`
`Koios Pharmaceuticals v. Medac
`
`Medac Exhibit 2054
`
`IPR2016-01370
`
`Page 00001
`
`Medac Exhibit 2054
`Koios Pharmaceuticals v. Medac
`IPR2016-01370
`Page 00001
`
`

`

`THE OXFORD HANDBOOK OF
`
`THE ECONOMICS
`
`OF THE
`
`BIOPHARMACEUTICAL
`
`INDUSTRY
`
`Page 00002
`
`Page 00002
`
`

`

`
`
`CONS ULTING EDITORS
`
`MICHAEL SZENBERG
`
`LUBIN SCHOOL OF BUSINESS, PACE UNIVERSITY
`
`LALL RAMRATTAN
`
`UNIVERSITY OF CALIFORNIA, BERKELEY EXTENSION
`
`Page 00003
`
`Page 00003
`
`

`

`THE OXFORD HANDBOOK OF
`
`THE ECONOMICS
`
`OF THE
`
`BIOPHARMACEUTICAL
`
`INDUSTRY
`
`Edited by
`
`PATRICIA M. DANZON
`
`AND
`
`SEAN NICHOLSON
`
`Page 00004
`
`Page 00004
`
`

`

`OXFORD
`UNIVERSITY PRESS
`
`Oxford University Press, Inc., publishes works lhat further
`Oxford University's objective of excellence
`in research, scholarship, and education.
`Oxford New York
`
`Auckland Cape Town Dar es Salaam lIong Kong Karachi
`Kuala Lumpur Madrid Melbourne Mexico City Nairobi
`New Delhi Shanghai Taipei Toronto
`With offices in
`
`Argentina Austria Brazil Chile Czech Republic France Greece
`Guatemala Hungary
`Italy
`Iapan Poland Portugal Singapore
`South Korea Switzerland Thailand Turkey Ukraine Vietnam
`
`Copyright © 2012 by Oxford University Press
`
`Published by Oxford University Press, Inc.
`193 Madison Avenue, New York, New York 10016
`www.0up.com
`
`Oxford is a registered trademark of Oxford University Press
`
`All rights reserved. No part of this publication may be reproduced,
`stored in a retrieval system, or transmitted, in any form or by anyineans,
`electronic, mechanical, photocopying, recording. or otherwise.
`without the prior permission of Oxford University Press.
`
`Library of Congre5s Cataloging—in-Publication Dala
`The Oxford handbook of the economics of the biopharmaceulica] industry 1‘
`edited by Patricia M. Danzon and Sean Nicholscm.
`p. cm.
`Includes bibliographical references and indexes.
`lSBN 978—0-19-974299-8 (cloth : alk. paper)
`1. Pharmaceutical industry.
`2. Biopharmaceutics-Economic aspects.
`I. Danzon, Patricia Munch, 1946—
`II. Nicholson, Sean.
`III. ‘I‘ille: Economies of the biopharmaceutical industry.
`HD9665.5.094 2012
`338.4’76157-dc23
`2011040815
`
`1 3 5 7 9 3 6 4 2
`Printed in the United States of A merica
`on acid—free paper
`
`Page 00005
`
`Page 00005
`
`

`

`CONTENTS
`
`Contributors
`
`1.
`
`Introduction
`
`PATRICIA M. DANZON AND SEAN NICHOLSON
`
`PART I PHARMACEUTICAL INNOVATION
`
`2. R&D Costs and Returns to New Drug Development:
`A Review of the Evidence
`
`JOSEPH A; DIMASI AND HENRY G. GRABOWSKI
`
`3. Financing Research and Development
`SEAN NICHOLSON
`
`4. Cost of Capital for Pharmaceutical, Biotechnology,
`and Medical Device Firms
`SCOTT E. HARRINGTON
`
`5. The Regulation of Medical Products
`ANUP MALANI AND TOMAS PHILIPSON
`
`6. Incentives to Innovate
`DARIUS LAKDAWALLA AND NEERA] SOOD
`
`7. Patents and Regulatory Exclusivity
`REBECCA S. EISENBERG
`
`vii
`
`1
`
`21
`
`47
`
`75
`
`'
`
`100
`
`143
`
`167
`
`PART II THE MARKET FOR PHARMACEUTICALS
`
`8. Pricing and Reimbursement in US Pharmaceutical Markets
`ERNST R. BERNDT AND JOSEPH P. NEWHOUSE
`
`9. Regulation of Price and Reimbursement for Pharmaceuticals
`PATRICIA M. DANZON
`
`201
`
`266
`
`Page 00006
`
`Page 00006
`
`

