Journal of Economic Perspectives—Volume 16, Number 4—Fall 2002—Pages 45– 66
`
`Pharmaceuticals in U.S. Health Care:
`Determinants of Quantity and Price
`
`Ernst R. Berndt
`
`T he U.S. pharmaceutical industry has again become the focus of consider-
`
`able public attention and controversy, in part because of its highly visible
`sales growth, averaging around 12–13 percent annually over the last
`15 years, along with its apparent high pro tability.
`To understand the economics underlying the pharmaceutical industry, the
`traditional microeconomic distinctions between short-, medium- and long-run costs
`turn out to be critical, as does the distinction between accounting and economic
`costs. Although the long-run supply of new pharmaceuticals depends critically on
`research and development efforts, in the short run, R&D costs are largely sunk or
` xed, while the marginal costs of manufacturing another tablet of a developed drug
`are typically very small. For patent-protected branded pharmaceuticals, in most
`cases, prices are several times greater than short-run marginal production costs,
`thereby making short-run gross pro t margin accounting measures for pharma-
`ceuticals among the highest in all U.S. industries.
`These different short- and long-run cost perspectives help one understand why
`the pharmaceutical industry is particularly vulnerable to public policies involving
`pricing. Although the con ict between static ef ciency (price new drugs low, near
`short-run marginal cost) versus dynamic ef ciency (price new drugs high, maintain
`incentives for innovation) is a deep and enduring one, as the costs of bringing new
`drugs to market have increased sharply in recent years, this tradeoff is becoming
`more severe. The resolution of this static versus dynamic ef ciency con ict is likely
`the single most important issue facing the pharmaceutical industry over the next
`decade.
`
`y Ernst R. Berndt is the Louis B. Seley Professor of Applied Economics, MIT Sloan School of
`Management, and Research Associate, National Bureau of Economic Research, both in
`Cambridge, Massachusetts.
`
`ALCON 2024
`Apotex Corp. v. Alcon Research, Ltd.
`Case IPR2013-00428
`
`

`

`46
`
`Journal of Economic Perspectives
`
`Table 1
`Expenditures on Health Care Services and Supplies, Percentage Share by
`Category, 1960–2000
`
`Category
`
`1960
`
`1970
`
`1980
`
`1990
`
`1995
`
`1998
`
`1999
`
`2000
`
`Hospital care
`Physician services
`Prescription drugs
`Nursing home care
`All other
`
`36.9%
`21.0
`10.7
`3.2
`28.2
`
`41.0%
`20.8
`8.2
`6.2
`23.8
`
`43.5%
`20.2
`5.1
`7.6
`23.6
`
`38.0%
`21.7
`5.6
`7.5
`27.2
`
`36.1%
`20.4
`5.9
`7.8
`29.8
`
`34.1%
`23.1
`7.8
`8.0
`26.8
`
`33.4%
`23.0
`8.8
`7.6
`27.2
`
`32.8%
`22.8
`9.7
`7.3
`27.4
`
`Notes: Data from the Centers for Medicare and Medicaid Services, National Health Accounts, various
`years. Expenditures exclude research and construction. “All other” includes dental and other profes-
`sional services, home health care, nonprescription drugs and medical durables, vision products, net cost
`of private health insurance and government public health activities.
`
`Pharmaceutical Expenditure Growth in Context
`
`Prescription pharmaceuticals (hereafter, pharmaceuticals) account for a mod-
`est but increasingly important share of total U.S. health care expenditures. As
`shown in Table 1, between 1960 and 2000, hospital care has consistently been the
`single largest component of health care expenditures, ranging from 33 to 44 per-
`cent. Physician services has been the second largest component, accounting for
`approximately 20 –23 percent of total health care expenditures. Although nursing
`home care was in third place during the 1980s and until about 1998, in 1999–2000
`the relatively rapid sales growth of outpatient prescription pharmaceuticals pro-
`pelled it into third place. By 2000, the outpatient pharmaceutical share constituted
`about 9.7 percent of total U.S. health care expenditures, slightly less than its
`10.7 percent share back in 1960, but almost twice its 1980 share of 5.1 percent. This
`9.7 percent share in 2000 translates into expenditures of $121.8 billion, or on average
`about $450 per person; for 2002, projected drug costs are $161 billion, or about
`$565 per person ( http://www.hcfa.gov/stats/NHE-Proj/proj2001/tables/t11.htm ).
`The distribution of drug expenditures across the entire population is quite skewed;
`for some chronically ill patients, the drug cost burden can be many times larger
`than the $565 average.
`It is useful to decompose pharmaceutical expenditure growth into price
`growth of incumbent products, quantity growth of incumbent products and expen-
`ditures on new products. Although expenditures on pharmaceuticals grew at
`roughly similar average annual rates of 11.9 percent between 1987 and 1994 and
`12.9 percent between 1994 and 2000, the primary driver of this spending growth
`differed in these two subperiods. As shown in Figure 1, from 1987 through 1994,
`price growth at 6.1 percent annually was responsible for slightly more than half of
`the 11.9 percent sales growth. However, from 1994 through 2000, price growth
`accounted for only about one- fth of revenue growth (2.7 percentage points out
`of 12.9 percent), with the remaining four- fths re ecting volume/mix changes
`
`

