Comments

Health Affairs encourages readers to engage in discussion via comments on our Web site.

• To RESPOND to a particular article: Click on the link "Submit a response to this article" in the box at the top right-hand corner of the article.
• To READ responses to a particular article: Click on the link "View responses" in the box at the top right-hand corner of the article.

Comments to:

Salomeh Keyhani, Marie Diener-West, and Neil Powe Are Development Times For Pharmaceuticals Increasing Or Decreasing? Health Affairs, March/April 2006; 25(2): 461-468. [Abstract] [Full Text] [Figures Only] [PDF] [Reprints & Permissions]

Comments:Submit a response to this article

Comments published:

The Real Significance of Drug Development Times

Joseph A. DiMasi, Ph.D.   ( 14 March 2006 )

Re: The Real Significance of Drug Development Times

Salomeh Keyhani   ( 15 March 2006 )

Re: The Real Significance of Drug Development Times

Donald W. Light   ( 21 March 2006 )

A Response To Light From DiMasi

Joseph A. DiMasi   ( 24 March 2006 )

High Drug Prices Set by Economic Theory?

Donald W. Light   ( 2 June 2006 )

A Response to Light's Rejoinder from DiMasi

Joseph A. DiMasi   ( 13 June 2006 )

The Real Significance of Drug Development Times 14 March 2006
Joseph A. DiMasi, Ph.D., Director of Economic Analysis Tufts Center for the Study of Drug Development, Tufts University Send comment to journal: Re: The Real Significance of Drug Development Times joseph.dimasi{at}tufts.edu Joseph A. DiMasi, Ph.D.	Keyhani et al. analyze clinical development and regulatory approval times in the United States for a sample of new drugs and biologics approved from 1992 through 2001 (the stated end date is just one day into 2002). Numerous studies on development and approval times covering various periods have been published (including many conducted by my center). None, as far I know, claimed a connection to drug pricing. Keyhani et al. set out to examine whether what they assert is industry's claim that prescription drug prices have been increasing in the last decade or so because development times have also been increasing is correct. The hypothesized connection links development times to R&D costs and then to drug prices. The support offered for this claim about the "industry" view comes from several documents published by the industry's U.S. trade association. The notion that these documents make a claim that rising development times are largely responsible for rising drug prices is, I think, at best tenuous, and, in any event, some industry leaders have explicitly refuted the view that there is a causal link between R&D costs and drug prices.[1] Regardless of whether one should attribute to industry as a whole or in part the claim asserted in Keyhani et al., we could appeal to the economics of drug development and pricing to help us draw conclusions. At the time that drug prices are determined, the associated R&D expenditures for a drug are sunk costs. Pricing to recoup their R&D costs, even when possible, will in general not be financially optimal for the firm. Thus, basic economic logic tells us that R&D costs do not determine prices. Market and regulatory (primarily outside the United States) factors are what determine drug prices. We do not need a study of development times to come to this conclusion. What we can say about R&D costs and pricing is that expected R&D (and other) costs together with expected prices (and associated quantities demanded) determine expected profitability. Expected profitability, in turn, determines the incentive to develop new therapies. Even if one accepts the underlying proposition that whether development time is an important explanatory factor in the determination of a drug price is a reasonable enough hypothesis to test, the evidence provided by Keyhani et al. would not resolve the issue. Development times are only one of the factors that effect R&D costs. Cash outlays, failure rates, and the cost of capital also help determine R&D costs. What's more, as the authors note, total development and approval time includes a period (pre-IND time) that they have not measured. The pre-IND period also often includes clinical testing on the compound conducted outside the United States. In addition, the time cost component of a total R&D cost estimate can increase substantially even if development time and the cost of capital remain constant, in the same way that total interest payments on a loan increase when the amount borrowed increases and the loan term and interest rate remain the same. Finally, while our data on U.S. clinical development times also show a decline for the mid- to late 1990s approvals, more recent data than Keyhani et al. examined (2002-2004) exhibit an increase in U.S. clinical development times (we also noted increases in clinical development times in general over several decades leading up to the late 1980s).