`

`
`
`vi
`
`10.
`
`IL
`
`11
`
`13.
`
`14.
`
`15.
`
`16.
`
`17.
`
`18.
`
`Drugs and Vaccines for Developing Countries
`ADRIAN Towss, ERIC KEUFFEL, HANNAH E. KETTLER,
`AND DAVID B. RIDLEY
`
`Insurance and Drug Spending
`MARK V. PAULY
`
`Consumer Demand and Health Effects of Cost-Sharing
`DANA P. GOLDMAN AND GEOFFREY F. JOYCE
`
`Measuring Value: Pharmacoeconomics Theory and Practice
`ADRIAN TOWSE, MICHAEL DRUMMOND,
`AND CORINNA SORENSON
`
`Price Indexes for Prescription Drugs: A Review of the Issues
`ANA AIZCORBE AND NICOLE NESTORIAK
`
`Empirical Evidence on the Value of Pharmaceuticals
`CRAIG GARTHWAITE AND MARK DUGGAN
`
`Promotion to Physicians and Consumers
`DON KENKEL AND ALAN MATHlos
`
`The Economics of Vaccines
`FRANK A. SLOAN
`
`Mergers, Acquisitions, and Alliances
`HENRY G. GRABOWSKI AND MARGARET KYLE
`
`Index
`
`CONTENTS
`
`302
`
`336
`
`355
`
`394
`
`463
`
`493
`
`524
`
`552
`
`579
`
`Page 00007
`
`Page 00007
`
`

`

`
`
`CHAPTER 5
`
`THE REGULATION OF
`
`MEDICAL PRODUCTS
`
`ANUP MALANI AND
`
`TOMAS PHILIPSON
`
`IMPROVEMENTS in health have been a major component of overall gain in eco-
`nomic welfare in the last century (Cutler and Richardson 1998; Murphy and Tape]
`2006). Part of these health gains are attributable to medical research (Cutler and
`McClellan 2001; Murphy and Topel 2003; Lichtenberg 2003; Cutler et a1. 2007). hi
`the same time, the cost of health care has tripled in real terms since 1965 and new
`accounts for more than 17 percent ofthe gross dornestic product (GDP). Evidence
`exists to suggest that a large share of this cost growth is driven by technologi-
`cal innovation (Newhouse 1992), including the cost ofmedical products such as
`drugs. The large role of innovation in explaining both improvements in health
`and health care cost growth suggests that it is important to understand the medi‘
`cal research and development (R&D) process and how it is regulated.
`In virtually all developed countries and many developing countries, the 3°“
`ernment provides regulatory oversight over the quality of products generated bl
`medical innovation. In the United States, this oversight is conducted by the US
`Food and Drug Administration (FDA), which regulates drugs, medical devices»
`biologics (products made from living organisms, such as vaccines and blood pIGd'
`ucts), cosmetics, radiation-emitting electronic products, veterinary products. and
`f90ds' According to the FDA: the products it regulates account for more than 0116'
`filth ofUS consumer spending. In the area of medical products, the agencyreview
`w ether drugs and many deVices are safe and effective both before and after lhfl
`have been cleared for sale.
`has :2:;::::;:: in which the FDA regulates the qualit),r of drugs and‘devilclt‘:
`Impact on the cost of their development. The FDA “3‘1"”est a
`
`3
`
`i
`
`!
`!
`l
`
` .4‘
`
`Page 00008
`
`Page 00008
`
`

