How Safe and Effective is the FDA?The hundredth anniversary of the FDA is by any measure a bittersweet moment. Born of a reformist zeal directed against the marketing of impure and adulterated products, the original mission of the FDA could hardly be regarded as controversial as a matter of first principle. Hidden defects in foods and drugs are hard for ordinary consumers to detect or correct on an individual basis. Therefore, some collective mechanism that substitutes for consumer weakness is hard to attack on grounds of first principle, even if we accept the narrowest account of the police power that allows the government to oversee private property and voluntary contracts only if they affect the health and safety of the public at large. By keeping food, drugs and cosmetics free of adulteration, the FDA discharges a market-reinforcing function that confers a systematic advantage to reputable firms that sell sound products.
Richard A. Epstein
Medical Progress Today
June 30, 2006
Over its history, however, the FDA has suffered from massive mission creep. The first expansion of its role took place in 1938 when new legislation allowed the FDA on a limited basis to monitor whether certain drugs were "safe" and pure. The 1938 legislation was prompted by the misuse of a diethelene glycol, a poisonous solvent that killed several hundred people. The older rules could have handled that problem by way of quality control, but instead more intrusive regulations were adopted. No longer is it sufficient to show that products are produced in accordance with their basic chemical formula; now it is also necessary to show that the formula is safe. Yet since absolute safety has always been unattainable in dealing with medical products, this simple command has over time morphed into one that asks whether a drug's risks outweigh its benefits or vice versa. That is no mean shift in emphasis. To give a comparison, the shift here is similar to that which took place in products liability law in the 1960s. No longer could a plaintiff win by showing that the unit did not correspond to the plan. Now it was possible to win by showing that the design of the unit, or the warnings supplied with it, did not conform to some external cost/benefit analysis. This is a huge shift in the intellectual requirement, but taken with scarce attention to the implicit difficulties.
The second stage in the expansion of the FDA's jurisdiction was more noticeable. Starting with the Kefauver amendments in 1962, drugs had to be shown to be effective as well as safe. The 1962 effectiveness rule was adopted in response to the thalidomide scourge on unborn babies. Ironically, the effectiveness standard was adopted in response to a a safety issue that was not caught because the risks in question only took place during the early stages of pregnancy. Yet determination of effectiveness requires yet a different kind of cost-benefit analysis, which asks not whether the drug makes a person worse, but whether in some sense it makes the person better off. As with safety, effectiveness will vary by individual, so that the new mission adds yet another layer of complexity to the overall situation.
There is, moreover, a deep political irony attached to the relentless expansion of the FDA's mission. The more the FDA is asked to do, the more likely it is that people will be in a happy position to say that it is doing something wrong. Put otherwise, traditional safety inspections have their risks, but these are nothing compared to the missteps that the FDA may well make in dealing with the safety and effectiveness issue that seem to many to define its overall purpose.
At this anniversary moment, the questions of the safety and effectiveness of Vioxx and Celebrex, along with other drugs, have given much ammunition to a slew of critics who claim that the FDA has fallen down on the job in discharging its statutory mandates on safety and effectiveness. Alarmed at the appearance of drugs that have been linked to thousands of deaths from such common ailments as heart attacks, strokes and liver disease, the uniform chorus calls for stronger scrutiny and more extensive clinical trials before new drugs are released into the market place. It then seeks to ratchet consumer protection up a notch by seeking strengthened powers to order stronger warning labels on products, to control the way in which new drugs may be marketed, and to order the removal of drugs that have made it through the net on the grounds that their subsequent safety record falls short of expectations.
The critics are right on one point. The FDA does not set the gold standard for a watchdog agency. But they are wrong about the direction of the mistake. The FDA keeps too many products off the market for too long. It orders too many products off the market at the first sign of adverse side effects. A record this consistent is not the result of bureaucratic indifference or random error. A pattern this conspicuous and dangerous has to lie in the basic approach that the FDA takes to drug warnings. In this instance it is not difficult to identify the philosophical culprit. The systematic error of the FDA is to overstate the risk of conspicuous losses and to overlook the major, if hidden, benefits forgone by unnamed persons when new therapies are kept off the market and older ones are removed at the first sign of trouble. The vestigial limbs from Thalidomide anyone can understand. The hidden spread of a liver tumor that some drug might have arrested if the FDA hadn't kept that drug off the market does not etch itself into the public consciousness in the same way.
The outcomes here are explained all too well by skewed political incentives rather than sound public policy. The FDA is always in the glare of Congress. Its officials do not have strong incentives to factor into account the losses to nameless individuals whose deaths or injuries are not readily visible or counted. Private parties find it hard to organize interest groups to secure the rapid release of life-saving drugs that have never made it into the market in the first place. Only very strong and unified groups, such as AIDS patients, are general able to push new drugs into the marketplace. The political calculations change once a drug does reach the market and develops a group of loyal users, as with Prozac patients who have long resisted any efforts to remove the drug from the marketplace or reduce its overall level of utilization. These groups are, moreover, not idle and uninformed individuals. Their families have struggled with chronic illness over prolonged periods of time, and they have developed a body of knowledge that is anything but casual and incomplete.
