Joe Allen, the political coordinator behind the Bayh-Dole Act, is seen by some as an expert on the law, because, well, he helped assemble the sausage. Me, I have become something of an expert on Bayh-Dole because I have had to work under its benefits and burdens for two decades and decided I ought to get to the know the law if I was going to live that long in its house. Joe, political pundit. Me, licensing guru. Take your pick of perspectives.
I have already grumped about Joe Allen’s op/ed piece defending monopoly prices for drugs developed from compounds and methods that received federal support. I will follow some of the same ground again, but elaborate a bit with some discussion of the past history of federal patent policy and Cisplatin. Allen’s argument amounts to something rather simple and brazen–any exploitation for profit of a patent on a matter of public welfare is a public good. The exploitation is a “commercial product.” The profit motive necessarily produced it. No other motives could possibly produce something to save lives. But without a monopoly, the profit motive would not be quite enough and we would have nothing. Money would move into other things–sports betting, mortgages, technology stocks. No health-related commercial products. No therapies at all. No drugs but street drugs. We would all shrivel and die (but we’d go down with some mighty hallucinations). He doesn’t out and say this. He’s way smarter than that. But that’s where the argument heads.
Put it another way. The highest and most defensible public good is when the public good consumes itself for the wealth that can be wrung out of the public. Well, the public good eaters don’t put it that way. From their point of view, they are wringing wealth out of insurance companies and the U.S. government, and as for the insurance companies, the public good eaters are in a gold rush competition for profits–insurance company shareholders mano a mano with drug company shareholders. If there’s huge profit to be had, then there’s a fight over that huge profit worth having. And it’s worth pointing out that among the shareholders doing the most profit-battling are the big pension funds and endowments of the nonprofits–especially the universities. There, profit is a moral good because if the pension fund or endowment has enough capital to pay its future obligations, that’s a good thing. For pension funds and endowments, profit is the only reason to exist. Social good is having enough money to pay pensions and make bigger endowments. Best to go find companies that share this belief system about profits.
There is no fun in such pension fund and endowment investing, probably by law. One can’t be a technologist or visionary investor and run a pension fund. Given that an investor comes in early to provide funds at risk to get something started, pension funds are more like speculators–betting on the future value of shares by trading them with other speculators. The pension funds make their money, as Thomas Piketty points out, by having more money to bet with and thus can hire fund managers with better access to deals than the deals available to ordinary people or even to the ordinary endowment or pension fund. The rich have better means to get richer than have the less rich, and the rich thus get richer. One reason is that there is more money to be made in speculation (so long as there is a bubble) than in productivity, more money in betting on things against others betting on the same things than in doing work that produces something other than money. In Piketty’s equation, r > g. The rate of return on capital is greater than economic growth.
To put it simply, Bayh-Dole is a law that allows private speculators to eat the public good that might arise from university research when supported by public funding. This is cast as a good thing because the money produced from such speculation funds university research and education (public service is expressly excluded–see Thomas Jones’s testimony in supported of the failed government-wide IPA in 1978:
Now read Bayh-Dole’s restriction on the use of royalties by nonprofits (35 USC 202(c)(7)(C)):
a requirement that the balance of any royalties or income earned by the contractor with respect to subject inventions, after payment of expenses (including payments to inventors) incidental to the administration of subject inventions, be utilized for the support of scientific research or education;
Yarp, they really did it. It goes to show you don’t really know what a text means until you see what’s been left out. You can’t just read for what’s presented and find that it’s consistent and spelled nicely. I once when I was starting out submitted a draft license for legal review and it came back with some helpful suggestions on consistent use of terms. But the attorney had failed to notice that I’d left out a termination clause–a huge blunder. Luckily, I caught it. And I realized that general purpose university attorneys didn’t have an IP checklist. They read for what they saw, not for what they knew should be there. That’s why I added a practice rule–always have two expert readers for any draft deal. Add another for each order of magnitude of expected value or possible liability, starting with $10^5. So at a million dollars, there ought to be at least another four expert readers. There’s you, the drafter. There’s someone to read for consistency. There’s someone to read for legal sufficiency, there’s someone to read for the deal–for what’s left out. There’s someone to decide on behalf of the organization whether this now apparently perfectly formed deal is actually with a business partner, at this time, with these terms, that’s worth doing. Ah, the “no” reviewer.
But Bayh-Dole doesn’t get the benefit, often, of anyone reading for what’s been left out–especially left out deliberately. That’s a special bonus of reading deeper into Research Enterprise articles. Uncovering what’s left out, what’s distorted, what’s just untrue–that often takes more work and more words than simply accusing Bayh-Dole supporters that they are wrong or bogus or worse. There is still some good in Dolth-Vayher, for all the woe it has caused. Unmasked, it might just be a round-faced daddy worried about research innovation, gone over to the bureaukleptic empire in the hope of doing some good. Or, perhaps it is just a crappy law punched through a lame-duck Congress that cared more about giving a sweet parting gift to Senator Bayh than paying attention to what the patent brokers were scheming up–a franken version of the Kennedy patent policy, the IPA program, and patent broker fantasies, swathed in a rich, spicy sauce of monopoly, commercialization, and public good.
