« A Million Years of Logic, the End of Economics, and the Sociological Future | Main | Odd Numbers blog, from Portfolio »
Should we waive legal liability for FDA-approved drugs?
So asks Megan McArdle. The argument runs as follows:
I don't understand quite why FDA approval of drugs and medical devices hasn't long provided legal safe harbor for their manufacturers. The defects that show up, such as the Vioxx and Fen-Phen problems, are discovered long after approval precisely because they're so rare that they don't show up in ordinary clinical trials. If the government experts, who are presumably highly motivated to avoid catastrophes, can't spot the danger, why do we expect the drug companies to?
I can think of three possible rebuttals:
1. We simply can't trust the bureaucrats to find the flaws with drugs. But note this is inconsistent with both the rhetoric of FDA defenders ("the FDA can work") and FDA critics, who argue we are overinvesting in drug safety as it is.
2. Lawsuits encourage the companies to look for problems once a drug is already approved. Regulation does not.
3. People need lawsuits as a way of emotionally striking back. If they are denied that privilege, they will demand ridiculously oversafe levels of regulation in the first place. In this view regulation is as much about building consumer confidence in a health care system as it is about protecting people.
I do not currently have a view on this matter. Do you? Kevin Drum is opposed.
Posted by Tyler Cowen on March 4, 2008 at 07:03 AM in Medicine | Permalink
Comments
I tend to be opposed.
One other reason is that companies may know more from their internal testing than the FDA knows. If they withheld knowledge of a flaw they certainly should be culpable even if they successfully tricked the FDA.
That said, I think the burden of proof probably should be higher against companies that passed the FDA standard in good faith. If there's no reason a company did know or should have known about the flaw, the case should be dismissed.
Posted by: Greg Sanders at Mar 4, 2008 7:42:23 AM
I didn't think FDA critics said that bureaucrats can't find the flaws in drugs. I thought they said that it's not a cost effective way of finding the flaws. It should then be noted that they tend to value time to market heavily. People dying because of drugs not being available due to slow approval processes are counted as a cost for the FDA critics, but not for the FDA defenders.
Lawsuits without FDA is more than likely much better than the other way around.
Posted by: Joe T at Mar 4, 2008 7:44:18 AM
My preference is to allow companies to put any drug they want to onto the market (with minimum toxicity testing) but be accountable via lawsuits for anything that goes wrong. Anything that goes through the full FDA-approval process should be off-limits from lawsuits.
However, my primary rationale for that position is to quicken the pace of medical innovation.
Posted by: Jody at Mar 4, 2008 7:52:11 AM
Perhaps a percentage cap the maximum cumulative penalty relative to the number of documented and verified cases of ill effect and the amount of revenue that particular drug has yielded.
Posted by: Vin at Mar 4, 2008 7:53:05 AM
Seeing Tyler say "I do not currently have a view on this matter" helps me to maintain agnosticism on more issues. Even hearing Bryan Caplan say that several times, it's still hard to do...
Posted by: shawn at Mar 4, 2008 7:55:05 AM
Isn't this a case of the Coase theorem, where it doesn't matter who is liable, so long as the laws are consistent? If the drug companies take on the liability, they're just going to pass it on to the consumer anyway, in the form of higher drug prices. If they don't take on the liability, then the consumers would pay less and could use the savings to buy insurance.
I suppose the argument could be made that leaving the consumer in charge of the liability allows them to choose whether to insure for the risks or not, whereas leaving the manufacturer in charge of the liability forces all the consumers to basically pay for such insurance.
With that in mind, I'd argue for an implicit but waiverable warranty. A drug sold FDA-approved-but-otherwise-AS-IS isn't eligible for these lawsuits, but any other drug is.
Posted by: Anthony at Mar 4, 2008 8:03:46 AM
I think that there is an assumption in Ms. McArdle's argument which is not true. The assumption is that if a drug which is approved has a hidden flaw which is discovered later, that the drug company is sued for something for which they no nothing about.
In fact when a drug company gets sued, it is not for a hidden side effect. It is only when a drug company KNOWS of a side effect, yet pushes the drug on the public anyway in large amounts, is when the lawsuits fly.
In the Vioxx example, the makers said that Vioxx was better than Aspirin, safer than Aspirin, and only 100 times as expensive as Aspirin. When in fact they knew there was small yet deadly cardiovascular side effect, they pushed the drug anyway and minted money. If the drug company had simply said "Vioxx is pretty good, but all new drugs have side effects, and you should only use it if Aspirin and other drugs don't work for you, and be careful of these side effects, etc." they would never have been sued. They also would not have made nearly as much money.
The real question that should be asked is:
Should a company who knowingly hides a side effect of a drug and makes a ton of money on it while killing it's customers be subject to lawsuits even if the drug has FDA approval?
I think that answer is yes.
Posted by: Jonathan L at Mar 4, 2008 8:25:49 AM
Can't we consider the safety risk as an externality that is not properly priced into the cost of the drug? If that is so, legal liability transfers the costs from that risk to the manufacturer, who then incorporates it into the price.
Posted by: Rob at Mar 4, 2008 8:27:13 AM
Vouching for the safety of drugs is only one function of FDA regulation. Another is vouching that they are at least somewhat effective.
Posted by: Cyrus at Mar 4, 2008 8:31:25 AM
I would argue that waiving legal liability for FDA-approved drugs only should make sense if they simultaneously allowed non-FDA-approved drugs to be on the market, and make the process become arduous enough that it becomes a "fair return" on investment to be protected from related torts...
Posted by: Michael Fisk at Mar 4, 2008 8:58:30 AM
Validation.
The one and only reason to be against FDA regulation. Many out of processes are still in use through pharma and biotech, simply because the cost of validating the new process in prohibitive. So a better safer way to manufacture the drug is not used.