[2] If development times are not causally related to drug prices, then what is the significance of drug development time trends? At least two fundamental reasons why it is important to examine drug development times come to mind. First, if development times lengthen, then patients will get access to new therapies later than they otherwise would. Second, longer development times increase R&D costs and shorten the period during which firms can earn the kinds of returns that they need in the long run to make investment in this area financially viable. In other words, other things equal, longer development times reduce innovation incentives. As a consequence, fewer new therapies may be developed. NOTES 1. R.V. Gilmartin, "Meeting the Challenge of Pharmaceutical Innovation in an Environment of Rising R&D Costs," November 31, 2001, http://www.merck.com/newsroom/executive_speeches/113001.html [accessed March 13, 2006]; and H.A. McKinnell, A Call to Action: Taking Back Healthcare for Future Generations (New York: McGraw-Hill, 2005). 2. J.A. DiMasi, "New Drug Development in the United States from 1963 to 1999," Clinical Pharmacology and Therapeutics 69, no. 5 (May 2001):286-296; and Tufts CSDD, "Longer Clinical Times Are Extending Time to Market for New Drugs in U.S.," Impact Report 7, no. 6 (Nov/Dec 2005).
Re: The Real Significance of Drug Development Times 15 March 2006
Salomeh Keyhani, Assistant Professor of Health Policy Mount Sinai School of Medicine Send comment to journal: Re: Re: The Real Significance of Drug Development Times salomeh.keyhani{at}mountsinai.org Salomeh Keyhani	In our study, we examined trends in development times and found that they were decreasing. We suggested that development times are important because they are one factor in R&D costs and because of industry claims that significant R&D costs are the main reason behind high drug prices. The statement by DiMasi that industry does not link drug prices to development time and research and development costs is to ignore decades of industry lobbying on this issue. PhRMA’s links to its reports “Why Do Prescription Drugs Costs so Much” are no longer available; however, a quick search of the Internet with the words “drug prices” and “research and development costs” yields many documents referring to this oft repeated claim by PhRMA representatives.[1,2] By conducting this study, we were hoping to raise questions about the connection between drug prices and research and development costs. DiMasi et al. estimated that it costs on average $802 million to develop a new drug.[3] Ironically, DiMasi’s study is the most widely cited by industry to justify high drug prices.[4,5] We agree with DiMasi that there is little direct evidence to link R&D costs and drug prices. DiMasi is correct in his contention that industry leaders are now refuting the link between R&D costs and drug prices. The new industry position appears to be that drugs are priced based on value.[6] This is a welcome development. We certainly hope that industry supports its new mantra of value with robust independent cost-effectiveness analyses so that the value of new drugs can be assessed. Hopefully, industry will not resist possible statutory changes that will allow the CMS to purchase drugs based on comparative clinical efficacy of new drugs and cost-effectiveness studies, which are the concrete tools necessary to measure value. Finally, DiMasi suggests that drug development time is again increasing in recent years. Unfortunately, research in this area is never transparent and the underlying data proprietary and not available to the public. To our knowledge, our study was the first study that revealed individual data on drugs. The underlying data on development times that DiMasi used to evaluate recent trends is not available to all researchers. Transparency is essential to the dissemination of credible research and the formulation of sound policy. Notes 1. C. Domrose, “Rising prices of prescriptions force older patients to make trade-offs”, October 30, 2000, http://www.nurseweek.com/news/features/00-10/drugs.asp [Accessed March 13, 2006]. 2. “Drug Companies Examine Themselves Amongst Waves of Bad Publicity,” February 7 2005, http://www.consumerwatchdog.org/healthcare/nw/?postId=964&pageTitle=Drug+Companies+Examine+Themselves+Amongst+Waves+of+Bad+Publicity [Accessed March 13, 2006]. 3. J.A. DiMasi, R.W. Hansen, H.G. Grabowski, “The Price of Innovation: New Estimates of Drug Development Costs,” Journal of Health Economics Mar 2003;22(2):151-185. 4. BNA’s Daily Report for Executives “Average Drug R&D Cost Is $802 Million,” December 3, 2001, http://lists.essential.org/pipermail/ip- health/2001-December/002498.html [Accessed March 13, 2006]. 5. L. Richwine, "Drug Development Costs $802 million on average” November 30, 2001, http://www.mult- sclerosis.org/news/Dec2001/HighCostOfDrugDevelopment.html [Accessed March 13, 2006]. 6. R.V. Gilmartin, "Meeting the Challenge of Pharmaceutical Innovation in an Environment of Rising R&D Costs," November 31, 2001, http://www.merck.com/newsroom/executive_speeches/113001.html [Accessed March 13, 2006].