Or maybe Bayh-Dole was a ring of power, a Gyges kind of ring, a Sauron kind of ring, an absolute kind of powerful ring that has an urge to corrupt absolute. One might note the absence of enforcement provisions. Goof-ball march-in procedures. Goof-ball preference for small business licensing procedures. Goof-ball oversight of exclusive licensing. Secrecy for all utilization reporting. No enforcement of the Congressional patent policy that limits the scope of the property right rather than limits the license dealings. Appearance that the law appropriates inventors’ rights without notice or consideration and hands these rights to universities to do whatever with. And notice the other thing happily left out of Bayh-Dole that was in the IPA–no concern whatsoever whether a university has a “viable technology transfer program” (to use Thomas Jones’s phrase)–universities under Bayh-Dole get to take without any credible policy or practice or capability for taking. Neat, if you are into arbitrary, judgment-free absolutism–er, “uniform, government-wide policy.”
In the usual narrative, Bayh-Dol has been wildly successful. For folks deeply committed to the usual narrative for political reasons, there’s no need to speak freely and all:
Then roar’d awhile the fire, its sharpen’d point
On either side waved, and thus breathed at last:
“If I did think my answer were to one
Who ever could return unto the world,
This flame should rest unshaken. But since ne’er,
If true be told me, any from this depth
Has found his upward way, I answer thee,
Nor fear lest infamy record the words.
Yeah, none of that, yet.
Here’s the thing. Allen argues that the march-in procedures in Bayh-Dole don’t have anything to do with drug pricing. Just with availability of the drug. Even though “practical application” requires benefits available to the public on reasonable terms, the march-in procedures don’t include that as a basis for action. Federal agencies can intervene for nonuse or if “needs” are not “reasonably satisfied.” And Allen is right, march-in was set up to fail to protect the public from monopoly pricing created by university licensing of patents. And universities were given the “incentive” (because public service was not sufficient–or, ooh, because public service was excluded from Bayh-Dole as an incentive–I’m glad we already covered that) to share revenues with monopolists. That is Bayh-Dole binds university administrator interest in profits with that of monopoly companies. More monopoly, longer monopoly, higher prices, protected sales, more to share, and oh, the more the greater public good as measured by income (meaning economic impact, not institutional greed, of course–you have to know the right vocabulary to use).
To get at Allen’s argument–to put it into context rather than merely contradict it–will take a bit of effort. It is a sophisticated defense of the usual narrative, and one does not simply disagree with such things. They laugh at disagreement, they are built to eat disagreement for lunch. They exist to destroy disagreement with all the happy tools of debate. And they win, once they are in power, by creating the appearance of success, of excellence, of good intentions, of the bad of the past, and if those don’t work, then by creating doubt or lack of will to act. Here, at Research Enterprise, we are more into the mystery of the future, getting at our narratives of how things have actually worked (as best we can, to the limits of what’s primitive and what feels good), and how things might work if we went at it in decent fashion.
While it would be really neat if Bayh-Dole resulted in public well-being, that’s different from Bayh-Dole being a success on its own terms–as a monopoly-inducing, profit-seeking confirming, speculation-attracting engine to shunt important research results to private efforts to “commercialize” publicly funded research, all without public accountability or oversight. That is, take invention property with impunity, prevent immediate research and local internal uses, limit variations and improvements and applications made by early adopters, and attempt to gather up the private resources to make product to sell, and failing that, to recover the costs of the license by trading on the future value of the patent rights that might be had from trolling industry outright for infringement.
Allen presents various graphs to show the number of new drugs being introduced in the U.S. compared to the rest of the world. But he doesn’t show how many of these drugs were supported by federal money and “invented” at a university. The U.S. is the big speculative market for drugs. Any company sniffing profits will try to get drugs approved here. That there is a huge market for drugs here is clear. Allen is right. But making a connection to Bayh-Dole is more difficult. Especially since Bayh-Dole (with franken-like cleverness) restores much of the same elements of the IPA program–which had over 70 university and other nonprofit participants–so the transition point ought to be 1968 (if not 1954), not 1981. But in 1975, WARF had only, er, 0 patented drugs on the market under its HEW IPA. And in 1981, Bremer reported there were only 4 products in 96 inventions under the HEW IPA program generally. There’s some serious work for anyone trying to make a case that Bayh-Dole makes a speck of difference in university licensing practices for drugs.
Here’s the Allen punch:
The successful integration of public research institutions into the economy is based on the Bayh-Dole Act, which inserted the incentives of patent ownership into the government R&D system.
This is, of course, political fiction. The whole point of Bayh-Dole was a claim that the IPA program run from 1968 to 1978 had “integrated” “public research institutions” into “the economy.” And before that, the reason that the IPA program was restarted was that universities reported that their licensing efforts, while not so good, actually, were getting much better and in any event there had been a string of research findings that had created all sorts of things of value for the public–from toothpaste with fluoride to tomatoes that withstood the rigors of tomato picker machines to air bags in cars (not elsewhere) to treatments if not cures for childhood leukemia and Hodgkin’s lymphoma. All sorts of stuff.
