Do Pharmaceutical Sales Respond to
`Scienti c Evidence?
`
`Pierre Azoulay
`Columbia University, New York, NY 10027-6902
`pierre.azoulay@columbia.edu
`
`I investigate how different sources of information inuence the diffusion of
`pharmaceutical innovations. In prescription-drug markets, both advertising
`and scientic information stemming from clinical trials can affect physicians’
`prescription choices. Using novel indices of clinical-research output, I nd
`that both marketing and scientic evidence directly inuence the diffusion
`process in the antiulcer-drug market, with marketing having a more pro-
`nounced inuence. I also nd evidence that clinical outputs are important
`drivers of rms’ marketing efforts, affecting sales indirectly. Taken together,
`the direct and indirect effects of science on demand imply strong private
`incentives for clinical research. I conclude that product-market competition
`in the pharmaceutical industry is shaped by both advertising rivalries and
`scientic rivalries. Moreover, drug advertising may perform an important
`informative function.
`
`Introduction
`1.
`How do different types of information inuence the diffusion of phar-
`maceutical innovation? The spread of technological advances is lim-
`ited by the extent to which relevant information is available among
`potential adopters. Furthermore, the information necessary for the dif-
`fusion of pioneer products may be different from that required for the
`market penetration of subsequent innovations.
`In most industries, one would expect underinvestment in the
`production of knowledge to limit the availability of objective sources
`of
`information about product characteristics, safety, and efcacy
`(Arrow, 1962). However, in prescription-drug markets, two features
`of the institutional environment—extensive, government-mandated
`
`For useful suggestions and support, I would like to thank audience participants at
`the MIT IO Lunch and the NBER Productivity Lunch, as well as Dan Ackerberg,
`Richard Caves, Peter Davis, John DeFigueiredo, Sara Ellison, Stan Finkelstein, Jeff Fur-
`man, David Genesove, Jerry Hausman, David Hsu, Rebecca Henderson, Kip King, Bob
`Pindyck, Robert Rubin, Otto Toivanen, and especially Scott Stern and Ernie Berndt. The
`usual disclaimer applies.
`
`© 2002 Massachusetts Institute of Technology.
`Journal of Economics & Management Strategy, Volume 11, Number 4, Winter 2002, 551–594
`
`SENJU EXHIBIT 2209
`INNOPHARMA v SENJU
`IPR2015-00903
`
`PAGE 1 OF 3
`
`

`
`Do Pharmaceutical Sales Respond to Scientic Evidence?
`
`573
`
`In what follows, I make the identifying assumption that clinical-
`research outputs and (science1 and science2) are uncorrelated with
`month-to-month changes to these unobserved characteristics (D »).16
`
`4.4 Results
`This section reports the empirical results on the direct competitive
`effects of advertising and science (Section 5 below estimates the
`magnitude of the indirect effect of scientic information on demand—
`through the promotion efforts induced by scientic activity). The
`ndings suggest that the levels of both variables drive diffusion
`and performance on the product market, with marketing activities
`having a more pronounced effect. Because of the semilog functional
`form of the logit model, coefcient estimates are not immediately
`interpretable as elasticities.
`Turning to the results of Table IV, model (1) presents OLS
`estimates of the diffusion equation ignoring the effect of scientic
`information (b 8
`0). The coefcient on stkdetailing and
`9
`stkjournal are positive and signicant, and the demand curve is
`downward sloping, as anticipated. Other product characteristics con-
`tribute signicantly to the model t, with signs conforming to priors,
`except for dosage and interactions.17
`Model (2) adds the effect of science. In this specication,
`b 5
`decreases by about 10%, and both science1 and science2 obtain
`positive and signicant coefcients. A likelihood-ratio test between
`models (1) and (2) easily rejects the former (lr
`62.334, df
`2).
`Interestingly,
`including the science measures causes the dosage
`coefcient to ip sign, while the coefcient on interactions is not
`statistically signicant.
`Model (3) addresses the issue of endogeneity by presenting
`2SLS estimates. The results are similar, except that the stkdetailing
`coefcient drops substantially.18 Because serial correlation is present
`
`16. I gain additional insight into this issue by examining whether variation in the
`ow of scientic information can be explained by differences in the characteristics of the
`rms selling these drugs. I report the results of specications that regress  ow (and
`also the count of published clinical studies) on a constant, the log of the US revenues of
`the rm outside the gastrointestinal therapeutic area, its stock of detailing minutes on
`all its other products, and its stock of journal advertising expenditures on all its other
`products. I observe no systematic relationship between these variables. See Azoulay
`(2001) for further details.
`17. One would expect consumers to prefer drugs with the lowest dosage frequency,
`ceteris paribus. Tagamet entered the market with a requirement of four daily doses, but
`was able to match Zantac’s twice-a-day dosage within a year of the new drug’s entry.
`18. The Hausman specication test decisively rejects the null hypothesis of exoge-
`nous regressors (Â2
`2065.144, df
`7).
`
`PAGE 2 OF 3
`
`b
`

`
`586
`
`Journal of Economics & Management Strategy
`
`Because advertising does not jam all other information channels avail-
`able to reach the population of prescribing physicians, pharmaceutical
`rms face strong private incentives to perform clinical research.
`
`6. Concluding Remarks
`The results presented here demonstrate that product-market competi-
`tion in the H2-antagonist therapeutic class was shaped by rival rms’
`advertising efforts and the quality of the scientic information con-
`cerning the four drugs. The paper provides an original methodology
`for computing indices of quality-adjusted scientic outputs. I nd that
`marketing had a more pronounced direct effect on demand than sci-
`ence, but the latter was still statistically and economically signicant.
`I introduce the distinction between market-expanding and compara-
`tive science, demonstrating that the second type was a particularly
`effective business-stealing weapon for the second mover Zantac. In
`addition, I nd evidence that clinical-research outputs were important
`drivers of rms’ promotion efforts, although detailing and journal-
`advertising expenditures also responded positively to the intensity
`of competitors’ marketing campaigns. Taken together, these results
`suggest that pharmaceutical advertising does not perform a purely
`persuasive function, nor does it jam professionally sanctioned infor-
`mation channels by preventing scientic results to get through to pre-
`scribing physicians.
`I take into account both the direct and the indirect effect of sci-
`ence on demand to compute the appropriate elasticities. The sum
`of the direct and indirect effects yields a level for the total market-
`expanding science elasticities of demand around 0.4 for the pioneer
`drug and its challenger, and positive and signicantly above zero
`for the two later entrants. These results imply strong private incen-
`tives for performing clinical research and suggest that controlled clin-
`ical trials do not accomplish the sole function of securing regulatory
`approval, but also represent investments whose effects on the product
`market are both substantial and long-lived. The results are consistent
`with long-run trends noted by industry practitioners (Carr, 1998). A
`growing number of drugs go into postapproval, so-called Phase IV
`trials. These are designed to extend the range of conditions for which
`a drug can be used, thereby making it more protable. Such trials
`also satisfy the need to accumulate evidence for use in persuading
`physicians to favor new drugs over older ones.
`The results of this paper are also of signicant interest in the
`continuing debate surrounding pharmaceutical advertising. Numer-
`ous academic critics of the industry have argued that promotion
`
`PAGE 3 OF 3