`

`2
`
`__
`
`-_
`
`I
`
`II
`
`1 l | lII
`
`I
`I
`
`=
`
`' I
`
`1 |
`
`
`
`Page 00009
`
`THEREGULATIONOFMEDICALPRODUCTS
`
`companies conduct clinical trials to demonstrate that their medical products are
`safe and effective. These trials account for a large portion of the total development
`costs of these products (DiMasi et al. 2003; Adams and Brantner 2006). In addi-
`tion, completion of trials does not guarantee that a product will be approved. This
`risk of nonapproval compounds the cost of product development (DiMasi et al.
`2003}.
`
`101
`
`Despite the central role of the FDA in regulating the quality and R&D costs of
`medical products, economists have conducted relatively little theoretical or empir-
`ical research on the efficiency of FDA policies. Ironically, if a product application
`Were presented to the FDA with the scant amount of evidenCe that currently exists
`on the efficiency of the policies of the agency itself, such an application would
`likely be rejected on the basis of insufficient evidence. In this chapter, we synthew
`size and extend, in a nontechnical manner, recent research on the FDA. Our aim is
`
`to shed light on whether the policies ofthe agency itselfare safe and effective when
`measured in terms of economic efficiency.
`The first section provides an overview of the role of the FDA in regulating
`
`pharmaceuticaldrugsandmedicaldevices.Thesecondsectionsurveystheexist-
`
`ing efficiency rationales for government regulation of the information about and
`the quality of medical products and then canvasses the literature for empirical
`studies on the effects of FDA regulation on innovation and costs. The final section
`examines the growing role of tort law—specifically, products liability litigation—in
`
`supplementing FDA regulation of drug quality.
`To understand the relationship of this chapter to others in this book, it is
`helpful to break government influence on medical product innovation into five
`parts. The first is the use of property rights over innovation to encourage invest-
`ment in R&D. The second is direct government spending on R&D. The third
`is premarketing screening of new drugs and devices by the FDA. The fourth is
`postmarketing regulation of medical products by courts (and perhaps the FDA).
`The last is demand for medical products by government-run health insurance
`systems such as Medicare and Medicaid. This chapter focuses on the third and
`fourth categories; that is, the role of quality and safety regulation in directing
`innovation.
`
`FDA REGULATION OF DRUGS, DEVICES,
`
`AND BIOLOGICS
`
`Drugs
`
`The US government began regulating drugs with the 1906 Food and Drug Act.
`That statute prohibited companies from selling misbranded or adulterated
`
`Page 00009
`
`

`

`PHARMACEUTICAL INNOVATION
`
`drugs. Using a modern example, aspirin would be misbranded ifits label stated
`that it was ibuprofen. A drug is adulterated ifit is tainted with a poison or some
`other dangerous substance. The 1906 Act was targeted mainly at quack pat-
`ent medicines and regulated drugs largely after they had already been sold to
`consumers.
`
`extremely toxic, are performed on terminally ill cancer patients.
`
`Since 1962, drug development can be broken down into two parts. The first
`is in vitro and animal testing. If this proves successful, a company may apply to
`the FDA for permission to conduct clinical trials on humans. The application is
`called an Investigational New Drug (IND) application (FD8IC Act, §505(i)). The
`FDA uses the power to approve human testing to regulate the nature of trials con—
`ducted by a company. The FDA works with companies to determine appropriate
`endpoints for clinical evaluation and the sample size required for trials. It may also
`provide guidance on the level ofefficacy or safety that would help clear the way for
`drug approval.
`Clinical testing in turn has three phases. Phase 1 involves studies to determine
`the toxicology ofa drug. A small number of healthy humans2 are gradually given
`increasing doses of a drug to determine how toxic it is. If the dosage can be raised
`to pharmacologically effective closes with tolerable side effects, the drug passes
`phase 1. Phase 2 involves medium—sized studies to identify early evidence that the
`drug is effective in humans. These studies often do not have a control group. Ifthe
`drug demonstratésefficacy relative to historical controls or the expectations of
`the company and the FDA, the drug passes phase 2. In phase 3, the company must
`typically conduct two large, well—controlled trials of a drug’s safety and efficacy.
`
`The United States government did not screen drugs before permitting them to
`be Sold to consumers until the enactment of the Federal Food, Drug, and Cosmetic
`Act (FD8:C Act) in 1938. That statute required that drugs be proved safe before they
`could be sold to consumers. Companies did not have to prove drugs were effective,
`though the FDA could prevent marketing of a drug if the agency could demon-
`strate it did not work. In practice, however, there was no premarket screening of
`drugs for efficacy. In 1962, following a public health scare in which a number of
`children were born with serious birth defects after their mothers used thalido—
`
`mide to reduce morning sickness during pregnancy, the US Congress amended
`the FDStC Act to require that a company provide substantial evidence that its new
`drug is both safe and effective before it could sell that drug (FD&C Act, §§505(a)
`and (d)).1
`
`. This is also the standard now applied to biologics, which are defined by the Public
`Health Service Act §351(i) as “any virus, therapeutic serum, toxin, antitOXin, vaccine,
`blood, blood component or derivative, allergenic product, or analogous product
`applicable to the prevention, treatment or cure ofa disease or c0ndition ofhuman
`beings.” That Act states that biologics are to be regulated under the FD8ZC Act, and the
`FDA has interpreted the term “drug” in the FD&C Act to cover biologics.
`. An exception is cancer drugs. Phase 1 studies of these drugs, because they are
`
`Page 00010
`
`