`

`Ernst R. Berndt
`
`47
`
`Figure 1
`The U.S. Prescription Pharmaceutical Market Total Annual Sales Growth and its
`Sources
`
`Source: IMS Health, Retail and Provider Perspective™, 2001.
`
`in utilization rates for incumbent drugs, as well as expenditures on new phar-
`maceuticals. Hence, in recent years, price increases have been relatively less
`important, and instead, quantity growth— greater utilization of incumbent and
`new products— has been the primary driver of increased spending.
`The relative importance of new products as drivers of total expenditure growth
`depends in part on how one de nes “new drugs.” IMS Health, a supplier of data on
`pharmaceutical use, de nes a new product as any product launched during the
`twelve months ending with the last calendar quarter and having a new National
`Drug Classi cation code. Although a one-year window de nition is rather narrow,
`it includes all new products, such as new generics, new branded products and new
`line extensions of existing molecules. While it would be preferable to have a new
`product de nition that excluded new generic drugs having comparable branded
`presentations and strengths, spending on new generics is typically much less than
`that for new brands, and thus the IMS de nition is a reasonable  rst-cut estimate
`of the impact of new products on drug spending. Using this de nition of new
`products, IMS Health (2002) reports that from 1997 through 2000, expenditures
`on new drugs accounted for on average 42 percent of total spending growth,
`although in 2001, this share fell to about 20 percent as the number of new
`
`

`

`48
`
`Journal of Economic Perspectives
`
`treatment approvals by the Food and Drug Administration decreased from a high
`of 53 in 1996 to 27 in 2000.
`Some of the new drugs have involved novel generations of treatments obso-
`lescing the old, vividly illustrating Schumpeter’s notion of dynamic competition.
`The proton pump inhibitors (Prilosec, Prevacid, Aciphex and Protonix), for ex-
`ample, have largely replaced the previous generation’s “wonder drugs,” the H2-
`antagonists (Tagamet, Zantac, Pepcid and Axid) for the treatment and prevention
`of ulcers and various gastrointestinal disorders. Due in large part to their increased
`safety or preferable side effect pro les, the new atypical antipsychotics such as
`Risperdal, Zyprexa, Seroquel and Geodon have largely replaced older generation
`products such as Haldol and Clozapine for the treatment of schizophrenia; IMS
`Health reports that in 2001, this therapeutic class had the highest annual spending
`growth at 33 percent.
`In other cases, new products have provided effective therapies where previ-
`ously few if any had existed: for example, the protease inhibitors for acquired
`immunode ciency syndrome, the human growth hormone Protopin, and Viagra
`for the treatment of erectile dysfunction. Still others have been new entrants in
`already crowded therapeutic classes, but their manufacturers have claimed lower
`prices and greater cost-effectiveness: for example, Lipitor, a statin lipid-lowering
`agent that in 2001 was the top selling drug in the United States at $5.2 billion, and
`Protonix, a proton pump inhibitor.
`A particularly remarkable class of new products has been the set of selective
`serotonin reuptake inhibitors, like Prozac, Zoloft, Paxil, Luvox and Celexa. Used
`initially primarily for the treatment of depression, these drugs have been shown to
`offset nonpharmaceutical medical costs by decreasing the intensity of costly psy-
`chotherapy (Berndt et al., 2002). Years after their initial launch in the late 1980s,
`developers of these drugs found new uses for the antidepressants, and now several
`have received FDA approval for the treatment of various anxiety disorders. Another
`antidepressant, Wellbutrin, has been approved for treatment of smoking cessation
`and is now manufactured in a lower dosage form and marketed under a different
`brand name, Zyban. While the new anxiety uses for the antidepressants would not
`be counted as new products by IMS Health, whereas Zyban likely would, it is clear
`that the pharmaceutical marketplace is a dynamic one, with both new products and
`new approved uses playing a very important role.
`Major changes in the use of pharmaceuticals do not always involve newer
`drugs. The off-patent beta blockers, for example, along with other old blood
`pressure medications, are now routinely used in treatment following a heart attack,
`reducing morbidity and mortality associated with heart disease, decreasing the
`probability of a second heart attack and reducing costs of cardiovascular-related
`hospitalizations and surgeries (National Committee for Quality Assurance, http://
`www.ncqa.org ). By contrast, results from recent highly publicized randomized
`studies suggest that the widely practiced use of hormone replacement therapy to
`reduce cardiovascular disease in postmenopausal women may not only be ineffec-
`tive, but may be harmful to certain women (Petitti, 2002).
`This background suggests basic questions about quantity, price and value in
`
`