If that stuff could come out pretty much on its own, via companies, via federal dedication to the public, via faculty publishing and consulting, via graduate students starting companies (hello, Hewlett and Packard), then (and this is the administrative brain at work) how much better all this would be if it were taken over by administrators, reduced to procedures to eliminate the uncertainties of personal judgment and the soil of having to negotiate based on equities and capabilities and purposes, so that administrators could get the lion’s share (administrators think of themselves as lions) of the money to be made from such inventions. Why, patent the stuff! and license the patents! create private monopolies, especially–because they pay more upfront! and make companies pay a share of their profits to university administrators (or, really, to patent brokers under contract with university administrators), and the administrators will pay a share (a small share) of that to the inventors, to induce them to participate in the scheme. Bill this as the only way that university research will ever get used by the public. Only by monopolies will research be commercialized. Universities don’t want to create such monopolies, according to the usual narrative, but they have to. It’s the only way. Then implement university policies on research and patents and conflict of interest to ensure that this is the only way. Now get federal law to confirm the scheme. Woohoo. It took a couple of decades, but the patent brokers carried the day. No wonder the elation among them after Bayh-Dole passed. Frabjous, simply frabjous.
But beyond the scheme, look at the pattern of emphasis. Go back awhile and one finds that for, say, treating cancer, there were competing arguments–a metabolic argument, a genetic argument, a mutation argument. Treat cancer with surgery, with radiation, with diet, with immune system. Treat cancer with prevention. Treat cancer by understanding the biophysical systems that create the conditions for it and fail to respond to it. Why drugs? The commercial success of drugs overshadows other possible modes of intervention. The profit in drugs lies in a statistical argument for efficacy, not generally in a scientific one about the actual mechanism of action in specific humans. It’s a plausible speculation on benefits that depends on the market for remedies continuing to exist. May as well be Dr. Pepper. Make a plausible claim. Desperate folks will seek you out. In any event, that the U.S. leads the world in approving drugs does not of itself mean that drugs are the thing that best favors health–though they are sure nice to have around, given that so little is being done to make them unnecessary!
The point being, university research was happily “integrated” into the American economy well before Bayh-Dole. The thing that made Bayh-Dole even remotely plausible was that universities were doing (claimed to be doing, at least) so much with stuff that did not have federal funding. Of course, for the federal funding stuff, universities weren’t doing things because they didn’t have to–the federal government, if it obtained patents, released inventions back for anyone to use, and only with a small portion found reason to allow a contractor or licensee to create a monopoly, and then often for a limited time, and even then with the expectation that the contractor or licensee would do a better job making the invention available than the government would do on its own. Things were working.
What wasn’t working was the university patent broker scheme of getting access to as many inventions as possible to make their scheme look better than the alternatives. The patent brokers were competing with the federal government’s policy of open release, just as they were competing with company sponsors expecting assignment of inventions supported by their funding of university projects. The point of refusing companies to take title to the inventions they sponsored was–arguably–so that the invention would be more broadly available rather than under a monopoly controlled by the company sponsor. This idea morphed into the argument that a company sponsor might not be best suited to “developing” an invention for commercial productization and so it should fall to university experts (or their patent brokers) to do this. And that point slipped into the idea that the university should license to create the most favorable monopoly to university interests. And from there we are back to the idea that a university takes patent positions for the money, with public benefit a side effect, rather than the other way around.
Allen’s next sentence:
Not a single new drug had been developed from NIH funded research under the patent destroying policies preceding Bayh-Dole.
Allen must be referring to the Kennedy patent policy and the HEW IPA program, but if so, he’s wrong. Those policies did not destroy patents. They limited how private contractors could acquire patents to stuff that the federal government had commissioned. The federal government used the same form of argument that university administrators now use to claim title to inventions made with federal support.
Feds: we paid for this research, the public should have the benefit
Admins: we paid for this research, the university should have the benefit
Except in the case of the Admins, it’s really:
Truth: we paid for this research with money we received from the government, plus we were reimbursed for our administrative and facilities costs in releasing our personnel from their duties to the university to do this work, which they volunteered to do knowing it was for the public welfare, but even so, the university should benefit so we should own it.
But that’s too accurate to be useful. University administrators just claim that even though the federal government has paid for the work leading to the invention, the university should get the monopoly. The good thing, they claim, about Bayh-Dole, is that it prevents the federal government from making inventions available to the public generally at no cost and the vast majority of the time without licensing delays (that’s changed somewhat with Stevenson-Wydler–for another time). Please understand: Bayh-Dole prevents university inventors from gaining access to their own inventions as members of the public. Bayh-Dole prevents principal investigators from gaining access to inventions made by their research staff outside their own institution–if they leave for another university, if they join a company, if they start a company. Under the federal patent policy before Bayh-Dole, principal investigators mostly had unrestricted access. Now they don’t. It’s a good thing, too, that they don’t, according to folks like Joe Allen. But every unlicensed invention, every exclusively licensed invention that has not made it to practical application, every commercial product that does not implement every possible variation or function under claims of an exclusively licensed patent is a matter of nonuse enforced by monopoly.
The number of claims that are actually practiced in an exclusively licensed patent is often–almost always–remarkably small compared to the expanse of the full set of claims. Unless an exclusive license’s scope extends only to the product actually produced, everything else in that scope is dedicated to nonuse–a violation of Bayh-Dole’s Congressional patent policy at 35 USC 200, but no one seems to care. And Joe Allen’s rhetoric works to keep it that way.