Re: The Real Significance of Drug Development Times 21 March 2006
Donald W. Light, Professor University of Medicine and Dentistry of New Jersey Send comment to journal: Re: Re: The Real Significance of Drug Development Times dlight{at}princeton.edu Donald W. Light	Joseph DiMasi is correct that, narrowly speaking, no research studies establish a connection between development times and drug prices, but the industry and its lobbyists have made this connection central to their decades-long “fact” that the long times and great costs of R&D are why prices need to be so high and why protections from normal price competition need to be extended. To pretend that this connection does not exist is particularly disingenuous, since DiMasi et al. have watched for years while the industry’s trade association, PhRMA, has misrepresented their research that the estimated R&D cost for the most costly subsample of all new drugs is $800 million, by proclaiming far and wide that it was the cost for the average new drug, without objecting to their research being so substantially mispresented. For these reasons, the disconnect between lengths of trials and prices in the article by Keyhani et al., the quality of their non-secret data, the size of their sample, and its detail discredit a core of the industry’s lobbying and campaigns, even if (like much else they claim) it is not based directly on any research study. The authors' more extensive article in Medical Care provides considerably more detail. DiMasi is equally disingenuous when he argues correctly that sunk costs are not related to price, first because the industry has used this claim for over 40 years, and second because the argument for high prices in order to pay for expected R&D costs are, of course, based on the very high estimates of past costs, which have been largely supplied by DiMasi and his industry-supported research center. To pretend that estimates of expected R&D costs are not related to estimates of past R&D costs is to ignore the politics of the pharmaceutical industry. And while "development times are only one of the factors that effect [sic] R&D costs,” they are a major one that affects cash outlays, the cost of capital, and how rapidly companies can begin their big launch to recover those costs. (The contents of this letter are mine and do not represent those of any institution to which I am affiliated.)
A Response To Light From DiMasi 24 March 2006
Joseph A. DiMasi, Director of Economic Analysis Tufts Center for the Study of Drug Development, Tufts University Send comment to journal: Re: A Response To Light From DiMasi joseph.dimasi{at}tufts.edu Joseph A. DiMasi	Donald Light’s letter badly distorts what I wrote in my letter about the Keyhani et al. article. Moreover, he makes erroneous claims, and, unfortunately, couches his prose in ad hominem terms. I did not write (or “pretend”) that pharmaceutical company executives or industry representatives have never linked R&D costs and individual drug prices or maintained that prices “need” to be high because of high R&D costs. What I did indicate in my letter was simply that, for at least a number of years now, some industry leaders have explicitly and publicly rejected the notion that individual drug prices are determined on the basis of the R&D costs specific to those drugs. This was by no means meant to be the essence of my comment, but surely it is a fact that should have been acknowledged in the paper. My own argument on the substance of this issue was based on economic logic (not, as Keyhani and Light suggested, on references to the results of empirical investigations). I find it remarkable that Light calls me disingenuous for making an argument that he acknowledges is correct and presumably is central to his concerns. The essential points made in my comment were about the implications of the study’s results and the significance of development times. The abstract to the paper, the text of the paper, and Keyhani’s letter make it clear that they were chiefly interested in establishing whether R&D costs are linked to individual drug prices. The following three points summarize the real substance of my letter: 1. This purported connection can be rejected on the basis of economic theory. 