`

`Pharmaceuticals in U.S. Health Care: Determinants of Quantity and Price
`
`49
`
`Table 2
`Share of Prescription Drug Spending by Source of Payment, Selected Years,
`1965–1998
`
`Year
`
`1965
`1970
`1975
`1980
`1985
`1990
`1995
`1996
`1997
`1998
`
`Out-of-Pocket
`
`Private Insurance
`
`Medicaid
`
`Other
`
`92.6%
`82.4
`75.4
`66.0
`55.4
`48.3
`33.9
`31.6
`29.1
`26.6
`
`3.5%
`8.8
`12.2
`20.1
`29.9
`34.4
`46.8
`48.8
`50.8
`52.7
`
`0.0%
`7.6
`10.8
`11.7
`11.8
`13.5
`15.8
`16.1
`16.5
`17.1
`
`3.9
`1.2
`1.6
`2.2
`2.9
`3.8
`3.4
`3.5
`3.6
`3.6
`
`Source: Report to the President: Prescription Drug Coverage, Spending, Utilization and Prices (Wash-
`ington: DHHS, April 2000, Table 2-30).
`
`the markets for prescription pharmaceuticals. What reasons explain the substantial
`increase in the quantity of prescription pharmaceuticals consumed since the mid-
`1990s? What factors affect the pricing of pharmaceuticals?
`
`Increased Utilization of Pharmaceuticals
`Although new products and new knowledge have surely had a substantial
`impact on the utilization of pharmaceuticals, here I instead focus on two of the
`more prosaic economic forces behind this quantity growth: increased drug insur-
`ance bene t coverage and enhanced marketing efforts, including in particular
`direct-to-consumer marketing.
`Insurance coverage tends to increase consumption of medical services (for
`example, Newhouse, 1993, pp. 165–171). In this context, it is useful to examine
`third party drug insurance coverage, which has grown rapidly in the United States,
`particularly since 1990. As seen in Table 2, in 1965, only 3.5 percent of dollars spent
`for prescription pharmaceuticals were paid for by private insurance, while 92.6 per-
`cent came from consumers’ out-of-pocket payments. In 1990, 25 years later, the
`private plus Medicaid insurance share had increased to 47.9 percent, about equal
`to the 48.3 percent from consumers’ out-of-pocket payments. During the 1990s, the
`private plus Medicaid insurance share increased even more—to 62.6 percent in
`1995 and 69.8 percent in 1998, even as the consumers’ out-of-pocket share fell to
`33.9 percent in 1995 and 26.6 percent in 1998.
`In part, this expanded drug insurance coverage re ected an increasingly tight
`U.S. labor market in which employers provided more generous nonwage bene ts
`to workers. But the increased drug bene t coverage also emanated from growth in
`the managed health care organizational form, which widely sought to attract
`insuree members by offering integrated drug and nondrug insurance coverage.
`Growth in drug bene t coverage was also greatly facilitated by technological
`progress. Developments in information and telecommunications technologies
`
`

