Federal policy on inventions made in federally supported research starts in a big way with Federal Security Agency Order 110-1, dated December 30, 1952. Norman Latker, patent counsel for the NIH, in 1978 testimony before Senator Nelson’s subcommittee, identified Order 110-1 as the basis for the Institutional Patent Agreement program, which in turn Latker later identified as the basis for his drafting of Bayh-Dole. Not that Latker necessarily complied with FSA 110-1, but rather found in its language sufficient ambiguity to, with a bit of spin, get the IPA program back up and going.
The Federal Security Agency was created in 1939 to bring together various federal agency activities pertaining to social welfare and public health, including what are now the NIH and Department of Education. The FDA was added in 1940, various bits were tossed off in reorganizations, until in 1953, the FSA was re-organized away into the Department of Health, Education, and Welfare. So order 110-1 was a beginning of federal invention policy created by an agency about to end.
Let’s have a look. We will take it slowly, so we can see just what Latker might have seen.
The Agency, as a matter of policy, takes the position that the results of research supported by grants of public moneys should be utilized in the manner which would best serve the public interest.
We stop here to wonder at this simple sentence that isn’t simple at all. The “as a matter of policy” qualification to what follows is odd. What other matters might be in play besides policy? Same for “takes the position that.” More oddnesses! The sense of these two constructions is that there are other matters and positions, but that the FSA has for whatever reason adopted just this one policy position on the matter of research results. Rather than stating what might be taken as obvious, the FSA here makes a restriction–that results of publicly funded health research should be used to best serve the public interest, even though other uses might also be possible.
Research results should be “utilized in the manner which would best serve the public interest.” Sounds nice, worth bouncing over, but if you think about it, this gets very strange! Not at all obvious. The policy insists that research results be used. And not only serve the public interest, but best serve it. And what is the public interest? And of the various things that might serve the public interest, which of those are the best? Spoiler: 110-1 doesn’t speak to the public interest but rather stipulates how results will be managed–and we are left to make the connection that this management of results does–er, is the position of the FSA–best serve the public interest.
Here we have the FSA, early in the expansion of federal funding for research, authorized to fund biomedical research. Of course, we might think, the results of this research ought to be utilized. But then things get funny-strange. What if the results are not worth using? What if the results are wrong, misstated, can’t be replicated, involve such complicated and expensive equipment that no one else can afford to use those results or simply doesn’t have access to the equipment and so cannot use them? What if the results lack clinical relevance? What if there are better technologies available elsewhere, so the results of the federal research are obsolete before they are ever announced? You get the idea. There are all sorts of research results–and some are defective, stupid, obsolete, impossible in practice, ineffective, wildly more expensive than alternatives, unable to scale, crappy to maintain, incompatible with everything they would have to work with to be truly useful. The push for the FSA to get its results used demands context–we might think, “use the results that ought to be used, and not other results that would be a waste of time and money for anyone to use because they are not clinically relevant.”
For patenting, we can also throw in inventions that will take more than the term of a patent to “develop” into anything worthwhile, and inventions that will be used, if they are, before a patent can even issue and in three years or so will be obsolete–and for which any delay whatsoever in uptake ruins the opportunity, even a delay to prepare a patent application for a patent that will never be relevant. Merely that something is an invention has–really–nothing to do with whether or how that something gets used, if at all. Inventing, when one is in the research game, is easy–and mostly meaningless. But here we have a federal agency out with a policy that each result–and necessarily, then, each invention that is also a result–should be used. Mind-bending time for bureaucrats. It sounds good only if one is arm’s length from reality and so can see no conflict of interest between policy and practice.
FSA order 110-1 is written to make it appear that anything that results from federally supported research is by its nature, by default, a public good if used. Federally supported research, like any other research, produces all sorts of nothingness, false leads, faulty analysis, overclaimed significance. We might go so far to propose–this may shock–that federally supported health research, when disassociated from clinical effort, is bound to be mostly useless. “Bench to bedside” might be the mantra of NIH research, but the whole concept of starting at the “bench” is sketchy, and it may well be expensive and “high risk” to try to find any available “bedside” where the research might eventually be used.
The use of research results ought not to be a concern of policy. The non-use of results, by contrast, might be of some concern–if the results would be worth using. Suppression of use. Monopolization of use. How do such things work in the context of “public interest”? And there’s the rub–a federal agency might fund “research” but that doesn’t mean the research produces meaningful results. Here’s Stanford professor John Ioannidis:
Overall, not only are most research findings false, but, furthermore, most of the true findings are not useful. Medical interventions should and can result in huge human benefit. It makes no sense to perform clinical research without ensuring clinical utility. Reform and improvement are overdue.
Ioannidis cites a 2014 article by Macleod, et al. in Lancet that estimates 85% of the spending on biomedical research is waste. If that’s the case, then one might expect a policy statement directed at trying to find ways to fund research that would have useful results, not in trying to make something of the results one has got, that in generally mostly are not useful. One can construct policy “incentives” to attract money or talent to such results, but what then is an agency doing? Distracting people and resources from useful results elsewhere to take a flyer on results produced in an agency’s research programs?
Put another way, following the Kreeger’s 1947 Attorney General’s report, if a research result is useful in biomedical research, then why isn’t the federal agency using that result itself? Why would it have to recruit others to use that result using policy inducements? We might say, then, that while a federal agency hopes that the results of the research it funds will be, say, true (that’s a stretch, apparently) and useful if true, it’s another thing to write a policy regarding inventions to try to induce use generally, without regard for either truth (repeatability, consistent not diminishing effect, and the like) or utility (true, but doesn’t scale, doesn’t translate to human physiology, cannot be used by others, data or protocols not available, small sample size, limitations of clinical trials contexts, and so on).