2. Even if one does not accept this logic as sufficient to dismiss the hypothesis, the Keyhani et al. analysis of U.S. clinical development times does not come remotely close to a standard of evidence that would be needed to reject it. No study of development times alone, no matter how complete or accurate, could do so. Thus, Light is wrong in claiming that the Keyhani et al. paper discredits this hypothesis. 3. There are reasons to study development times separately, but Keyhani et al.’s primary reason is not one of them. Light veers off from the points that I actually made in my letter and from the Keyhani et al. article to make erroneous claims about our R&D cost study and the validity of PhRMA’s representation of it. Light asserts that our study estimated costs for “the most costly subsample of all new drugs” and that we “watched” while PhRMA "substantially misrepresented” our research by proclaiming that our estimate is an average for new drugs. If industry describes our result as an estimate of the average cost per approved new drug, then industry is correct. Here, it is Light who is being disingenuous. He knows that we view a “new drug” to be a new active ingredient (as do pharmacologists).[1] New drugs typically come in many product presentations (drug strengths and forms), including often as multiple presentations with an original marketing approval. The presentations that are approved after original marketing approval are new “drug products,” but not new drugs. It makes no more sense to essentially divide total R&D expenditures by the number of all of these product presentations than it does to provide an estimate of the cost of designing and developing a new car model by dividing total R&D costs by the number of trim levels that the manufacturer happens to offer for sale. The reader can find a fuller discussion of the flaws of this and other criticisms in a paper that my coauthors and I have posted on the Web.[2] Finally, Light’s suggestion that I pretended that “estimates of expected R&D costs are not related to estimates of past R&D costs” is puzzling. There is no discussion in my letter to this effect. In any event, Light skirts cogent discussions of the important underlying economic issues, resorting instead to overheated rhetoric about politics and lobbying and to personal attacks. While economic theory indicates that individual drug prices will not be set on a cost-plus basis, economic theory also suggests that significant price regulation can have a negative effect on the incentives to develop new drugs. If actual cost and price outcomes are reasonable proxies for prior expectations and if entry in drug markets is sufficient to result in near-total rent dissipation (as studies of ex-post profitability suggest), then one could argue that in some aggregate sense the price structure that was realized in the marketplace (or something relatively close to it) was indeed “needed” to incentivize firms to incur the development costs for the drugs that they did investigate, even if no price was set based on sunk costs. There is, however, more to consider. Light does not use the term “justified” in relation to prices (although Keyhani did). It has often been said that the industry has “justified” its prices by reference to its high R&D costs. Need and justification are two quite different concepts in this setting. While particular price levels may be “needed” to induce development of a given set of innovations, whether a particular innovation is worth its price is a separate issue that deserves, and likely will receive, additional attention. NOTES 1. J.A. DiMasi, R.W. Hansen, H.G. Grabowski, “Extraordinary Claims Requires Extraordinary Evidence: Reply,” Journal of Health Economics Sep 2005;24(4):1034-1044; J.A. DiMasi, R.W. Hansen, H.G. Grabowski, “Setting the Record Straight on Setting the Record Straight: Response to the Light and Warburton Rejoinder,” Journal of Health Economics Sep 2005;24(4):1049-1053. 2. J.A. DiMasi, R.W. Hansen, H.G. Grabowski, “Assessing Claims about the Cost of New Drug Development: A Critique of the Public Citizen and TB Alliance Reports,” November 1, 2004, http://csdd.tufts.edu/_documents/www/Doc_231_45_735.pdf [Accessed March 23, 2005].
High Drug Prices Set by Economic Theory? 2 June 2006