`

`50
`
`Journal of Economic Perspectives
`
`made possible ef cient and low-cost instantaneous processing of covered insurance
`transactions at retail pharmacies, replacing now antiquated “shoe box” methods by
`which insurees previously had been reimbursed for covered drug expenditures only
`after  lling out various forms and providing proof of purchase. These technological
`developments also fostered the growth of new organizations—pharmaceutical ben-
`e t managers—that assisted payers and managed care organizations in part by
`closely monitoring prescription drug dispensing behavior in real time.
`Until very recently, changes in the form of cost sharing between insurer and
`insuree have ampli ed the incentives that insurance provides to increase use of
`prescription drugs. Copeland (2000) reports that from 1990 to 1994, private health
`insurers increased pharmaceutical expenditures at an average annual growth rate
`of 17.3 percent, while consumers’ out-of-pocket expenditures increased by only
`4.8 percent per year. From 1994 to 1998, differentials in these average annual
`growth rates became even greater at 18.9 percent and 2.7 percent, respectively.
`With consumers directly footing an ever smaller marginal share of prescription
`drug costs, and insurers bearing the ever larger residual share, it is not surprising
`that the utilization of prescription drugs has increased dramatically.
`In recent years, private third party insurers have attempted to control rising
`drug expenditures in a number of ways. One approach has been to steer use to
`preferred drugs on the insurer’s list of approved medications (the “formulary”).
`During the early and mid-1990s, a common “two-tier” method was to charge
`consumers a low copayment for a generic drug (say, $5) and a higher copayment
`(like $10) for a branded drug in the same therapeutic class; in some relatively
`uncommon cases, a branded drug was not covered at all by the third party insurer,
`unless the physician provided evidence of medical necessity.
`However, varying the copayments between generic and branded drugs in this
`two-tier formulary structure still meant that consumers typically faced no differen-
`tial copayment among various brands within the same therapeutic class— even
`though to third party payers, the costs of various branded drugs within the same
`therapeutic class were not necessarily the same. Already by the early 1990s, private
`third party payers began to negotiate with brand drug manufacturers, offering
`them assured formulary access and/or preferential status on the payer’s formulary
`in return for price discounts (rebates and chargebacks). To a brand manufacturer,
`being excluded from preferential status on a formulary could result in a substantial
`loss of market share. In this way, private third party payers exercised their power of
`exclusion to hold down drug prices and make aggregated demand for prescription
`pharmaceuticals more price elastic (Elzinga and Mills, 1997).
`The ability of third party payers to obtain discounts from brand manufacturers
`was of course greatly affected by payers’ perceived ability to in uence the prescrib-
`ing behavior of physicians. This ability to affect physician prescribing behavior
`varies among organizational forms. In staff-model health maintenance organiza-
`tions, physicians’ prescribing behavior can be in uenced quite effectively, for in
`those organizations, physicians work solely for the organization that employs them,
`and drug prescriptions are dispensed primarily via in-house pharmacies. However,
`
`

