1962 amendments) as well as crucial standard-setting activities. The development of criteria for effectiveness judgments gave rise to newly cemented conceptions of experimental control, based upon reflections of how to separate treatment from control groups in clinical experiments—not only in the legitimation of randomization but also in the emergence of new conceptions of balanced treatment and control groups in both randomized and nonrandomized studies. In addition, the DES mounted a square attack on “therapeutic equivalence” as a concept separable from chemical equivalence and as a necessary component of any demonstration of effectiveness (Carpenter and Tobbell 2011).
For yet another feature, DESI provides unique opportunities to assess the long-term market and public health impacts of standardization. The scope of the National Academy of Sciences-National Research Council panels’ work, for one, was vast. Some estimates suggest that DESI “affected…the indications and the ingredients for 75 percent of the products on the marketplace at the time” (Crout 1998). The closed process that the panels followed and that FDA maintained enabled recommendations to manifest as an abrupt shock to the market, not something that firms could anticipate and adjust to preemptively. The recommendations the panels offered and the decisions FDA subsequently promulgated were ultimately binding.
What was the market impact of DESI? There is much to be gained by further empirical study of DESI and its consequences. Theoretically, DESI evokes competing predictions. From the standpoint of libertarian economic theory, the forced removal of a product is a supply restriction that should generate at least three consequences. First, equilibrium price in the market from which products are removed should rise. Second, because in an unregulated marketplace consumers (patients) will have been optimally matched to heterogeneous products, consumer value (and perhaps social welfare generally) should decline. Finally, because withdrawal and standards of entry dampen incentives to producers to invest in specific markets, innovation might also be retarded.23 These are of course posed as the simplest propositions from an account of efficient market operation, yet we note that all three of them have been stated in various ways in the economics, law and economics, and policy literature, as characterizations of FDA’s program of efficacy-based drug regulation (Epstein 2006; Gieringer 1985; Peltzman 1973, 1974).
Another set of arguments, based upon “screening” and “approval regulation” theories, suggests that when there are clear standards of efficacy, market removals may increase social welfare. First, behavioral dynamics in the marketplace for new therapeutics (placebo effects, inefficiencies in prescribing, poor physician and patient learning about quality, etc.) may lead patients to be matched suboptimally to therapies. If some of these therapies are of poor quality and can be judged more efficiently by an independent agent (say, on the basis of randomized, double-blind, placebo controlled trials), then the quality of treatment may rise because the worst products are excluded from the market. The exclusion property of screening regulation then carries two properties: (1) the “worst” tail of the product distribution is excluded or dampened, thereby generating higher average product quality and reduced expectations of the worst quality (which may matter especially when patients are risk averse); (2) a system by which the inclusion and exclusion decisions are based upon randomized-controlled trials will reduce consumer uncertainty about the quality of the drugs that remain—more data is available about them, diffused through medical journals, labels, news, and other features—which can generate an increased likelihood of optimal consumption. Second, by a reversal of the “market for lemons” dynamic by which bad quality products drive out good ones (the famous argument of George Akerlof), the removal of “lemons” from the marketplace may clear space for high-value innovations. The total amount of innovation may actually rise, but the quality-weighted amount of innovation under this scenario will certainly rise.
Debate over these fundamentally divergent views of regulation remains robust. Future empirical work on DESI’s long-term market and public health impact has the potential to offer new insight for these long-standing debates and to do so from the perspective of FDA’s ability to assert regulatory power on the basis of efficacy. Scholars and policy makers continue to assess ways in which FDA policies and procedures impact drug safety (IOM 2006), and our emerging analysis is uniquely positioned to draw attention to the neglected impact that regulating on the basis of efficacy can have on safety as well. Empirical evidence from FDA that addresses these competing views of regulatory impact, moreover, has the potential to generalize well beyond the FDA case and speak to the impact of regulation through standardization in other policy domains.