And that shows. In some situations, as with Astra-Zenica's lung cancer treatment, Iressa, the FDA is often prepared to split the difference by allowing persons who are taking the drug to continue to use it, while new patients are told to wait until new trials have made it better. Janet Woodcock, then acting (now regular) deputy commissioner of the FDA, has said that the public should rest assured that the FDA works "tirelessly to ensure that the balance of benefit and risk to cancer patients is in their favor." But most patients are not sure why or how their prospects of survival are magically improved when they are denied one clinical option simply because a second drug, in this instance, Tarceva, has made its way onto the market. It's no coincidence that user groups have flayed the FDA, and for reasons that bear special note:
As to Tarceva being an adequate replacement, that remains to be seen. The FDA does not know if Tarceva will provide the same benefit to the same patients that Iressa helps. The FDA deals only in averages, and the averages don't predict the response of any one individual to either drug.
This, from a letter to the Wall Street Journal in 2005, gives us a clear indication that the FDA errs so often not just because of the politics of public oversight, but because of a fundamental misconception of the FDA's mission. Despite the FDA's run of over 40 years in administering the 'safe and effective' standard, it has not been able to come up with sensible ways to decide either. The root of the difficulty is that the statutory standards make it appear as though the FDA could find some sharp line to divide safe from unsafe, and ineffective and ineffective products, just as the double yellow line sharply divides northbound from southbound traffic. But those sharp dichotomies are not maintainable in the shifting sands of pharmaceutical development.
In principle, it looks easy to ban one drug with adverse side effects when a new drug promises the same therapeutic effects without any of the side effects. Unfortunately, medical situations are rarely this clean. Remember, virtually all effective drugs have side effects, and the willingness to tolerate these effects depends at a minimum on the severity of the targeted underlying condition and the availability of other treatments that can be taken with or instead of the proposed drugs. Yet once the decision on drug use depends on multiple factors, the word "reasonable" has to be inserted in front of both "safety" and "effectiveness" to signal that the relevant judgments now turn on matters of degree, not kind. The FDA therefore has to scramble to get ever more information as to the pattern of drug benefits and side effects, which in turn leads to the expensive clinical trials and endless delays, all of which translate into reduced return to drug and device innovators, and fewer useful products for the public.
Worse still, the necessary shift to reasonableness determinations demonstrates why it is unwise to rely on the FDA to act as an omnipotent gatekeeper. As the Iressa tale painfully illustrates, the FDA can make its cost-benefit judgments only at the wholesale level. As such it must ignore the inevitable variations in response that take place across individuals who differ in age, sex, race, medical history, and a thousand other variables. Nothing is more commonplace than for a drug to be unsafe and ineffective for one person, and easily tolerated and highly beneficial for another. Different strokes for different folks is not just a euphonious phrase. It is the distillation of much wisdom on how to make collective decisions.
To be sure, when the FDA exercises its power to warn of various side effects from given treatment, it does not prevent the downstream sorting that results in drugs being taken chiefly by the individuals who will benefit from them. But when the FDA exercises its power to ban a new drug from the market, its decisions are no longer subject to downstream correction by informed users. The drug is kept from those who benefit from it most. These costs, moreover, only increase when an established drug is pulled from the market by FDA order or under FDA pressure. After all, the people who are forced to give up the drug are those for whom it has been shown to work. One need only read the heart-rending testimony before the FDA by doctors who had prescribed Rezulin with success before it was pulled: "You can't buy this drug back from these patients," Dr. Steven V. Edelman testified before the FDA in March, 1999. No surprise here.
To make matters worse, the FDA uses the wrong standard in deciding which drugs to let on the market and which to pull off. As the Iressa users note, the FDA standards don't take into account the wide variation among potential users. It only asks whether the average person who takes the new drug is better off than the average person who takes a placebo. But so what? The real question is whether there is any subgroup of potential users who are better off with the new drug than they were with the old. If so, then it should be let onto the market. The FDA uses a standard that is too restrictive and thus exercises a regulatory paternalism at the wrong level of the process. Physicians and family members should help patients decide which drugs should be taken. The FDA should limit its power to ban to those products that promise uniform harms—that is, to those products that are never marketed in the first place.
Richard A. Epstein is a professor at the University of Chicago, and a senior fellow at the Hoover Institution. His new book: "Overdose: How Excessive Regulation Stifles Pharmaceutical Innovation" will be published this coming fall by Yale University Press. He has also worked as a consultant for PhRMA and Pfizer.
1. Janet Woodcock, FDA to Cancer Patients: We're in Your Corner, Wall St. J., July 20, 2005, at A13
2. See Steven Walker, Iressa: The Reality vs. the FDA's Version, WALL ST. J., July 26, 2005, at A25.