`

`Ernst R. Berndt
`
`51
`
`staff-model health maintenance organizations are now only a small part of the
`market; in the  rst six months of 2001, they accounted for less than 1 percent of all
`prescription drug expenditures. Instead,
`insurees have migrated toward more
` exible preferred provider and independent practice association networks where
`control over physician prescribing behavior is less stringent. Most solo and group
`physician practices, for example, have contracts with more than one payer. Data
`from the 1996–1997 Community Tracking Survey of Physicians indicate that 61 per-
`cent of primary care physicians and 64 percent of specialists had contracts with six
`or more managed care organizations. When a physician is confronted with six or
`more distinct formularies, the result is often that the physician simply follows his or
`her own preferences, rather than tracking the formulary preferences of each
`patients’ payer (Hellerstein, 1998). As a result, efforts by third party payers to
`control rising brand drug costs by focusing on physician prescribing behavior have
`had smaller effects than payers had hoped.
`An alternative mechanism by which private payers have attempted to constrain
`rising pharmaceutical expenditures is to exert  nancial pressures on the patient
`(rather than the physician) by altering cost-sharing provisions between insurer and
`insuree. This involves changing the insuree’s copayment (a  xed dollar amount) or
`coinsurance rate (a  xed percentage).
`In the most recent formulary incarnations, called “three tier” formularies,
`many payers have not only raised the levels of patient copayments across both all
`generic and all brand prescriptions, but they have also implemented differential
`copayments amongst the various branded drugs within a given therapeutic class. In
`2000, the average patient retail copayment for a generic ( rst tier) drug was $7.17,
`that for preferred brands in a second tier was $14.14, while that for all other
`nonpreferred brands on the third tier was $27.35; these copayment levels represent
`successively larger percentage increases from 1998 levels across the three tiers, at
`17 percent, 27 percent and 56 percent, respectively (Takeeda-Lilly, 2001).
`Another recent phenomenon is using a separate and more attractive copay-
`ment structure to encourage mail order dispensing of certain drugs, particularly for
`medications used to treat long-term chronic conditions. On a per prescription
`basis, dispensing fees paid by insurers to mail order providers are only about half
`that paid retail pharmacies— on average $1.15 vs. $2.31 in 2000, down about
`8 percent and 37 percent from 1995, respectively (Takeeda-Lilly, 2001). By shifting
`insurees needing long-term pharmaceutical treatment to mail order prescriptions
`(with, say, 90 days’ supply rather than the usual retail pharmacy 30 days’ supply),
`payers/insurers have been able to realize signi cant dispensing cost reductions
`per day of therapy. This use of mail order dispensing facilities by third party payers
`has of course exerted downward cost pressures on traditional retail pharmacy
`dispensing.
`Three-tier formularies involving coinsurance (not copayment) have also grown
`recently. On average in 2000, the  rst, second and third tier coinsurance rates were
`19, 21 and 42 percent for retail dispensing, respectively, and 18, 21 and (as
`expected, a much higher) 58 percent for mail order. Both three-tier formulary
`
`

`

`52
`
`Journal of Economic Perspectives
`
`schemes have diffused rapidly. In 2000, three-tier copayment structures were of-
`fered by about 35 percent of employers surveyed, and three-tier coinsurance by
`7 percent; the remaining 58 percent still had two-tier bene t designs. Some
`observers believe that because multiple-tier coinsurance schemes more  nely and
`continuously affect consumers’ marginal payments, they are increasingly likely to
`replace multiple-tier copayment designs. To date, very little is known about whether
`three-tier bene t design structures are more effective than the traditional two-tier
`schemes in shifting market shares among various medications, their impacts on
`overall drug expenditures and their effects on patients’ health outcomes.
`
`Aggressive Marketing Efforts
`A second major factor affecting growth in prescription drug utilization is
`greater marketing efforts by pharmaceutical  rms. The intensity and growth of
`marketing efforts by pharmaceutical  rms in the last decade has generated con-
`siderable controversy, but little careful analysis.
`For the U.S. pharmaceutical
`industry between 1996 and 2000, aggregate
`annual marketing to sales ratios ranged between 13.6 and 15.3 percent (Rosenthal
`et al., 2002a). To put this  gure into perspective, it is instructive to consider a
`distinction between experience and search goods proposed by Nelson (1974). A
`pure search good does not require one’s own consumption experience to gain
`reliable information on its traits; the potential purchaser need only examine its
`product speci cations, which are often quanti able. An example might be a DVD
`player or a credit card contract. By contrast, the characteristics and effectiveness of
`experience goods are largely idiosyncratic and unpredictable, so if a person wishes
`to obtain reliable information on the consumption aspects of a pure experience
`good, the person must consume it directly. Since for many pharmaceuticals, patient
`response in terms of ef cacy, side effects and adverse interactions is idiosyncratic,
`pharmaceuticals often have the characteristics of experience goods. Nelson hypoth-
`esized that marketing-to-sales ratios would be greater for experience than for
`search goods, since with an experience good, sellers need to entice a consumer
`continuously to engage in a trial of the product and, to preempt defection to
`another product, remind the consumer of previous successful experiences.
`Data on advertising/sales ratios are consistent with viewing prescription phar-
`maceuticals as experience goods. In 1998, advertising-sales ratios for primarily
`search good companies were relatively low: Home Depot, 1.6 percent; Phillips
`Electronics, 2.4 percent; American Express, 3.2 percent; Circuit City, 4.5 percent;
`Sony Corporation, 4.7 percent; and Intel, 5.8 percent. By contrast, 1998 ratios for
`experience good companies were as high if not higher than for pharmaceuti-
`cals: Ralston Purina, 14.9 percent; Wendy’s International, 16.7 percent; Coors,
`18.8 percent; McDonald’s, 21.1 percent; Estee Lauder Cosmetics, 22.2 percent;
`Revlon, 24.0 percent; and L’Oreal, 26.5 percent (Berndt, 2001).
`Perhaps more controversial is the relatively recent changing composition of
`pharmaceutical marketing efforts. Traditionally, pharmaceutical  rms focused
`most of their marketing on physicians, either in the form of physician of ce visits
`
`