For yet another feature, DESI highlights the human impact of product regulation. The ultimate legacies of DESI involved transformations in pharmacology, the practice of experiment, administrative law, and health and safety regulation. Yet we conclude with the observation here that those legacies were tangibly human as well—chemicals coursing through bodies, processes of metabolism, acts and sensations of healing, and side effects. All these processes had occurred for decades and were now, with the removals of hundreds of compounds from the American marketplace, arrested and replaced by others. There are social science and economic implications of these transformations, to be sure, that we will examine in the coming years. It is plausible that the screening mechanism of DESI worked in such a way that approval regulation institutions were employed to rid a market of “lemons,” with plausibly beneficial effects on health in the years following the policy intervention. Yet to miss the truly human, emotive, and felt transformations would be also to miss a larger picture in which DESI remade individual and collective experiences of suffering and therapy.
The institutional story of DESI is clear. The joint screening and standardization exercise of DESI combined multiple forms of regulatory power as exercised by the American state: gatekeeping power (exercised retroactively) and conceptual power (patterns of standard and protocol that congealed in the stark rupture of existing patterns of claims making, therapy, and marketing). The legacies of DESI remain even more multidimensional and more durable than has been detected in academic inquiry.
A version of this chapter was presented at the STANDEX workshop (Standard Exchanges Programme; workshop international sur la standardisation en histoire de la santé et de la médicine), Strasbourg, France, December 7, 2012. Our thanks to Christian Bonah, Nils Kessel, Tricia Koenig, Carsten Timmerman, and others for helpful remarks and criticisms. Kyle Giddon provided helpful research assistance. Please direct all comments, criticisms, and corrections to dcarpenter@gov.harvard.edu.
1. In the simplest sense of microeconomic theory, standardization combined with licensure shapes the population of those agents (whose behavior) and products (whose marketing) determine the “supply function.”
2. For the 1938 Federal Food, Drug, and Cosmetic Act, see 21 USC § 301 et seq., as amended by Drug Amendments of 1962, Pub. L. No. 87-781.
3. Rettig, Earley, and Merrill (1992: 49–52, 56) offer a helpful summary of the development of DESI, upon which we draw in this section.
4. Rettig et al. (1992: 51) note that the ineffective designation could be limited to “ineffective as a fixed combination.” On the rigor and use of the two-trial standard in the DESI review process, see Carpenter (2010: 348, n. 61). For a rather ideologically tilted (but occasionally informative) critique of the quality of evidence used in the review process, see Shorter (2008).
5. On keeping NAS-NRC reports confidential until FDA decisions had been reached, see Bryan and Stern (1970). On closed meetings, see Rettig et al. (1992: 56).
6. For discussion of closure in FDA committees during this time period, see Moffitt (2014).
7. Goddard noted that the names of the panel members other than the chairmen were even withheld from FDA until after the panel review process was complete.
8. For further discussion on this point and related analysis of the lawsuit brought by the American Public Health Association and the National Council of Senior Citizens, see Carpenter (2010: 353).
9. Edwards noted, however, that designations of “ineffective” were common in the earliest reviews—40 percent of the first 900 rulings. He noted that FDA sought to address ineffective designations first. See p. 8 of speech.
10. In fact, the possibility of fixed combinations of other NDAs created a fifth category, “ineffective as a fixed combination.” For purposes here, we fold these into Category D, as this reflects the typological understanding in 1966 when Trexler began to explain the arrangement and procedures to various audiences. Yet the effectiveness of fixed combinations was a principal source of contention, not least because it raised the possibility that FDA would be judging the relative effectiveness of a fixed combination of previous NDAs to the NDAs themselves.
11. The published orders in the Federal Register were not part of conventional notice-and-comment rulemaking either. FDA did not solicit public comment on its list of withdrawal decisions.
12. See Carpenter (2010:329–30). On arguments about the denial of due process, Carpenter refers to Lasagna to Arthur Kallet, September 2, 1971; LSG. See also, Lasagna to Goodman, October 25, 1971; LSG. Lasagna also called for an “evidentiary hearing” and rehearsed the due-process claim in an April 1971 Wall Street Journal editorial, “FDA ‘Efficacy’ Rule: Does It Work?” WSJ, April 8, 1971, 12.
13. On Panalba’s removal from the market, court cases more generally, and reaction to DESI reports and FDA decisions, see Carpenter (2010:351–57).