Allen claims no drugs with federal funding made it to market under pre-Bayh-Dole policies. That’s simply not true. We should preface things a bit. Not many drugs made it to market with federal funding, period. Not all that many drugs made it to market anyway. From 1961 to 1963–the year of Kennedy’s patent policy but too soon for that policy to have had any effect–the number of approved drugs stepped down from over 200 a year to more like 60, bouncing from 20s and 30s to 80s and 90s per year. The issue wasn’t federal patent policy. Something else stepped things down. Federal patent policy might have played a tiny role in keeping things stepped down, but then, after the IPA program started, things didn’t step back up. Before Bayh-Dole, in 1979 to 1981–summer of 1981, when Bayh-Dole came into effect–drug approvals stepped up to about 100 or so per year, and have remained at that level. Drug approvals went down before federal patent policy changed, and recovered only a little before Bayh-Dole changed federal patent policy again. Did the prospect of Bayh-Dole have a beneficial effect five years before it came into effect? Ah, a mystery for the advocates of Bayh-Dole.
It would appear the changes in FDA drug approval numbers point to two candidate causes: one, changes in regulatory practices; two a patent gold rush to control the amazing discoveries coming from federal and university research labs.
Consider (from the FDA’s timeline website)
1962
Thalidomide, a new sleeping pill, is found to have caused birth defects in thousands of babies born in western Europe. News reports on the role of Dr. Frances Kelsey, FDA medical officer, in keeping the drug off the U.S. market, arouse public support for stronger drug regulation.Kefauver-Harris Drug Amendments passed to ensure drug efficacy and greater drug safety. For the first time, drug manufacturers are required to prove to FDA the effectiveness of their products before marketing them. The new law also exempts from the Delaney proviso animal drugs and animal feed additives shown to induce cancer but which leave no detectable levels of residue in the human food supply.
Consumer Bill of Rights is proclaimed by President John F. Kennedy in a message to Congress. Included are the right to safety, the right to be informed, the right to choose, and the right to be heard.
Hmmm. That sure looks like regulatory policy not patent policy coming into play. A big disaster with thalidomide. Not as big as the aspirin disaster, but much harder to suppress. If these are “patent destroying” policies, well, that would appear to be a good thing.
1973
The U.S. Supreme Court upholds the 1962 drug effectiveness law and endorses FDA action to control entire classes of products by regulations rather than to rely only on time-consuming litigation.
More hmmm. Looks like a drug company (CIBA) didn’t like that regulation. One might say that while a patent right gives a drug company a monopoly to exclude others, the FDA has the right to exclude the drug company’s public practice of the invention it controls. Think of the FDA right as a super-right that prevents the operation of the profit-making core of a patent monopoly. A patent has the attributes of private property. But the underlying invention is also subject to regulation, when its use is a matter of public health. All the more so when those proposing the use have a profit motive that must dominate their corporate thinking.
1976
Vitamins and Minerals Amendments (“Proxmire Amendments”) stop FDA from establishing standards limiting potency of vitamins and minerals in food supplements or regulating them as drugs based solely on potency.
According to one commentary:
Finally, in a few instances, legislation has weakened government’s authority over food and drug products. For example, the 1976 Vitamins and Minerals Amendments precluded the FDA from establishing standards that limited the potency of vitamins and minerals added to foods.
That would not account for the uptick in drug approvals in the late 1970s, but does show a move to push back on FDA regulation of some things–limiting the scope of what’s a “drug” and what’s “diet.”
But consider the rise of the Japanese pharmaceutical industry in the 1970s. By the early 1980s, a National Academy of Engineering report was worrying that the U.S. might lose its global leadership in drug production to Japan, according to Henry Grabowski (“Innovation and International Competitiveness in Pharmaceuticals” in Evolving Technology and Market Structure: Studies in Schumpterian Economics). The U.S. had established itself as the clear global leader in drugs by the early 1960s. Consider Grabowski’s summary (without needing Table 2 US 353, France 271, West Germany 201, Japan 155):
So the U.S. pharma industry was doing globally fine under those “patent destroying” federal policies–it was the world leader already. Bayh-Dole didn’t give it that position. It doesn’t appear to have done anything other than destroy the federal competition for alternative ways to develop drugs and make them available for public use. That is, the monopoly approach does more than protect a product from competition so price is set by the company not the market–it prevents anyone from using the drug outside of product formulations, prevents any other variants of the drug’s active compounds from being developed–even if more effective, more specific, having fewer side effects, easier to administer, cheaper to produce, for other conditions–whatever. It’s a double whammy monopoly when the patent covers the compound for any biological use. The monopoly permits development of a drug without proximate competition–not the same drug, not a variant, not any other use of the compound for biological purposes. One wipes out all other uses in favor of the use–if ever–designed for maximum profit.
So perhaps–it’s only a hypothesis–the FDA drug approval uptick in the late 1970s is the result of international competition for the U.S. market. Or, it may be a blip from the 1972 transfer of 114 radioactive drug applications from the AEC to the FDA and 94 applications for amphetamines to treat obesity–that year there were 448 drug applications in total (99 in 1971, 146 in 1973). About five years later, one might expect to see these applications becoming approved. An uptick. Not anticipation of Bayh-Dole. Seems plausible. Maybe Joe Allen has a better explanation. Somehow I don’t think he much cares.
