FDA payola?

The FDA will soon stipulate that researchers who accept more than $50,000 in corporate grants, contracts and consulting fees cannot sit on FDA advisory committees.  This will rule out many current advisors.

First, I wonder how this fits into the old Sam Peltzmann story that the FDA is too conservative in approving new drugs.

Second, what if we reformed in the opposite direction?  Why not do away with all the mandatory drug trials and the like, and simply let drug companies purchase approval for new drugs?  Think of the companies as posting bonds, and of course they still can be sued ex post if the drug harms somebody.  The companies still will have reason to conduct their own tests.  Set the price high if you wish.

To be sure, how much a company will pay for approval will depend on expected profits, not social welfare.  But even with market power there is usually some connection between those two magnitudes.  Or maybe the fear of lawsuits won’t deter poorly capitalized companies, but at the very least we could let the corporate giants take this path.

Some companies might be too overconfident about their drugs.  If you believe that, I hope you are buying puts on them.  Other companies might have excessively short time horizons.  If you believe that, I hope you are loading up on drug companies with heavy R&D and raking in your excess returns.

So does this idea have any takers?  If not, why not? 

Addendum: Matt Yglesias argues regulation is a substitute for litigation.


We rely on government approval of medicines because it is nearly impossible for individuals singly to judge the effectiveness or long-term risks of taking a drug. Sure, if a thousand people died in a month after taking ViagCialOxyStatin then the victims’ estates might file a class action against the manufacturer, the press could pick up the story, and the transgressor would be appropriately punished. However you'd also end up with thousands of patent medications of diluted water and food coloring with tons of marketing and no benefit at all.

We pay our government to regulate medicines because we believe that they can qualify and hire scientists who are adept at reviewing the effectiveness and cost/benefit of new drugs better than we can ourselves. Allowing a free market would be akin to getting rid of the requirement for individuals to obtain a driver's license -- the logic is parallel.

On the other hand, the trust that the public has placed in the government to qualify and hire scientists has been breached by examples of manipulation of drug testing results and secretive approval processes. In this case we'd be better served by opening up the operations by which we hire and qualify the government inspectors, and by providing thorough and complete transparency to the clinical-trial testing and approval processes.

-- Jeff

Litigation does replace regulation. Either regulate and pass tort reform, or deregulate and let the court's punish companies that cheat.

I favor a two tiered approach. We need something like underwriters labratory and consumer reports as an independent drug agency. It can be funded by a small tax on medicines. Drugs must get approval from at least this body. The other path (and a company could do both) is to go before the FDA. The FDA will take longer, but once approved the company has zero legal liability (assuming they did not decieve the FDA) and anyone harmed by an FDA approved drug must accept a standard government payout.

I'll go for this. Frankly I think the FDA is silly: let the people take whatever drugs they want, just make sure that manufacturers aren't lying about what chemicals they are selling. The flat fee for approval can cover those oversight costs.

Of course, many more early adopters will get killed/harmed if this is done. But no one is forcing you to be an early adopter, and the rest of us get the benefit of more drug testing for free.

Better get rid of public health care before you try that, though.

Right now we seem to have the worst of both worlds - regulation plus litigation. The FDA approval process is slow and very expensive - it can cost over $100 million to bring a new drug to market. Then, if there are adverse results, the companies still get sued. Why should we care? Well, we, the consumers, are the ones who ultimately get stuck with the bills. It's a major reason for the high cost of drugs in this country. Personally, I don't have much use for either trial lawyers or bureaucrats, but, if forced to pick, I'd choose the latter. The FDA is a huge, risk-averse bureaucracy, because patients who suffer horrible side effects from an FDA-approved medication make it into the headlines, while patients who suffer or die because the FDA won't approve their needed medications do not.

I would slim down the FDA and speed up the approval process but then grant tort immunity to companies for drugs that get FDA approval. But if the companies lie, cheat or fudge data during the approval process, the immunity is removed and the tort lawyers can have at them. This would give the drug companies strong incentive to play by the rules.

Important early hurdles to new drugs are clinical trials, which are staged. Fall at an early fence and that is that. Further, the FDA places the bar very high--patients who are sickest, and who have progressed furthest, must be helped by the new therapy or it won't be approved. Yet in a world of increasingly individualized and targeted medecine, something effective at an early stage, a desirable and cost effective new tool, could be screened out by the FDA.

Expertise and information are at the heart of the problem. The FDA hires agents--medical experts--to judge the research of other medical experts. Patients are supposed to know too little to be allowed input. In fact, in the case of clinical trials there is very often a sub-population of patients who know a great deal about the risks and benefits of a therapy, they have an incentive to do so, after all.

So let me make modest proposal for an improvement in the clinical trials process. The FDA judges an early stage trial to have failed if it doesn't help enough people (for example, Genasense, a cancer therapy, was almost killed-off after a stage I trial because "the data showed a 10 percent benefit. More simply put, one in 10 patients who received it had real improvement and often a very significant remission. For the FDA, this was not significant because they saw too many others who didn’t get much benefit and thought the expense wasn’t worth it.." (quote from Andrew Schorr at HealthTalk). Of course, although the FDA didn't like the drug, my guess is that any patient among the 10% helped was pretty happy with it. As it happens, patient driven complaints may have helped to have this case reconsidered.

What if there was a patient driven escape clause in clinical trials as a matter of course? If an FDA approved trial at an early stage failed to pass the agreed bar, what if the drug company got one shot at the next stage if they could round up enough informed patients to participate? Their willingness would itself be a test. This could also require participation by one academic medical center. Such a trial would, at least, collect more data, it would be consumer driven, it might actually help a very sick person or two, and have the effect of taking the FDA out of the hot seat.

"We pay our government to regulate medicines because we believe that they can qualify and hire scientists who are adept at reviewing the effectiveness and cost/benefit of new drugs better than we can ourselves."
NO, we don't. With all due respect, Mr. Chapman, you cannot begin to imagine how frustrating and dehumanizing it is to live with an illness and have your most important decisions taken out of your hands. Risk/benefit tradeoffs are unique to the individual. No FDA advisor has ever walked a mile in my shoes, much less the last 20 years. They simply cannot make that decision better than we can ourselves.

Expertise is relevant to give advice; not to make another person's decisions for them. If I'm the expert, should whether you need aspirin or Actiq be determined by my fears that you will become a fentanyl addict, or your level of pain and your personal belief about whether the risk is acceptable? As for the information asymmetry, information is a good like any other, and markets will provide the information people demand. The FDA doesn't provid certainty or complete information. If anything, it gives people a false sense of security.

Shavell has the classic piece on this talking about ex ante regulation versus ex post litigation as substitutes. Kolstad and Ulen have a piece on this in AER 10 or 15 years ago, arguing that litigation and regulation both substitutes and complements.

Some people here are making a common error about new drugs and clinical trials. The problem is - and I speak, unfortunately, from experience here - is that we don't actually know all the things that might happen when a new compound reaches the market. Clinical trials will uncover the largest sets of side effects (or they should!), but no trial can mimic what will happen in the real world. The difference is the diversity of patients out there - and more importantly, their sheer number.

Never forget this: the people at every single drug company hold their collective breath when a new compound goes into animal toxicity testing. And again when it goes into phase I, and phase II, and again in phase III. And they hold it again when the compound hits the market. It's quite a business to be in, let me tell you.

If the FDA will me make it alright for companies to buy their drug into the market, then the reputation of the FDA will be scrutinized because the american public is led to believe that the drugs that the FDA approves are safe. There will be no reason for a companie that wants to make a drug that everyone would want with little testing, put it on the market, receive the profits and disappear.

Comments for this post are closed