Overwhelming consensus says that patents discourage innovation.

Fair warning: this is a post about patents and IP. I say that because in the past, I've practically been able to see the readers fleeing in real time when such things come up. Two or three of them in a row, and there would probably be nobody left around at all. But it's a big part of drug discovery, like it or not, and some people don't.

Perhaps the editors of The Economist fall into that category? I say that because of a big article (the cover story, with lead editorial) in this week's issue. The subhead is "Patents are protected by governments because they are held to promote innovation. But there is plenty of evidence that they do not". And it's worth seeing the passage from an 1851 issue of the magazine, which held that the patent system "excites fraud, stimulates men to run after schemes that may enable them to levy a tax on the public, begets disputes and quarrels betwixt inventors, provokes endless lawsuits [and] bestows rewards on the wrong persons". It's certainly still capable of doing all of those, for sure.

Here's their summary of the idea behind granting patents, and it's exactly as I would have put it myself. The second paragraph is where it gets into the territory of this site (and my job, and the jobs of many of the readers here):

The public-good position on patents is simple enough: in return for registering and publishing your idea, which must be new, useful and non-obvious, you get a temporary monopoly—nowadays usually 20 years—on using it. This provides an incentive to innovate because it assures the innovator of some material gain if the innovation finds favour. It also provides the tools whereby others can innovate, because the publication of good ideas increases the speed of technological advance as one innovation builds upon another.

This sounds plausible. But is it true? There is much room for doubt. The evidence that the current system encourages companies to invest in research in a way that leads to innovation, increased productivity and general prosperity is surprisingly weak. A growing amount of research in recent years, including a 2004 study by America’s National Academy of Sciences, suggests that, with a few exceptions such as medicines, society as a whole might even be better off with no patents than with the mess that is today’s system.

There we go: even people who think the current patent system is broken (and business method and software/algorithm patents make a strong case for that), and some who are philosophically opposed to the entire idea, tend to regard the pharmaceutical part of it as the one that actually comes closest to working the way it's supposed to. I've had several attorneys over the years tell me that patent quality is noticeably higher in the drug field - prior art gets more of a look, inventions tend to actually be enabled, and tend to actually be inventions. None of these hold as often as they should, of course, but they are even worse in other areas.

Here's a case against the whole idea of patents, from some economists at the Federal Reserve. The ties between patent regimes and innovation in general are not as strong as you might think, as they show with a good deal of evidence. And there's no doubt that patents, as a proxy for innovation itself, aren't much use. That's partly an observer effect - some companies, universities, and whole countries have decided that since patents are correlated with invention (dubious assumption though that may be), that if they can just file more patent applications, then they must be innovating more. It's so convenient, too, because you can get an exact count of just how innovative you're being! This is an ancient mistake, applicable to more than patents, and any medicinal chemist who's worked in a department where people were rewarded for sheer number of compounds synthesized will know it well.

The Economist does get around to the drug-patenting system, of course:

Even if many industries do not really need patents—and a fair few might be better off without them—there is still a strong belief that in some businesses they are vital. The example always touted is pharmaceuticals. Drugs have to undergo exceptionally expensive and long-drawn-out testing procedures to demonstrate that they are safe and effective. And once a company has laboriously demonstrated that a molecule does its job with few or manageable ill effects, its rivals could make far cheaper copies were it not for patent protection. . .

However, the history of the industry raises doubts about such arguments. Until 1967 German drug companies could only patent the way they made drugs, not the formulae of the drugs themselves. Anyone could sell copies of the medicines if they found another method of making them. Yet Mr Boldrin and Mr Levine [the Federal Reserve link above - DBL] say German drugmakers produced more innovations than British ones (remember where aspirin was invented). Another interesting case is Italy, which had no patent protection for drugs until 1978. One study showed it invented a larger proportion of the world’s new medicines before that date than afterwards. Before the “reform” it had lots of copycat firms, but the biggest of these also did research on drugs of their own. They were largely wiped out once they had to pay royalties on their copycat drugs.

I'd be interested in knowing, under that previous German system, just how many times a rival actually did usurp an existing drug. It might have been possible in theory, but very hard to realize in practice, and in that case there would be little real-world difference compared to a system with outright chemical matter patents. That's a subject for further research (and if anyone has any insights, please send them along!) And I also would like to track down that Italian study, and see what historical period we're talking about. Inventing a new drug and getting it onto the market used to be a lot easier than it is today, and since the sheer complexity of the regulatory process can be an argument in itself for something like patent protection, that argument may well have gotten stronger over the years just through those means.

And now here's where we get down to talking dollars:

In 2005 Dean Baker, an economist at the Centre for Economic and Policy Research, a think-tank in Washington, DC, took a much simpler, but still rather striking, approach: he just compared the costs imposed by the patent system with the innovation that system bought.

America’s health systems, he noted, spent $210 billion on prescription drugs that year. Based on how much cheaper generic drugs were than patented ones, Mr Baker calculated that a competitive patent-free market might have provided the same drugs for no more than $50 billion. That represented a saving of $160 billion.