`

`
`
`THE REGULATION OF MEDICAL PRODUCTS
`
`
`Preclinical
`Clinical
`LApproval
`Market
`.
`y
`*3
`Q
`Toxtcology E gag
`;
`
`
`
`
`
`
`a a a.
`“5: 2 3
`H
`
`
`
`j"
`'3
`
`.
`
`.
`
`_
`
`.I
`
`-.
`'
`
`
`:
`_' :
`I
`
`_
`-
`
`5
`5
`
`_
`
`'
`
`;
`I
`l
`
`i
`
`
`
`
`
`=
`._.
`u
`:
`3
`a“
`s
`«I:
`
`‘“
`safety
`efficacy
`.
`side eflects
`
`H
`1‘3
`.12:
`tn
`
`safety
`
`safety
`dosing
`‘
`c[l1cacy
`
`a E
`i- "'2
`2
`g a
`4:
`
`w
`s :3 “3
`E E g
`E .1. 3
`m H
`9* a a
`
`$15.2 million $23.5 million $86.5 million
`21.6 months lg] months 30.5 months
`61.011 years
`
`0.6 to 2 yearsl 11 to 14 years
`
`Expenses
`Time
`lto 6 years
`Overall probability ofsuccess
`30%
`14%
`9%
`8%
`
`"h-
`—p
`b
`r
`r— —*
`It
`
`Conditional probability of success
`90%
`64%
`48%
`40%
`75%
`
`-+— + —+
`v— +— —r
`
`r-
`
`Figure 5.1 Overview of the drug development process.
`Source: From Philipson and Sun, zoos.
`
`The trials are typically randomized and blinded and use placebos or conventional
`treatments as controls.
`Ifa new drug passes phase 3 with significant evidence of efficacy and a reason-
`able safety profile, the company may file a New Drug Application (NDA) to obtain
`approval from the FDA to market the drug. The FDA reviews the data to ensure
`that there is substantial evidence of efficacy and that the drug has been demon—
`strated to be safe. In practice, the FDA balances the value of efficacy from the
`drug against the cost in terms of side effects when deciding to approve a drug. For
`example, the agency may tolerate greater side effects from a drug that treats a more
`serious medical condition or a drug that dramatically improves health outcomes.
`The overall review process, from preclinical testing to NBA approval, is illustrated
`in Figure 5.1, which is reproduced from Philipson and Sun (2008}.
`Drugs are approved for particular uses. This does not prevent doctors from
`prescribing an approved drug for alternative uses. Such “off-label” uses are quite
`common, for example, in oncology. Use—specific approval does, however, prevent
`companies from advertising unapproved uses to doctors and patients.3 If a com-
`pany seeks to market a drug for as yet unapproved uses, it must conduct studies to
`demonstrate that the drug is safe and effective for those uses.
`
`3. Because of First Amendment protection for truthful commercial speech. however,
`companies can share peer—reviewed articles about unapproved uses, so long as they are
`not accompanied by nonacademic marketing [Washington Legal Foundation v Henney,
`1999}.
`
`=
`
`_
`
`;
`
`_
`
`
`
`Page 00011
`
`Page 00011
`
`