And now about this idea that there were no NIH-funded drugs approved before Bayh-Dole. What is Joe Allen smoking? Something not approved by the FDA, I fear. I will mention a few drugs to show that Allen is simply wrong.
Methotrexate–identified by Sidney Farber at Harvard as a treatment for a form of cancer, later adapted as a treatment for rheumatoid arthritis. In use in the late 1940s, further developments in the 1950s.
MOPP–a drug cocktail that cures Hodgkin’s lymphoma. In use in the 1960s, developed by researchers at NCI. So, not developed at a university, but surely with federal funding.
Cisplatin–a platinum molecule that cures some cancers, identified by Barnett Rosenberg at Michigan State with federal funding, patents assigned to Research Corporation, licensed exclusively to Bristol-Myers, approved 1978.
There, three instances. Joe Allen might modify his claim–“only a few” drugs were approved from university research supported by federal funds prior to Bayh-Dole. But how many since? Allen won’t give the numbers. Under the IPA program, 1968-1981, according to Howard Bremer’s records, only 4 products made it to market out of 96 inventions claimed by universities. I don’t know what those products were, but they were all funded by HEW, so that’s the max. And the IPA program was essentially Bayh-Dole with oversight, not that the oversight did much good. How are we doing now? It’s tough to say, since the data on utilization are secret. But Allen paints a dire picture:
No one is going to spend billions of dollars and more than a decade of effort turning early stage inventions into new drugs or fund a life science startup company without strong patent protection. Yet the patent system and Bayh-Dole are precisely what the critics seek to undermine. If they succeed we can expect the number of drugs developed here to decline to the levels of the rest of the world. What that would mean for protecting human health is not pleasant to imagine. But that doesn’t faze the critics.
Yet people were creating such companies and developing useful drugs with “patent protection” that was, apparently, good enough. Consider methotrexate, MTX. The compound wasn’t patented. Applications of it were. But this left the compound free, out in the clear, to be adapted from cancer use to rheumatoid arthritis uses. A patent on the compound would have prevented such lateral shifts. Who would have thought to try a compound identified for cancer to treat arthritis? Such a shift reflects the excitement at the time (speculators would call it a gold rush). Science–especially medicinal chemistry–was opening up many compounds to consider. The federal government started screening programs to look for bioactivity–and initially pharma companies sent compounds to the federal labs for screening. That is, the companies were looking for new possible uses (er, markets). It is hard to explain how many possible chemical variants there are. It is not just a lot, not a gazillion. It’s more than that. No brute force method of screening can possibly begin to reduce the total. Screening is a matter of luck, intuition, and maybe some science.
Then they figured out screening was important enough that they would do their own, and offered to do the screening for university chemists for free-absolutely no charge–with the hope that if they found something, they would get an exclusive license to any patent. It’s just that a bunch of that chemistry work was funded by NIH, and NIH didn’t see the reason why pharma companies should get monopolies on compounds that NIH had funded the work to develop. NIH insisted that it should own any patents on such compounds, that screening didn’t justify a monopoly on the compounds, and the NIH would make them available to everyone–you know, like methotrexate and the like. Pharma companies objected, and then boycotted NIH’s medicinal chemistry program. Perhaps it is this boycott that Joe Allen refers to as a “patent destroying” patent policy.
Allen might be clearer, if that’s the case. Certainly pharma objected to not getting a monopoly on a compound. Their monopolies would be limited to the specific applications and formulations they pursued that were non-obvious over the prior art (including the compound itself). They could find new ways to synthesize, new analogues, new formulations, new dosage requirements, new combinations with other drugs and therapies–whatever, just not get a full monopoly on the compound. But the objection to the monopoly wasn’t the big problem. The big problem, according to the Harbridge House report, was the possibility of “contamination” of proprietary pharma research with NIH compounds and claims. When NIH funded work in medicinal chemistry, it wanted not only the foreground patent rights but any background rights that would be necessary to use the candidate compound. What is the point of getting foreground rights, if someone is going to hold out on you? The point is public availability of the invention, not of a patent right on the invention, not on a slice of the invention served up as a monopoly product.
Look at warfarin for a moment–first characterized in the 1940s by agricultural chemists at the University of Wisconsin trying to find a way to make sweet clover hay (itself a product introduction promoted by the US government) taste less bitter–and then trying to track down the compound that was making cattle bleed to death when they ate moldy sweet clover hay. Sweet-smelling hay death. What did the chemists do? They synthesized 150 analogue chemicals–variations on the theme, and screened them for bioactivity. Anticoagulants were a big thing–think heparin. But the compound variant they chose to introduce as a blood thinner for humans–didn’t work so well. Meanwhile, they chose another of the variants and created a rat poison. This was called warfarin. After a series of accidents, it turned out warfarin was a better human blood thinner than their first candidate compound. The applications got patented, and WARF made a pile of money in royalties. But the 150 analogue chemicals were not all patented–just the ones that became products.