14. In Capitalism, Socialism and Democracy, 2008, Schumpeter presents creative destruction as a core feature of capitalist economies in which innovation habitually supplants existing structures and processes, yielding (according to theory) more productive economies.
15. Marvin Seife, the first head of FDA’s Generic Drug Division, linked the birth of the abbreviated new drug application (ANDA) and the Generic Drug Division to the implementation of DESI in 1970. FGDAP, part 2, p. 4.
16. Material in this section adapted with permission from chapters 4 and 7 of Greene (2014).
17. For more on the divisions between these advisory panels on generic equivalence, see also Carpenter and Tobbell (2011).
18. It was unethical, they argued, to expose research subjects to placebo-controlled trials for a drug already known to work; it was economically wasteful to require a full NDA—which could reach thirty or forty volumes of paper for a single submission. For more on the role of FDA in market making and unmaking, see Carpenter (2010).
19. FDA was initially unwilling to play such a role, as noted by Albert Lavender to Robert Shulman, November 23, 1970, vol. 3, AF10 156, FDAAF.
20. “ANDA Inspection,” January 17, 19, 22, 23, and 26, 1973, vol. 4 AF10 156, FDAAF.
21. On these concepts, see Carpenter (2010), passim and especially Introduction and chapter 1.
22. For a sense of the stakes involved in one of the conditional designations —“Effective, But”—see “Agenda,” Guidelines Conference for Drugs Ranked as “Effective, But,” April 5, 1971; “The Drug Efficacy Study of the National Research Council’s Division of Medical Sciences, 1966–1969,” Series 1, NAS Archives. In a sign of the importance of the meeting, FDA Commissioner Charles C. Edwards and Bureau of Drugs Director Henry E. Simmons attended, as did PMA representatives (Joseph Stetler and John G. Adams, Director of Scientific Activities), and Joseph Pisani from the Proprietary Association. Former FDA Bureau of Medicine Director Ralph G. Smith reviewed “Categories of Problems Encountered in Evaluation of Claims Ranked as ‘Effective, But,’” and also presents “Typical Judgmental Processes.”
23. It is possible that a kind of market space was opened by withdrawal, but in the libertarian version of economic theory, this substitution effect is usually more than outweighed by the dampening of incentives for market entry.
REFERENCES
Archives of the Drug Efficacy Study, National Academy of Sciences, Washington, D.C.
Records of the U.S. Food and Drug Administration, Record Group 88, National Archives. “Agenda.” Guidelines Conference for Drugs Ranked as “Effective, But.” April 5, 1971.
“ANDA Inspection.” January 17, 19, 22, 23, and 26, 1973, vol. 4, AF10 156, FDAAF.
Bryan, Paul A. and Lawrence H. Stern. 1970. “The Drug Efficacy Study, 1962–1970.” FDA Papers.
Carpenter, Daniel. 2010. Reputation and Power: Organizational Image and Pharmaceutical Regulation at the FDA. Princeton, N.J.: Princeton University Press.
Carpenter, Daniel, Jeremy A. Greene, Susan Moffitt, and Jonathan Warsh. “Therapeutic Implications of Efficacy-Based Drug Withdrawals: The Drug Efficacy Study Implementation.” Unpublished manuscript.
Carpenter, Daniel and Gisela Sin. 2007. “Policy Tragedy and the Emergence of Regulation: The Food, Drug and Cosmetic Act of 1938.” Studies in American Political Development 21(2):149–80.
Carpenter, Daniel and Dominique Tobbell. 2011. “Bioequivalence: The Regulatory Career of a Pharmaceutical Concept.” Bulletin of the History of Medicine 85(1):93–131.
Ciba Corp. v. Weinberger, 412 U.S. 640 (1973).
Edwards, Charles C., MD, Commissioner of Food and Drugs. “Positive and Rational Drug Therapeutics.” Presented at the Symposium on Drugs—Hospital Pharmacists. September 26, 1970.
Epstein, Richard. 2006. Overdose: How Excessive Government Regulation Stifles Pharmaceutical Innovation. New Haven: Yale University Press.