`

`
`
`104 PHARMACEUTICAL INNOVATION
`
`The incomplete regulation of off-label use highlights an important limitation
`in the federal regime for drug regulation. A drug need only be approved for one use
`to be available for almost any use that doctors may pursue. This affects a drug com-
`pany’s decision about the specific use for which to seek approval from the FDA. A
`company will consider the use for which proving safety and efficacy is easiest as
`well as the use for which marketing is most important. It also affects the value of
`the FDA’s regulation. If only a fraction of uses are screened by the FDA, the drug
`market is only partially regulated by the FDA.‘
`The FD&C Act does not merely require a drug to be safe and effective; it also
`rquires that the drug be labeled so that a doctor can prescribe the drug to patients
`in a manner that will ensure proper sorting of drugs to patients. The FDA screens
`a new drug’s label before the drug is approved for marketing and may require that
`the label be amended as the agency receives new information about a drug after
`approval. The label must describe not only indications for use but also inter alia
`contraindications and side effects.
`Clinical testing does not end once a drug has been approved for marketing.
`I Although the FDA does not have the formal authority to require clinical testing
`after a drug is approved, it may hold up drug approval unless the drug company
`“voluntarily” agrees to conduct so-called phase 4 clinical trials. Although some of
`these phase 4 obligations involve placebo—controlled trials, many require long—term
`and large-scale observational studies. The purpose of these studies is to identify
`side effects that are serious but so rare that even phase 3 trials are not sufficiently
`
`provided.
`
`large to detect them.
`One problem identified soon after passage of the 1962 Amendments to the
`FDSrC Act was that companies stopped developing and fully testing drugs for rare
`conditions-The financial burden of clinical testing exceeded the small expected
`revenue from these drugs. Congress addressed this problem in 1983 with the
`Orphan Drug Act. That statute provided companies that developed drugs for rare
`conditions, defined as diseases that affect fewer than 200,000 people annually, tax
`credits to offset their clinical testing costs and seven "years of market exclusivity
`(FDSrC Act §§526, 527).
`-
`i
`'
`.-
`,
`A second "set of_ problems with the 1962 Amendments was that they both
`reduced effective patent life for pioneer new drugs and hampered entry of generic
`versions of these drugs into the market. The long time needed for the FDA to
`review NDAs reduced effective patent life, defined as the number of remaining
`years a company has to market a drug under its patent monopoly following FDA
`approval. This in turn reduced the incentive to innovate. Once a patent expired,
`
`4. This gap in the regulatory regime is partly addressed by the tendency of insurance
`companies to refuse reimbursement for off—label prescription of drugs. But such
`off~label use is difficult to monitor and can often be evaded by administering drugs
`on an inpatient basis at a hospital (rather than in a physician’s office), because hospital
`reimbursement is largely based on a patient’s diagnosis, not the specific treatment
`
`Page 00012
`
`