It is important to understand the context of discovery. A research invention–like finding a compound–is not made in isolation. There may be a thousand similar compounds nearby–analogues that may have remarkably different properties. There may be many possible applications. If one files a patent that’s broad–that claims everything possible–that sure sounds like the road to wealth. Everyone will have to pass through your patent to do anything with the invention, and with a capable patent attorney, one can claim any variation, any functional equivalent, regardless of application. This is a powerful monopoly. From an investor’s point of view (or speculator’s, or university administrator’s), it’s pretty sweet to gobble up all that opportunity and not have to digest most of it. Just shave off some bit, develop it, sell it. That may be enough. A partially digested invention. The rest of the rights–all those broad claims–those need not be used. The payoff is the money made less the cost of the patent plus the cost of development. One tiny part of the invention-in-the-world covered by the patent claims becomes a product. And the rest of the patent is waste, is scorched earth, is exclusion of others.
Compare with methotrexate, with warfarin. An early monopoly patent cuts off access to all variations. Cuts off the incentives and opportunities of others to work without having to pay, or risk infringement claims, or face delays. While an early, broad monopoly may be great for speculation–stake a claim to all the gold of Northern California if you can–it is not so good for research, further discovery, creation of variations, lateral shifts in application, serendipity. In fact, as the certainty of broad patent ownership goes up, the likelihood of serendipity goes down. Think about that in a world that has combinations beyond imagination–that serendipity is way, way more important as a tool of discovery than is management efficiency or investor spread sheets on ROI.
There were problems with NIH patent policy in the 1960s and 1970s. But the idea that early stage discoveries ought to be generally open, that the public domain ought to be broadened–that wasn’t such a dumb idea. If it was “patent destroying” there was a good reason for it. Expand the public domain, limit the role of patents to developing specific applications, not to dominating the entire area of research. No wonder the cost of new drug development skyrocketed. Here’s an account of Lorne Brandes’s efforts to get a drug candidate through clinical trials on his own. The compound failed at stage 3–but at way less cost that the pharma industry makes out. His compound is a variation tamoxifen–again, made available because there was no monopoly on all such variations (and no doubt there are thousands more to be had). Or consider Barnett Rosenberg’s comments about cisplatin development:
At the time, Rosenberg said he was “euphoric” over the life-saving capabilities of cisplatin. However, he also said it was “disturbing” that a discovery he had made more than 25 years earlier remained the gold standard for cancer treatment.
“For years I’ve been saying this is the first platinum-based drug we discovered,” he said. “It can’t possibly be the best one. It’s disappointing that the scientific community has not been able to find better ones.”
Well, let’s see here. First there’s a patent on the compound and an exclusive license, then MSU and Research Corporation fight to get a second set of claims approved that would extend the life of the patent (claims later invalidated). Consider the request by Research Corporation to extend the term of exclusivity for its exclusive license with Bristol-Myers. Research Corporation reports that Bristol-Myers had invested $46 million in “developing uses” for cisplatin. There’s no indication of how much of this amount was for developing the compound into a drug and how much was for marketing and sales. Although $46 million might sound like a large amount, keep in mind that Bristol-Myers, by the end of cisplatin’s patent run, was making about $100 million per year on the drug. Over the course of its patent, cisplatin earned Bristol-Myers over $1 billion.
Research Corporation’s argument for extending the term of exclusivity does not depend on Bristol-Myers recovering its $46 million investment or even its proposed additional $28 million of work. Rather, the argument for monopoly is tied to the claim that Bristol-Myers will stop working on the compound if it loses its monopoly, and no one else will work on the compound, then, either.
By developing, here, Research Corporation means “inventing” and “obtaining additional patents.” The “new uses” would mean “new monopolies” tied to payments from people (or on behalf of) people who would otherwise die.
The argument Research Corporation makes here is that without being able to extend its monopoly on the original compound, even with the prospect of dying people who would receive care, Bristol-Myers would not invest even a few months’ profits to develop new patents that would be free of Research Corporation’s control. This is amazing. Clearly, this is an argument for Research Corporation’s interests, not Bristol-Myers’s interests or the public interest in lower drug prices. While the original manufacturer’s interest in the licensed compound might decline, why would its interest in new applications, if those required significant “development”? Perhaps it goes without saying that Bristol-Myers would never consider developing any drug to save lives if that drug did not have a profile for maximum profits. That would be consistent with always seeking shareholder value. But such a position–were Bristol-Myers-Squibb to take it–is not consistent with the public purpose in developing drugs to save lives.
If the federal government created companies to develop such drugs as generics, with only such profit motive as necessary to recover the costs of development and maintain high quality control over manufacturing, it would appear that the public could obtain drugs at a fraction of the prices that monopoly drug producers charge. When the drug industry reports that it costs north of $1 billion to develop a new drug, they include in that figure the funding for a bunch of failed efforts, without indicating the nature of those “failures.” Winners pay for losers–but clearly there is plenty of profit as well to spread around.
Again, Research Corporation makes it clear that the development required is “research” and that Bristol-Myers won’t conduct that research without a monopoly on the original compound.
The case for monopoly drug prices is important. According to one account, when competition enters a market, drug income drops significantly:
Sales losses for branded products are always significant when generic competition is set in motion. Losses for branded products can range from 80% to 90%.