Donald W. Light, Professor University of Medicine and Dentistry of New Jersey Send comment to journal: Re: High Drug Prices Set by Economic Theory? dlight{at}princeton.edu Donald W. Light	Please allow me to apologize for calling disingenuous Joseph DiMasi’s argument that the sunk costs and long development times involved in testing new drugs do not affect the high prices charged for them. Economic theory makes clear they do not, because they are sunk costs, and theory holds that prices are set by what the market will bear, not sunk costs. DiMasi concludes, “Thus, basic economic logic tells us that R&D costs do not determine prices.” Disingenuous came to mind when DiMasi continued, that “expected R&D (and other) costs together with expected prices (and associated quantities demanded) determine expected profitability. Expected profitability, in turn, determines the incentive to develop new therapies.” But on what else would expected R&D costs usually be based than studies of past R&D costs, principally by the author and his colleagues? And development times directly affect the other causes of R&D costs that DiMasi lists, such as cash outlays, failure rates, and cost of capital. Thus, diminishing the importance of Keyhani, Diener-West, and Powe’s research by arguing that development times are only one of several factors affecting costs and that sunk costs are irrelevant anyway seemed calculated to divert attention from the central influence of trial lengths on those factors and on expected R&D costs. In his response to my letter, DiMasi claims that he did not discuss sunk R&D costs as irrelevant to prices while expected R&D costs are central to them, as if they were not related. Have I once again “badly distorted” what he wrote, even though I’ve used his own words? See for yourself. This exchange would make good material for classroom discussion about pharmaceutical policy and argumentation. The pharmaceutical industry’s long, intensive campaign to defend and extend regulatory protections from normal market-based pricing has drawn centrally on studies by DiMasi and his colleagues to emphasize that prices must be high because the sunk costs of R&D are huge and cannot be recovered without more regulatory protections. Those huge cost estimates are based, as the authors made clear in their key studies, on the select sample of new chemical entities (NCEs) researched and developed entirely in house that have R&D costs several times greater than others. Ever since then, however, the pharmaceutical industry’s trade associations, writers, and lobbyists have cited the high R&D cost estimate of this subsample as the cost per average new drug, a substantial inflation of what DiMasi and his colleagues stated was their conclusion. In his letter dated 24 March, DiMasi slides by my point that the authors have left this substantial inflation of their estimate unchallenged for years. Instead, he employs a distinction the industry has never bothered with, between “new drug” meaning a newly approved chemical entity, and “new drug” meaning a newly approved product for which a new drug approval was submitted. My point still stands. If we pull back from DiMasi’s letters and look at them analytically, we can see they involve multiple switches between different models of argumentation. DiMasi’s first letter contains distortions and bold assertions that cannot be supported by evidence, on one hand, and close, precise adhering to academic purity, on the other. For example, Keyhani et al. did not “make a claim that rising development times are largely responsible for rising drug prices,” a set-up distortion used by DiMasi to discredit their research. The authors made no such claim, but so what? If it works, go for it. Then comes the “R&D costs do not determine prices” passage, based on invoking pure academic theory. Once again, no researcher has claimed that R&D costs determine prices -- another straw man that ignores 40 years of the industry claiming that they do.[1,2] More likely is that marketing costs drive up prices, given that many drug companies spend three times more on marketing than R&D, and someone has to pay for drug reps wooing physicians to prescribe more high-price drugs. Next DiMasi asserts that expected R&D costs, prices, and demand determine expected profitability. No, they don’t. How can an expected something, minus an expected something else, be said to determine anything? Only by definition -- that is, by employing theory as self-justifying prophecy for high prices by using the regulatory powers of government to stave off market-based pricing for many years, to incentivize firms to develop new drugs. This is the kind economic theory as self-fulfilling prophesy that Amartya Sen and other distinguished economists have critiqued for years. It gives health economics a bad name. Keyhani et al. have employed a far more transparent data set on drug development than the highly secretive set that the industry provides DiMasi and his teammates, and they document that drug development times are much shorter than pharmaceutical officers, trade associations, and lobbyists claim in justifying high prices. (For a recent example, see the State Department report written by Pfizer.)[3] Drug companies earn back all their R&D expenditures every year, with good profits, at Canadian and European prices.