`

`
`
`_THE REGULATION OF MEDICAL PRODUCTS 105
`
`
`
`however, the Amendments required a generic drug company to perform the same
`Clinical trials that were performed to gain approval for the pioneer drug. The low
`expected profits on generic drugs made such testing unprofitable. Although the
`second problem offset the first, it was not the how Congress intended the patent
`system and the FDSIC Act to interact.
`'
`In 1984 Congress addressed both problems with the so—called Hatch-Waxman
`Act. This statute extended a drug company’s patent on a pioneer drug to partly
`offset the time required for clinical testing and FDA review. The extension was
`capped at 5 years and could never extend effective patent life beyond 15 years.5 In
`return, generic drug companies were not required to conduct trials demonstrating
`that their drugs were safe and effective. Instead, they could market their drugs so
`long as they demonstrated that their drugs were bioequivalent to a pioneer drug
`whose patent had expired or had been invalidated (FD8<C Act §5o5(j)).
`Congress once more took aim at long FDA review times in 1992 with the
`Prescription Drug Users Fee Act (PDUFA). That Act required drug companies to
`pay a user fee to fund resources the FDA could use to speed up the drug approval
`process. In return, the Act required the FDA to commit to deadlines for complet-
`ing its drug reviews. PDUFA was written to expire in five years, but it has repeat-
`edly been renewed.
`
`years.
`
`Devices
`
`Medical device regulation has lagged behind drug regulation in the United
`States. Whereas the 1906 Food and Drug Act prohibited the misbranding or
`adulteration of drugs, devices did not have such restrictions until passage of the
`1938 FDSIC Act. Likewise, whereas the 1938 Act and then the 1962 Amendments
`introduced premarket screening of drugs, there was no premarket screening of
`devices until 1976.
`The 1976 Medical Device Amendments introduced three tiers of regulation
`that finally provided for prernarket screening of certain devices. This new regime
`first sorted devices into three classes and then prescribed different levels of regu-
`lation for the difference classes (FD&C Act §513). Class I devices are Subject only
`to prohibitions on misbranding and adulteration. Class 11 devices are subject to
`these prohibitions and also to certain performance standards developed by the
`FDA. For example, for a device to be labeled an x-ray machine it must meet cer-
`tain guidelines for the quality of its beam and the amount of radiation that the
`
`5. The statute also guaranteed that the FDA would not permit any generic drug
`manufacturer to market a drug within 5 years. This was important to pioneer
`companies, because a generic maker is permitted to sue to invalidate a pioneer’s drug _
`patent before it formally expires. The delayed FDA approval ensured that, even if a
`generic won its patent suit, the pioneer company would enjoy market exclusivity for 5
`
`Page 00013
`
`

`

`
`
`106 PHARMACEUTICAL INNOVATION
`
`EVALUATION OF FDA REGULATION
`
`Rationale for Government Regulation
`
`FDA regulation of medical products has two basic components. The first is pro-
`duction and dissemination of information and analysis about the quality of drugs
`and devices. This is embodied in the FDA’s requirements that companies produce
`data to demonstrate safety and efficacy and that drug and device labels report indi-
`cations as well as contraindications and adverse effects. The second is a minimum
`quality regulation. This is directly embodied in the FDA’s rule that drugs and cer-
`tain devices cannot be sold until they have been demonstrated to be safe and effec—
`tive. It is indirectly implemented in the FDA’s rule that class II devices must meet
`«a
`certain performance standards.
`
`machine may leak (21 CFR 1020.30). Finally, class III devices are regulated much
`like new drugs. Manufacturers must provide data that demonstrate both safety and
`effectiveness. The Amendments sort devices into a class based largely on whether
`the FDA believes that the level of regulation implied by that class is Sufficient to
`reasonably ensure that the device is safe and effective for consumers. Roughly 30
`percent of devices fall in class I, 60 percent in class II, and 10 percent in class III
`(Hutt et al. 2007).
`Since the 1976 Amendments, the development of device regulation has largely
`tracked that of drug regulation. The Orphan Drug Act and the Hatch—Waxman
`Act included provisions for devices. And in 2002, Congress extended PDUFA tO
`apply to medical devices.
`
`(Malani et al. 2009).
`
`Production of Information
`In order to justify the government provision of information, One must explain
`why markets do not produce enough information and why the government should
`step in to do so. Perhaps the production of information is a public good and free—
`riding causes inefficiently iow levels of information to be produced (Musgrave
`1959}. Alternatively, the government may be more efficient at providing informa-
`tion. Perhaps the government has more expertise, or perhaps patients (and doctors
`they might hire to help them) have cognitive limitations.‘S Or perhaps there are
`
`6. This concept is embedded in the average patient standard that the FDA employs. By
`barring drugs that may benefit some patients but not the average patient, the FDA is
`assuming that they are better at sorting different patients to a drug than doctors are
`
`Page 00014
`
`