Some of this drop has to do with generic drugs taking market share. But the reason the generics take market share is that they sell from significantly reduced prices. The patent creates a monopoly, creates a bubble, requires life-saving need to pass through corporate hands, creates substantial profits, rewards owners of stock, and creates the opportunity to extend the monopoly with more research. Here is a Department of Justice report on the effects of generic drug competition:
Competition between brand-name and generic pharmaceutical manufacturers provides consumers enormous savings. Pharmaceutical industry studies indicate that the first generic competitor enters the market at a price that averages approximately 80 percent of the brand-name counterpart, and gains substantial share from the brand-name product in a short period of time. Subsequent generic firms may enter at even lower prices—often discounted 80 percent or more off the price of the brand-name drug—and prompt the earlier generic entrants to reduce their prices. Thus, as the number of generics increases, prices to consumers decrease even further. As a result of price competition, as well as the policies of public and private health plans and state laws that encourage the use of generic drugs, generic sellers typically capture approximately 90 percent of brand-name sales within the first full year after a generic product launches.
There’s a point embedded in there about the cost to move through the FDA regulatory process. That’s expensive. And it makes sense that there should be some period of time that a company that does pay the costs to get through the process be able to recover those costs. And no one I know argues that a company should have to sell its drugs below cost or with no profit whatsoever after that. Even among generic drug manufacturers there is still a reasonable profit. It’s just not a monopoly income. It’s just the profit arising from pricing that’s 80% less than that of the monopoly price–and there’s still profit. The business is still viable. But perhaps, too, the generic business does not care about new research, new monopolies. Perhaps it feeds only off what enters the public domain, twenty years afterwards.
All this raises some policy questions. What is the length of time from first commercial sale of a drug to the point at which that drug’s development and regulatory costs have been met? Not the losses from everything that has failed. Not including more research to explore new indications and new variations. What’s the break-even time, on average?
Further, what’s a reasonable return on investment? That is, the amount that’s profit? In the very old days, a stock corporation might be formed to cover a venture–send a fleet to get spices, hope some get back and sell off their cargo, pocket the profit less expenses. Now corporations may live forever, skipping from project to project, product to product. The initial investors have long since sold off their shares. The shareholders are buyers and sellers of stock based on their anticipation of the fortunes of the company. What sort of dividends should these speculators expect? The share price doesn’t much matter to the operations of a company unless it has, Enron-style, pledged shares against debt or something silly like that. If developing drugs to save lives is just a speculative lark by investors looking for the biggest upside possible, and all companies in the pharma business would fail without them, then we really should look for a different way to develop drugs.
If generic–that is, non-monopoly drugs–are only 20% the cost of monopoly drugs, then the policy matter is to deal with the cost of the research and development for a drug and the recovery of those costs from sales (if not from some other source). The rest is a matter of dealing with profit–whether that profit goes to more research or to pay out salaries or to be sent to shareholders.
For innovation from research, however, the most important policy question is whether there’s any public benefit served by patents on fundamental inventions–such as compounds and their basic methods of acting on human physiology. There appears to be an argument that drug development was much less expensive, and more diverse, with more outcomes of interest, when the patenting shifted later–that is, a discovery is made, a shared platform or library or inventory of characterizations, methods of production, variations, and applications are explored. This is all done as research–competitive (for priority) and collaborative (sharing of access, data, technique, and outcomes). But without monopoly. Then there’s a time for competition without free-riders. An industry might move the platform collectively to a standard–that is, ten companies cross license their patents and share the cost to get a drug in a base form approved by the FDA–and then the companies compete for the market based on pricing, availability, function, formulation, support, and the like. It’s not impossible. Other industries do this. There’s still profit, but it is contained by competition. There is still development, but no one takes over the common platform or tools.
In this approach, no one company has to play the monopolist from start to finish. Competition begins after there’s a common platform, after there is use established. The research that’s competitive then is in specializations that make production more efficient, make outcomes better, extend the applications of the compound, and the like. There, take a monopoly position. In standards talk, such positions are “non-essential” claims–valuable claims but not ones that block access to the core platform technology. In industries that use these methods, costs go down over time for the consumer. Profits stay nice. Speculators still have a good time–since they are betting against each other and don’t have much to do with companies unless the companies, unlike sports teams, start to dabble in the betting pools as a source of funding. Then it can get bad, because the share price becomes essential to operations of the company, rather than “merely” representing the fortunes of capital betting on the future of the company.
For Bristol-Myers-Squibb and cisplatin (marketed as Platinol), the monopoly proved too enticing. In 2003, the Federal Trade Commission accused the company of anticompetitive behaviors:
Over the course of the past decade, BMS engaged in a series of anticompetitive acts across the BuSpar, Taxol, and Platinol product lines. Among other things, BMS: paid a would-be generic competitor millions of dollars to abandon its patent challenge and agree to withhold competition until patent expiry; misled the United States Food and Drug Administration (“FDA”) about the scope, validity, and enforceability of its patents and abused FDA regulations to block generic entry; breached its duty of candor and good faith before the Patent and Trademark Office (“PTO”) while pursuing patent applications purportedly related to the branded BMS products; and filed objectively baseless patent infringement lawsuits in federal court against would-be generic competitors. BMS’s pattern of conduct evidences a scheme to abuse competitive and 2 government processes for the purpose of maintaining its branded drug monopolies. As a result of these anticompetitive acts, BMS thwarted low-cost generic competition to these monopolies for many months or years, forcing consumers to overpay by hundreds of millions of dollars for vital prescription drug products.