The drug companies reckoned at the time that they were spending $25 billion on R&D; the government was spending $30 billion on basic medical research. The money it would have been able to save buying drugs for Medicare and Medicaid in a patent-free world have allowed the government to double that research spending, more than replacing industry’s R&D, while still leaving $130 billion in public benefit.

It's that second paragraph where I wonder if this train of thought might not have gone off the rails. I don't quite see how you can go from generic drug prices, under the current regime, to what a patent-free market might have been able to deliver. Generic drugs have already had most of their costs borne by the original inventors, who undertook those projects under the current patent system, with full intent to use it in their pricing decisions. Generic drug prices, as far as I can make out, are useful for the thought experiment of "what if a bunch of useful drugs dropped from the sky, with instructions on what they're good for and how to make them", but they're not so useful, I think, for the further problem of how you figure those things out in the first place.

So if I have a problem with the way the numbers in the second paragraph were generated, I have to leave the third one alone. I will note, though, that the $30 billion on basic medical research mentioned does not necessarily go into the same basket as industrial R&D, as it assumes. I've gone on about this issue many times (here are three posts from 2004, for example!) Academic work, of the kind traditionally funded by the NIH, is vital, but it addresses basic research problems rather than applied ones. (I'll leave aside the recent translational research push there for now). Let's take "By what mechanisms does muscle tissue respond to insulin, and how do these malfunction in Type II diabetes?" as an example. That's an excellent set of questions, important parts of which are still in dispute. But they are different questions than "What chemical compound would be a useful drug for Type II diabetes?". The research efforts spent on the first topic will most certainly affect work on the second one, but it's also possible to work on the second one without a clear understanding of what's going on at a molecular level at all. That, for example, is how the glitazones (PPAR-gamma ligands) were discovered, and metformin, too: by screening in cell and animal models. Rosiglitazone was deep in Phase III before its target was even discovered, and metformin's may not even be known now, years after it's gone generic. Those two compound illustrate something else - that drug industry research can, in turn, stimulate basic biological research through providing tools that are known to produce well-defined effects.

And even when the two are coupled, the story isn't simple. I'll take a diabetes project that I worked on years ago as an example. Hormone-sensitive lipase is known to be a key enzyme in lipolysis, the liberation of free fatty acids (FFAs) from fat stores in tissue. And FFAs are key players in type II diabetes - their blood levels are too high, because a type II diabetic's tissues, some of them, act as if they're fasting, even though there's plenty of glucose and insulin washing around (there's that insulin resistance thing - the signaling is broken). They start mobilizing stored fat, inappropriately, and the resulting FFAs go on to stimulate the liver (most inappropriately indeed) to actually start making more glucose and dumping it into the blood, the last thing you need, because it interprets the situation as a fasting state, too. So interrupting lipolysis was not such a bad idea, therapeutically.

Now, all that background I just flew through was established by basic research, the great majority of it, I'm sure, in academia, and no small part of it probably funded over the years by the NIH. But none of it was news when we started our project, and in fact, hormone-sensitive lipase (HSL) had been known enzymatically for decades. But no one had tried it out as a drug target, or not to any effect that we could see. In the late 1990s and early 2000s, though, more than one drug company with an interest in diabetes gave it a whack, and we (among others) eventually developed potent and selective HSL inhibitors. And what we found, unfortunately, was that they were not going to be drugs: there's a mechanism-based toxic effect that I won't go into, but is not without interest, that would seem to keep this particular drug target off the list. We demonstrated that in animals with our new compounds, and then went on to other things. Academic labs, in turn, wrote to us requesting samples of our best compound, because it was the best HSL inhibitor known, and a valuable tool for in vitro systems.

So just who, in this case, was ripping off whom? Because that's what a lot of these arguments come down to, that the drug industry is just piggybacking on NIH money. This particular patent argument isn't making this point in so many words, but by treating the money spent in both camps as part of the same ledger item, it's making it implicitly, and at the very least confusing two different things. And my point is, they really are different things, with different purposes, that help advance each other's goals. Basing money-saving arguments on this foundation calls their conclusions into doubt.

The Economist's possible prescriptions for drug research (even though they admit that these are too radical, for now, to be put into effect) hinge on awarding prizes, and that's not a crazy idea. The current system provides a monetary reward for discovery, a prize system would (presumably) do the same. There are various arguments about how the prize money will come out of "all the money that'll be saved", and (as indicated above) I don't know how much that is, because I don't think that current calculations are realistic. (Note that the prizes are going to have to be pretty substantial in order to have a real effect). But in the end, the whole patent system is an incentive, and if there are different ways to provide that incentive - especially different ways that avoid some of the problems we have now - then they're worth considering. Politically, it would be an uphill fight, but uphill fights have been won before, with time. One has to keep Ambrose Bierce in mind, though, when he said that a conservative was someone enamored of the existing evils, while a liberal wished to replace them with fresh ones.