Lower drug development costs than industry reports–shock!

At the SciScip discussion group notice has been given to an article published in September 2017 in JAMA that has determined the median cost to develop a new cancer drug at $648m, much lower than the pharmaceutical industry reports for the cost to develop new drugs in general. After FDA approval, these ten drugs generated a median revenue of $1.7 billion. Since the article is paywalled, I don’t know what time period that income was over.

The authors looked at 10 drugs developed by companies to reach their finding. We might then wonder if the “sample size” is too small to draw a general conclusion. Fine. But if there are 10 more drugs out there to bring the median up to the billion or so dollars that the industry touts as its numbers, those ten drugs will have to be running on the up side of a billion dollars–say $1.4 billion or so–for the industry’s figure to be close to accurate.

The authors then use this finding to conclude that there’s more profit than it costs to develop the drugs. This in itself is about as bland a “meaning” as we could have. We might wonder if the authors first looked for drugs that had modestly good commercial lives, then did their analysis for costs, and then reached the conclusion that the costs were less than income. If so, big wow on the methodology.

But there’s a different point I will make. This study examines only drug development costs for for-profit companies operating on a monopoly product model. The study finds that in this model, for the selected drugs, the median cost to develop is $648m. Even if this figure is relatively accurate, it’s not “the cost to develop cancer drugs”–it’s only the “cost to develop cancer drugs using a monopoly patent model.”

In this monopoly patent model, we might look then at the “effect of Bayh-Dole.” Actually, it won’t be the effect of Bayh-Dole, but rather the effect of university patent brokers unleashed by a misrepresentation of Bayh-Dole and unable to restrain themselves from the idea that every invention must become an institutional monopoly. The effect–what has happened in the past 35 years–is that university administrators have created patent monopolies around “early” research everywhere.

“Early” is a difficult term, though it is the one that is most often used. Often, the research invention isn’t “early” at all. It may be “mature” as far as a research tool goes. A patent held as a monopoly (and not licensed non-exclusively, even royalty-free) blocks the use of the research tool, especially in industry but also can create barriers in non-profit based research. Or an invention may be “rapidly developing” as part of a networked, non-market accumulation of research findings and technology. The patent breaks off that networked development of a cumulative technology with contributions by a number of participants. Instead, each institution grasps at its piece of monopoly, holding that piece for a speculative investment by someone who wants to keep that monopoly in a single piece–no non-exclusive licensing, certainly not royalty-free, such as to a standard.

Anywhere an invention can be used with only nominal development efforts, a patent held as a monopoly thwarts that use. Processes are often in this category, as are compositions of matter, as are research tools–protocols, specialty instruments, methods of analysis.

The effect, then, of Bayh-Dole’s unleashed patent brokers is to patent inventions to create monopolies before an invention has a chance to participate in the development of a cumulative technology platform, before the invention might be shared broadly with those who could readily use the invention in their own research (including company research), and before an invention might be considered for a standard or for use in commercial settings (in-house production uses, sale of commercial product uses). In these contexts “early” is a misnomer–the actual description is “blocking” or “disrupting” invention. The “early” invention, patented and held as a monopoly asset, blocks and disrupts the activity that otherwise would take place frequently with regard to published inventions made in university research settings.

We will not expect AUTM or Bio to conduct a survey to discover the absence of rapid use of research inventions as a result of “early” blocking and disrupting patents. We won’t even have the benefit of a control group where university patent brokers don’t patent “early” and often but rather allow inventions to move to the public domain, or if it is apparent that someone will immediately block development of a public technology platform, patent and license non-exclusively, but only to those willing to also license non-exclusively into the platform or commons. So we won’t be able to compare the control group with the “monopoly for commercialization via speculative investment” to see which approach results in the more interesting and useful public assets.

But we can see that every patent reported by a university is evidence of blocking and disrupting until the university demonstrates that it is not–that the patent is made available non-exclusively, royalty-free. Or that the patent has been licensed exclusively to a company but research uses (including in industry) are licensed non-exclusively, royalty-free. And that there is indeed timely introduction of a commercial product, the benefits of which are available to the public on reasonable terms.

Again–every university patent blocks and disrupts until there’s evidence that it doesn’t. That’s what a patent does–it excludes. Thus, metrics that recite the number of patents held and the number of new patents obtained are direct evidence of blocking and disrupting. And here’s the reality:

There is no evidence rebutting the idea that university patents almost always block and disrupt what otherwise would be typical, frequent research uses, development of cumulative technology in platforms and commons, contributions to standards, and rapid commercial uptake. The evidence, then, that is public is that university patenting programs–because tied to a patent monopoly approach to making money and fixated on licensing for “commercialization”–block the movement of research findings (discoveries, inventions) to those who would otherwise use or develop these findings.

There need be no deeper analysis to demonstrate that things that would have diffused have not. It’s rather a meaningless inquiry, since one would have to examine what hasn’t happened. Instead, one recognizes what a patent does, legally, and any account of movement of research findings in the context of university patents must necessarily be concerned with infringement or licensing, and since universities almost never successfully license inventions and discoveries, the inquiry will be one of infringement. Who would willingly participate in such an inquiry. “Well, yes, we routinely infringe university patents.” Right. Any deeper analysis will go nowhere. If universities can produce evidence that their patenting does not block or disrupt, they should do so. But they can’t–because they don’t have that evidence.

We might say, then, that another effect of “Bayh-Dole” has been to confirm a patent monopoly approach to the development of prescription drugs–reserving for financial speculation this key area concerning public health. Bayh-Dole has given legitimacy to the pharmaceutical industry’s effort to corner the market on all new potential therapeutics, to make the available first and only for monopoly patent investment. This is represented to the public and to policy makers as a really good thing, better than most anything else anyone could imagine. High drug prices, supported by monopoly patents on inventions made with public money, is a consequence of permitting Bayh-Dole exploiting university administrators to operate unchallenged by competing approaches to innovation.

This entry was posted in Commons, Metrics, Policy, Technology Transfer. Bookmark the permalink.