Bristol-Myers-Squibb then settled, paying a $670 million fine. Given the billions in earnings ($2 billion per year, according to the Wall Street Journal) on these products–cisplatin (Platinol) was making $100M per year)– the fine represents maybe a 5% hit on the company’s profits. Hardly enough to deter anyone from trying the same thing, if they can.
But behind all this, there were university patents on an invention made with federal support, that got licensed exclusively and for a long time. Michigan State made good money–$160 million (and even sued Research Corporation to try to get more). Research Corporation made even better money (its usual deal was along the lines of costs plus 40%–I have no idea what the MSU-RC deal was–it dated from the 1950s). And of course, Bristol-Myers made out wonderfully, at the expense of the public, the government, and insurance companies. Research Corporation sought to extend its monopoly and argued that the government should permit it to extend Bristol-Myers’s monopoly. And Bristol-Myers gamed the system in all sorts of ways to extend its monopoly yet further. When there’s $2 billion per year coming in, many things appear reasonable that otherwise would not get considered.
Joe Allen is right, however, that price controls are not something built into Bayh-Dole’s march-in procedures. They could be–but that’s going to be difficult to implement. It is certainly not the NIH director’s role to monitor drug prices and intervene at will. But Bayh-Dole does require competition, and it does it in multiple ways, through its statement of Congressional policy and objective (35 USC 200), which is made a part of patent law and thus constrains the scope of the private monopoly available for patents on subject inventions:
patent rights must promote use
thus, they cannot be used to block “infringers” if there is no use
patent rights must promote free competition and enterprise
thus, they cannot be assigned or licensed exclusively to restrict competition
patent rights must not interfere with research and discovery
thus, they cannot be used to block such activities even in favor of products
patent rights must not support nonuse
thus, claims that do not result in practical application must go non-exclusive
These are all limitations on the scope of patent property rights in subject inventions, in inventions that the federal government has been willing to support based on the prospect of public benefit. While these patent property rights are limited compared to those available for an ordinary patentable invention, it makes sense that if we intend to expand the public domain with new scientific knowledge and we aim to make results available for use–and exploitation, and profit–then those managing patent rights have an obligation to work within these constraints. There is not maximum profit here–but this is not a place to look for maximum profit, nor is selling water during a drought, or food in a time of shortfall. We call that gouging. Nor do we want the market for innovation to be forestalled or regrated or engrossed or trolled by speculators.
Neither university faculty nor university administrators require a profit incentive to do work in the public interest. If they want to get rich, work outside federal support. If what they are doing is so very profitable, capital markets will find it. That’s what capital markets do. Find things to buy into that will create wealth faster than doing something productive. It’s okay, it’s what we have. A happy ecosystem has its shares of predators, scavangers, and pikas. The policy issue is not a matter of attack on patents–as Joe Allen makes it appear–but rather a recognition of the limited scope given by Bayh-Dole to patents on subject inventions. That’s where to turn to limit the monopolies available to drug companies working with inventions made with federal support. It may well be that the NIH and FDA will have to work together to create a regulatory system for non-profit drug development and approval. That indeed would be an innovation worth developing, as it would provide yet another source of competition for companies operating on a profit motive in areas of public suffering. There, we don’t need monopolies, nor competition, nor commercialization–we need compassion and commitment, and these, too, can produce quality outcomes and even financial rewards.
Bayh-Dole already attacked the patent system–the part used to create a federal research and development commons. The problem is that Bayh-Dole has never been enforced. For thirty years, it was grossly misrepresented as a vesting statute. It gobbled up all meaningful research outputs from universities. It has reduced the public domain, created bureaucratic bottlenecks, allowed the formation of anticompetitive private monopolies, and has positioned these monopolies too soon the development of anything new, with fragmented rather than collective interests, with profits before service.
It’s time to restore some balance. Keep Bayh-Dole as it is? Okay. Start by enforcing the law. Change Bayh-Dole? Reopen university and federal research to the public domain, to research and user commons and communities, and let commercial services and products ride on top of these commons–as we have, say, with the internet and telecommunications. We can do the same thing with chemicals that might become drugs. The existing pharma model doesn’t like it–but the existing pharma model does not need federally funded research. It has plenty of money. It can fund its own research. If that research is but a tiny portion of its overall expenditures, then it’s a tiny hit on profits, too. Let the federal effort alone. Leave the university effort alone. There are good reasons to explore an interface between the two systems–but there’s no good reason to sell out public research to corporate monopolies, or destroy the research commons and public use commons merely to prop up winner-take-all approaches that delay the common good for two decades (eight years of profits sounds plenty) and then leave products to waste because they lacked common support all those years and fail when the monopolist loses interest.
There are better ways. Bayh-Dole can play a role. But to get there, those advocating so ardently for the law need to start be more thoughtful about the law, and more careful about the history of innovation, and more reflective on whose interests they have chosen to serve.