[4] There are no foreign “free riders,” only high U.S. prices that pay for large budgets for marketing and lobbying to protect high prices. In their other major article, Keyhani et al. [5] document that drugs for cancer and HIV/AIDS have the shortest development times (and lowest research costs). The very high prices charged for these drugs simply exploit the desperately ill, and the rest of us who help pay their bills. NOTES 1. Temin P. Taking Your Medicine: Drug Regulation in the United States. Cambridge, MA: Harvard University Press, 1980. 2. Pharmaceutical Research and Manufacturers of America. Why Do Prescription Drugs Cost So Much…And Other Questions About Your Medicine. Washington DC: PhRMA (Pharmaceutical Research and Manufacturers of America), 2005. 3. Masia N. The Cost of Developing a New Drug. Washington D.C.: U S Department of State, International Information Programs, 2006: 1-4. 4. Light DW, Lexchin J. Foreign free riders and the high price of US medicines. BMJ 2005;331:958-60. 5. Keyhani S, Diener-West M, Powe N. Do drug prices reflect development time and government investment? Medical Care 2005;43:753-62.
A Response to Light's Rejoinder from DiMasi 13 June 2006
Joseph A. DiMasi, Director of Economic Analysis Tufts Center for the Study of Drug Development Send comment to journal: Re: A Response to Light's Rejoinder from DiMasi joseph.dimasi{at}tufts.edu Joseph A. DiMasi	Donald Light’s second letter once again demonstrates a proclivity to make ad hominem and innuendo-laden arguments. My letters lay out my own thinking in some detail on topics related to the Keyhani et al. article. Readers can determine for themselves what I really wrote and meant. I stand behind my comments in both letters. I believe that we have reached a point where further back-and-forth here would not be productive. So, I will not open up for discussion new areas that Light brings in that are unrelated to my comments about the Keyhani et al. article (e.g., the economics of marketing, foreign country contributions to global R&D, what drug industry officials purportedly were thinking or implying when they referred to the cost of getting a new drug from discovery to the marketplace, etc.). I will, however, point out here a fresh set of misleading and false statements that are in Light’s last letter. 1. Light claims that I was disingenuous for stating that expected costs and returns determine expected profitability and that development incentives depend on expected profitability. I do not see how anyone could interpret as disingenuous this basic, and one would think unobjectionable, statement about innovation incentives for for-profit firms. In connection to this, Light asks on what else would expected R&D costs be based than on our past studies of average new drug development costs. When considering the development prospects for a particular drug a pharmaceutical firm will map out a development plan specific to that drug, project costs for the myriad activities called for by that plan, specify subjective probabilities for proceeding through various development phases based on their views of the prospects for that particular drug, and factor in expectations about future sales under various labeling scenarios should the drug reach the market. Managers may use some of our component estimates as starting-off points when thinking about how to formulate their drug-specific decision analyses, but this would only speak to the confidence that the firms themselves had in our estimates. In any event, how is any of this relevant to the points that I made in my letters? Light does not and cannot explain that. 2. Keyhani et al. correctly note in their paper that development times impact R&D costs through their effect on opportunity cost measurements. Ironically, this is a concept that Light has dismissed elsewhere.[1] Light instead asserts here that development time is an important direct causal factor in determining cash outlays, failure rates, and the cost of capital. Development times cannot be said to affect any of these other factors. At best, Light is confusing causation with correlation in the case of cash outlays. However, shorter development times for any particular period may or may not be associated with lower cash outlays. One cannot assume that cash outlays will go in one direction or another in relation to development times. Our last two studies of R&D costs, for example, showed substantially higher cash outlays for the later study period, but a nearly constant average total time from discovery to approval across the periods. It is hard to even conceive of a correlation, let alone causation, between development times and failure rates. Finally, Light’s claim that development times directly affect the cost of capital suggests that he has little, if any, grasp of finance theory (to say nothing of the fact that his ascribing any importance to the cost of capital in this context logically contradicts his prior dismissal of opportunity costs). The cost of capital, as determined by a standard technique such as the application of the Capital Asset Pricing Model (which we used for our studies), is dependent on nondiversifiable investment risk, not on development times (or, for that matter, on failure rates or cash outlays). 