In a sense, then, focusing on inventions–and patent rights–arising from research avoids the problem of research design; access to methods, materials, and data; and transfer of ideas and qualifications regarding results–a kind of scientific candor, a Feynman-class scientific honesty:
I’m talking about a specific, extra type of integrity that is not lying, but bending over backwards to show how you’re maybe wrong, that you ought to have when acting as a scientist. And this is our responsibility as scientists, certainly to other scientists, and I think to laymen.
Perhaps the federal policy ought to be directed to scientific honesty–an “extra type of integrity.” Instead of worrying that the “best companies” in biomedical research might not want to accept federal money to do research (as the NIH apparently did worry), a federal agency might instead worry about attracting scientists with extra integrity. Is there a problem in doing so? Is medical science worse off if the scientists that don’t bother with the extra integrity are attracted to participate in federally supported research? Does the reputation of companies matter to the advance of medical science? Or was Feynman wrong about integrity? Is Ioannidis wrong about clinical research?
Now about this “public interest.” What is the “public interest”? We might expect that public interest is not necessarily a private interest, though there are some that argue that the public interest ends up being optimized by a bunch of private interests all doing their own wild and crazy things. We might think that public interest considers a benefit to an entire population rather than a benefit to only a few. That makes some sense, but only if not thought of too hard. If we can cure one person, is that not a good thing? Might we learn from doing so to figure out how to cure another? Would not that be better than refusing to do the research to cure one because it holds no direct prospect to cure everyone?
The FSA’s policy use of “public interest” seems off. Perhaps public interest in the pens of institutional officials will always seem off, if not be off. One would think that the point of biomedical research would be to learn how people can stay healthy, how they can prevent diseases, and how to cure diseases when they get them. No need for an abstraction like “public interest.” It’s not like the public might be confused that the FSA might fund research against the public interest–to attack the public, to create diseases, say, and release them on a foolishly trusting public. “Public interest,” then, is something of a placeholder for any number of other institutional intentions with regard to allocating money for research.
The 110 policy does not stop at use, however. The policy also demands proper tone for the use–the use has to be “best.” Of all the uses one can imagine, the policy drives in typical bureaucratic fashion to demand the “best” use. Again, why? Why isn’t a nominal use good enough, a serviceable use, even a “they also serve who only stand and wait” use? Because bureaucrats! policy!? Perhaps “best” just sounds like a better policy goal than does plain old use. Do possible uses form a well ordered set, so that uses can be ranked and the best use distinguished from the rest? This is not a technical nit–it’s essential. The policy assumes that best can be distinguished from the rest. How? Who? What then happens to uses that aren’t the best, but are still uses?
The FSA had to deal with multiple issues. First, how might federal funding for research change the commercial nature of public health products? Should the federal government in choosing who gets funding pick “winners” and so disrupt competition and put all other companies at a disadvantage? Second, what about how the terms of research contracts attract some companies and put off others? Should the FSA adopt a policy on results that makes some companies–perhaps those with the best commercial positions already–from participating in federally supported research? And third, what about patents? Should inventors or their employers get to take out patents on what is invented in federally supported programs, or should the federal government use the patent system to document inventions but not use the patent system to suppress use of inventions or to demand payment for use or to set conditions for that use?
Even if the FSA restricted its funding to nonprofits, these questions still remain open. A nonprofit or an inventor could do an exclusive deal with a company to convey all rights in research results, and the company could then suppress uses or demand payments or put conditions on use. Then the nonprofit or inventor picks the “winners,” and the public doesn’t even get a say, nor does the public agency funding the work. Thus, even funding nonprofits does not get at the problems in federal funding of who gets rights, on what terms research should be contracted, and whether patents should be used to exclude others and extract rents.
The FSA as a matter of policy takes the position that among the possible uses of research results, the uses that best serve the public interest are ones that arise from open access.
And this is the heart of the debate over patents on inventions made in federally supported work. It is easy enough to argue instead that the public interest is served only when some research result is used such that some new thing may be mass produced and distributed to all who would use it. If one goes this way, then it also seems reasonable that if open access would make some companies–even capable ones–uninterested in mass production and distribution, then the public interest would not be served if no other companies came forward to mass produce. Worse if the federal government does not also fund “development” of results from their scientific and laboratory basis into things that really do treat disease or injury and do it effectively, and safely, and–if companies are going to be involved–in a way that results in a profit. The less the federal government funds development, then the greater the profit payoff we might expect from companies that choose to get involved. If the results are sketchy to begin with–untested, not replicated by others not in on possible deals, in need of many other things to also be done for the first time–then demanding immediate commercial use might raise the requirement for a profit payoff that much higher. And there you might see then how federal agencies, in a last-ditch effort to get the results they have funded into use (to justify, as it were, spending the money in the first place), might be led to offer patent monopoly rights to companies (and to universities to shop to companies) as the best of all incentives.
But the FSA doesn’t go there. Order 110-1 proposes that open access best serves the public interest. There’s way more structure in practice to the 110-1 default–is the government picking “winners” in an industry in the public interest? Is the government ceding control of the terms under which results are suppressed or licensed in the public interest? Will the public benefit from such results only if those results are placed in the hands of a company motivated to exercise a monopoly and so attempt to mass produce a product based on those results? And what if a federal agency spreads the work in a given area around, so that ten nonprofits and three companies all make inventive contributions to the treatment of a given health issue? What then does one do with that argument about patent monopolies to incentivize mass production–but now there are thirteen players all with their patent bazookas loaded, and no easy mechanism by which they can all cross-license their results and compete on some basis other than exclusion of all others. But wait–that would mean federal open access makes good sense. And so it goes.