Another lesser reason is off-label use. Once a drug is approved it can be used for anything, not only it's approved use. The FDA never evaluated the efficacy (phase III trials - the biggest and most costly) of the drugs for off label use, the market does.
Posted by: Marc at Mar 4, 2008 8:59:15 AM
Omigod. The FDA, like many "regulatory" agencies, is largely run by and for the industry it's supposed to regulate.
Unless and until safeguards are implemented to prevent that being the case (& I'm not sure they are feasible), the McArdle proposal is ... well, McArdlian.
How about that "new" study on how Prozac et al. work only for the severely depressed? Where'd they get that info? ... from the FDA, which sat on it for years.
Posted by: Anderson at Mar 4, 2008 9:32:33 AM
Isn't this like saying you're not liable for damages caused by car accidents you get into if you're a licensed driver? By licensing me, the government said I'm a safe driver, so any accident is the DMV's fault.
Posted by: Mo at Mar 4, 2008 9:43:04 AM
By God, that would give you wonderful returns on bribing a bureaucrat.
Posted by: dearieme at Mar 4, 2008 9:44:21 AM
Two ideas (probably already mentioned somewhere by someone else)
1) Turn the FDA into an "honest" broker of information. Provide expert clinical trials, witnesses, and the like.
2) Change the FDA's role to that of an Intrade-like betting market manager (and judge of contracts).
Posted by: Eric H at Mar 4, 2008 9:53:30 AM
One lawyer’s $.02:
The whole point of the tort system is to make those who suffer damage whole through monetary payment without assigning moral blame to the tortfeasor. (When we set out to assign moral blame, we make a criminal law). With immunity, who makes the victim whole? Regulation is an exercise of the police power in service of public health. Money well spent. It is not, however, a system for addressing consumers when things go wrong. It is the tort system that both protects the consumer by recompensing them for damages and by acting as disincentive to manufacture dangerous drugs. It also allows the manufacturer to reasonably manage risk. (And no, I'm not a plaintiffs' attorney).
Posted by: robert at Mar 4, 2008 9:58:09 AM
I believe drug companies can select which clinical trials to report and usually don't disclose the bad results. This reminds me of the hilarious quote from Anchorman, "60% of the time it works every time." I've never heard a compelling argument for letting the companies keep information on some of the trials private, and it seems like this allowance is probably the result of the pharmaceutical lobby.
Posted by: Adam H at Mar 4, 2008 10:14:29 AM
4. Tort law serves an insurance function. This idea has been around for several decades. The notion is that large business enterprises are in the best position to 1) obtain information about the potential harms of their products, and 2) spread the cost of any harms over the entire run of that product, thus basically incorporating a small insurance component into the price of the product. You can argue endlessly about whether tort law actually does serve this insurance function very well (i.e., whether it compensates the right people), but the insurance argument isn't affected by whether or not the FDA approved the drug. The point isn't assigning blame, after all, but spreading risk.
Posted by: Stuart Buck at Mar 4, 2008 10:21:06 AM
Apparently Ms. Mcardle doesn't follow the Vioxx litigation very closely. The CV problems of Vioxx were identified in a clinical trial, and they even changed the label instructions to indicate that.
The most salient and important criticism I can come up with against an FDA approval safe harbor is that industry capture is too much of a reality.
Posted by: Eric at Mar 4, 2008 10:40:30 AM
If you are going to criticize Ms McArdle, at least read her entire comment, which would only take a minute. If you did, you would read this:
I mean, obviously, if you mislead the Feds, you should be subject to criminal prosecution.
Posted by: anomdebus at Mar 4, 2008 10:51:04 AM
The criticism of the FDA is not necessarily that it results in strict overinvestment in safety, but that it doesn't have incentives to avoid malinvestment or to do a proper cost-benefit analysis.
Posted by: Brandon Berg at Mar 4, 2008 11:09:58 AM
So long as the drug companies are doing the testing, they must accept liability. If the government does the testing, then I could accept the government taking on at least some liability.
Posted by: Evan Robinson at Mar 4, 2008 11:20:52 AM
Mo:
Isn't this like saying you're not liable for damages caused by car accidents you get into if you're a licensed driver?
No, because the quality of your driving is not consistent. The DMV can test the quality of your driving during the test, but there's no guarantee that you will continue to drive as safely after the test. The quality of drugs is far more consistent--a drug is likely to be as safe on the market as it is during the trial, unless the company cuts corners on the manufacturing process, in which case it should be held liable.
Posted by: Brandon Berg at Mar 4, 2008 11:25:14 AM
Banning lawsuits would decrease or eliminate altogether the mechanism through which third parties investigate and and test adverse effect claims. In other words, if someone suspects that they were hurt by a drug, there would be no incentive for lawyers to take on the case (unless they can pin the blame on someone else, see below) and therefore no money to pay experts to look into it. In the absence of this sort of investigation, it is likely others will also be hurt who would not have been had a lawsuit been possible.
Banning lawsuits against pharmaceutical companies would also increase lawsuits against healthcare providers and probably lead to new theories of liability.
Posted by: Jorge Schmidt at Mar 4, 2008 11:27:20 AM
Banning lawsuits would decrease or eliminate altogether the mechanism through which third parties investigate and and test adverse effect claims. In other words, if someone suspects that they were hurt by a drug, there would be no incentive for lawyers to take on the case (unless they can pin the blame on someone else, see below) and therefore no money to pay experts to look into it. In the absence of this sort of investigation, it is likely others will also be hurt who would not have been had a lawsuit been possible.
Banning lawsuits against pharmaceutical companies would also increase lawsuits against healthcare providers and probably lead to new theories of liability.
Posted by: Jorge Schmidt at Mar 4, 2008 11:27:52 AM