`

`Pharmaceuticals in U.S. Health Care: Determinants of Quantity and Price
`
`53
`
`by manufacturers’ representatives (called “detailers,” often dropping off product
`information and free samples) or by advertising in medical journals. Beginning in
`the early and mid-1990s, however, pharmaceutical  rms greatly expanded their
`direct-to-consumer marketing efforts, not only in print media, but, especially, on
`television.
`Although the aggregate marketing-to-sales ratio in the U.S. pharmaceutical
`industry was essentially unchanged at 14.1 percent in 1996 and 14.0 percent in
`2000, the direct-to-consumer marketing-sales ratio rose from 1.2 to 2.2 percent
`between 1996 and 2000 (Rosenthal et al., 2002a). This one percentage point
`increase for direct-to-consumer marketing intensity came at the expense of the
`physician of ce-based detailing to sales ratio, which fell from 11.3 to 10.7 percent,
`and from the hospital-based plus medical journal advertising to sales ratio, which
`between 1996 and 2000, fell proportionally more from 1.1 to 0.6 percent. These
`changing ratios suggest, in keeping with the classic Dorfman-Steiner (1954) theo-
`rem, that elasticities of demand with respect to marketing increased for direct-to-
`consumer relative to other marketing instruments. Factors likely increasing the
`impact of direct-to-consumer marketing include the baby boom cohort becoming
`older and experiencing identi able symptoms treatable with prescription drugs, as
`well as more general secular trends such as increased consumer empowerment.
`It may not be coincidental, however, that direct-to-consumer advertising began
`to grow substantially in the early 1990s even as drug insurance bene ts covered an
`ever larger proportion of all drug expenditures. Over time, the reach of direct-to-
`consumer marketing included an ever larger number of drug insurance bene cia-
`ries, eventually reaching threshold levels where direct-to-consumer marketing be-
`came more pro table. With insured consumers paying relatively little at the margin
`for prescriptions, and with third party payers attempting to shift physician prescrib-
`ing behavior toward the payers’ formulary-preferred drugs, it is quite likely that
`direct-to-consumer marketing elasticities increased, at least relative to the elastici-
`ties of physician marketing.
`An additional factor likely affecting direct-to-consumer marketing efforts is the
`Food and Drug Administration’s 1997 clari cation of what constituted “adequate
`provision” of information on radio and television advertising. The FDA’s 1997
`guidelines indicated that, in addition to meeting obligations requiring balanced
`information on bene ts and risks, a necessary condition for ful lling the adequate
`provision requirement was to refer in the broadcast to four sources from which
`consumers could obtain further information: their physician, a toll-free number, a
`print advertisement and a website. This clari cation meant that pharmaceutical
` rms could now employ consumer-focused broadcast media more intensively in
`their marketing efforts, without fear of violating FDA marketing regulations.
`While direct-to-consumer marketing has grown considerably, it is largely con-
`centrated on a small portion of drugs. Rosenthal et al. (2002b) report that the
`median direct-to-consumer spending across drugs was zero between 1996 and 1999
`and that in 2000, the top 20 advertised drugs accounted for about 60 percent of
`total direct-to-consumer marketing spending.
`
`

