Well, here's a short argument at Nature that there are good economic (and social) reasons to insist that drug companies show efficacy before they release drugs. Some extracts:
Knowledge of the history is important. The 1938 US Food, Drug, and Cosmetic Act required only that drug safety be demonstrated. In 1962, new legislation demanded that marketed drugs also go through well-controlled studies to test for therapeutic benefit. More than 1,000 medical products were subsequently withdrawn after reviews found little or no evidence of efficacy1. The free market that existed before 1962 revealed no connection between a drug's ability to turn a profit and its clinical usefulness. The same is likely to be true of any future deregulated market....And here's the key point:
An overly stringent system will err by withholding or delaying safe and effective 'good' drugs from patients. Critics of existing regulations often point to the case of a treatment for Hunter syndrome — a rare, inherited degenerative disease in which the absence of a crucial enzyme can be fatal. Trials of the enzyme-replacement drug Elaprase (idursulfase) meant that, for a year, a group of children received a placebo instead of the drug that was eventually shown to be effective2.
Conversely, a lax regulatory system will subject patients to 'bad' drugs that may be toxic. The iconic example is the more than 10,000 birth defects caused worldwide by the drug thalidomide, a late 1950s remedy for nausea during pregnancy. Even in the past dozen years, initially promising drugs, such as torcetrapib (for reducing cholesterol and heart-disease risk) and semagacestat (for improving cognition in people with Alzheimer's disease), were found to cause harm only after they had been tested in large, mandatory trials — effects that were not seen in the smaller trials3.
The most extreme proponents of deregulation argue that the market can serve as the sole arbiter of utility: if a medicine is selling well, it must be delivering value4. A more moderate view is that reliable information on efficacy can be collected after a drug goes on sale, through uncontrolled observational studies and other post hoc analyses.
There is a third type of error that these arguments neglect (see ‘The good, the bad and the useless’). Untested drugs can be reasonably safe but provide no benefit.
Arguments for deregulation fail to recognize that valuable information has a cost. Drug companies cannot afford to generate reliable evidence for efficacy unless their competitors are all held to the same high standards. Efficacy requirements level the playing field and ensure that the health sector receives the data needed to inform good therapeutic and economic decisions. The government, insurers, patients and others need to know whether medicines are likely to provide benefits. Patients and physicians must have access to reliable information to make educated and ethical choices.Jason - that article is definitely tweet worthy, no?
Rigorous clinical studies are still the best way to learn whether a drug works, and regulation is essential to ensure that these studies are conducted. Pre-specified endpoints, controls, randomization and blinding cannot be discarded without sacrificing actionable clinical information5.
Once a drug is on the market, it is hard to gather solid efficacy data....
The FDA's gatekeeper role makes the medical marketplace function. The economic benefits of good research and a healthier population will be lost without incentives to find truly effective drugs.
Update: I just Googled up an article at Vox from a couple of months ago that explained the pro FDA argument from a medical point of view. A lot of this read like what John just said in comments:
Thiel, a libertarian iconoclast, has repeatedly made the case that the FDA gets in the way of drug innovation by making it too difficult for new medicines to get to the market. Some of the FDA candidates he’s identified — including Silicon Valley’s Jim O’Neill and Balaji Srinivasan — have similarly argued that the agency should dump its requirement that drugs be proven effective before reaching the market, and that we’d be better off if the FDA operated more like a “Yelp for drugs.” In other words, bringing the same speedy and disruptive approach to medical regulation that Silicon Valley brought to the taxi and hotel industries, for example, will unlock cures — fast.But Thiel and his pals miss a very important point about developing new drugs: Manipulating biology isn’t the same as manipulating computer code. It’s much, much harder. Speeding up medical innovation will take a lot more than just stripping down the FDA — it’ll take huge leaps forward in our understanding of biochemistry and the body. Health care is also different from taxis and hotels in another key way: Consumers can’t really judge the safety and quality of medical products by themselves....
...I asked a longtime pharmaceutical scientist (and conservative), Derek Lowe, for his views. In his 28 years in the lab, Lowe has seen hundreds of thousands of compounds tested on a huge variety of drug targets, and never, not once, has he brought a drug to market.
The reason? “We don’t know how to find drugs that work,” he said.
For every 5,000 compounds discovered at this "preclinical" phase of drug development, only about five are promising enough to be tried in humans. That’s a success rate of 0.1 percent.
Drug innovation comes from painstaking tinkering and a dash of luck. “It’s very tempting for someone who has come out of IT to say, ‘DNA is code, and cells are the hardware; go in and debug it’,” Lowe said. “But this is wrong.”
In Silicon Valley, humans have designed the hardware, software, and computer code they’re working with. In medical research, scientists do not have that advantage, Lowe said. “We have 3 billion years of spaghetti-tangled gibberish to deal with. And unless you’ve done [drug development], it’s very hard to get across how hard it is. I don’t know of anything that’s harder.” Biochemistry and cell biology are “like alien nanotechnology,” he added.
Update 2: from another article, talking about the effect of having an FDA that insists on showing efficacy as well as safety:So the real hurdle researchers face when it comes to finding new drugs for people isn’t overcoming a stringent regulator; it’s grappling with that “alien nanotechnology” in the lab.
Pharmaceutical executives complain about the drug approval process, but usually don’t want to go anywhere close to a safety-only path. In practice, what they want is for the FDA to return their calls, for bureaucratic delays to be reduced, and to find the fastest and least expensive way to prove safety and efficacy.
Many biotech entrepreneurs are actually fans of a tough FDA. Pharmaceutical billionaire Leonard Schleifer, the founder and chief executive of Regeneron Pharmaceuticals, said that he was against “making it really easy to get your drug approved” at the Forbes Healthcare Summit last week, before news of that O’Neill was under consideration leaked.
Schleifer said that he couldn’t compete with companies like Pfizer or Eli Lilly, which have 10 to 100 times as many salespeople as Regeneron. But he can compete to get approved first, or to have a better drug that has more uses that the FDA allows it to advertise based on science.
“Having a high bar is a good thing, in my opinion, because it allows innovators to compete,” Schleifer said.