Food and Drug Administration Press Release, May 1, 1966. 1992. In Food and Drug Administration Advisory Committees, ed. Richard A. Rettig, Laurence E. Earley, and Richard A. Merrill, 50. Washington, D.C.: National Academy Press.
Gieringer, Dale. 1985. “The Safety and Efficacy of New Drug Approval.” CATO Journal (Spring/Summer):177–201.
Goddard, James L., to Keith Cannan, March 31, 1966. 1992. “Efficacy Review of Pre-1962 Drugs.” In Food and Drug Administration Advisory Committees, ed. Richard A. Rettig, Laurence E. Earley, and Richard A. Merrill, 50. Washington, D.C.: National Academy Press.
Greene, Jeremy A. 2011. “What’s in a Name? Generics and the Persistence of the Pharmaceutical Brand in American Medicine.” Journal of the History of Medicine & Allied Sciences 66(4):425–67.
Greene, Jeremy A. 2014. Generic: The Unbranding of Modern Medicine. Baltimore: Johns Hopkins University Press.
Greene, Jeremy A., and Aaron S. Kesselheim. 2011. “Why Do the Same Drugs Look Different? Pills, Trade Dress, and Public Health.” New England Journal of Medicine 365(1):83–89.
Institute of Medicine (IOM). 2006. The Future of Drug Safety. Washington, D.C.: National Academy of Sciences.
Jackson, Charles O. 1970. Food and Drug Legislation in the New Deal. Princeton, N.J.: Princeton University Press.
Lavender, Albert to Robert Shulman, November 23, 1970, vol. 3, AF10 156, FDAAF.
Marks, Harry M. 1996. The Progress of Experiment. New York: Cambridge University Press.
Moffitt, Susan L. 2010. “Promoting Agency Reputation Through Public Advice: Advisory Committee Use in the FDA.” Journal of Politics 72(3):880–93.
Moffitt, Susan L. 2014. Making Policy Public: Participatory Bureaucracy in American Democracy. New York: Cambridge University Press.
Peltzman, Sam. 1973. “An Evaluation of Consumer Protection Legislation: The 1962 Drug Amendments.” Journal of Political Economy 81(5):1049–91.
Peltzman, Sam. 1974. Regulation of Pharmaceutical Innovation: The 1962 Amendments. Washington, D.C.: American Enterprise Institute.
Rettig, Richard A., Laurence E. Earley, and Richard A. Merrill (ed.). 1992. Food and Drug Administration Advisory Committees. Washington, D.C.: National Academy Press.
“Report of Intensified Inspection, 4 November 1968.” March 20, 1969, vol. 2, AF10 156, FDAAF.
Schumpeter, Joseph. 2008. Capitalism, Socialism, and Democracy. New York: Harper Perennial Modern Classics.
Shorter, Edward. 2008. “The Liberal State and the Rogue Agency: FDA’s Regulation of Drugs for Mood Disorders, 1950s–1970s.” Journal of International Psychiatry 31(2):126–35.
Shulman, Robert to Henry Simmons (FDA). October 30, 1970, vol. 3, AF10 156, FDAAF.
Trexler, Duke C. 1966. “The Drug Efficacy Study of the National Academy of Sciences–National Research Council,” Drug Information Bulletin (October-December):12.
Trexler, Duke C. to DES Panelists. January 24, 1969. Quoted in Drug Efficacy Study: A Report to the Commissioner of Food and Drugs. 1969. Washington, D.C: National Academy of Sciences.
USV Pharmaceuticals v. Weinberger, 412 U.S. 455 (1973).
Wall Street Journal editorial. 1971. “FDA ‘Efficacy’ Rule: Does It Work?” Wall Street Journal (April 8), 12.
Warner, Edward. “Violative Inspection—for Information Purposes,” May 14–16, 1968, vol. 2, AF10 156, FDAAF.
Weinberger v. Bentex Pharmaceuticals, Inc., 412 U.S. 645 (1973).
Weinberger v. Hynson, Westcott & Dunning, 412 U.S. 609 (1973).
