Medical Care Output
`and Productivity
`
`Edited by
`
`David M. Cutler and
`Ernst R. Berndt
`
`trch
`md Wealth
`
`'i~i-.e University of Chicago Press
`
`Chicago and London
`
`~ARUCH COLLEGE UBRARY
`
`
`
`PAGE 1 OF 3
`
`SENJU EXHIBIT 2198
`INNOPHARMA v SENJU
`IPR2015-00903
`
`

`
`456
`
`lain M. Cockburn and Aslam H. Anis
`
`variables are based only on the most recently published trial in each year,
`while in models 7 and 9 they are computed as a three-year moving average
`of published trial results. In both cases the drug effect is calculated relative
`to placebo, but very similar results are obtained using just the change rela(cid:173)
`tive to the baseline values.
`Results in models 6 and 7 are encouraging. The signs of the coefficients
`on the characteristics variables conform to our priors, with increased toxic(cid:173)
`ity negatively associated with market share, and increased efficacy posi(cid:173)
`tively associated. Though the coefficient on price is insignificant, and cor(cid:173)
`responds to a very small elasticity, it is at least negative in model6. A very
`small price effect is also consistent with our interpretation of results from
`estimating the price equation.
`Models 8 and 9 include fixed drug effects in the estimation to control
`for drug-specific problems in measuring market share or characteristics.
`Several of these dummies are highly significant, and they markedly im(cid:173)
`prove the fit of the model, suggesting that we do indeed have systematic
`problems in measuring market shares. Furthermore the estimated coeffi(cid:173)
`cients on the other variables change substantially when we include fixed
`drug effects, indicating that the equations omit significant variables driving
`quantities consumed, either quality characteristics of drugs or other drug(cid:173)
`specific factors which_ determine demand.
`
`11.7 Conclusion
`
`Economic considerations appear to play a relatively minor role in the
`market for DMARDs. Information from published clinical trials relating
`to key quality characteristics of these drugs (efficacy and toxicity) is statis(cid:173)
`tically assopiated with changes in their quantity shares in this market, but
`has no cortsistent impact on relative prices. Given the nature of RA, these
`results may not be too surprising. They do~ however, point to some inter(cid:173)
`esting economic issues which we have not attempted to address in this
`study.
`First, there is the question of using prices to measure the impact of
`technical change on consumer welfare in markets such as this one. Most
`prior work on innovation, quality change, and pricing has examined the
`prices of new goods which embody tecpnological change in the form of
`improvements to tangible aspects of-quality. Here the technical change
`takes a rather. unusual form: R&D generates revisions to the intangible
`information set possessed by physicians and patients, affecting perceived
`quality rather than physical characteristics such as speed, durability,
`weight, and so on. R&D surely improves welfare in this context, but the
`fact that relative prices in this market change very little (and are most
`likely determined exogenously) and that demand ~ppears to be quite price
`inelastic means that its impact is very difficult to s€'e in price space. Rather,
`
`•
`
`PAGE 2 OF 3
`
`

`
`.. ~,
`~~· .:s··· .. ,r~.··. h~­
`·.:1'··!•.: ... ,.,
`
`I
`
`·~ .
`
`;hed trial in each year,
`e-year moving average
`::t is calc~lated relative
`1g just the change rela-
`
`gns of the coefficients
`:, with increased toxic(cid:173)
`creased efficacy posi(cid:173)
`insignificant, and cor(cid:173)
`ive in model6. A very
`tation of results from
`
`estimation to control
`tre or characteristics.
`d they markedly im(cid:173)
`:ieed have systematic
`the estimated coeffi(cid:173)
`hen we include fixed
`:ant variables driving
`drugs or other drug-
`
`~ly minor role in the
`linical trials relating
`nd toxicity) is statis(cid:173)
`s in this market, but
`nature of RA, these
`point to some inter(cid:173)
`j to address in this
`
`1sure the impact of
`has this one. Most
`g has examined the
`1nge in the form of
`le technical change
`1s to the intangible
`affecting perceived
`: speed, durability,
`1is context, but the
`ittle (and are most
`trs to be quite price
`?rice space. Rather,
`
`Hedonic Analysis of Arthritis Drugs
`
`457
`
`the most visible direct effect of changes in quality is seen in movements in
`quantities, which has significant implications for how we should interpret
`movements in, for example, a fixed-weight price index.
`Second, these results hint at an interesting variety of non price competi(cid:173)
`tion. Rents to producers in this market are determined initially by the level
`of prices (which to a rough approximation they set once in real terms,
`often based upon conditions prevailing in unrelated markets) and then by
`the evolution of quantities as consumers and/or their agents respond to
`exogenous changes in perceived quality. In such circumstances the role
`played by marketing and promotional activity may well be very important.
`Our analysis here is based on the generation of new information about
`product quality in the form of publication of research results in peer re(cid:173)
`viewed journals by (hopefully) impartial authors. The question of how this
`information reaches practicing physicians and their patients has not been
`examined here. In future work we hope to extend our analysis of this mar(cid:173)
`ket to include marketing and promotional activity by producers of these
`drugs, which may shed light on the interesting question of the relative im(cid:173)
`portance of objective versus persuasive information in drug choices.
`
`References
`
`Anis, A. H., and Q. Wen. 1998. Price regulation of pharmaceuticals in Canada.
`Journal of Health Economics 17 (1): 21-38.
`Berndt, E. R., I. M. Cockburn, and Z. Griliches. 1997. Pharmaceutical innova(cid:173)
`tions and market dynamics: Tracking effects on price indexes for antidepressant
`drugs. Brookings Papers on Economic Activity, Microeconomics, 133-88.
`Berndt, E. R., and S. N. Finkelstein. 1992. Price indexes for anti-hypertensive
`drugs that incorporate quality change: A progress report on a feasibility study.
`MIT Program on th~.:Pharmaceutical Industry, Working Paper no. 6-92.
`Berry, S. B. 1994. Estimating discrete-choice models of product differentiation.
`RAND Journal of Economics 25:242-62.
`Berry, S. B., J. Levinsohn, and A. Pakes. 1995. Automobile prices in market equi(cid:173)
`librium. Econometrica 63:841-90.
`Brewerton, D. 1994. All about arthritis. Cambridge, Mass.: Harvard University
`Press.
`Cash, J. M., and J. H. Klippel. 1994. Second-line drug therapy for rheumatoid ar(cid:173)
`thritis. New England Journal of Medicine 330:1368-75.
`Felson, D. T., J. J. Anderson, and R. F. Meenan. 1990 .. The comparative efficacy
`and toxicity of second-line drugs in rheumatoid arthritis: Results of two meta(cid:173)
`analyses. Arthritis and Rheumatism 33:1449-59.
`Griliches, A., and I. M. Cockburn. 1995. Generics and new goods in pharmaceuti(cid:173)
`cal price indexes. American Economic Review 84:1213-32.
`IMS America Inc. 1980-94. National drug and therapeutic index-Drugs. Plym(cid:173)
`outh Meeting, Pa.: IMS America.
`- - - . 1980-94. National drug and therapeutic index-Diagnosis. Plymouth
`Meeting, Pa.: IMS America.
`~~-.
`
`PAGE 3 OF 3