`

`54
`
`Journal of Economic Perspectives
`
`To date, very little research has been disseminated examining the effects of
`direct-to-consumer marketing on the sales of individual drugs within a therapeutic
`class, on sales of the entire class, and on health outcomes. Preliminary evidence
`suggests that direct-to-consumer marketing primarily increases total demand for a
`therapeutic class, with little shifting of market shares among the within-class indi-
`vidual products (Rosenthal et al., 2002b). One interpretation of this  nding is
`consistent with the view that direct-to-consumer marketing encourages previously
`untreated patients to ask their physicians about treatment, but that the choice of a
`particular drug within a therapeutic class is determined primarily by other factors
`such as physician experience/preferences and by insurance coverage. Other things
`equal, brands having largest market share within a therapeutic class may therefore
`face the greatest incentive to increase direct-to-consumer marketing efforts, for
`they can expect to appropriate the corresponding share of increased class sales.
`
`Summing Up Reasons Behind the Rise in Quantity Consumed
`Currently there are no quantitative estimates of how much each possible factor
`contributed to the rise in consumption of prescription pharmaceuticals since the
`mid-1990s. My conjecture is that the rise in insurance coverage has been more
`important than increased marketing efforts, in part since aggregate marketing to
`sales ratios have been relatively constant, but these two factors interact positively.
`Moreover, both factors should be viewed in the larger context where signi cant
`new product introductions have replaced older therapies and expanded treatment
`possibilities, and where demographic changes led by baby boomers have increased
`overall demand for health care services.
`In the future, if prescription drug insurance coverage is extended to the
`elderly via expanded Medicare bene ts, I would expect a substantial insurance-
`induced growth in prescription drug utilization, but how large this effect would be
`is unclear. On the other hand, if in the future the U.S. labor market is not as tight
`as it has been in the last few years, one might expect to see private sector employers
`cut back on their contribution to drug cost-sharing, perhaps implementing greater
`patient copayments or coinsurance bene t design changes. This could have a
`substantial negative impact on growth in drug spending. It is also quite possible that
`employers will shift away from de ned bene t schemes for drug coverage, where
`the insurance plan speci es what will be covered, and instead move toward de ned
`contribution plans, where the plan speci es how much the employer will spend but
`gives employees greater choice in selecting prescription drug coverage. In such an
`environment, there could well be even greater incentives for pharmaceutical  rms
`to increase direct-to-consumer marketing efforts.
`Finally, a number of states are now considering implementation of formulary-
`like provisions in their Medicaid programs, specifying that certain branded drugs
`be designated as “preferred” and where physicians would need case-by-case autho-
`rization to prescribe nonpreferred brands. Notably, unlike the private sector, up
`until very recently, the states’ Medicaid administrators have not utilized such
`prescription steering procedures. This may now change, and whether the states
`have authority to do so is being litigated. States are also attempting to expand drug
`
`

`

`Ernst R. Berndt
`
`55
`
`coverage eligibility beyond current Medicaid criteria to the “near poor,” thereby
`likely affecting both drug utilization and pricing.
`
`The Pricing of Pharmaceuticals
`
`The pricing of pharmaceuticals is complex and controversial. Nonetheless,
`basic economic theory provides many instructive insights. A useful starting point is
`to distinguish between prescription drugs that are patent-protected from those
`whose patent protection has expired.
`For branded drugs, patent protection implies that the innovating  rm has
`rights to market exclusivity for a limited time period; in many cases, an effective
`patent life is between 12 and 15 years, although this range can vary substantially
`according to the amount of time needed for FDA regulatory approval and other
`factors. Market exclusivity for a certain brand of a prescription pharmaceutical
`drug does not usually generate a pure monopoly situation, because most branded
`drugs face competition from other brands within a given therapeutic class. Within
`many therapeutic classes of drugs, a number of possible substitute medications
`exist, and in such cases, the market structure is more appropriately depicted by the
`differentiated product oligopoly framework. In such a setting, it is useful to envis-
`age the optimal pro t maximizing price as equaling marginal cost plus a positive
`margin, where the margin depends on bene ts and attributes (including prices) of
`the  rm’s own drug relative to other drugs in the therapeutic class, on attributes of
`nondrug therapies, patient heterogeneity and other demand-side considerations.
`Since branded pharmaceuticals often have patent protection and oligopolistic
`differentiated products, one should expect this margin of price over marginal cost
`to be positive. But how large? What counts as marginal cost?
`
`The Cost Structure of Pharmaceuticals
`What one includes within marginal costs depends critically on the time period,
`as well as on the stage of the product life cycle. For pharmaceutical  rms, R&D costs
`are very large and are increasing rapidly. A study released in November 2001 by the
`Tufts Center for the Study of Drug Development ( http://www.tufts.edu/med/
`csdd ) estimates that on average, it now costs about $802 million and takes 12 years
`to bring a new drug to market,
`including prepatent application R&D. This
`$802 million is more than twice the $350 million estimated a decade earlier, and in
`large part re ects increased costs of clinical development rather than basic R&D. In
`the long run, R&D costs