`

`107
`THE REGULATION OF MEDICAL PRODUCTS
`
`economies of scale in data analysis. It may be more efficient for one entity to ana-
`lyze the data than for each patient to do it on her own.
`Even if one concludes that the government should provide information ab0ut a
`product, there remain important questions about which party—the medical prod-
`uct company, a third private party, or the government—is best suited to produce__ .
`that information and who should pay for the production of that information. The
`producer of the medical product may have private information about the product
`that places it in the best position to test the product. For example, it knows which
`patient group is most likely to benefit. The drug producer and its consumers are
`also entities who gain from information that the product is safe and effective.
`An alternative to testing by the producer is testing by a third private party or
`the government. One concern with permitting medical product companies to test
`their own products is that their studies may be biased in favor of their products.
`Studies have documented that clinical trials conducted by authors with financial
`conflicts from drug companies tend to report larger treatment effects than studies
`without such conflicts (Bekelman et a1. 2003; Lexchin et al. 2003). However, gov—
`ernment—run studies may take longer to complete, and this is a problem because
`delay is costly to producers, who lose monopoly profits under their patents, but not
`to the government. Another problem is that government studies may have a bias
`toward finding that drugs have side effects. Indeed, the FDA has been criticized for
`being overly cautious with drugs, because although the agency is not given credit
`for new effective drugs, it is blamed when it permits an unsafe drug to enter the
`market. (To be fair, the FDA has also been criticized for being captured by the drug
`
`them to do so. A more precise cost-benefit analysis would weigh the cost ofthe
`
`industry.)
`The question of who should fund the FDA’s mission to produce information is
`in part a question about efficiency and in part a question about equity. There is a
`large optimal taxation literature that discusses which sorts oftaxes are able to raise
`a given quantum of revenue with minimum distortions. The lessons there apply to
`the FDA context. However. one consideration that favors taxing medical product
`companies to fund clinical research is that doing so may bypass the administrative
`costs of collecting taxes from third parties. Questions about equity are perhaps
`beyond the ken of economists. But one factor implying that a company and its
`consumers should not be disproportionately taxed to pay for research on the com-
`pany’s product is that information suggesting that a product is not safe or effective
`benefits individuals who do not consume the company’s product and perhaps the
`company’s competitors. This favors a broader tax, although perhaps not so broad
`as to extend to individuals who are not in the market for a medical product.
`Whereas efficient and fair production of information is important, the cen-
`tral concern with the government’s decision to produce information is whether
`the benefits of that information are worth the costs? In chapter 2 of this volume,
`
`7. Medical product companies would conduct R&D and thus produce some information-'-
`about product quality even in the absence ofany government regulation requiring
`
`Page 00015
`
`

`

`
`
`108 PHARMACEUTICAL INNOVATION
`
`DiMasi and Grabowski document the large and growing costs of clinical testing.
`According to DiMasi et al. (2003), a commonly cited study, the cost of completing
`three phases of clinical testing is roughly $200 million. Ofcourse many drugs are
`abandoned before testing is completed. Typically only one in five completes testing
`and is approved bythe FDA for sale. If the cost of testing is normalized by the num-
`ber of successful drugs, one obtains estimates of clinical testing costs in the range
`of $800 million per approved drug.
`Whichever estimate is used, this cost must be balanced against the benefits of
`clinical testing. Note that this is not the same question as whether the revenue from
`a nieuical product exceeds the cost ofclinical testing. The revenue reflects both the
`value of the medical product and the value of information about the product. To
`the extent that doctors would use the medical product without the information, the
`revenue does not reflect the value of information. If, however, doctors would use
`the product on a different population ofpatients when provided with the informa-
`tion, then the additional benefit to this new population of users minus the avoided
`costs to the old users is a benefit ofinformation.
`In an early, seminal paper on the value of FDA regulation, Peltzman (1973)
`attempted to capture some of these ideas in a graph along the lines of that in
`Figure 5.2, panels A and B. Suppose that a typical consumer’s demand for a drug
`before the FDA provides any information on the drug is represented by the line
`DD. The price of the drug is p, so the equilibrium quantity is (1.3 (These initial
`demand and supply curves are drawn in bold to distinguish them.) The graph
`does not capture all facets ofthe drug market, specifically the fact that there may
`be a monopoly producer that charges a price higher than marginal cost and that
`changes price in response to demand shifts. But it is a useful simplification that
`makes the following important point: because the consumer may not have accu-
`rate information about the value ofthe drug, we cannot conclude from the graph
`that the surplus is the triangular area under DD but above p?
`After the FDA provides the consumer information, demand might shift either
`in (panel A) or out {panel B) to line D’D’ (Peltzman 1973; McGuire et a1. 1975). A shift
`inward could occur if the FDA demonstrated that the drug is less safe and effective
`than previously thought. A shift outward could reflect either an FDA assessment that
`
`welfare calculations.
`
`additional information the government requires against the benefit of that additional
`information.
`_
`
`. To keep matters simple, we ignore complications from, for example, health insurance
`(Lakdawalla and Sood 2006, 2009) and pharmaceutical advertising (Lakdawalla et al.
`2006). These features make the calculation of welfare more complicated.
`. We ignore philosophical issues such as whether welfare can be calculated by using
`subjective expected utility. With subjective utility, consumer perceptions are important
`to welfare and the area under DD but abovep may provide some measure ofwelfare.
`Instead, for simplicity, we assume that only objectively verified benefits matter for
`
`Page 00016
`
`