3. Light asserts: “DiMasi claims that he did not discuss sunk R&D costs as irrelevant to prices while expected R&D costs are central to them, as if they were not related.” He then asks sardonically if he has once again badly distorted what I wrote, claiming that he is using my own words. The answer is a resounding yes! I wrote nothing of the kind. I wrote simply that sunk R&D costs do not determine prices and that expected R&D costs are one factor contributing to innovation incentives. It is not even clear what criticism Light could possibly be trying to make with his assertion. Light’s rhetoric is constructed so that on the surface it sounds like it might be a meaningful criticism, but on close inspection it makes no sense. Light invites readers to check for themselves. I do as well. 4. Light claims that my first letter contains "distortions and bold assertions that cannot be supported by evidence.” What is his evidence for this dramatic claim? As an example, he offers a quote from my letter (that Keyhani et al. supposedly “'make a claim that rising development times are largely responsible for rising development prices’”), and he states that Keyhani et al. made no such claim. Light does not provide proper context here. The quote was in relation to industry documents, not what Keyhani et al. themselves believe or found regarding a relationship between development times and drug prices. In relation to the documents and development times, Keyhani et al. use the phrase “an important factor.” In retrospect, the adverb “largely” may have been a bit too strong as a descriptor, although this is of no consequence here. I would thus replace “largely” with “an important factor,” but this does not change in any way the essence of my comment or the validity of my point. 5. Light then accuses me of constructing a straw man when I argued that sunk costs do not determine individual drug prices. Of course, this is precisely the argument that Light had argued in his first letter was disingenuous even though he acknowledged it to be correct, an obvious inconsistency for which he now apologizes in the first paragraph of his second letter. In my letter, the objective of this discussion was to point out that, on the basis of economic logic, testing whether a relationship exists between development times and drug prices, as Keyhani et al. sought to do, is unnecessary. I further argued that even if, on whatever rationale, one maintains that it is reasonable to test this hypothesis, what Keyhani et al. did was inadequate to that objective. How can this possibly be construed as a straw man argument? 6. Light’s final bit of “evidence” in this section is that my assertion that expected profits equal expected returns minus expected costs somehow constitutes employing theory as a “self-justifying prophecy for high prices.” This construction is a non sequitur. There is no logical basis for Light’s conclusion. A proper reading of my response letter indicates that whether realized price structures (or something close to them) were, in an aggregate sense, necessary to induce development that did occur is actually an empirical issue, having to do with whether what economists refer to as excess profits exist for this industry to a significant and long-lasting extent (the evidence suggests that they do not). 7. In Light’s final paragraph, he claims that the Keyhani et al. analysis shows that drug development times are much shorter than what is claimed in a report by Masia [2]. This is false. The only quantitative mention of development time in the Masia report is a range for the average time from discovery to marketing approval. The Keyhani et al. analysis only deals with a portion of this period (U.S. clinical development and regulatory approval times). 8. Finally, Light claims that in their article published in Medical Care [3], Keyhani et al. “document” that research costs are lowest for cancer and HIV/AIDS drugs. This also is a false claim. Keyhani et al. provided information on U.S. clinical development and regulatory approval times, but they had no information on research costs. Notes 1. D.W. Light, J. Lexchin, “Will Lower Drug Prices Jeopardize Drug Research? A Policy Fact Sheet,” The American Journal of Bioethics Winter 2004;4(1):W3-W6. 2. Masia N, The Cost of Developing a New Drug. Washington, DC: U.S. Department of State, International Information Program, 2006:1-4. 3. S. Keyhani, M. Diener-West, N. Powe, “Do Drug Prices Reflect Development Time and Government Investment?,” Medical Care 2005;43(8):753-62.