`

`
`
`
`
`
`Units
`qq‘
`D
`0’ Units
`5'
`i
`ofdrug
`ofdrug
`.
`g”
`A: Decrease in demand following disclosure
`B: Increase in demand following disclosure
`'
`
`'
`
`-
`
`I
`
`
`
`
`
`U$iunit
`
`D
`
`UMunir
`
`D
`
`D'
`
`q
`
`0
`
`q’
`
`L] D'
`
`D
`
`Units
`Units
`*1)
`of drug
`ofdrug
`'
`D: Minimum qualily standard above [3
`C: Minimum quality standard below p
`Figure 5.2 Welfare effect from production ofinformation and
`minimum quality standards.
`
`I
`,
`the drug is more safe or effective than previously thought,” patients’increased con-
`fidence in estimates about a drug’s value, or (looking past our model of a represen~
`tative patient) a better sorting of patients to drugs. These shifted demand curves can
`be used to calculate welfare because they reflect “true” information. Whichever way
`demand shifts. the price of the drug may rise to p’ because of the cost ofeither pro-
`ducing information or conveying that information to consumers.11
`
`10. Peltzman {1975) argued that this scenario is unlikely because companies have an
`incentive not to undersell their drugs prior to regulation. However, this does not rule
`out an increase in demand resulting from the other causes listed.
`11.Producing information can entail both fixed and variable costs, which may not be
`reflected in price in a straightforward fashion. For example, the cost of information_
`may depend on the quantity of the drug consumed in equilibrium. To keep matters
`simple, we ignore this comple‘irity in Figure 5.2.
`
`-_:
`
`.
`
`-_:'
`I
`
`r
`
`-
`
`'i
`
`'
`
`_
`
`'
`
`"
`
`'
`
`_
`
`-
`
`'
`
`-
`
`g
`
`;
`
`I
`
`Page 00017
`
`Page 00017
`
`

`

`If demand shifts inward to DU as in panel A. the consumer’s surplus before
`production of information (“preregulatiOn”) has two parts. The first part is the
`area under DD and above p. This triangle represents the “true” surplus from con-
`sumption up to quantity q*. The second part is actually a loss, equivalent to the
`gray triangle. This is consumption between (3* and q that costs more than it is
`truly worth to the consumer. After the production of information (“postregula—
`tion”), the consumer’s surplus is the area under D'D' but above p’. The higher price
`reflects the cost of producing informatiOn. The net gain from the production of
`information is the gray—shaded triangle minus the line-shaded area. The former
`represents the avoidance of excessively costly consumption, and the latter reflects
`the cost of informatiOn. The net gain could be positive and thus increase welfare or
`
`negative and thus decrease it.
`If, instead, demand shifts out to D’D’ as in panel B after the production of
`information, the preregulation surplus is the area under D’D’ and above p, but it is
`bounded on the right by the quantity 9*. This area represents the true surplus from
`preregulation consumption of quantity 9*. The postregulation surplus is the area
`under D'D’ and above p'. This is consumer surplus at the elevated level of demand
`after accounting for higher prices due to the cost of producing information. The
`net gain from regulation is the gray-shaded triangle minus the line-shaded rectan—
`gle. The former is extra surplus extracted clu